Quantitative thermodynamic study of superconducting and normal states in UTe2 under pressure
Extracted source text · arXiv 2603.29760
\documentclass[aps,prl,twocolumn,amsmath,amssymb,floatfix,superscriptaddress,citeautoscript]{revtex4-2}
%\documentclass[twocolumn,showpacs,%
% nofootinbib,aps,superscriptaddress,%
% eqsecnum,prd,notitlepage,showkeys,10pt]{revtex4-1}
\usepackage{amssymb}
\usepackage{amsmath}
\usepackage{graphicx}
\usepackage{dcolumn}
\usepackage{hyperref}
\usepackage{siunitx}
\usepackage{bm}
\usepackage{xcolor}
\usepackage{natbib}
\newcommand{\warn}[1]{\textcolor{red}{[#1]}}
\newcommand{\gray}[1]{\textcolor{gray}{[#1]}}
\def\etal.{et\penalty50\ al.}
\usepackage[commentmarkup=todo, authormarkup=none, highlightmarkup=background]{changes}
\definechangesauthor[name = Timothee, color=blue]{T}
\definechangesauthor[name = Georg, color=orange]{GK}
\newcommand{\pheliqs}{Univ. Grenoble Alpes, CEA, Grenoble INP, IRIG, PHELIQS, F-38000 Grenoble, France}
\newcommand{\neel}{Univ. Grenoble Alpes, CNRS, Institut Néel, F-38000 Grenoble, France}
\newcommand{\lncmiG}{Univ. Grenoble Alpes, INSA Toulouse, Univ. Toulouse Paul Sabatier, EMFL, CNRS, LNCMI, F-38000 Grenoble, France}
\newcommand{\imr}{Institute for Materials Research, Tohoku University, Oarai, Ibaraki, 311-1313, Japan}
\newcommand{\TU}{Institute for Solid State and Materials Physics, TU Dresden University of Technology, 01062 Dresden, Germany}
\newcommand{\MPG}{Max Planck Institute for Chemical Physics of Solids, 01187 Dresden, Germany}
\newcommand{\KIT}{Institute for Quantum Materials and Technologies, Karlsruhe Institute of Technology, Kaiserstra\ss e 12, 76131 Karlsruhe, Germany }
\begin{document}
\title{Quantitative thermodynamic study of superconducting and normal states in \texorpdfstring{UTe$_2$}{UTe2} under pressure}
\author{T.~Vasina}
\thanks{These authors contributed equally to this work.}
\affiliation{\pheliqs}
\author{M.~Pfeiffer}
\thanks{These authors contributed equally to this work.}
\affiliation{\KIT}
\affiliation{\MPG}
\affiliation{\TU}
\author{R.~Borth}
\affiliation{\MPG}
\author{M.~Nicklas}
\affiliation{\MPG}
\author{M.~Amano Patino}
\affiliation{\pheliqs}
\affiliation{\neel}
\author{G.~Lapertot}
\affiliation{\pheliqs}
\author{J.-P.~Brison}
\affiliation{\pheliqs}
\author{E.~Hassinger}
\affiliation{\KIT}
\affiliation{\MPG}
\author{G.~Knebel}
\email[E-mail me at: ]{georg.knebel@cea.fr}
\affiliation{\pheliqs}
\author{D.~Braithwaite}
\email[E-mail me at: ]{daniel.braithwaite@cea.fr}
\affiliation{\pheliqs}
\begin{abstract}
We report a quantitative calorimetric study of UTe$_2$ under pressure with a direct measurement of the Sommerfeld $\gamma$ coefficient, showing a three-fold enhancement of electronic effective mass when approaching the critical pressure where superconductivity is suppressed and ordered states occur. We analyse the evolution of $\gamma$ with the amplitude of the jumps in the specific heat at the two superconducting transitions, and the superconducting critical temperature with pressure. This analysis would suggest that the high pressure superconducting phase nucleates only on a fraction of the Fermi surface. It also points to the possible major role of a quantum critical point of the unidentified phase that has been called weak magnetic order, rather than to the critical pressure of the antiferromagnetic phase. Just at the border of long-range antiferromagnetic order, where superconductivity emerges from the weak magnetic order phase, a significant increase in the specific heat jump for both superconducting transitions is found, accompanied by a noticeable change of their shapes.
\end{abstract}
\maketitle
%\section{Introduction}
In the landscape of unconventional superconductors, UTe$_2$ stands out as a remarkable material. Not only is it a very probable candidate for stabilizing spin-triplet superconductivity \cite{Ran2019,Aoki2019}, but it also exhibits several distinct superconducting states, a rare feature observed in only a handful of compounds \cite{Khim2021, Sauls1994, Ott1985}. These distinct superconducting phases can be induced by tuning the magnetic field \cite{Rosuel2023}, applying hydrostatic pressure \cite{Braithwaite2019, honda_pressure-induced_2023}, or varying the direction of the magnetic field \cite{Ran2019a}. Each superconducting phase of UTe$_2$ is thought to be stabilized by a distinct mechanism \cite{Rosuel2023, Wu2025}.
When applying pressure on UTe$_2$, a new superconducting phase (SC2) is induced above $\SI{0.2}{\giga\pascal}$, and its critical temperature reaches a maximum near $\SI{1}{\giga\pascal}$ with a higher $T_\text{c}$ than that of the superconducting state SC1 at ambient pressure. When further increasing pressure to a critical pressure $p_\mathrm{c}$, superconductivity is abruptly suppressed by a first-order like transition, and a long-range antiferromagnetic (AFM) order is promoted below $T_N \approx 3.5$~K \cite{knafo_incommensurate_2025}. Another phase transition is detected by resistivity \cite{Ran2020, Aoki2021}, magnetization \cite{Li2021}, and calorimetry \cite{Thomas2020} at a higher temperature. This phase was proposed to be a short-range magnetic order and named “Weak Magnetic Order” (WMO), although its microscopic nature is yet to be determined. It was recently shown that the pressure-induced phase SC2 is identical to the high-field-induced superconducting phase for $\bm{H}$ along the $\bm{b}$-axis \cite{vasina_connecting_2025}. This field-induced phase for $\bm{H}\parallel\bm{b}$ has been extensively studied, with a quantitative specific heat study bringing important information on the superconducting order parameter and pairing mechanism \cite{Rosuel2023}. In addition to clearly showing the existence of two distinct superconducting phases, this study revealed an increase of the normal state Sommerfeld coefficient $\gamma$ by a factor of two with increasing field up to the metamagnetic transition at $H_m \approx 34.5$~T. This indicates an enhanced effective mass $m^*$ from magnetic fluctuations on approaching $H_m$ \cite{Tokunaga2023}. The same mechanism controlling this mass enhancement might also trigger the increase of the superconducting transition with field for $\bm{H} \parallel \bm{b}$. This also suggests a different pairing mechanism of SC2 from that responsible for low field superconductivity SC1. This quantitative study allowed a careful analysis of the height and width of the jumps in $C/T$ related to the two superconducting transitions, providing insights into the SC2 superconducting phase.
As SC2 can also be stabilized by pressure, a quantitative calorimetric study under pressure could similarly help refine our understanding of this SC2 phase.
Several ac calorimetry studies under pressure have already been performed \cite{Braithwaite2019, Thomas2020, Aoki2020, Wu2025PRL}. They revealed the existence of two superconducting phases, the appearance of the AFM and WMO phases, and some qualitative information on changes in the size of the jumps of $C/T$ at the onset of superconductivity. However ac calorimetry under pressure does not provide reliable quantitative values of the specific heat. A qualitative indication of enhanced electronic correlations on increasing pressure is seen in transport properties \cite{Valiska2021, Braithwaite2019, Thomas2020}, as well as microscopic nuclear magnetic resonance (NMR) measurements \cite{Kinjo2023PRB, Ambika2026}. A reliable determination of an increase of $\gamma$ under pressure, revealing the enhancement of the density of states and effective mass $m^*$ of the quasi-particles, together with its relationship to the appearance of AFM and WMO orders, and possible quantum critical effects, is lacking. A quantitative determination of the evolution of the relative $C/T$ jumps at the superconducting transitions would also clarify the superconducting phase diagram, namely the fact that three second-order transition lines meet in one point, and give more insight on the SC2 phase.
\begin{figure}[t]
\centering
\includegraphics[width=\linewidth]{IMG/0T_pc_dresde.pdf}
\caption{\label{fig:raw}Specific heat as a function of temperature at different pressures. Blue and orange triangles represent superconducting transitions (to resp. SC1 and SC2). Violet and grey triangles represent the magnetic phases (resp. AFM and WMO).}
\end{figure}
Here, we present a quantitative specific heat study of UTe$_2$ under pressure. In order to preserve a sizable sample contribution against the addenda, we used a rather large single crystal ($m=\SI{23.73}{\milli\gram}$), hence grown by Chemical Vapor Transport (CVT). It was placed in a small piston-cylinder pressure cell. The total heat capacity of the sample and pressure cell was measured using a quasi-adiabatic technique \cite{Nicklas2015}. The specific heat of UTe$_2$ was obtained by subtracting the contribution of the pressure cell, pressure transmitting medium (Fluorinert FC-72), and teflon cap, taking into account pressure effects on the specific heat of the latter two. Before the experiment, the specific heat of the sample was measured at ambient pressure. This measurement, and more details on the methods, are given in the Supplemental Material \cite{suppl}. Measurements were performed from close to ambient pressure up to $\SI{1.48}{\giga\pascal}$, above the critical pressure $p_\text{c}$ where long-range AFM order occurs.
The results of this study are shown in Fig.~\ref{fig:raw}. At nearly ambient pressure, a single phase transition is observed at about $T_\text{c1} = \SI{1.75}{\kelvin}$, corresponding to the transition between the paramagnetic phase and the superconducting SC1 phase. When pressure is increased to $\SI{0.55}{\giga\pascal}$, a second transition appears at higher temperatures, corresponding to the pressure-induced SC2 phase. At first, $T_\text{c2}$ is enhanced by pressure, but above $\approx \SI{1}{\giga\pascal}$, $T_\text{c2}$ starts decreasing. When reaching $p=\SI{1.43}{\giga\pascal}$, $T_\text{c1}$ and $T_\text{c2}$ are at their lowest. At this pressure, a broad anomaly has appeared at higher temperatures, corresponding to the WMO phase. Finally, at the highest pressure of $\SI{1.48}{\giga\pascal}$, the sample is above the critical pressure and superconductivity is completely suppressed, while antiferromagnetic order is stabilized at $T_\text{AFM} = \SI{3.3}{\kelvin}$. The anomaly associated with the magnetic WMO phase is now characterized by a larger specific heat jump, at around $\SI{4.5}{\kelvin}$.
\begin{figure}
\centering
\includegraphics[width=0.85\linewidth]{IMG/0T_P-T_max.pdf}
\caption{Zero-field $p-T$ phase diagram of UTe$_2$ obtained from quantitative specific heat measurements in the piston cylinder cell (red squares), and ac calorimetry in the diamond anvil cell (green circles). The pressure is normalized to the critical pressure where the AFM order appears (1.45 GPa and 1.62 GPa for the piston-cylinder and DAC respectively. The WMO phase is stabilized in a significant pressure range before the critical pressure where the AFM state appears, with the maximum $T_\text{c}$ lying close to the extrapolated critical pressure $p_\text{WMO}$.}
\label{fig:P-T}
\end{figure}
These results indicate that the WMO phase is stabilized above the superconducting SC2 phase, at pressures clearly below the antiferromagnetic critical pressure. We also confirmed this on a high-quality MSF sample measured in a diamond anvil cell with in-situ pressure tuning to probe the phase diagram in more details. The resulting phase diagram for the two studies is presented in Fig.~\ref{fig:P-T}. Because of the different experimental conditions, the pressure scale is normalized to the critical pressure $p_{\rm c}$ of the AFM long range order for each case. This shows a significant pressure range where both, superconductivity and the WMO phase are found. It suggests that the WMO transition temperatures can be extrapolated down to $\SI{0}{\kelvin}$, at a putative critical pressure $p_\text{WMO}$, here equal to roughly $\SI{1.2}{\giga\pascal}$, close to the maximum of SC2. However no sign of the WMO transition was observed within the superconducting domain, suggesting that the onset of superconductivity could expel the WMO phase.
Coming back to the absolute measurements of Fig.~\ref{fig:raw} , we extract the normal-state electronic specific heat and the pressure evolution of the Sommerfeld coefficient $\gamma$. The latter is obtained from the constant value of $C/T$ in the paramagnetic Fermi-liquid regime above the ordered phases, after subtraction of a weak phonon contribution determined at ambient pressure and assumed to be pressure-independent. Further details of the analysis are given in \cite{suppl}). This is shown in Fig.~\ref{fig:compil}(a). The electronic specific heat increases with pressure and reaches a maximum at $\SI{1.29}{\giga\pascal}$, where the Sommerfeld coefficient $\gamma$ is enhanced by a factor of three. This behavior closely parallels the field-induced enhancement of $\gamma$ for $\bm{H}\parallel\bm{b}$, supporting a correlation between the mass renormalization and the reinforcement of pairing in the SC2 phase under both tuning parameters. At the two highest pressures, the determination of $\gamma$ is complicated by the emergence of the WMO and AFM phases. Assuming that the paramagnetic Fermi-liquid regime is recovered at $T=\SI{6}{\kelvin}$, we extract $\gamma$ from the corresponding $C/T$ value and find that it decreases markedly with further increasing pressure.
\begin{figure}[t]
\centering
\includegraphics[width=0.95\linewidth]{IMG/compil_en.pdf}
\caption{\label{fig:compil}(a) Relative evolution of gamma coefficient and entropy as a function of pressure. The red dotted line is a guide for the eye, the orange line a prediction from a strong coupling model with an increase of the fraction ($\alpha$ - blue dashed line on right axis) of the Fermi surface affected by the pairing mechanism of SC2 (see text) (b) Specific heat jumps for both superconducting transitions. The BCS jump expected from gamma is also shown in grey. The orange arrow shows the extrapolated crossing point of SC1 and SC2. (c) Zero-field phase diagram obtained from Fig.~\ref{fig:raw}. Dashed lines: strong coupling constants (right axis) reproducing the pressure dependence of the critical temperatures of SC1 and SC2.}
\end{figure}
In Fig.~\ref{fig:compil}(a), we also plot the entropy, calculated by integrating the electronic contribution of $C/T$ (after subtraction of the phonon contribution) up to a temperature in the normal state. This was usually 4K except for the highest two pressures where we used 6K to be above the WMO and AFM transition temperatures. At ambient pressure, the entropy balance at the superconducting transition is well respected, with $S/T = \gamma \approx \SI{120}{\milli\joule\per(\mole\cdot\kelvin\squared)}$. When increasing the pressure, both $\gamma$ and the entropy rise in a similar fashion, respecting entropy balance even under pressure up to $\SI{1.29}{\giga\pascal}$. For the two points at higher pressure, there is clearly an excess of entropy compared to the value of $\gamma$ determined above $\SI{6}{\kelvin}$, but in the presence of the WMO and AFM phases, there could be additional contributions to entropy.
For both superconducting transitions, the associated jump size in specific heat is plotted in Fig.~\ref{fig:compil}(b) as a function of pressure. We also show the expected BCS value $\Delta C /T_\text{c} (p) = 1.43 \gamma (p)$, taking into account the evolution of $\gamma$ with pressure. At ambient pressure, the specific heat jump size of SC1 is slightly higher than the BCS value. Indeed, ambient pressure measurements on high-quality single crystals have shown that UTe$_2$ is in the strong coupling regime, with $\Delta C/T_\text{c} \approx 2.3 \gamma $ \cite{aoki_molten_2024}. In this study, the jump size appears smaller compared to measurements on Molten Salt Flux (MSF)-grown crystals, due to the lower quality of CVT-grown crystals and the lower $T_\text{c1}$. The jump size and $T_\text{c1}$ are well correlated \cite{Cairns2020, Rosa2022a}. As pressure increases, the jump associated with SC1 decreases.
Regarding SC2, an extrapolation to low pressures indicates that the jump size is close to zero at the pressure where the SC2 transition line meets the SC1 boundary (marked with an orange arrow), in agreement with previous studies \cite{Braithwaite2019, vasina_connecting_2025}.
With increasing pressure, the jump associated with SC2 rapidly grows up to $\SI{1.3}{\giga\pascal}$, like the electronic specific heat coefficient but in contrast to the behavior of $T_\text{c2}$, which has a maximum at around 1 GPa. Finally, at $\SI{1.43}{\giga\pascal}$, the highest pressure where superconductivity is still present in this study, both the SC1 and the SC2 phases show a sudden increase in their jump size even though $\gamma$ decreases.
At this pressure, where the SC2 phase emerges in the presence of WMO, we also notice a change in the shape of both transitions.
The SC1 transition gets extremely sharp with an enhanced jump, while the SC2 anomaly increases too but evolves into a broader, almost symmetrical peak. Large pressure inhomogeneities are ruled out by the sharpness of the SC1 transition, although, as we are close to the pressure where superconductivity is abruptly suppressed, small inhomogeneities may have a strong effect on SC2. A recent ultrasound attenuation study proposed a surprisingly large extension of the superconducting fluctuation regime under pressure near $\SI{0.77}{\giga\pascal}$ for the SC2 transition \cite{kamat2026}. But we do not find no a similar effect in our specific heat data in this pressure range. On the contrary, at such pressures the specific heat anomaly is sharper than at lower pressures, as found already in previous work \cite{vasina_connecting_2025}. A significant broadening and change of behavior of the specific heat anomaly is only observed in a narrow pressure range around $p_\text{c}$,
and will be further discussed below.
Focusing again on the pressure range below $\SI{1.4}{\giga\pascal}$, where the SC2 phase emerges from a normal state without WMO, the value of $\gamma(p)$, of the specific heat jumps, and the critical temperatures should be related. We can use a simple model where the evolution of $T_\text{c}$ under pressure of SC2 and SC1 is only governed by that of the pairing strength of their respective mechanism (see Supplemental Material \cite{suppl} for more details).
The large specific heat jump of the SC1 phase at ambient pressure and the strong increase of the Sommerfeld coefficient under pressure point to a strong coupling regime for both superconducting phases. In Fig.~\ref{fig:compil}(c), we have shown a typical pressure dependence of the strong coupling constants $\lambda_{\text{SC2}}$ and $\lambda_{\text{SC1}}$ reproducing $T_\text{c}(p)$ for both phases.
Strong coupling effects should then lead to a renormalization of the electronic specific heat coefficient as $C/T = \gamma(p) = \gamma_0 [1+ \lambda_{\text{SC1}}(p)][1+ \lambda_{\text{SC2}}(p)]$. This predicts a smooth maximum around $\SI{1}{\giga\pascal}$, far from matching the experimental $\gamma(p)$ reported on Fig.~\ref{fig:compil}(a) \cite{suppl}.
Moreover, in such a simple strong coupling picture, the specific heat jump of SC2 should always be larger than the BCS jump, whereas
the amplitude of $\Delta C/T$ for SC2 seemingly vanishes at low pressure. To account for these discrepancies we suggest a model where the SC2 phase nucleates only on a small fraction of the Fermi surface.
We can take the data of the absolute specific heat jump as measure of this fraction ($\alpha$) of the Fermi surface \cite{suppl}.
Then, renormalization of
$\gamma(p)$ by $\lambda_{\text{SC2}}$ should happen only on that fraction of the Fermi surface on which SC2 condenses.
We report on Fig.~\ref{fig:compil}(a) the resulting behavior calculated for $\alpha(p)$ and $\gamma(p)$ (more details in \cite{suppl}):
the prediction for the pressure dependence of $\gamma(p)$, and notably the continuous increase beyond the pressure of the maximum of $T_\text{c}$ for SC2, is much closer to the experimental data.
Everything happens as if indeed, the SC2 phase first nucleates on a small portion of the Fermi surface, rapidly increasing under pressure. This description can explain how the critical temperature of SC2 decreases at $\SI{1.29}{\giga\pascal}$ whereas $\Delta C/T$ and $\gamma$ increase. It also accounts for the vanishing value of $\Delta C/T$ for SC2 where the SC1 and SC2 lines meet. This fact is noteworthy, as this condition is required for a thermodynamically consistent phase diagram in which three second-order transition lines meet at a single point \cite{Braithwaite2019, Wu2025PRL}.
What this empirical analysis does not solve, is the puzzle of a large strong coupling constant ($\lambda_{\text{SC2}}(p) \geq 1.5$) for the SC2 phase when it emerges near $\SI{0.3}{\giga\pascal}$, with a vanishing weight on the Fermi surface:
superconductivity with very small carrier numbers exist in other systems, but it requires very specific, often elusive, pairing mechanisms \cite{CollignonARCMP2019, KimSciAdv2018}.
Regarding the pressure where SC2 emerges from a WMO phase, the strong increase in the size of the jump in $C/T$ at both the SC2 and SC1 transitions at $\SI{1.43}{\giga\pascal}$ whereas $\gamma$ and $T_\text{c}$ are decreasing, might appear surprising. A possible explanation is suggested by the large entropy measured at 6\,K around the critical pressure: it points to the presence of strong magnetic fluctuations in the WMO phase. Then, the gap opening on the Fermi surface induced by the transition into the superconducting state might also trigger magnetic entropy reduction, hence a larger specific heat jump than predicted even by strong coupling scenarios.
If, as our empirical model suggests, SC2 concerns only a portion of the Fermi surface even close to the critical pressure (see Fig.~\ref{fig:compil}(a)),
further reduction of the magnetic entropy should occur at the SC1 phase transition (taking place on the whole Fermi surface), explaining also its large jump at $\SI{1.43}{\giga\pascal}$ (Fig.~\ref{fig:compil}(b)).
Microscopically, the existence of a maximum under pressure where $\gamma$ is increased by a factor of three gives a direct indication of enhanced electronic correlations, and possible quantum criticality under pressure. Previously, this had been indirectly inferred mainly from transport measurements: the $A$ coefficient of the quadratic Fermi-liquid temperature dependence ($\rho$=$\rho_0+AT^2$) was found to increase by a factor of six \cite{Braithwaite2019} to eight \cite{Valiska2021}. Although the value of the $A$ coefficient depends critically on the current direction \cite{Eo2022, Thebault2022, Knebel2024}, the usual relationship ($m^* \propto A^2$) is compatible with our direct determination of a threefold enhancement of $\gamma$. Another study, analyzing the resistivity with a non-Fermi-liquid law ($\rho = \rho_0 + A_nT^n$), found a minimum of the resistivity exponent $n \approx 1$ under pressure, indicating possible quantum criticality. Microscopic evidence of the increase of AFM spin fluctuations with pressure has been obtained by NMR measurements under pressure \cite{Kinjo2023PRB, Ambika2026}, which suggests the importance of these fluctuations for the development of the SC2 phase.
A striking aspect of our study is that $\gamma$, and thus the electronic correlations that are associated with it, reaches its highest value at a pressure significantly lower than the critical pressure $p_\textit{c}$ where long-range AFM order is established. This quite surprising fact was already suggested from the previous transport studies where the maximum of the $A$ coefficient \cite{Braithwaite2019}, and the minimum of the exponent $n$ \cite{Thomas2020}, were also found at pressures below the AFM critical pressure, possibly coinciding with an extrapolated critical pressure for the WMO phase. Indeed, there appears to be a near coincidence of the three phenomena, the maximum of $\gamma$, the maximum in $T_\text{c}$ of SC2, and the extrapolated critical pressure of the WMO phase $p_\text{WMO}$. This could indicate a significant role of the WMO phase in stabilizing SC2 through a quantum critical point. This is not incompatible with a simultaneous competition between the two phases as suggested by our scenario for the enhanced specific heat jumps at $\SI{1.43}{\giga\pascal}$. Indeed superconductivity is known to sometimes suppress the ordered phase whose quantum critical fluctuations are believed to mediate pairing in AFM systems \cite{Knebel2011} and charge density wave systems \cite{Fradkin2015, Chen_2016}.
While this, and several other points, remain to be confirmed or clarified, our quantitative calorimetry study of UTe$_2$ under pressure brings solid new elements to the previous qualitative studies, determining precisely the threefold enhancement of the electronic effective mass with pressure, and quantitative analysis of the pressure effects on the height of the $C/T$ jumps at the superconducting transitions. The direct measurement of a maximum of $\gamma$ under pressure, seemingly linked to
the emergence of the WMO phase above SC2, and the maximum of the SC2 critical temperature, possibly linked to the WMO critical pressure,
provide evidence of a strong interplay between the WMO and the high-pressure superconducting phase.
\begin{acknowledgements}
We thank D. Aoki, M. Houzet and M. Zhitomirsky for fruitful discussions. This work was supported by the French National Agency for Research (ANR) within the projects FRESCO No. ANR-20-CE30-0020 and SCATE No. ANR-22-CE30-0040.
The quantitative heat capacity experiments under hydrostatic pressure were performed at the Max Planck Institute for Chemical Physics of Solids (MPI CPfS) in Dresden, Germany. T.V. thanks MPI CPfS for their hospitality. E. H. acknowledges funding by the DFG through CRC1143 (Project No. 247310070) and the W\"{u}rzburg- Dresden Cluster of Excellence on Complexity and Topology in Quantum Matter—ct.qmat (EXC 2147, Project 6 ID 390858490) and by the ERC Consolidator Grant Ixtreme (GA 101125759).
\end{acknowledgements}
\bibliographystyle{apsrev4-2}
\bibliography{ute2}
\end{document}