Interest in the study of the lanthanides has recently been enhanced by the observation in some of their compounds of unconventional superconductivity appearing in the vicinity of a quantum critical point, where the magnetic order is destroyed. The originality of samarium, within the lanthanides, lies in two facts: i) the competition between its two stable valence states, nonmagnetic Sm2+ and magnetic Sm3+, can lead to the formation of intermediate valence compounds, showing peculiar electronic properties [1], and ii) the presence of low-energy excited multiplets for Sm3+ can lead to a ground state which is not an eigenstate of the total angular momentum J (J = 5/2 for the ground multiplet). Moreover, the charge state can be tuned towards trivalency by the application of pressure, with the expected onset of long-range magnetic order owing to the fact that Sm3+ has an odd number of 4f electrons (e.g., it is a Kramers ion). We have investigated these aspects by 149Sm Nuclear Forward Scattering (NFS) of synchrotron radiation on beamlines ID18 and ID22N.

Fig. 8: Temperature dependence of the magnetic hyperfine field Bhf and of the electric quadrupole interaction EQ of SmMn2Ge2. The full circles refer to the magnetic component while the full triangles refer to the nonmagnetic one, which is present only for T > 100 K.

An intriguing intermetallic compound of samarium is SmMn2Ge2, which has two magnetic sublattices (Mn and Sm) and shows multiple magnetic phase transitions with temperature (ferromagnetic for T < 100 K and 150 < T < 350 K, antiferromagnetic for 100K < T < 150 K and 350 K < T < 385 K, and paramagnetic for T > 385 K) [2]. The role of samarium in these transitions is not clear, because of difficulties in separating its small magnetic moment from that of Mn by neutron diffraction. Our 149Sm NFS investigations were performed on beamline ID18. Figure 8 shows the temperature dependence of the hyperfine magnetic field and of the electric quadrupole interaction at the 149Sm nuclei. These results show that Sm is in a trivalent state with fairly pure J = 5/2 character (magnetic moment ~ 0.65 µB), contrary to most other Sm compounds. They give the first direct evidence that samarium has an ordered magnetic moment at all temperatures between 4 and 250 K, spanning the three main magnetic phases, and rule out the possibility that the antiferromagnetic ferromagnetic phase transition at 100 K could be due to the ordering of the samarium sublattice. However, above 100K, only about 50% of the samarium ions keep the ordered moment, the rest becoming paramagnetic. This suggests that the ordering of the 4f moments in this temperature range is favoured by the non-vanishing molecular field created by the Mn moments at about half of the lattice sites occupied by Sm.

Examples of intermediate valence Sm compounds are the high pressure ‘golden’ phases of the Sm monochalcogenides (SmS, SmSe and SmTe) and SmB6. They belong to the class of Kondo insulators or narrow-gap semiconductors, which behave at high temperature like an array of independent localised moments interacting with itinerant conduction electrons, whereas at low temperature they develop clear narrow-gap properties. Pressure was shown to increase the Sm valency in all of these compounds, so it was guessed that a trivalent state could be reached, where magnetic order would appear at low temperatures. However this order could never be observed. We have started a systematic investigation of the correlations between magnetism, electrical transport and valence in these compounds. The NFS measurements as a function of temperature and pressure on SmS and SmB6 powders were performed on beamline ID22N. Our results are the first direct proof that pressure induces a magnetically ordered state for both compounds. This order likely appears when the semiconducting gap closes, for pressures which are unexpectedly much lower than those required to reach the fully trivalent state. The importance of short range magnetic correlations in the semiconducting phases could also be evidenced thanks to the use of a local probe like NFS. The phase diagrams which could be established for the two compounds by our and previous measurements are shown in Figure 9.

Fig. 9: The phase diagrams of SmS and SmB6 as a function of temperature and pressure, with the magnetically-ordered phases as discovered by Nuclear Forward Scattering. Legend: SRMO = short-range magnetic order, LRMO = long-range magnetic order, S = semiconductor. The shaded region between 50 and 100 K in the phase diagram of SmB6 indicates the smooth transition between the SRMO and the paramagnetic phases.

Our investigations show that NFS can be an invaluable tool for the study of magnetism of strongly correlated electron systems, thanks to its high sensitivity and its compatibility with the diamond anvil cell high pressure technique.



[1] P. Wachter, Handbook on the Physics and Chemistry of Rare Earths, vol. 19, ed. K. A. Gschneidner et al. (Amsterdam: North-Holland), p. 383 (1994).
[2] S. Chaudhary, M.K. Chattopadhyay, K.J. Singh, S.B. Roy, P. Chaddah, and E.V. Sampathkumaran, Phys. Rev. B 66, 014424 (2002)

Principal Publication and Authors

A. Barla (a,b), J. Derr (a), J.P. Sanchez (a), B. Salce (a), G. Lapertot (a), B.P. Doyle (d), R. Rüffer (b), R. Lengsdorf (c), M.M. Abd-Elmeguid (c) and J. Flouquet (a), Phys. Rev. Lett. 94, 166401 (2005).
(a) Département de Recherche Fondamentale sur la Matière Condensée, CEA, Grenoble (France)
(b) ESRF
(c) II. Physikalisches Institut, Universität zu Köln, Cologne (Germany)
(d) Laboratorio TASC-INFM, Trieste (Italy)