Fe4Si2Sn7O16 has a complex but well-ordered composite structure that incorporates perfectly hexagonal kagomé lattices of high-spin Fe2+ (S = 2) separated by non-magnetic Sn4+. Below TN = 3.5 K, conventional crystallographic analysis of the Bragg intensity in neutron powder diffraction data show that Fe4Si2Sn7O16 adopts a unique k = (0, ½, ½) “striped” magnetic structure [1,2] that breaks the hexagonal symmetry of the kagomé lattice with partial long-range order on 2/3 of the HS Fe2+ sites in “stripes”. The other 1/3 show no evidence of long-range order, suggesting a highly unusual “partial spin-liquid” state.
Here we will present new insights obtained from quantitative fitting of diffuse magnetic scattering to short-range magnetic spin ordering in Fe4Si2Sn7O16. Polarised neutron diffraction data collected on D7 at the ILL show that 1/3 of the paramagnetic Q = 0 magnetic scattering above TN persists as structured diffuse scattering down to at least 38 mK, consistent with the partially ordered crystallographic average structure. We used a Monte Carlo “big-box” magnetic approach with the program Spinvert [3] to show that the first-, second- and third-nearest neighbour magnetic interactions (J1, J2 and J3) are all predominantly antiferromagnetic (AFM) above TN, with only J2 switching to ferromagnetic (FM) below TN in a compromise forced by the triangular nature of the kagomé lattice. A spin-Hamiltonian approach using the program Spinteract [4] independently gave AFM J1, J2 and J3 above TN.
Ab initio (DFT) calculations using the high-level meta-GGA SCAN [5] functional further supported these signs of the short-range magnetic interactions, which explain why Fe4Si2Sn7O16 does not adopt either of the conventional long-range-ordered kagomé states that preserve hexagonal symmetry: k = 0 “in-out”, in which third-nearest neighbours are strictly FM; or k = √3 × √3 “spiral”, in which second-nearest neighbours are strictly FM. The partially ordered k = (0, ½, ½) ground state of Fe4Si2Sn7O16, in which there are no strictly FM relationships, appears to be the best compromise when J1-3 are all strongly AFM in character. We will discuss our experimental result in the context of the very recent theoretical prediction that the partially ordered ground state will emerge when J1 is FM and J2-3 are AFM.