Neutron scattering study of crystal fields inCeRhIn5

Abstract

Neutron scattering results for the tetragonal compound CeRhIn5 give evidence for two crystal field (CF) excitations at 6.9 and 23.6 meV. The scattering can be fit assuming a set of CF parameters B 0 2 = -1.03 meV, B 0 4 = 0.044 meV and B 4 4 = 0.122 meV. To compare our results to previous work, we calculate the susceptibility and specific heat for this CF scheme, including a molecular field term λ =35 mol/emu to account for the Kondo effect. We also include a calculation based on these CF parameters that uses the non-crossing approximation to the Anderson model to estimate the effect of Kondo physics on the susceptibility, specific heat and neutron linewidths. PACS numbers: 75.30.Mb 75.20.Hr 71.27.+a 71.28.+d 61.10.Ht CeRhIn 5 crystallizes in the same tetragonal HoCoGa 5 structure as the heavy fermion superconductors CeIrIn 5 and CeCoIn 5 1,2,3 . At ambient pressure CeRhIn 5 undergoes a transition to an antiferromagnetic (AF) state at T N = 3.8 K 1,4 . With application of hydrostatic pressure the Néel temperature remains essentially constant until antiferromagnetism disappears and superconductivity appears at pressures above 15 kbar 1 . Recently, Pagliuso et al. 5 have suggested the importance of CF splitting to the ground state properties of the CeMIn 5 family of heavy fermion superconductors, underscoring the fact that the ultimate ground state achieved by a particular member of the family must grow out of the ground state crystal field doublet. Thus a careful determination of both the CF splitting and wavefunctions is important. To that end we have begun to directly probe the CF energy level splitting in the CeMIn 5 family using inelastic (IE) neutron scattering. The first step in our investigations has been determining the crystal field level scheme in CeRhIn 5 .In CeRhIn 5 , as in the other members of CeMIn 5 family, the crystal field Hamiltonian in tetragonal symmetry can be writtenwhere O m l and B m l are the Stevens operators and CF parameters respectively. The Ce 3+ J = 5/2 wavefunction splits into three doublets, Γ (1) 7 = {α[±5/2 + β[∓3/2 }, Γ (2) 7 = {β[±5/2 − α[∓3/2 } and Γ 6 = [±1/2 5 . An analysis of susceptibility and thermal expansion results 6 suggested crystal field levels Γ (2) 7 , Γ (1) 7and Γ 6 at E=0, 5.86 meV (68 K) and 28.43 meV (300 K) respectively, with β = 0.969 (yielding a nearly pure [±5/2 ground state). A subsequent study 5 based on an analysis of the susceptibility and specific heat suggested a similar scheme, but with splittings 6 and 12 meV (70 and 140 K). In this paper we report the results of an analysis of neutron scattering data for CeRhIn 5 which indicate that these initial estimates are nearly correct; our results have somewhat different values for the splittings and a smaller value for the mixing parameter β, i.e., a greater admixture of [∓3/2 into the [±5/2 ground state. To assist in comparison of our results to those of Pagliuso et al. 5 and Takeuchi et al. 6 , we report calculations of the specific heat and magnetic susceptibility based on our CF parameters which in...

High-magnetic-field thermal expansion and elastic properties ofCeRhIn5

“…For B = 18 T, the separation between the two lowest singlet states is ∆ S 1 (18 T) ∼ 21 K. Hence, for the whole field range we can access (up to 18 T), ∆ S 1 (B) < T ef f K ∼ 25 K (Ref. [11]), and both states overlap. Moreover, the expected value of the magnetic moment gµ B J z along B (= 0.92 µ B for zero field) for both the ground and first excited singlet states are amazingly different: 1.52 and 0.02 µ B , respectively.…”

“…A peak in the INS spectra related to transitions from the ground state doublet to the second excited doublet (forbidden if η = 0), is unambigously detected. 11 The other contribution that we are perhaps neglecting is the influence of the Kondo interaction on the CEF effects and vice versa. The Kondo scattering is reduced by the CEF splitting giving rise to an effective Kondo temperature T ef f K for the ground doublet state that can be much lower than the Kondo temperature T K for the whole J = 5/2 multiplet.…”

“…9,10,11,12,16,18 Figure 5 shows α c at zero field and 18 T along the caxis in an extended temperature range. As previously observed 12 in Ce 0.6 La 0.4 RhIn 5 , the CEF-related negative peak around 25 K moves towards higher T when the field is increased.…”

“…This is an anomalous behavior if we consider the CEF scheme obtained from inelastic neutron scattering measurements. 11 According to these data, the CEF induced splitting of the Ce 3+ J = 5/2 multiplet results in a Γ (2) 7 = (1 − η 2 ) | ± 5/2 − η | ∓ 3/2 ground state doublet with a mixing parameter η = 0.6. However, as shown in Ref.…”

“…10,19 On the other hand, as a result of the hybridization between the conduction and the f -electrons, the CEF levels become dispersive bands whose width is T ef f K (Ref. [11]). According to the CEF scheme from INS results, a magnetic field splits the three doublets and mixes together the ground state and the first excited doublet.…”