betaFeedback

WorksPhysics and Astronomy

Theoretical prediction of Curie temperature in two-dimensional ferromagnetic monolayer

Abstract

Theoretical prediction of Curie temperature (TC) is of vital importance for designing the spintronic devices in two-dimensional (2D) ferromagnetic materials. Herein, based on the extensive investigation of Monte Carlo simulations, we summary and propose an improved method to estimate TC more precisely, which includes the different contributions of multiple near-neighbor interactions. Taking monolayer CrI3 as an example, the trends of TC with biaxial strain are investigated via Monte Carlo simulations, mean-field formulas and our method. Besides, our method is not only accurate and convenient to predicting the TC in 2D ferromagnetic honeycomb lattice CrI3 but it can be extended for predicting the TC of other 2D lattices. Our work paves the way to accelerate the prediction and discovery of novel 2D ferromagnets for spintronic applications.

Introduction

The Curie temperature (TCT_{\mathrm{C}}) is one of the essential properties of two-dimensional (2D) ferromagnetic (FM) materials [1, 2, 3, 4]. Moreover, the method of predicting TCT_{\mathrm{C}} is extremely important not only for the deep understanding of the phase transition from ferromagnetic to paramagnetic phase but also for further exploring novel 2D FM materials in spintronic devices[5, 6, 7]. Typically, mean-field and Heisenberg exchange theories are commonly used to explain the process of ferromagnetic phase transformation. These two theories do lead to different pictures of certain aspects of the collective characteristics and the ferromagnetic phase transformation. Mean-field theory considers the complex interactions in the magnet as a powerful “molecular field” (Weiss field), which exists inside the magnet, and that an individual particle should respond to the average field generated by interactions with surrounding particles [8]. While Heisenberg exchange theory considers long-range exchange interactions between atomic spins mediated by the overlap of the atomic orbitals in the absence of an applied field.

Many efforts have been done to develop an accurate theoretical method for calculating TCT_{\mathrm{C}}. However, most methods based on mean-field theory, generally overestimate the TCT_{\mathrm{C}} of 2D ferromagnets due to their failure to consider the sufficient number of the contributions of near-neighbor magnetic interactions [9, 10, 11, 12, 13]. Monte Carlo (MC) simulation is recognized as a more accurate method for systems where stochastic processes occur [14, 15], but the results of MC are largely affected by the initial lattice size and the number of MC steps per site, leading to a large amount of time cost [16]. Thus, a reliable method for theoretically predicting TCT_{\mathrm{C}} of 2D ferromagnets remains a serious challenge. Early studies showed that the magnitude of TCT_{\mathrm{C}} is strongly influenced by the exchange interactions, lattice shapes, and spins. For a determined lattice type of a crystal with spin S, its TCT_{\mathrm{C}} is proportional to exchange interaction JJ [17]. Additionally, for the simple body-centered and face-centered cubic lattices, the formula KBTC=5(Z−1)(11X−1)/96JK_{\mathrm{B}}T_{\mathrm{C}} = 5(Z-1)(11X-1)/96J reproduces the estimated TCT_{\mathrm{C}} accurately, where X=S(S+1)X = S(S+1) and ZZ is the lattice coordination-number [18]. However, the neglect of relatively complex correlation among spin moments leads to a large deviation between the theoretical magnetic behavior and the experiments near TCT_{\mathrm{C}}. Therefore, there is the potential significance for developing an accurate method to predict TCT_{\mathrm{C}} under an appropriate assumption.

According to the established Mermin-Wagner theorem [19], long-range ferromagnetism is not almost existent in strictly 2D isotropic systems. Excitingly, long-range FM order in the 2D limit has been recently experimentally confirmed in monolayer CrI3_3 (TC≈45T_{\mathrm{C}} \approx45 K). Meanwhile, theoretical results have been reported that the TCT_{\mathrm{C}} of monolayer CrI3_3 is 107 K by Ising model [20, 21], and 109 K by mean-field theory [22, 23], indicating that Ising model and mean-field theory may overestimate TCT_{\mathrm{C}} a lot [24]. Thus, based on abundant MC studies, we propose a dependable method (3NN-MF) for theoretical prediction of TCT_{\mathrm{C}} with considering first-, second-, and third-nearest neighboring exchange interactions, which relies on the parameterization of first-principles calculations to construct a 2D ferromagnetic Heisenberg Hamiltonian.

Our paper is structured as follows. Firstly, based on the results of MC simulations, we propose a more efficient method to estimate TCT_{\mathrm{C}} that incorporates multiple near-neighbor exchange interactions and the single-ion anisotropy (SIA), which is obtained from DFT calculations. To eliminate potential finite-size effects of TCT_{\mathrm{C}}, the periodic boundary conditions and the Binder cumulant are adopted in our calculation. Furthermore, by taking monolayer CrI3_3 as an example, we investigate the physical origin of observed magnetic behavior under biaxial strain, and then compare our results with two mean-field methods and MC simulations. These studies will provide a guideline for strain-induced modulation of TCT_{\mathrm{C}} in 2D ferromagnets and prediction of the high TCT_{\mathrm{C}} materials ahead of experiments.

Calculational Details

The density-functional theory (DFT) calculations are performed using projected augmented wave (PAW) method as implemented in the Vienna ab initio Simulation Package (VASP) [25, 26]. In all of these calculations, we use the generalized gradient exchange-correlation function of the Perdew-Burke-Ernzerhof (PBE) [27] flavor. The Brillouin zone is sampled by 6×6×16 \times6 \times1 k-point grid mesh [28] and a plane wave cutoff energy of 400 eV is used. In addition, a vacuum of 15 Å is applied along the z-axis to avoid any artificial interactions between images. Relaxations are performed until the Hellmann-Feynman force on each atom becomes smaller than 0.05 eV/Å and the total energy is converged to be within 10−510^{-5} eV. Spin-polarization is considered to reproduce the semiconducting nature of this system. For the biaxial strain study of CrI3_3, the lattice constants are changed accordingly for each strain while the atomic positions are fully optimized. In addition, the exchange interactions and the SIA for strained cells are obtained from DFT calculations. Schematizations of crystal structures are prepared by using VESTA [29].

Moreover, the temperature evolutions of the magnetization and susceptibility are obtained by Monte Carlo simulations using the Hamiltonian of Eq. (1). In order to eliminate potential finite-size effects in calculating the TCT_{\mathrm{C}}, we apply the periodic boundary conditions in our calculation, and then the Binder cumulant [30] of CrI3_3 with different lattice sizes also calculated for a finite size scaling analysis. In addition, to obtain reliable simulated data with high accuracy, we start with ferromagnetic configurations and use 600000 sweeps to sufficiently thermalize the system into equilibrium at each temperature, and all statistical results are obtained from next 80000 sweeps.

According to spin Heisenberg model including the magnetic anisotropy factor, we can obtain the magnetic parameters of 2D ferromagnets. The Hamiltonian equation can be written as

Hspin=∑i,jJnSiSj+∑iA(Siz)2,(1)H_{\mathrm{spin}}=\sum_{i,j}J_nS_iS_j+\sum_iA(S_i^z)^2, \tag*{(1)}

where Si(j)\mathbf{S}_{i(j)} is the spin operator on site ii (jj), JnJ_n is the nnth near-neighbor exchange interaction between sites ii and jj, SizS_i^z is the spin component parallel to the zz-direction, and AA is the easy-axis single-ion anisotropy (SIA). The different near neighbors exchange interactions JijJ_{ij} and AA can be calculated based on DFT with local spin density approximation plus Hubbard parameter U level. In addition, we find that it is reasonable to take U = 1.0 eV when investigate the magnetic properties of CrI3_3 with biaxial strain (as shown in Figs. S1 and S2 of supplementary material).

Results and Discussion

Curie temperature in 2D ferromagnet

Previously, the Mermin-Wagner theorem states that long-range ferromagnetic ordering cannot exist in a 2D isotropic Heisenberg spin system with a finite temperature [19]. However, in most 2D FM materials, the atomic arrangement and energy of different directions are anisotropic. The SIA is several orders of magnitude weaker than the exchange interaction in most 2D FM systems, meaning that the Curie temperature is not sensitive to it in light of energy[31]. In a relatively small range of A/J1A/J_1: 0.004 ~ 0.5, the error of KBTC/(J1∣S∣2)K_{\mathrm{B}}T_{\mathrm{C}}/(J_1|\mathbf{S}|^2) is less than 0.4, as shown in Figure 1(a). Nonetheless, we should emphasize that the magnetic anisotropy does play an indispensable role in a long-range magnetic stability under thermal condition.

Image not extracted

Figure 1. The Curie temperature as functions of (a) single-ion anisotropy AA and (b) the first-nearest neighboring exchange interactions J1J_1 in different 2D lattices with S=3/2S = 3/2. (c) The constants β1\beta_1, β2\beta_2 and β3\beta_3 as a function of 1/L1/L for honeycomb lattice, where βn=KBTC/(Jn∣S∣2)\beta_n = K_{\mathrm{B}}T_{\mathrm{C}}/(J_n|S|^2) derived from Eq. (5). (d) The variation of β1\beta_1, β2\beta_2 and β3\beta_3 with 2D honeycomb lattice (Z1=3Z_1 = 3), simple cubic lattice (Z1=4Z_1 = 4), and hexagonal close-packed lattice (Z1=6Z_1 = 6) are represented by black, red and blue bars, respectively.

To make a better prediction of TCT_{\mathrm{C}} in 2D materials, based on the theoretical foundation of mean-filed theory and MC study, we propose an empiric method called 3NN-MF for the prediction of TCT_{\mathrm{C}}. Firstly, the classical Heisenberg spin Hamiltonian of FM material is given by:

EFM=∑n=1∞[ZnJnSi⋅Sj+A(Siz)2]+E0,(2)E_{\mathrm{FM}}=\sum_{n=1}^{\infty}\left[Z_nJ_n\mathbf{S}_i\cdot\mathbf{S}_j+A(S_i^z)^2\right]+E_0, \tag*{(2)}

where ZnZ_n is the lattice coordination-number and E0E_0 is the non-magnetic atomic energy (not sensitive for different magnetic states). When JnJ_n is negative, its magnitude is proportional to the TCT_{\mathrm{C}}[32].

Similarly, the corresponding antiferromagnetic (AFM) energy can be written as:

EAFM=∑n=1∞[(Zn−2m)JnSi⋅Sj+A(Siz)2]+E0,(3)E_{\mathrm{AFM}}=\sum_{n=1}^{\infty}\left[(Z_n-2m)J_n\mathbf{S}_i\cdot\mathbf{S}_j+A(S_i^z)^2\right]+E_0, \tag*{(3)}

where mm is the number of the neighboring anti-parallel spin, which is related to the spin at site ii. According to Eqs. (2) and (3), the energy difference of phase transition from FM to AFM state can be written as:

EFM−EAFM=2∑n=1∞mJn∣S∣2.(4)E_{\mathrm{FM}}-E_{\mathrm{AFM}}=2\sum_{n=1}^{\infty}mJ_n|\mathbf{S}|^2. \tag*{(4)}

Additionally, we then explore the correlation between the multiple exchange interactions JnJ_n and TCT_{\mathrm{C}}. It is well known that TCT_{\mathrm{C}} is proportional to the J1J_1[32, 33], which is also consistent with the results of Figure 1(b). The relationship between J1J_1, J2J_2, J3J_3 and TCT_{\mathrm{C}} are still well nearly linear dependent. Thus, it is possible to find that the magnetic contributions of multiple nearest neighbor interactions to obtain the approximated formula:

KBTC≈−∑n=1∞βnJn∣S∣2,(5)K_{\mathrm{B}}T_{\mathrm{C}}\approx-\sum_{n=1}^{\infty}\beta_n J_n\lvert S\rvert^2, \tag*{(5)}

where βn\beta_n are a set of coefficients and we can obtain the values of βn\beta_n for a certain type of lattice by setting the supercell sizes (L×\timesL) between L = 4 and L =112. If we examine the relationship between L and β1\beta_1, from Figure 1(c), we will find that β1\beta_1 (L→∞\to\infty) is smaller than that of in a finite size. Moreover, one can extrapolate the values of β1\beta_1 with the thermodynamic limit (L→∞\to\infty). Furthermore, as shown in Figure 1(d), we can obtain β1≈0.40\beta_1\approx0.40, 0.63 and 1.20, β2≈1.52\beta_2\approx1.52, 1.00, 1.42 and β3≈0.60\beta_3\approx0.60, 1.67, 1.87 for 2D honeycomb lattice, simple cubic lattice, and hexagonal close-packed lattice, respectively. For example, in 2D hexagonal close-packed of MnBi2_2Te4_4, we adopt the parameters J1=−0.92J_1=-0.92 meV, S=1S=1, and Z1=6Z_1=6, substitute them into Eq. (5) and then estimate the TCT_{\mathrm{C}} is 11.85 K, which is in good agreement with the MC result of TC≈12.0T_{\mathrm{C}}\approx12.0 K [34].

Alternatively, the TCT_{\mathrm{C}} also can be roughly estimated according to Eqs. (4) and (5) as the following (it is convenient for preliminary estimate but not for great accuracy):

KBTC≈−β12n(EFM−EAFM).(6)K_{\mathrm{B}}T_{\mathrm{C}}\approx\frac{-\beta_1}{2n}(E_{\mathrm{FM}}-E_{\mathrm{AFM}}). \tag*{(6)}

The TCT_{\mathrm{C}} is estimated by the energy difference between EFME_{\mathrm{FM}} and EAFME_{\mathrm{AFM}} for 2D FM materials. Obviously, it is different from the traditional mean-filed formulas TC=2(EDLM−EFM)/(3KB)T_{\mathrm{C}}=2(E_{\mathrm{DLM}}-E_{\mathrm{FM}})/(3K_{\mathrm{B}}) [35, 36] or TC=2(E0−EFM)/(3KB)T_{\mathrm{C}}=2(E_0-E_{\mathrm{FM}})/(3K_{\mathrm{B}}) [16],which estimate TCT_{\mathrm{C}} by considering the energy difference between disordered local moment (DLM) EDLME_{\mathrm{DLM}} and ferromagnetic state EFME_{\mathrm{FM}}, and the exchange interactions JJ can be obtained from EDLM−EFME_{\mathrm{DLM}}-E_{\mathrm{FM}} or E0−EFME_0-E_{\mathrm{FM}} in many previous works [16, 35, 36]. It should be noticed, at the Curie point, the thermal fluctuations are required to balance with the energy difference between FM and DLM states, there is still some local magnetization, the traditional mean-field formulas does not capture the local magnetic order above the transition temperature, leading to the overestimation of TCT_{\mathrm{C}} [37]. Therefore, it is evident that the neglect of local magnetic order may give rise to lager deviations of theoretical predicting TCT_{\mathrm{C}} [38, 39, 40, 41].

On the other hand, from a lattice structure point of view, the ferromagnetism may vanish (EAFM−EFM≈0E_{\mathrm{AFM}}-E_{\mathrm{FM}}\approx0) at a certain strain (the inevitable result of the competition between direct interactions and superexchange interactions), which means that the TCT_{\mathrm{C}} is close to zero at the critical transition point between FM and AFM, and this similar conclusion has also been widely reported in other material [42, 43, 44]. Thus, we can obtain the TCT_{\mathrm{C}} by the energy difference between FM and various AFM configurations: KBTC∝EAFM−EFMK_{\mathrm{B}}T_{\mathrm{C}}\propto E_{\mathrm{AFM}}-E_{\mathrm{FM}}.

Curie temperature in CrI3_3

Next, we will focus on theoretical prediction of TCT_{\mathrm{C}} and explore the effects of biaxial strain on the magnetic properties in monolayer CrI3_3. Figure 2(a) shows the crystal structure of monolayer CrI3_3. The Cr ions form a honeycomb network sandwiched by two atomic planes of I atoms. As shown in Figure 2(b), four different magnetic configurations are considered to evaluate the magnetic ground state by comparing their total energies, when the spin orientations are all set to +z+z direction to represent FM configuration, while the adjacent Cr ion spins in the zz-axis are anti-parallel to represent three different AFM configurations: Néel-AFM, Stripy-AFM and Zigzag-AFM, respectively.

Top and side views of monolayer $\mathrm{CrI}_{3}$ and four magnetic orders

Figure 2. (a) Top and side views of monolayer CrI3\mathrm{CrI}_{3}. The black solid lines indicate the supercell of 1×2×11 \times2 \times1. (b) The four magnetic orders, namely FM, Néel-AFM, Stripy-AFM, Zigzag-AFM, and non-magnetic atoms are not shown in schematic diagram.

Here, we take 2D honeycomb magnetic lattice as an example to predict TCT_{\mathrm{C}}. According to Eqs. (2) and (3), a 1×2×11\times2\times1 supercell contains four magnetic atoms as shown in Figure 2, the corresponding energy difference per unit cell can be derive as:

EFM−EAFMNeˊel=(12J1+0J2+12J3)∣S∣2/4,EFM−EAFMStripy=(8J1+16J2+0J3)∣S∣2/4,EFM−EAFMZigzag=(4J1+16J2+12J3)∣S∣2/4.(7)\begin{aligned} E_{\mathrm{FM}}-E_{\mathrm{AFM}}^{\mathrm{Néel}}&=(12J_1+0J_2+12J_3)|S|^2/4,\\ E_{\mathrm{FM}}-E_{\mathrm{AFM}}^{\mathrm{Stripy}}&=(8J_1+16J_2+0J_3)|S|^2/4,\\ E_{\mathrm{FM}}-E_{\mathrm{AFM}}^{\mathrm{Zigzag}}&=(4J_1+16J_2+12J_3)|S|^2/4. \tag*{(7)} \end{aligned}

As mentioned above, the Curie temperature can be preliminarily determined by Eq. (6). When n=1n=1, for Néel-AFM, Stripy-AFM and Zigzag-AFM configurations, m=3m=3, 2, 1, respectively, β1=0.40\beta_1=0.40 for Z1=3Z_1=3. Generally speaking, EFM−EAFMZigzagE_{\mathrm{FM}}-E_{\mathrm{AFM}}^{\mathrm{Zigzag}} tends to achieve the minimum energy difference between FM and AFM configurations when TCT_{\mathrm{C}} is close to 0 at certain strain, and the exchange interactions (J2J_2 and J3J_3) with a longer distance are much weaker than the first-nearest neighbor interactions (J1J_1). Therefore, we can estimate the TCT_{\mathrm{C}} by the energy difference between EFME_{\mathrm{FM}} and EAFMZigzagE_{\mathrm{AFM}}^{\mathrm{Zigzag}}: KBTC≈−0.2(EFM−EAFMZigzag)K_{\mathrm{B}}T_{\mathrm{C}}\approx-0.2(E_{\mathrm{FM}}-E_{\mathrm{AFM}}^{\mathrm{Zigzag}}). For more accurate studies, J2J_2 and J3J_3 should be considered in 3NN-MF formula. With S=3/2S=3/2 in monolayer CrI3_3, we can get the following formula by Eq. (5) and the results of Figure 1(c):

KBTC≈−0.90J1−3.42J2−1.35J3.(8)K_{\mathrm{B}}T_{\mathrm{C}}\approx-0.90J_1-3.42J_2-1.35J_3. \tag*{(8)}

we notice that β2/β1≈3.8\beta_2/\beta_1\approx3.8 (L→∞L\rightarrow\infty) in honeycomb lattice, β2\beta_2 is just about four times as large as β1\beta_1, which is in good agreement with KBTC∝EAFMZigzag−EFMK_{\mathrm{B}}T_{\mathrm{C}}\propto E_{\mathrm{AFM}}^{\mathrm{Zigzag}}-E_{\mathrm{FM}} (the coefficient of J2J_2 is four times higher than that of J1J_1 in Eq. (7).

As a starting point, we investigate the TCT_{\mathrm{C}} dependence of biaxial strain. The strain ε\varepsilon is defined as ε=(a−a0)/a0×100%\varepsilon=(a-a_0)/a_0\times100\%, where a0a_0 and aa are the lattice constants for the unstrained and strained systems, respectively. As shown in Figure 3 (a), the optimized lattice constant a0a_0 of monolayer CrI3_3 is 7.00 Å. Both the first-nearest neighboring Cr-Cr bond length and Cr-I-Cr bond angle are increasing approximately linearly as the growing biaxial strain. According to the well-known Goodenough-Kanamori-Anderson rules [45, 46, 47], the super-exchange interaction among Cr-I-Cr is the direct consequence of the FM state, especially for materials with the bond angles of super-exchange close to 90°. The bond angle of CrI3_3 with ε=0\varepsilon=0 is around 98°, which probably induces a ferromagnetic super-exchange interaction and makes it FM order.

Image not extracted

Figure 3. (a) The first-nearest neighboring Cr-Cr bond length and Cr-I-Cr bond angle, (b) energy difference ΔE\Delta E (meV) between FM (EFME_{\mathrm{FM}}) and AFM (EAFMNeˊelE_{\mathrm{AFM}}^{\mathrm{Néel}}, EAFMStripyE_{\mathrm{AFM}}^{\mathrm{Stripy}} and EAFMZigzagE_{\mathrm{AFM}}^{\mathrm{Zigzag}}) configurations of CrI3_3 in dependence of biaxial strain.

Then, we investigate the energy difference ΔE\Delta E (ΔE=EFM−EAFM\Delta E = E_{\mathrm{FM}} - E_{\mathrm{AFM}}) to study the magnetic ground state dependence of strain in CrI3_3. As illustrated in Figure 3(b), the FM state of CrI3_3 favors stable from ε=−3%\varepsilon= -3\% to +6%+6\%. By calculating the energies of FM state and three AFM states, one also can determine the exchange interactions J1J_1, J2J_2, and J3J_3 with Eq. (7). As shown in Figure 4, the magnitude of ∣J1∣\lvert J_1\rvert and ∣J2∣\lvert J_2\rvert decrease with increasing strain, while the third-nearest neighboring exchange interaction (J3>0J_3 > 0) maintains AFM coupling, which weaken the ferromagnetism as strain increases. As shown in Table 1, the magnetic ground states are Néel-AFM with strain larger than ∼−3%\sim-3\%, which are also consistent with the previous studies [48, 30]. Notably, the drastic variations of J1J_1 (and J3J_3) for AFM states are probably attributable to the occurrence of phase transitions between FM and AFM states. These similar changes have been reported in monolayer CrGaTe3_3 [49].

Strain (%)J1J_1 (meV)J2J_2 (meV)J3J_3 (meV)Ground stateTC/NT_{\mathrm{C/N}}(MC)(K)TCT_{\mathrm{C}}(3NN-MF)(K)
−417.23−0.61−5.98Néel121.43-
−3.916.78−0.61−5.83Néel118.01-
−3.515.25−0.59−5.36Néel107.14-
−3−2.66−0.600.19FM53.2250.91
−2−2.64−0.570.16FM52.3250.40
0−2.12−0.520.11FM47.8743.89
2−1.46−0.480.10FM34.9734.53
4−1.04−0.430.09FM27.3628.31
6−0.64−0.400.06FM23.3823.03

Table 1. The exchange interactions (J1J_1, J2J_2 and J3J_3), magnetic ground states, and transition temperature (TC/NT_{\mathrm{C/N}} for Curie/Néel temperatures) in CrI3_3. The last 2 columns are the MC and 3NN-MF results.

The evolutions of the three exchange interactions $J_1$, $J_2$ and $J_3$ as a function of biaxial strain in monolayer CrI$_3$

Figure 4. The evolutions of the three exchange interactions J1J_1, J2J_2 and J3J_3 as a function of biaxial strain in monolayer CrI3_3.

As we known, in most case, the lattice size of 2D ferromagnets is easily altered by the biaxial strain, but the lattice symmetry (the numbers of the nnth-nearest neighbors of magnetic atoms) hardly changes. Thus, neglecting the effect of lattice slight deformation which induced by the biaxial strain, we can estimate the TCT_{\mathrm{C}} of CrI3_3 via our proposed 3NN-MF method. To further investigate the dependence of biaxial strain on TCT_{\mathrm{C}}, and verify the validity of 3NN-MF method [Eq. (8)], here three different methods to predict TCT_{\mathrm{C}} of CrI3_3 are chosen for comparison in Figure 5: MC simulations [50], mean-field theory (MFT) formula: KBTC=Z1J1S(S+1)/3K_{\mathrm{B}}T_{\mathrm{C}} = Z_1J_1S(S + 1)/3 and mean-field approximation (MFA) formula: KBTC=3/2J1K_{\mathrm{B}}T_{\mathrm{C}} = 3/2J_1 [32, 33]. The strain-free TCT_{\mathrm{C}} of CrI3_3 calculated by our 3NN-MF method is 43.89 K, and the values calculated by MFT and MFA are 92.29 K and 36.91 K, respectively. Obviously, our method is more closer to that of the experimental result of TC=45T_{\mathrm{C}} = 45 K [51, 52]. Notably, we derive KBTC=15/4J1K_{\mathrm{B}}T_{\mathrm{C}} = 15/4J_1 based on MFT (Z1=3Z_1=3, S=3/2S=3/2 for CrI3_3), which apparently overestimates the TCT_{\mathrm{C}} a lot because of the neglect of the disordered local moments. However, if we consider a simple average ⟨cos⁡2θ⟩=1/2\langle\cos^2\theta\rangle= 1/2, where θ\theta is the angle between spin and zz direction, the MFT would be KBTC=15/8J1K_{\mathrm{B}}T_{\mathrm{C}} = 15/8J_1, and gives an acceptable TC≈46.14T_{\mathrm{C}} \approx46.14 K. However, the MFT and MFA have their considerable limitations where they only include the effect of nearest exchange interaction J1J_1. As we discussed earlier, TCT_{\mathrm{C}} is linearly dependent for JnJ_n: TC∝−J1−αJ2T_{\mathrm{C}} \propto-J_1 - \alpha J_2, we obtain the coefficients α\alpha of J2J_2 is 3.42, whereas the α\alpha are 3/2 for MFA and 15/4 for MFT, respectively. Particularly, β2\beta_2 is about four times larger than β1\beta_1, suggesting that the contribution of J2J_2 cannot be ignored for a reliable prediction. Meaningfully, our work is significantly important for searching high-temperature 2D ferromagnets.

Curie temperatures of $\mathrm{CrI}_{3}$ as a function of biaxial strain

Figure 5. Curie temperatures of CrI3\mathrm{CrI}_{3} as a function of biaxial strain by means of MC simulations, 3NN-MF method, and two mean-field approximation formulas (MFT, MFA).

The SIA is essential for the stabilization of magnetic order in 2D materials. Since there are two magnetic atoms in a primary cell of CrI3_3, the SIA can be estimated as follows: SIA = MAE / (2∣S∣22\lvert S\rvert^2), where the MAE is the magnetic anisotropy energy (MAE), which is defined as the difference between energies corresponding to the magnetization in in-plane and off-plane directions (MAE = E∥−E⊥E_{\parallel} - E_{\perp}). The SIA of CrI3_3 is ∼−0.369\sim-0.369 meV (as shown in Fig. S1 of supplemental material), which is consistent with the previous results [23, 53, 54]. The TCT_{\mathrm{C}} is ∼47.87\sim47.87 KK for CrI3CrI_3 via Binder cumulant crossing method [30] obtained from the crossing point of the different curves that proper averaging with the large system size LL of 50, 60, and 70 (as shown in Figs. S3 (a) and S3 (b) of supplemental material). Moreover, the TCT_C is nearly unchanged when LL is greater than 50.

However, we should emphasize that the 3NN-MF formula and the mean field formulas adopted in our manuscript have the limitations that the neglect of magnetic anisotropy and dipole-dipole interaction could cause the deviation of calculated TCT_C. Generally speaking, magnetic anisotropy and dipole-dipole interaction play merely the role of a “trigger” in the Curie temperature, which is not sensitive to them according to our MC study and previous studies [31]. The 3NN-MF formula can give an approximate TCT_C with a small magnetic anisotropy. For ferromagnets with a certain larger magnetic anisotropy, the MFT and MFA formula may also give a reasonable value of TCT_C with an accepting error. Besides, we only focus on the Curie temperature of 2D ferromagnets monolayer, do not investigate it for antiferromagnets, because our assumption [Eq. (5)] is only suitable for ferromagnet monolayer and it is not consistent with the mean-filed formula and MC results for antiferromagnet. Moreover, the exchange interactions JnJ_n are related to the energy differences between the FM and AFM configurations, but this processing may not be applicable to itinerant magnetic metals which fit the Stoner model rather than the Heisenberg model to understand the mechanism yielding the magnetic order [55].

Summary

In summary, based on MC simulations with anisotropic Heisenberg model in 2D ferromagnets, a theoretical study is conducted on the relationship between multiple exchange interactions and TCT_C, and we investigate the βn\beta_n for different lattice types by fitting the MC results to predict TCT_C. As an example, we calculate TCT_C of monolayer CrI3CrI_3 under a range of biaxial strain by MC simulations, mean-field formulas and our 3NN-MF formula. The exchange interactions J1J_1, J2J_2, J3J_3 and SIA for different strains obtained from DFT calculations, especially J2J_2 is important and should be included in the TCT_C estimation. Under the assumptions we proposed, the 3NN-MF formula including the magnetic contributions of multiple nearest-neighbor interactions are of reliable and convenient. In addition, the method we proposed can be extended to the TCT_C estimation for other types of 2D ferromagnetic lattices. Our work is expected to theoretical predict TCT_C and to give insight into future research to improve the TCT_C of intrinsic 2D ferromagnetic materials.

Supplementary Material

See supplementary material for more information of the effect of Hubbard U on single-ion anisotropy, exchange interactions, as well as Curie temperature and the results for the Binder cumulants in monolayer CrI3\mathrm{CrI}_3.

Data Availability

The data used to support the findings of this study are available within the article and its supplementary material.

Acknowledgments

This work was supported by National Natural Science Foundation of China (No.11804301), the Natural Science Foundation of Zhejiang Province (No.LY21A040008), and the Fundamental Research Funds of Zhejiang Sci-Tech University (No.2021Q043-Y, LGYJY2021015).

References

  1. [1]N. A. Spaldin, "Magnetic materials: fundamentals and applications" (Cambridge university press, Cambridge, UK, 2010).
  2. [2]J. M. Coey, "Magnetism and Magnetic Materials" (Cambridge university press, Cambridge, UK, 2010).
  3. [3]Q. Ma, W. Wan, Y. Li, and Y. Liu, "First principles study of 2D half-metallic ferromagnetism in Janus Mn₂XSb (X = As, P) monolayers", Appl. Phys. Lett. 120, 112402 (2022).DOI
  4. [4]X. Huang, J. Xu, X. Nie, C. Chen, W. Wang, G. Song, X. Jiang, and J.-M. Liu, "Significantly enhanced interlayer ferromagnetic coupling in van der Waals Fe₃GeTe₂ bilayer by Be-ion intercalation", Appl. Phys. Lett. 120, 073106 (2022).DOI
  5. [5]B. Akgenc, E. Vatansever, and F. Ersan, "Tuning of electronic structure, magnetic phase, and transition temperature in two-dimensional Cr-based Janus MXenes", Phys. Rev. B 5, 083403 (2021).DOI
  6. [6]Y. Z. Abdullahi, Z. D. Vatansever, F. Ersan, U. Akinci, O. U. Akturk, and E. Akturk, "Ferromagnetic TM 2 BC (TM = Cr, Mn) monolayers for spintronic devices with high Curie temperature", Phys. Chem. Chem. Phys. 23, 6107 (2021).DOI
  7. [7]Y. Z. Abdullahi, F. Ersan, Z. D. Vatansever, E. Aktürk, and O. Ü. Aktürk, "Exploring the potential of MnX (S, Sb) monolayers for antiferromagnetic spintronics: A theoretical investigation", J. Appl. Phys. 128, 113903 (2020).DOI
  8. [8]J. W. Negele, "The mean-field theory of nuclear structure and dynamics", Rev. Mod. Phys. 54, 913 (1982).
  9. [9]B. Peng and Y. X. Yu, "A Density Functional Theory with a Mean-field Weight Function: Applications to Surface Tension, Adsorption, and Phase Transition of a Lennard-Jones Fluid in a Slit-like Pore", J. Phys. Chem. B 112, 15407 (2008).DOI
  10. [10]E. Şaşıoğlu, L. M. Sandratskii, and P. Bruno, "Pressure dependence of the Curie temperature in Ni₂MnSn Heusler alloy: A first-principles study", Phys. Rev. B 71, 214412 (2005).DOI
  11. [11]R. F. Sabiryanov and S. S. Jaswal, "Ab Initio Calculations of the Curie Temperature of Complex Permanent-Magnet Materials", Phys. Rev. Lett. 79, 155 (1997).DOI
  12. [12]M. Pajda, J. Kudrnovský, I. Turek, V. Drchal, and P. Bruno, "Ab initio calculations of exchange interactions, spin-wave stiffness constants, and Curie temperatures of Fe, Co, and Ni", Phys. Rev. B 64, 174402 (2001).DOI
  13. [13]A. Azouaoui, N. Benzakour, A. Hourmatallah, and K. Bouslyk hane, "Structural, magnetic properties and phase stability of CrAs in the zinc-blende structure: Investigated by first principle calculations and Monte Carlo simulation", Solid State Commun. 323, 114101 (2021).
  14. [14]J. Kotze, "Introduction to Monte Carlo methods for an Ising Model of a Ferromagnet", arXiv: 0803.0217 (2008).arxiv.org/abs/0803.0217
  15. [15]Y. Zhang, B. Wang, Y. Guo, Q. Li, and J. Wang, "A universal framework for metropolis Monte Carlo simulation of magnetic Curie temperature", Comput. Mater. Sci. 197, 110638 (2021).DOI
  16. [16]K. Sato, L. Bergqvist, J. Kudrnovsky, P. H. Dederichs, O. Eriksson, I. Turek, B. Sanyal, G. Bouzerar, H. Katayama-Yoshida, V. A. Dinh, T. Fukushima, H. Kizaki, and R. Zeller, "First-principles theory of dilute magnetic semiconductors", Rev. Mod. Phys. 82, 1633 (2010).
  17. [17]M. D. Kuz'min, "Extended high-temperature series for the Spin- Heisenberg ferromagnet", Philos. Mag. Lett. 99, 338 (2019).
  18. [18]G. S. Rushbrooke and P. J. Wood, "On the Curie points and high temperature susceptibilities of Heisenberg model ferromagnetics", Mol. Phys. 1, 257 (1958).DOI
  19. [19]N. D. Mermin and H. Wagner, "Absence of Ferromagnetism or Antiferromagnetism in One- or Two-Dimensional Isotropic Heisenberg Models", Phys. Rev. Lett. 17, 1133 (1966).
  20. [20]J. Liu, Q. Sun, Y. Kawazoe, and P. Jena, "Exfoliating biocompatible ferromagnetic Cr-trihalide monolayers", Phys. Chem. Chem. Phys. 18, 8777 (2016).
  21. [21]N. Sivadas, M. W. Daniels, R. H. Swendsen, S. Okamoto, and D. Xiao, "Magnetic ground state of semiconducting transition-metal trichalcogenide monolayers", Phys. Rev. B 91, 235425 (2015).DOI
  22. [22]V. Kumar Gudelli and G.-Y. Guo, "Magnetism and magneto-optical effects in bulk and few-layer CrI₃: a theoretical GGA +Ustudy", New J. Phys. 21, 053012 (2019).DOI
  23. [23]Y. Fang, S. Wu, Z.-Z. Zhu, and G.-Y. Guo, "Large magneto-optical effects and magnetic anisotropy energy in two-dimensional Cr₂Ge₂Te₆", Phys. Rev. B 98, 125416 (2018).DOI
  24. [24]S. Tiwari, M. L. Van de Put, B. Soree, and W. G. Vandenberghe, "Critical behavior of the ferromagnets CrI₃, CrBr₃, and CrGeTe₃ and the antiferromagnet FeCl₂: A detailed first-principles study", Phys. Rev. B 103, 014432 (2021).DOI
  25. [25]G. Kresse and D. Joubert, "From ultrasoft pseudopotentials to the projector augmented-wave method", Phys. Rev. B 59, 1758 (1999).DOI
  26. [26]G. Kresse and J. Furthmüller, "Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set", Phys. Rev. B 54, 11169 (1996).
  27. [27]J. P. Perdew, K. Burke, and M. Ernzerhof, "Generalized Gradient Approximation Made Simple", Phys. Rev. Lett. 77, 3865 (1996).
  28. [28]H. J. Monkhorst and J. D. Pack, "Special points for Brillouin-zone integrations", Phys. Rev. B 13, 5188 (1976).
  29. [29]K. Momma and F. Izumi, "VESTA 3 for three-dimensional visualization of crystal, volumetric and morphology data", J. Appl. Crystallogr. 44, 1272 (2011).DOI
  30. [30]K. Binder, "Critical properties from Monte Carlo coarse graining and renormalization", Phys. Rev. Lett. 47, 693 (1981).DOI
  31. [31]P. Bruno, "Magnetization and Curie temperature of ferromagnetic ultrathin films: The influence of magnetic anisotropy and dipolar interactions", Mat. Res. Soc. Symp. Proc. 231, 299 (1991).DOI
  32. [32]B. D. Cullity and C. D. Graham, "Introduction to magnetic materials" (John Wiley & Sons, 2011).
  33. [33]L. Webster and J. A. Yan, "Strain-tunable magnetic anisotropy in monolayer CrCl₃, CrBr₃, and CrI₃", Phys. Rev. B 98, 144411 (2018).
  34. [34]M. M. Otrokov, I. P. Rusinov, M. Blanco-Rey, M. Hoffmann, A. Y. Vyazovskaya, S. V. Eremeev, A. Ernst, P. M. Echenique, A. Arnau, and E. V. Chulkov, "Unique Thickness-Dependent Properties of the van der Waals Interlayer Antiferromagnet MnBi₂Te₄ Films", Phys. Rev. Lett. 122, 107202 (2019).
  35. [35]B. L. Gyorffy, A. J. Pindor, J. Staunton, G. M. Stocks, and H. Winter, "A first-principles theory of ferromagnetic phase transitions in metals", J. Phys. Condens. Matter 15, 1337 (1985).
  36. [36]L. Bergqvist and P. H. Dederichs, "A theoretical study of half-metallic antiferromagnetic diluted magnetic semiconductors", J. Phys.: Condens. Matter 19, 216220 (2007).DOI
  37. [37]F. Körmann, A. Dick, B. Grabowski, B. Hallstedt, T. Hickel, and J. Neugebauer, "Free energy of bcc iron: Integrated ab initio derivation of vibrational, electronic, and magnetic contributions", Phys. Rev. B 78, 033102 (2008).
  38. [38]E. E. Kokorina and M. V. Medvedev, "Magnetization and magnetic entropy change of a three-dimensional isotropic ferromagnet near the Curie temperature in the random phase approximation", Phys. B 416, 29 (2013).DOI
  39. [39]T. Oguchi, K. Terakura, and N. Hamada, "Magnetism of iron above the Curie temperature", J. Phys. F: Met. Phys. 13, 145 (1983).
  40. [40]K. S. Chana, J. H. Samson, M. U. Luchini, and V. Heine, "Magnetic short-range order in iron above Tc? Statistical mechanics with many-atom interactions", J. Phys. Condens. Matter 3, 6455 (1991).
  41. [41]A. V. Ruban and O. E. Peil, "Impact of thermal atomic displacements on the Curie temperature of 3d transition metals", Phys. Rev. B 97, 174426 (2018).DOI
  42. [42]M. Pizzochero and O. V. Yazyev, "Inducing Magnetic Phase Transitions in Monolayer CrI₃ via Lattice Deformations", J. Phys. Chem. C 124, 7585 (2020).DOI
  43. [43]S. Rijal, C. Xu, and L. Bellaiche, "Designing frustration in two-dimensional magnetic systems via the application of uniaxial strain", Phys. Rev. B 103, 014442 (2021).DOI
  44. [44]J. Zhang, B. Zhao, C. Ma, and Z. Yang, "Bipolar ferromagnetic semiconductors and doping-tuned room-temperature half-metallicity in monolayer MoX₃(X=Cl, Br,I): An HSE06 study", Phys. Rev. B 103, 075433 (2021).DOI
  45. [45]J. B. Goodenough, "Theory of the role of covalence in the perovskite-type manganites [La, M (II)] MnO₃", Phys. Rev. 100, 564 (1955).DOI
  46. [46]J. Kanamori, "Crystal distortion in magnetic compounds", J. Appl. Phys. 31, S14 (1960).
  47. [47]P. W. Anderson, "New approach to the theory of superexchange interactions", Phys. Rev. 115, 2 (1959).
  48. [48]B. Yang, X. Zhang, H. Yang, X. Han, and Y. Yan, "Strain controlling transport properties of heterostructure composed of monolayer CrI3", Appl. Phys. Lett. 114, 192405 (2019).DOI
  49. [49]M. Yu, X. Liu, and W. J. P. C. C. P. Guo, "Novel two-dimensional ferromagnetic semiconductors: Ga-based transition-metal trichalcogenide monolayers", Phys. Chem. Chem. Phys. 20, 6374 (2018).DOI
  50. [50]J. Zhang, W. W. Yao, L. Sang, X. W. Pan, X. Z. Wang, W. L. Liu, L. F. Wang, and X. M. Ren, "Multi-step structural phase transitions with novel symmetry breaking and inverse symmetry breaking characteristics in a [Ag₄₆](-2) cluster hybrid crystal", Chem. Commun. 56, 462 (2020).
  51. [51]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. D. Xu, "Layer-dependent ferromagnetism in a van der Waals crystal down to the monolayer limit", Nature 546, 270 (2017).DOI
  52. [52]C. Gong, L. Li, Z. L. Li, H. W. Ji, A. Stern, Y. Xia, T. Cao, W. Bao, C. Z. Wang, Y. A. Wang, Z. Q. Qiu, R. J. Cava, S. G. Louie, J. Xia, and X. Zhang, "Discovery of intrinsic ferromagnetism in two-dimensional van der Waals crystals", Nature 546, 265 (2017).DOI
  53. [53]W. B. Zhang, Q. Qu, P. Zhu, and C. H. Lam, "Robust intrinsic ferromagnetism and half semiconductivity in stable two-dimensional single-layer chromium trihalides", J. Mater. Chem. C 3, 12457 (2015).
  54. [54]54C. Song, W. Xiao, L. Li, Y. Lu, P. Jiang, C. Li, A. Chen, and Z. Zhong, "Tunable band gap and enhanced ferromagnetism by surface adsorption in monolayer Cr₂Ge₂Te₆", Phys. Rev. B 99, 214435 (2019).DOI
  55. [55]55H. L. Zhuang, P. Kent, and R. G. J. P. R. B. Hennig, "Strong anisotropy and magnetostriction in the two-dimensional Stoner ferromagnet Fe₃GeTe₂", Phys. Rev. B 93, 134407 (2016).DOI

Paper details

vlnc:W_5db7zmx3
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