This paper was converted on www.awesomepapers.org from LaTeX by an anonymous user.
Want to know more? Visit the Converter page.

thanks: This manuscript has been authored by UT-Battelle, LLC under Contract No. DE-AC05-00OR22725 with the U.S. Department of Energy. The United States Government retains and the publisher, by accepting the article for publication, acknowledges that the United States Government retains a non-exclusive, paid-up, irrevocable, world-wide license to publish or reproduce the published form of this manuscript, or allow others to do so, for United States Government purposes. The Department of Energy will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (http://energy.gov/downloads/doe-public-access-plan).

Weakly coupled alternating S=1/2S=1/2 chains in the distorted honeycomb lattice compound Na2Cu2TeO6

Shang Gao Materials Science & Technology Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA Neutron Scattering Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA    Ling-Fang Lin Department of Physics and Astronomy, University of Tennessee, Knoxville, Tennessee 37996, USA    Andrew F. May Materials Science & Technology Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA    Binod K. Rai Materials Science & Technology Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA    Qiang Zhang Neutron Scattering Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA    Elbio Dagotto Materials Science & Technology Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA Department of Physics and Astronomy, University of Tennessee, Knoxville, Tennessee 37996, USA    Andrew D. Christianson Materials Science & Technology Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA    Matthew B. Stone Neutron Scattering Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA
Abstract

Spin-1/2 chains with alternating antiferromagnetic (AF) and ferromagnetic (FM) couplings exhibit quantum entanglement like the integer-spin Haldane chains and might be similarly utilized for quantum computations. Such alternating AF-FM chains have been proposed to be realized in the distorted honeycomb-lattice compound Na2Cu2TeO6, but to confirm this picture a comprehensive understanding of the exchange interactions including terms outside of the idealized model is required. Here we employ neutron scattering to study the spin dynamics in Na2Cu2TeO6 and accurately determine the coupling strengths through the random phase approximation and density functional theory (DFT) approaches. We find the AF and FM intrachain couplings are the dominant terms in the spin Hamiltonian, while the interchain couplings are AF but perturbative. This hierarchy in the coupling strengths and the alternating signs of the intrachain couplings can be understood through their different exchange paths. Our results establish Na2Cu2TeO6 as a weakly-coupled alternating AF-FM chain compound and reveal the robustness of the gapped ground state in alternating chains under weak interchain couplings.

pacs:

Introduction. Spin-1/2 chains with alternating AF and FM couplings are known to exhibit gapped excitations and exponentially decaying correlations [1, 2], which are different from those of the spin-1/2 Bethe chains with uniform couplings [3] but more similar to the spectrum of integer-spin Haldane chains [4, 4]. Assuming alternating couplings of J1J_{1} (AF) and J2=βJ1J_{2}=\beta J_{1} (β<1\beta<1), the Hamiltonian of an alternating chain can be written as =jJ1𝑺2j1𝑺2j+jJ2𝑺2j𝑺2j+1\mathcal{H}=\sum_{j}J_{1}\bm{S}_{2j-1}\cdot\bm{S}_{2j}+\sum_{j}J_{2}\bm{S}_{2j}\cdot\bm{S}_{2j+1}. In the special case of β=0\beta=0, an alternating chain can be viewed as disconnected spin dimers, and the local singlet-triplet (triplon) excitations over the dimers account for the excitation gap in the spectrum. Non-zero J2J_{2} couplings will introduce dispersion for the triplon excitations. At β=1\beta=1, one recovers the gapless character of a Bethe chain. In the whole range of <β<1-\infty<\beta<1, theoretical calculations have revealed a hidden string order that is protected by the Z2×Z2\rm{Z_{2}}\times\rm{Z_{2}} global rotation symmetry [6, 7], showing that the gapped ground state of alternating chains is a symmetry-protected topological state of the same type as that of Haldane chains [8]. Considering the prospect of Haldane chains as the resource ground state for measurement-based quantum computation [9], the additional degrees of freedom in S=1/2S=1/2 alternating chains may be important to explore further flexibility in qubit operations.

Refer to caption
Figure 1: (a, b) Crystal structure of Na2Cu2TeO6 viewed along the bb (a) and cc^{*} (b) axes with the Cu2+ ions forming chains along the bb axis. The size of the unit cell is indicated by black lines. The Cu2+ spins are coupled through alternating J1J_{1} (red) and J2J_{2} (blue) interactions along the chain and J3J_{3} interactions (grey) between the chains. Cu-O bonds of different lengths are indicated in the CuO6 octahedron at the left bottom corner of panel (b), with b1 = 1.978(1) Å, b2 = 1.999(1) Å, and b3 = 2.533(1) Å.

Experimental realizations of alternating AF-FM chains are limited to a few compounds, including CuNb2O6 [10], DMACuCl3 [11], Na3Cu2SbO6 [12], and BaCu2V2O8 [13] where the exchange couplings have been accurately determined through neutron scattering. Recently, a new candidate compound, Na2Cu2TeO6, was proposed [14, 15, 16, 17, 18]. Similar to Na3Cu2SbO6, the Cu2+ ions (S=1/2S=1/2) in Na2Cu2TeO6 form a distorted honeycomb lattice in the abab plane, which are separated by the Na+ layers along the cc axis (see Fig. 1). Magnetic susceptibility measurements on a powder sample of Na2Cu2TeO6 revealed a spin gap Δ127\Delta\sim 127 K [14], which has been attributed to strong AF couplings J1J_{1} in the DFT calculations [14, 16, 17, 18]. However, controversy remains as to the sign of the intrachain coupling J2J_{2}: magnetic susceptibility data can be fitted equally well by the FM or AF J2J_{2} model [15], and this ambiguity is not resolved by contradictory DFT results that support either AF [14, 16] or FM [17, 18] J2J_{2} couplings. The magnitude of J3J_{3} also remains to be determined, as comparable J2J_{2} and J3J_{3} couplings might invalidate a spin chain scenario.

To establish whether Na2Cu2TeO6 represents a rare realization of the alternating AF-FM spin chains, here we perform inelastic neutron scattering (INS) experiments on a single crystal sample of Na2Cu2TeO6 to determine the exchange coupling strengths. We confirm the spin gap originates from the dominant AF coupling J1J_{1} (22.8 meV), and the interchain coupling J3J_{3} is found to be much weaker (1.3 meV). Most importantly, we reveal the intrachain coupling J2J_{2} to be FM with a strength (8.7-8.7 meV) that is much higher than J3J_{3}, thus establishing Na2Cu2TeO6 as a weakly-coupled alternating AF-FM chain compound. Through the DFT calculations, the alternating signs of the interchain couplings can be attributed to their different exchange paths.

Experimental Details. Polycrystalline Na2Cu2TeO6 was utilized as a source material for crystal growth in a flux based on TeO2 and Na2CO3. Details for the polycrystal synthesis and characterization can be found in the Supplemental Materials [19]. Powders of these materials in a ratio of 2(Na2Cu2TeO6):1(Na2CO3):4(TeO2) were mixed and loaded into a Pt crucible that was covered with a Pt lid. The crucible was heated rapidly to 900 C where the melt was homogenized for 12 h in air. The furnace was then cooled at 2 C/h to 500 C at which point it was turned off to cool. The translucent green crystals were recovered by boiling the product-filled crucible in a hot aqueous solution of potassium hydroxide, followed by additional rinsing in deionized water.

INS experiments on Na2Cu2TeO6 were performed on the fine-resolution Fermi chopper spectrometer SEQUOIA at the SNS of the ORNL. A single crystal (mass \sim140 mg) was aligned with the (001) vector vertical. A closed cycle refrigerator (CCR) was employed to reach temperatures TT down to 5 K. The incident neutron energy was Ei=60E_{i}=60 meV, and the Fermi chopper frequency of 180 Hz and 420 Hz was selected in the high-intensity and high-resolution modes, respectively. Data were acquired by rotating the sample in 1 steps, covering a total range of 200 at 5 K and 100 at higher temperatures. Data reductions and projections were performed using MANTID [20] and HORACE softwares [21].

Refer to caption
Figure 2: (a) Scattering intensity of Na2Cu2TeO6 measured along (1 0.5Ql)(-1\ 0.5\ Q_{l}) at T=5T=5 K with integration widths of ΔQh=0.2\Delta Q_{h}=0.2 and ΔQk=0.1\Delta Q_{k}=0.1. Measurements were performed in the high-resolution configuration. (b) Intensity map in the (QhQk 0)(Q_{h}\ Q_{k}\ 0) plane integrated over ranges of [17, 28][17,\ 28] meV in EE and [2, 2][-2,\ 2] in QlQ_{l}. Measurements were performed in the high-intensity setup at T=5T=5 K. (c) Scattering intensity along QkQ_{k} integrated from the map in panel (b) over the range of [2.1, 1.9][-2.1,\ 1.9] in QhQ_{h}. The intensity can be described by the structure factor of the r1r_{1}-dimer model (red solid line) instead of the r2r_{2}-dimer model (blue dashed line), where r12b/3r_{1}\approx 2b/3 and r2b/3r_{2}\approx b/3 are the bond distances for J1J_{1} and J2J_{2}, respectively.
Refer to caption
Figure 3: (a-d) Intensity maps of Na2Cu2TeO6 measured along (1 QkQ_{k} 0), (2 QkQ_{k} 0), (QhQ_{h} 0.5 0), and (QhQ_{h} 1 0) at T=5T=5 K using the high-intensity configuration. The integration widths are ΔQh=0.1\Delta Q_{h}=0.1, ΔQk=0.1\Delta Q_{k}=0.1, and ΔQl=2\Delta Q_{l}=2. (e-h) Intensity maps calculated through the random phase approximation with fitted coupling strengths J1=22.78(2)J_{1}=22.78(2) meV, J2=8.73(4)J_{2}=-8.73(4) meV, and J3=1.34(3)J_{3}=1.34(3) meV. An instrumental energy resolution of 3.0 meV was convolved with the calculated spectrum. Panels (a)-(d) and (e)-(h) are shown on the identical intensity scale, respectively.

First-principles calculations using the projector augmented wave (PAW) method were performed based on the DFT as implemented in the Vienna ab initio Simulation Package (VASP) code [22, 23, 24]. The generalized gradient approximation (GGA) and the revised Perdew-Burke-Ernzerhof (PBEsol) function were used to treat the electron exchange-correlation potential [25, 26]. Based on the ab initio ground-state wave function, the Cu-site centered Wannier functions (WFs) with orbital dx2y2d_{x^{2}-y^{2}} were constructed using the WANNIER90 code [27, 28].

Results. As shown in Fig. 1(a), the large separation of c=5.92c=5.92 Å between the honeycomb layers indicates likely negligible interlayer couplings. This is immediately confirmed in our single-crystal INS experiment. Figure 2(a) plots a representative slice of the measured spectrum along the QlQ_{l} direction that is perpendicular to the abab plane. A single excitation mode is observed at 18\sim 18 meV, which is flat within the instrument resolution (1.6 meV at the elastic line). In contrast, strong dispersion in the energy range of E=E= [17, 28] meV are observed in the abab plane as summarized in Fig. 3(a-d), thus confirming all the related couplings to be within the honeycomb layers.

The lack of dispersion out of the abab plane allows us to integrate data along QlQ_{l} for better statistics. Fig. 2b plots the scattering intensity integrated in the range of QlQ_{l} in [-2, 2] (r.l.u.) and EE in [17, 28] meV. Along the QkQ_{k} direction, intensity is strongly modulated with a periodicity of 1.5\sim 1.5 (r.l.u.). For dimer systems, it is established that intensity of the triplon excitations is modulated by the dimer structure factor S(𝑸)[1cos(𝑸𝒓)]S(\bm{Q})\propto[1-\cos(\bm{Q}\cdot\bm{r})], where 𝒓\bm{r} is the vector that connects the two spin sites within the dimer [29, 30, 31]. Therefore, the modulation along QkQ_{k} indicates that dimers in Na2Cu2TeO6 are forming along the bb axis, and its periodicity tells the bond distance rr within the dimers. As shown in Fig. 2(c), the model that assumes dimers forming over the J1J_{1} bonds with a distance of r1=2b/3r_{1}=2b/3 accurately describes the intensity modulation, thus confirming the J1J_{1} couplings to be AF and dominant as proposed by the DFT calculations [15, 16, 17, 18].

Figure 3(a-d) summarizes the dispersion along QhQ_{h} and QkQ_{k} in the abab plane. Along the chain direction (QkQ_{k}), the triplon band reaches its lowest energy of 18\sim 18 meV at Qk=N+1/2Q_{k}=N+1/2 with integer NN, indicating the FM character of the interdimer J2J_{2} couplings [32, 11]. Different from the conventional chain compounds, the top of the dispersion varies at successive integer QkQ_{k} positions, which might arise from the interchain couplings. As compared in Fig. 3(c) and (d), although the triplon band looks flat along (QhQ_{h} 0.5 0) at the bottom of the band, dispersion with a bandwidth of 3\sim 3 meV is observed along (QhQ_{h} 1 0) at the top of the band, which suggests weak but non-negligible J3J_{3} couplings.

Refer to caption
Figure 4: (a) Energy dependence of the scattering intensity at (1 0.5 0) measured at T=5T=5 (circles), 50 (squares), 100 (diamonds), 150 (up-pointing triangles), and 250 (right-pointing triangles) K. Solid lines are gaussian fits. (b) Temperature dependence of the fitted gap size Δ\Delta (circles, righ axis) and full width at half maximum FWHM (squares, left axis). Solid lines are fits using the empirical formula in Eq. (3). The fitted parameters are α=2.3(8)\alpha=2.3(8) meV, γ=6(2)\gamma=6(2) meV, and Γ0=2.8(2)\Gamma_{0}=2.8(2) meV. An instrumental energy resolution of 3 meV was convolved with the calculated spectrum.

The triplon dispersion in dimer systems can often be analyzed through the random phase approximation [33, 29, 34, 35, 36]. Under this approximation, the dispersion relation can be written as

ω(𝑸)=J12+J1𝒥(𝑸)R(T),\hbar\omega(\bm{Q})=\sqrt{J_{1}^{2}+J_{1}\mathcal{J}(\bm{Q})R(T)}\rm{,} (1)

where R(T)R(T) describes the population difference between the singlet and triplet states [29, 34] and for Na2Cu2TeO6 can be approximated by 1 at T=5T=5 K due to the large excitation gap. 𝒥(𝑸)\mathcal{J}(\bm{Q}) is the Fourier sum of interactions beyond dimer exchange:

𝒥(𝑸)=J2cos(2πQk)2J3cos(πQh)cos(πQk).\mathcal{J}(\bm{Q})=-J_{2}\cos(2\pi Q_{k})-2J_{3}\cos(\pi Q_{h})\cos(\pi Q_{k})\rm{.} (2)

Experimental dispersion values were extracted from Gaussian fits to constant QQ scans at 140 points throughout the measured reciprocal space volume, which were then fitted by the dispersion relation in Eq. (1). The fitted coupling strengths are J1=22.78(2)J_{1}=22.78(2) meV, J2=8.73(4)J_{2}=-8.73(4) meV, and J3=1.34(3)J_{3}=1.34(3) meV. The calculated spectra are summarized in Fig. 3(e-h) for comparison with the experimental data. The FM character of the J2J_{2} couplings is thus unambiguously established, and the weakness of the interchain J3J_{3} justifies the description of Na2Cu2TeO6 as a weakly-coupled alternating AF-FM chain compound.

Refer to caption
Figure 5: (a) Wannier functions of the Cu 3dx2y23d_{x^{2}-y^{2}} orbital, an antibonding combination of the Cu 3dx2y23d_{x^{2}-y^{2}} and O 2p2p states, viewed along the cc axis. Different colors represent the +/- signs of the Wannier function. (b) Diagrams for the super-superexchange and superexchange couplings between the nearest-neighbor Cu 3dx2y23d_{x^{2}-y^{2}} orbitals via oxygen 2p2p ligands. For the J1J_{1} path, the Cu-O\cdotsO-Cu super-superexchange leads to the AF alignment of the nearest-neighbor Cu ions. For the J2J_{2} path, Cu-O-Cu superexchange with a bonding angle of 9090^{\circ} results in a FM exchange between the nearest-neighbor Cu ions.

In Haldane chains the existence of a spin gap is known to protect the spin entanglement [37]. Na2Cu2TeO6 can serve as a model system to test that the same ideas are valid for alternating spin chains. In long-range ordered magnets, the excitation gap often decreases at higher TT due to reduced ordering moments. As a contrast, the gap of Haldane chains increases with TT due to the reduction in the coherent length and the consequent finite-size effect [38, 39]. The EE scans at 𝑸=\bm{Q}= (1, 0.5, 0) shown in Fig. 4(a) indeed reveal an increased gap at elevated TT. We parameterize the lineshape using a Gaussian function that is independently fit at each temperature, and the value of Δ\Delta is the centroid of the Gaussian peak. As summarized in Fig. 4(b), in a large range below 150\sim 150 K, the fitted gap size Δ\Delta and full width at half maximum (FWHM) Γ\Gamma exhibit an activated behavior that are characteristic of Haldane chains [40, 41]:

Δ(T)\displaystyle\Delta(T) =Δ0+αTexp(Δ0/T)\displaystyle=\Delta_{0}+\sqrt{\alpha T}\exp(-\Delta_{0}/T)
Γ(T)\displaystyle\Gamma(T) =Γ0+γexp(Δ0/T),\displaystyle=\Gamma_{0}+\gamma\exp(-\Delta_{0}/T)\rm{,} (3)

where α\alpha, γ\gamma, and Γ0\Gamma_{0} are fitting parameters, Δ0\Delta_{0} is the gap size at 0 K and is fixed at 18.0 meV. The validity of the activated behavior further confirms the similarity between the S=1/2S=1/2 alternating AF-FM chains and the integer-spin Haldane chains, thus revealing the robustness of the topological ground state against weak interchain couplings.

Discussion. The emergence of spin chains in Na2Cu2TeO6 can be ascribed to the distortion of the honeycomb lattice. As shown in Fig. 1(b), the exchange paths of J3J_{3} involve the longest Cu-O bonds b3 of the distorted octahedra. Therefore, the J3J_{3} couplings are expected to be weak as the unoccupied Cu 3dx2y23d_{x^{2}-y^{2}} orbitals disfavor the elongated bond direction, which reduces the electron hopping between the chains.

The sign of the interchain couplings J2J_{2} is more subtle, and different scenarios exist in the previous DFT calculations [14, 16, 17, 18]. As summarized in the Supplemental Materials, our DFT calculations confirm the alternating FM and AF intrachain couplings in agreement with the experiment. The contrasting J1J_{1} and J2J_{2} couplings can be understood through the Wannier functions (WFs) as plotted in Fig. 5(a). Due to the contributions from the O 2p2p states, the WFs overlap directly over the J1J_{1} paths but are almost orthogonal over the J2J_{2} path. Therefore, the J1J_{1} path, in spite of its longer distance, develops a stronger coupling than that over the J2J_{2} path. Based on the WFs, the signs of the couplings can be understood through the Goodenough-Kanamori-Anderson rules [42, 43, 44, 45] as shown in Fig. 5(b). For the J1J_{1} couplings, the Cu-O\cdotsO-Cu super-superexchange leads to an AF interaction between the Cu2+ spins. While for the J2J_{2} couplings, the interaction become FM as the angle of ∠Cu-O-Cu is close to 90\sim 90^{\circ}, which means a pair of orthogonal O 2p2p orbitals with parallel spins are involved in the virtual electron hopping.

Conclusions. Neutron scattering experiments have been performed on the honeycomb-lattice compound Na2Cu2TeO6 to study its spin correlations. A triplon excitation mode was observed, which exhibits strong dispersion along the chain but weak dispersion perpendicular to the chain. Under the random phase approximation, the intrachain couplings were found to be alternating AF and FM, and a weak interchain coupling was also established. The emergence of spin chains in Na2Cu2TeO6 was ascribed to the distortion of the honeycomb lattice, and the alternating intrachain couplings were understood through the DFT calculations. Our works establish the existence of weakly-coupled alternating AF-FM spin-1/2 chains in Na2Cu2TeO6 and reveal a robust gapped ground state that is similar to that of the integer-spin Haldane chains.

Acknowledgments. We acknowledge helpful discussions with Jyong-Hao Chen, Tong Chen, and Tao Hong. This work was supported by the U.S. Department of Energy, Office of Science, Basic Energy Sciences, Materials Sciences and Engineering Division. This research used resources at the Spallation Neutron Source (SNS), a DOE Office of Science User Facility operated by the Oak Ridge National Laboratory (ORNL).

References

  • Hida [1992a] K. Hida, Ground-state phase diagram of the spin-1/2 ferromagnetic-antiferromagnetic alternating heisenberg chain with anisotropy, Phys. Rev. B 46, 8268 (1992a).
  • Hida [1994] K. Hida, Excitation spectrum of the spin-1/2 ferromagnetic-antiferromagnetic alternating Heisenberg chain, J. Phys. Soc. Jpn 63, 2514 (1994).
  • Bethe [1931] H. Bethe, Zur Theorie der Metalle. I. Eigenwerte und Eigenfunktionen der linearen Atomkette, Z. Phys. 71, 205 (1931).
  • Haldane [1983] F. D. M. Haldane, Nonlinear field theory of large-spin Heisenberg antiferromagnets: Semiclassically quantized solitons of the one-dimensional easy-axis Néel state, Phys. Rev. Lett. 50, 1153 (1983).
  • Affleck et al. [1987] I. Affleck, T. Kennedy, E. H. Lieb, and H. Tasaki, Rigorous results on valence-bond ground states in antiferromagnets, Phys. Rev. Lett. 59, 799 (1987).
  • Hida [1992b] K. Hida, Crossover between the Haldane-gap phase and the dimer phase in the spin-1/2 alternating Heisenberg chain, Phys. Rev. B 45, 2207 (1992b).
  • Kohmoto and Tasaki [1992] M. Kohmoto and H. Tasaki, Hidden Z×2{}_{2}\timesZ2 symmetry breaking and the Haldane phase in the S=1/2{S}=1/2 quantum spin chain with bond alternation, Phys. Rev. B 46, 3486 (1992).
  • Pollmann et al. [2010] F. Pollmann, A. M. Turner, E. Berg, and M. Oshikawa, Entanglement spectrum of a topological phase in one dimension, Phys. Rev. B 81, 064439 (2010).
  • Miyake [2010] A. Miyake, Quantum computation on the edge of a symmetry-protected topological order, Phys. Rev. Lett. 105, 040501 (2010).
  • Kodama et al. [1999] K. Kodama, H. Harashina, H. Sasaki, M. Kato, M. Sato, K. Kakurai, and M. Nishi, Neutron scattering study on the quasi-one-dimensional spin-gap system CuNb2O6J. Phys. Soc. Jpn 68, 237 (1999).
  • Stone et al. [2007] M. B. Stone, W. Tian, M. D. Lumsden, G. E. Granroth, D. Mandrus, J.-H. Chung, N. Harrison, and S. E. Nagler, Quantum spin correlations in an organometallic alternating-sign chain, Phys. Rev. Lett. 99, 087204 (2007).
  • Miura et al. [2008] Y. Miura, Y. Yasui, T. Moyoshi, M. Sato, and K. Kakurai, Magnetic excitations of spin-gap system Na3Cu2SbO6 with distorted honeycomb structure, J. Phys. Soc. Jpn 77, 104709 (2008).
  • Klyushina et al. [2018] E. S. Klyushina, A. T. M. N. Islam, J. T. Park, E. A. Goremychkin, E. Wheeler, B. Klemke, and B. Lake, S=1/2S=1/2 dimerized antiferromagnetic-ferromagnetic quantum spin chain BaCu2V2O8Phys. Rev. B 98, 104413 (2018).
  • Xu et al. [2005] J. Xu, A. Assoud, N. Soheilnia, S. Derakhshan, H. L. Cuthbert, J. E. Greedan, M. H. Whangbo, and H. Kleinke, Synthesis, structure, and magnetic properties of the layered copper(II) oxide Na2Cu2TeO6Inorg. Chem. 44, 5042 (2005).
  • Miura et al. [2006] Y. Miura, R. Hirai, Y. Kobayashi, and M. Sato, Spin-gap behavior of Na3Cu2SbO6 with distorted honeycomb structure, J Phys. Soc. Jpn. 75, 084707 (2006).
  • Derakhshan et al. [2007] S. Derakhshan, H. L. Cuthbert, J. E. Greedan, B. Rahaman, and T. Saha-Dasgupta, Electronic structures and low-dimensional magnetic properties of the ordered rocksalt oxides Na3Cu2SbO6 and Na2Cu2TeO6Phys. Rev. B 76, 104403 (2007).
  • Koo and Whangbo [2008] H.-J. Koo and M.-H. Whangbo, Analysis of the spin lattice model for the spin-gapped layered compounds Na3Cu2SbO6 and Na2Cu2TeO6 on the basis of electronic structure calculations, Inorg. Chem. 47, 128 (2008).
  • Schmitt et al. [2014] M. Schmitt, O. Janson, S. Golbs, M. Schmidt, W. Schnelle, J. Richter, and H. Rosner, Microscopic magnetic modeling for the Na3Cu2SbO6 and Na2Cu2TeO6Phys. Rev. B 89, 174403 (2014).
  • [19]  See Supplemental Materials for details on the preparation and characterizations of the powder sample and the DFT calculations.
  • Arnold et al. [2014] O. Arnold, J. Bilheux, J. Borreguero, A. Buts, S. Campbell, L. Chapon, M. Doucet, N. Draper, R. Ferraz Leal, M. Gigg, V. Lynch, A. Markvardsen, D. Mikkelson, R. Mikkelson, R. Miller, K. Palmen, P. Parker, G. Passos, T. Perring, P. Peterson, S. Ren, M. Reuter, A. Savici, J. Taylor, R. Taylor, R. Tolchenov, W. Zhou, and J. Zikovsky, Mantid—Data analysis and visualization package for neutron scattering and μ\muSR experiments, Nucl. Instr. Methods Phys. Res. Sec. A 764, 156 (2014).
  • Ewings et al. [2016] R. Ewings, A. Buts, M. Le, J. van Duijn, I. Bustinduy, and T. Perring, Horace: Software for the analysis of data from single crystal spectroscopy experiments at time-of-flight neutron instruments, Nucl. Instrum. Methods Phys. Res. A 834, 132 (2016).
  • Kresse and Furthmüller [1996] G. Kresse and J. Furthmüller, Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set, Phys. Rev. B 54, 11169 (1996).
  • Kresse and Joubert [1999] G. Kresse and D. Joubert, From ultrasoft pseudopotentials to the projector augmented-wave method, Phys. Rev. B 59, 1758 (1999).
  • Blöchl [1994] P. E. Blöchl, Projector augmented-wave method, Phys. Rev. B 50, 17953 (1994).
  • Perdew et al. [1996] J. P. Perdew, K. Burke, and M. Ernzerhof, Generalized gradient approximation made simple, Phys. Rev. Lett. 77, 3865 (1996).
  • Perdew et al. [2008] J. P. Perdew, A. Ruzsinszky, G. I. Csonka, O. A. Vydrov, G. E. Scuseria, L. A. Constantin, X. Zhou, and K. Burke, Restoring the density-gradient expansion for exchange in solids and surfaces, Phys. Rev. Lett. 100, 136406 (2008).
  • Marzari and Vanderbilt [1997] N. Marzari and D. Vanderbilt, Maximally localized generalized wannier functions for composite energy bands, Phys. Rev. B 56, 12847 (1997).
  • Mostofi et al. [2008] A. A. Mostofi, J. R. Yates, Y.-S. Lee, I. Souza, D. Vanderbilt, and N. Marzari, wannier90: A tool for obtaining maximally-localised wannier functions, Comput. Phys. Commun. 178, 685 (2008).
  • Leuenberger et al. [1984] B. Leuenberger, A. Stebler, H. U. Güdel, A. Furrer, R. Feile, and J. K. Kjems, Spin dynamics of an isotropic singlet-ground-state antiferromagnet with alternating strong and weak interactions: An inelastic-neutron-scattering study of the dimer compound Cs3Cr2Br9Phys. Rev. B 30, 6300 (1984).
  • Xu et al. [2000] G. Xu, C. Broholm, D. H. Reich, and M. A. Adams, Triplet waves in a quantum spin liquid, Phys. Rev. Lett. 84, 4465 (2000).
  • Cavadini et al. [2001] N. Cavadini, G. Heigold, W. Henggeler, A. Furrer, H.-U. Güdel, K. Krämer, and H. Mutka, Magnetic excitations in the quantum spin system TlCuCl3Phys. Rev. B 63, 172414 (2001).
  • Tennant et al. [2003] D. A. Tennant, C. Broholm, D. H. Reich, S. E. Nagler, G. E. Granroth, T. Barnes, K. Damle, G. Xu, Y. Chen, and B. C. Sales, Neutron scattering study of two-magnon states in the quantum magnet copper nitrate, Phys. Rev. B 67, 054414 (2003).
  • Haley and Erdös [1972] S. B. Haley and P. Erdös, Standard-basis operator method in the Green’s-function technique of many-body systems with an application to ferromagnetism, Phys. Rev. B 5, 1106 (1972).
  • Stone et al. [2008a] M. B. Stone, M. D. Lumsden, S. Chang, E. C. Samulon, C. D. Batista, and I. R. Fisher, Singlet-triplet dispersion reveals additional frustration in the triangular-lattice dimer Ba3Mn2O8Phys. Rev. Lett. 100, 237201 (2008a).
  • Stone et al. [2008b] M. B. Stone, M. D. Lumsden, Y. Qiu, E. C. Samulon, C. D. Batista, and I. R. Fisher, Dispersive magnetic excitations in the S=1{S}=1 antiferromagnet Ba3Mn2O8Phys. Rev. B 77, 134406 (2008b), publisher: American Physical Society.
  • Allenspach et al. [2020] S. Allenspach, A. Biffin, U. Stuhr, G. S. Tucker, S. Ohira-Kawamura, M. Kofu, D. J. Voneshen, M. Boehm, B. Normand, N. Laflorencie, F. Mila, and C. Rüegg, Multiple magnetic bilayers and unconventional criticality without frustration in BaCuSi2O6Phys. Rev. Lett. 124, 177205 (2020).
  • Wierschem and Sengupta [2014] K. Wierschem and P. Sengupta, Characterizing the haldane phase in quasi-one-dimensional spin-1 heisenberg antiferromagnets, Mod. Phys. Lett. B 28, 1430017 (2014).
  • Jolicœur and Golinelli [1994] T. Jolicœur and O. Golinelli, σ\sigma model study of Haldane-gap antiferromagnets, Phys. Rev. B 50, 9265 (1994).
  • Nightingale and Blöte [1986] M. P. Nightingale and H. W. J. Blöte, Gap of the linear spin-1 Heisenberg antiferromagnet: A Monte Carlo calculation, Phys. Rev. B 33, 659 (1986).
  • Zheludev et al. [1996] A. Zheludev, S. E. Nagler, S. M. Shapiro, L. K. Chou, D. R. Talham, and M. W. Meisel, Spin dynamics in the linear-chain SS=1 antiferromagnet Ni(C3H10N2)2N3(ClO4), Phys. Rev. B 53, 15004 (1996).
  • Bera et al. [2015] A. K. Bera, B. Lake, A. T. M. N. Islam, and A. Schneidewind, Critical properties of coupled anisotropic Haldane spin chains in a magnetic field, Phys. Rev. B 92, 060412(R) (2015).
  • Anderson [1950] P. W. Anderson, Antiferromagnetism. Theory of superexchange interaction, Phys. Rev. 79, 350 (1950).
  • Goodenough [1955] J. B. Goodenough, Theory of the role of covalence in the perovskite-type manganites [La, MM(II)]MnO3Phys. Rev. 100, 564 (1955).
  • Goodenough [1958] J. B. Goodenough, An interpretation of the magnetic properties of the perovskite-type mixed crystals La1-xSrxCoO3-λJ. Phys. Chem. Solids 6, 287 (1958).
  • Moskvin et al. [1975] A. S. Moskvin, N. S. Ovanesyan, and V. A. Trukhtanov, Angular dependence of the superexchange interaction Fe3+-O2--Cr3+Hyperfine Interactions 1, 265 (1975).

Weakly-coupled alternating S=1/2S=1/2 chains in the distorted honeycomb-lattice compound Na2Cu2TeO6
Supplementary Information

I Polycrystal synthesis and characterization

Polycrystalline samples of Na2Cu2TeO6 were prepared using a solid-state method. Powders of Na2CO3 (99.997%\%, Alfa Aesar), TeO2 (99.99%\%, Beantown Chemical) and CuO (99.995%\%, Alfa Aesar) were mixed in a molar ratio of 1.03:1:2. The mixed powders were ground thoroughly in air and pressed into pellets with 12′′\frac{1}{2}^{\prime\prime} diameter. A furnace was preheated to 660 C before the pellets were loaded in the furnace to minimize the loss of Na2CO3. The temperature was held at 660 C for 8 hours before cooling down to room temperature.

Neutron diffraction experiments on a powder sample of Na2Cu2TeO6 were performed on the time-of-flight powder diffractometer POWGEN at the Spallation Neutron Source (SNS) of the Oak Ridge National Laboratory (ORNL). The powder (mass \sim8.4 g) was placed in a vanadium can. Data were acquired at 300 K and 20 K using the 1.5 Å instrumental configuration with an incoming neutron wavelength band of 0.967-2.033 Å. Rietveld refinements of the neutron diffraction data were performed using the FULLPROF program [1].

Refer to caption
Figure S1: Refinement result of the powder neutron diffraction data measured at T=20T=20 K. Data points are shown as red circles. The calculated pattern is shown as the black solid line. The vertical bars from top to bottom indicate the positions of the structural Bragg peaks for Na2Cu2TeO6, Cu3TeO6, Cu2O, CuO, and the magnetic Bragg peaks for CuO. The blue line at the bottom shows the difference of measured and calculated intensities.

Figure S1 summarizes the refinement results for the powder neutron diffraction data collected at T=20T=20 K, which is far below the 127\sim 127 K gap revealed in magnetic susceptibility. The Na2Cu2TeO6 phase (fractional weight 88\sim 88%) can be refined using the monoclinic C2/mC2/m space group with lattice parameters a=5.6951(1)a=5.6951(1) Å, b=8.6560(1)b=8.6560(1) Å, c=5.9236(1)c=5.9236(1) Å, and β=113.741(1)\beta=113.741(1)^{\circ}, which are consistent with the previous report [2]. About 9%9\% deficiency was observed at the Na sites. The Cu ions occupy the 4g4g sites at [0, 0.665(1), 0], leading to distances of d1=5.806(3)d_{1}=5.806(3) Åb/3~{}\approx b/3 and d2=2.850(3)d_{2}=2.850(3) Å 2b/3\approx 2b/3 for the J1J_{1} and J2J_{2} bonds, respectively. Minority phases of Cu3TeO6 (0.3\sim 0.3 %), Cu2O (0.9\sim 0.9 %), and CuO (9.8\sim 9.8 %) were detected in the powder sample. The magnetic phase of CuO was also included in the refinement [3], which contributes to the weak reflection at Q0.84Q\sim 0.84 Å-1. The overall goodness-of-fit parameters are Rp=10.8%R_{p}=10.8\%, Rwp=7.2%R_{wp}=7.2\%, and χ2=24.9\chi^{2}=24.9. Most importantly, no magnetic reflection of Na2Cu2TeO6 can be discerned in our diffraction data at T=20T=20 K, which confirms the quantum disordered ground state of the gapped spin system [4].

II DFT calculations

To better understand the magnetic properties of Na2Cu2TeO6, we introduced the electron correlation by using GGA plus UeffU_{\rm eff} (Ueff=UJU_{\rm eff}=U-J) with the Dudarev format on the Cu site [5]. As shown in Fig. S2, we considered four different antiferromagnetic (AF) configurations together with the ferromagnetic (FM) state.

Refer to caption
Figure S2: Antiferromagnetic configurations considered in the DFT analysis. (a) Néel AFM (N-AFM), (b) stripe-AFM (S-AFM), (c) zigzag-AFM (Z-AFM), and (d) double-stripe AFM (D-AFM). The gray rectangle represents the minimum magnetic unit cell for each configuration. To eliminate the effect of choosing different cell sizes, the unit cell as shown in (d) is adopted in our calculations to compute energies for different magnetic configurations.

Using the experimental crystal structure, we calculated the energies of five magnetic configurations as a function of Ueff=0U_{\rm eff}=0, 2, 4, 6, and 8 eV in Table S1. The exchange interactions can be estimated by mapping the calculated total energies for each magnetic state to the classical Heisenberg model. Four AFM equations are used to calculate four unknown parameters and the extracted spin exchange parameters can be found in Table S2.

Table S1: List of energy equations and calculated energies of the five collinear spin configurations used to estimate the magnetic exchange integrals, where S=12S=\frac{1}{2} is the magnetic moment. The double-stripe AFM state is taken as the energy reference. All energies in units of meV.
Confg. Energy equations Energy (meV)
UeffU_{\rm eff} (eV)= 0 2 4 6 8
FM E0+4J1S2+4J2S2+8J3S2E_{0}+4J_{1}S^{2}+4J_{2}S^{2}+8J_{3}S^{2} 269 172 116 77 49
N-AFM E04J1S24J2S28J3S2E_{0}-4J_{1}S^{2}-4J_{2}S^{2}-8J_{3}S^{2} 38 33 28 22 16
S-AFM E0+4J1S2+4J2S28J3S2E_{0}+4J_{1}S^{2}+4J_{2}S^{2}-8J_{3}S^{2} 252 160 107 71 44
Z-AFM E04J1S24J2S2+8J3S2E_{0}-4J_{1}S^{2}-4J_{2}S^{2}+8J_{3}S^{2} 55 46 38 29 21
D-AFM E04J1S2+4J2S2E_{0}-4J_{1}S^{2}+4J_{2}S^{2} 0 0 0 0 0
Table S2: List of magnetic exchange couplings obtained in our DFT calculations, in units of meV.
Coupling UeffU_{\rm eff} (eV) = 0 2 4 6 8
J1J_{1} 130.3 83.3 56.0 37.3 23.3
J2J_{2} -23.3 -19.8 -16.3 -12.8 -9.3
J3J_{3} 4.3 3.3 2.5 1.8 1.3

For all of the values of UeffU_{\rm eff} employed here, the super-superexchange coupling J1J_{1} is AF, and its coupling strength is several times higher than that of the FM coupling J2J_{2}. The J3J_{3} coupling is AF and is much weaker in strength than J1J_{1} and J2J_{2}. Thus, the two strongest spin exchange couplings J1J_{1} and J2J_{2} form alternating chains, which is consistent with the experimental results in the main text. With UeffU_{\rm eff} in the range of [0, 8] eV, the ratio of J2/J1J_{2}/J_{1} has maximum value of 0.399-0.399 and a minimum value of 0.179-0.179, covering the experimentally determined value of 0.383(2)-0.383(2). The calculated value of J3/J1J_{3}/J_{1} has a maximum value of 0.056 and a minimum value of 0.033. The experimentally determined ratio of J3/J1J_{3}/J_{1} was found to be 0.059(1), which is close to the upper limit of the calculated value.

All our present DFT results show good agreement with the previous theoretical works [6, 7]. As explained in Ref. [6], the DFT calculations in Ref. [2] neglects the ferromagnetic contributions to the exchange coupling, thus predicting the wrong sign for J2J_{2} when the coupling strength is weak [8]. While in Ref. [9], although the authors emphasized that the J2J_{2} coupling should be AF by comparing their total energy with different spin arrangements, not sufficient information about the details of the calculation, such as energies or spin arrangements used, were presented in that publication. By contrast, all the details of the energies and mapping equations for different magnetic configurations are listed clearly in the present work.

References

  • Rodriguez-Carvajal [1993] J. Rodriguez-Carvajal, Recent advances in magnetic structure determination by neutron powder diffraction, Physica B: Conden. Matter 192, 55 (1993).
  • Xu et al. [2005] J. Xu, A. Assoud, N. Soheilnia, S. Derakhshan, H. L. Cuthbert, J. E. Greedan, M. H. Whangbo, and H. Kleinke, Synthesis, structure, and magnetic properties of the layered copper(II) oxide Na2Cu2TeO6Inorg. Chem. 44, 5042 (2005).
  • Wang et al. [2016] Z. Wang, N. Qureshi, S. Yasin, A. Mukhin, E. Ressouche, S. Zherlitsyn, Y. Skourski, J. Geshev, V. Ivanov, M. Gospodinov, and V. Skumryev, Magnetoelectric effect and phase transitions in CuO in external magnetic fields, Nat. Commun. 7, 10295 (2016).
  • Affleck et al. [1987] I. Affleck, T. Kennedy, E. H. Lieb, and H. Tasaki, Rigorous results on valence-bond ground states in antiferromagnets, Phys. Rev. Lett. 59, 799 (1987).
  • Dudarev et al. [1998] S. L. Dudarev, G. A. Botton, S. Y. Savrasov, C. J. Humphreys, and A. P. Sutton, Electron-energy-loss spectra and the structural stability of nickel oxide: An LSDA+U study, Phys. Rev. B 57, 1505 (1998).
  • Koo and Whangbo [2008] H.-J. Koo and M.-H. Whangbo, Analysis of the spin lattice model for the spin-gapped layered compounds Na3Cu2SbO6 and Na2Cu2TeO6 on the basis of electronic structure calculations, Inorg. Chem. 47, 128 (2008).
  • Schmitt et al. [2014] M. Schmitt, O. Janson, S. Golbs, M. Schmidt, W. Schnelle, J. Richter, and H. Rosner, Microscopic magnetic modeling for the Na3Cu2SbO6 and Na2Cu2TeO6Phys. Rev. B 89, 174403 (2014).
  • Whangbo et al. [2003] M.-H. Whangbo, H.-J. Koo, and D. Dai, Spin exchange interactions and magnetic structures of extended magnetic solids with localized spins: theoretical descriptions on formal, quantitative and qualitative levels, J. Solid State Chem. 176, 417 (2003).
  • Derakhshan et al. [2007] S. Derakhshan, H. L. Cuthbert, J. E. Greedan, B. Rahaman, and T. Saha-Dasgupta, Electronic structures and low-dimensional magnetic properties of the ordered rocksalt oxides Na3Cu2SbO6 and Na2Cu2TeO6Phys. Rev. B 76, 104403 (2007).