Abstract
The Kitaev spin liquid, a ground state of the bond-dependent Kitaev model in a honeycomb lattice has been a center of attraction, since a microscopic theory to realize such an interaction in solid-state materials was discovered. A challenge in real materials though is the presence of the Heisenberg and another bond-dependent Gamma interactions detrimental to the Kitaev spin liquid, and there have been many debates on their relative strengths. Here we offer a strategy to extract the Kitaev interaction out of a full microscopic model by utilizing the symmetries of the Hamiltonian. Two tilted magnetic field directions related by a two-fold rotational symmetry generate distinct spin excitations originated from a specific combination of the Kitaev and Gamma interactions. Together with the in- and out-of-plane magnetic anisotropy, one can determine the Kitaev and Gamma interactions separately. Dynamic spin structure factors are presented to motivate future experiments. The proposed setups will advance the search for Kitaev materials.
Similar content being viewed by others
Introduction
An electron’s orbital motion in an atom generates a magnetic field that influences its spin moment, known as spin-orbit coupling. When the coupling is strong in heavy atoms, the effective Hamiltonian is described by the spin-orbit-entangled pseudospin wave-function and the interactions among magnetic ions are highly anisotropic different from the standard Heisenberg interaction1,2,3,4,5,6. A fascinating example is the Kitaev model with a bond-dependent interaction in a two-dimensional honeycomb lattice, whose ground state is a quantum spin liquid (QSL) with Majorana fermions and Z2 vortex excitations7. There have been extensive studies on the model because in the Kitaev QSL non-Abelian excitations emerge under a magnetic field, and their braidings provide topological computation. Since a microscopic mechanism to generate such an interaction was uncovered8, intense efforts toward finding QSLs including a variety of candidate materials from spin S = 1/29,10,11,12,13,14,15,16,17,18 to higher-spin S systems have been made19,20,21,22. Despite such efforts, a confirmed Kitaev QSL is still missing.
One challenge in finding the Kitaev QSL in magnetic materials is the presence of other spin interactions which may generate magnetic orderings or other disordered phases23,24,25,26,27,28,29. A generic nearest neighbor (n.n.) model in an ideal honeycomb was derived which revealed the isotropic Heisenberg interaction and another bond-dependent interaction named the Gamma (Γ)25. Furthermore, there exist further neighbor interactions such as second and third n.n. Heisenberg interactions, which makes it difficult to single out the Kitaev interaction itself. There have been many debates on the relative strengths, especially between the dominant Kitaev and Gamma interactions in Kitaev candidate materials13,18,28,30, and an experimental guide on how to extract the Kitaev interaction out of a full Hamiltonian is highly desirable.
In this work, we present a symmetry-based experimental strategy to determine the Kitaev interaction. Our proposal is based on the π-rotation around the a axis perpendicular to one of the bonds in the honeycomb plane, denoted by C2a symmetry that is broken by a specific combination of the Kitaev and Γ interactions. This broken C2a can be easily detected with the help of a magnetic field applied within the a−c plane where the c-axis is perpendicular to the honeycomb plane; spin excitations under the two field angles of θ and −θ, measured away from the honeycomb plane as shown in Fig. 1a, are distinct due to the combination of the Kitaev and Gamma interactions. The two field angles are related by the π-rotation around a axis, i.e., C2a operation. Such differences are based on the symmetry and signal the relative strengths of these interactions. A magnetic ordering that further enhances the broken C2a symmetry does not alter the asymmetry, but quantifying the interaction strengths requires the size of the magnetic ordering. For this reason, a polarized state in the high-field region would be ideal for our purpose.
To determine each of the interactions, one needs to use the conventional in- vs. out-of-plane anisotropy in spin excitations. We note that the Gamma interaction affects the conventional anisotropy, but the Kitaev does not when the field is large enough to compensate for the order by disorder effect31. Thus subtracting the Gamma contribution deduced from the conventional anisotropy allows us to estimate the Kitaev interaction from the measured spin excitations under the field angles of θ and −θ. Both the conventional anisotropy and the π-rotation-related spin excitations can be measured by angle-dependent ferromagnetic resonance (FMR) or inelastic neutron scattering (INS) techniques while sweeping the magnetic field directions in the a−c plane containing the C2a rotation axis.
Below we present the microscopic model and main results based on the π-rotation symmetry around a-axis. To demonstrate our theory, we also show the FMR and dynamical spin structure factors (DSSF) obtained by exact diagonalization (ED). We analyze the different spin excitations under the two field angles at finite momenta using the linear spin-wave theory (LSWT), which further confirms our results based on the symmetry argument. Our results will guide a future search of Kitaev materials.
Result
Model
The generic spin-exchange Hamiltonian among magnetic sites with strong spin-orbit coupling for the ideal edge-sharing octahedra environment in the octahedral x−y−z axes shown in Fig. 1a contains the Kitaev (K), Gamma (Γ), and Heisenberg (J) interactions25:
where \({{{{{{{\bf{S}}}}}}}}=\frac{1}{2}\overrightarrow{\sigma }\) with ℏ ≡ 1 and \(\overrightarrow{\sigma }\) is Pauli matrix, 〈ij〉 denotes the n.n. magnetic sites, and αβ(γ) denotes the γ bond taking the α and β spin components (α, β, γ ∈ {x, y, z}). The x-, y-, and z-bonds are shown in red, blue, and green colors, respectively in Fig. 1a. Further neighbor interactions and trigonal distortion allowed interactions, and their effects will be discussed later.
To analyze the symmetry of the Hamiltonian, we rewrite the model in the a−b−c axes32,33,34:
where \({\phi }_{\gamma }=0,\frac{2\pi }{3}\), and \(\frac{4\pi }{3}\) for γ = z-, x-, and y-bond respectively, and the exchange interactions are given by
The Hamiltonian \({{{{{{{\mathcal{H}}}}}}}}\) is invariant under π-rotation around the b axis denoted by C2b and \(\frac{2\pi }{3}\)-rotation around the c axis by C3c in addition to the inversion and time-reversal symmetry.
Our proposed experimental design is based on the observation that the \({{{{{{{\mathcal{H}}}}}}}}\) is not invariant under π-rotation about the a axis C2a due to the presence of only Jac, i.e., if Jac = 0, C2a is also a symmetry of \({{{{{{{\mathcal{H}}}}}}}}\). Since the C2a is broken by Jac, if there is a way to detect the broken C2a, that will signal the strength of Jac. We note that the magnetic field sweeping from the c axis to a axis within the a−c-plane does the job. The fields with angles of θ (blue line) and −θ (red line) for \(0 \, < \, \theta \, < \, \frac{\pi }{2}\) shown in Fig. 1b, c are related by C2a rotation, and thus measuring the spin excitation difference between these two field directions will detect the strength of Jac.
To prove our symmetry argument, we consider a full model with a magnetic field. Under a magnetic field, the total Hamiltonian including the Zeeman term is given by
where the external field \(\overrightarrow{h}\) has the polar angle θ measured away from the a−b honeycomb plane and the azimuthal angle ϕ from the a-axis as shown in Fig. 1b. The magnetic anisotropy in the spin excitation energies is defined as ωn(θ) = En(θ) − E0(θ), where En and E0 are the excited and ground-state energy respectively. This anisotropy is affected by all interactions other than the isotropic Heisenberg limit (JXY = JZ), making it difficult to quantify the effect of individual interactions. However, if we compare the two excitation anisotropies, ωn(θ) and ωn(−θ) for a given strength h and ϕ = 0 as shown in Fig. 1c, related by C2a symmetry transformation, we can eliminate the effects of all other interactions except Jac thanks to symmetries of the model. Since our theory relies on the symmetry of the Hamiltonian, the ground state should break the C2a symmetry only explicitly from the Jac term. The magnetic field also contributes to the C2a breaking, but by comparing two angles of θ and −θ, the effect of Jac is isolated.
We focus on the lowest energy excitation n = 1 which gives a dominant resonance at low temperatures, and drop the n in ωn from now on for simplicity, even though our proposal works for all n. We define the excitation anisotropy between the magnetic field with angles of θ and −θ as δωK(θ) ≡ ω(θ) − ω( −θ) for \(0 \, < \, \theta \, < \, \frac{\pi }{2}\), and the conventional anisotropy between in- and out-of-plane fields as \(\delta {\omega }_{A}\equiv \omega (\theta =0)-\omega (\theta =\frac{\pi }{2})\). Below we first show how δωK arises from Jac under the field in the a−c plane based on the symmetry.
Symmetry analysis
To understand the origin of a finite δωK for ϕ = 0 under the magnetic field sweep, we first begin with a special case when \(\phi =\frac{\pi }{2}\), i.e, when the external field is in the b−c plane. This is a special case where δωK = 0 for the following reason.
The Zeeman terms due to the field with the angle θ and with −θ are related by a π rotation of the field about the \(\hat{b}\) axis, denoted by
The same can be achieved by a π-rotation of the lattice,
which also indicates \({{{{{{{\mathcal{H}}}}}}}}\) is invariant under C2b. While \({{{{{{{{\mathcal{H}}}}}}}}}_{B}\) breaks the C2b symmetry of \({{{{{{{\mathcal{H}}}}}}}}\), the total Hamiltonian \({{{{{{{\mathcal{H}}}}}}}}+{{{{{{{{\mathcal{H}}}}}}}}}_{B}(\theta )\) and \({{{{{{{\mathcal{H}}}}}}}}+{{{{{{{{\mathcal{H}}}}}}}}}_{B}(-\theta )\) are related by C2b and therefore, share the same eigenenergies, i.e., δωK = 0. The difference due to the field is simply removed by a π rotation of the eigenstates about the \(\hat{b}\) axis. The magnetic field sweeping from θ to −θ in the other planes equivalent to b−c plane by C3c symmetry also gives δωK = 0.
Now let us consider when the magnetic field sweeps in the a−c plane. Similarly, the magnetic field directions θ and −θ are related by
Considering a π rotation of the lattice about the \(\hat{a}\) axis,
we find JXY, JZ, Jab, terms are invariant under C2a, while the Jac terms transform as
By the same argument, if Jac = 0, \({{{{{{{\mathcal{H}}}}}}}}\) is invariant under C2a, and the eigenenergies of the total Hamiltonian for θ and −θ are the same, i.e., δωK = 0. If Jac ≠ 0, the total Hamiltonian \({{{{{{{\mathcal{H}}}}}}}}+{{{{{{{{\mathcal{H}}}}}}}}}_{B}(\theta )\) and \({{{{{{{\mathcal{H}}}}}}}}+{{{{{{{{\mathcal{H}}}}}}}}}_{B}(-\theta )\) cannot be related by C2a, and therefore, δωK ≠ 0. We need to change the sign of Jac for the C2a relation to hold, i.e., the transformation of the external field angles of θ to −θ is equivalent to the change of Jac to −Jac. Thus, the lack of C2a symmetry allows us to single out the Jac interaction through δωK.
Since Jac contains a combination of the Kitaev and Γ interactions, we need other methods to subtract the Γ contribution. The in- and out-of-plane anisotropy, δωA offers precisely the other information. We note that the in- and out-of-plane anisotropy δωA is determined by JZ − JXY = Γ. Thus, for the ideal edge-sharing octahedral environment, we can first estimate Γ from the measured δωA, and then extract the Kitaev strength by subtracting the Γ contribution from the measured δωK(θ).
Below we show numerical results of spin excitations obtained by ED on a 24-site cluster which can be measured by angle-dependent FMR and INS techniques under magnetic field angles of θ and −θ with ϕ = 0.
Angle-dependent ferromagnetic resonance
FMR is a powerful probe to study ferromagnetic or spin correlated materials. FMR spectrometers record the radio-frequency (RF) electromagnetic wave that is absorbed by the sample of interest placed under an external magnetic field. To observe the resonance signal, the resonant frequency of the sample is changed to match that of the RF wave under a scan of the external magnetic field, so the excitation anisotropy δω(θ) leads to the anisotropy in the resonant magnetic field. FMR provides highly resolved spectra over a large energy range and has been used to investigate exchange couplings35,36,37,38 and anisotropies39,40 due to its dependence on the magnetic field angle. Here, for simplicity, we calculate the excitation energy probed by the RF field (details can be found in the Methods) with a set magnetic field strength for spin \(\frac{1}{2}\) using ED on a C3-symmetric 24-site cluster.
We set our units the magnetic field h = 1 and g = μB ≡ 1, leading to the excitation energy of a free spin, ω0 = gμBh = 1, so the excitation energies calculated are normalized by ω0. A few sets of different interaction parameters (in units of ω0) are investigated. Figure 2(a) shows the J = −1 and K = Γ = 0.5 case with no δωK(θ) between −π/2 < θ < 0 (red line) and 0 < θ < π/2 (blue line), since Jac = 0. The conventional anisotropy δωA is finite, because the Γ interaction generates a strong anisotropy between the plane θ = 0 and the c-axis θ = π/2, i.e., JXY ≠ JZ due to a finite Γ contribution. The black line is for only J = −1 showing a uniform FMR independent of angles which serves as a reference. Figure 2b shows the J = −1, K = 1, and Γ = 0 case, which shows a finite δωK(θ) between −π/2 < θ < 0 and 0 < θ < π/2 in the a−c plane. On the other hand, no δωK(θ) by sweeping θ in the b−c plane (up and down triangles with green line) is observed, consistent with the symmetry analysis presented above. Note the conventional anisotropy δωA in both a−c and b−c planes are not exactly zero, because the Kitaev interaction selects the magnetic moment along the cubic axes in the ferromagnetic state via order by disorder31,41. This leads to a tiny anisotropy between the plane θ = 0 and the c-axis θ = π/2 when Γ = 0 and JXY = JZ. This anisotropy becomes weaker when the magnetic field increases, i.e, when the moment polarization overcomes the order by disorder effect. Supplementary Note 1 shows that the anisotropy is almost gone when the field is increased by three times with the same set of parameters, where the Heisenberg limit (black line) is added for reference. When Γ becomes finite favouring either the a−b plane or the c axis depending on the sign of the Γ, this conventional anisotropy is determined by the Γ interaction as shown in Fig. 2c, d, and the order by disorder effect becomes silent. Figure 2c shows the J = −0.5, Γ = 0.5, and K = 0 case. The Γ interaction alone can generate a finite δωK due to the broken C2a by Jac. In addition, the Γ interaction generates a large δωA, different from Fig. 2b. Figure 2d presents the J = −0.1, K = −1, and Γ = 0.5 case, which is close to a set of parameters proposed for J\({}_{{{{{{{{\rm{eff}}}}}}}}}=\frac{1}{2}\) Kitaev candidate materials28. Clearly, δωK(θ) is significant due to a finite Jac, and δωA is also large due to a finite Γ. While a magnetic field of strength h = 1 is used to polarize the ground state where the finite-size effect is small as shown in Supplementary Note 2, our symmetry argument works for any finite field. However, we note that the finite-size effect of ED is minimal when the ground state is polarized.
Inelastic neutron scattering
Complementary to FMR, INS can measure excitations between different points in the reciprocal space based on the momentum transfer of the scattered neutrons. The magnon dispersions of the ordered states of magnetic materials measured via INS have been used to determine the spin-exchange Hamiltonian parameters10,17,42,43,44,45,46. Figure 3a, b show the spin excitations at accessible wavevectors on a C3-symmetric 24-site cluster with the same exchange parameters for Fig. 2d and with h = 1 and h = 8, respectively. The cluster and the accessible momenta are shown in c, d, respectively. We set the magnetic field angles θ = 30∘ (blue) and θ = −30∘ (red) in a−c plane. The square boxes denote the excitation energies obtained by the ED, and the color bars indicate the intensity of DSSF ∑αSαα(q, ω) (details can be found in the Methods). The structure factor is convolved with a Gaussian of finite width to emulate finite experimental resolution. We observe a clear difference between the two field directions, δωK at every momentum points. In particular, δωK is the largest at M2-point, while it is tiny at the K1-point. Note that M1 and M3 are related by the C2b and inversion.
To gain more insights of δωK(θ) at finite momenta obtained by ED, we also perform LSWT calculations with the magnetization making an angle θM as indicated in Fig. 1b. θM is found via minimizing the classical ground state energy (details can be found in the Methods); the LSWT with the set of parameters used for Fig. 3a’s ED results leads to θM ~ 12.1∘. The spin excitations within the LSWT are shown as dashed lines together with the ED results in Fig. 3a. The mismatch between LSWT and ED is visible at every momentum, which implies the significant effects of nonlinear terms47.
However, when the field increases, the difference should decrease, since the magnetic polarization increases at a higher field. In Fig. 3b, we show both ED and LSWT with h = 8 and θM ~ 25.8∘, where the two results match well as expected, and the nonlinear terms become less significant. In particular, the anisotropy δωK at the K-point at the high-field limit given by the leading terms in 1/h, is simplified as
where θM(θ) → θ when h → ∞. This shows that both Jac and Jab should be finite for a finite δωK at the K-point, which explains no splitting of δωK at the K-point in Fig. 3b, as our choice of parameters gives Jab = 0, i.e, Γ = −K/2. On the other hand, at the M2-point, there is no simple expression, but the leading terms of δωK(θ) in δθa/c around the a- and c axis (δθa = 0 − θ and δθc = θ − π/2) are given by
where A and C are functions of other interactions given in Supplementary Note 3. Clearly, δωK(θ) appears as odd powers of Jac and δθa/c, consistent with the symmetry analysis presented above.
So far, we have focused on the ideal octahedra environment. However, trigonal distortion is often present, albeit small, which introduces extra exchange interactions. Below we discuss other contributions to δωA complicating the isolation of K from Jac and our resolution of such complication in order to estimate the Kitaev interaction out of a full Hamiltonian.
Effects of trigonal distortion and further neighbor interactions
In principle, there are other small but finite interactions; few examples in \(\delta {{{{{{{{\mathcal{H}}}}}}}}}^{\prime}\) include
where \({{{\Gamma }}}^{\prime}\) is introduced when a trigonal distortion is present26; J2 and J3 are the second and third n.n. Heisenberg interactions respectively. It is natural to expect that they are smaller than the n.n. Kitaev, Gamma, and Heisenberg interactions18,27,28. Several types of interlayer exchange interactions are present, but they are even smaller than the terms considered in Eq. (12)18.
Let us investigate how they affect the above analysis done for the ideal n.n. Hamiltonian. First of all, the isotropic interactions such as further neighbor J2, J3, and the interlayer Heisenberg do not make any change to our proposal, since they do not contribute to δωA nor δωK. On the other hand, the \({{{\Gamma }}}^{\prime}\) modifies the exchange parameters as follows:
The conventional anisotropy δωA is now due to \({{\Gamma }}+2{{{\Gamma }}}^{\prime}\) obtained from JZ − JXY. Thus to single out the Kitaev interaction, one has to find both Γ and \({{{\Gamma }}}^{\prime}\), as Jac is a combination of K, Γ, and \({{\Gamma }}^{\prime}\). Once the trigonal distortion is present, the g-factor also becomes anisotropic, i.e., the in-plane ga is different from the c axis gc, which affects δωA.
However, the g-factor anisotropy does not affect the δωK, since the field angles of θ and −θ involve the same strength of in- and out-of-plane field components, i.e, \({{{{{{{\bf{h}}}}}}}}(\theta )={h}_{a}\hat{a}+{h}_{c}\hat{c}\) and \({{{{{{{\bf{h}}}}}}}}(-\theta )=-{h}_{a}\hat{a}+{h}_{c}\hat{c}\). Thus we wish to extract the information of K and \({{\Gamma }}-{{{\Gamma }}}^{\prime}\) from δωK, as it is free from the g-factor anisotropy.
We note that δωK at the K-point, Eq. (10) offers both Jac and Jab from the first term independent of the field and the next term proportional to 1/heff (\({h}_{{{{{{{{\rm{eff}}}}}}}}}=h\sqrt{{g}_{a}^{2}{\cos }^{2}\theta +{g}_{c}^{2}{\sin }^{2}\theta }\)). Once Jac and Jab are deduced, K and \({{\Gamma }}-{{\Gamma }}^{\prime}\) can be estimated from Eq. (13). The measurements of δωK at the K-point with a large magnetic field then determine K and \({{\Gamma }}-{{{\Gamma }}}^{\prime}\) separately. Further neighbor Heisenberg interactions, J2 and J3 do not modify Eq. (10) in the high-field limit, so they do not affect our procedure.
Discussion
We propose an experimental setup to single out the Kitaev interaction for honeycomb Mott insulators with edge-sharing octahedra. In an ideal octahedra cage, the symmetry-allowed n.n. interactions contain the Kitaev, another bond-dependent Γ, and Heisenberg interactions. We prove that the magnetic anisotropy related by the π-rotation around the a-axis denoted by δωK occurs only when a combination of K and Γ, i.e. K − Γ, is finite. This can be measured from the spin excitation energy differences under the magnetic field of angle sweeping from above to below the honeycomb plane using the FMR or INS techniques. Since the in- and out-of-plane magnetic anisotropy, δωA is determined solely by Γ, one can estimate Γ strength first from δωA and then extract the Kitaev interaction from δωK.
While the trigonal distortion introduces an additional interaction, the Kitaev interaction is unique as it is the only interaction that contributes to δωK without altering δωA. Our theory is applicable to all Kitaev candidate materials including an emerging candidate RuCl3. In particular, since the two dominant interactions are ferromagnetic Kitaev and positive Γ interactions in RuCl33,5,18,27, leading to a large Jac and a small Jab, we predict that δωK independent of the g-factor anisotropy is significant except at the K-point. Supplementary Note 4 shows the FMR and INS of a set of parameters with a small negative \({{{\Gamma }}}^{\prime}\) interaction to stabilize a zero-field zig-zag ground state as in RuCl318,27,28. Another relevant perturbation in some materials is the effect of monoclinic structure which loses the C3c symmetry of \(R\bar{3}\), making the z-bond different from the x- and y-bonds. The current theory of finite δωK due to a finite Jac still works for C2/m structure. However, since the z-bond of \({J}_{ac}^{z}(={K}_{z}/3-{{{\Gamma }}}_{z}/3)\) is no longer the same as the x- and y-bonds of \({J}_{ac}^{x}(={J}_{ac}^{y})\) and C2a symmetry relates between the x- and y-bonds, the anisotropy δωK at different momenta, detecting both \({J}_{ac}^{x}={K}_{x}/3-{{{\Gamma }}}_{x}/3 \;{{{\mbox{and }}}}\; {J}_{ac}^{z}\), is required to determine different x- and z-bond strengths.
The symmetry-based theory presented here is also valid for higher-spin models with the Kitaev interaction such as S = 3/2 CrI3 including a nonzero single-ion anisotropy19,21,22,48 which generates a further anisotropy in δωA but does not affect the δωK. The next n.n. Dzyaloshinskii–Moriya interaction with the d-vector along the c-axis49 is also invariant under the C2a symmetry. Further studies for higher-spin models remain to be investigated to identify higher-spin Kitaev spin liquid. We would like to emphasize that the proposed setup is suitable for other experimental techniques such as low-energy terahertz optical and nuclear magnetic resonance spectroscopies that probe spin excitations in addition to the angle-dependent FMR and INS spectroscopy shown in this work as examples.
Methods
Exact Diagonalization Simulations
Numerical ED was used to compute spin excitations under a magnetic field. ED was performed on a 24-site honeycomb cluster with periodic boundary conditions, where the Lanczos method50,51 was used to obtain the lowest-lying eigenvalues and eigenvectors of the Hamiltonian in Eq. (2). The 24-site honeycomb shape and accessible momentum points in the Brillouin zone are shown in Fig. 3c, d. The probability of the spin excitation of momentum q and energy ω is proportional to the DSSF52 given by
where the Lehmann representation is used; \(\left|\lambda \right\rangle\) and \(\left|{\lambda }^{\prime}\right\rangle\) are the eigenstates with the thermal population factor pλ, and Sα,β are the spin operators. In the low temperatures, we take \(\left|\lambda \right\rangle\) to be the ground state \(\left|0\right\rangle\) and we are interested in the lowest energy excitation to \(\left|1\right\rangle\) with a nonzero probability. For optical spectroscopies such as FMR, α = β = direction of the RF electromagnetic field and q = 0, so \(\left|0\right\rangle\) and \(\left|1\right\rangle\) belong to the same momentum sector. The structure factor simplifies to
For INS, the finite q must match the difference in the momenta of \(\left|0\right\rangle\) and \(\left|1\right\rangle\). For simplicity, we calculate the DSSF for α = β,
Linear spin-wave theory
The Hamiltonian in Eq. (2) is bosonized by the standard Holstein-Primakoff transformation53 expanded to linear order in the spin S:
where the quantization axis is parallel to the c-axis. The Fourier transforms are \({a}_{j}=\frac{1}{\sqrt{N}}{\sum }_{{{{{{{{\boldsymbol{k}}}}}}}}}{e}^{i{{{{{{{\boldsymbol{k}}}}}}}}\cdot {{{{{{{{\boldsymbol{r}}}}}}}}}_{j}}{a}_{{{{{{{{\boldsymbol{k}}}}}}}}}\) for sublattice A and \({b}_{j}=\frac{1}{\sqrt{N}}{\sum }_{{{{{{{{\boldsymbol{k}}}}}}}}}{e}^{i{{{{{{{\boldsymbol{k}}}}}}}}\cdot ({{{{{{{{\boldsymbol{r}}}}}}}}}_{j}+\delta )}{b}_{{{{{{{{\boldsymbol{k}}}}}}}}}\) for sublattice B, where δ is the vector pointing to nearest neighbors. The resulting quadratic Hamiltonian has the form \({{{{{{{\mathcal{H}}}}}}}}={\sum }_{{{{{{{{\boldsymbol{k}}}}}}}}}{{{{{{{{\rm{X}}}}}}}}}^{{{{\dagger}}} }H({{{{{{{\boldsymbol{k}}}}}}}}){{{{{{{\rm{X}}}}}}}}\), where \({{{{{{{{\rm{X}}}}}}}}}^{{{{\dagger}}} }=(\begin{array}{cccc}{a}_{{{{{{{{\boldsymbol{k}}}}}}}}}^{{{{\dagger}}} },&{b}_{{{{{{{{\boldsymbol{k}}}}}}}}}^{{{{\dagger}}} },&{a}_{-{{{{{{{\boldsymbol{k}}}}}}}}},&{b}_{-{{{{{{{\boldsymbol{k}}}}}}}}}\end{array})\). Diagonalizing this BdG Hamiltonian following standard methods54 gives two spin-wave excitation branches.
For a general field, the Hamiltonian in Eq. (2) is first written in new axes \({{{{{{{{\boldsymbol{a}}}}}}}}}^{\prime}{{{{{{{{\boldsymbol{b}}}}}}}}}^{\prime}{{{{{{{{\boldsymbol{c}}}}}}}}}^{\prime}\). \({{{{{{{{\boldsymbol{a}}}}}}}}}^{\prime}=(\sin {\theta }_{M}\cos {\phi }_{M},\sin {\theta }_{M}\sin {\phi }_{M},-\cos {\theta }_{M})\), \({{{{{{{{\boldsymbol{b}}}}}}}}}^{\prime}=(-\sin {\phi }_{M},\cos {\phi }_{M},0)\) and \({{{{{{{{\boldsymbol{c}}}}}}}}}^{\prime}=(\cos {\theta }_{M}\cos {\phi }_{M},\cos {\theta }_{M}\sin {\phi }_{M},\sin {\theta }_{M})\). \({{{{{{{{\boldsymbol{c}}}}}}}}}^{\prime}\) is parallel to the magnetization S(θM, ϕM), which is not in the same direction as the magnetic field, unless the field is very large to fully polarize the moment. The magnetization angles (θM, ϕM) are obtained by minimizing the classical ground state energy, and LSWT is applied to the ground state47. Arbitrary \({{{{{{{{\boldsymbol{a}}}}}}}}}^{\prime}\) and \({{{{{{{{\boldsymbol{b}}}}}}}}}^{\prime}\) axes obtained by rotation around \({{{{{{{{\boldsymbol{c}}}}}}}}}^{\prime}\) are valid and do not affect the result.
Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Code availability
The code used to generate the data used in this study is available from the corresponding author upon reasonable request.
References
Witczak-Krempa, W., Chen, G., Kim, Y. B. & Balents, L. Correlated quantum phenomena in the strong spin-orbit regime. Annu. Rev. Condens. Matter Phys. 5, 57–82 (2014).
Rau, J. G., Lee, E. K.-H. & Kee, H.-Y. Spin-orbit physics giving rise to novel phases in correlated systems: iridates and related materials. Annu. Rev. Condens. Matter Phys. 7, 195–221 (2016).
Winter, S. M. et al. Models and materials for generalized Kitaev magnetism. J. Phys. Condens. Matter 29, 493002 (2017).
Hermanns, M., Kimchi, I. & Knolle, J. Physics of the Kitaev model: fractionalization, dynamic correlations, and material connections. Annu. Rev. Condens. Matter Phys. 9, 17–33 (2018).
Janssen, L. & Vojta, M. Heisenberg-kitaev physics in magnetic fields. J. Phys. Condens. Matter 31, 423002 (2019).
Takayama, T., Chaloupka, J., Smerald, A., Khaliullin, G. & Takagi, H. Spin–orbit-entangled electronic phases in 4d and 5d transition-metal compounds. J. Phys. Soc. Jpn. 90, 062001 (2021).
Kitaev, A. Anyons in an exactly solved model and beyond. Ann. Phys. 321, 2–111 (2006).
Jackeli, G. & Khaliullin, G. Mott insulators in the strong spin-orbit coupling limit: from Heisenberg to a quantum compass and Kitaev models. Phys. Rev. Lett. 102, 017205 (2009).
Singh, Y. et al. Relevance of the Heisenberg-Kitaev model for the honeycomb lattice iridates A2IrO3. Phys. Rev. Lett. 108, 127203 (2012).
Choi, S. K. et al. Spin waves and revised crystal structure of honeycomb iridate Na2IrO3. Phys. Rev. Lett. 108, 127204 (2012).
Plumb, K. W. et al. α-RuCl3: a spin-orbit assisted mott insulator on a honeycomb lattice. Phys. Rev. B 90, 041112(R) (2014).
Modic, K. A. et al. Realization of a three-dimensional spin-anisotropic harmonic honeycomb iridate. Nat. Commun. 5, 4203 (2014).
Kim, H.-S., Shankar V., V., Catuneanu, A. & Kee, H.-Y. Kitaev magnetism in honeycomb α-RuCl3 with intermediate spin-orbit coupling. Phys. Rev. B 91, 241110(R) (2015).
Sears, J. A. et al. Magnetic order in α-RuCl3: a honeycomb-lattice quantum magnet with strong spin-orbit coupling. Phys. Rev. B 91, 144420 (2015).
Sandilands, L. J., Tian, Y., Plumb, K. W., Kim, Y.-J. & Burch, K. S. Scattering continuum and possible fractionalized excitations in α-RuCl3. Phys. Rev. Lett. 114, 147201 (2015).
Johnson, R. D. et al. Monoclinic crystal structure of α-RuCl3 and the zigzag antiferromagnetic ground state. Phys. Rev. B 92, 235119 (2015).
Banerjee, A. et al. Proximate Kitaev quantum spin liquid behaviour in a honeycomb magnet. Nat. Mater. 15, 733 (2016).
Kim, H.-S. & Kee, H.-Y. Crystal structure and magnetism in α-RuCl3: an ab initio study. Phys. Rev. B 93, 155143 (2016).
Stavropoulos, P. P., Pereira, D. & Kee, H.-Y. Microscopic mechanism for a higher-spin kitaev model. Phys. Rev. Lett. 123, 037203 (2019).
Lado, J. L. & Fernández-Rossier, J. On the origin of magnetic anisotropy in two dimensional CrI3. 2D Mater. 4, 035002 (2017).
Xu, C., Feng, J., Xiang, H. & Bellaiche, L. Interplay between kitaev interaction and single ion anisotropy in ferromagnetic CrI3 and CrGeTe3 monolayers. npj Comput. Mater. 4, 57 (2018).
Lee, I. et al. Fundamental spin interactions underlying the magnetic anisotropy in the kitaev ferromagnet CrI3. Phys. Rev. Lett. 124, 017201 (2020).
Chaloupka, Jcv, Jackeli, G. & Khaliullin, G. Kitaev-Heisenberg model on a honeycomb lattice: possible exotic phases in iridium oxides A2IrO3. Phys. Rev. Lett. 105, 027204 (2010).
Chaloupka, Jcv, Jackeli, G. & Khaliullin, G. Zigzag magnetic order in the iridium oxide Na2IrO3. Phys. Rev. Lett. 110, 097204 (2013).
Rau, J. G., Lee, E. K.-H. & Kee, H.-Y. Generic spin model for the honeycomb iridates beyond the Kitaev limit. Phys. Rev. Lett. 112, 077204 (2014).
Rau, J. G. & Kee, H.-Y. Trigonal distortion in the honeycomb iridates: proximity of zigzag and spiral phases in Na2IrO3. arxiv https://arxiv.org/abs/1408.4811 (2014).
Winter, S. M., Li, Y., Jeschke, H. O. & Valentí, R. Challenges in design of Kitaev materials: magnetic interactions from competing energy scales. Phys. Rev. B 93, 214431 (2016).
Janssen, L., Andrade, E. C. & Vojta, M. Magnetization processes of zigzag states on the honeycomb lattice: identifying spin models for α-RuCl3 and Na2IrO3. Phys. Rev. B 96, 064430 (2017).
Luo, Q., Zhao, J., Kee, H.-Y. & Wang, X. Gapless quantum spin liquid in a honeycomb γ magnet. npj Quant. Mater. 6, 57 (2021).
Maksimov, P. A. & Chernyshev, A. L. Rethinking α−RuCl3. Phys. Rev. Res. 2, 033011 (2020).
Chaloupka, J. & Khaliullin, G. Magnetic anisotropy in the kitaev model systems Na2IrO3 and RuCl3. Phys. Rev. B 94, 064435 (2016).
Onoda, S. Effective quantum pseudospin-1/2 model for Yb pyrochlore oxides. J. Phys.:Conf. Ser. 320, 012065 (2011).
Ross, K. A., Savary, L., Gaulin, B. D. & Balents, L. Quantum excitations in quantum spin ice. Phys. Rev. X 1, 1–10 (2011).
Chaloupka, Jcv & Khaliullin, G. Hidden symmetries of the extended kitaev-heisenberg model: implications for the honeycomb-lattice iridates A2IrO3. Phys. Rev. B 92, 024413 (2015).
Z, Z., L, Z., PE, W. & K, O. Angular dependence of ferromagnetic resonance in exchange-coupled Co/Ru/Co trilayer structures. Phys. Rev. B, Condens. Matter 50, 6094–6112 (1994).
B, H. et al. Dynamic exchange coupling in magnetic bilayers. Phys. Rev. Lett. 90, 4 (2003).
Nascimento, V. P. et al. Ferromagnetic resonance study of the exchange bias field in NiFe, FeMn, NiFe trilayers. J. Appl. Phys. 99, 08C108 (2006).
Lenz, K., Kosubek, E., Toliński, T., Lindner, J. & Baberschke, K. In situ ferromagnetic resonance in coupled ultrathin trilayers with perpendicularly oriented easy axes. J. Phys. Condens. Matter 15, 7175 (2003).
Díaz de Sihues, M., Durante-Rincón, C. A. & Fermin, J. R. A ferromagnetic resonance study of NiFe alloy thin films. J. Magn. Magn. Mater. 316, e462–e465 (2007).
VG, G., A, D., BD, S. & SP, K. A study of the magnetic resonance in a single-crystal Ni(50.47)Mn(28.17)Ga(21.36) alloy. J. Phys. Condens. Matter: Inst. Phys. J. 18, 7613–7627 (2006).
Stavropoulos, P. P., Catuneanu, A. & Kee, H.-Y. Counter-rotating spiral order in three-dimensional iridates: signature of hidden symmetry in the kitaev-Γ model. Phys. Rev. B 98, 104401 (2018).
Shirane, G., Minkiewicz, V. J. & Nathans, R. Spin waves in 3d metals. J. Appl. Phys. 39, 383 (1968).
Endoh, Y., Shirane, G., Birgeneau, R. J., Richards, P. M. & Holt, S. L. Dynamics of an span class. Phys. Rev. Lett. 32, 170 (1974).
Ran, K. et al. Spin-wave excitations evidencing the kitaev interaction in single crystalline α-RuCl3. Phys. Rev. Lett. 118, 107203 (2017).
Do, S. H. et al. Majorana fermions in the Kitaev quantum spin system α-RuCl3. Nat. Phys. 13, 1079–1084 (2017).
Banerjee, A. et al. Excitations in the field-induced quantum spin liquid state of α-rucl3. npj Quant. Mater. 3, 8 (2018).
Cônsoli, P. M., Janssen, L., Vojta, M. & Andrade, E. C. Heisenberg-kitaev model in a magnetic field: 1/s expansion. Phys. Rev. B 102, 155134 (2020).
Stavropoulos, P. P., Liu, X. & Kee, H.-Y. Magnetic anisotropy in spin-3/2 with heavy ligand in honeycomb mott insulators: application to CrI3. Phys. Rev. Res. 3, 013216 (2021).
Chen, L. et al. Magnetic field effect on topological spin excitations in CrI3. Phys. Rev. X 11, 031047 (2021).
Lanczos, C. An iteration method for the solution of the eigenvalue problem of linear differential and integral operators. J. Res. Natl. Bur. Stand. 45, 282 (1950).
Weiße, A & Fehske, H. Exact Diagonalization Techniques. 529–544 (Springer Berlin Heidelberg: Berlin, Heidelberg, 2008). .
Lovesey, S. W. S. W. Theory of neutron scattering from condensed matter (Clarendon Press, 1984).
Holstein, T. & Primakoff, H. Field dependence of the intrinsic domain magnetization of a ferromagnet. Phys. Rev. 58, 1098 (1940).
Colpa, J. H. Diagonalization of the quadratic boson hamiltonian. Phys. A: Stat. Mech. its Appl. 93, 327–353 (1978).
Acknowledgements
We thank J. Gordon, I. Lee, C. Hammel, S. Nagler, and A. Tennant for useful discussions. This work was supported by the Natural Sciences and Engineering Research Council of Canada and the Canada Research Chairs Program. This research was enabled in part by support provided by Sharcnet (http://www.sharcnet.ca) and Compute Canada (http://www.computecanada.ca). Computations were performed on the GPC and Niagara supercomputers at the SciNet HPC Consortium. SciNet is funded by: the Canada Foundation for Innovation under the auspices of Compute Canada; the Government of Ontario; Ontario Research Fund - Research Excellence; and the University of Toronto.
Author information
Authors and Affiliations
Contributions
Exact diagonalization and linear spin-wave theory calculations were performed by J.C. The symmetry analysis was done by H.-Y.K. and J.C. H.-Y.K. planned and supervised the project. All authors wrote the manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests. Hae-Young Kee is an Editorial Board Member for Communications Physics, but was not involved in the editorial review of, or the decision to publish this article.
Peer review
Peer review information
Communications Physics thanks Lukas Janssen, Willian Natori, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Cen, J., Kee, HY. Strategy to extract Kitaev interaction using symmetry in honeycomb Mott insulators. Commun Phys 5, 119 (2022). https://doi.org/10.1038/s42005-022-00893-4
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s42005-022-00893-4
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.