The first-order magnetoelastic transition in Eu2In: A 151Eu Mössbauer study

Our Eu Mössbauer investigation of Eu2In and Eu2Sn shows that the europium in both materials is fully divalent. We confirm the distinct thermodynamic orders of the magnetic transitions and reveal a remarkable difference between the magnetic environments of the europium atoms in the two compounds. Possible structural and electronic origins for these differences are discussed using DFT calculations. © 2019 Author(s). All article content, except where otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). https://doi.org/10.1063/1.5129683., s


I. INTRODUCTION
First-order magnetic transitions (FOMTs) with large differences in magnetisation are important in the development of advanced functional magnetic materials, e.g. magnetocalorics, and isosymmetric magnetoelastic FOMTs that do not perturb the crystal symmetry are especially rare. Eu 2 In exhibits a particularly sharp ferromagnetic (FM) to paramagnetic (PM) transition at Tc=55 K (FWHM < 1 K) with a remarkable combination of large latent heat and magnetocaloric response, small volume discontinuity, and negligible hysteresis. 1 By contrast the nearly isostructural Eu 2 Sn exhibits a conventional second-order antiferromagnetic transition at TN=31 K.
Here we study the differences in local magnetic behaviour between Eu 2 In and Eu 2 Sn using 151 Eu Mössbauer spectroscopy and band structure calculations. We confirm the different thermodynamic nature of the transitions and find that the magnetic environments of the europium in the two isostructural compounds are completely different.
Magnetic measurements were carried out in a 7 T magnetic property measurement system MPMS (Quantum Design) magnetometer equipped with a reciprocating sample option (RSO).
Samples were prepared for Mössbauer spectroscopy by hand grinding under hexane in a nitrogen-filled glove-box to reduce the possibility of oxidation. However as will be seen in ARTICLE scitation.org/journal/adv the spectra presented below, some oxidation occurred during mounting. The 151 Eu Mössbauer spectroscopy measurements were carried out using a 4 GBq 151 SmF 3 source, driven in sine mode and calibrated using a standard 57 CoRh/α-Fe foil. Isomer shifts are quoted relative to EuF 3 at ambient temperature. The 21.6 keV gamma rays were recorded using a thin NaI scintillation detector. The sample was cooled in a vibration-isolated closed-cycle helium refrigerator with the sample in a helium exchange gas. The spectra were fitted to a sum of Lorentzian lines with the positions and intensities derived from a full solution to the nuclear Hamiltonian. 3

III. RESULTS
The magnetisation vs. temperature shown in Fig. 1 serves to emphasise the remarkable difference between the two materials. The FOMT in Eu 2 In is marked by a large discontinuous step in the magnetisation at 55 K while Eu 2 Sn exhibits a cusp at 31 K marking a conventional second-order transition.
The 151 Eu Mössbauer spectrum of Eu 2 Sn at 5.6 K shown in Fig. 2 is relatively simple, consisting of a single magnetic pattern, despite there being two crystallographically distinct sites occupied by the europium. The isomer shift of −9.71(3) mm/s is consistent with fully divalent europium, however the hyperfine field (B hf ) is remarkably small, being only 12.7(1) T. A linear fit to the high temperature susceptibility yields an effective paramagnetic moment of 7.94μB 1 and DFT calculations give europium moments of 7μB at both sites. 1 If the europium does indeed carry the full 7μB expected for Eu 2+ then the observed B hf of 12.7(1) T in Eu 2 Sn may be the smallest in any metallic Eu 2+ system, and values of 20 T -40 T are more typical. 4 The thermal evolution of the spectra is unremarkable (Fig. 3) and B hf (T) is well fitted using a J= 7 2 Brillouin function, yielding TN=30.4(1) K, fully consistent with the cusp observed in M(T).
By contrast, the 151 Eu Mössbauer spectra of Eu 2 In shown in Fig. 4  components present with hyperfine fields of 27.2(1) T and 16.8(2) T. Not only are the fields at the two sites remarkably different, they are also much larger than the single site field seen in Eu 2 Sn. Saturation magnetisation measurements give 7.2 μB/Eu while DFT calculations give 7μB/Eu. 1 These results combined with the observed isomer shift of −9.01(4) mm/s all point to fully divalent europium being present in Eu 2 In. The evolution of the spectra with temperature confirms the first-order nature of the transition at 55 K. As can be seen in Fig. 3, B hf (T) at both sites declines slowly as the temperature is increased, and by 55 K they are down by about 25%, then they vanish at the FOMT. Fitting the temperature dependence to a J= 7 2 Brillouin function suggests that the low temperature ferromagnetic phase would have an ordering temperature of ∼80 K if the first-order structural transition did not occur. In addition, the Eu 2+ component exhibits a small (2-3 T) broadening above the FOMT that extends to about 100 K suggesting that some weak residual order persists well above the FOMT. This is consistent with the pedestal apparent in M(T) above the FOMT in Fig. 1.

IV. DISCUSSION
Given that x-ray absorption spectroscopy (XAS) demonstrates purely divalent state of Eu in Eu 2 In, 1 the significant (17(1)%) Eu 3+ contribution seen in Fig. 4 near 0 mm/s, which is much larger than the detection limit of XAS, is attributed to an impurity phase, likely the sesquioxide, Eu 3+ hydroxide, or a mixture of both, formed during grinding and mounting. The Eu 3+ impurity is much more limited in Eu 2 Sn (4(1)%), indicating its lower reactivity with moist air. The present Mössbauer data thus provide an additional experimental confirmation of Eu 2+ state in Eu 2 Sn.
The evolution of the hyperfine fields (Fig. 3) in Eu 2 In and Eu 2 Sn is markedly different, with the former exhibiting clear discontinuities for two distinct europium sites. These differences are in agreement with earlier density functional theory (DFT) calculations that show much reduced density of states (DOS) at the Fermi level, EF, as well as lower induced moments on Eu 6s, 6p, and 5d states of Eu 2 Sn. 1 Replacement of In with Sn in the Co 2 Si-type structure shifts the total DOS further below the Fermi level leading to semimetallic-like topology at EF (Fig. 5). Consequently, negligible spin polarization due to lack of electronic states at EF does not support ferromagnetism in Eu 2 Sn in either the Co 2 P-type 1 or Co 2 Si-type (this study) structures.
For Eu 2 In the ratios of the lattice parameters a/c ∼0.723 and b/a ∼0.749 are typical of other rare earth intermetallics adopting the Co 2 Si-type structure. On the other hand, a/c ∼0.690 of Eu 2 Sn is closer to that of the Co 2 P-type structure. 5 Further, one of the Eu(II)−Sn bond distances extends by nearly 10% compared to its analog in Eu 2 In (4.463Å vs. 4.071Å, respectively) and the bond is essentially broken in Eu 2 Sn, which provides an additional reason to treat the Co 2 Si and Co 2 P structure types differently. 5 These anisotropic differences in lattice parameters and atomic arrangements lead to modified local environments around the Eu atoms, although most of the changes are minor. For example, the average Eu−Eu interatomic distance in the Eu(I) nearest neighbor environment is shorter by 0.3% in Eu 2 In (3.871Å) compared to Eu 2 Sn (3.885Å), even though the unit cell volume of the former is larger by 2.4% compared to the latter. Crystallographic differences together with the modifications of the electronic structure are, therefore, responsible for the variation of magnetic behavior of Eu 2 In and Eu 2 Sn.
Analysis of the differences in the interatomic bonding between Eu(I) and Eu(II) in Eu 2 In reveals closer Eu(I)−Eu(I) (

V. CONCLUSIONS
While 151 Eu Mössbauer spectroscopy shows that the europium in both Eu 2 In and Eu 2 Sn is fully divalent and confirms the first-order (Eu 2 In) and second-oder (Eu 2 Sn) natures of the respective magnetic transitions, It also reveals some striking differences in the ordered states of these two compounds. Eu 2 In exhibits two clearly distinct Eu sites with hyperfine fields that differ by more than 50%. By contrast not only are the two sites not resolved in Eu 2 Sn, but the observed hyperfine field is remarkably small. The isomer shifts for the two compounds are also distinct, −9.01(4) mm/s and −9.71(3) mm/s for In and Sn respectively, pointing to significant differences in the local density of states around the europium in these compounds.
Neutron diffraction measurements are needed both to establish the actual magnetic structures of these compounds and to determine the moments on the two europium sites in each one.