WorksPhysics and Astronomy
Experimental and theoretical thermodynamic studies in BaMgReO – the ground state in the context of Jahn-Teller effect
Abstract
We address the degeneracy of the ground state multiplet on the 5 Re ion in double perovskite BaMgReO using a combination of specific heat measurements and density functional calculations. For BaMgReO, two different ground state multiplets have previously been proposed - a quartet (with degeneracy =4) [1] and a doublet (=2) [2]. Here we employ two independent methods for the estimation of phonon contribution in heat capacity data to obtain the magnetic entropy , which reflects the degeneracy of the ground state multiplet through ln. In both cases, we obtain a better fit to ln2 indicating evidence of =2 degeneracy in the range from 2 to 120~K. The detailed nature of the ground state multiplet in BaMgReO remains an open question.
Introduction
In recent years there has been an increase in interest in materials which exhibit strong spin-orbit coupling (SOC) with both experimental and theoretical studies revealing exotic phases, such as spin-liquids [3, 4] or systems with multipolar orders [5, 6, 7]. Since the strength of SOC grows rapidly with increasing atomic number , it is important in materials containing ions, where the entanglement between spin and orbital degrees of freedom profoundly affects the low-energy excitation spectrum.
One such family of compounds comprises double perovskites containing elements Ta, Re and Os with formally single electron occupying their orbitals (). These have general chemical formula , with the magnetic ion surrounded by six anions, creating a highly symmetric octahedron. The octahedral crystal electronic field (CEF) splits the five-fold degenerate orbitals into characteristic and orbital sets (see Figure 1). Under the influence of SOC the three-fold degenerate set (with an effective orbital momentum ) is split further into a lower-lying quartet () and a doublet () at higher energy. Up to this point, the single ion ground state multiplet is the quartet () proposed in [1] paper. However, further investigation of some double perovskites showed the structural distortion from cubic to tetragonal unit cell and elongation or compression of the ideal -O octahedron along -axis [8, 9]. In theory, this is known as Jahn-Teller effect that leads to another splitting of ground quartet into two Kramers doublets, where , and which translates into value of entropy.

Figure 1. Proposed energy level splitting scheme for electronic configuration in DPs. Elongation of ideal Re-O octahedron along c-axis [8], known as Jahn-Teller (J-T) effect, lifts degeneracy of orbitals and splits original quartet into two Kramer doublets. The splitting of higher energy levels is not shown for clarity.
The coupling between magnetic moments has been considered to arise from a super- superexchange mechanism, comprising both ferromagnetic () and antiferromagnetic () type and leading to a long-range magnetic order at [5]. Additionally, it has been suggested that systems could allow investigation of higher-order multipoles due to their small dipole magnetic moments [5]. For an electronic quadrupolar interaction that is large enough, the prediction of a quadrupolar order at has been made [5]. This causes a small lattice distortion from a perfect cubic environment.
Several double perovskites have been previously investigated in this context. In BaNaOsO, where the Os ion is in a formally 7+ valence state, a long-range magnetic ordering has been found to occur around K [10, 11, 12, 13], preceded by a breaking of local point symmetry, indicated by the splitting of NMR spectra below 12 K [12]. Specific heat studies revealed a sharp transition at [13, 14] and a small, broad shoulder extending towards higher temperatures. Authors assigned this shoulder to the nematic (orthogonal paramagnetic) phase based on temperature derivative of specific heat data [13]. For only a single transition has been observed [10, 11] and no detailed heat studies have been reported. In the rhenium family several compounds have been investigated, all with Re in a formally 6+ valence state and the site occupied by 2+ ions. For , occurs at 18 K [1, 2], with a broad shoulder revealed in calorimetric data around 33 K. This shoulder has been associated with a cubic to tetragonal structural distortion and argued to indicate the quadrupolar order [8]. DFT calculations further demonstrated the ordering of the proposed charge quadrupoles and suggested an additional spin-canting dependent ordered charge quadrupolar component[15]. A similar sequence of features has been revealed in BaZnReO ( K, K) [16] and BaCdReO ( K, K) [17], pointing to a universal behavior across the family. Additionally, charge quadrupoles have also been recently discussed in the more ionic vacancy-ordered double perovskite halides such as CsTaCl [18].
Coming back to earlier discussion about the ground state multiplet of systems, when the measured specific heat is used to extract the magnetic entropy, it has been observed in several reports to amount to a value significantly lower than the expected [2, 14, 16]. For BaNaOsO it has been suggested that the missing entropy should be recovered at higher temperatures [14], although the NMR spectra show that the local environment is not distorted above K [12]. In contrast, recent single crystal experiments on BaMgReO [1] indicate a complete recovery, with the magnetic entropy being released all the way up to 80 K, much higher than K. Similarly to the case of BaNaOsO, an ideal cubic environment has been found in high resolution synchrotron diffraction experiment at K [8]. Thus, there still remains an open question of the relationship between the magnetic entropy, ordering and ground state in double perovskites.
In this article we focus our attention on the phonon background in BaMgReO, which is a crucial component for obtaining the magnetic specific heat. There are several ways to estimate the phonon contribution including analytical approaches, first-principle calculations and by measuring a nonmagnetic analog in which the magnetic ion has been replaced with a non magnetic one. Ref. [1] used an analytical approach and showed that the sum of two Debye functions could describe the data above 80 K, although, there was no further clarification for use of these two Debye temperatures (299 K and 796 K) in particular, nor was the optical contribution addressed through Einstein modes. In this work, we determine of the phonon contribution by comparing measurments on newly synthesized powders of BaMgReO and its nonmagnetic analog BaMgWO. We complement our measurements with density-functional calculations of the phonon density of states, from which we extract the phonon specific heat.
Methods
Experimental Details
Polycrystalline samples of BaMgReO and BaMgWO were prepared by a standard solid-state method following published papers [19, 20]. For BaMgReO, stoichiometric amounts of BaO, MgO and ReO were mixed and ground in an agate mortar under a protective argon atmosphere in a glove box. The mixture was placed into silica tubes, evacuated and sealed. The tubes were annealed in a box furnace for 24 hours at 900 °C and cooled at the furnace cooling rate. For BaMgWO, the polycrystalline powder was prepared by repeated annealing and intermediate grinding. Powders of BaCO, MgO and WO were mixed in 2:1:1 ratios and ground in an agate mortar. The mixture was compressed into the pellet and heated up to 850 °C for 24 hours. The initial reaction was followed by two annealing cycles with intermediate grinding and compression into a new pellet. First annealing was done
at 1150 °C for 20 hours and the second one at 1150 °C for 12 hours.
The chemical composition of the polycrystalline samples was determined by powder X-ray diffraction on a Panalytical Empyrean diffractometer equipped with a PIXcel-1D detector, Bragg-Brentano beam optics and utilising a monochromated Cu K source ( Å) at 300 K. The refinements of all samples using the HighScore Plus software were consistent with previous reports on powder samples [19, 20].
Specific heat measurements were performed using the relaxation method in a commercial apparatus (PPMS, Quantum Design). All polycrystalline samples were compressed into pellets with diameter 3.2 mm, thickness under 0.5 mm and mass mg.
Computational Details
Density functional theory (DFT) calculations were performed using a plane wave basis set and projected augmented wave (PAW) potentials as implemented in the Vienna – simulation package (VASP) [21, 22, 23, 24]. We used the following pseudopotentials: Ba (), Mg (), Re (), and O (). The Perdew–Burke–Ernzerhof (PBE) generalized gradient approximation was utilized to account for exchange and correlation, with an effective on-site Hubbard U = 1.8 eV correction applied to the d orbitals of the magnetic ions within the Dudarev scheme[25]. In addition, we included spin-orbit coupling using a full relativistic scheme[26]. An energy cutoff of 600 eV and a gamma-centered point mesh of were used. Electronic and structural optimizations were carried out with energy convergence criteria of eV. The crystal structure including the internal atomic positions were relaxed, while the experimental ratio was constrained. The relaxed lattice parameters are Å and Å. The magnetic structure, canted in-plane antiferromagnetism, was imposed as described in Ref. 15 by constraining the magnetic moments on the Re sites. The vibrational properties of BaMgReO were calculated using the open-source PHONOPY package[27], which constructs a force-constant matrix (calculated by VASP) by displacing symmetry-independent atoms by Å from their equilibrium positions. Since DFT is a zero kelvin technique, phonon and therefore specific heat calculations are only calculated for the tetragonal ground-state crystal structure.
Experimental Results
Phonon contribution to specific heat using the non-magnetic analogue
In Figure 2(a) we present the measured specific heat of BaMgReO and BaMgWO powders. The Re compound exhibits two features reported in previous publications, a sharp peak at K and a broad feature around K. On the other hand the W compound displays a smooth behavior across the investigated range. As BaMgReO and BaMgWO are isostructural compounds, it is reasonable to expect similar phonon contribution of the specific heat as a function of temperature. In order to use the BaMgWO specific heat as estimate of the phonon specific heat in the BaMgReO compound, the W amplitude was scaled by a factor of 1.08 to match the high temperature part of the Re specific heat. There is also a small shift between the two curves along the temperature axis that is the result of different atomic masses of the two compounds, causing dissimilar characteristic Debye temperatures. We dismiss errors in heat capacity measurements that are lower than 0.5%. Minimal amount of low-temperature grease was used with negligible contribution to data inaccuracy. We subtract this renormalized W curve from the Re data, presented in Figure 2(b). The magnetic ordering at and the broad bump around are now emphasized, and the tail of the magnetic contribution can be seen to diminish above K. The magnetic entropy is then obtained by integrating , and the result is presented in Figure 2(c). The saturation value approaches 8 J/Kmol, which is significantly lower than the expected J/Kmol.

Figure 2. (a) Temperature dependence of specific heat divided by temperature measured for BaMgReO powder (red circles), and BaMgWO powder (5d), black circles. The solid blue line represents BaMgWO data scaled by 1.08. (b) Difference between magnetic and non-magnetic sample. (c) Entropy plot obtained from integrated for the BaMgReO powder.
In Figure 3 we plot the magnetic entropy as a function of temperature using different scaling factors for phonon part of specific heat to observe which values do not have physical meaning.

Figure 3. Entropy plot obtained from integrated using phonon background of non-magnetic BaMgWO studied in limits of different normalization factors - from 1.06 to 1.10. Negative slope or rapid increase of entropy in region above 60 K does not have physical meaning. The normalization factor of 1.08 was used as most suitable in Figure 2.
physical meaning. The value of 1.08 is chosen in order to have the high temperature entropy approximately saturated, implying that the magnetic degrees of freedom are engaged only at low temperatures.
Phonon contribution to specific heat using first principles calculations
The calculated phonon specific heat is shown in Figure 4(a), black circles, and presents similar temperature dependence as data. As with the non-magnetic analog, we observe a slight mismatch between the maxima in , in this case arising from two possible factors - one being an error in the calculation of interatomic forces and consequently the dynamic matrix, another is related to the approximation in phonon calculations (in case of DFT, the interatomic forces are calculated at zero kelvin and then phonons are elucidated using a harmonic approximation). Therefore we employ a renormalization factor to match the high temperature experimental data of BaMgReO.

Figure 4. (a) Temperature dependence of specific heat divided by temperature - experimental data for BaMgReO (red) and the phonon contribution for BaMgReO calculated from DFT (black). Solid blue line represents scaled DFT data. (b) shows the difference between experimental BaMgReO and DFT data. (c) Entropy plot obtained from integrating . The differences between two methods in Figure 2 and Figure 4 are very subtle - in case of non-magnetic analogue method, the second anomaly in data is bigger and decreases to zero values slower.
Figure 4(b) displays the subtraction of the experimental data and the calculated phonon background, revealing a similar shape as the in Figure 2(b), with small differences above the magnetic transition - the non-magnetic analogue approach results in a slightly stronger peak and a tail decaying slower towards high temperature. Above 50 K, the magnetic contribution oscillates very closely around value and start to increase slightly at 130 K meaning that the fit of high temperature data is not perfect (the deviation is less than 3% at 200 K). We attribute this mismatch to the discrepancies between maxima of DFT and experimental data at higher temperatures. The resulting entropy in the range up to 120 K, Figure 4(c), saturates above 50 K to a value within 95% of .
Discussion
The difference in extracted utilising two methods in our work illustrates that the determination of the intrinsic phonon background of BaMgReO is not a straight forward process. Both the non-magnetic analog BaMgWO and the DFT calculations have their maximum in slightly shifted towards higher temperatures compared to BaMgReO. This results in an imperfect subtraction, causing values to oscillate around for temperatures above 120 K. Nevertheless, it is evident from the extracted results that the magnetic entropy up to 120 K is significantly below the expected for N=4.
Our results are in general agreement with several other publications where magnetic entropy was extracted in 5^{1}$ systems, covering both Os [14] and Re compounds [2, 28]. This indicates that the choice of the phonon background presented in Ref. [1] should not be regarded as a definitive proof of the total magnetic entropy reaching . Another point to make is the presence of a strong SOC for this family of compounds and structural distortions of ideal octahedron present in these compounds at that might cause further splitting of single ion ground state doublet into two Kramers doublets. As more 5 systems are emerging, it opens possibilities to study this ground state multiplet further, even employing more powerful experimental methods. The determination of the phonon part of heat capacity is non-trivial and require careful consideration.
Summary
We determined the phonon contribution to the heat capacity of BaMgReO by measuring its non-magnetic analog BaMgWO as well as calculating it from first principles. We find that both approaches give a slightly mismatched maximum in , indicating a non-trivial problem of establishing phonon contribution at high temperatures. Both approaches result in total magnetic entropy reaching values significantly lower than , which will hopefully instigate more detailed investigations to resolve this controversial issue.
Acknowledgements
This work was funded by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program projects HERO (Grant No. 810451). Jana Pásztorová acknowledges support from the Federal Commission for Scholarships for Foreign Students for the Swiss Government Excellence Scholarship (ESKAS No. 2019.0041) for the academic year 2019-20.1
References
- [1]D. Hirai and Z. Hiroi. Journal of the Physical Society of Japan, 88:064712, 2019.
- [2]C. A. Marronier, C. M. Thomson, and G. Sala et al. Inorganic Chemistry, 55:10701–10713, 2016.
- [3]A. Krimmel, M. Mucksch, V. Tsurkan, M. M. Koza, H. Mutka, and A. Loidl. Physical Review Letters, 94:237402, 2005.
- [4]C. R. Wiebe, J. E. Greedan, G. M. Luke, and J. S. Gardner. Physical Review B, 65:144413, 2002.
- [5]G. Chen, R. Pereira, and L. Balents. Physical Review B, 82:174440, 2010.
- [6]G. Chen and L. Balents. Physical Review B, 84:094420, 2011.
- [7]A. Paramekanti, D. D. Maharaj, and B. D. Gaulin. Physical Review B, 101:054439, 2020.
- [8]D. Hirai, H. Sagayama, G. Hao, H. Ohsumi, G. Chen, T. Arima, and Z. Hiroi. Physical Review Research, 2:022063, 2020.
- [9]H. Ishikawa et al. Physical Review B, 100:045142, 2019.
- [10]K. E. Stitzer, M. D. Smith, and H. C. Zur Loye. Solid State Science, 4:311–316, 2002.
- [11]Andrew J. Steele, Peter J. Baker, Tom Lancaster, Francis L. Pratt, Isabel Franke, Saman Ghannadzadeh, Paul A. Goddard, William Hayes, D. Prabhakaran, and Stephen J. Blundell. Low-moment magnetism in the double perovskites ba₂moso₆ (m = Li, Na). Phys. Rev. B, 84:144416, Oct 2011.
- [12]L. Lu, M. Song, W. Liu, A.P. Reyes, P. Kuhns, H.O. Lee, I.R. Fisher, and V.F. Mitrovic. Nature Communication, 8:14407, 2017.
- [13]K. Willa, R. Willa, I. R. Fisher U. Welp, A. Rydh, W.-K. Kwok, and Z. Islam. Physical Review B, 100:041108, 2019.
- [14]A. S. Erickson, G. J. Miller S. Misra, R. R. Gupta, and Z. Schlesinger et al. Physical Review Letters, 99:061404, 2007.
- [15]Aria Mansouri Tehrani and Nicola A. Spaldin. Untangling the structural, magnetic dipole, and charge multipolar orders in Ba₂MgReO₆. Physical Review Materials, 5(10):104410, October 2021.DOI
- [16]Victor da Cruz Pinha Barbosa, Jie Xiong, Phuong Minh Tran, Michael A. McGuire, Jiaqiang Yan, Matthew T. Warren, Rolando Valdes Aguilar, Wenjuan Zhang, Mohit Randeria, Nandini Trivedi, Daniel Haskel, and Patrick M. Woodward. The impact of structural distortions on the magnetism of double perovskites containing 5d¹ transition-metal ions. Chemistry of Materials, 34(3):1098–1109, 2022.
- [17]D. Hirai and Z. Hiroi. J. Phys.: Condensed Matter, 33:135603, 2021.
- [18]Aria Mansouri Tehrani, Jian-Rui Soh, Jana Pásztorová, Maximilian E. Merkel, Ivica Živković, Henrik M. Rønnow, and Nicola A. Spaldin. Charge multipoles correlations and order in Cs₂TaCl₆, July 2022. arXiv:2207.10042 [cond-mat].
- [19]D. E. Bugaris, J. P. Hodges, A. Huq, and H. C. Zur Loye. Journal of Solid State Chemistry, 184:2293–2298, 2011.
- [20]K. G. Bramnik, H. Ehrenberg, J. K. Dehn, and H. Fuess. Solid State Science, 5:235–341, 2002.
- [21]G. Kresse and J. Hafner. Ab initio molecular dynamics for liquid metals. Phys. Rev. B, 47(1):558–561, January 1993.
- [22]G. Kresse and J. Furthmüller. Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Phys. Rev. B, 54(16):11169–11186, October 1996.pubmed.ncbi.nlm.nih.gov/9984901
- [23]G. Kresse and D. Joubert. From ultrasoft pseudopotentials to the projector augmented-wave method. Phys. Rev. B, 59(3):1758–1775, January 1999.DOI
- [24]P. E. Blöchl. Projector augmented-wave method. Phys. Rev. B, 50(24):17953–17979, December 1994.pubmed.ncbi.nlm.nih.gov/9976227
- [25]S. L. Dudarev, G. A. Botton, S. Y. Savrasov, C. J. Humphreys, and A. P. Sutton. Electron-energy-loss spectra and the structural stability of nickel oxide: An LSDA+U study. Phys. Rev. B, 57(3):1505–1509, January 1998.DOI
- [26]Soner Steiner, Sergii Khmelevskyi, Martijn Marsmann, and Georg Kresse. Calculation of the magnetic anisotropy with projected-augmented-wave methodology and the case study of disordered Fe₁₋ₓCoₓ alloys. Phys. Rev. B, 93(22):224425, June 2016.DOI
- [27]Atsushi Togo and Isao Tanaka. First principles phonon calculations in materials science. Scr. Mater., 108:1–5, November 2015.arxiv.org/abs/1506.08498
- [28]Hajime Ishikawa, Daigorou Hirai, Akihiko Ikeda, Masaki Gen, Takeshi Yajima, Akira Matsuo, Yasuhiro H. Matsuda, Zenji Hiroi, and Koichi Kindo. Phase transition in the 5d¹ double perovskite ba₂careo₆ induced by high magnetic field. Phys. Rev. B, 104:174422, Nov 2021.DOI