betaFeedback

WorksPhysics and Astronomy

Experimental and theoretical thermodynamic studies in Ba2_2MgReO6_6 – the ground state in the context of Jahn-Teller effect

Abstract

We address the degeneracy of the ground state multiplet on the 5d1d^1 Re6+^{6+} ion in double perovskite Ba2_{2}MgReO6_{6} using a combination of specific heat measurements and density functional calculations. For Ba2_{2}MgReO6_{6}, two different ground state multiplets have previously been proposed - a quartet (with degeneracy NN=4) [1] and a doublet (NN=2) [2]. Here we employ two independent methods for the estimation of phonon contribution in heat capacity data to obtain the magnetic entropy SmagS_{mag}, which reflects the degeneracy of the ground state multiplet NN through Smag=RS_{mag}=RlnNN. In both cases, we obtain a better fit to Smag=RS_{mag}=Rln2 indicating evidence of NN=2 degeneracy in the range from 2 to 120~K. The detailed nature of the ground state multiplet in Ba2_{2}MgReO6_{6} 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 ZZ, it is important in materials containing 5d5d 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 5d5d elements Ta, Re and Os with formally single electron occupying their dd orbitals (5d15d^1). These have general chemical formula A2BB′O6A_2BB' O_6, with the magnetic B′B' ion surrounded by six anions, creating a highly symmetric octahedron. The octahedral crystal electronic field (CEF) splits the five-fold degenerate dd orbitals into characteristic ege_g and t2gt_{2g} orbital sets (see Figure 1). Under the influence of SOC the three-fold degenerate t2gt_{2g} set (with an effective orbital momentum leff=−1l_{\rm eff}=-1) is split further into a lower-lying quartet (jeff=3/2j_{\rm eff}=3/2) and a doublet (jeff=1/2j_{\rm eff}=1/2) at higher energy. Up to this point, the single ion ground state multiplet is the quartet (N=4N=4) proposed in [1] paper. However, further investigation of some double perovskites A2BB′O6A_2BB'O_6 showed the structural distortion from cubic to tetragonal unit cell and elongation or compression of the ideal B′B'-O6_6 octahedron along cc-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 N=2N=2, and which translates into Rln⁡2R\ln2 value of entropy.

Proposed energy level splitting scheme for 5d1 electronic configuration in double perovskites

Figure 1. Proposed energy level splitting scheme for 5d15d^1 electronic configuration in DPs. Elongation of ideal Re-O6_6 octahedron along c-axis [8], known as Jahn-Teller (J-T) effect, lifts degeneracy of dd orbitals and splits original jeff=3/2j_{\mathrm{eff}}=3/2 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 (JFMJ_{FM}) and antiferromagnetic (JAFMJ_{AFM}) type and leading to a long-range magnetic order at TmT_m [5]. Additionally, it has been suggested that 5d15d^1 systems could allow investigation of higher-order multipoles due to their small dipole magnetic moments [5]. For an electronic quadrupolar interaction VV that is large enough, the prediction of a quadrupolar order at Tq>TmT_q>T_m has been made [5]. This causes a small lattice distortion from a perfect cubic environment.

Several 5d15d^1 double perovskites have been previously investigated in this context. In Ba2_2NaOsO6_6, where the Os ion is in a formally 7+ valence state, a long-range magnetic ordering has been found to occur around Tm=7T_m=7 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 TmT_m [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 B=LiB=\mathrm{Li} 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 BB site occupied by 2+ ions. For B=MgB=\mathrm{Mg}, TmT_m 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 Ba2_2ZnReO6_6 (Tm=16T_m=16 K, Tq=23T_q=23 K) [16] and Ba2_2CdReO6_6 (Tm=4T_m=4 K, Tq=25T_q=25 K) [17], pointing to a universal behavior across the family. Additionally, charge quadrupoles have also been recently discussed in the more ionic vacancy-ordered 5d15d^1 double perovskite halides such as Cs2_2TaCl6_6 [18].

Coming back to earlier discussion about the ground state multiplet of 5d15d^1 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 Rln⁡4R\ln4 [2, 14, 16]. For Ba2_2NaOsO6_6 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 ∼12\sim12 K [12]. In contrast, recent single crystal experiments on Ba2_2MgReO6_6 [1] indicate a complete Rln⁡4R\ln4 recovery, with the magnetic entropy being released all the way up to 80 K, much higher than Tq=33T_q = 33 K. Similarly to the case of Ba2_2NaOsO6_6, an ideal cubic environment has been found in high resolution synchrotron diffraction experiment at T>Tq=33T > T_q = 33 K [8]. Thus, there still remains an open question of the relationship between the magnetic entropy, ordering and ground state in 5d15d^1 double perovskites.

In this article we focus our attention on the phonon background in Ba2_2MgReO6_6, 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 Ba2_2MgReO6_6 and its nonmagnetic analog Ba2_2MgWO6_6. 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 Ba2_2MgReO6_6 and Ba2_2MgWO6_6 were prepared by a standard solid-state method following published papers [19, 20]. For Ba2_2MgReO6_6, stoichiometric amounts of BaO, MgO and ReO3_3 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 Ba2_2MgWO6_6, the polycrystalline powder was prepared by repeated annealing and intermediate grinding. Powders of BaCO3_3, MgO and WO3_3 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 Kα1,2\alpha_{1,2} source (λ1=1.5406\lambda_1 = 1.5406 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 ∼10\sim10 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 abab–initioinitio simulation package (VASP) [21, 22, 23, 24]. We used the following pseudopotentials: Basv_{\mathrm{sv}} (5s25p66s25s^2 5p^6 6s^2), Mgpv_{\mathrm{pv}} (2p63s22p^6 3s^2), Repv_{\mathrm{pv}} (5p65d56s25p^6 5d^5 6s^2), and O (2s22p42s^2 2p^4). 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 kkpoint mesh of 6×6×66 \times6 \times6 were used. Electronic and structural optimizations were carried out with energy convergence criteria of 1×10−61 \times10^{-6} eV. The crystal structure including the internal atomic positions were relaxed, while the experimental c/ac/a ratio was constrained. The relaxed lattice parameters are a=5.8a = 5.8 Å and c=8.2c = 8.2 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 Ba2_2MgReO6_6 were calculated using the open-source PHONOPY package[27], which constructs a force-constant matrix (calculated by VASP) by displacing symmetry-independent atoms by ±0.01\pm0.01 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 Ba2_2MgReO6_6 and Ba2_2MgWO6_6 powders. The Re compound exhibits two features reported in previous publications, a sharp peak at Tm=18T_m = 18 K and a broad feature around Tq=33T_q = 33 K. On the other hand the W compound displays a smooth behavior across the investigated range. As Ba2_2MgReO6_6 and Ba2_2MgWO6_6 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 Ba2_2MgWO6_6 specific heat as estimate of the phonon specific heat in the Ba2_2MgReO6_6 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 TmT_m and the broad bump around TqT_q are now emphasized, and the tail of the magnetic contribution can be seen to diminish above ∼60\sim60 K. The magnetic entropy is then obtained by integrating Smag=∫(Cmag/T)dTS_{\mathrm{mag}} = \int(C_{\mathrm{mag}}/T)\mathrm{d}T, and the result is presented in Figure 2(c). The saturation value approaches 8 J/Kmol, which is significantly lower than the expected Rln⁡4=11.5R\ln4 = 11.5 J/Kmol.

Temperature dependence of specific heat and magnetic entropy for Ba2MgReO6 and Ba2MgWO6

Figure 2. (a) Temperature dependence of specific heat divided by temperature C/TC/T measured for Ba2_2MgReO6_6 powder (red circles), and Ba2_2MgWO6_6 powder (5d0^0), black circles. The solid blue line represents Ba2_2MgWO6_6 data scaled by 1.08. (b) Difference Cmag/TC_{\mathrm{mag}}/T between magnetic and non-magnetic sample. (c) Entropy plot obtained from integrated Cmag/TC_{\mathrm{mag}}/T for the Ba2_2MgReO6_6 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.

Entropy plot obtained from integrated magnetic specific heat

Figure 3. Entropy plot obtained from integrated Cmag/TC_{\mathrm{mag}}/T using phonon background of non-magnetic Ba2_2MgWO6_6 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 C/TC/T, 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 Ba2_2MgReO6_6.

Temperature dependence of specific heat, magnetic contribution, and magnetic entropy

Figure 4. (a) Temperature dependence of specific heat divided by temperature C/TC/T - experimental data for Ba2_2MgReO6_6 (red) and the phonon contribution for Ba2_2MgReO6_6 calculated from DFT (black). Solid blue line represents scaled DFT data. (b) Cmag/TC_{\mathrm{mag}}/T shows the difference between experimental Ba2_2MgReO6_6 and DFT data. (c) Entropy plot obtained from integrating Cmag/TC_{\mathrm{mag}}/T. 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 Cm/TC_{\mathrm{m}}/T 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 Cmag/TC_{\mathrm{mag}}/T in Figure 2(b), with small differences above the magnetic transition - the non-magnetic analogue approach results in a slightly stronger TqT_q peak and a tail decaying slower towards high temperature. Above 50 K, the magnetic contribution Cmag/TC_{\mathrm{mag}}/T oscillates very closely around Cmag/T=0C_{\mathrm{mag}}/T=0 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 C/TC/T 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 Rln⁡2R\ln2.

Discussion

The difference in extracted SmagS_{\mathrm{mag}} utilising two methods in our work illustrates that the determination of the intrinsic phonon background of Ba2_2MgReO6_6 is not a straight forward process. Both the non-magnetic analog Ba2_2MgWO6_6 and the DFT calculations have their maximum in C/TC/T slightly shifted towards higher temperatures compared to Ba2_2MgReO6_6. This results in an imperfect subtraction, causing Cmag/TC_{\mathrm{mag}}/T values to oscillate around Cmag/T=0C_{\mathrm{mag}}/T=0 for temperatures above 120 K. Nevertheless, it is evident from the extracted results that the magnetic entropy up to ∼\sim 120 K is significantly below the expected Rln⁡4R\ln4 for N=4.

Our results are in general agreement with several other publications where magnetic entropy was extracted in 5dd^{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 Rln⁡4R\ln4. 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 TqT_q that might cause further splitting of single ion ground state doublet into two Kramers doublets. As more 5d1d^1 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 Ba2_2MgReO6_6 by measuring its non-magnetic analog Ba2_2MgWO6_6 as well as calculating it from first principles. We find that both approaches give a slightly mismatched maximum in C/TC/T, indicating a non-trivial problem of establishing phonon contribution at high temperatures. Both approaches result in total magnetic entropy reaching values significantly lower than Rln⁡4R\ln4, 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. [1]D. Hirai and Z. Hiroi. Journal of the Physical Society of Japan, 88:064712, 2019.
  2. [2]C. A. Marronier, C. M. Thomson, and G. Sala et al. Inorganic Chemistry, 55:10701–10713, 2016.
  3. [3]A. Krimmel, M. Mucksch, V. Tsurkan, M. M. Koza, H. Mutka, and A. Loidl. Physical Review Letters, 94:237402, 2005.
  4. [4]C. R. Wiebe, J. E. Greedan, G. M. Luke, and J. S. Gardner. Physical Review B, 65:144413, 2002.
  5. [5]G. Chen, R. Pereira, and L. Balents. Physical Review B, 82:174440, 2010.
  6. [6]G. Chen and L. Balents. Physical Review B, 84:094420, 2011.
  7. [7]A. Paramekanti, D. D. Maharaj, and B. D. Gaulin. Physical Review B, 101:054439, 2020.
  8. [8]D. Hirai, H. Sagayama, G. Hao, H. Ohsumi, G. Chen, T. Arima, and Z. Hiroi. Physical Review Research, 2:022063, 2020.
  9. [9]H. Ishikawa et al. Physical Review B, 100:045142, 2019.
  10. [10]K. E. Stitzer, M. D. Smith, and H. C. Zur Loye. Solid State Science, 4:311–316, 2002.
  11. [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. [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. [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. [14]A. S. Erickson, G. J. Miller S. Misra, R. R. Gupta, and Z. Schlesinger et al. Physical Review Letters, 99:061404, 2007.
  15. [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. [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. [17]D. Hirai and Z. Hiroi. J. Phys.: Condensed Matter, 33:135603, 2021.
  18. [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. [19]D. E. Bugaris, J. P. Hodges, A. Huq, and H. C. Zur Loye. Journal of Solid State Chemistry, 184:2293–2298, 2011.
  20. [20]K. G. Bramnik, H. Ehrenberg, J. K. Dehn, and H. Fuess. Solid State Science, 5:235–341, 2002.
  21. [21]G. Kresse and J. Hafner. Ab initio molecular dynamics for liquid metals. Phys. Rev. B, 47(1):558–561, January 1993.
  22. [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. [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. [24]P. E. Blöchl. Projector augmented-wave method. Phys. Rev. B, 50(24):17953–17979, December 1994.pubmed.ncbi.nlm.nih.gov/9976227
  25. [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. [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. [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. [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

Paper details

vlnc:W_8nsjdmzy
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

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

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents