First-Principles Evidence for Strongly Correlated Superconductivity Driven by Structural Variations in La$_3$Ni$_2$O$_7$
Extracted source text · arXiv 2502.19501
\documentclass[reprint, superscriptaddress, amsmath, amssymb, aps, physrev]{revtex4-2}
\usepackage{graphicx}
\usepackage{xcolor}
\usepackage[percent]{overpic}
\usepackage{subfigure}
\usepackage{braket}
\usepackage{tikz}
\usepackage{hyperref}
\begin{document}
\title{Evidence for Strongly Correlated Superconductivity in \texorpdfstring{La$_3$Ni$_2$O$_7$}{La3Ni2O7} from First Principles}
\author{Daan Verraes}
\affiliation{Center for Molecular Modeling, Ghent University, Technologiepark 46, 9052 Zwijnaarde, Belgium}
\affiliation{Department of Physics and Astronomy, Ghent University, Krijgslaan 281, 9000 Ghent, Belgium}
\author{Tom Braeckevelt}
\affiliation{Center for Molecular Modeling, Ghent University, Technologiepark 46, 9052 Zwijnaarde, Belgium}
\author{Nick Bultinck}
\email{Contact author: Nick.Bultinck@UGent.be}
\affiliation{Department of Physics and Astronomy, Ghent University, Krijgslaan 281, 9000 Ghent, Belgium}
\author{Veronique Van Speybroeck}
\email{Contact author: Veronique.VanSpeybroeck@UGent.be}
\affiliation{Center for Molecular Modeling, Ghent University, Technologiepark 46, 9052 Zwijnaarde, Belgium}
\begin{abstract}
We conduct first-principles simulations of \texorpdfstring{La$_3$Ni$_2$O$_7$}{La3Ni2O7}, a nickelate in which a recent series of experiments found signs of high-temperature superconductivity. In the hydrostatic pressure range where superconductivity is observed, we find a significant increase in effective on-site repulsion in the maximally localized Wannier functions comprising the Ni 3$d$ $e_g$ bands crossing the Fermi energy, as determined by density functional theory. We attribute this increase to reduced screening by the spacer La $5d_{x^2-y^2}$ bands -- an effect that is sensitive to the pressure-driven crystal structure. Our results therefore indicate that the superconducting region in the \texorpdfstring{La$_3$Ni$_2$O$_7$}{La3Ni2O7} phase diagram coincides with a region of enhanced electron correlations. \textit{Ab initio} molecular dynamics simulations extend these trends to finite temperatures, providing insights into the experimentally observed transition lines. Finally, we leverage intrinsic chemical pressure by applying the methodology to \texorpdfstring{Ac$_3$Ni$_2$O$_7$}{Ac3Ni2O7}.
\end{abstract}
\maketitle
\section*{Introduction} \label{sec: Introduction}
\input{Sections/Introduction}
\section*{Results} \label{sec: Results}
\input{Sections/Results}
\section*{Acknowledgments}
This work is supported by the Research Board of Ghent University (BOF) through a Concerted Research Action (BOF23/GOA/021). V. V. S. acknowledges the Research Board of Ghent University (BOF). The computational resources (Stevin Supercomputer Infrastructure) and services used in this work were provided by VSC (Flemish Supercomputer Center), funded by Ghent University, FWO, and the Flemish Government—department EWI. N. B. was supported by the European Research Council under the European Union Horizon 2020 Research and Innovation Programme via Grant Agreement No. 101076597-SIESS, and by a grant from the Simons Foundation (SFI-MPS-NFS-00006741-04).
\section*{Supporting Information Available}
Computational data supporting the results of this work are available from the online GitHub repository at https://github.com/daanverraes/Supporting-Information or upon request from the authors. A more elaborate discussion of the results can be found in the Supplementary Information.
\section*{Author contributions}
D. V., T. B., N. B. and V. V. S. initiated the discussion, designed the paper, and were involved in the discussion of the results. D. V. and N. B. wrote the manuscript with contributions of all authors. D. V. performed all simulations.
\section*{Competing Interests}
The authors declare no competing interests.
\section*{Methods} \label{sec: Methodology}
\input{Sections/Methodology}
\bibliographystyle{naturemag}
\bibliography{ref}
\clearpage
\input{Sections/Supplementary}
\end{document}
\typeout{get arXiv to do 4 passes: Label(s) may have changed. Rerun}
Nearly four decades after the discovery of high-transition-temperature (high-$T_\text{c}$) superconductivity in hole-doped cuprates, a recent breakthrough occurred with the observation of superconductivity up to $80$ K in bulk samples of the nickelate La$_3$Ni$_2$O$_7$ under hydrostatic pressures above 14 GPa \cite{bednorz_possible_1986, sun_signatures_2023, zhang_high-temperature_2024}. Subsequent transport measurements suggest a right-triangular superconducting region in the pressure-temperature (PT) phase diagram between 14 and 80 GPa, with a maximum $T_\text{c}$ at 18 GPa \cite{li_pressure-driven_2024}. The structural and electronic differences with the cuprates have thereby motivated a plethora of theoretical works to unmask the pairing mechanism at play in La$_3$Ni$_2$O$_7$ \cite{zhang_electronic_2023, luo_bilayer_2023, gu_effective_2023, luo_high-tc_2024, yang_possible_2023, lechermann_electronic_2023, shen_effective_2023, sakakibara_possible_2024, lu_interlayer_2024, liao_electron_2023, qu_bilayer_2024, yang_interlayer_2023, wu_superexchange_2024, huang_impurity_2023, jiang_high-temperature_2024, lu_superconductivity_2023, oh_type_2023, zhang_structural_2024,kaneko_pair_2024, YangHui_2024, YangHuiOh_2024, shilenko_correlated_2023, tian_correlation_2024, liu_-wave_2023, cao_flat_2024, qin_high-T_c_2023, chen_charge_2024, jiang_pressure_2024, christiansson_correlated_2023, zhang_trends_2023, yi_antiferromagnetic_2024, chen_orbital-selective_2024, jiang_theory_2025}.
As a member of the Ruddlesden-Popper nickelates, the conventional polymorph of La$_3$Ni$_2$O$_7$ features intact bilayers of NiO$_6$ octahedra (Fig.\ \ref{fig: introduction}(a)), with a formal nickel oxidation of Ni$^{2.5+}$ ($3d^{7.5}$). This hole concentration indicates a multi-orbital low-energy configuration, which differs significantly from the Ni$^{+}$ ($3d^{9}$) cations forming the square-planar NiO$_2$ layers in infinite-layer nickelates, or from the Cu$^{2+}$ ($3d^{9}$) cations of the CuO$_2$ planes in cuprates \cite{li_superconductivity_2019, bednorz_possible_1986}. Density functional theory (DFT) calculations indicate that the electron density near the Fermi energy ($E_\text{F}$) is comprised mostly of the Jahn-Teller distorted Ni $e_g$ orbitals, which exhibit hybridization with adjacent O $2p$ orbitals along their symmetry-aligned axes \cite{luo_bilayer_2023, gu_effective_2023, zhang_electronic_2023}. These studies suggest that the active Ni $d_{z^2}$ orbitals form strong $\sigma$-bonds with the apical O $p_z$ orbitals connecting the bilayers, giving rise to a large interlayer hopping which splits the $d_{z^2}$ orbitals into bonding and antibonding states, while the $d_{x^2-y^2}$ orbitals remain approximately decoupled between monolayers. This orbital-selective (2Ni)$^{5+}$ configuration, shown in Fig.\ \ref{fig: introduction}(b), suggests the construction of an effective dimer model comprising the four $e_g$ orbitals (the $d_{z^2}$ and $d_{x^2-y^2}$ orbitals in each layer) for which many bilayer Hubbard and \textit{t-J} models have been proposed and studied in literature \cite{luo_bilayer_2023, luo_high-tc_2024, gu_effective_2023, yang_possible_2023, lechermann_electronic_2023, shen_effective_2023, sakakibara_possible_2024, lu_interlayer_2024, liao_electron_2023, qu_bilayer_2024, yang_interlayer_2023, wu_superexchange_2024, huang_impurity_2023, jiang_high-temperature_2024, lu_superconductivity_2023, oh_type_2023, zhang_structural_2024, kaneko_pair_2024,YangHui_2024, YangHuiOh_2024}.
\begin{figure}[t]
\centering
\begin{overpic}[width=\linewidth]{Figures/introduction.png}
% subfigure labels:
\put(2,84){(a)}
\put(45,84){(b)}
\put(54,42){(c)}
% atom types:
\put(45,17){O}
\put(45,11.5){Ni}
\put(45,6){La}
% compass a:
\put(3,11.5){\textbf{c}}
\put(0.5,0){\textbf{a}}
\put(16,0){\textbf{b}}
% compass b:
\put(53,11){\textbf{b}}
\put(66,0){\textbf{a}}
% electronic configuration:
\put(71.5,79){$e_g$}
\put(71.5,59){$t_{2g}$}
\put(90,82){$d_{z^2,-}$}
\put(90,76){$d_{x^2-y^2}$}
\put(90,68.5){$d_{z^2,+}$}
\put(90,58.5){$d_{xy}$}
\put(83,48){$d_{xz}, d_{yz}$}
\end{overpic}
\caption{\textbf{Crystal structure and low-energy configuration of bilayer La$_3$Ni$_2$O$_7$.}
(a) Orthorhombic unit cell of the $Cmcm$ phase, characterized by tilted NiO$_6$ octahedra, shown in gray.
(b) Crystal field splitting and orbital-selective $e_g$ configuration (blue) of two Ni$^{2.5+}$ cations in a dimer. The $d_{z^2}$ orbitals form bonding $d_{z^2,+}$ and anti-bonding $d_{z^2,-}$ states.
(c) Top view of a NiO$_2$ bilayer. Solid and dashed black frames represent the $Fmmm$ and $Cmcm$ unit cells, respectively.
}
\label{fig: introduction}
\end{figure}
On the experimental side, the crystal structure of La$_3$Ni$_2$O$_7$ at superconducting conditions is not yet completely resolved due to resolution constraints in the reported X-ray diffraction (XRD) data. At room temperature, the bilayer polymorph exhibits a first-order structural transition between two orthorhombic ($a \neq b$) phases, symmetrizing from space group $Cmcm$ to subgroup $Fmmm$ above a hydrostatic pressure of 14 GPa \cite{sun_signatures_2023}. A key structural change at the transition is the $c$-axis alignment of NiO$_6$ octahedra, depicted in Fig.\ \ref{fig: introduction}(c). This octahedral alignment coincides with a metallization of the $\sigma$-bonding states, which has been suggested to stabilize bilayer superconductivity at low temperatures \cite{sakakibara_orbital_2014}. Subsequent \textit{in situ} synchrotron XRD identified the high-pressure phase at a much lower temperature of $\sim 40$ K to be a third, tetragonal ($a=b$) phase with space group $I4/mmm$ \cite{wang_structure_2024}, in agreement with DFT predictions at 0 K \cite{geisler_structural_2024}. However, the precise boundaries between these structural phases in the complete pressure-temperature phase diagram remain to be located.
In this work, we conduct first-principles calculations of La$_3$Ni$_2$O$_7$ to link the right-triangular superconducting region in the electronic phase diagram to structural variations and their effect on the low-energy dimer ($e_g$) subspace. We perform full-cell enthalpic optimizations to study the structural stability at zero temperature under varying hydrostatic pressure. Next, we examine how the pressure-driven crystal structure affects the electronic band structure and, in particular, the low-energy physics. To arrive at a quantitative statement, we derive the effective interactions in the dimer model by downfolding the optimized electronic structures using the constrained random phase approximation (cRPA) \cite{aryasetiawan_frequency-dependent_2004}. Our results furthermore provide a starting point for future studies exploring the pairing mechanism in bilayer La$_3$Ni$_2$O$_7$ via low-energy models constructed from first principles. Subsequent \textit{ab initio} molecular dynamics (AIMD) simulations allow us to explicitly include finite temperatures, offering insights into the structural phase diagram and the right-triangular shape of the superconducting dome. Finally, we investigate the effect of replacing all La with Ac, which provides a higher intrinsic chemical pressure.
\subsection{Structural Optimization} \label{subsec: structural optimization}
Fig.\ \ref{fig: bilayer geometries}(a) shows the optimized lattice parameters of the orthorhombic unit cell of La$_3$Ni$_2$O$_7$, defined in Fig.\ \ref{fig: introduction}(a), at hydrostatic pressures ranging from 0 to 100 GPa. We focus on the bilayer structure, as the experimentally observed alternating monolayer-trilayer polymorph exhibits a higher enthalpy everywhere within the considered pressure range (as shown in the Supplementary Information).
\begin{figure}[ht]
\centering
\begin{overpic}[width=\linewidth]{Figures/geometries.png}
\put(32, 36){\includegraphics[height=0.24\linewidth]{Figures/angles.png}}
\put(16, 93){(a)}
\put(16, 55){(b)}
\end{overpic}
\caption{\textbf{Enthalpic optimization of bilayer La$_3$Ni$_2$O$_7$.}
(a) Lattice parameters of the orthorhombic unit cell at hydrostatic pressures between 0 and 100 GPa.
(b) The three inequivalent Ni-O-Ni bond angles of the $Cmcm$ phase. The two intralayer angles $\theta_2$ and $\theta_3$ in the $Cmcm$ phase (gray) become equal in the $I4/mmm$ phase (above 10 GPa), and coincidentally the NiO$_6$ octahedra align along the $c$-axis ($\theta_1=180^\circ$).}
\label{fig: bilayer geometries}
\end{figure}
Two distinct regimes in both degeneracy and slopes can be observed. In the low-pressure region up to $\sim 10$ GPa, the in-plane lattice constants $a$ and $b$ as defined in Fig.\ \ref{fig: introduction}(c), are non-degenerate. At higher external pressures, these become equal, thereby eliminating the Jahn-Teller distortion in the $ab$-plane. This explicitly confirms the absence of a zero-temperature, high-pressure orthorhombic \textit{Fmmm} phase where $a\neq b$. Fig.\ \ref{fig: bilayer geometries}(b) shows the three inequivalent Ni-O-Ni bond angles in the $Cmcm$ phase. The interlayer (apical) bond angle $\theta_1$ becomes equal to $180^\circ$ at the transition to the tetragonal phase around 10 GPa, indicating the alignment of the NiO$_6$ octahedra along the $c$-axis. Furthermore, the two inequivalent intralayer (basal) bond angles in the low-pressure structures, $\theta_2$ and $\theta_3$, become equal and both face inwards (\textit{i.e.} into the bilayer), which restores periodicity in the $ab$-plane, as depicted in Fig.\ \ref{fig: introduction}(c). The combined effect of symmetry restoration in the $ab$-plane and the alignment of the NiO$_6$ octahedra reproduces the structural transition from the low-pressure phase with space group $Cmcm$ to the high-pressure tetragonal phase with subgroup $I4/mmm$, consistent with previous work \cite{geisler_structural_2024}. The discrepancy in transition pressure between our calculations and experiment may reflect limitations of the DFT approximation. We attribute the experimental stabilization of a zero-temperature intermediate $Fmmm$ phase to unintended anisotropic stress, highlighting the need for a detailed first-principles analysis of the full external stress tensor.
In the high-pressure limit the cell contracts isotropically, as can be seen from the parallel curves corresponding to the three lattice constants in Fig.\ \ref{fig: bilayer geometries}(a). However, the crystal structure in the $I4/mmm$ phase compresses anisotropically as the intralayer angle $\theta_2$ in Fig.\ \ref{fig: bilayer geometries}(b) reaches its maximum value of $\sim 177.4^\circ$ near the structural transition and then decreases monotonically with increasing pressure. This buckling effect can be explained by the La ions from the neighboring rock-salt (spacer) layers, closing in on the bilayers along the $c$-axis as the pressure increases. In the next section we examine the impact of the structural phase transition and this high-pressure compression on the low-energy electronic structure.
\subsection{Electronic Structure and Dimer Model} \label{subsec: electronic structure}
To understand the origin of superconductivity, it is crucial to develop an effective Hamiltonian that adequately describes the electrons near the Fermi surface. Fig.\ \ref{fig: electronic}(b) presents the full electronic band structure of La$_3$Ni$_2$O$_7$ in the $I4/mmm$ phase at 40 GPa, referenced to $E_\text{F}$ and projected onto the Ni $e_g$ and spacer La $5d_{x^2-y^2}$ orbitals. Additional $lm$-decompositions are provided in the Supplementary Information. The four bands closest to $E_\text{F}$ exhibit clear $e_g$ character, with the individual contributions from $d_{z^2}$ and $d_{x^2-y^2}$ varying significantly across the Brillouin zone. To capture the low-energy physics of these bands, we employ the multi-scale \textit{ab initio} scheme for correlated electrons, incorporating screening from high-energy states through many-body perturbation theory \cite{aryasetiawan_frequency-dependent_2004, imada_electronic_2010}. This yields a generalized Hubbard model defined on a bilayer square lattice with two orbitals per site, commonly referred to as the \emph{dimer model} (see Fig.\ \ref{fig: electronic}(f)). The negligible contribution from other spherical harmonics in Fig.\ \ref{fig: electronic}(b) allows us to disentangle the $e_g$ bands from higher-energy states using the projection method for Wannierization of entangled bands \cite{souza_maximally_2001}. The resulting maximally localized Wannier functions (MLWFs) accurately span the low-energy subspace, as confirmed by the Wannier-interpolated band structure in Fig.\ \ref{fig: electronic}(b).
\begin{figure*}[t]
\centering
\begin{overpic}[width=\linewidth]{Figures/electronic.png}
% subfigure labels:
\put(0, 54){(a)}
\put(0, 41){(b)}
\put(0, 20){(c)}
\put(39, 52){(d)}
\put(39, 24){(e)}
\put(83, 47){(f)}
\put(71, 24){(g)}
\end{overpic}
\caption{
\textbf{Pressure dependence of the band structure and dimer model of bilayer La$_3$Ni$_2$O$_7$.}
(a) Brillouin zone of the $I4/mmm$ primitive unit cell with high-symmetry points labeled.
(b) Band structure at 40 GPa, shown as solid lines, with the color scale indicating the overlap with the different Ni $e_g$ orbitals. The projection onto the spacer La $d_{x^2-y^2}$ is superimposed as purple circles, with radii representing the weight. The Wannier-interpolated bands obtained from the four MLWFs are shown as dashed lines.
(c) DFT band structures at pressures between 20 GPa and 100 GPa.
(d) Pressure dependence of the largest effective parameters in the dimer model shown in (f). The on-site interactions $U_{x^2-y^2}$, $U_{z^2}$ ($V_{x^2-y^2}$, $V_{z^2}$) are the screened (bare) Coulomb interactions for two electrons in MLWFs with $d_{x^2-y^2}$ and $d_{z^2}$ character. The screened (bare) interaction between different on-site MLWFs is $U_{12}$ ($V_{12}$). $\Delta = \epsilon (d_{x^2-y^2}) - \epsilon (d_{z^2})$ is the energy difference between orbitals, and the tight-binding parameters $t^x_{x^2-y^2}$ and $t^z_{z^2}$ represent nearest-neighbor intralayer hopping between $d_{x^2-y^2}$ orbitals and interlayer hopping between $d_{z^2}$ orbitals, respectively.
(e) Ratios of screened interaction strengths to their corresponding hopping parameters in the dimer model.
(g) Ratios of bare interaction strengths to their corresponding screened interaction in the dimer model.
}
\label{fig: electronic}
\end{figure*}
For the high-pressure $I4/mmm$ phase, the evolution of the DFT band structure with pressure is shown in Fig.\ \ref{fig: electronic}(c). We see that most (high-energy) bands tend to move away from $E_\text{F}$ under increasing pressure, while the dimer $e_g$ and spacer La 5$d_{x^2-y^2}$ bands become more dispersive as reported in earlier work \cite{jiang_theory_2025}. Remarkably, near the $\Gamma$ and Z points, we observe that above 45 GPa, the lowest spacer La 5$d_{x^2 - y^2}$ conduction band moves below the $e_g$ bands. Above 75 GPa, it even drops below $E_\text{F}$, thereby forming an additional electron pocket. As we will show next, this lowering of the spacer La bands has a significant effect on the effective low-energy physics. For completeness, the pressure-dependence of the DFT band structure in the $Cmcm$ phase is shown in the Supplementary Information.
Localized model parameters are now obtained by downfolding the electronic structure onto the corresponding MLWFs. By doing this for every optimized crystal structure, the pressure dependence of the dimer model is given in Fig.\ \ref{fig: electronic}(d). We see that the on-site repulsion energies, $U_{x^2-y^2}, U_{z^2}$ and $U_{12}$ in Fig.\ \ref{fig: electronic}(d), are significantly enhanced (up to $\sim 30\%$) with the structural phase transition near 10 GPa, and gradually decrease with cell compression at higher pressures. In contrast, the kinetic terms, $\Delta$, $t^x_{x^2-y^2}$ and $t^z_{z^2}$, increase monotonically and at a slower rate. As a measure of correlation strength in the dimer model, various $U_i/t_i$ ratios are given in Fig.\ \ref{fig: electronic}(e). These interaction-to-hopping ratios are abruptly enhanced near the structural phase transition, peaking around 14 GPa before gradually decreasing. Interestingly, the parameters near the structural transition at 10 GPa are nearly equivalent -- up to a global scale -- to those around 40–45 GPa. From the perspective of this effective dimer model, these similar electronic conditions can be qualitatively associated with the critical pressures of 14 and 51 GPa, between which the strongest signs of zero resistance have been observed \cite{li_pressure-driven_2024}. While the DFT approximation and the simplified nature of the dimer model may shift the zero temperature critical pressures, the overall trend is robust: the smaller $U_i/t_i$ ratios in both the low-pressure $Cmcm$ phase and at extreme pressures clearly signal a weakening of electronic correlations.
We can understand the pressure-dependence of the dimer model parameters as follows. In the low-pressure limit, the $Cmcm$ phase hosts a checkerboard pattern with two distinct tilted NiO$_6$ octahedra (see Fig. \ref{fig: introduction}(c)). Consequently, the nearest-neighbor hopping integrals are weakened due to the tilted bond angles which reduce orbital overlap. The reduced symmetry also separates the intralayer hopping into two components: one corresponding to the inward-facing and one for the outward-facing angles. However, the difference between these parameters is minimal, remaining below $2\%$. The change in the effective interaction $U_i$ can be understood from the screening by high-energy electrons, quantified by various bare to screened interaction ratios, $V_i/U_i$, in Fig.\ \ref{fig: electronic}(g). The enhanced screening at lower pressures in the $Cmcm$ phase can be attributed the spacer La 5$d_{x^2-y^2}$ orbitals, which overlap more strongly with the neighboring $e_g$-like MLWFs due to the tilted octahedral configuration. However, the decrease in $U_i$ (at lower pressure) cannot be entirely assigned to screening effects, as we find that the bare interactions $V_i$ also decrease in the $Cmcm$ phase (see Fig.\ \ref{fig: electronic}(d)). This can be explained by the same octahedral tilts, which lead to a reduced atomic character and associated delocalization (increased spreads) of the MLWFs in the $Cmcm$ phase.
In the high-pressure limit, the compression of the crystal structure in the $I4/mmm$ phase (Fig.\ \ref{fig: bilayer geometries}) enhances the overlap between nearby MLWFs, which monotonically increases the hopping integrals in Fig.\ \ref{fig: electronic}(d). The compression also decreases the MLWF spreads, which leads to increasing values of $\Delta$ (the energy difference between $d_{z^2}$ and $d_{x^2-y^2}$ orbitals). At the same time, the screening in Fig.\ \ref{fig: electronic}(g), increases with high pressures, and becomes dominant above 40 GPa, after which the effective interaction parameters start to decrease. We attribute the enhanced screening at high pressures to the spacer La cations, whose 5$d_{x^2-y^2}$ conduction bands energetically enclose the low-energy bands near the $\Gamma$ and Z points (Fig.\ \ref{fig: electronic}(c)), and which spatially enclose the $e_g$ MLWFs, as indicated by the decreasing intralayer angle $\theta_2$ in Fig.\ \ref{fig: bilayer geometries}(b).
\subsection{Molecular Dynamics} \label{subsec: aimd}
Next we include finite temperature (10–100 K) effects through AIMD simulations, extending pressure-driven structural and electronic trends along the temperature axis. The top panel in Fig.\ \ref{fig: AIMD}(a) shows the time-averaged absolute difference between the in-plane lattice constants $a$ and $b$. This is a measure for the finite-temperature fluctuations in $a-b$ away from zero. We see that $\langle |a - b| \rangle$ is on the order of $\sim$0.1 Å at low temperatures and low pressures, corresponding to the orthorhombic $Cmcm$ phase. Upon increasing pressure $\langle |a-b|\rangle$ quickly drops to $\sim$ 0.01 Å in the tetragonal $I4/mmm$ phase. The region of rapid decrease in $\langle |a-b|\rangle$ extends almost vertically in the PT-phase diagram, with a slope of approximately -75 K/GPa, up to the maximum temperature of $100$ K used in this work. Fig.\ \ref{fig: AIMD}(a) furthermore implies that the temperature-driven transition from the high-pressure $I4/mmm$ tetragonal phase found in DFT to the orthorhombic $Fmmm$ phase observed in XRD at room temperature must occur at temperatures above $100$ K, placing this transition at a comfortable distance from the superconducting region. This can be further confirmed by looking at $a-b$ histograms, as shown for six PT-points in the bottom panel of Fig.\ \ref{fig: AIMD}(a). In the region of large $\langle |a-b|\rangle$ at low pressure (exemplified at 0 GPa), we see that the $a-b$ histograms are centered around $a-b \approx$ 0.1 Å, both at $T=10$ K and $T=50$ K. For higher pressures, on the other hand, the $a-b$ histograms are clearly centered around $0$ at both $T=10$ K and $T=50$ K.
\begin{figure*}[t]
\centering
\begin{overpic}[width=\linewidth]{Figures/AIMD_together.png}
% subfigure labels:
\put(16, 0){(a)}
\put(50, 0){(b)}
\put(84, 0){(c)}
\end{overpic}
\caption{\textbf{AIMD simulations of bilayer La$_3$Ni$_2$O$_7$ in the range of 10–100 K and 0–100 GPa.}
The upper row shows the time-averages of the full diagram, while the lower row displays frequency histograms at six indicated pressure-temperature points.
Panels show: (a) the difference between lattice constants $a$ and $b$, (b) the interlayer angle $\theta_1$, and (c) the inwards intralayer angle $\theta_2$.
}
\label{fig: AIMD}
\end{figure*}
The time-averaged interlayer angle $\theta_1$ is shown in the top panel of Fig.\ \ref{fig: AIMD}(b). Note that $\theta_1$ is defined as the internal bond angle, with $\theta_1 = 180^\circ$ representing a fully straightened bond and $\langle \theta_1\rangle$ a measure for the fluctuations away from $\theta_1=180^\circ$. The reduced bin heights near $\theta_1=180^\circ$ reflects the unscaled space angle, resulting from azimuthal freedom. We see from Fig.\ \ref{fig: AIMD}(b) that $\langle \theta_1\rangle$ exhibits two distinct trends. First, approximately independent of temperature, $\langle\theta_1\rangle$ quickly increases from $\sim 165^\circ$ to $\sim 180^\circ$ around a pressure of $10$ GPa. This we attribute to the octahedral tilt which disappears upon going from $Cmcm$ to $I4/mmm$, which can be verified explicitly from the histograms in the lower panel of Fig.\ \ref{fig: AIMD}(b): the low-pressure histograms are centered around $\sim 165^\circ$ while the high-pressure histograms are centered around $\sim 180^\circ$. The verticality of the region where $\langle \theta_1\rangle$ quickly rises agrees with the structural transition discussed above. This result indicates that the experimentally observed superconducting region is contained entirely within the $I4/mmm$ phase. Secondly, inside the $I4/mmm$ phase, increasing the temperature reduces $\langle \theta_1\rangle$ due to enhanced thermal fluctuations. At pressures right after the structural transition ($14-20$ GPa), we see that increasing the temperature significantly affects the interlayer angle. Higher hydrostatic pressure, in turn, enhances the robustness of octahedral alignment against thermal fluctuations.
Also $\langle \theta_2\rangle$, the average of the intralayer Ni-O-Ni bond angle, exhibits a narrow region of rapid increase at the $Cmcm$ -- $I4/mmm$ phase boundary, as shown in the top panel of Fig.\ \ref{fig: AIMD}(c). At higher pressures, we see that crystal compression leads to a decreasing value of $\langle \theta_2\rangle$, due to the La ions approaching the NiO layers along the $c$-axis as observed at 0 K in Fig.\ \ref{fig: bilayer geometries}(b). Increasing temperature also reduces $\langle \theta_2\rangle$. From the histograms in the bottom panel of Fig.\ \ref{fig: AIMD}(c) we see that $\theta_2$ is sufficiently far away from $180^\circ$ to be able to conclude that the decrease in $\langle \theta_2\rangle$ is not a fluctuation effect, but indicates a true shift in average value of the intralayer bond angle.
These AIMD results, together with the discussion in Subsection \ref{subsec: electronic structure}, suggest following picture underlying the experimentally observed right-triangular superconducting region in the PT-phase diagram \cite{li_pressure-driven_2024}. First, the rapid increase of the superconducting transition temperature $T_\text{c}$ at the critical pressure can be related to the almost vertical $Cmcm$ -- $I4/mmm$ phase boundary where the octahedral tilt disappears. Secondly, for pressures larger than $\sim18$ GPa, $T_\text{c}$ is observed to start decreasing, which coincides with the region of decreasing $\theta_2$. Recall that latter can be taken as a proxy for the spatial overlap between the La states and the MLWFs of the $e_g$ bands: smaller $\theta_2$ means larger overlap and hence more screening. Besides thermal fluctuations of the superconducting condensate itself, another detrimental finite-temperature effect for superconductivity could therefore be that the intralayer angle decreases upon raising the temperature, giving rise to increased screening at higher temperature. This effect could contribute to the decrease of $T_\text{c}$ with pressure.
\subsection{Chemical Substitution} \label{subsec: chemical substitution}
Our analysis of bilayer La$_3$Ni$_2$O$_7$ can now be used as a guiding principle for the design of bulk nickelates superconducting near ambient pressure. The most straightforward adaptation is an isoelectronic substitution of the rare-earth $A$-site cation La$^{3+}$, which plays a pivotal role in the structural phase transition and corresponding critical pressure. Doping with smaller atoms was initially expected to stabilize the tetragonal phase at lower pressures via chemical precompression. However, DFT simulations on $A_3$Ni$_2$O$_7$ showed the opposite: smaller atoms increased the transition pressure from the orthorhombic to the tetragonal structure \cite{geisler_structural_2024}. Consequently, larger substituents were proposed to stabilize the $I4/mmm$ structure by the inherent chemical pressure, replacing external hydrostatic pressure. Among (valence) isoelectronic elements, only Ac$^{3+}$ possesses a larger ionic radius, for which Ac$_3$Ni$_2$O$_7$ was indeed found to be structurally stable in the $I4/mmm$ phase at ambient pressure, as confirmed by DFT+$U$ calculations \cite{rhodes_structural_2024, wu_ac_3ni_2o_7_2024}.
\begin{figure}[ht]
\centering
\begin{overpic}[width=\linewidth]{Figures/ANO.png}
\put(42, 35){\includegraphics[height=0.24\linewidth]{Figures/angles.png}}
% subfigure labels:
\put(17, 93){(a)}
\put(17, 58){(b)}
\end{overpic}
\caption{\textbf{Enthalpic optimization of bilayer Ac$_3$Ni$_2$O$_7$.}
(a) Lattice constants of the orthorhombic unit cell (as shown for La$_3$Ni$_2$O$_7$ in Fig.\ \ref{fig: introduction}(a)) at hydrostatic pressures between 0 and 10 GPa.
(b) The three unique Ni-O-Ni bond angles of the $Cmcm$ phase.
}
\label{fig: ANO geometries}
\end{figure}
Enthalpic optimization of bilayer Ac$_3$Ni$_2$O$_7$ at external pressures between 0 and 10 GPa, as shown in Fig.\ \ref{fig: ANO geometries}, confirms the absence of Jahn-Teller distortion in the $ab$-plane, which is likely due to the increased chemical pressure of the larger Ac$^{3+}$ cations. However, we find that the intralayer spatial symmetry is still (slightly) broken at pressures up to 3.2 GPa. The two inequivalent intralayer angles $\theta_2$ and $\theta_3$ become equal with NiO$_6$ octahedral alignment in the $I4/mmm$ phase above 3.2 GPa. The presence of a low-pressure $Cmcm$ phase could arise from limitations in the PBE-D3 level, as opposed to the DFT+$U$ methodology used in Ref. \cite{rhodes_structural_2024}, which takes more electronic correlations into account and found an $I4/mmm$ phase already at $0$ GPa. Nonetheless, the significantly lower critical pressure found in this work, resulting from the enlarged ionic radius, confirms the important role of chemical pressure exerted by the $A$-site cations. The intralayer angle $\theta_2 = \theta_3$ in the $I4/mmm$ phase furthermore decreases faster than in La$_3$Ni$_2$O$_7$, which can also be ascribed to the enhanced chemical pressure of the larger Ac$^{3+}$ cations.
\begin{table}[t]
\caption{\textbf{Dimer models for La$_3$Ni$_2$O$_7$ and Ac$_3$Ni$_2$O$_7$} Comparison of the dimer models for bilayer La$_3$Ni$_2$O$_7$ at 20 GPa and bilayer Ac$_3$Ni$_2$O$_7$ at 4 GPa.}
\begin{ruledtabular}
\begin{tabular}{ccc}
\multicolumn{1}{c}{\textrm{Parameter [eV]}}&
\textrm{La$_3$Ni$_2$O$_7$}&
\textrm{Ac$_3$Ni$_2$O$_7$}\\
\colrule
$U_{x^2-y^2}$ & 3.4084 & 3.0284 \\
$U_{z^2}$ & 3.1518 & 2.9291 \\
$U_{12}$ & 2.0541 & 1.7342 \\
$t^x_{x^2-y^2}$ & 0.5298 & 0.4268 \\
$t^z_{z^2}$ & 0.6937 & 0.5797 \\
$\Delta$ & 0.3381 & 0.2930 \\
\end{tabular}
\label{tab: parameters}
\end{ruledtabular}
\end{table}
Apart from its crucial role in pressure-driven structural deformation, we have argued above that the La $5d$ bands are also important in providing significant screening effects, which we suggest reduce the superconducting $T_\text{c}$ at higher external pressures and could therefore be altered by this substitution. The band structures of La$_3$Ni$_2$O$_7$ and Ac$_3$Ni$_2$O$_7$ are known to be nearly identical \cite{rhodes_structural_2024}. The low-energy parameters of the dimer model are compared for La$_3$Ni$_2$O$_7$ at 20 GPa and Ac$_3$Ni$_2$O$_7$ at 4 GPa in Table \ref{tab: parameters}. The on-site screened Coulomb interactions in $4$ GPa Ac$_3$Ni$_2$O$_7$ are clearly smaller in value than in ideally pressurized La$_3$Ni$_2$O$_7$, which can be explained by the larger lattice constants and consequent decreased localization of the MLWFs. However, the hopping integrals are also smaller, which can be explained by the reduced MLWF overlap as a result of the larger Ni-Ni distances. Thus, up to an overall rescaling, the dimer models shown in Table \ref{tab: parameters} are very similar. This indeed hints at the possibility of ambient pressure superconductivity in Ac$_3$Ni$_2$O$_7$, and confirms the important role of the rare-earth element in determining the screened interaction parameters for the electrons in the $e_g$ bands. Even though Ac$_3$Ni$_2$O$_7$ is not a viable option for practical use because of actinium's intense radioactivity, our results nevertheless confirm the essential role of the $A$-site cations as discussed throughout this work.
\paragraph*{\textbf{DFT Calculations}}
The first-principles DFT calculations are performed using the Vienna Ab initio Simulation Package (VASP) employing the Perdew-Burke-Ernzerhof (PBE) exchange-correlation functional and the projector-augmented-wave formalism with a plane-wave energy cut-off of 600 eV \cite{kresse_ab_1993, kresse_efficient_1996, perdew_generalized_1996, blochl_projector_1994}. All calculations are performed in the non-magnetic (spin-unpolarized) case. The convergence criterion for the electronic self-consistent loops is set to 10$^{-8}$ eV. Brillouin-zone sampling is $\Gamma$-centered, with a density that depends on the unit cell (as detailed below), where convergence is systematically validated. \\
\paragraph*{\textbf{Structural Optimization}}
Structural optimizations are performed at fixed hydrostatic pressures by minimizing the enthalpy using the conjugate gradient algorithm with respect to all ionic and cell degrees of freedom until the change in total energy subceeds 10$^{-7}$ eV and all Hellmann-Feynman forces subceed 10 meV/\AA. Pairwise van der Waals dispersion corrections are included using the DFT-D3 method with Becke-Johnson (BJ) damping \cite{grimme_consistent_2010}. To avoid finite size effects, the charge densities are calculated on a 5$\times$5$\times$1 $k$-point grid in the largest unit cell of all experimentally known phases. Initial cells contain distorted Wyckoff positions from experimental data \cite{sun_signatures_2023}. \\
\paragraph*{\textbf{Downfolding}}
The single-point electronic structures are recalculated on a $\Gamma$-centered 5$\times$5$\times$9 and 9$\times$9$\times$9 $k$-point grid for the primitive unit cells in the $Cmcm$ and $I4/mmm$ phases, respectively. The 12 GPa structure is calculated in the $Cmcm$ phase to ensure no bias from unit cell choice. The resulting 336 and 168 lowest bands, respectively, are then downfolded into the $e_g$ dimer configuration using the multi-scale \textit{ab initio} scheme for correlated electrons developed by Imada \textit{et al.} \cite{imada_electronic_2010}. This scheme includes the (projector) cRPA and tight-binding methodology to obtain the screened Coulomb interactions and hopping integrals, respectively \cite{aryasetiawan_frequency-dependent_2004, projector_Hanke_Sham, kaltak_projector}. Static interactions were assumed, as frequency dependence from retarded screening was found to be negligible (see Supplementary Information). To construct a sufficiently localized basis spanning this low-energy subspace, MLWFs are constructed from projecting the Kohn-Sham orbitals on the $e_g$ orbitals \cite{marzari_maximally_localized_1997, pizzi_wannier90_2020, souza_maximally_2001}. The dispersion of the Wannier-interpolated bands is systematically compared to the DFT bands, confirming a faithful disentanglement. \\
\paragraph*{\textbf{Molecular Dynamics}}
The AIMD simulations are conducted on a 3$\times$3$\times$1 $k$-point grid for the orthorhombic unit cell (Fig.\ \ref{fig: introduction}(a)) in the isothermal-isobaric (NPT) ensemble, using the Nosé-Hoover thermostat and the Parrinello-Rahman barostat \cite{nose_unified_1984, shuichi_constant_1991, hoover_canonical_1985, parrinello_crystal_1980, parrinello_polymorphic_1981}. A barostat mass of 2000 amu was employed to maintain stable pressure control. Langevin damping coefficients of 10 ps$^{-1}$, 5 ps$^{-1}$, and 2 ps$^{-1}$ were applied to the La, Ni, and O atoms, respectively, while a lattice damping coefficient of 10 ps$^{-1}$ was used. A time step of 1 fs was adopted, with initial configurations equilibrated for 2 ps prior to data collection. All input files and post-processing scripts are available from the \href{https://github.com/daanverraes/Supporting-Information}{online GitHub repository}.
\onecolumngrid
\renewcommand{\thefigure}{S\arabic{figure}}
\renewcommand{\thesection}{S\arabic{section}}
\begin{center}
\textbf{\large Supplementary Information:}
\vspace{0.5 cm}
\textbf{\large Evidence for Strongly Correlated Superconductivity in \texorpdfstring{La$_3$Ni$_2$O$_7$}{La3Ni2O7} from First Principles}
\vspace{0.5 cm}
Daan Verraes$^{1,2}$, Tom Braeckevelt$^1$, and Nick Bultinck$^2$, Veronique Van Speybroeck$^1$
\vspace{0.2 cm}
\textit{$^1$Center for Molecular Modeling, Ghent University, Technologiepark 46, 9052 Zwijnaarde, Belgium} \\
\textit{$^2$Department of Physics and Astronomy, Ghent University, Krijgslaan 281, 9000 Ghent, Belgium}
\vspace{0.5 cm}
\end{center}
\section{Polymorphic Stability}
Experimental studies show a biphasic equilibrium between bilayer and monolayer-trilayer polymorphs of La$_3$Ni$_2$O$_7$ \cite{puphal_unconventional_2024, chen_polymorphism_2024, wang_long-range_2024}. To assess their stability, each structure was optimized separately, as the high energy barrier prevents interconversion. As shown in Fig.\ \ref{fig: enthalpic_difference}(b), the bilayer is consistently more stable across 0–100 GPa, with increasing enthalpy difference at higher pressures, extending results from \cite{zhang_electronic_2024, ouyang_absence_2024}.
\begin{figure}[ht]
\centering
\begin{overpic}[width=0.9\linewidth]{Figures/SI/Polymorphs_stability.png}
\put(16,0){(a)}
\put(52,0){(b)}
\put(86,0){(c)}
\end{overpic}
\caption{\textbf{Polymorphic stability in La$_3$Ni$_2$O$_7$ from first principles.}
Perovskite layer stacking in the (a) bilayer (BL) and (c) monolayer-trilayer (ML-TL) polymorph shown in their high-symmetry (tetragonal) phases.
(b) The enthalpic difference per structural formula, $H$(ML-TL)-$H$(BL), with respect to the applied hydrostatic pressure ranging from 0 to 100 GPa.}
\label{fig: enthalpic_difference}
\end{figure}
\section{Electronic Structure and Dimer Model}
\subsection{Band Structure}
In the low-pressure \textit{Cmcm} phase the primitive unit cell contains twice the number of structural formula units, as illustrated in Fig.\ \ref{fig: rainbow Cmcm}(a). The single-point electronic structure is calculated on a $5\times5\times9$ $k$-mesh. The resulting (pressure-dependent) band structures -- including at 12 GPa, which was found stable in the $I4/mmm$ subgroup -- in Fig.\ \ref{fig: rainbow Cmcm}(b) show band folding as a consequence of the broken spatial symmetry in the $ab$-plane.
\begin{figure}[ht]
\centering
\subfigure[]{
\begin{overpic}[height=0.32\linewidth]{Figures/SI/Primitive_Cmcm.png}
% atom types:
\put(8,88){O}
\put(8,82){Ni}
\put(8,76){La}
% compass:
\put(14,5){\textbf{a}}
\put(14,25){\textbf{b}}
\put(-1,11){\textbf{c}}
\end{overpic}
}
\hspace{0.03\linewidth}
\subfigure[]{
\includegraphics[height=0.3\linewidth]{Figures/SI/rainbow_cmcm.png}
}
\caption{\textbf{Primitive unit cell of La$_3$Ni$_2$O$_7$ in the \textit{Cmcm} phase and pressure-dependent band structures.}
(a) The crystal structure in the primitive unit cell of the \textit{Cmcm} phase containing one structural formula. The atom types are displayed as represented in Fig.\ \ref{fig: enthalpic_difference}.
(b) DFT band structures at pressures between 0 GPa and 12 GPa. The high-symmetry $k$-points in the Brillouin zone are defined as:
$\Gamma$ (0.0, 0.0, 0.0),
S (0.5, 0.0, 0.0),
R (0.5, 0.0, 0.5),
Z (0.0, 0.0, 0.5),
T (0.5, 0.5, 0.5),
Y (0.5, 0.5, 0.0).
}
\label{fig: rainbow Cmcm}
\end{figure}
In the high-pressure \textit{I4/mmm} phase, the primitive unit cell contains the structural formula once, as seen in Fig.\ \ref{fig: rainbow I4/mmm}(a). The single-point electronic structure is calculated on a $9\times9\times9$ mesh. The resulting (pressure-dependent) band structure, as discussed in the main text, is seen in Fig.\ \ref{fig: rainbow I4/mmm}(b).
\begin{figure}[ht]
\centering
\subfigure[]{
\begin{overpic}[height=0.32\linewidth]{Figures/SI/Primitive_I4mmm.png}
% atom types:
\put(8,86){O}
\put(8,80){Ni}
\put(8,74){La}
% compass:
\put(10,25){\textbf{a}}
\put(-2,25){\textbf{b}}
\put(4,0){\textbf{c}}
\end{overpic}
}
\hspace{0.03\linewidth}
\subfigure[]{
\includegraphics[height=0.3\linewidth]{Figures/SI/rainbow_I4mmm.png}
}
\caption{\textbf{Primitive unit cell of La$_3$Ni$_2$O$_7$ in the \textit{I4/mmm} phase and pressure-dependent band structures.}
(a) The crystal structure in the primitive unit cell of the \textit{I4/mmm} phase containing one structural formula. The atom types are displayed as represented in Fig.\ \ref{fig: enthalpic_difference}.
(b) DFT band structures at pressures between 20 GPa and 100 GPa. The high-symmetry $k$-points in the Brillouin zone are defined as:
$\Gamma$ (0.0, 0.0, 0.0),
X (0.0, 0.0, 0.5),
P (0.25, 0.25, 0.25),
N (0.0, 0.5, 0.0),
Z (0.5, 0.5, –0.5),
S (0.2591, 0.7409, –0.2591).
}
\label{fig: rainbow I4/mmm}
\end{figure}
The character of the bands near the Fermi energy ($E_\textrm{F}$) can be revealed by projecting the corresponding Kohn-Sham orbitals $\psi_{n\textbf{k}}$ onto different atomic orbitals centered on the different ion types (represented by spherical harmonics $Y^\alpha_{lm}$ that are non-zero within spheres around a given atom type $\alpha$). The obtained overlap $|\bra{Y^\alpha_{lm}}\ket{\psi_{n\textbf{k}}}|^2$ gives rise to the so-called $lm$-decomposition of the initial Kohn-Sham states as seen in the band structure at 40 GPa for different atom types $\alpha$ in Fig.\ \ref{fig: projections atoms}.
\begin{figure}[ht]
\centering
\noindent\makebox[\textwidth]{
\includegraphics[width=\textwidth]{Figures/SI/projection_atoms.png}}
\caption{\textbf{The $lm$-decomposition of bilayer La$_3$Ni$_2$O$_7$ at 40 GPa on the different atom types.} The $lm$-decomposition of the band structure on the different atom types for the range $[-3 \text{ eV}, 4 \text{ eV}]$ relative to the Fermi energy $E_\textrm{F}$.
The color of the band points corresponds to the absolute overlaps $|\bra{Y^\alpha_{lm}}\ket{\psi_{n\textbf{k}}}|^2$.
In the right figure the contributions of all three (equivalent) oxygen atoms in the unit cell are combined.}
\label{fig: projections atoms}
\end{figure}
The $lm$-decomposition onto all the atomic orbitals for the target bands around $E_\textrm{F}$ shows by far the strongest character centered on the Ni ions in the middle panel of Fig.\ \ref{fig: projections atoms} and some weak character for the surrounding oxygen centers. The overlap can be divided even more onto the different atomic orbitals labeled by the quantum numbers $l$ and $m$ as seen in Fig.\ \ref{fig: projection nickel}.
\begin{figure}[ht]
\centering
\noindent\makebox[\textwidth]{
\includegraphics[width=\textwidth]{Figures/SI/projection_nickel.png}}
\caption{\textbf{The $lm$-decomposition of bilayer La$_3$Ni$_2$O$_7$ at 40 GPa on the different atomic orbitals.} The $lm$-decomposition of the band structure on (a) the $d_{x^2-y^2}$ orbitals on the La ions, (b) the $t_{2g}$ orbitals of Ni and (c) the $e_{g}$ orbitals of Ni for the range $[-3 \text{ eV}, 4 \text{ eV}]$ relative to the Fermi energy $E_\textrm{F}$.
The color of the band points corresponds to the absolute overlaps $|\bra{Y^\alpha_{lm}}\ket{\psi_{n\textbf{k}}}|^2$.
}
\label{fig: projection nickel}
\end{figure}
The $lm$-decomposition onto different groups of atomic orbitals as seen in Fig.\ \ref{fig: projection nickel}(b-c) reproduces the crystal field splitting of the $3d$-orbitals into $t_{2g}$ and $e_g$ orbitals under the octahedral geometry. This confirms the presumption of a correlated subspace spanned by the $e_g$ bands and allows, through Wannierization, for a quasi-two-dimensional local, effective Hamiltonian as discussed in the main text.
\subsection{Wannierization}
To construct a sufficiently localized basis functions spanning the low-energy subspace, maximally localized Wannier functions (MLWFs) are constructed from the Kohn-Sham orbitals representing the $e_g$ orbitals near $E_\text{F}$. Wannier functions are defined as \cite{marzari_maximally_localized_1997}
\begin{align}
\phi_{\alpha\mathbf{R}}(\mathbf{r}) = \frac{V}{(2\pi)^3}\int e^{-i\mathbf{k}\cdot\mathbf{R}}\psi^{(w)}_{\alpha\mathbf{k}}(\mathbf{r})d\mathbf{k} \, , \label{eq:wann}
\end{align}
where the integral is over the first Brillouin zone, $V$ is the real-space primitive cell volume, $\mathbf{R}$ is a real-space lattice vector, and
\begin{align}
\psi^{(w)}_{\alpha\mathbf{k}}(\mathbf{r}) = \sum_\beta T_{\alpha\beta}^{(\mathbf{k})}\psi_{\beta\mathbf{k}}(\mathbf{r})
\end{align}
is a linear combination of Bloch functions $\psi_{\beta\mathbf{k}}(\mathbf{r})$, which are approximated by the Kohn-Sham orbitals. For a discretized Brillouin zone, \textit{i.e.} a set of $N_k$ $k$-points, as is the case in the reciprocal sampling approach, \eqref{eq:wann} becomes
\begin{align}
\phi_{\alpha\mathbf{R}}(\mathbf{r}) = \frac{1}{N_k}\sum_{\beta\mathbf{k}} e^{-i\mathbf{k}\cdot\mathbf{R}}T_{\alpha\beta}^{(\mathbf{k})}\psi_{\beta\mathbf{k}}(\mathbf{r}) .
\end{align}
The expansion coefficients, $T_{\alpha\beta}^{(\mathbf{k})}$, are uniquely obtained by imposing that the spread,
\begin{align} \label{eq: MLWF spread}
\Omega = \sum_\alpha \left(\bra{\phi_{\alpha\mathbf{0}}}r^2\ket{\phi_{\alpha\mathbf{0}}} - \bra{\phi_{\alpha\mathbf{0}}}\mathbf{r}\ket{\phi_{\alpha\mathbf{0}}}^2 \right),
\end{align}
is minimized, resulting in Wannier functions that are maximally localized in real space. However, the $e_g$ corresponding to this target state are not isolated from the lower $t_{2g}$ bands and higher La 5$d$ bands (see also Fig.\ \ref{fig: projections atoms}), so that both the construction of the MLWFs and the screening approach will require disentanglement of these Kohn-Sham bands. The disentanglement scheme of Souza \textit{et al.} \cite{souza_maximally_2001} initially constructs a set of $J$ smooth Bloch-like states from $\mathcal{J}_\textbf{k} \geq J $ original Bloch bands, where $\mathcal{J}_\textbf{k}$ can vary throughout the first Brillouin zone, and subsequently constructs a orthonormal basis of MLWFs as defined in \eqref{eq:wann}. The $\mathcal{J}_\textbf{k}$ entangled Bloch bands at every calculated point in reciprocal space $\textbf{k}$ have an energy within the so-called outer energy window, which is fixed by a constant minimum and maximum for the full first Brillouin zone.
Within this outer window, an inner window can be defined where the original Bloch states with an energy in this inner window remain unchanged. In this way, $\mathcal{J}_\textbf{k}$ is reduced with the number of original Bloch states in the inner window for every reciprocal point \textbf{k}. The disentanglement to $J$ Bloch-like states $\ket{\psi_{n\textbf{k}}}$ is then performed by selecting a set of $J$ localized trial orbitals $g_n(\textbf{r})$, which in this case represent the $e_g$ atomic orbitals, and projecting them onto the remaining $\mathcal{J}_\textbf{k}$ original Bloch states $\ket{\psi^0_{m\textbf{k}}}$,
\begin{equation}
\ket{\psi_{n\textbf{k}}} = \sum_{m=1}^{\mathcal{J}_\textbf{k}} \ket{\psi^0_{m\textbf{k}}} \braket{\psi^0_{m\textbf{k}}|g_n} ,
\end{equation}
which becomes after normalization
\begin{equation} \label{MLWF_DISENT}
\ket{\psi^{(w)}_{m\textbf{k}}} = \sum_{m=1}^{J} \ket{\psi_{m\textbf{k}}} (S^{-1/2}_\textbf{k})_{mn}
\end{equation}
with the normalization matrix $(S_\textbf{k})_{mn}=\braket{\psi_{m\textbf{k}}|\psi_{n\textbf{k}}}_V = (A^\dagger_\textbf{k}A_\textbf{k})_{mn}$ where $(A_\textbf{k})_{mn} = \braket{\psi_{m\textbf{k}}|g_n}$ is the overlap matrix of dimension $\mathcal{J}_\textbf{k} \times J$.
Given a set of $J$ MLWFs and a Hamiltonian operator $\hat{H}$, a so-called Wannier interpolation can be done on a denser $k$-mesh, which requires diagonalisations that are computationally cheap compared to diagonalisations of for example Hamiltonian matrices in the plane wave basis set \cite{qiao_calculation_2018}. The $J \times J$ Hamiltonian matrix in the Wannier gauge can be defined as
\begin{equation}
H_{nm}^{(w)}(\textbf{q}) = \bra{\psi^{(w)}_{n\textbf{q}}} \hat{H} \ket{\psi^{(w)}_{m\textbf{q}}}.
\end{equation}
Transforming this Hamiltonian matrix to real-space using Fourier transform results in
\begin{equation}
H_{nm}^{(w)}(\textbf{R}) = \frac{1}{N^3_q} \sum_\textbf{q} e^{-i \textbf{q}.\textbf{R}} H_{nm}^{(w)}(\textbf{q})
\end{equation}
with $N_q$ the (small) number of reciprocal points used in the calculation of the initial Bloch states. The interpolation to a denser $k$-mesh can then be done by performing the inverse Fourier transform,
\begin{equation} \label{MLWF_INTERPOL}
H_{nm}^{(w)}(\textbf{k}) = \sum_\textbf{R} e^{i \textbf{k}.\textbf{R}} H_{nm}^{(w)}(\textbf{R}),
\end{equation}
on arbitrary $k$-point \textbf{k}. The success of this interpolation is due to the localized nature of the MLWFs and thereby of $H_{nm}^{(w)}(\textbf{R})$. The final step is then to diagonalise $H_{nm}^{(w)}(\textbf{k})$ to extract the band structure \cite{qiao_calculation_2018}. The discussed MLWF construction en interpolation is performed with VASP via the Wannier90 interface \cite{kresse_ab_1993, kresse_efficient_1996, pizzi_wannier90_2020}.
For the \textit{Cmcm} phase, the octahedral tilts entail two major precautions. First of all, the $z$-orientation of the trial orbitals is not fixed $z$-axis, and is aligned with the out-of-plane Ni-O bonds to maximally match the interpolation with the Kohn-Sham bands. Secondly, the resulted angles increased hybridization with other entangled orbitals of no $e_g$ character, which should be maximally excluded as made possible with adequate inner windows. This was done for all pressure points, again to optimize interpolation. All outer and inner windows are given, relative to $E_\textrm{F}$ in Fig.\ \ref{fig: Wannier interpolation}(a). The outer windows are the smallest energy intervals in which the $e_g$ character (in Fig.\ \ref{fig: projection nickel}(c)) is fully included, while the inner windows are the largest energy intervals with only $e_g$ character. Latter window is always upper bounded by the spacer La 5$d_{x^2-y^2}$ bands (Fig.\ \ref{fig: projection nickel}(a)) and lower bounded by the Ni 3$d$ $t_{2g}$ bands (Fig.\ \ref{fig: projection nickel}(b)). As an example of the projected Wannierization and subsequent interpolation, Fig.\ \ref{fig: Wannier interpolation} shows the band structure, disentanglement windows and interpolated bands for La$_3$Ni$_2$O$_7$ at 2 GPa (\textit{Cmcm}) and at 40 GPa (\textit{I4/mmm}).
\begin{figure}[ht]
\centering
\subfigure[]{\includegraphics[width=0.32\textwidth]{Figures/SI/windows.png} \label{fig:windows}}
\subfigure[]{\includegraphics[width=0.32\textwidth]{Figures/SI/interpolation_2GPa.png} \label{fig:interpolation_2GPa}}
\subfigure[]{\includegraphics[width=0.32\textwidth]{Figures/SI/interpolation_40GPa.png} \label{fig:interpolation_40GPa}}
\caption{\textbf{Disentanglement windows and Wannier-interpolated band structures in bilayer La$_3$Ni$_2$O$_7$.}
(a) Disentanglement energy windows, with the outer windows indicated in red and the inner windows in blue.
(b) and (c) PBE band structures (solid blue) for crystal structures optimized at 2 GPa and 40 GPa, respectively. Horizontal lines denote the disentanglement windows. Wannier-interpolated bands, obtained by projecting onto the $e_g$ orbitals, are shown as dashed lines.}
\label{fig: Wannier interpolation}
\end{figure}
Isosurfaces of the two inequivalent MLWFs per Ni ion can be seen for the crystal structure optimized at 2 GPa in Fig.\ \ref{fig: isosurfaces}(a), and display strong $d_{x^2-y^2}$ and $d_{z^2}$ character. The individual spreads of each $d_{x^2-y^2}$ and $d_{z^2}$ projected MLWF, as defined in \eqref{eq: MLWF spread}, can be seen in Fig.\ \ref{fig: isosurfaces}(b). In general, the cell compression decreases the spread of both MLWFs. In the $Cmcm$ phase (up to approximately 10 GPa), the $d_{x^2-y^2}$ MLWFs become less localized with increasing octahedral tilt.
\begin{figure}[ht]
\centering
\begin{overpic}[width=0.7\linewidth]{Figures/SI/MLWFs.png}
\put(2,44){(a)}
\put(2,17){(b)}
\put(41,44){(c)}
\end{overpic}
\caption{\textbf{Maximally localized Wannier functions of bilayer La$_3$Ni$_2$O$_7$.}
The isosurface plots of the MLWFs centered on a Ni ion and projected on the (a) $d_{x^2-y^2}$ and (b) $d_{z^2}$ orbitals for the crystal structure of La$_3$Ni$_2$O$_7$ optimized at 2 GPa.
(c) Pressure-dependence of the spreads as defined in \eqref{eq: MLWF spread}.}
\label{fig: isosurfaces}
\end{figure}
\subsection{Downfolding}
The Kohn-Sham bands are downfolded to the orbital-selective (2Ni)$^{5+}$ dimer configuration, spanned by the $e_g$ MLWFs using the multi-scale \textit{ab initio} scheme for correlated electrons \cite{aryasetiawan_frequency-dependent_2004, imada_electronic_2010}. For bands further away from the Fermi level, it is assumed that correlation is weak and, consequently, that DFT describes the high-energy screening adequately. When a suitable low-energy space is obtained, the most important parameters defining a effective multiband Hubbard model can be obtained. In many-body perturbation theory, the random phase approximation is used to typically used to approximate the screened interaction in function of the bare interaction $v$ as
\begin{align}
W = \frac{v}{1-vP} .
\end{align}
The polarization function in the frequency domain is then given by
\begin{equation} \label{eq: POLAR}
P(\mathbf{r},\mathbf{r'},\omega) = -\sum_{\alpha \mathbf{k}}^{\text{occ}}\sum_{\alpha '\mathbf{k'}}^{\text{unocc}} \frac{\psi_{\alpha \mathbf{k}}(\mathbf{r})\psi_{\alpha '\mathbf{k'}}^*(\mathbf{r}) \psi_{\alpha '\mathbf{k'}}(\mathbf{r'})\psi_{\alpha \mathbf{k}}^*(\mathbf{r'})}{\omega +\varepsilon_{\alpha '\mathbf{k'}} -\varepsilon_{\alpha \mathbf{k}} - i\eta} \nonumber
+\sum_{\alpha \mathbf{k}}^{\text{occ}}\sum_{\alpha '\mathbf{k'}}^{\text{unocc}} \frac{\psi_{\alpha \mathbf{k}}^*(\mathbf{r})\psi_{\alpha '\mathbf{k'}}(\mathbf{r}) \psi_{\alpha '\mathbf{k'}}^*(\mathbf{r'})\psi_{\alpha \mathbf{k}}(\mathbf{r'})}{\omega -\varepsilon_{\alpha '\mathbf{k'}} +\varepsilon_{\alpha \mathbf{k}} + i\eta}
\end{equation}
with $\varepsilon_{\alpha \mathbf{k}}$ the eigenvalue corresponding to orbital $\psi_{\alpha \mathbf{k}}(\mathbf{r})$ and $\eta$ an infinitesimally small positive number. The summations over the occupied and unoccupied states can be interpreted as a summation over all possible transitions from below the Fermi energy to above.
The main idea of the constrained random phase approximation (cRPA) is then to split this sum into the transitions within the chosen low-energy subspace and the remaining transitions, $P = P_l + P_r$. As the screening due to states within the target space is included in the effective model, we exclude them from the screened interaction to arrive at a partially screened interaction,
\begin{equation}\label{eq: screened interaction}
W_r = \frac{v}{1-vP_r} ,
\end{equation}
which is to be identified with the Hubbard $U$. However, separating the sum into $P_l$ and $P_r$ is not trivial when the bands of the two subspaces are entangled with each other. Similar to the disentanglement of the Kohn-Sham bands for the Wannierization in \eqref{MLWF_DISENT}, the projector method \cite{projector_Hanke_Sham} initially projects both wave functions on the propagators in \eqref{eq: POLAR} individually. The polarization function within the low-energy subspace then becomes
\begin{equation} \label{PROJ_P_L}
P_l(\mathbf{r},\mathbf{r'},\omega) = \sum_m^{occ}\sum_{m'}^{unocc} A_{mm'}(\mathbf{r}) \times B_{mm'}(\mathbf{r'})
\left[ \frac{1}{\omega-\epsilon_{m'}+\epsilon_{m}+i\eta}-\frac{1}{\omega+\epsilon_{m'}-\epsilon_{m}-i\eta}\right]
\end{equation}
with the projected products $A$ and $B$ defined as
\begin{align}
A_{mm'}(\mathbf{r}) &= \sum_{m_1,m_2'} \mathcal{P}^*_{m_1m} \psi_{m_1}^*(\textbf{r})\psi_{m_2'}(\textbf{r}) \mathcal{P}_{m_2'm'} \\
B_{mm'}(\mathbf{r}) &= \sum_{m_1',m_2} \mathcal{P}^*_{m_2m'} \psi_{m_2}^*(\textbf{r})\psi_{m_1}(\textbf{r}) \mathcal{P}_{m_1m}
\end{align}
and the target projectors $\mathcal{P}_{mn}$ defined as
\begin{equation}
\mathcal{P}_{mn} = \sum_{l} \braket{\phi_l|\psi_m}\braket{\psi_n|\phi_l} .
\end{equation}
These projectors $\mathcal{P}_{mn}$ map the initial, entangled wave functions $\psi_{m}(\textbf{r})$ to a new set of states $\psi_{n}(\textbf{r})$, with expansion coefficients being the overlap with the MLWFs spanning the low-energy subspace. The projected products $A$ and $B$ therefore filter out relative contributions of the low-energy subspace to every initial wave function $\psi_m$. Correct projectors $\mathcal{P}_{mn}$ are therefore essential for accurately projecting onto the target subspace, which requires carefully optimized outer and inner windows.
Furthermore, the projector method introduces a regularization to the projectors that does not conserve the total number of electrons \cite{kaltak_projector}. As a result, the long-wavelength limit—evaluated in VASP using \textbf{k·p} perturbation theory—can become negative, leading to unphysical screening behavior and severely impaired $k$-point convergence. In our system, the choice of appropriate windows is particularly critical, as the high-energy La $5d_{x^2-y^2}$ spacer bands become metallic at pressures above 75 GPa. If these bands are incorrectly included in the projector basis, they introduce significant errors during regularization. This is illustrated in Fig.\ \ref{fig: DIP}, where the orange stars represent the same Wannierization as in the main text but \emph{without} a frozen (inner) window. This omission allows for non-zero La-band components in the Wannier gauge. The red diamonds and orange stars, both with dotted lines, correspond to calculations that exclude the long-wavelength correction from \textbf{k·p} perturbation theory, confirming that the discrepancy near 80 GPa originates from this correction. However, the deviation seen between 12–14 GPa arises due to $k$-point convergence issues when the long-wavelength correction is omitted. Finally, the blue circles, obtained using the weighted cRPA method by {\c{S}}a{\c{s}}{\i}o{\u{g}}lu \textit{et al.}, do not suffer from regularization problems and yield a slope consistent with the green triangles and red diamonds, supporting the accuracy of the projector method with a frozen window.
\begin{figure}[ht]
\centering
\includegraphics[width=0.7\linewidth]{Figures/SI/U1_DIS.png}
\caption{\textbf{Comparison of effective interaction $U_{x^2-y^2}$ under different disentanglement schemes.} The green triangles show results using the corrected projector method with a frozen window, as in the main text. The blue circles employ the weighted cRPA approach. The red diamonds show projector method results with a frozen window but without the long-wavelength correction; the orange stars show results without both the frozen window and the correction; and the purple squares show results without the frozen window but including the correction.
}
\label{fig: DIP}
\end{figure}
Although frequency dependence in \eqref{eq: screened interaction} can have non-negligible effects, the parameters are chosen to be static, valued at $\omega=0$. This can be justified by the retarded screening effects being negligible for small $\omega$, as seen by the horizontal slopes of the on-site Hubbard interactions of La$_3$Ni$_2$O$_7$ at 40 and 70 GPa in Fig.\ \ref{fig: FREQ}.
\begin{figure}[ht]
\centering
\includegraphics[width=0.5\linewidth]{Figures/SI/FREQ.png}
\caption{\textbf{Frequency dependence of cRPA screening}
Frequency dependence of the real and imaginary part of the on-site screened Coulomb interaction $U_{x^2-y^2}$ of bilayer La$_3$Ni$_2$O$_7$ optimized at 40 and 70 GPa.}
\label{fig: FREQ}
\end{figure}
\begin{thebibliography}{10}
\expandafter\ifx\csname url\endcsname\relax
\def\url#1{\texttt{#1}}\fi
\expandafter\ifx\csname urlprefix\endcsname\relax\def\urlprefix{URL }\fi
\providecommand{\bibinfo}[2]{#2}
\providecommand{\eprint}[2][]{\url{#2}}
\bibitem{bednorz_possible_1986}
\bibinfo{author}{Bednorz, J.~G.} \& \bibinfo{author}{Müller, K.~A.}
\newblock \bibinfo{title}{{Possible high $T_\text{c}$ superconductivity in the Ba-La-Cu-O system}}.
\newblock \emph{\bibinfo{journal}{Z. Phys. B}} \textbf{\bibinfo{volume}{64}}, \bibinfo{pages}{189--193} (\bibinfo{year}{1986}).
\bibitem{sun_signatures_2023}
\bibinfo{author}{Sun, H.} \emph{et~al.}
\newblock \bibinfo{title}{{Signatures of superconductivity near 80 K in a nickelate under high pressure}}.
\newblock \emph{\bibinfo{journal}{Nature}} \textbf{\bibinfo{volume}{621}}, \bibinfo{pages}{493--498} (\bibinfo{year}{2023}).
\bibitem{zhang_high-temperature_2024}
\bibinfo{author}{Zhang, Y.} \emph{et~al.}
\newblock \bibinfo{title}{{High-temperature superconductivity with zero resistance and strange-metal behaviour in La$_3$Ni$_2$O$_{7-\delta}$}}.
\newblock \emph{\bibinfo{journal}{Nat. Phys.}} \textbf{\bibinfo{volume}{20}}, \bibinfo{pages}{1269--1273} (\bibinfo{year}{2024}).
\bibitem{li_pressure-driven_2024}
\bibinfo{author}{Li, J.} \emph{et~al.}
\newblock \bibinfo{title}{{Identification of Superconductivity in Bilayer Nickelate La$_3$Ni$_2$O$_7$ under High Pressure up to 100 GPa}}.
\newblock \emph{\bibinfo{journal}{Natl. Sci. Rev.}} (\bibinfo{year}{2025}).
\bibitem{zhang_electronic_2023}
\bibinfo{author}{Zhang, Y.}, \bibinfo{author}{Lin, L.-F.}, \bibinfo{author}{Moreo, A.} \& \bibinfo{author}{Dagotto, E.}
\newblock \bibinfo{title}{{Electronic structure, dimer physics, orbital-selective behavior, and magnetic tendencies in the bilayer nickelate superconductor La$_3$Ni$_2$O$_7$ under pressure}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{108}}, \bibinfo{pages}{L180510} (\bibinfo{year}{2023}).
\bibitem{luo_bilayer_2023}
\bibinfo{author}{Luo, Z.}, \bibinfo{author}{Hu, X.}, \bibinfo{author}{Wang, M.}, \bibinfo{author}{W\'u, W.} \& \bibinfo{author}{Yao, D.-X.}
\newblock \bibinfo{title}{{Bilayer Two-Orbital Model of La$_3$Ni$_2$O$_7$ under Pressure}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. Lett.}} \textbf{\bibinfo{volume}{131}}, \bibinfo{pages}{126001} (\bibinfo{year}{2023}).
\bibitem{gu_effective_2023}
\bibinfo{author}{Gu, Y.}, \bibinfo{author}{Le, C.}, \bibinfo{author}{Yang, Z.}, \bibinfo{author}{Wu, X.} \& \bibinfo{author}{Hu, J.}
\newblock \bibinfo{title}{{Effective model and pairing tendency in the bilayer Ni-based superconductor La$_3$Ni$_2$O$_7$}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{111}}, \bibinfo{pages}{174506} (\bibinfo{year}{2025}).
\bibitem{luo_high-tc_2024}
\bibinfo{author}{Luo, Z.}, \bibinfo{author}{Lv, B.}, \bibinfo{author}{Wang, M.}, \bibinfo{author}{Wú, W.} \& \bibinfo{author}{Yao, D.-X.}
\newblock \bibinfo{title}{{High-$T_\text{c}$ superconductivity in La$_3$Ni$_2$O$_7$ based on the bilayer two-orbital $t$-$J$ model}}.
\newblock \emph{\bibinfo{journal}{npj Quantum Mater.}} \textbf{\bibinfo{volume}{9}}, \bibinfo{pages}{1--7} (\bibinfo{year}{2024}).
\bibitem{yang_possible_2023}
\bibinfo{author}{Yang, Q.-G.}, \bibinfo{author}{Wang, D.} \& \bibinfo{author}{Wang, Q.-H.}
\newblock \bibinfo{title}{{Possible $s$±-wave superconductivity in La$_3$Ni$_2$O$_7$}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{108}}, \bibinfo{pages}{L140505} (\bibinfo{year}{2023}).
\bibitem{lechermann_electronic_2023}
\bibinfo{author}{Lechermann, F.}, \bibinfo{author}{Gondolf, J.}, \bibinfo{author}{B\"otzel, S.} \& \bibinfo{author}{Eremin, I.~M.}
\newblock \bibinfo{title}{{Electronic correlations and superconducting instability in La$_3$Ni$_2$O$_7$ under high pressure}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{108}}, \bibinfo{pages}{L201121} (\bibinfo{year}{2023}).
\bibitem{shen_effective_2023}
\bibinfo{author}{Shen, Y.}, \bibinfo{author}{Qin, M.} \& \bibinfo{author}{Zhang, G.-M.}
\newblock \bibinfo{title}{{Effective bi-layer model Hamiltonian and density-matrix renormalization group study for the high-$T_\text{c}$ superconductivity in La$_3$Ni$_2$O$_7$ under high pressure}}.
\newblock \emph{\bibinfo{journal}{Chin. Phys. Lett.}} \textbf{\bibinfo{volume}{40}}, \bibinfo{pages}{127401} (\bibinfo{year}{2023}).
\bibitem{sakakibara_possible_2024}
\bibinfo{author}{Sakakibara, H.}, \bibinfo{author}{Kitamine, N.}, \bibinfo{author}{Ochi, M.} \& \bibinfo{author}{Kuroki, K.}
\newblock \bibinfo{title}{{Possible High $T_\text{c}$ Superconductivity in La$_3$Ni$_2$O$_7$ under High Pressure through Manifestation of a Nearly Half-Filled Bilayer Hubbard Model}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. Lett.}} \textbf{\bibinfo{volume}{132}}, \bibinfo{pages}{106002} (\bibinfo{year}{2024}).
\bibitem{lu_interlayer_2024}
\bibinfo{author}{Lu, C.}, \bibinfo{author}{Pan, Z.}, \bibinfo{author}{Yang, F.} \& \bibinfo{author}{Wu, C.}
\newblock \bibinfo{title}{{Interlayer coupling driven high-temperature superconductivity in La$_3$Ni$_2$O$_7$ under pressure}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. Lett.}} \textbf{\bibinfo{volume}{132}}, \bibinfo{pages}{146002} (\bibinfo{year}{2024}).
\bibitem{liao_electron_2023}
\bibinfo{author}{Liao, Z.} \emph{et~al.}
\newblock \bibinfo{title}{{Electron correlations and superconductivity in La$_3$Ni$_2$O$_7$ under pressure tuning}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{108}}, \bibinfo{pages}{214522} (\bibinfo{year}{2023}).
\bibitem{qu_bilayer_2024}
\bibinfo{author}{Qu, X.-Z.} \emph{et~al.}
\newblock \bibinfo{title}{{Bilayer $t$-$J$-$J_{\perp}$ model and magnetically mediated pairing in the pressurized Nickelate La$_3$Ni$_2$O$_7$}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. Lett.}} \textbf{\bibinfo{volume}{132}}, \bibinfo{pages}{036502} (\bibinfo{year}{2024}).
\bibitem{yang_interlayer_2023}
\bibinfo{author}{Yang, Y.-f.}, \bibinfo{author}{Zhang, G.-M.} \& \bibinfo{author}{Zhang, F.-C.}
\newblock \bibinfo{title}{{Interlayer valence bonds and two-component theory for high-$T_\text{c}$ superconductivity of La$_3$Ni$_2$O$_7$ under pressure}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{108}}, \bibinfo{pages}{L201108} (\bibinfo{year}{2023}).
\bibitem{wu_superexchange_2024}
\bibinfo{author}{Wú, W.}, \bibinfo{author}{Luo, Z.}, \bibinfo{author}{Yao, D.-X.} \& \bibinfo{author}{Wang, M.}
\newblock \bibinfo{title}{{Superexchange and charge transfer in the nickelate superconductor La$_3$Ni$_2$O$_7$ under pressure}}.
\newblock \emph{\bibinfo{journal}{Sci. China Phys. Mech. Astron.}} \textbf{\bibinfo{volume}{67}}, \bibinfo{pages}{117402} (\bibinfo{year}{2024}).
\bibitem{huang_impurity_2023}
\bibinfo{author}{Huang, J.}, \bibinfo{author}{Wang, Z.~D.} \& \bibinfo{author}{Zhou, T.}
\newblock \bibinfo{title}{{Impurity and vortex states in the bilayer high-temperature superconductor La$_3$Ni$_2$O$_7$}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{108}}, \bibinfo{pages}{174501} (\bibinfo{year}{2023}).
\bibitem{jiang_high-temperature_2024}
\bibinfo{author}{Jiang, K.}, \bibinfo{author}{Wang, Z.} \& \bibinfo{author}{Zhang, F.-C.}
\newblock \bibinfo{title}{{High-Temperature Superconductivity in La$_3$Ni$_2$O$_7$}}.
\newblock \emph{\bibinfo{journal}{Chinese Phys. Lett.}} \textbf{\bibinfo{volume}{41}}, \bibinfo{pages}{017402} (\bibinfo{year}{2024}).
\bibitem{lu_superconductivity_2023}
\bibinfo{author}{Lu, D.-C.} \emph{et~al.}
\newblock \bibinfo{title}{{Superconductivity from doping symmetric mass generation insulators: application to La$_3$Ni$_2$O$_7$ under pressure}}.
\newblock \bibinfo{note}{Preprint at \url{https://arxiv.org/abs/2308.11195} (2023)}.
\bibitem{oh_type_2023}
\bibinfo{author}{Oh, H.} \& \bibinfo{author}{Zhang, Y.-H.}
\newblock \bibinfo{title}{{Type II $t$-$J$ model and shared antiferromagnetic spin coupling from Hund's rule in superconducting La$_3$Ni$_2$O$_7$}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{108}}, \bibinfo{pages}{174511} (\bibinfo{year}{2023}).
\bibitem{zhang_structural_2024}
\bibinfo{author}{Zhang, Y.}, \bibinfo{author}{Lin, L.-F.}, \bibinfo{author}{Moreo, A.}, \bibinfo{author}{Maier, T.~A.} \& \bibinfo{author}{Dagotto, E.}
\newblock \bibinfo{title}{{Structural phase transition, $s$±-wave pairing, and magnetic stripe order in bilayered superconductor La$_3$Ni$_2$O$_7$ under pressure}}.
\newblock \emph{\bibinfo{journal}{Nat. Commun.}} \textbf{\bibinfo{volume}{15}}, \bibinfo{pages}{2470} (\bibinfo{year}{2024}).
\bibitem{kaneko_pair_2024}
\bibinfo{author}{Kaneko, T.}, \bibinfo{author}{Sakakibara, H.}, \bibinfo{author}{Ochi, M.} \& \bibinfo{author}{Kuroki, K.}
\newblock \bibinfo{title}{{Pair correlations in the two-orbital Hubbard ladder: Implications for superconductivity in the bilayer nickelate La$_3$Ni$_2$O$_7$}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{109}}, \bibinfo{pages}{045154} (\bibinfo{year}{2024}).
\bibitem{YangHui_2024}
\bibinfo{author}{Yang, H.}, \bibinfo{author}{Oh, H.} \& \bibinfo{author}{Zhang, Y.-H.}
\newblock \bibinfo{title}{{Strong pairing and symmetric pseudogap metal in double Kondo lattice model: from nickelate superconductor to tetralayer optical lattice}}.
\newblock \bibinfo{note}{Preprint at \url{https://arxiv.org/abs/2408.01493} (2024)}.
\bibitem{YangHuiOh_2024}
\bibinfo{author}{{Yang}, H.}, \bibinfo{author}{{Oh}, H.} \& \bibinfo{author}{{Zhang}, Y.-H.}
\newblock \bibinfo{title}{{Strong pairing from a small Fermi surface beyond weak coupling: Application to La$_3$Ni$_2$O$_7$}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{110}}, \bibinfo{pages}{104517} (\bibinfo{year}{2024}).
\bibitem{shilenko_correlated_2023}
\bibinfo{author}{Shilenko, D.~A.} \& \bibinfo{author}{Leonov, I.~V.}
\newblock \bibinfo{title}{{Correlated electronic structure, orbital-selective behavior, and magnetic correlations in double-layer La$_3$Ni$_2$O$_7$ under pressure}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{108}}, \bibinfo{pages}{125105} (\bibinfo{year}{2023}).
\bibitem{tian_correlation_2024}
\bibinfo{author}{Tian, Y.-H.}, \bibinfo{author}{Chen, Y.}, \bibinfo{author}{Wang, J.-M.}, \bibinfo{author}{He, R.-Q.} \& \bibinfo{author}{Lu, Z.-Y.}
\newblock \bibinfo{title}{{Correlation effects and concomitant two-orbital $s$±-wave superconductivity in La$_3$Ni$_2$O$_7$ under high pressure}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{109}}, \bibinfo{pages}{165154} (\bibinfo{year}{2024}).
\bibitem{liu_-wave_2023}
\bibinfo{author}{Liu, Y.-B.}, \bibinfo{author}{Mei, J.-W.}, \bibinfo{author}{Ye, F.}, \bibinfo{author}{Chen, W.-Q.} \& \bibinfo{author}{Yang, F.}
\newblock \bibinfo{title}{{$s$±-Wave Pairing and the Destructive Role of Apical-Oxygen Deficiencies in La$_3$Ni$_2$O$_7$ under Pressure}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. Lett.}} \textbf{\bibinfo{volume}{131}}, \bibinfo{pages}{236002} (\bibinfo{year}{2023}).
\bibitem{cao_flat_2024}
\bibinfo{author}{Cao, Y.} \& \bibinfo{author}{Yang, Y.-f.}
\newblock \bibinfo{title}{{Flat bands promoted by Hund's rule coupling in the candidate double-layer high-temperature superconductor La$_3$Ni$_2$O$_7$}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{109}}, \bibinfo{pages}{L081105} (\bibinfo{year}{2024}).
\bibitem{qin_high-T_c_2023}
\bibinfo{author}{Qin, Q.} \& \bibinfo{author}{Yang, Y.-f.}
\newblock \bibinfo{title}{{High-$T_\text{c}$ superconductivity by mobilizing local spin singlets and possible route to higher $T_\text{c}$ in pressurized La$_3$Ni$_2$O$_7$}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{108}}, \bibinfo{pages}{L140504} (\bibinfo{year}{2023}).
\bibitem{chen_charge_2024}
\bibinfo{author}{Chen, X.}, \bibinfo{author}{Jiang, P.}, \bibinfo{author}{Li, J.}, \bibinfo{author}{Zhong, Z.} \& \bibinfo{author}{Lu, Y.}
\newblock \bibinfo{title}{{Charge and spin instabilities in superconducting La$_3$Ni$_2$O$_7$}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{111}}, \bibinfo{pages}{014515} (\bibinfo{year}{2025}).
\bibitem{jiang_pressure_2024}
\bibinfo{author}{Jiang, R.}, \bibinfo{author}{Hou, J.}, \bibinfo{author}{Fan, Z.}, \bibinfo{author}{Lang, Z.-J.} \& \bibinfo{author}{Ku, W.}
\newblock \bibinfo{title}{{Pressure Driven Fractionalization of Ionic Spins Results in Cupratelike High-$T_\text{c}$ Superconductivity in La$_3$Ni$_2$O$_7$}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. Lett.}} \textbf{\bibinfo{volume}{132}}, \bibinfo{pages}{126503} (\bibinfo{year}{2024}).
\bibitem{christiansson_correlated_2023}
\bibinfo{author}{Christiansson, V.}, \bibinfo{author}{Petocchi, F.} \& \bibinfo{author}{Werner, P.}
\newblock \bibinfo{title}{{Correlated electronic structure of La$_3$Ni$_2$O$_7$ under pressure}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. Lett.}} \textbf{\bibinfo{volume}{131}}, \bibinfo{pages}{206501} (\bibinfo{year}{2023}).
\bibitem{zhang_trends_2023}
\bibinfo{author}{Zhang, Y.}, \bibinfo{author}{Lin, L.-F.}, \bibinfo{author}{Moreo, A.}, \bibinfo{author}{Maier, T.~A.} \& \bibinfo{author}{Dagotto, E.}
\newblock \bibinfo{title}{{Trends in electronic structures and $s_{\pm}$-wave pairing for the rare-earth series in bilayer nickelate superconductor La$_3$Ni$_2$O$_7$}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{108}}, \bibinfo{pages}{165141} (\bibinfo{year}{2023}).
\bibitem{yi_antiferromagnetic_2024}
\bibinfo{author}{Yi, X.-W.} \emph{et~al.}
\newblock \bibinfo{title}{{Antiferromagnetic Ground State, Charge Density Waves and Oxygen Vacancies Induced Metal-Insulator Transition in Pressurized La$_3$Ni$_2$O$_7$}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{110}}, \bibinfo{pages}{L140508} (\bibinfo{year}{2024}).
\bibitem{chen_orbital-selective_2024}
\bibinfo{author}{Chen, J.}, \bibinfo{author}{Yang, F.} \& \bibinfo{author}{Li, W.}
\newblock \bibinfo{title}{{Orbital-selective superconductivity in the pressurized bilayer nickelate La$_3$Ni$_2$O$_7$: An infinite projected entangled-pair state study}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{110}}, \bibinfo{pages}{L041111} (\bibinfo{year}{2024}).
\bibitem{jiang_theory_2025}
\bibinfo{author}{Jiang, K.-Y.}, \bibinfo{author}{Cao, Y.-H.}, \bibinfo{author}{Yang, Q.-G.}, \bibinfo{author}{Lu, H.-Y.} \& \bibinfo{author}{Wang, Q.-H.}
\newblock \bibinfo{title}{{Theory of Pressure Dependence of Superconductivity in Bilayer Nickelate La$_3$Ni$_2$O$_7$}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. Lett.}} \textbf{\bibinfo{volume}{134}}, \bibinfo{pages}{076001} (\bibinfo{year}{2025}).
\bibitem{li_superconductivity_2019}
\bibinfo{author}{Li, D.} \emph{et~al.}
\newblock \bibinfo{title}{{Superconductivity in an infinite-layer nickelate}}.
\newblock \emph{\bibinfo{journal}{Nature}} \textbf{\bibinfo{volume}{572}}, \bibinfo{pages}{624--627} (\bibinfo{year}{2019}).
\bibitem{sakakibara_orbital_2014}
\bibinfo{author}{Sakakibara, H.} \emph{et~al.}
\newblock \bibinfo{title}{{Orbital mixture effect on the Fermi surface-$T_\text{c}$ correlation in the cuprate superconductors --- bilayer vs single layer}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{89}}, \bibinfo{pages}{224505} (\bibinfo{year}{2014}).
\bibitem{wang_structure_2024}
\bibinfo{author}{Wang, L.} \emph{et~al.}
\newblock \bibinfo{title}{{Structure Responsible for the Superconducting State in La$_3$Ni$_2$O$_7$ at High-Pressure and Low-Temperature Conditions}}.
\newblock \emph{\bibinfo{journal}{J. Am. Chem. Soc.}} \textbf{\bibinfo{volume}{146}}, \bibinfo{pages}{7506--7514} (\bibinfo{year}{2024}).
\bibitem{geisler_structural_2024}
\bibinfo{author}{Geisler, B.}, \bibinfo{author}{Hamlin, J.~J.}, \bibinfo{author}{Stewart, G.~R.}, \bibinfo{author}{Hennig, R.~G.} \& \bibinfo{author}{Hirschfeld, P.~J.}
\newblock \bibinfo{title}{{Structural transitions, octahedral rotations, and electronic properties of $A_3$Ni$_2$O$_7$ rare-earth nickelates under high pressure}}.
\newblock \emph{\bibinfo{journal}{npj Quantum Mater.}} \textbf{\bibinfo{volume}{9}}, \bibinfo{pages}{1--9} (\bibinfo{year}{2024}).
\bibitem{aryasetiawan_frequency-dependent_2004}
\bibinfo{author}{Aryasetiawan, F.} \emph{et~al.}
\newblock \bibinfo{title}{Frequency-dependent local interactions and low-energy effective models from electronic structure calculations}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{70}}, \bibinfo{pages}{195104} (\bibinfo{year}{2004}).
\bibitem{imada_electronic_2010}
\bibinfo{author}{Imada, M.} \& \bibinfo{author}{Miyake, T.}
\newblock \bibinfo{title}{Electronic {Structure} {Calculation} by {First} {Principles} for {Strongly} {Correlated} {Electron} {Systems}}.
\newblock \emph{\bibinfo{journal}{J. Phys. Soc. Jpn.}} \textbf{\bibinfo{volume}{79}}, \bibinfo{pages}{112001} (\bibinfo{year}{2010}).
\bibitem{souza_maximally_2001}
\bibinfo{author}{Souza, I.}, \bibinfo{author}{Marzari, N.} \& \bibinfo{author}{Vanderbilt, D.}
\newblock \bibinfo{title}{Maximally localized {Wannier} functions for entangled energy bands}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{65}}, \bibinfo{pages}{035109} (\bibinfo{year}{2001}).
\bibitem{rhodes_structural_2024}
\bibinfo{author}{Rhodes, L.~C.} \& \bibinfo{author}{Wahl, P.}
\newblock \bibinfo{title}{Structural routes to stabilise superconducting {La$_3$Ni$_2$O$_7$} at ambient pressure}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. Materials}} \textbf{\bibinfo{volume}{8}}, \bibinfo{pages}{044801} (\bibinfo{year}{2024}).
\bibitem{wu_ac_3ni_2o_7_2024}
\bibinfo{author}{Wu, S.} \emph{et~al.}
\newblock \bibinfo{title}{{Ac$_3$Ni$_2$O$_7$ and La$_2$AeNi$_2$O$_6$F (Ae = Sr, Ba): Benchmark Materials for Bilayer Nickelate Superconductivity}}.
\newblock \bibinfo{note}{Preprint at \url{https://arxiv.org/abs/2403.11713} (2024)}.
\bibitem{kresse_ab_1993}
\bibinfo{author}{Kresse, G.} \& \bibinfo{author}{Hafner, J.}
\newblock \bibinfo{title}{Ab initio molecular dynamics for liquid metals}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{47}}, \bibinfo{pages}{558--561} (\bibinfo{year}{1993}).
\bibitem{kresse_efficient_1996}
\bibinfo{author}{Kresse, G.} \& \bibinfo{author}{Furthmüller, J.}
\newblock \bibinfo{title}{Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{54}}, \bibinfo{pages}{11169--11186} (\bibinfo{year}{1996}).
\bibitem{perdew_generalized_1996}
\bibinfo{author}{Perdew, J.~P.}, \bibinfo{author}{Burke, K.} \& \bibinfo{author}{Ernzerhof, M.}
\newblock \bibinfo{title}{Generalized {Gradient} {Approximation} {Made} {Simple}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. Lett.}} \textbf{\bibinfo{volume}{77}}, \bibinfo{pages}{3865--3868} (\bibinfo{year}{1996}).
\bibitem{blochl_projector_1994}
\bibinfo{author}{Blöchl, P.~E.}
\newblock \bibinfo{title}{Projector augmented-wave method}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{50}}, \bibinfo{pages}{17953--17979} (\bibinfo{year}{1994}).
\bibitem{grimme_consistent_2010}
\bibinfo{author}{Grimme, S.}, \bibinfo{author}{Antony, J.}, \bibinfo{author}{Ehrlich, S.} \& \bibinfo{author}{Krieg, H.}
\newblock \bibinfo{title}{A consistent and accurate ab initio parametrization of density functional dispersion correction ({DFT}-{D}) for the 94 elements {H}-{Pu}}.
\newblock \emph{\bibinfo{journal}{J. Chem. Phys.}} \textbf{\bibinfo{volume}{132}}, \bibinfo{pages}{154104} (\bibinfo{year}{2010}).
\bibitem{projector_Hanke_Sham}
\bibinfo{author}{Hanke, W.} \& \bibinfo{author}{Sham, L.~J.}
\newblock \bibinfo{title}{Many-particle effects in the optical spectrum of a semiconductor}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{21}}, \bibinfo{pages}{4656--4673} (\bibinfo{year}{1980}).
\bibitem{kaltak_projector}
\bibinfo{author}{Kaltak, M.}
\newblock \emph{\bibinfo{title}{Merging {GW} with {DMFT}}}.
\newblock Ph.D. thesis, \bibinfo{school}{University of Vienna}, \bibinfo{address}{Wien} (\bibinfo{year}{2015}).
\newblock \bibinfo{note}{Universit{\"a}t Wien, Dissertation}.
\bibitem{marzari_maximally_localized_1997}
\bibinfo{author}{Marzari, N.} \& \bibinfo{author}{Vanderbilt, D.}
\newblock \bibinfo{title}{Maximally-localized generalized {Wannier} functions for composite energy bands}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{56}}, \bibinfo{pages}{12847--12865} (\bibinfo{year}{1997}).
\bibitem{pizzi_wannier90_2020}
\bibinfo{author}{Pizzi, G.} \emph{et~al.}
\newblock \bibinfo{title}{Wannier90 as a community code: new features and applications}.
\newblock \emph{\bibinfo{journal}{J. Phys.: Condens. Matter}} \textbf{\bibinfo{volume}{32}}, \bibinfo{pages}{165902} (\bibinfo{year}{2020}).
\bibitem{nose_unified_1984}
\bibinfo{author}{Nosé, S.}
\newblock \bibinfo{title}{A unified formulation of the constant temperature molecular dynamics methods}.
\newblock \emph{\bibinfo{journal}{J. Chem. Phys.}} \textbf{\bibinfo{volume}{81}}, \bibinfo{pages}{511--519} (\bibinfo{year}{1984}).
\bibitem{shuichi_constant_1991}
\bibinfo{author}{Shuichi, N.}
\newblock \bibinfo{title}{Constant {Temperature} {Molecular} {Dynamics} {Methods}}.
\newblock \emph{\bibinfo{journal}{Prog. Theor. Exp. Phys.}} \textbf{\bibinfo{volume}{103}}, \bibinfo{pages}{1--46} (\bibinfo{year}{1991}).
\bibitem{hoover_canonical_1985}
\bibinfo{author}{Hoover, W.~G.}
\newblock \bibinfo{title}{Canonical dynamics: {Equilibrium} phase-space distributions}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. A}} \textbf{\bibinfo{volume}{31}}, \bibinfo{pages}{1695--1697} (\bibinfo{year}{1985}).
\bibitem{parrinello_crystal_1980}
\bibinfo{author}{Parrinello, M.} \& \bibinfo{author}{Rahman, A.}
\newblock \bibinfo{title}{Crystal {Structure} and {Pair} {Potentials}: {A} {Molecular}-{Dynamics} {Study}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. Lett.}} \textbf{\bibinfo{volume}{45}}, \bibinfo{pages}{1196--1199} (\bibinfo{year}{1980}).
\bibitem{parrinello_polymorphic_1981}
\bibinfo{author}{Parrinello, M.} \& \bibinfo{author}{Rahman, A.}
\newblock \bibinfo{title}{Polymorphic transitions in single crystals: {A} new molecular dynamics method}.
\newblock \emph{\bibinfo{journal}{J. Appl. Phys.}} \textbf{\bibinfo{volume}{52}}, \bibinfo{pages}{7182--7190} (\bibinfo{year}{1981}).
\bibitem{puphal_unconventional_2024}
\bibinfo{author}{Puphal, P.} \emph{et~al.}
\newblock \bibinfo{title}{{Unconventional {Crystal} {Structure} of the {High}-{Pressure} {Superconductor} La$_3$Ni$_2$O$_7$}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. Lett.}} \textbf{\bibinfo{volume}{133}}, \bibinfo{pages}{146002} (\bibinfo{year}{2024}).
\bibitem{chen_polymorphism_2024}
\bibinfo{author}{Chen, X.} \emph{et~al.}
\newblock \bibinfo{title}{{Polymorphism in the {Ruddlesden}–{Popper} {Nickelate} La$_3$Ni$_2$O$_7$: {Discovery} of a {Hidden} {Phase} with {Distinctive} {Layer} {Stacking}}}.
\newblock \emph{\bibinfo{journal}{J. Am. Chem. Soc.}} \textbf{\bibinfo{volume}{146}}, \bibinfo{pages}{3640--3645} (\bibinfo{year}{2024}).
\bibitem{wang_long-range_2024}
\bibinfo{author}{Wang, H.}, \bibinfo{author}{Chen, L.}, \bibinfo{author}{Rutherford, A.}, \bibinfo{author}{Zhou, H.} \& \bibinfo{author}{Xie, W.}
\newblock \bibinfo{title}{{Long-{Range} {Structural} {Order} in a {Hidden} {Phase} of {Ruddlesden}–{Popper} {Bilayer} {Nickelate} La$_3$Ni$_2$O$_7$}}.
\newblock \emph{\bibinfo{journal}{Inorg. Chem.}} \textbf{\bibinfo{volume}{63}}, \bibinfo{pages}{5020--5026} (\bibinfo{year}{2024}).
\bibitem{zhang_electronic_2024}
\bibinfo{author}{Zhang, Y.}, \bibinfo{author}{Lin, L.-F.}, \bibinfo{author}{Moreo, A.}, \bibinfo{author}{Maier, T.~A.} \& \bibinfo{author}{Dagotto, E.}
\newblock \bibinfo{title}{{Electronic structure, self-doping, and superconducting instability in the alternating single-layer trilayer stacking nickelates La$_3$Ni$_2$O$_7$}}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{110}}, \bibinfo{pages}{L060510} (\bibinfo{year}{2024}).
\bibitem{ouyang_absence_2024}
\bibinfo{author}{Ouyang, Z.}, \bibinfo{author}{Gao, M.} \& \bibinfo{author}{Lu, Z.-Y.}
\newblock \bibinfo{title}{{Absence of electron-phonon coupling superconductivity in the bilayer phase of La$_3$Ni$_2$O$_7$ under pressure}}.
\newblock \emph{\bibinfo{journal}{npj Quantum Mater.}} \textbf{\bibinfo{volume}{9}}, \bibinfo{pages}{1--6} (\bibinfo{year}{2024}).
\bibitem{qiao_calculation_2018}
\bibinfo{author}{Qiao, J.}, \bibinfo{author}{Zhou, J.}, \bibinfo{author}{Yuan, Z.} \& \bibinfo{author}{Zhao, W.}
\newblock \bibinfo{title}{Calculation of intrinsic spin {Hall} conductivity by {Wannier} interpolation}.
\newblock \emph{\bibinfo{journal}{Phys. Rev. B}} \textbf{\bibinfo{volume}{98}}, \bibinfo{pages}{214402} (\bibinfo{year}{2018}).
\end{thebibliography}