Experimental Realization of a Spin-1/2Triangular-Lattice Heisenberg Antiferromagnet

Abstract

We report the results of magnetization and specific heat measurements on Ba3CoSb2O9, in which the magnetic Co 2+ ion has a fictitious spin-1 2, and show evidence that a spin-Heisenberg antiferromagnet on a regular triangular lattice is actually realized in Ba3CoSb2O9. We found that the entire magnetization curve including the one-third quantum magnetization plateau is in excellent agreement with theoretical calculations at a quantitative level.Exploring the ground state of a frustrated quantum magnet has been one of the main subjects of condensed matter physics 1-3 . A long theoretical debate reached a consensus that a two-dimensional (2D) spin-1 2 TLHAF has an ordered ground state of the 120• spin structure with an extremely small sublattice magnetization 4-7 . Although the zero-field ground state is qualitatively the same as that for the classical spin, the ground state in a magnetic field H cannot be determined uniquely only from the classical model. The classical equilibrium condition is given by S 1 + S 2 + S 3 = gµ B H/(3J) with the sublattice spin Si. Because there are an infinite number of spin states that satisfy this condition, the classical ground state is continuously degenerate. The ground state of a small spin TLHAF in a magnetic field is essentially determined by the quantum fluctuation energy.A remarkable quantum effect is that an up-up-down spin state, which appears in a magnetic field for the classical model, can be stabilized in a finite magnetic field range, so that the magnetization curve has a plateau at onethird of the saturation magnetization 8-14 .The nature of the quantum mechanical ground state in a magnetic field is strongly reflected in the magnetization process. The magnetization process for a 2D spin-1 2 TLHAF, which exhibits the most pronounced quantum effect, was calculated energetically by means of spin wave theory 9,11 , coupled cluster method 12 and exact diagonalization 13,14 . The calculated magnetization curves are greatly different from that for the classical spin. However, experimental verification of the theoretical results has not been conducted at a quantitative level.Experimentally, Cs 2 CuCl 4 15 , Cs 2 CuBr 4 16,17 and κ-(BEDT-TTF) 2 Cu 2 (CN) 3 18 have been actively investigated as spin-1 2 TLHAFs. However, the triangular lattice in these substances is not regular but distorted, and thus, the exchange interaction is spatially anisotropic. Cs 2 CuCl 4 and Cs 2 CuBr 4 also exhibit a large antisymmetric interaction of the Dzyaloshinsky-Moriya (DM) type. Although the quantum magnetization plateau has been actually observed in Cs 2 CuBr 4 16,17 , the magnetization process is anisotropic and the magnetization plateau is not observed for a magnetic field perpendicular to the triangular lattice plane. In this letter, we present the results of magnetization and specific heat measurements on Ba 3 CoSb 2 O 9 and provide evidence that Ba 3 CoSb 2 O 9 closely approximates to the ideal spin- This substance crystallizes in a highly symmetric hexagonal structure, P 6 3 /mmc, whi...

The Magnetization Process of the Spin-One Triangular-Lattice Heisenberg Antiferromagnet

“…For the extreme quantum case, i.e., spin quantum number s ¼ 1=2, this plateau at T ¼ 0 has been widely discussed. Very recently an almost perfect experimental realization of an s ¼ 1=2 triangular-lattice Heisenberg antiferromagnet has been reported for Ba 3 CoSb 2 O 9 , 17) where the entire measured magnetization curve MðHÞ including the 1/3 plateau is in excellent agreement with theoretical predictions. In particular, the theoretical magnetization data obtained by high-order coupled-cluster approximation 11) and by large-scale exact diagonalization approach 15) almost coincide with the measured ones over a wide range of the magnetic field.…”

“…Motivated by the excellent agreement of theoretical predictions based on exact diagonalization (ED) and coupled-cluster method (CCM) and experimental results for the spin-half compound Ba 3 CoSb 2 O 9 reported in Ref. 17 we apply in this paper these methods to the s ¼ 1 model to provide theoretical data to compare with experimental results.For the s ¼ 1 model the Lanczos ED for N ¼ 27 sites and the CCM-SUBn-n approximation 20) for n ¼ 2; 4; 6; 8 are used to calculate the field-dependent properties of the model.In this short note we will not explain details of the Lanczos ED and the CCM. We refer the interested reader e.g., to Refs.…”

“…Motivated by the excellent agreement of theoretical predictions based on exact diagonalization (ED) and coupled-cluster method (CCM) and experimental results for the spin-half compound Ba 3 CoSb 2 O 9 reported in Ref. 17 we apply in this paper these methods to the s ¼ 1 model to provide theoretical data to compare with experimental results.…”

“…One spectacular feature of this model system for a strongly frustrated magnet is the unconventional magnetization process of the triangular-lattice Heisenberg antiferromagnet. [3][4][5][6][7][8][9][10][11][12][13][14][15][16][17] While for the classical isotropic Heisenberg model at zero temperature the magnetization M increases linearly with the applied magnetic field H, thermal or quantum fluctuations induce a plateau at 1/3 of the saturation magnetization M sat (''order from disorder'' phenomenon). For the extreme quantum case, i.e., spin quantum number s ¼ 1=2, this plateau at T ¼ 0 has been widely discussed.…”

Ground states of spin-12triangular antiferromagnets in a magnetic field

“…Several experimental spin-1/2 materials with the triangular lattice structure have indeed been observed to support a 1/3 magnetization plateau, including the well documented material Cs 2 CuBr 4 53,54 as well as Ba 3 CoSb 2 O 9 , studied more recently 55 . A notable exception is Cs 2 CuCl 4 , which is isostructural to Cs 2 CuBr 4 , but does not exhibit a magnetization plateau 56 .…”