betaFeedback

WorksPhysics and Astronomy

On the Ferrimagnetic State of CrCl2_2(pyz)2_2

Abstract

Van der Waals layered ferromagnetic compounds with high two-dimensional electronic conductivity holds strong potential for quantum computing, future unconventional superconductors, catalysts, batteries, and fuel cells. We suggest a minimal theoretical model to understand the magnetic properties of the metal-organic framework CrCl2_2(pyz)2_2 (pyz=pyrazine). Using a Hubbard model we show that the groundstate is dominated by a specific configuration of delocalized electrons on the pyz sites with a ferrimagnetic coupling to the localized spins on the Cr sites. This model suggests a magnetic moment of 2μB2μ_B which is remarkably close to the experimental value of 1.8μB1.8 μ_B [K. S. Pedersen et al., Nat. Chem. 10, 1056-1061 (2018)]. From Weiss mean-field theory we predict a weak ferromagnetic Cr-Cr coupling of ≈0.9\approx 0.9 meV. This is consolidated by second order perturbation theory of the RKKY interaction yielding a Cr-Cr coupling of ≈5\approx 5 meV. Understanding the interactions in these types of compounds can facilitate designs of metal-organic compounds with tailored magnetic properties.

Introduction

Metal-organic frameworks (MOFs) have been predicted to serve as a platform for the next generation of quantum technology [1, 2]. The diversity of organic ligands compared to metal ligands allow for enhanced tunability of electronic and magnetic properties. These compounds may be key to harvest renewed insights into and exploitation of quantum phenomena such as the Quantum Hall Effect [3, 4, 5], topological insulators [6] and superconductivity [7]. The putatively first organic-inorganic hybrid 2D MOF, CrCl2_2(pyz)2_2 (pyz = pyrazine), was successfully synthesized in 2018 [8]. CrCl2_2(pyz)2_2 exhibits both long-range magnetic order and high 2D electronic conductivity. The compound has a Curie temperature of 55 K, and its magnetization at low temperatures saturates to 1.8μB1.8\mu_B in strong external fields. This indicates an intralayer ferrimagnetic coupling in agreement with predictions by density functional theory (DFT) [9, 10]. First principle calculations indicate that CrCl2_2(pyz)2_2, like other hybrid materials [11, 12, 13], can be exfoliated making it relevant for 2D applications [9]. The majority of 2D materials are non-magnetic in their pristine form [14]. A magnetic 2D material, like CrCl2_2(pyz)2_2, could have applications within spintronics [15], multiferroics [16, 17, 18] and magnetoelectronics [19]. This material could possibly even have applications within dark matter detection [20, 21, 22]. The sibling coordination solids VCl2_2(pyz)2_2 and TiCl2_2(pyz)2_2 display contrasting physical properties. In the former, divalent V ions facilitate an antiferromagnetically ordered ground state, while in the latter trivalent Ti ions favor paramagnetism and a correlated Fermi liquid state [23]. In principle, this opens the possibility for tuning through exotic phase transitions via doping. This may be compared to the magnetic phase transitions proposed for Mn2−x_{2-x}Cr2_2Sb [24, 25] and Mn2_2Sb1−x_{1-x}Bix_x [26]. Finally, the perovskite S2_2FeMoO6_6 may be highlighted as another material in which the coexistence of itinerant and localized electrons cause an ordered, in this case putatively ferromagnetic, ground state [27, 28].

A crucial step towards synthesizing materials with desired magnetic properties is to understand the mechanism yielding these properties. Here we provide a simple explanation of the ferrimagnetic coupling and explain why the magnetic moment of CrCl2_2(pyz)2_2 is approximately 2μB2\mu_B. As such, we propose a minimal model for monolayer CrCl2_2(pyz)2_2 that captures the essential features of chromium couplings mediated by intermediate and itinerant pyrazine electrons.

First, we compute a tight-binding model using a Slater–Koster procedure to obtain hopping parameters. From this we propose an effective model relevant for the valence electrons. The ground state configuration of the model is investigated before a magnetic field is introduced and the magnetic moment is computed. Using mean-field theory an estimate of the Cr-Cr coupling is found. Finally, the exchange interaction between electrons on neighboring chromium and pyrazine sites is calculated and an expression of the indirect exchange interaction coupling between neighboring chromium sites is obtained using second order perturbation of the RKKY interaction. From this we predict an estimate of the coupling between the localized Cr spins and the itinerant pyz electrons.

Tight Binding Model of Monolayer CrCl2_2(pyz)2_2

Monolayer CrCl2_2(pyz)2_2 has an approximately four-fold rotational in-plane symmetry. Each layer is coupled through Van der Waals interactions. The pyrazine rings are tilted with two alternating orientations [8], see FIG. S1 in the Supplementary Material. One unit cell contains two Cl atoms, one Cr atom, and two pyrazine rings, see a fragment of a unit cell in Figure 1 a).

Fragment of one unit cell, monolayer lattice, and band structure

Figure 1. a) Illustration of a fragment of one unit cell of CrCl2_2(pyz)2_2. b) Monolayer CrCl2_2(pyz)2_2 viewed along the Cl-Cr-Cl axis. The chromium and pyrazine sites are illustrated with black and white circles, respectively. Each pyz ring has been approximated by a single sites. Further, the unit cell is marked with a dashed square and the relevant lattice vectors, a1\mathbf{a}_1 and a2\mathbf{a}_2, are indicated. c) Band structure of monolayer CrCl2_2(pyz)2_2 obtained from tight binding along the path M →\to Γ →\to X. Inset show the Brillouin zone and the relevant symmetry points.

We compute an effective tight-binding model for monolayer CrCl2_2(pyz)2_2, as summarized in Figure 1 b). The lattice is approximated as a square lattice in the Cr-pyz plane, and each pyrazine ring is modeled by a single site with two orbitals tilted by 45° compared to the in-plane Cr lattice, see Figure S1. Due to the tilting of the pyrazine sites, the Cr sites and the pyz sites form a Lieb lattice [29]. A study of the symmetries of the model complex CrCl2_2(pyz)4_4 using a group theoretical approach can be found in the Supplementary Material. The basis includes five d-orbitals on the Cr site, three p-orbitals on both Cl sites, and two pz′_{z}^{\prime} orbitals tilted in two different orientations on the two pyz sites, Φk=ΦkCr⊗ΦkCl⊗Φkpyz\Phi_k = \Phi_k^{\mathrm{Cr}} \otimes\Phi_k^{\mathrm{Cl}} \otimes\Phi_k^{\mathrm{pyz}}. Here z′ refers to the direction perpendicular to the plane spanned by the pyz rings. The Hamiltonian

Illustration of the model complex CrCl2(pyz)4

Figure S1. Illustration of the model complex CrCl2(pyz)4\mathrm{CrCl}_{2}(\mathrm{pyz})_{4}. The nomenclature of the six relevant atoms, the coordinate system in use, and the orientation of the pyrazine rings are illustrated.

H=∑kΦk†H^(k)Φk(1)H = \sum_k \Phi_k^\dagger\hat{H}(k)\Phi_k \tag*{(1)}

describes inter- and intraorbital hopping between the aforementioned orbitals, as encoded by the 15×1515 \times15 matrix H^\hat{H}. The form of H^\hat{H} is listed in Eq. (40) in the Supplementary Material. The basis is not spin-resolved. The elements in the Hamiltonian has been obtained using the Slater–Koster decomposition of the tight binding model [30, 31].

The band structure along the path M →\to Γ →\to X is displayed in Figure 1 c). The Fermi Level is located at 0 eV. This captures the qualitative, semi-metallic properties in agreement with first principle calculations [8]. The two empty d-bands of Cr are located above 1.5 eV, whereas the filled d-bands are located below the p-bands (not visible in the plot). This is also seen from first principle calculations of the PDOS [8] and could indicate a strong π-d hybridization. The flatness of the d-bands indicates a localization of the d-electrons in accordance with the behavior of a Mott insulator. The p-levels from the pyz sites are dispersive which indicates a delocalization of the p-electrons. The Cl levels are located far below the Fermi level and will be safely neglected in the minimal model. From first principle calculations the summed spin population on the pyrazine rings is calculated to −0.68 delocalized over four pyz sites of the model complex CrCl2_2(pyz)4_4 [8].

In this paper the two pyrazine rings in a unit cell will be approximated as two single sites over which one electron delocalize. Further, Ref. [8] concludes that the total spin on the Cr site is S=3/2S = 3/2. For these reasons, the Cr sites will be assumed to have a localized spin S=3/2S = 3/2, and we assume one spin S=1/2S = 1/2 to be localized over the two pyz sites.

Minimal Model

To shed light on the microscopic mechanism of the magnetic ground state of CrCl2_2(pyz)2_2 we investigate the model shown in Figure 2. This is not identical to two unit cells of a single layer of CrCl2_2(pyz)2_2 since the pyz sites are moved with respect to the two chromium atoms. The model does, however, capture the competition between kinetic energy and exchange energy, which comprise the essential low-energy physics relevant for explaining the magnetic properties of CrCl2_2(pyz)2_2. In this model the two chromium atoms are assumed to be localized S=3/2S = 3/2 spins. There are two electrons that can delocalize over the four pyrazine sites. The Cl sites are omitted in this description, yet effectively taken into account by reducing the number of electrons on the Cr sites. We include the exchange coupling between the chromium spins and the spins of the two pyrazine electrons. Moreover, the pyrazine electrons are allowed to hop between the pyrazine sites (indicated with dashed lines in Figure 2). Lastly, we include an on-site repulsive Hubbard interaction, UU, penalizing electrons residing on the same site.

Model containing two chromium sites and four pyrazine sites

Figure 2. Model containing two chromium sites (red circles labeled A and B) and four pyrazine sites (blue circles labeled 1-4). Each chromium spin couples only to three out of the four pyrazine sites (A to 1-3 and B to 2-4).

The effective Hamiltonian can thus be written as

H^=−2JS^A⋅(S^1+S^2+S^3)−2JS^B⋅(S^2+S^3+S^4)−t∑σ=↑,↓(c^1σ†c^2σ+c^2σ†c^3σ+c^1σ†c^3σ+c^2σ†c^4σ+c^3σ†c^4σ+h.c.)+2U∑i=14n^i↑n^i↓,(2)\hat{H}=-2J\hat{\mathbf{S}}_{A}\cdot\left(\hat{\mathbf{S}}_{1}+\hat{\mathbf{S}}_{2}+\hat{\mathbf{S}}_{3}\right)-2J\hat{\mathbf{S}}_{B}\cdot\left(\hat{\mathbf{S}}_{2}+\hat{\mathbf{S}}_{3}+\hat{\mathbf{S}}_{4}\right) -t\sum_{\sigma=\uparrow,\downarrow}\left(\hat{c}_{1\sigma}^{\dagger}\hat{c}_{2\sigma}+\hat{c}_{2\sigma}^{\dagger}\hat{c}_{3\sigma}+\hat{c}_{1\sigma}^{\dagger}\hat{c}_{3\sigma}+\hat{c}_{2\sigma}^{\dagger}\hat{c}_{4\sigma}+\hat{c}_{3\sigma}^{\dagger}\hat{c}_{4\sigma}+\mathrm{h.c.}\right)+2U\sum_{i=1}^{4}\hat{n}_{i\uparrow}\hat{n}_{i\downarrow}, \tag*{(2)}

where S^A\hat{\mathbf{S}}_{A} and S^B\hat{\mathbf{S}}_{B} represent the spin of the two chromium atoms and S^i\hat{\mathbf{S}}_{i} with i=1,2,3,4i=1,2,3,4 labeling the spin of the four pyrazine sites. Moreover, c^iσ†\hat{c}_{i\sigma}^{\dagger} and c^iσ\hat{c}_{i\sigma} are the creation and annihilation operators of an electron on site ii and spin σ\sigma, respectively. n^iσ=c^iσ†c^iσ\hat{n}_{i\sigma}=\hat{c}_{i\sigma}^{\dagger}\hat{c}_{i\sigma} is the spin density on site ii with spin σ\sigma. JJ is the exchange coupling and tt the hopping constant.

The spins of each chromium atom can take four SzS_z values, i.e., −3/2-3/2, −1/2-1/2, 1/21/2 and 3/23/2. Details of the diagonalization are provided in the Supplementary Material. The basis is given by ∣mSA,mSB⟩⊗∣ψ⟩\lvert m_{S_A},m_{S_B}\rangle\otimes\lvert\psi\rangle, where ∣ψ⟩\lvert\psi\rangle and ∣mSA,mSB⟩\lvert m_{S_A},m_{S_B}\rangle are the given states for the pyrazine and chromium electrons, respectively. To invoke an antiferromagnetic coupling between the spin of the pyrazine electrons and the chromium spins, JJ must be negative. The eigenstates are generally superpositions of 448 states but when all three variables, JJ, tt and UU, are comparable in size the ground state is two-fold degenerate. Surprisingly, however, two basis states in these superpositions turns out to dominate, namely

∣−3/2,−3/2⟩⊗∣0↑↑0⟩and∣3/2,3/2⟩⊗∣0↓↓0⟩.(3)\lvert-3/2,-3/2\rangle\otimes\lvert0\uparrow\uparrow0\rangle\quad\text{and}\quad\lvert3/2,3/2\rangle\otimes\lvert0\downarrow\downarrow0\rangle. \tag*{(3)}

This is a robust feature of the model checked for the physically relevant parameter ranges of tt, UU and JJ. For t=U=−J=1t=U=-J=1 the amplitudes of these states are −0.747-0.747, and is ≥0.5\ge0.5 for all parameters considered. A

two level model is exploited in the Supplementary Material which indicates that one unit cell acts as a spin-2 particle. For t=U=−J=1t=U=-J=1 the occupation probability of the basis states for one of these degenerate states is illustrated, see Figure 3, clearly showing the dominance of one basis state.

Sorted occupation probabilities of basis states

Figure 3. A sorted list of the occupation probability of each basis state for one of two degenerate ground states for t=U=−J=1t=U=-J=1. For the illustrated ground state the dominant basis state is ∣−3/2,−3/2⟩⊗∣0↑↑0⟩\lvert-3/2,-3/2\rangle\otimes\lvert0\uparrow\uparrow0\rangle. It is evident that the probability of all other basis states is small compared to the most dominant basis state. For symmetry reasons the plot of occupation probabilities for the other ground state (when sorted) looks identical, though the dominant basis state here is ∣3/2,3/2⟩⊗∣0↓↓0⟩\lvert3/2,3/2\rangle\otimes\lvert0\downarrow\downarrow0\rangle.

This result has a simple, intuitive explanation. The exchange interaction is maximal when both the pyrazine electrons and the chromium spins are antiparallel and maximal. Electrons located on sites 2 and 3 couple to both chromium spins, unlike electrons located on site 1 and 4 which only couple to one of the Cr spins. Thus, the exchange interaction is further maximized by the electrons being located at site 2 and 3. For t=0t=0 this effect is further enhanced and the amplitudes of the states shown in Eq. (3) increases to −0.845-0.845. When t≠0t\ne0 the electrons can gain energy by delocalizing and the amplitudes of the triplet states with Sz=0S_z=0 and the singlet states increases. The two pyz electrons can delocalize and still avoid each other in two different configurations for a given spin configuration. One electron can delocalize over sites 1 and 2 while the other delocalizes over 3 and 4, see Figure 4 a), or one electron delocalizes over sites 1 and 3 while the other one delocalizes over sites 2 and 4, see Figure 1 b).

Two different configurations of the spins of the pyrazine sites in the minimal model

Figure 4. The two different configurations of the spins of the pyrazine sites in the minimal model. One electron delocalizes over two sites with different exchange energy and in this way gain kinetic energy while still minimizing the on-site repulsion.

A magnetic field can be introduced by adding the term

H^B=gμBB⋅S^tot=gμBBzS^totz,(4)\hat{H}_{B}=g\mu_{B}\mathbf{B}\cdot\hat{\mathbf{S}}_{\mathrm{tot}}=g\mu_{B}B^{z}\hat{S}_{\mathrm{tot}}^{z}, \tag*{(4)}

to the Hamiltonian. We choose the magnetic field to point out of the plane. The expectation value of total spin in the minimal model is found to be ⟨S^totz⟩=2\langle\hat{S}_{\mathrm{tot}}^{z}\rangle=2. This is derived in the Supplementary Material. Since the model contains the same number of pyrazine rings and chromium atoms as two unit cells of

For CrCl2(pyz)2\mathrm{CrCl}_{2}(\mathrm{pyz})_{2} this would correspond to one unit cell having S=1S=1 which corresponds to a magnetic moment of

μmodel=2μB.(5)\mu_{\mathrm{model}}=2\mu_{B}. \tag*{(5)}

This is true both for t=0t=0 and t≠0t\ne0. This is strikingly similar to both experimental results [8] and DFT results [8, 9]. Experimentally, it is found that the saturation magnetization at high pressure and low temperature is 1.8μB1.8\mu_{B} [8]. The slightly lower value can be ascribed to the canted radical spins in the real compound and the higher spin occupation on the pyrazines in DFT calculations (0.68 compared 0.5 in this model). Thus, we have here made a minimal model focusing on the relevant physics in the system that is quantitatively precise in its predictions on physical properties.

If we generalize this model to a layer of CrCl2(pyz)2\mathrm{CrCl}_{2}(\mathrm{pyz})_{2} we would expect the ground state to look something like what is depicted in Figure 5. Again, all Cr spins are localized and couple antiferromagnetically to the spins on the pyrazine electrons which have delocalized over two neighboring sites. Obviously, this can be configured in several different ways, but it will focus on one of these, since it will be very costly to change configuration.

A sketch of one of the configurations the delocalized spins on the pyrazine ligands could choose

Figure 5. A sketch of one of the configurations the delocalized spins on the pyrazine ligands could choose.

Weiss mean-field theory for a Cr3+\mathrm{Cr}^{3+} square lattice that ignores the pyrazine rings yields a fairly weak ferromagnetic coupling, JCr−Cr=3kBTCμ0zS(S+1)≈0.9 meVJ_{\mathrm{Cr-Cr}}=\frac{3k_{B}T_{C}}{\mu_{0}zS(S+1)}\approx0.9\ \mathrm{meV} (JCr−Cr/hc≈7 cm−1J_{\mathrm{Cr-Cr}}/hc\approx7\ \mathrm{cm}^{-1}), see the Supplementary Material for further details. A more sophisticated estimate of the Cr-Cr coupling can be obtained by means of second order perturbation theory of the RKKY interaction mediated by intermediate pyrazine rings. This calculation yields JCr−Cr≈0.15J2/2tJ_{\mathrm{Cr-Cr}}\approx0.15J^{2}/2t, where tt is an effective square lattice hopping strength. Using a strong antiferromagnetic coupling of JCr−pyz≡J≈−0.25 eVJ_{\mathrm{Cr-pyz}}\equiv J\approx-0.25\ \mathrm{eV} as suggested by ab initio calculations [8] and t≈1 eVt\approx1\ \mathrm{eV} (see Figure 1), we predict that JCr−Cr≈5 meVJ_{\mathrm{Cr-Cr}}\approx5\ \mathrm{meV} (JCr−Cr/hc≈40 cm−1J_{\mathrm{Cr-Cr}}/hc\approx40\ \mathrm{cm}^{-1}). Inserting an effective Cr spin of S=1S=1, accounting for the presence of the pyz spins, brings the mean-field result closer to the perturbative estimate of the Cr-Cr coupling. Here, two independent calculations unify in a meV range prediction of the ferromagnetic Cr-Cr coupling. This coupling may be probed by neutron scattering techniques and spin wave experiments. Uniaxial strain experiments on CrCl2(pyz)2\mathrm{CrCl}_{2}(\mathrm{pyz})_{2} may be employed to strain-tune tt and reveal potentially interesting insight into the importance of the itinerant electrons and the broken four-fold symmetry. Derivations for both expressions of JCr−CrJ_{\mathrm{Cr-Cr}} is included in the Supplementary Material.

Conclusions

We have proposed both realistic and minimal models capable of explaining the ferrimagnetic state of monolayer CrCl2(pyz)2\mathrm{CrCl}_{2}(\mathrm{pyz})_{2}. From the Slater–Koster-derived band structure monolayer CrCl2(pyz)2\mathrm{CrCl}_{2}(\mathrm{pyz})_{2} is found to exhibit an insulating ground state, in agreement with first principle calculations and the experimental results [8]. Furthermore, the d-bands are observed to be flat indicative of these electrons being localized whereas the pyrazine electrons are found to be very delocalized.

To investigate the magnetic properties of monolayer CrCl2(pyz)2\mathrm{CrCl}_{2}(\mathrm{pyz})_{2} a simple model was proposed. Here, the electrons on the chromium sites are viewed as localized spins, whereas the electrons on the pyrazine sites can delocalize over a few sites. This suggests a magnetic moment of one unit cell to be 2μB2\mu_{B} which corresponds to an antiferromagnetic interaction between the chromium spins and the pyrazine spins. Due to the uneven magnitudes of the spin on the Cr and pyz sites, this results in a ferrimagnetic ordering. The calculated magnetic moment is strikingly close to the experimental value of 1.8μB1.8\mu_{B} [8]. Weiss mean-field theory estimates a direct ferrimagnetic coupling between neighboring Cr sites of JCr−Cr≈0.9 meVJ_{\mathrm{Cr-Cr}}\approx0.9\ \mathrm{meV} (JCr−Cr/hc≈7 cm−1J_{\mathrm{Cr-Cr}}/hc\approx7\ \mathrm{cm}^{-1}). Second order perturbation theory of the RKKY interaction between Cr and pyz yields a Cr-Cr coupling of the same order of magnitude, namely JCr−Cr≈5 meVJ_{\mathrm{Cr-Cr}}\approx5\ \mathrm{meV}. This Cr-Cr coupling may be probed by neutron scattering and spin wave experiments.

Acknowledgments

FLG acknowledges helpful conversations with Ivano E. Castelli and elaborate discussions and proofreading by Henrik S. Guttesen.

References

— Supplementary Material —

This is based on results that are previously reported in Ref. [32].

Group Theory for Model Complex CrCl2(pyz)4\mathrm{CrCl}_{2}(\mathrm{pyz})_{4}

First, we investigate the energy levels and symmetries of the model complex CrCl2(pyz)4\mathrm{CrCl}_{2}(\mathrm{pyz})_{4}, see Figure S1. For this to apply to monolayer CrCl2(pyz)2\mathrm{CrCl}_{2}(\mathrm{pyz})_{2}, the stoichiometry used in this section will be that of CrCl2(pyz)2\mathrm{CrCl}_{2}(\mathrm{pyz})_{2}. One unit cell of CrCl2(pyz)2\mathrm{CrCl}_{2}(\mathrm{pyz})_{2} consists of one Cr ion, two Cl ions, and two pyrazine rings. In the main text and when investigating monolayer CrCl2(pyz)2\mathrm{CrCl}_{2}(\mathrm{pyz})_{2} in the following this corresponds to only including pyz1,r^{1,r} and pyz2,t^{2,t} of the pyz sites in Figure S1. Each chlorine atom reduce the number of electrons on the chromium atom by one. The two pyrazine rings reduce the Cr atom of an extra electron, such that three electrons remain in an unfilled dd-shell on the Cr ion. If these spins occupy three distinct energy levels and align parallel to each other, to obey Hund’s rules, the spin on the Cr atoms is S=3/2S = 3/2.

As for Benzene, the 6 atoms in the pyrazine rings are sp2\mathrm{sp}^{2} hybridized so we are left with one unhybridized p-orbital per site perpendicular to the plane of the ring. This will be referred to as the z′z^{\prime} direction. Due to the presence of the N atoms in pyrazine, compared to Benzene, there are no degenerate states within the ring. The 6 energy levels for a pyrazine ring, where the H atoms are neglected and only the unhybridized p-orbitals are included, is shown in Figure S2. In Figure S2 the eigenstates for the n=1,…,4n = 1, \ldots, 4 levels are illustrated. Each eigenstate is visualized as circles plotted on top of the pyrazine ring with a radius proportional to the amplitude of the p-orbital on the given atom. The red and blue color corresponds to a positive and negative sign on the basis state, respectively. For the HOMO level (n=3n = 3) and the LUMO+1 level (n=5n = 5) the amplitude on the N atoms are equal to zero. These orbitals do not couple to the Cr atoms and will be discarded in the following calculations. The coupling between the carbon atoms in these states and the chromium atom will be negligibly small due to distance. The n=1n = 1 (n=6n = 6) level is so high (low) in energy that these can be neglected. Therefore, we only include the two levels n=2n = 2 and n=4n = 4, and each pyrazine ring is safely approximated as a single atom with pz′p_{z^{\prime}}-orbitals of two different amplitudes, α2\alpha_{2} and α4\alpha_{4}, where α2\alpha_{2} and α4\alpha_{4} are the amplitudes on the N atoms of the n=2n = 2 and n=4n = 4 levels, respectively. The amplitude on the C atoms are neglected. Note, that the former is an odd linear combination whereas the latter is an even linear combination of the two N atoms.

Energy levels of a pyrazine ring and eigenstates of intermediate energy levels

Figure S2. a) Energy levels of a pyrazine ring calculated from tight binding. We have excluded the 4 hydrogen atoms and only included the unhybridized p-orbitals on the 4 carbon atoms and 2 nitrogen atoms. b) The eigenstates of the 4 intermediate energy levels n=2n = 2 to n=4n = 4. The eigenstates are illustrated by circles with a radius proportional to the amplitude of the orbital occupying the given atom with a positive (red) or negative (blue) sign, respectively.

The pyrazine rings are canted relative to the Cl-Cr-Cl axis. In Ref. 8 the rotation angles are measured to 42.5∘42.5^\circ and 43.9∘43.9^\circ with respect to the crystallographic cc direction for the pyrazine ring 1 and 2, respectively. This will be approximated as 45∘45^\circ in both cases. This means that the pz′p_z^\prime-orbitals of the pyrazine “atoms” are instead linear combinations of both pxp_x, pyp_y and pzp_z orbitals with respect to the global coordinate system, see Figure S1.

The energy spectrum of model complex CrCl2_2(pyz)4_4 can be obtained using the tight binding model, which in second quantization can be written as

H^t=ξ∑i,σc^i,σ†c^i,σ−t∑⟨i,j⟩∑σ(c^i,σ†c^j,σ+h.c.).(6)\hat{H}_t = \xi\sum_{i,\sigma} \hat{c}^{\dagger}_{i,\sigma}\hat{c}_{i,\sigma} - t \sum_{\langle i,j\rangle}\sum_{\sigma}\left(\hat{c}^{\dagger}_{i,\sigma}\hat{c}_{j,\sigma} + h.c.\right). \tag*{(6)}

Here ξ\xi refer to the onsite energies, tt is the hopping integral and h.c. includes the Hermitian conjugate terms. c^i,σ†\hat{c}^{\dagger}_{i,\sigma} and c^i,σ\hat{c}_{i,\sigma} are the creation and annihilation operators, respectively, for an electron on site ii with spin σ\sigma. Only nearest neighbor interactions are included. For simplicity the spin-indices will be omitted in the following.

The full Hamiltonian for a single molecule can be written in terms of the Slater–Koster elements in the basis

Ψ={dxy,dyz,dzx,dx2−y2,dz2,pxt,pyt,pzt,pxb,pyb,pzb,pz1,r(α2),pz1,r(α4),pz2,t(α2),pz2,t(α4),pz1,l(α2),pz1,l(α4),pz2,b(α2),pz2,b(α4)}.(7)\Psi= \{d_{xy},d_{yz},d_{zx},d_{x^2-y^2},d_{z^2},p_x^t,p_y^t,p_z^t,p_x^b,p_y^b,p_z^b, p_z^{1,r}(\alpha_2),p_z^{1,r}(\alpha_4),p_z^{2,t}(\alpha_2),p_z^{2,t}(\alpha_4),p_z^{1,l}(\alpha_2),p_z^{1,l}(\alpha_4),p_z^{2,b}(\alpha_2),p_z^{2,b}(\alpha_4)\}. \tag*{(7)}

where the first five elements refer to the d-orbitals of the chromium atom followed by the three p-orbitals of both chlorine atoms (top and bottom), and lastly the pz′p_z^\prime-orbitals for the four pyrazine “atoms”. The full Hamiltonian can be written in matrix form as

H=[HdVdpVpdHp].(8)\mathcal{H} = \begin{bmatrix} \mathcal{H}_d & V_{dp}\\ V_{pd} & \mathcal{H}_p \end{bmatrix}. \tag*{(8)}

Hd\mathcal{H}_d is a diagonal 5×55 \times5 matrix with ξd\xi_d on the diagonal whereas Hp\mathcal{H}_p is a 14×1414 \times14 matrix with ξp\xi_p on the diagonal, where ξd\xi_d and ξp\xi_p are the onsite energies of the d- and p-orbitals, respectively. Thus, the onsite energies of all p-orbitals are assumed to be degenerate prior the interaction with the d-orbitals and similar for the d-orbitals. The two off-diagonal blocks of the Hamiltonian VdpV_{dp} and VpdV_{pd} couple the d-orbitals to the p-orbitals and vice versa.

Subspace of the d-orbitals

The Hamiltonian describing the system can be split up into a part that describes the d-orbitals, a part that describes the p-orbitals and then a term that describes the interaction between these orbitals,

H^=H^p+H^d+V^.(9)\hat{H} = \hat{H}_{p} + \hat{H}_{d} + \hat{V}. \tag*{(9)}

An eigenstate can be expressed as a sum of a p- and d-part of the state as

∣ψ⟩=∣ψp⟩+∣ψd⟩.(10)|\psi\rangle= |\psi_{p}\rangle+ |\psi_{d}\rangle. \tag*{(10)}

The projection operators P^d\hat{P}_{d} and P^p\hat{P}_{p} for the d-subspace and the p-subspace, respectively, can be applied to rewrite the Hamiltonian. These projection operators must sum to identity, P^p+P^d=1\hat{P}_{p} + \hat{P}_{d} = 1, and are defined to project the eigenstate of the full Hamiltonian into the subspace of either the d- or the p-orbitals, as

P^d∣ψ⟩=E∣ψd⟩ ,P^p∣ψ⟩=E∣ψp⟩.(11)\hat{P}_{d}|\psi\rangle= E|\psi_{d}\rangle\,, \qquad\hat{P}_{p}|\psi\rangle= E|\psi_{p}\rangle. \tag*{(11)}

Using these two operators yields

H^d∣ψd⟩+V^dp∣ψp⟩=E∣ψd⟩,(12)\hat{H}_{d}|\psi_{d}\rangle+ \hat{V}_{dp}|\psi_{p}\rangle= E|\psi_{d}\rangle, \tag*{(12)}
H^p∣ψp⟩+V^pd∣ψd⟩=E∣ψp⟩.(13)\hat{H}_{p}|\psi_{p}\rangle+ \hat{V}_{pd}|\psi_{d}\rangle= E|\psi_{p}\rangle. \tag*{(13)}

Here VdpV_{dp} and VpdV_{pd} refer to the 5×145 \times14 and 14×514 \times5 off-diagonal blocks of Eq. (8), respectively. Isolating ∣ψp⟩|\psi_{p}\rangle in Eq. (13) and inserting this in Eq. (12) leads to the expression

H^d∣ψd⟩+V^dp1E−H^pV^pd∣ψd⟩=E∣ψd⟩.(14)\hat{H}_{d}|\psi_{d}\rangle+ \hat{V}_{dp}\frac{1}{E-\hat{H}_{p}}\hat{V}_{pd}|\psi_{d}\rangle= E|\psi_{d}\rangle. \tag*{(14)}

Using that ∣ψd⟩|\psi_{d}\rangle and ∣ψp⟩|\psi_{p}\rangle are eigenstates of H^d\hat{H}_{d} and H^p\hat{H}_{p}, respectively, and assuming all p-levels to be degenerate prior to the interaction, this can be written as

ξd∣ψd⟩+ξ2E−ξp∣ψd⟩=E∣ψd⟩.(15)\xi_{d}|\psi_{d}\rangle+ \frac{\xi^{2}}{E-\xi_{p}}|\psi_{d}\rangle= E|\psi_{d}\rangle. \tag*{(15)}

Here, we have introduced the eigenenergy of VdpVpdV_{dp}V_{pd} as ξ2\xi^{2}. Setting ξd=0\xi_{d}=0, EE is isolated as

E±=ξp2(1±1+(2ξξp)2).(16)E_{\pm} = \frac{\xi_{p}}{2}\left(1 \pm\sqrt{1+\left(\frac{2\xi}{\xi_{p}}\right)^{2}}\right). \tag*{(16)}

Since we are seeking the bonding orbitals and ξp\xi_{p} is chosen to be positive, the E+E_{+} solution is discarded. Knowing the eigenenergies of the 5×55 \times5 Hamiltonian Vpd†VpdV_{pd}^{\dagger}V_{pd}, namely ξ2\xi^{2}, in the d-subspace allow us to solve Eq. (16). Investigating this matrix analytically we find that the matrix can be written in a block diagonal in the basis {dyz,dzx,dxy,dx2−y2,dz2}\{d_{yz},d_{zx},d_{xy},d_{x^{2}-y^{2}},d_{z^{2}}\},

VdpVpd=(M11M12000M12M2200000M11M34M3500M34M44M4500M35M45M55).(17)V_{dp}V_{pd} = \begin{pmatrix} M_{11} & M_{12} & 0 & 0 & 0 \\ M_{12} & M_{22} & 0 & 0 & 0 \\ 0 & 0 & M_{11} & M_{34} & M_{35} \\ 0 & 0 & M_{34} & M_{44} & M_{45} \\ 0 & 0 & M_{35} & M_{45} & M_{55} \end{pmatrix}. \tag*{(17)}

with

M11=12∣Vpdπ∣2(4+α),(18)M_{11}=\frac{1}{2}|V_{pd\pi}|^2(4+\alpha), \tag*{(18)}
M12=−32VpdσVpdπ†α,(19)M_{12}=-\frac{\sqrt{3}}{2}V_{pd\sigma}V_{pd\pi}^{\dagger}\alpha, \tag*{(19)}
M22=32∣Vpdσ∣2α,(20)M_{22}=\frac{3}{2}|V_{pd\sigma}|^2\alpha, \tag*{(20)}
M34=14Vpdπ†α(2Vpdπ−3Vpdσ),(21)M_{34}=\frac{1}{4}V_{pd\pi}^{\dagger}\alpha\left(\sqrt{2}V_{pd\pi}-\sqrt{3}V_{pd\sigma}\right), \tag*{(21)}
M35=−14Vpdπ†α(6Vpdπ+Vpdσ),(22)M_{35}=-\frac{1}{4}V_{pd\pi}^{\dagger}\alpha\left(\sqrt{6}V_{pd\pi}+V_{pd\sigma}\right), \tag*{(22)}
M45=18[43∣Vpdσ∣2−α(23Vpdπ+2Vpdσ)(Vpdπ†−32Vpd†)],(23)M_{45}=\frac{1}{8}\left[4\sqrt{3}|V_{pd\sigma}|^2-\alpha\left(2\sqrt{3}V_{pd\pi}+\sqrt{2}V_{pd\sigma}\right)\left(V_{pd\pi}^{\dagger}-\sqrt{\frac{3}{2}}V_{pd}^{\dagger}\right)\right], \tag*{(23)}
M44=18[123∣Vpdσ∣2+α(2Vpdπ−6Vpdσ)(Vpdπ†−32Vpd†)],(24)M_{44}=\frac{1}{8}\left[12\sqrt{3}|V_{pd\sigma}|^2+\alpha\left(2V_{pd\pi}-\sqrt{6}V_{pd\sigma}\right)\left(V_{pd\pi}^{\dagger}-\sqrt{\frac{3}{2}}V_{pd}^{\dagger}\right)\right], \tag*{(24)}
M55=18[4∣Vpdσ∣2+α(Vpdπ†(6Vpdπ+6Vpdσ)+Vpdσ†(Vpdσ+6Vpdπ))],(25)M_{55}=\frac{1}{8}\left[4|V_{pd\sigma}|^2+\alpha\left(V_{pd\pi}^{\dagger}\left(6V_{pd\pi}+\sqrt{6}V_{pd\sigma}\right)+V_{pd\sigma}^{\dagger}\left(V_{pd\sigma}+\sqrt{6}V_{pd\pi}\right)\right)\right], \tag*{(25)}

where α≡∣α2∣2+∣α4∣2\alpha\equiv|\alpha_2|^2+|\alpha_4|^2 has been defined. Here, VijbV_{ij}^{b} is the b=σb=\sigma or b=πb=\pi bond integral between two orbitals i=(s,p,d,f,…)i=(\mathrm{s},\mathrm{p},\mathrm{d},\mathrm{f},\ldots) and j=(s,p,d,f,…)j=(\mathrm{s},\mathrm{p},\mathrm{d},\mathrm{f},\ldots). Note that M11=M33M_{11}=M_{33} by symmetry. The two blocks in Eq. (17) are trivial to diagonalize. It is evident from Eq. (17) that the dyzd_{yz} orbital only couples to the dzxd_{zx} orbital, and vice versa, and further that dxyd_{xy}, dx2−y2d_{x^2-y^2}, and dz2d_{z^2} only couples to each other. To understand this, we elaborate on the symmetries of the molecule in the following.

The fact that the pyrazine rings are rotated by two different angles decreases the symmetry of the molecule C4h→C2hC_{4h}\to C_{2h}. As always, the group describing the spatial symmetries of CrCl2_2(pyz)4_4 includes identity. Only three other operations leave the molecule invariant, namely inversion (i), a π\pi rotation of about the zz-axis (C2(z)C_2(z)) and a reflection in the xyxy plane (σh\sigma_h), see Figure S1. These four symmetry-operations form the point group C2hC_{2h}. This group is Abelian and thus both the order of the group and the number of irreducible representations are equal to 4. The character table can now be constructed, see Table I.

C2hC_{2h}EEC2(z)C_2(z)iiσh\sigma_hQuadratic Functions
AgA_g1111x2x^2, y2y^2, z2z^2, xyxy
BgB_g1-11-1yzyz, zxzx
AuA_u11-1-1
BuB_u1-1-11

Table I. Character table of the C2hC_{2h} group.

From the character table it is observed that yzyz and zxzx belong to the same representation and similarly for the last three dd-orbitals. This explains the block diagonal form of Eq. (17). Further, ⟨xy∣H∣xy⟩=⟨yz∣H∣yz⟩\langle xy|H|xy\rangle=\langle yz|H|yz\rangle which is due to the chosen coordinate system conforming with the symmetry of the model complex such that these matrix elements are identical.

Tight Binding Model of Monolayer CrCl2_2(pyz)2_2

Monolayer CrCl2_2(pyz)2_2, see Figure 1 b), is now modeled. Using the Rietveld refinement technique [33] the three lattice constants are in Ref. [8] found to be a=6.90351(4)a=6.90351(4) Å, b=6.97713(4)b=6.97713(4) Å and c=10.82548(6)c=10.82548(6) Å. The close affinity of aa and bb motivates the lattice to be approximated as a square lattice in the Cr-pyz plane. The relevant symmetry-points are thus approximated to be

Γ=(0,0,0),M=(0,0,12),X=(−12,0,12).(26)\Gamma=(0,0,0),\qquad M=\left(0,0,\frac{1}{2}\right),\qquad X=\left(-\frac{1}{2},0,\frac{1}{2}\right). \tag*{(26)}

These symmetry points are indicated on the Brillouin zone illustrated in Figure 1 c).

A unit cell includes one chromium atom, two chlorine atoms and two pyrazine rings, which will be labeled 1 and 2. Assuming that the two pyrazine rings have the same onsite energy, and similarly for the chlorine atoms, a tight binding Hamiltonian for this system in real-space, including only nearest neighbor hopping between p- and d-orbitals, can be written as

H^=ξCr∑ri∑μ=15d^i,μ†d^i,μ+ξCl∑ri∑ν=13[(p^i,νt)†p^i,νt+(p^i,νb)†p^i,νb]+ξpyz∑ri[(p^i,z′1)†p^i,z′1+(p^i,z′2)†p^i,z′2]+12∑ri∑rj=ri±a1∑μ=15[tμ,α1,rd^i,μ†p^j+1+tμ,α1,ld^i,μ†p^j−1+h.c.]+12∑ri∑rj=ri±a2∑μ=15[tμ,α2,td^i,μ†p^j+2+tμ,α2,bd^i,μ†p^j−2+h.c.]+∑ri∑rj=ri±a3∑μ=15[tμ,νtd^i,μ†p^j,νt+tμ,νbd^i,μ†p^j,νb+h.c.].(27)\begin{aligned} \hat{H}={}&\xi^{\mathrm{Cr}}\sum_{\mathbf{r}_i}\sum_{\mu=1}^{5}\hat{d}_{i,\mu}^{\dagger}\hat{d}_{i,\mu}+\xi^{\mathrm{Cl}}\sum_{\mathbf{r}_i}\sum_{\nu=1}^{3}\left[(\hat{p}^{t}_{i,\nu})^{\dagger}\hat{p}^{t}_{i,\nu}+(\hat{p}^{b}_{i,\nu})^{\dagger}\hat{p}^{b}_{i,\nu}\right]\\ &+\xi^{\mathrm{pyz}}\sum_{\mathbf{r}_i}\left[(\hat{p}^{1}_{i,z^{\prime}})^{\dagger}\hat{p}^{1}_{i,z^{\prime}}+(\hat{p}^{2}_{i,z^{\prime}})^{\dagger}\hat{p}^{2}_{i,z^{\prime}}\right]\\ &+\frac{1}{2}\sum_{\mathbf{r}_i}\sum_{\mathbf{r}_j=\mathbf{r}_i\pm\mathbf{a}_1}\sum_{\mu=1}^{5}\left[t_{\mu,\alpha}^{1,r}\hat{d}_{i,\mu}^{\dagger}\hat{p}^{1}_{j+}+t_{\mu,\alpha}^{1,l}\hat{d}_{i,\mu}^{\dagger}\hat{p}^{1}_{j-}+\mathrm{h.c.}\right]\\ &+\frac{1}{2}\sum_{\mathbf{r}_i}\sum_{\mathbf{r}_j=\mathbf{r}_i\pm\mathbf{a}_2}\sum_{\mu=1}^{5}\left[t_{\mu,\alpha}^{2,t}\hat{d}_{i,\mu}^{\dagger}\hat{p}^{2}_{j+}+t_{\mu,\alpha}^{2,b}\hat{d}_{i,\mu}^{\dagger}\hat{p}^{2}_{j-}+\mathrm{h.c.}\right]\\ &+\sum_{\mathbf{r}_i}\sum_{\mathbf{r}_j=\mathbf{r}_i\pm\mathbf{a}_3}\sum_{\mu=1}^{5}\left[t_{\mu,\nu}^{t}\hat{d}_{i,\mu}^{\dagger}\hat{p}^{t}_{j,\nu}+t_{\mu,\nu}^{b}\hat{d}_{i,\mu}^{\dagger}\hat{p}^{b}_{j,\nu}+\mathrm{h.c.}\right]. \tag*{(27)} \end{aligned}

Here d^i,μ\hat{d}_{i,\mu} (d^i,μ†\hat{d}_{i,\mu}^{\dagger}) creates (annihilates) an electron on the chromium site at ri\mathbf{r}_i in unit cell ii and orbital μ\mu. Similarly, the operator (p^j,νt,b)†(\hat{p}_{j,\nu}^{t,b})^{\dagger} and p^j,νt,b\hat{p}_{j,\nu}^{t,b} creates and annihilates an electron in the orbital ν\nu at the chlorine atom at site ri±a3\mathbf{r}_i\pm\mathbf{a}_3, respectively. Lastly, the operators (p^j,z′1(2))†(\hat{p}_{j,z^{\prime}}^{1(2)})^{\dagger} and p^j,z′1(2)\hat{p}_{j,z^{\prime}}^{1(2)} creates and annihilates an electron in a pz′p_{z^{\prime}} orbital on the pyrazine site 1 (2) at ri+a1\mathbf{r}_i+\mathbf{a}_1 and ri+a2\mathbf{r}_i+\mathbf{a}_2, respectively.

The three first terms of the Hamiltonian states the onsite energies of the d- and p-orbitals of the five atoms in the unit cell. The lower three lines of Eq. (27) describe electrons hopping from a p-orbital to any nearest neighboring d-orbital. The factor 1/21/2 compensate for double counting. For simplicity the spin indices have been omitted. The following on-site energies and hopping integrals are defined as

ξegCr=⟨ri;dμ∣H^t∣ri;dμ⟩,(28)\xi_{\mathrm{eg}}^{\mathrm{Cr}}=\langle\mathbf{r}_i;d_\mu|\hat{H}_t|\mathbf{r}_i;d_\mu\rangle, \tag*{(28)}
ξCl=⟨ri±a3;pν∣H^t∣ri±a3;pν⟩,(29)\xi^{\mathrm{Cl}}=\langle\mathbf{r}_i\pm\mathbf{a}_3;p_\nu|\hat{H}_t|\mathbf{r}_i\pm\mathbf{a}_3;p_\nu\rangle, \tag*{(29)}
ξpyz=⟨ri+a1;pz′(α′)∣H^t∣ri+a1;pz′(α′)⟩,(30)\xi^{\mathrm{pyz}}=\langle\mathbf{r}_i+\mathbf{a}_1;p_{z^{\prime}}(\alpha^{\prime})|\hat{H}_t|\mathbf{r}_i+\mathbf{a}_1;p_{z^{\prime}}(\alpha^{\prime})\rangle, \tag*{(30)}
tμ,νt=⟨ri;dμ∣H^t∣ri+a3;pν⟩,(31)t_{\mu,\nu}^{t}=\langle\mathbf{r}_i;d_\mu|\hat{H}_t|\mathbf{r}_i+\mathbf{a}_3;p_\nu\rangle, \tag*{(31)}
tμ,νb=⟨ri;dμ∣H^t∣ri−a3;pν⟩,(32)t_{\mu,\nu}^{b}=\langle\mathbf{r}_i;d_\mu|\hat{H}_t|\mathbf{r}_i-\mathbf{a}_3;p_\nu\rangle, \tag*{(32)}
tμ,z′1,r=⟨ri;dμ∣H^t∣ri+a1;pz′(α′)⟩,(33)t_{\mu,z^{\prime}}^{1,r}=\langle\mathbf{r}_i;d_\mu|\hat{H}_t|\mathbf{r}_i+\mathbf{a}_1;p_{z^{\prime}}(\alpha^{\prime})\rangle, \tag*{(33)}
tμ,z′1,l=⟨ri;dμ∣H^t∣ri−a1;pz′(α′)⟩,(34)t_{\mu,z^{\prime}}^{1,l}=\langle\mathbf{r}_i;d_\mu|\hat{H}_t|\mathbf{r}_i-\mathbf{a}_1;p_{z^{\prime}}(\alpha^{\prime})\rangle, \tag*{(34)}
tμ,z′2,t=⟨ri;dμ∣H^t∣ri+a2;pz′(α′)⟩,(35)t_{\mu,z^{\prime}}^{2,t}=\langle\mathbf{r}_i;d_\mu|\hat{H}_t|\mathbf{r}_i+\mathbf{a}_2;p_{z^{\prime}}(\alpha^{\prime})\rangle, \tag*{(35)}
tμ,z′2,b=⟨ri;dμ∣H^t∣ri−a2;pz′(α′)⟩.(36)t_{\mu,z^{\prime}}^{2,b}=\langle\mathbf{r}_i;d_\mu|\hat{H}_t|\mathbf{r}_i-\mathbf{a}_2;p_{z^{\prime}}(\alpha^{\prime})\rangle. \tag*{(36)}

H^t\hat{H}_t includes hopping between the two relevant orbitals. Here, we have used that α′=α2,α4\alpha^{\prime}=\alpha_2,\alpha_4. The numerical values for these can be obtained using the Slater–Koster decomposition [30, 31]. ttt^t and tbt^b are 3×53\times5 matrices that includes hopping between the 3 p-orbitals on the Cl atoms and the 5 d-orbitals on the Cr atoms. The last four lines in Eq. (36) includes the 2×52\times5 hopping matrices between pyrazine orbitals and the chromium atoms that take the form

tμ,z′1,r=[α2Ez′,xyα2Ez′,yzα2Ez′,zxα2Ez′,x2−y2α2Ez′,z2α4Ez′,xyα4Ez′,yzα4Ez′,zxα4Ez′,x2−y2α4Ez′,z2]μ,z′,(37)t_{\mu,z^{\prime}}^{1,r}= \begin{bmatrix} \alpha_2E_{z^{\prime},xy} & \alpha_2E_{z^{\prime},yz} & \alpha_2E_{z^{\prime},zx} & \alpha_2E_{z^{\prime},x^2-y^2} & \alpha_2E_{z^{\prime},z^2}\\ \alpha_4E_{z^{\prime},xy} & \alpha_4E_{z^{\prime},yz} & \alpha_4E_{z^{\prime},zx} & \alpha_4E_{z^{\prime},x^2-y^2} & \alpha_4E_{z^{\prime},z^2} \end{bmatrix}_{\mu,z^{\prime}}, \tag*{(37)}
tμ,z′1,l=[−α2Ez′,xy−α2Ez′,yz−α2Ez′,zx−α2Ez′,x2−y2−α2Ez′,z2α4Ez′,xyα4Ez′,yzα4Ez′,zxα4Ez′,x2−y2α4Ez′,z2]μ,z′.(38)t_{\mu,z^{\prime}}^{1,l}= \begin{bmatrix} -\alpha_2E_{z^{\prime},xy} & -\alpha_2E_{z^{\prime},yz} & -\alpha_2E_{z^{\prime},zx} & -\alpha_2E_{z^{\prime},x^2-y^2} & -\alpha_2E_{z^{\prime},z^2}\\ \alpha_4E_{z^{\prime},xy} & \alpha_4E_{z^{\prime},yz} & \alpha_4E_{z^{\prime},zx} & \alpha_4E_{z^{\prime},x^2-y^2} & \alpha_4E_{z^{\prime},z^2} \end{bmatrix}_{\mu,z^{\prime}}. \tag*{(38)}

The above Slater–Koster elements in the global coordinate system for d-orbital μ\mu on Cr and the pz′p_{z^{\prime}}-orbitals on pyz is given by

Ez′,μ=[14(2−2)Ex,μ+2Ey,μ+(2+2)Ez,μ].(39)E_{z^{\prime},\mu}=\left[\frac{1}{4}\left(\sqrt{2}-2\right)E_{x,\mu}+2E_{y,\mu}+\left(2+\sqrt{2}\right)E_{z,\mu}\right]. \tag*{(39)}

The bonding matrices t2,tt^{2,t} and t2,bt^{2,b} take on a similar form. The Hamiltonian can be formulated in reciprocal space as

H^=∑k[ξCr∑μ=15d^k,μ†d^k,μ+ξCl∑ν=13((p^k,νt)†p^k,νt+(p^k,νb)†p^k,νb)+ξpyz((p^k,z′1)†p^k,z′1+(p^k,z′2)†p^k,z′2)+∑μ=15[d^k,μ†p^k,z′1(tμ,z′1,teik⋅a1+tμ,z′1,be−ik⋅a1)+h.c.]+∑μ=15[d^k,μ†p^k,z′2(tμ,z′2,teik⋅a2+tμ,z′2,be−ik⋅a2)+h.c.]+∑μ=15∑ν=13[tμ,νtd^k,μ†p^k,νteik⋅a3+tμ,νbd^k,μ†p^k,νbe−ik⋅a3+h.c.]}.(40)\begin{aligned} \hat{\mathcal{H}} ={}& \sum_{\mathbf{k}}\left[\xi^{\mathrm{Cr}}\sum_{\mu=1}^{5}\hat{d}^{\dagger}_{\mathbf{k},\mu}\hat{d}_{\mathbf{k},\mu}+\xi^{\mathrm{Cl}}\sum_{\nu=1}^{3}\left((\hat{p}^{t}_{\mathbf{k},\nu})^{\dagger}\hat{p}^{t}_{\mathbf{k},\nu}+(\hat{p}^{b}_{\mathbf{k},\nu})^{\dagger}\hat{p}^{b}_{\mathbf{k},\nu}\right)\right.\\ &+\xi^{\mathrm{pyz}}\left((\hat{p}^{1}_{\mathbf{k},z'})^{\dagger}\hat{p}^{1}_{\mathbf{k},z'}+(\hat{p}^{2}_{\mathbf{k},z'})^{\dagger}\hat{p}^{2}_{\mathbf{k},z'}\right)\\ &+\sum_{\mu=1}^{5}\left[\hat{d}^{\dagger}_{\mathbf{k},\mu}\hat{p}^{1}_{\mathbf{k},z'}\left(t^{1,t}_{\mu,z'}e^{i\mathbf{k}\cdot\mathbf{a}_{1}}+t^{1,b}_{\mu,z'}e^{-i\mathbf{k}\cdot\mathbf{a}_{1}}\right)+\mathrm{h.c.}\right]\\ &+\sum_{\mu=1}^{5}\left[\hat{d}^{\dagger}_{\mathbf{k},\mu}\hat{p}^{2}_{\mathbf{k},z'}\left(t^{2,t}_{\mu,z'}e^{i\mathbf{k}\cdot\mathbf{a}_{2}}+t^{2,b}_{\mu,z'}e^{-i\mathbf{k}\cdot\mathbf{a}_{2}}\right)+\mathrm{h.c.}\right]\\ &\left.+\sum_{\mu=1}^{5}\sum_{\nu=1}^{3}\left[t^{t}_{\mu,\nu}\hat{d}^{\dagger}_{\mathbf{k},\mu}\hat{p}^{t}_{\mathbf{k},\nu}e^{i\mathbf{k}\cdot\mathbf{a}_{3}}+t^{b}_{\mu,\nu}\hat{d}^{\dagger}_{\mathbf{k},\mu}\hat{p}^{b}_{\mathbf{k},\nu}e^{-i\mathbf{k}\cdot\mathbf{a}_{3}}+\mathrm{h.c.}\right]\right\}. \tag*{(40)} \end{aligned}

The eigenenergies of the 15×1515 \times15 matrix H^\hat{\mathcal{H}} are obtained numerically for the path M →\to Γ →\to X in reciprocal space, yielding the band structure in Figure 1 c). The following values, obtained from a manual fit to DFT results [8], were used to calculate the energy bands, ξt2gCr=−5.7 eV\xi^{\mathrm{Cr}}_{t_{2g}}=-5.7\ \mathrm{eV}, ξegCr=1.56 eV\xi^{\mathrm{Cr}}_{e_g}=1.56\ \mathrm{eV}, ξCl=−6.24 eV\xi^{\mathrm{Cl}}=-6.24\ \mathrm{eV}, ξα2pyz=−1.02 eV\xi^{\mathrm{pyz}}_{\alpha_{2}}=-1.02\ \mathrm{eV}, ξα4pyz=−0.08 eV\xi^{\mathrm{pyz}}_{\alpha_{4}}=-0.08\ \mathrm{eV}, Vpdσ=1.7 eVV_{pd\sigma}=1.7\ \mathrm{eV}, Vpdπ=−4.75 eVV_{pd\pi}=-4.75\ \mathrm{eV}, α2=0.67\alpha_{2}=0.67, α4=−0.54\alpha_{4}=-0.54. Knowing that 3 of the d-levels are half-filled and the last two are empty, we employ two different on-site energies for the chromium orbitals; one value for the three t2gt_{2g} orbitals and one for the two ege_g orbitals. Further, we know from Figure S2 that the two “orbitals” on the pyrazine sites have two distinct values. All Cl orbitals are assumed to have the same on-site energy.

The empty d-bands are located above 1.5 eV (the ege_g orbitals) whereas the three filled d-bands (t2gt_{2g}) are located below the p-levels (not visible in the plot). This is evident when studying the eigenstates of each band but even more clearly seen from first principle calculations of the PDOS [8]. These d-bands are very flat and thus, the d-electrons are localized on the Cr sites. This could indicate a strong π-d hybridization. In contrary the p-bands mainly occupied by pyz electrons, located above and below the Fermi energy, are dispersive indicating that the pyz electrons are delocalized.

The Minimal Model and its ground states

Here, we present an effective model of monolayer CrCl2_2(pyz)2_2, see Figure 2. Note that this is not identical to two unit cells of a single layer of CrCl2_2(pyz)2_2 since it contain fewer pyrazine sites and they are shifted with respect to the two chromium atoms. However, it essentially describes the competition between the kinetic energy and the exchange energy. In this model the two chromium atoms are viewed as localized S=3/2S=3/2 spins. There are two electrons that can delocalize over the four pyrazine sites. The chlorine sites are neglected in this model, except for the fact that they have reduced the number of electrons on the chromium sites.

We include the exchange coupling between the chromium spins and the spins of the two pyrazine electrons. Furthermore, the pyrazine electrons can hop between the pyrazine sites (indicated with dashed lines in Figure 2). Lastly, we include onsite Hubbard repulsion, U>0U>0, for electrons located on the same site. The full Hamiltonian can thus be written as

H^=−2JS^A⋅(S^1+S^2+S^3)−2JS^B⋅(S^2+S^3+S^4)−t∑σ=↑,↓(c^1σ†c^2σ+c^2σ†c^3σ+c^1σ†c^3σ+c^2σ†c^4σ+c^3σ†c^4σ+h.c.)+U∑i=14n^i↑n^i↓.(41)\hat{\mathcal{H}}=-2J\hat{\mathbf{S}}_{A}\cdot\left(\hat{\mathbf{S}}_{1}+\hat{\mathbf{S}}_{2}+\hat{\mathbf{S}}_{3}\right)-2J\hat{\mathbf{S}}_{B}\cdot\left(\hat{\mathbf{S}}_{2}+\hat{\mathbf{S}}_{3}+\hat{\mathbf{S}}_{4}\right)-t\sum_{\sigma=\uparrow,\downarrow}\left(\hat{c}^{\dagger}_{1\sigma}\hat{c}_{2\sigma}+\hat{c}^{\dagger}_{2\sigma}\hat{c}_{3\sigma}+\hat{c}^{\dagger}_{1\sigma}\hat{c}_{3\sigma}+\hat{c}^{\dagger}_{2\sigma}\hat{c}_{4\sigma}+\hat{c}^{\dagger}_{3\sigma}\hat{c}_{4\sigma}+\mathrm{h.c.}\right)+U\sum_{i=1}^{4}\hat{n}_{i\uparrow}\hat{n}_{i\downarrow}. \tag*{(41)}

where S^A\hat{\mathbf{S}}_{A} and S^B\hat{\mathbf{S}}_{B} represent the spin of the two chromium atoms and S^i\hat{\mathbf{S}}_{i} with i=1,2,3,4i=1,2,3,4 represent the spin of the four pyrazine electrons. The operators c^iσ†\hat{c}^{\dagger}_{i\sigma} and c^iσ\hat{c}_{i\sigma} create and annihilate an electron on site ii with spin σ\sigma, respectively. JJ is the exchange integral, tt the hopping constant and UU is the strength of the Hubbard repulsion.

Basis

The spins of each chromium atom can take four SzS_z values, i.e., −3/2-3/2, −1/2-1/2, 1/21/2 and 3/23/2. The basis for the two electrons on the pyrazine sites will be written in terms of the singlet and triplet states. For this purpose I will introduce the following notation. A singlet state will be represented by a line, e.g.,

∣− 00⟩=12(∣↑↓ 00⟩−∣↓↑ 00⟩),(42)\lvert-\,00\rangle= \frac{1}{\sqrt{2}}\left(\lvert\uparrow\downarrow\,00\rangle-\lvert\downarrow\uparrow\,00\rangle\right), \tag*{(42)}
∣/0 0⟩=12(∣↑ 0↓ 0⟩−∣↓ 0↑ 0⟩),(43)\lvert/0\,0\rangle= \frac{1}{\sqrt{2}}\left(\lvert\uparrow\,0\downarrow\,0\rangle-\lvert\downarrow\,0\uparrow\,0\rangle\right), \tag*{(43)}
∣− 000⟩=12(∣↑↓ 000⟩−∣↓↑ 000⟩).(44)\lvert-\,000\rangle= \frac{1}{\sqrt{2}}\left(\lvert\uparrow\downarrow\,000\rangle-\lvert\downarrow\uparrow\,000\rangle\right). \tag*{(44)}

The first state is a singlet state between an electron on site 1 and an electron on site 2, the second between two electrons on site 1 and 3, and the last is a singlet state of two electrons both occupying site 1. Similarly, we can represent a triplet state with Sz=0S_z=0 as an arrow,

∣→ 00⟩=12(∣↑↓ 00⟩+∣↓↑ 00⟩),(45)\lvert\to\,00\rangle= \frac{1}{\sqrt{2}}\left(\lvert\uparrow\downarrow\,00\rangle+\lvert\downarrow\uparrow\,00\rangle\right), \tag*{(45)}
∣/0 0⟩=12(∣↑ 0↓ 0⟩+∣↓ 0↑ 0⟩).(46)\lvert/0\,0\rangle= \frac{1}{\sqrt{2}}\left(\lvert\uparrow\,0\downarrow\,0\rangle+\lvert\downarrow\,0\uparrow\,0\rangle\right). \tag*{(46)}

The reason for representing the triplet state as an arrow can be seen by anticommuting operators, c^1↑†c^2↓†+c^1↓†c^2↑†=−(c^2↓†c^1↑†+c^2↑†c^1↓†)\hat{c}^{\dagger}_{1\uparrow}\hat{c}^{\dagger}_{2\downarrow}+\hat{c}^{\dagger}_{1\downarrow}\hat{c}^{\dagger}_{2\uparrow}=-\left(\hat{c}^{\dagger}_{2\downarrow}\hat{c}^{\dagger}_{1\uparrow}+\hat{c}^{\dagger}_{2\uparrow}\hat{c}^{\dagger}_{1\downarrow}\right), such that the overall sign is meaningfully giving the direction of the arrow. For singlets the two states would be identical. All triplet states are defined such that the operator to the left creates an electron on a site labeled by a lower numerical value than the operator to the right (also applicable to Sz=1S_z=1 and Sz=−1S_z=-1 triplets). The basis for the two electrons are thus

∣ψ⟩={∣− 00⟩,∣0−0⟩,∣00−⟩,∣/0 0⟩,∣0/0⟩,∣00/⟩,∣− 000⟩,∣0−00⟩,∣00−0⟩,∣000−⟩,∣↓↓ 00⟩,∣0↓↓ 0⟩,∣00↓↓⟩,∣↓ 0↓ 0⟩,∣0↓ 0↓⟩,∣↓ 00↓⟩,∣→ 00⟩,∣0→ 0⟩,∣00→⟩,∣/0 0⟩,∣0/0⟩,∣00/⟩,∣↑↑ 00⟩,∣0↑↑ 0⟩,∣00↑↑⟩,∣↑ 0↑ 0⟩,∣0↑ 0↑⟩,∣↑ 00↑⟩}.(47)\lvert\psi\rangle= \left\{ \begin{aligned} &\lvert-\,00\rangle,\lvert0-0\rangle,\lvert00-\rangle,\lvert/0\,0\rangle,\lvert0/0\rangle,\lvert00/\rangle,\\ &\lvert-\,000\rangle,\lvert0-00\rangle,\lvert00-0\rangle,\lvert000-\rangle,\\ &\lvert\downarrow\downarrow\,00\rangle,\lvert0\downarrow\downarrow\,0\rangle,\lvert00\downarrow\downarrow\rangle,\lvert\downarrow\,0\downarrow\,0\rangle,\lvert0\downarrow\,0\downarrow\rangle,\lvert\downarrow\,00\downarrow\rangle,\\ &\lvert\to\,00\rangle,\lvert0\to\,0\rangle,\lvert00\to\rangle,\lvert/0\,0\rangle,\lvert0/0\rangle,\lvert00/\rangle,\\ &\lvert\uparrow\uparrow\,00\rangle,\lvert0\uparrow\uparrow\,0\rangle,\lvert00\uparrow\uparrow\rangle,\lvert\uparrow\,0\uparrow\,0\rangle,\lvert0\uparrow\,0\uparrow\rangle,\lvert\uparrow\,00\uparrow\rangle \end{aligned} \right\}. \tag*{(47)}

The full state is then given by ∣msA⟩⊗∣msB⟩⊗∣ψ⟩\lvert ms_A\rangle\otimes\lvert ms_B\rangle\otimes\lvert\psi\rangle and the Hilbert space is 448-dimensional.

Hamiltonian representation

The Hamiltonian is diagonalized numerically. First, we rewrite the Hamiltonian, using S^1⋅S^2=S^1zS^2z+12(S^1+S^2−+S^1−S^2+)\hat{S}_1\cdot\hat{S}_2=\hat{S}_1^z\hat{S}_2^z+\frac{1}{2}\left(\hat{S}_1^+\hat{S}_2^-+\hat{S}_1^-\hat{S}_2^+\right), such that

H^=−2JS^Az(S^1z+S^2z+S^3z)−JS^A+(S^1−+S^2−+S^3−)−JS^A−(S^1++S^2++S^3+)−2JS^Bz(S^2z+S^3z+S^4z)−JS^B+(S^2−+S^3−+S^4−)−JS^B−(S^2++S^3++S^4+)−t∑σ=↑,↓(c^1σ†c^2σ+c^2σ†c^3σ+c^1σ†c^3σ+c^2σ†c^4σ+c^3σ†c^4σ+h.c.)+U∑i=14∑σ,σ′=↑,↓c^iσ†c^iσ′†c^iσc^iσ′.(48)\begin{aligned} \hat{H}={}&-2J\hat{S}_A^z\left(\hat{S}_1^z+\hat{S}_2^z+\hat{S}_3^z\right)-J\hat{S}_A^+\left(\hat{S}_1^-+\hat{S}_2^-+\hat{S}_3^-\right)-J\hat{S}_A^-\left(\hat{S}_1^++\hat{S}_2^++\hat{S}_3^+\right)\\ &-2J\hat{S}_B^z\left(\hat{S}_2^z+\hat{S}_3^z+\hat{S}_4^z\right)-J\hat{S}_B^+\left(\hat{S}_2^-+\hat{S}_3^-+\hat{S}_4^-\right)-J\hat{S}_B^-\left(\hat{S}_2^++\hat{S}_3^++\hat{S}_4^+\right)\\ &-t\sum_{\sigma=\uparrow,\downarrow}\left(\hat{c}_{1\sigma}^{\dagger}\hat{c}_{2\sigma}+\hat{c}_{2\sigma}^{\dagger}\hat{c}_{3\sigma}+\hat{c}_{1\sigma}^{\dagger}\hat{c}_{3\sigma}+\hat{c}_{2\sigma}^{\dagger}\hat{c}_{4\sigma}+\hat{c}_{3\sigma}^{\dagger}\hat{c}_{4\sigma}+\mathrm{h.c.}\right)\\ &+U\sum_{i=1}^{4}\sum_{\sigma,\sigma'=\uparrow,\downarrow}\hat{c}_{i\sigma}^{\dagger}\hat{c}_{i\sigma'}^{\dagger}\hat{c}_{i\sigma}\hat{c}_{i\sigma'}. \tag*{(48)} \end{aligned}

In the following we will explore how the full Hamiltonian can be written in the basis described above by investigating some specific examples. The less enthusiastic reader can potentially skip this part and jump to section VII C.

We focus on the first three exchange terms including S^Az\hat{S}_A^z. The chromium part of this is trivial in the basis {∣−3/2⟩,∣−1/2⟩,∣1/2⟩,∣3/2⟩}\{\lvert-3/2\rangle,\lvert-1/2\rangle,\lvert1/2\rangle,\lvert3/2\rangle\}:

MCrz=[−320000−12000012000032].(49)M_{\mathrm{Cr}}^{z}= \begin{bmatrix} -\frac{3}{2} & 0 & 0 & 0\\ 0 & -\frac{1}{2} & 0 & 0\\ 0 & 0 & \frac{1}{2} & 0\\ 0 & 0 & 0 & \frac{3}{2} \end{bmatrix}. \tag*{(49)}

Now, let us see how this affects the basis states for the electrons on the pyrazine sites. Applying this operator on basis states where neither of the two electrons are located on site 4 results in the expected Sz=0S_z=0 for a singlet state, e.g.,

(S^1z+S^2z+S^3z)∣−00⟩=12(12∣↑↓00⟩−12∣↓↑00⟩−12∣↑↓00⟩+12∣↓↑00⟩)=0.(50)\left(\hat{S}_1^z+\hat{S}_2^z+\hat{S}_3^z\right)|-00\rangle=\frac{1}{\sqrt{2}}\left(\frac{1}{2}|\uparrow\downarrow00\rangle-\frac{1}{2}|\downarrow\uparrow00\rangle-\frac{1}{2}|\uparrow\downarrow00\rangle+\frac{1}{2}|\downarrow\uparrow00\rangle\right)=0. \tag*{(50)}

The same applies to triplet states with Sz=0S_z=0 with no electrons occupying site 4 and for singlet states where both electrons occupy the same site. On the other hand, when one of the electrons is occupying site 4, singlet and triplet states couple. For instance:

(S^1z+S^2z+S^3z)∣00−⟩=12[12∣00↑↓⟩−(−12∣00↓↑⟩)]=12∣00→⟩.(51)\left(\hat{S}_1^z+\hat{S}_2^z+\hat{S}_3^z\right)|00-\rangle=\frac{1}{\sqrt{2}}\left[\frac{1}{2}|00\uparrow\downarrow\rangle-\left(-\frac{1}{2}|00\downarrow\uparrow\rangle\right)\right]=\frac{1}{2}|00\rightarrow\rangle. \tag*{(51)}

The triplet states with Sz=1S_z=1 or Sz=−1S_z=-1 are eigenstates of this operator; the eigenvalue depends on whether one of the electrons is occupying site 4 or not,

(S^1z+S^2z+S^3z)∣0↓↓0⟩=−∣0↓↓0⟩,(52)\left(\hat{S}_1^z+\hat{S}_2^z+\hat{S}_3^z\right)|0\downarrow\downarrow0\rangle=-|0\downarrow\downarrow0\rangle, \tag*{(52)}
(S^1z+S^2z+S^3z)∣00↓↓⟩=−12∣00↓↓⟩,(53)\left(\hat{S}_1^z+\hat{S}_2^z+\hat{S}_3^z\right)|00\downarrow\downarrow\rangle=-\frac{1}{2}|00\downarrow\downarrow\rangle, \tag*{(53)}
(S^1z+S^2z+S^3z)∣0↑↑0⟩=∣0↑↑0⟩,(54)\left(\hat{S}_1^z+\hat{S}_2^z+\hat{S}_3^z\right)|0\uparrow\uparrow0\rangle=|0\uparrow\uparrow0\rangle, \tag*{(54)}
(S^1z+S^2z+S^3z)∣00↑↑⟩=12∣00↑↑⟩.(55)\left(\hat{S}_1^z+\hat{S}_2^z+\hat{S}_3^z\right)|00\uparrow\uparrow\rangle=\frac{1}{2}|00\uparrow\uparrow\rangle. \tag*{(55)}

This procedure is similar for the SBzS_B^z part of the Hamiltonian but in this case singlet and triplet states are coupled when one of the electrons is occupying site 1.

Now, let us look at the other part of the exchange interaction, i.e., the terms −JS^A−(S^1++S^2++S^3+)−JS^A+(S^1−+S^2−+S^3−)-J\hat{S}_A^-\left(\hat{S}_1^++\hat{S}_2^++\hat{S}_3^+\right)-J\hat{S}_A^+\left(\hat{S}_1^-+\hat{S}_2^-+\hat{S}_3^-\right). The raising and lowering operators for the chromium spins can be written as

MCr+=[0000300002000030]andMCr−=[0300002000030000].(56)M_{\mathrm{Cr}}^+= \begin{bmatrix} 0 & 0 & 0 & 0\\ \sqrt{3} & 0 & 0 & 0\\ 0 & 2 & 0 & 0\\ 0 & 0 & \sqrt{3} & 0 \end{bmatrix} \quad\text{and}\quad M_{\mathrm{Cr}}^-= \begin{bmatrix} 0 & \sqrt{3} & 0 & 0\\ 0 & 0 & 2 & 0\\ 0 & 0 & 0 & \sqrt{3}\\ 0 & 0 & 0 & 0 \end{bmatrix}. \tag*{(56)}

For the electrons on the pyrazine site it is not as simple. As before, the outcome depends on whether one of the electrons is occupying site 4 or not. Focusing on the electron part including raising operators for the pyrazine electrons, we look at a few cases to see how the basis states are affected by this operator,

(S^1++S^2++S^3+)∣−00⟩=0,(57)\left(\hat{S}_1^++\hat{S}_2^++\hat{S}_3^+\right)|-00\rangle=0, \tag*{(57)}
(S^1++S^2++S^3+)∣00−⟩=−12∣00↑↑⟩,(58)\left(\hat{S}_1^++\hat{S}_2^++\hat{S}_3^+\right)|00-\rangle=-\frac{1}{\sqrt{2}}|00\uparrow\uparrow\rangle, \tag*{(58)}
(S^1++S^2++S^3+)∣→00⟩=2∣↑↑00⟩,(59)\left(\hat{S}_1^++\hat{S}_2^++\hat{S}_3^+\right)|\rightarrow00\rangle=\sqrt{2}|\uparrow\uparrow00\rangle, \tag*{(59)}
(S^1++S^2++S^3+)∣00→⟩=12∣00↑↑⟩,(60)\left(\hat{S}_1^++\hat{S}_2^++\hat{S}_3^+\right)|00\rightarrow\rangle=\frac{1}{\sqrt{2}}|00\uparrow\uparrow\rangle, \tag*{(60)}
(S^1++S^2++S^3+)∣00↓↓⟩=∣00↑↓⟩,=12(∣00↑↓⟩+∣00↓↑⟩)+12(∣00↑↓⟩−∣00↓↑⟩)=12∣00→⟩+12∣00−⟩.(61)\begin{aligned} \left(\hat{S}_1^++\hat{S}_2^++\hat{S}_3^+\right)|00\downarrow\downarrow\rangle&=|00\uparrow\downarrow\rangle,\\ &=\frac{1}{2}\left(|00\uparrow\downarrow\rangle+|00\downarrow\uparrow\rangle\right)+\frac{1}{2}\left(|00\uparrow\downarrow\rangle-|00\downarrow\uparrow\rangle\right)\\ &=\frac{1}{\sqrt{2}}|00\rightarrow\rangle+\frac{1}{\sqrt{2}}|00-\rangle. \tag*{(61)} \end{aligned}

The final example is quite interesting since this results in a superposition between a triplet state and a singlet state. The same method can be used to obtain the resulting states when applying the operator (S^1−+S^2−+S^3−)\left(\hat{S}_1^-+\hat{S}_2^-+\hat{S}_3^-\right) and similarly for the B-part of the exchange coupling.

Now, we look at the tight binding part of the Hamiltonian. Since the spin is unchanged by this part, the singlet and the triplet states will not couple, and thus we can look at these parts of the basis states separately. First, let us focus on the 10 singlet states. Here, one example is sufficient to illustrate what happens to all of them:

H^t∣−00⟩=−t(2∣−000⟩+2∣0−00⟩+∣∘00⟩+∣0−0⟩+∣00⟩).(62)\hat{H}_{t}|-00\rangle=-t\left(\sqrt{2}|-000\rangle+\sqrt{2}|0-00\rangle+|\circ00\rangle+|0-0\rangle+|00\rangle\right). \tag*{(62)}

What is important to note here is that a factor of 2\sqrt{2} arises when a singly occupied state become a doubly occupied site, since

∑σc^1σ†c^2σ∣−00⟩=∑σc^1σ†c^2σ12(c^1↑†c^2↓†−c^1↓†c^2↑†)∣vac⟩=12(c^1↑†c^1↓†−c^1↓†c^1↑†)∣vac⟩=2∣−00⟩.(63)\begin{aligned} \sum_{\sigma}\hat{c}_{1\sigma}^{\dagger}\hat{c}_{2\sigma}|-00\rangle=\sum_{\sigma}\hat{c}_{1\sigma}^{\dagger}\hat{c}_{2\sigma}\frac{1}{\sqrt{2}}\left(\hat{c}_{1\uparrow}^{\dagger}\hat{c}_{2\downarrow}^{\dagger}-\hat{c}_{1\downarrow}^{\dagger}\hat{c}_{2\uparrow}^{\dagger}\right)|\mathrm{vac}\rangle\\ &=\frac{1}{\sqrt{2}}\left(\hat{c}_{1\uparrow}^{\dagger}\hat{c}_{1\downarrow}^{\dagger}-\hat{c}_{1\downarrow}^{\dagger}\hat{c}_{1\uparrow}^{\dagger}\right)|\mathrm{vac}\rangle=\sqrt{2}|-00\rangle. \tag*{(63)} \end{aligned}

Here ∣vac⟩|\mathrm{vac}\rangle represents the vacuum state. Due to the antisymmetry of a singlet state we do not need to think about signs. Again, one example is sufficient:

H^t∣↑↑00⟩=∑σ(c^3σ†c^1σ+c^3σ†c^2σ+c^4σ†c^2σ)∣↑↑00⟩=(c^3↑†c^1↑c^2↑†c^2↑†+c^3↑†c^2↑c^1↑†c^2↑†+c^4↑†c^2↑c^1↑†c^2↑†)∣vac⟩=(−c^2↑†c^3↑†+c^1↑†c^3↑†+c^1↑†c^4↑†)∣vac⟩=−∣0↑↑0⟩+∣↑0↑0⟩+∣↑00↑⟩.(64)\begin{aligned} \hat{H}_{t}|\uparrow\uparrow00\rangle=\sum_{\sigma}\left(\hat{c}_{3\sigma}^{\dagger}\hat{c}_{1\sigma}+\hat{c}_{3\sigma}^{\dagger}\hat{c}_{2\sigma}+\hat{c}_{4\sigma}^{\dagger}\hat{c}_{2\sigma}\right)|\uparrow\uparrow00\rangle\\ &=\left(\hat{c}_{3\uparrow}^{\dagger}\hat{c}_{1\uparrow}\hat{c}_{2\uparrow}^{\dagger}\hat{c}_{2\uparrow}^{\dagger}+\hat{c}_{3\uparrow}^{\dagger}\hat{c}_{2\uparrow}\hat{c}_{1\uparrow}^{\dagger}\hat{c}_{2\uparrow}^{\dagger}+\hat{c}_{4\uparrow}^{\dagger}\hat{c}_{2\uparrow}\hat{c}_{1\uparrow}^{\dagger}\hat{c}_{2\uparrow}^{\dagger}\right)|\mathrm{vac}\rangle\\ &=\left(-\hat{c}_{2\uparrow}^{\dagger}\hat{c}_{3\uparrow}^{\dagger}+\hat{c}_{1\uparrow}^{\dagger}\hat{c}_{3\uparrow}^{\dagger}+\hat{c}_{1\uparrow}^{\dagger}\hat{c}_{4\uparrow}^{\dagger}\right)|\mathrm{vac}\rangle\\ &=-|0\uparrow\uparrow0\rangle+|\uparrow0\uparrow0\rangle+|\uparrow00\uparrow\rangle. \tag*{(64)} \end{aligned}

The last term in the Hamiltonian (onsite repulsion) is trivial to include for the four singlet states. Using this basis the 448×448448\times448 Hamiltonian can be diagonalized numerically yielding the results described in the main text. The main conclusion when diagonalizing this Hamiltonian is that the two degenerate ground state are dominated by the two basis states viewed in Eq. (3) and Figure 4.

Two Level Model

Focusing on the dominant basis state in the ground state obtained in the model described above, see Eq. (3) and Figure 4, i.e., going to the limit J≫tJ\gg t, we can make a simple model to understand how the ground state energy changes when increasing the hopping parameter tt. For t=0t=0 the two electrons will localize on sites 2 and 3 with spins antiparallel compared to the chromium spins. When tt increases the two electrons can then hop between two different sites and still avoid each other, which they can do in two different configurations. One electron can hop between site 1 and 2 while the other hop between 3 and 4, see Figure 4 a), or one electron can hop between site 1 and 3 while the other can hop between 2 and 4, see Figure 4 b).

For an electron hopping between two sites, e.g., 1 and 2, the two levels do not have the same energy, since the electron can couple to both Cr spins at site 2 but only couple to one Cr spin at site 1. The gap between these two levels is Δ=3J4\Delta=\frac{3J}{4} since one level has the energy (32+32)J=3J\left(\frac{3}{2}+\frac{3}{2}\right)J=3J and the other has energy 32J\frac{3}{2}J. Thus, we employ the simple Hamiltonian

H^=[3J4tt−3J4],(65)\hat{H}=\begin{bmatrix}\frac{3J}{4}&t\\t&-\frac{3J}{4}\end{bmatrix}, \tag*{(65)}

with the eigenvalues

E=±(3J4)2+t2.(66)E=\pm\sqrt{\left(\frac{3J}{4}\right)^{2}+t^{2}}. \tag*{(66)}

The negative solution is the ground state. In Figure S3 we show this function alongside the exact eigenvalues as a function of tt with J=−0.2529 eVJ=-0.2529\ \mathrm{eV} and U=0.2529 eVU=0.2529\ \mathrm{eV}. The value chosen for JJ is motivated by [8], with UU chosen to be of the same magnitude as JJ. Note that Eq. (66) has been shifted to fit with the exact eigenvalues in the figure. The exact ground state energies follow this simple model quite closely in the entire tt regime investigated. The exact solution is a bit lower in energy, which can be explained by the fact that in this model the system has two possible configurations with the same low energy, see Figure 4. This only lowers the energy slightly since it has to be coordinated such that the electrons can still avoid each other. These features are robust towards changes in UU (even turning off repulsion completely) since electrons occupying the same site is not favorable in this model.

Five lowest degenerate eigenenergies plotted as a function of $t$

Figure S3. The five lowest, degenerate eigenenergies plotted as a function of tt with J=−0.2529J=-0.2529 eV and U=0.2529U=0.2529 eV, compared with the function in Eq. (66), shifted to match at t=0t=0.

When t=0t=0 the ground state energy becomes five-fold degenerate. Setting t≠0t\ne0 this energy splits in three of which two are two-fold degenerate. The five-fold degeneracy in the limit where J≫tJ\gg t indicates that the system behaves as a spin-2 particle. This makes sense since

S^totz∣3/2⟩⊗∣3/2⟩⊗∣0↓↓⟩=(S^Az+S^Bz+S^2z+S^3z)∣3/2⟩⊗∣3/2⟩⊗∣0↓↓⟩=(32+32−12−12)∣3/2⟩⊗∣3/2⟩⊗∣0↓↓⟩=2∣3/2⟩⊗∣3/2⟩⊗∣0↓↓⟩.(67)\hat{S}_{\mathrm{tot}}^{z}|3/2\rangle\otimes|3/2\rangle\otimes|0\downarrow\downarrow\rangle=\left(\hat{S}_{A}^{z}+\hat{S}_{B}^{z}+\hat{S}_{2}^{z}+\hat{S}_{3}^{z}\right)|3/2\rangle\otimes|3/2\rangle\otimes|0\downarrow\downarrow\rangle \\ =\left(\frac{3}{2}+\frac{3}{2}-\frac{1}{2}-\frac{1}{2}\right)|3/2\rangle\otimes|3/2\rangle\otimes|0\downarrow\downarrow\rangle \\ =2|3/2\rangle\otimes|3/2\rangle\otimes|0\downarrow\downarrow\rangle. \tag*{(67)}

All these five lowest states are plotted in Figure S3.

Zeeman splitting

To see how these states split up in a magnetic field we add the term

H^B=gμBB⋅S^tot=gμBBzS^totz,(68)\hat{H}_{B}=g\mu_{B}\mathbf{B}\cdot\hat{\mathbf{S}}_{\mathrm{tot}}=g\mu_{B}B^{z}\hat{S}_{\mathrm{tot}}^{z}, \tag*{(68)}

where we choose the magnetic field to point in the zz-direction, i.e. out of the plane in Figure 2. As seen in Figure S4 the five-fold degeneracy of the ground state for t=0t=0 is lifted when a magnetic field is applied. Furthermore, we see that for t≠0t\ne0 the energy splits into three, two of which are two-fold degenerate. The two doubly degenerate states act as a doublet and the non-degenerate state behaves as a singlet.

The five lowest eigenenergies as a function of $B_z$ for $t=0$ and $t=0.2529$ eV

Figure S4. The five lowest eigenenergies of Eq. (41) + Eq. (68) depicted as a function of BzB_z with J=−0.2529J=-0.2529 eV=−U=-U for t=0t=0 eV (a) and t=0.2529t=0.2529 eV (b). The degeneracy is lifted by Zeeman splitting for B≠0B\ne0.

Adding a magnetic field also allows us to calculate the expectation value of the total spin using

⟨Stot⟩=∂E∂B.(69)\langle S_{\mathrm{tot}}\rangle=\frac{\partial E}{\partial B}. \tag*{(69)}

Both for t=0t=0 and t≠0t\ne0 this is found numerically to be 2.

Estimate of the effective chromium coupling from mean-field theory

From mean-field theory we can derive an expression of the exchange coupling constant between neighboring chromium spins Si\mathbf{S}_i. We depart from

H^=−JCr−Cr∑⟨i,j⟩Si⋅Sj.(70)\hat{\mathcal{H}}=-J_{\mathrm{Cr-Cr}}\sum_{\langle i,j\rangle}\mathbf{S}_i\cdot\mathbf{S}_j. \tag*{(70)}

Using mean-field theory and neglecting the constant shift, the interaction of the iith spin is

H^i=−JCr−Crz⟨S⟩⋅Si,(71)\hat{\mathcal{H}}_i=-J_{\mathrm{Cr-Cr}}z\langle\mathbf{S}\rangle\cdot\mathbf{S}_i, \tag*{(71)}

where zz denotes the number of neighboring spins (the coordination number), i.e., z=4z=4 for a square lattice. Defining

H≡−zJCr−CrgμB⟨S⟩,(72)\mathbf{H}\equiv-\frac{zJ_{\mathrm{Cr-Cr}}}{g\mu_{\mathrm{B}}}\langle\mathbf{S}\rangle, \tag*{(72)}

the Hamiltonian for the iith spin can be written as

H^i=gμBSi⋅H.(73)\hat{\mathcal{H}}_i=g\mu_{\mathrm{B}}\mathbf{S}_i\cdot\mathbf{H}. \tag*{(73)}

Thus, H\mathbf{H} can be viewed as an exchange field affecting the iith spin, which arises from the polarization of the spins on the neighboring chromium sites. The magnetization from this exchange field is then given by

M=−gμBN⟨S⟩,(74)\mathbf{M}=-g\mu_{\mathrm{B}}N\langle\mathbf{S}\rangle, \tag*{(74)}

where NN is the number of magnetic chromium atoms in the sample. From this an expression for ⟨S⟩\langle\mathbf{S}\rangle can be obtained, such that

H=AM,withA=zJCr−Crg2μB2N.(75)\mathbf{H}=A\mathbf{M},\qquad\text{with}\qquad A=\frac{zJ_{\mathrm{Cr-Cr}}}{g^2\mu_{\mathrm{B}}^2N}. \tag*{(75)}

Including an external magnetic field B\mathbf{B} the total magnetization magnitude can be written as

M=CT(B+AM),(76)M=\frac{C}{T}(B+AM), \tag*{(76)}

where CC is the Curie constant, which for L=0L=0, where LL is the total orbital angular momentum, is given by

C=μ0μB23kBNg2S(S+1).(77)C=\frac{\mu_0\mu_{\mathrm{B}}^2}{3k_{\mathrm{B}}}Ng^2S(S+1). \tag*{(77)}

The magnetization magnitude can now be isolated:

M=CT−ACB.(78)M=\frac{C}{T-AC}B. \tag*{(78)}

Using the Curie Weiss law, we can conclude that AC=TCAC=T_C, where TCT_C is the Curie temperature, from which we obtain the expression

JCr−Cr=3kBTCμ0zS(S+1).(79)J_{\mathrm{Cr-Cr}}=\frac{3k_{\mathrm{B}}T_C}{\mu_0zS(S+1)}. \tag*{(79)}

Exchange Interaction between Cr and pyz

Here we study the interaction between chromium spins and neighboring pyrazine spins to provide an explanation as to why CrCl2(pyz)2\mathrm{CrCl}_{2}(\mathrm{pyz})_{2} is ferrimagnetic. An expression for the change in energy, that arise from the interaction, is obtained through second order perturbation theory [34],

ΔE(2)=−∣⟨ψ∣H^t∣ϕ⟩∣2−Δ+U−∣⟨ψ∣H^t∣ϕ⟩∣2Δ=−∣⟨ψ∣H^t∣ϕ⟩∣2(1(1−Δ/U)Δ)≡J.(80)\Delta E^{(2)}=-\frac{\left|\langle\psi|\hat{H}_{t}|\phi\rangle\right|^{2}}{-\Delta+U}-\frac{\left|\langle\psi|\hat{H}_{t}|\phi\rangle\right|^{2}}{\Delta}=-\left|\langle\psi|\hat{H}_{t}|\phi\rangle\right|^{2}\left(\frac{1}{(1-\Delta/U)\Delta}\right)\equiv J. \tag*{(80)}

∣ψ⟩|\psi\rangle represents the highest occupied energy level on the chromium site in the environment. This is expressed as a superposition of the five d-orbitals and is obtained from the band structure studied previously. ∣ϕ⟩|\phi\rangle can be written as

∣ϕ⟩=12(∣ϕ1⟩+∣ϕ2⟩),(81)|\phi\rangle=\frac{1}{\sqrt{2}}\left(|\phi_{1}\rangle+|\phi_{2}\rangle\right), \tag*{(81)}

where ∣ϕ1⟩|\phi_{1}\rangle and ∣ϕ2⟩|\phi_{2}\rangle are the eigenstates of the relevant energy level on the two pyrazine sites in the unit cell, also obtained from the band structure. We assume that J/t≪1J/t \ll1. Δ\Delta denotes the difference in energy of the two relevant energy levels and UU is the strength of the Hubbard repulsion for two electrons occupying the same chromium site, see Figure S5. The Hubbard repulsion for the pyrazine sites is neglected due to their much larger spatial extent compared to chromium.

Figure S5. Diagram illustrating the two ways in which an electron on the chromium site of one unit cell (five left energy levels in each illustration) can exchange location with an electron on a pyrazine site (6 right energy levels).

In the exchange interaction two electrons exchange location with one intermediate step. This can be done in two ways, both are illustrated in Figure S5. The exchange interaction constant JJ is a sum over these two possibilities, which results in Eq. (80). For the coupling between the Cr sites and the pyrazine sites to be antiferromagnetic, it is required that Δ>U\Delta> U.

RKKY-Interaction

In this section, we derive an expression for the effective chromium coupling, as determined by means of second order perturbation theory of the RKKY interaction mediated through intermediate pyrazine sites. The starting point of this is the exchange interaction between the spin of the Cr ions, S\mathbf{S}, and the spin of the valence electron, s^\hat{\mathbf{s}}, on the pyrazine sites, described above, with the interaction Hamiltonian

H^′=−J∑i,δSi⋅s^i+δ(82)\hat{\mathcal{H}}'=-J\sum_{i,\delta}\mathbf{S}_i\cdot\hat{\mathbf{s}}_{i+\delta} \tag*{(82)}

for unit cell ii. δ\delta sums over the vectors pointing to the four neighboring pyrazine sites. We treat this interaction as a perturbation to the ground state configuration. The second order correction to the energy is given by

Ek(2)=∑k′≠k∣⟨ψk′0∣H^′∣ψk0⟩∣2ξk0−ξk′0=∑k′≠k⟨ψk0∣H^′∣ψk′0⟩⟨ψk′0∣H^′∣ψk0⟩ξk0−ξk′0,(83)E_k^{(2)}=\sum_{k'\ne k}\frac{\left|\langle\psi_{k'}^0|\hat{\mathcal{H}}'|\psi_k^0\rangle\right|^2}{\xi_k^0-\xi_{k'}^0}=\sum_{k'\ne k}\frac{\langle\psi_k^0|\hat{\mathcal{H}}'|\psi_{k'}^0\rangle\langle\psi_{k'}^0|\hat{\mathcal{H}}'|\psi_k^0\rangle}{\xi_k^0-\xi_{k'}^0}, \tag*{(83)}

where ∣ψk0⟩|\psi_k^0\rangle is the ground state of the delocalized electrons, i.e., Bloch waves, and where ξk0\xi_k^0 is the normal-state, unperturbed dispersion relation. The chromium spins in H^′\hat{\mathcal{H}}' can be taken outside of the averages since these are not considered as a part of the perturbation. Now, we are left with a product of two correlation functions. Recall that we can write the spin of the valence electrons as

∑is^i+δ=12c^i+δ,σ†τσσ′c^i+δ,σ′=12N∑kk′c^kσ†τσσ′c^k′σ′.(84)\sum_i\hat{\mathbf{s}}_{i+\delta}=\frac{1}{2}\hat{c}_{i+\delta,\sigma}^{\dagger}\boldsymbol{\tau}_{\sigma\sigma'}\hat{c}_{i+\delta,\sigma'}=\frac{1}{2N}\sum_{kk'}\hat{c}_{k\sigma}^{\dagger}\boldsymbol{\tau}_{\sigma\sigma'}\hat{c}_{k'\sigma'}. \tag*{(84)}

In the textbook version of the RKKY model the valence electrons are considered as a free electron gas. In this case we consider the electrons to be delocalized but anchored to the four pyz sites. For the spin operators in each correlation function we must sum over all bonds to the relevant chromium spin. The dominant terms here are thus the terms where the neighboring chromium atoms couple through an electron located at a pyz site in between. Thus, we find

En(2)=JCr−Cr=J22N2S1⋅S2χ,(85)E_n^{(2)}=J_{\mathrm{Cr-Cr}}=\frac{J^2}{2N^2}\mathbf{S}_1\cdot\mathbf{S}_2\chi, \tag*{(85)}

where χ\chi is the Lindhard function given by

χ=[loop diagram]=∑kk′nF(ξk)−nF(ξk′)ξk−ξk′.(86)\chi=\text{[loop diagram]}=\sum_{kk'}\frac{n_F(\xi_k)-n_F(\xi_{k'})}{\xi_k-\xi_{k'}}. \tag*{(86)}

The factor of 1/N21/N^2 stems from the Fourier transformation and the factor of two stems from the fact that

⟨ψkσ∣ταβ∣ψkσ⟩=Tr⁡[ταβ]=2δαβ.(87)\langle\psi_{k\sigma}|\tau^{\alpha\beta}|\psi_{k\sigma}\rangle=\operatorname{Tr}[\tau^{\alpha\beta}]=2\delta_{\alpha\beta}. \tag*{(87)}

At T=0T=0 the Fermi functions in Eq. (86) are step functions. The Lindhard function can then be written as

χ=2∑k>kF∑k′<kF1ξk−ξk′=2∫ξF4tdξ∫−4tξFdξ′D(ξ)D(ξ′)ξ−ξ′,(88)\chi=2\sum_{k>k_F}\sum_{k'<k_F}\frac{1}{\xi_k-\xi_{k'}}=2\int_{\xi_F}^{4t}d\xi\int_{-4t}^{\xi_F}d\xi'\frac{D(\xi)D(\xi')}{\xi-\xi'}, \tag*{(88)}

where D(ξ)D(\xi) is the density of states. The density of states for a square lattice with spectrum ξ=2t(cos⁡kx+cos⁡ky)\xi=2t(\cos k_x+\cos k_y) is given by

D(ξ)=N22(πt)2K[1−(ξ4t)2].(89)D(\xi)=\frac{N^2}{2(\pi t)^2}K\left[1-\left(\frac{\xi}{4t}\right)^2\right]. \tag*{(89)}

Here KK is the complete elliptic integral of the first kind. In the subspace of the pyrazine sites there is 1/21/2 electron in each unit cell and thus the band has filling 1/41/4 due to spin.

The density of states is normalized such that ∫−4t4tdξD(ξ)=1\int_{-4t}^{4t}d\xi D(\xi)=1. Thus, the Fermi energy can be found by requiring that

∫−4tξFdξD(ξ)=14.\int_{-4t}^{\xi_F}d\xi D(\xi)=\frac{1}{4}.

The Fermi energy is obtained numerically and the Lindhard function is found to have a numerical value of 0.150.15 for t=1t=1, such that

JCr−Cr=0.15J22t.(90)J_{\mathrm{Cr-Cr}}=0.15\frac{J^2}{2t}. \tag*{(90)}

The value ξF/t=0.15\xi_F/t=0.15 obtained within the above approximations demonstrates that the Fermi surface is well-approximated by a small circle.

References

  1. [1]J. Pitcairn, A. Iliceto, L. Cañadillas-Delgado, O. Fabelo, C. Liu, C. Balz, A. We ilhard, S. P. Argent, A. J. Morris, and M. J. Cliffe, Low-dimensional metal-organic magnets as a route toward the s = 2 haldane phase, Journal of the American Chemical Society 145, 1783 (2023).
  2. [2]G. Graziano, Guilty and charged, Nature Reviews Chemistry 2, 256 (2018).
  3. [3]K. von Klitzing, G. Dorda, and M. Pepper, New method for high-accuracy determination of the fine-structure constant based on quantized hall resistance, Phys. Rev. Lett. 45, 494 (1980).
  4. [4]S. Wu, V. Fatemi, Q. D. Gibson, K. Watanabe, T. Taniguchi, R. J. Cava, and P. Jarillo-Herrero, Observation of the quantum spin hall effect up to 100 kelvin in a monolayer crystal, Science 359, 76 (2018).
  5. [5]F. Crasto de Lima, G. J. Ferreira, and R. H. Miwa, Quantum anomalous hall effect in metal-bis(dithiolene), magnetic properties, doping and interfacing graphene, Phys. Chem. Chem. Phys. 20, 22652 (2018).
  6. [6]Y. Tokura, K. Yasuda, and A. Tsukazaki, Magnetic topological insulators, Nature Reviews Physics 1, 126 (2019).
  7. [7]D. Wang, L. Kong, P. Fan, H. Chen, S. Zhu, W. Liu, L. Cao, Y. Sun, S. Du, J. Schneeloch, R. Zhong, G. Gu, L. Fu, H. Ding, and H.-J. Gao, Evidence for majorana bound states in an iron-based superconductor, Science 362, 333 (2018).
  8. [8]K. S. Pedersen, P. Perlepe, M. L. Aubrey, D. N. Woodruff, S. E. Reyes-Lillo, A. Reinholdt, L. Voigt, Z. Li, K. Borup, M. Rouzières, D. Samohvalov, F. Wilhelm, A. Rogalev, J. B. Neaton, J. R. Long, and R. Clérac, Formation of the layered conductive magnet CrCl2(pyrazine)2 through redox-active coordination chemistry, Nature Chemistry 10, 1056 (2018).
  9. [9]W. Hu, K. Yang, A. Stroppa, A. Continenza, and H. Wu, 2D hybrid CrCl2(N2C4H4)2 with tunable ferromagnetic half-metallicity, J. Mater. Chem. C 9, 5985 (2021).
  10. [10]H. Xie, Y. Qie, I. Muhammad, and Q. Sun, 2D CrCl2(pyrazine)2 monolayer: high-temperature ferromagnetism and half-metallicity, Journal of Physics: Condensed Matter 32, 135801 (2019).DOI
  11. [11]M. Zhao, Y. Huang, Y. Peng, Z. Huang, Q. Ma, and H. Zhang, Two-dimensional metal-organic framework nanosheets: synthesis and applications, Chem. Soc. Rev. 47, 6267 (2018).DOI
  12. [12]Y. Luo, M. Ahmad, A. Schug, and M. Tsotallas, Rising up: Hierarchical metal-organic frameworks in experiments and simulations, Advanced Materials 31, 1901744 (2019).
  13. [13]P. Perlepe, I. Oyarzabal, A. Mailman, M. Yquel, M. Platunov, I. Dovgaliuk, M. Rouzières, P. Négrier, D. Mondieig, E. Suturina, M.-A. Dourges, S. Bonhommeau, R. Musgrave, K. Pedersen, D. Chernysov, F. Wilhelm, A. Rogalev, C. Mathonière, and R. Clérac, Metal-organic magnets with large coercivity and ordering temperatures up to 242°C, Science (New York, N.Y.) 370, 587 (2020).
  14. [14]B. Huang, G. Clark, E. Navarro-Moratalla, D. R. Klein, R. Cheng, K. L. Seyler, D. Zhong, E. Schmidgall, M. A. McGuire, D. H. Cobden, W. Yao, D. Xiao, P. Jarillo-Herrero, and X. Xu, Layer-dependent ferromagnetism in a van der waals crystal down to the monolayer limit, Nature 546, 270 (2017).
  15. [15]W. Han, R. K. Kawakami, M. Gmitra, and J. Fabian, Graphene spintronics, Nature Nanotechnology 9, 794–807 (2014).DOI
  16. [16]L. C. Gómez-Aguirre, B. Pato-Doldán, J. Mira, S. Castro-García, M. A. Señarís-Rodríguez, M. Sánchez-Andújar, J. Singleton, and V. S. Zapf, Magnetic Ordering-Induced Multiferroic Behavior in [CH3NH3][Co(HCOO)3] Metal–Organic Framework, Journal of the American Chemical Society 138, 1122–1125 (2016).
  17. [17]M. Fiebig, T. Lottermoser, D. Meier, and M. Trassin, The evolution of multiferroics, Nature Reviews Materials 1, 16046 (2016).DOI
  18. [18]N. A. Spaldin and R. Ramesh, Advances in magnetoelectric multiferroics, Nature Materials 18, 203 (2019).DOI
  19. [19]Y. Tian, S. Shen, J. Cong, L. Yan, S. Wang, and Y. Sun, Observation of resonant quantum magnetoelectric effect in a multiferroic metal–organic framework, Journal of the American Chemical Society 138, 782–785 (2016).
  20. [20]T. Trickle, Z. Zhang, and K. M. Zurek, Detecting light dark matter with magnons, Phys. Rev. Lett. 124, 201801 (2020).DOI
  21. [21]A. Mitridate, T. Trickle, Z. Zhang, and K. M. Zurek, Detectability of axion dark matter with phonon polaritons and magnons, Phys. Rev. D 102, 095005 (2020).
  22. [22]S. Pavaskar, R. Penco, and I. Rothstein, An effective field theory of magneto-elasticity, SciPost Physics 12, 10.21468/scipostphys.12.5.155 (2022).
  23. [23]P. Perlepe, I. Oyarzabal, L. Voigt, M. Kubus, D. N. Woodruff, S. E. Reyes-Lillo, M. L. Aubrey, P. Négrier, M. Rouzières, F. Wilhelm, A. Rogalev, J. B. Neaton, J. R. Long, C. Mathonière, B. Vignolle, K. S. Pedersen, and R. Clérac, From an antiferromagnetic insulator to a strongly correlated metal in square-lattice MCl2(pyrazine)2 coordination solids, Nat. Commun. 13, 5766 (2022).
  24. [24]H. S. Jarrett, Mechanism of the Ferrimagnetic to Antiferromagnetic Transition in Mn2−xCrxSb, Phys. Rev. 134, A942 (1964).
  25. [25]T.-W. Chen, S.-S. Li, F. Tang, J.-S. Ying, Y. Zhang, Z.-C. Wang, S.-J. Zhang, F.-S. Luo, L. Chen, Y. Fang, and R.-K. Zheng, Magnetotransport and magnetic properties of Cr-modified Mn2Sb epitaxial thin films, Phys. Chem. Chem. Phys. 25, 5785 (2023).
  26. [26]Q. Shen, I. Batashev, F. Zhang, H. Ojiyed, N. van Dijk, and E. Brück, The antiferromagnetic to ferrimagnetic phase transition in Mn₂Sb₁₋ₓBiₓ compounds, Journal of Alloys and Compounds 866, 158963 (2021).
  27. [27]García-Landa, C. Ritter, M. Ibarra, J. Blasco, P. Algarabel, R. Mahendiran, and J. García, Magnetic and magnetotransport properties of the ordered perovskite Sr₂FeMoO₆, Solid State Communications 110, 435 (1999).DOI
  28. [28]M. Tovar, M. T. Causa, A. Butera, J. Navarro, B. Martínez, J. Fontcuberta, and M. C. G. Passeggi, Evidence of strong antiferromagnetic coupling between localized and itinerant electrons in ferromagnetic Sr₂FeMoO₆, Phys. Rev. B 66, 024409 (2002).DOI
  29. [29]A. Pires, Transport on the ferrimagnetic lieb lattice, Journal of Magnetism and Magnetic Materials 547, 168941 (2022).DOI
  30. [30]W. A. Harrison, Electronic Structure and the Properties of Solids, Vol. 1 (Dover Publications, 1989).
  31. [31]J. C. Slater and G. F. Koster, Simplified lcao method for the periodic potential problem, Phys. Rev. 94, 1498 (1954).DOI
  32. [32]F. L. Guttesen, Electronic and Magnetic Properties of YbOCl, VS₂ and CrCl₂(pyz)₂, Master’s thesis, University of Copenhagen (2020), available at https://nbi.ku.dk/english/theses/masters-theses/guttesen/Freja_Liebach_Guttesen.pdf.
  33. [33]H. M. Rietveld, A profile refinement method for nuclear and magnetic structures, Journal of Applied Crystallography 2, 65–71 (1969).
  34. [34]H. Bruus and K. Flensberg, Many-body quantum field theory in condensed matter physics: an introduction (Oxford University Press, United Kingdom, 2003).

Paper details

vlnc:W_69b8brct
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents