Orbital and spin exchange in LiNiO ₂

Abstract

We study the planar spin exchange couplings in LiNiO2 using a perturbative approach. We show that the inclusion of the trigonal crystal field splitting at the Oxygen sites leads to the appearance of antiferromagnetic exchange integrals in deviation from the Goodenough-Kanamori-Anderson rules for this 90 degree bond. That gives a microscopic foundation for the recently observed coexistence of ferromagnetic and antiferromagnetic couplings in the orbitally-frustrated state of LiNiO2 (F. Reynaud et al, Phys. Rev. Lett. 86, 3638).The compound LiNiO 2 which was first synthesized in 1958 [1] is known to be a good ionic conductor and therefore suitable as a material for rechargeable batteries. Despite its wide use as electrode material, its electronic and magnetic structures are not yet completely understood. Especially intriguing is the absence of any kind of long range magnetic or orbital order [2,3] at low temperature even in the purest samples synthesized up to now. That is especially remarkable since the isostructural compound NaNiO 2 shows orbital order and a collective Jahn-Teller transition at T o = 480 K from the trigonal high-temperature phase to the monoclinic low-temperature one, followed by an antiferromagnetic (AFM) order of ferromagnetic (FM) planes at the Néel temperature of T N = 20 K [4,5]. To explain the strange behavior of LiNiO 2 the proposal of an orbital liquid was pursued in terms of the SU(4) model [6,7,8]. There, a symmetry between orbital and spin degrees of freedom is assumed with equal amplitudes for the corresponding coupling terms. In reality, however, the energy scale for orbital interactions is one order of magnitude larger than those for spin exchange interactions as shown by experimental [3] and theoretical studies [9]. This is also indicated by the difference of T o and T N in NaNiO 2 . More in details, the magnetic susceptibility of LiNiO 2 shows a transition at T of = 400 K towards an orbitally frustrated state at low temperature. Given an orbital disorder that is frozen in, the magnetic properties at low temperature can be phenomenologically explained assuming FM exchange couplings between neighbouring orbitals of different kind (J do = (−6.2 ± 2) meV) and AFM couplings between identical ones (J so = (6.9 ± 2) meV) [3]. However, that contradicts seemingly the known Goodenough-Kanamori-Anderson (GKA) rules [10] which allow only FM couplings for the 90 degree Ni-O-Ni bond in LiNiO 2 [9].The trigonal (rhombohedral) crystal structure R3m of LiNiO 2 can be understood by starting from the cubic situation with oxygen and Ni/Li on the sites of the cubes, and with the cubic axesx,ỹ, andz. Perpendicular to the cube diagonal z =x +ỹ +z in fig. 1 one finds alternating planes of Li, Ni, and O. The electronically active NiO 2 layer (see fig. 1) contains a triangular lattice of the magnetic Ni ions. The O 2− ions have a completely filled 2p shell, whereas the Ni 3+ ion is in the low-spin electronic configuration (t 6 2g e 1 g ) with spin 1/2 [3]. The trigonal distortion changes the bond angle fro...

Decoupling of orbital and spin degrees of freedom in Li1−xNaxNiO2

“…Our experimental observations support previous analysis 10 and recent theoretical calculations 11 showing that, in the ϳ90°TM-O-TM systems, orbitals and spins are essentially decoupled. Departure from the ideal 90°angle, and the contribution from extra overlappings 9 or crystal-field splitting effects 12 have been invoked to strengthen the coupling between spins and orbitals in these systems, but the results reported here indicate that those hypotheses are not appropriate in the case of LiNiO 2 .…”

“…It presents a long-range AFM order below 20 K. 2,3 It has been speculated that a different oxygen crystal-field splitting would inhibit these AFM in-plane Ni-Ni interactions in NaNiO 2 , yielding its different magnetic properties. 12 We perform here a crystallographic and magnetic study of LiNiO 2 , NaNiO 2 , and Li 0.3 Na 0.7 NiO 2 . 14 Our main point is that samples with this intermediate composition do not undergo the JT transition of NaNiO 2 , but keep a rhombohedral structure like LiNiO 2 even at low temperature.…”

Orbital and Spin Order in the Triangular S=1/2 Layered Compound (LI,NA)NIO2

“…The interesting calculation of Daré et al [34] suggests that a large enough trigonal distortion can generate antiferromagnetic interactions in the Ni planes in addition to the ferromagnetic interactions derived from the Goodenough-Kanamori-Anderson rules. This could be important since the coexistence of ferromagnetic and antiferromagnetic interactions could explain the lack of magnetic ordering in LiNiO 2 .…”

Ferrodistorsive orbital ordering in the layered nickelate NaNiO2: A density-functional study

“…The 90°Ni-O-Ni exchange within the layer has been examined in great detail in two recent papers. 13,14 For the exchange between the layers, we treat the Ni-O-Na hopping as an effective Ni-Na hopping between the Ni͑3z 2 − r 2 ͒ and Na͑s͒ orbitals on the Ni-O-Na-O-Ni superexchange path. The model Hamiltonian for the effective NiNa-Ni path then reads…”