Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

High throughput computational screening for 2D ferromagnetic materials: the critical role of anisotropy and local correlations

High throughput computational screening for 2D ferromagnetic materials: the critical role of... High throughput computational screening for 2D ferromagnetic materials: the critical role of anisotropy and local correlations 1 1, 2 1 Daniele Torelli, Kristian S. Thygesen, and Thomas Olsen Computational Atomic-scale Materials Design (CAMD), Department of Physics, Technical University of Denmark, DK-2800 Kgs. Lyngby, Denmark Center for Nanostructured Graphene (CNG), Department of Physics, Technical University of Denmark, DK-2800 Kongens Lyngby, Denmark (Dated: February 18, 2020) The recent observation of ferromagnetic order in two-dimensional (2D) materials has initiated a booming interest in the subject of 2D magnetism. In contrast to bulk materials, 2D materials can only exhibit magnetic order in the presence of magnetic anisotropy. In the present work we have used the Computational 2D Materials Database (C2DB) to search for new ferromagnetic 2D materials using the spinwave gap as a simple descriptor that accounts for the role of magnetic anisotropy. In addition to known compounds we nd 12 novel insulating materials that exhibit magnetic order at nite temperatures. For these we evaluate the critical temperatures from classical Monte Carlo simulations of a Heisenberg model with exchange and anisotropy parameters obtained from rst principles. Starting from 150 stable ferromagnetic 2D materials we nd ve candidates that are predicted to have critical temperatures exceeding that of CrI . We also study the e ect of Hubbard corrections in the framework of DFT+U and nd that the value of U can have a crucial in uence on the prediction of magnetic properties. Our work provides new insight into 2D magnetism and identi es a new set of promising monolayers for experimental investigation. I. INTRODUCTION rst principles calculations, but in most cases the predic- tions have not yet been con rmed by experiments and estimates of the critical temperatures are often unjusti- The nature of magnetic order in two-dimensional (2D) ed or very crude. materials is fundamentally di erent from the three- Two-dimensional CrI has proven to comprise a highly dimensional case. In 3D, magnetic order arises from 3 versatile material. For example, an applied electric spontaneously broken symmetry of the magnetization eld can induce Dzyaloshinskii-Moriya interactions, direction and the magnetic anisotropy only plays a 11,12 and switch the magnetic state in bilayer samples. In marginal role. In 2D, however, the Mermin-Wagner addition, it has been demonstrated that one can obtain theorem prohibits a broken symmetry phase at nite control of in-plane conductivity and valley polarization temperatures and the spin rotational symmetry has to by constructing heterostructures of CrI /graphene and be broken explicitly by magnetic anisotropy. 3 CrI /WSe respectively. van der Waals heterostruc- 3 2 In 2017, two examples of 2D ferromagnetic insulators tures of 2D materials involving magnetic layers thus con- were discovered experimentally. 1) A monolayer of CrI stitute a highly exible platform for designing spin tunnel that exhibits magnetic order below 45 K. 2) A bilayer junctions and could provide new ways to build nanos- of Cr Ge Te with a Curie temperature of 20 K. In 2 2 6 9,15,16 tructured spintronics devices. However, in order to the case of CrI the the magnetic order is driven by a make 2D magnetism technologically relevant there is a strong out-of-plane magnetic anisotropy - a case that is pressing need to nd new 2D materials that exhibit mag- often referred to as Ising-type ferromagnet. In contrast, netic order at higher temperatures. Cr Ge Te has a rather weak magnetic anisotropy and 2 2 6 the magnetic order is maintained in bilayer structures A theoretical search for materials with particular prop- by interlayer exchange couplings. Subsequently, ferro- erties may be based on either experimental databases magnetic order at room temperature has been reported such as the Inorganic Crystal Structure Database (ICSD) 4 5 in monolayers of MnSe and VSe . Both of these are or computational databases where rst principles sim- 2 2 itinerant (metallic) ferromagnets and the origin of mag- ulations are employed to predict new stable materials. netism in these materials is still not completely clari ed. The former approach has been applied to predict new In particular, VSe has an easy plane, which implies lack 2D materials rooted in exfoliation energies of 3D materi- 17{19 of magnetic order by virtue of the Mermin-Wagner theo- als in the ICSD and several materials was found to rem. However, such a two-dimensional spin system may have a magnetically ordered ground state (at T = 0 K). comprise an example of a Kosterlitz-Thouless phase, An example of the latter approach is the Computational 20,21 which is known to display magnetic order due to nite 2D Materials Database (C2DB); presently contain- 7 8 size e ects. More recently, Fe GeTe was reported to ing 3712 2D materials of which 20 % are predicted to be 3 2 host itinerant ferromagnetic order below 130 K, which stable. One advantage of using a theoretical database is originates from strong out-of-plane magnetic anisotropy. that the search is not restricted to materials that are Several other 2D materials have been predicted to exhibit experimentally known in a 3D parent van der Waals either ferromagnetic or anti-ferromagnetic order based on structure. However, materials predicted from theoreti- arXiv:1903.11466v2 [cond-mat.mtrl-sci] 17 Feb 2020 2 cal databases may pose severe challenges with respect to II. METHOD AND COMPUTATIONAL DETAILS synthetization and experimental characterization; even if they are predicted to be stable by rst principles calcu- lations. The materials in C2DB have been found by perform- ing rst principles calculations of hypothetical 2D mate- rials in the framework of density functional theory (DFT) with the Perdew-Burke-Ernzerhof (PBE) exchange- Regarding the magnetic properties of materials, a ma- correlation functional using the electronic structure pack- jor diculty stems from the fact that standard rst prin- 25,26 age GPAW. The geometry of all materials are fully ciples methods can only predict whether or not mag- optimized and the dynamical stability is obtained based netic order is present at T = 0 K. For 2D materials on phonon frequencies at the the center and corners of the Mermin-Wagner theorem implies that magnetic or- the Brillouin zone. The heat of formation is calculated der at T = 0 vanishes at any nite temperature in the with respect to standard references and a material is absence of magnetic anisotropy. A rst principles pre- regarded as thermodynamically stable if it situated less diction of magnetic order at T = 0 is therefore irrele- than 0.2 eV above the convex hull de ned by the 2807 vant unless other properties of the material are taken most stable binary bulk compounds from the OQMD. into account. The question then arises: how to calculate We nd that more than 700 materials in the C2DB are the critical temperature for magnetic order given a set predicted to have a ferromagnetic ground state and 150 of exchange and anisotropy parameters for a particular of these are thermodynamically and dynamically stable. material. It is clear that the Mermin-Wagner theorem The DFT calculations show whether or not the mate- disquali es any standard mean- eld approach because rials have a ferromagnetic ground state at T = 0 and such methods neglect the uctuations that are respon- for insulators the critical temperature can then be ob- sible for deteriorating magnetic order at nite tempera- tained from the descriptor derived in Ref. 24 or Monte tures in the absence of magnetic anisotropy. On the other Carlo simulations. The procedure requires knowledge of hand, the importance of having magnetic anisotropy and exchange and anisotropy parameters, which can be ob- an easy axis for the magnetization (as opposed to an tained from an energy mapping analysis including spin- easy plane) has led many authors to derive the mag- orbit coupling non-selfconsistently. We will brie y out- netic properties from an Ising model for which the critical 22,23 line the approach below. temperatures are known for all Archimedian lattices. The magnetic properties of a system of localized spins However, the Ising model only provides a good magnetic are commonly analyzed in terms of the Heisenberg model. model in the limit of in nite single-ion anisotropy and The most basic ingredient in the model is the isotropic ex- simply provides an upper bound for the critical tem- change interactions arising between neighboring spins as perature in general. For example, in the case of CrI , a consequence of Coulomb repulsion and Pauli exclusion. which is regarded as an Ising-type ferromagnet, the Ising In addition, spin-orbit coupling may lead to magnetic model overestimates the critical temperature by a factor anisotropy, which manifests itself through anisotropic ex- of three. The e ect of nite anisotropy was analyzed in change interactions as well as single-ion anisotropy. 2D Ref. 24, where Monte Carlo simulations and renormal- materials often exhibit (nearly) isotropic magnetic inter- ized spin-wave theory were applied to obtain a simple actions in the plane of the materials and in the following expression for the critical temperature of Ising-type fer- we will restrict ourselves to models of the form romagnets. The expression only depends on the number of nearest neighbors, the nearest neighbor exchange inter- X X X 1 1 z 2 z z actions, and two anisotropy parameters. In the present H = J S S A (S ) B S S ; (1) ij i j i ij i i j 2 2 work we have applied this expression to search the C2DB i6=j i i6=j for ferromagnetic materials with nite critical tempera- where the sums over i and j run over all magnetic sites. tures. For some materials in the C2DB, the magnetic J denotes the isotropic exchange between spins at site i structure is not well approximated by an Ising-type fer- ij and j, B is the anisotropic exchange for spins pointing romagnet and we have performed full Monte Carlo sim- ij out of the plane (here assumed to be the z-direction), ulations to obtain the critical temperatures of these ma- and A is the single-ion anisotropy. We will also assume terials. i that the model is composed of a single kind of magnetic atom, which is characterized by a half-integer S, yielding the maximum possible eigenvalue of S for any site. The The paper is organized as follows. In Sec. II we sum- most general form of the exchange interaction between marize the computational details and discuss the Heisen- sites i and j can be written as S J S , where i ij j berg model, which forms the basis for calculations of crit- ical temperatures in the present work. In Sec. III we J is a 3 3 tensor for a given pair of i and j. This in- ij present the magnetic materials found by searching the cludes the Dzyaloshinkii-Moriya interactions as the anti- C2DB and discuss and compare the calculated critical symmetric part as well as symmetric o -diagonal compo- temperatures with previous works. Sec. IV contains a nents that give rise to Kitaev interactions. Such terms conclusion and outlook. are neglected in the present work since, we are mainly in- 3 a) b) energy di erences between in-plane and out-of-plane spin con gurations for ferromagnetic(anti-ferromagnetic) structures and N is the number of nearest FM(AFM) neighbors with aligned(anti-aligned) spins in the anti- ferromagnetic con guration. For bipartite lattices c) d) N = 0 and N is simply the number of nearest FM AFM neighbors, but the triangular magnetic lattices (for ex- ample the MoS crystal structure) have no natural anti- ferromagnetic con gurations and one has to consider a frustrated con guration where each atom has aligned as FIG. 1. Examples of spin con gurations for the calculation of well as anti-aligned nearest neighbors. ? ? Heisenberg parameters A, B and J : (a) E , (b) E , (c) AFM FM Once the parameters have been determined the Curie k k E and (d) E . AFM FM temperatures can be calculated from an expression ob- tained in Ref. 24, which was derived by tting to classi- cal Monte Carlo (MC) simulations of the model (1) com- terested in critical temperatures which is dominated by bined with renormalized spinwave theory. This simpli- the terms included in Eq. (1). However, we emphasize es the procedure signi cantly compared to running MC that the neglect of such terms as well as the assumption calculations for all materials and still assures a good re- of in-plane magnetic isotropy is an approximation we will 24 liability compared to available experimental data. The make to reduce the set of parameters needed for the iden- expression for the Curie temperature is. ti cation of promising candidates. Below we discuss a Ising few important exceptions exempli ed by materials with S JT T = f (5) large critical temperatures that are not well described by k J (2S 1) the model (1). In order to obtain the magnetic properties of a given where material based on the model (1), one needs to extract the 1=4 parameters J , A and B . In the case of a single mag- ij i ij f (x) = tanh log(1 + x) (6) nn netic element we have A = A and restricting ourselves to nearest neighbor interactions we take J = J and ij is a tted function with = 0:033 and N the num- nn B = B if i; j are nearest neighbors and J = B = 0 Ising ij ij ij ber of nearest neighbors. T is the critical tem- otherwise. The parameters can then be obtained by map- perature of the corresponding Ising model (in units of ping the model to rst principles calculations based on JS =k ), which has been tabulated for all the Archime- density functional theory. In particular, the three pa- dian lattices, and are given by 1.52, 2.27, and 3.64 for rameters can be obtained from the total energies of the honeycomb, square and quadratic lattices respectively. ?(k) ?(k) four spin con gurations E and E , where E FM FM AFM is the energy of a fully ferromagnetic con guration and = A(2S 1) + BSN (7) nn E is an anti-ferromagnetic state that involves anti- AFM is the spinwave gap derived from spinwave theory. It fol- parallel spin alignment. The superscripts ? and k in- lows from the Mermin-Wagner theorem that a positive dicates whether the spinors are lying in the plane of the spinwave gap is a minimal requirement for magnetic or- materials or perpendicular to the plane. A ferromagnetic der in 2D, and  thus provides a crucial parameter that material with E < E , will thus have an out-of-plane FM FM can be used for a rough screening for materials that ex- easy axis. The four con gurations are illustrated in Fig. hibit magnetic order at nite temperature. It should be 1. All energies are evaluated with the geometry obtained noted that for S = 1=2, the single-ion anisotropy alone from the relaxed ferromagnetic ground state. For materi- cannot open a gap in the spinwave spectrum and mag- als with a single magnetic atom in the unit cell, we have netic order thus requires anisotropic exchange. We note doubled the unit cell in order to accommodate the anti- that the present approach can lead to situations where a ferromagnetic con guration. Comparing with Eq. (1) material with an out-of-plane easy axis (E > 0) has FM and approximating the spin operators by classical vec- a negative spinwave gap indicating that the ground state tors we can obtain the parameters as is unstable. For example, for a honeycomb lattice with N N FM FM N = N = 3 and S = 1 Eqs. (2)-(4) one obtains E (1 ) + E (1 + ) nn AFM FM AFM N N AFM AFM A = ; (2) < 0 if E > 3E . This is due to the factor AFM FM 2S of 2S 1, which replaces a factor of 2S when quantum E E FM AFM B = ; (3) corrections to the anisotropy terms are taken into ac- N S AFM 3,24 count in renormalized spinwave theory. In principle k k E E this is inconsistent with the energy mapping approach, AFM FM J = ; (4) which is based on a classical treatment of the Heisenberg N S AFM model. However, a full quantum mechanical energy map- where E = E E are the ping analysis is beyond the scope of the present work. In FM(AFM) FM(AFM) FM(AFM) 4 0.2 Metals Insulators 0.1 0.0 0.1 0.05 0.00 0.05 0 20 40 60 T [K] Δ [eV] FIG. 2. MC calculations of the magnetic moment per atom FIG. 4. Distribution of the calculated parameters J and  for and heat capacity (dE=dT ) calculated as a function of tem- 87 metallic and 270 insulating materials obtained with PBE. perature for CrI . The dashed vertical line at T = 31 K, in- dicates the predicted critical temperature obtained from Eq. (5). have an out-of-plane ferromagnetic magnetization and a negative spinwave gap thus implies that the ground state a) b) c) d) must have in-plane magnetization. The transition metal halogen chalcogen (TMHC) com- prises another crystal structure that deserves an addi- tional comment here. These materials display an atomic structure that resembles a distorted hexagonal magnetic lattice arranged over two layers. Although at least two comparable - but distinct - exchange paths are identi- able, MC calculations show that we can obtain rough estimates of the critical temperatures from the model (5) FIG. 3. Top and side view of a (a) square, (b) honeycomb, by treating it as an hexagonal lattice with a single ef- (c) triangular, (d) TMHC crystal structures. Magnetic atoms fective nearest neighbour coupling obtained from the en- are in blue. ergy mapping analysis. For example, for CrIS we obtain T = 118 K from the model (5), which is in decent agree- Fig. 2 we compare the magnetization and heat capac- ment with the MC results of 140 K including both nearest and next-nearest neighbour exchange interactions. ity obtained from MC calculations of CrI as well as the model result from Eq. (5). The critical temperature can be obtained from the position of the peak in the heat III. RESULTS AND DISCUSSION capacity. The parameters A, B, and J (Eqs. (2)-(4)) and criti- cal temperatures (Eqs. (5)-(7)) have been calculated for In Fig. 5 we show the exchange coupling J and spin- the nearly 550 materials listed in the C2DB database, wave gap  for the stable ferromagnetic materials with which display honeycomb, square or triangular magnetic  > 0. We have performed the calculations for insulat- lattices, including stable as well as unstable materials. ing as well as metallic materials. For metals, the value The calculations were performed with the same plane of S is ill-de ned and here we have simply used the mag- wave cuto and k-point sampling as used for the mag- netic moment localized on the magnetic atoms, which is netic anisotropy calculations in the database. Examples obtained by integrating the magnetization density over of such structures are shown in Fig. 3 and includes the the PAW spheres. Moreover, the Heisenberg model is transition metal dichalcogenides (TMD) in the 1T phase not a reliable starting point for itinerant magnets and to and in the 2H phase (triangular magnetic lattice), com- our knowledge there is no simple method to obtain crit- pounds adopting the FeSe crystal structure (square mag- ical temperatures for metallic ferromagnetic materials. netic lattice), and transition metal trihalides such as CrI For this reason we will not discuss metallic materials any (honeycomb magnetic lattice). In Fig. 4 we show all the further in the present work, but simply note that large calculated parameters J and  for insulators and metals anisotropies and exchange couplings indicate that metal- with triangular, square or honeycomb lattice. The spin- lic compounds such as CoBr , VBr , NiI , and NiBr 3 3 3 3 wave gap was calculated by taking the ground state to could potentially exhibit very high critical temperatures. M [μ ] z B dE/dTk J [eV] 5 FIG. 5. PBE calculations of exchange coupling J (triangles) and spinwave gap  (squares) of stable ferromagnetic materials with  > 0. Green background indicates insulating materials. Formula Structure J [meV]  [meV] S [~] T [K] main reason for the high value of T is the large magnetic FeCl MoS 15.2 0.056 2.0 208 2 2 moment of 4  per Fe atom and an exchange coupling CuCl BiI 15.3 0.058 1.0 37 3 3 of J = 15 meV, which is one of the largest values found CrI BiI 2.3 0.96 1.5 35 3 3 in the present study. Previous ab initio calculations have CoCl CdI 2.0 0.058 1.5 31 2 2 reported that FeCl in the CdI crystal structure (which 2 2 CrBr BiI 2.0 0.23 1.5 23 3 3 is metallic) is more stable compared to the MoS crys- MnO CdI 0.54 0.31 1.5 19 2 2 31,32 tal structure and the Curie temperature was esti- NiCl CdI 7.2 0.001 1.0 14 2 2 mated to 17 K based on mean- eld theory. Our cal- CrCl BiI 1.4 0.033 1.5 13 3 3 culations con rm the stability hierarchy and predict an RuCl MoS 18.7 2.3 2.0 606 2 2 even more stable prototype GeS (formation energy re- RuBr MoS 16.1 1.77 2.0 509 2 2 duced by  30 meV/atom compared to the MoS phase). However, we do not expect out-of-plane long-range fer- TABLE I. List of 2D magnetic insulating materials with posi- romagnetic order in either the CdI or the GeS crystal 2 2 tive exchange coupling J and positive spinwave gap . Struc- structures, since the spinwave gaps are negative in both ture denotes the prototypical crystal structure and S is the cases. Interestingly, FeCl in the CdI crystal structure 2 2 spin carried by each magnetic atom. The critical tempera- has positive single-ion anisotropy (A), which could in- ture T is obtained from Eq. (5). The top part of the table dicate magnetic order. However, a negative anisotropic contains dynamically and thermodynamically stable materi- exchange coupling (B) yields an overall negative spin- als. The lower part of the table contains materials that are wave gap and the material thus serves as a good example not expected to be stable in their pristine form but exhibit of a case where the single-ion anisotropy is not a good high critical temperatures. indicator of magnetic order. To our knowledge there is no experimental reports of isolated 2D FeCl com- pounds. However, FeCl in the BiI crystal structure has 3 3 In general we observe that most compounds contain tran- been intercalated in bulk graphite exhibiting a ferromag- sition metal atoms with 3d valence electrons. In partic- netic transition at temperature T = 8:5 K. . Recently ular Cr, Mn, Fe, Ni and Co, which are all well-known it has also been employed as functional intercalation in elements in magnetic materials. In addition, most of the few-layer graphene compounds to weaken restacking of compounds contain halides, albeit with a few important 34 35 graphene sheets and bilayer graphene compounds to exceptions (for example MnO ). promote magnetic order in graphene. Nevertheless, ac- In Tab. I we show a list of the all the insulating fer- cording to our calculations, the FeCl crystal structure romagnetic materials and the calculated critical temper- is less stable than the FeCl ones and is not expected to atures. The top part of the table contains the stable exhibit ferromagnetic order as free standing layers due materials and in the lower part we have included a few to a negative value of the spinwave gap. In bulk form examples of materials that exhibit very high critical tem- FeCl is known in the CdI crystal structure with in- 2 2 peratures but which are predicted to be unstable in their plane ferromagnetic order, but the long range order is pristine form. stabilized by interlayer anti-ferromagnetic exchange cou- pling, which supports our assertion that exfoliated layers of this type will not exhibit magnetic order. Bulk FeCl FeCl has also been reported to form di erent stacking poly- mormphs of the BiI crystal structure, but the magnetic The largest Curie temperature is found for FeCl in the properties of these materials are not known. MoS crystal structure where we obtain T = 202 K. The 2 C 6 MnO XCl 2 2 Monolayers of MnO in the CdI crystal structure have Bulk CoCl and NiCl are both known to display anti- 2 2 2 2 been exfoliated in 2003, but the magnetic properties ferromagnetic interlayer coupling and in-plane ferromag- have not yet been thoroughly investigated experimen- netic order . As seen in Tab. I, our calculations predict tally. Our calculations con rm a ferromagnetic ground the materials to exhibit out-of-plane order. For NiCl , in agreement with previous calculations , where a criti- however, one should be a bit cautious due to the ex- cal temperature of 140 K was predicted. However, that tremely small value of the spinwave gap  and more result were obtained from energy mapping analysis using accurate calculations could lead to a ground state with PBE+U DFT calculations (U = 3.9 eV) and MC cal- in-plane magnetic order. Experimental measurements on culations based on the Ising model. From simulations bulk samples indicate anomalies in the heat capacity re- of the Heisenberg model - explicitly including the nite lated to magnetic phase transitions at 24 K and 52 K. anisotropy - we obtain a T ) of 63 K using Heisenberg While the rst result is in good agreement with our pre- parameters from a pure PBE calculations. The e ect of dicted properties, the second one is signi cantly higher Hubbard correction will be discussed in the next section. and could be related to an additional phase transition in the 3D structure. YCl Metastable high-T compounds A critical temperature of T = 55 K is found for YCl C 2 in the MoS prototype. There is neither experimental The lower part of Tab. I shows two materials that or theoretical reports on this material and it could pose we do not predict to be stable, but may be of interest an interesting new 2D magnetic compound. Our calcula- due to the large predicted critical temperatures. Here tions indicate that it is highly stable in the ferromagnetic we comment brie y on the case of RuCl in the MoS 2 2 con guration with magnetic moment of 1  per Y atom. crystal structure, which we predict to be a dynamically However, since the material comprises a spin-1/2 system stable insulator with a critical temperature of 598K. It the classical MC calculations of the critical temperature is, however, situated 0.5 eV above the convex hull, which may not be very accurate. will mostly likely pose an obstacle to experimental syn- thesis. Nevertheless, the calculations show that very high values of critical temperatures are indeed possible CrX in 2D materials with realistic atomic-scale parameters. It should be mentioned that monolayers of RuCl in the As reported in a previous study employing the same BiI crystal structure have been exfoliated and character- 24 40 41 ized experimentally. Moreover, in a recent study the method we predict CrI in the BiI structure to have 3 3 T = 31 K, while the similar compounds and CrCl and critical temperature of monolayer RuCl was calculated C 3 using DFT and MC simulations based on the Heisenberg CrBr have T of 9 K and 19 K respectively. We note 3 C that our calculated critical temperature for CrI is some- model and found T = 14:21 K. However, a Hubbard term is required to open a gap and RuCl in the BiI what lower than the experimental value of 45 K. This is 3 3 42,43 mainly due to the fact that PBE tends to underestimate crystal structure is metallic within PBE, which is why we do not include it in Tab. I. the exchange coupling and can be improved by using a PBE+U scheme as discussed below. CrCl and CrBr 3 3 have not previously been described in their 2D form, but In-plane anisotropy are known as ferromagnetic bulk compounds consisting of layers in the BiI crystal structure with out-of-plane magnetization. The experimental Curie temperatures As mentioned above, materials with the TMDH crys- of bulk CrCl , CrBr , and CrI are 27 K, 47 K, and 70 K 3 3 3 tal structure have been considered as e ective triangular respectively. Our calculated values show the same hierar- magnetic lattices with a single nearest-neighbour cou- chy, but are reduced compared to the bulk values due to pling. However, this model can only be used for a rough the lack of stabilization by interlayer exchange coupling. screening of materials. For example, CrIS exhibits a strong in-plane anisotropy and the axis of magnetiza- tion are ordered (from the hardest axis to the easiest) CuCl as: x, z and y. In Eq. (1), in-plane anisotropy is not considered and we thus extend the model with the full For CuCl in the BiI crystal structure we nd a criti- set of anisotropy parameters A , A , B , and B that 3 3 x y x y cal temperature of 33 K, which is similar to the calculated measures the single-ion anisotropy and anisotropic ex- value of CrI . The material does, however, lie above the change with respect to both x and y directions (relative convex hull by 0.15 eV per atom, which could complicate to the z-direction). These parameters can be found by experimental synthesis and characterization. generalizing the energy mapping analysis Eqs. (2)-(3) to 7 J J A A B B Easy axis T 1 2 x y x y C CrIS 5.71 4.85 0.084 -0.223 0.025 0.033 x,z,y 140 J MnClN 2.66 5.76 0.023 0.044 0.022 0.012 x,y,z 75 CrClO 1.08 0.74 -0.010 0.034 0.004 0.001 y,x,z 15 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 0.1 0.0 TABLE II. Ferromagnetic materials in the TMDH crystal 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 0.3 structure. The rst two columns show nearest neighbour and B 0.2 next-nearest neighbour exchange coupling constants in meV. Columns three to six display anisotropy parameters calculated 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 with respect to the two in-plane directions x and y in meV. 1.25 In the second last column we state the crystallographic direc- 1.00 tions of magnetization listed from the hardest to the easiest 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 axis. The last column shows the critical temperature in K obtained from MC calculations with these parameters. 0 1 2 3 4 U [eV] include di erent in-plane directions. We then run MC calculations using the full set of parameters to nd the FIG. 6. Calculated magnetic parameters of CrI as a func- critical temperatures, including nearest and next-nearest 3 tion of U. J , A and B,  are in units of meV and critical neighbours couplings J and J . We nd three insulat- 1 2 temperatures T are in units of K. ing materials in this crystal structure that shows ferro- magnetic order. The results are shown in Tab. II. In Element Fe Mn Cr Co Ni V Cu particular, CrIS is predicted to have a critical tempera- U [eV] 4.0 3.8 3.5 3.3 6.4 3.1 4.0 ture of 140 K. We also note that we obtain a Curie tem- perature of 15 K for CrClO, which has previously been predicted to have a Curie temperature of 160 K based TABLE III. Hubbard parameters employed in the PBE+U on an Ising model approach. Again, this comparison calculations. emphasizes that the magnetic anisotropy cannot simply be regarded as a mechanism that xes the magnetization to the out-of-plane direction: approximating magnetic material the structure was relaxed with the given value properties by the Ising model may yield a critical tem- of U, but the stability analysis was based on bare PBE. perature that is wrong by an order of magnitude. The inclusion of U can have a rather dramatic e ect on the results. For example, the magnetic con guration of the ground state or the magnetic moment localized on A. Hubbard U the transition metal ion may change. The results for the stable materials are shown in Fig. 7, while the insulating systems are listed in Tab. IV. Including a Hubbard term Almost half of the materials present in C2DB contain in the entire work ow (from the relaxation step onward) at least one element with a partially lled d-shell. Local a ects quantitatively and in some cases also qualitatively and semi-local xc-functionals such as PBE are known to the ground-state. This means that the magnetic moment, overestimate delocalization of correlated electrons, due to the energy gap or the sign of  may change, making the the uncompensated Coulomb self-interaction of the elec- comparison with Tab. I meaningful only for a subset of tron. In the Hubbard model a term is introduced that materials. Compounds that are also present in Tab. I acts as an e ective electronic on-site repulsion and pro- are shown in bold face for comparison. vides a penalty to delocalization. In order to determine the in uence of the Hubbard correction we have recalcu- lated exchange and anisotropy parameters for CrI for a E ect of U on T range of U values in the PBE+U scheme. The structure was fully relaxed for each value of U and the results are shown in Fig. 6. We observe that an increasing value of For MnO , the main e ect of adding a Hubbard correc- U leads to an overall increase of both  and J , which tion is to increase the exchange parameter J by a factor result in higher critical temperatures. The dependence of two. Interestingly the anisotropy parameters A and B of T on U is roughly linear with T increasing by 5 K decrease and the spinwave gap  becomes less than half C C per eV that U is increased. the value obtained with PBE. Nevertheless, the overall In order to gain more insight into the general in u- e ect is an increase of the critical temperature from 63 ence of U for the calculations of magnetic properties, we K to 82 K. This number can be compared to the result have recalculated the magnetic parameters and critical in Ref. 39 where a critical temperature of 140 K was temperature for all magnetic materials in the C2DB con- estimated from the Ising model. taining 3d valence electrons. We used the optimal values For CrI we predict a critical temperature of 50 K with determined in Ref. 45 and listed in Tab. III. For each PBE+U. This is signi cantly closer to the experimental 8 FIG. 7. PBE+U calculations of exchange coupling J (triangles) and spinwave gap  (squares) of stable ferromagnetic materials with  > 0. Green background indicates insulating materials. 37,47 Formula Structure J [meV]  [meV] S [~] T [K] moments being aligned in the individual planes. This MnO CdI 6.43 0.125 1.5 82 could indicate that PBE+U provides a more accurate 2 2 CoCl CdI 3.21 0.249 1.5 57 2 2 description than PBE, which does not predict magnetic CrI BiI 3.95 1.280 1.5 50 3 3 order for MnI . For MnBr in the CdI crystal struc- 2 2 2 CrBr BiI 2.82 0.185 1.5 24 3 3 ture, neutron scattering experiments on the bulk par- MnI CdI 0.40 0.081 2.5 21 2 2 ent structure revealed an anti-ferromagnetic order below MnBr CdI 0.41 0.024 2.5 16 2 2 T = 2:16 K with magnetic moments lying in-plane. CrCl BiI 2.19 0.016 1.5 10 3 3 4 However, this is not necessarily in contradiction with our NiCl CdI 5.69  10 1.0 7 2 2 calculations since the observed anti-ferromagnetic con g- FeBr BiI 0.04 0.124 2.5 2 3 3 uration is "double-striped", a con guration that has not CoO FeSe 106.54 0.199 1.5 520 been considered in the present study. For FeBr the Hub- FeS FeSe 28.99 0.591 2.0 413 3 bard term makes the spin jump from ~=2 to 5~=2per Fe atom and opens a spinwave gap. A previous investigation TABLE IV. List of 2D non-metal materials with positive of this material showed that it is predicted to be a quan- exchange coupling J and spin wave gap , obtained from tum spin Hall insulator with PBE while PBE+U predicts PBE+U calculations. Structure denotes the prototypical a trivial insulator above a critical value of U= 0:18 eV. crystal structure and S is the spin carried by each magnetic atom. The structures in bold are present also in I for com- parison. The top part contains stable materials, whereas the lower part contains materials with large critical temperatures that may be unstable in their pristine form. Metastable high-T compounds The lower part of Tab. IV lists materials, which are not value than the 31 K obtained with bare PBE. Similarly predicted to be completely stable in their pristine form T increases from 19 K to 24 K for CrBr while the T C 3 C according to PBE calculations (we have not performed of CrCl is increased from 9 K to 10 K. The critical tem- a full stability analysis with PBE+U). Bulk CoO has an perature of CoCl is almost una ected. These results anti-ferromagnetic rock-salt structure with a critical tem- indicate that it is non-trivial to predict how the inclu- perature of 293 K . According to PBE calculations the sion of a Hubbard U in uences the calculated critical most stable 2D phase is a metallic CdI crystal structure temperatures in general. (parameters J and  are shown in Fig. 5). In the FeSe For the compounds MnI , MnBr , and FeBr , which 2 2 3 crystal structure, CoO has a low dynamic stability but all have large magnetic moment of 5 per magnetic we report it here due to the very high critical tempera- atom we obtain rather low critical temperatures of 21, ture of 520 K originating from the extraordinarily large 16 and 2 K respectively. This is mainly due to small exchange coupling predicted by PBE+U. values of exchange coupling J for these materials. The inclusion of U in MnI and MnBr increases the elec- FeS in the FeSe crystal structure has a non-magnetic 2 2 tronic gap as well as the spinwave gap. But most impor- ground state with PBE, but is predicted to be highly tantly, it yields a ferromagnetic ground state, while the stable and is situated on the convex hull. With PBE+U ground state is anti-ferromagnetic without the inclusion the ground state becomes ferromagnetic and we predict a of U. For MnI The result appears to be in qualita- high critical temperature of 413 K. According to previous tive agreement with neutron scattering experiments on calculations, however, the true ground state is a striped the bulk compounds, which reports a helical magnetic anti-ferromagnetic con guration, which is not taken into structure below a critical temperature of 3.4 K, with the account in this work. 9 J J A A B B Easy axis T 1 2 x y x y C correlated systems. We have thus tested how the re- CrBrO 1.12 0.78 0.043 -0.010 0.001 0.001 x,z,y 35 sults are modi ed if the parameters are evaluated with CrIO 0.49 -1.46 0.586 -0.123 -0.008 -0.003 x,z,y 25 PBE+U instead and we nd that the predictions do in- CrClO 1.38 1.27 0.007 0.016 0.001  10 y,x,z 20 deed change in a non-systematic way. For the hexagonal and honeycomb systems three materials that were pre- dicted to be ferromagnetic (at nite temperature) are TABLE V. Parameters and results for TMHC structures ob- no longer predicted to show magnetic order when the tained from PBE+U calculations. Symbols and units are the PBE+U scheme is employed and three materials that same as in Tab. II. Materials in bold are the ones listed in were not magnetic with PBE become magnetic with both tables. PBE+U. For the ve materials that are magnetic with both PBE and PBE+U the critical temperatures are slightly di erent in the two approximations. The biggest In-plane anisotropy di erence is seen for CrI where inclusion of U increases the critical temperature from 31 K to 50 K, which is In Tab. V we list Heisenberg parameters and crit- closer to the experimental value of 45 K. ical temperatures for TMHC structures obtained from In the present work we have mainly focused on insu- PBE+U calculations and MC calculations following the lators. This restriction is rooted in the simple fact that same procedure as in the previous section where no Hub- we do not have a reliable way to estimate Curie temper- bard correction was included. Comparing the results atures of metallic 2D magnetic materials. Metallic ferro- with Tab. II, it is noted that MnClN and CrIS are not magnetism in 2D is, however, a highly interesting subject predicted to be ferromagnetic insulators with PBE+U. In and we note that room-temperature magnetism has re- particular, MnClN is predicted to be a metal and CrIS 5 4 cently been reported in the 2D metals VSe and MnSe . 2 2 exhibits a negative spinwave gap. On the other hand, two Moreover, Figs. 4 and 5 indicate that in the C2DB the new materials - CrIO and CrBrO - are predicted to ex- largest values of both spinwave gaps and exchange cou- hibit ferromagnetic order at 25 K and 35 K respectively. plings are found in metallic materials. Clearly, there is pressing need for theoretical developments of 2D itinerant magnetism that can be applied in conjunction with rst principles simulations to provide accurate predictions of IV. CONCLUSIONS the magnetic properties of 2D metallic materials. Finally, we have restricted ourselves to ferromag- We have presented a high throughput computational netic order. Nevertheless, the C2DB contains 241 anti- screening for magnetic insulators based on the Compu- ferromagnetic entries - 50 of which are predicted to be tational 2D Materials Database. In contrast to several stable. The prediction of a novel 2D anti-ferromagnetic previous studies of magnetism in 2D, we have empha- compound would certainly comprise an important step sized the crucial role of magnetic anisotropy and used the forward in the study of 2D magnetism, but the theoretical spinwave gap as a basic descriptor that must necessarily treatment is complicated by the possibility of several or- be positive in order for magnetism to persist at nite dered structures that may coexist at a given temperature temperatures. This criterion severely reduces the num- - in particular for non-bipartite lattice such as the trian- ber of relevant candidates and we end up with 12 stable gular one. We will leave the study of anti-ferromagnetism candidate materials for which the critical temperatures in 2D to future work. were calculated from classical MC simulations. Seven of the materials were predicted to have Curie temperatures V. ACKNOWLEDGEMENTS exceeding that of CrI . The classical MC simulations appear to comprise an accurate method for obtaining the critical temperatures T. Olsen and D. Torelli acknowledges supported from for insulating materials with S > 1=2. However, the the Independent Research Fund Denmark, grant number Heisenberg parameters that enter the simulations may be 6108-00464A. K. Thygesen acknowledges support from sensitive to the approximations used to calculate them. the European Research Council (ERC) under the Euro- In the C2DB all calculations are performed with the PBE pean Union's Horizon 2020 research and innovation pro- functional, which may have shortcomings for strongly gram (Grant No. 773122, LIMA). N. D. Mermin and H. Wagner, Phys. Rev. Lett. 17, 1133 Herrero, and X. Xu, Nature 546 (2017), 10.1038/na- (1966). ture22391. 2 3 B. Huang, G. Clark, E. Navarro-Moratalla, D. Klein, C. Gong, L. Li, Z. Li, H. Ji, A. Stern, Y. Xia, T. Cao, R. Cheng, K. Seyler, D. Zhong, E. Schmidgall, W. Bao, C. Wang, Y. Wang, Z. Q. Qiu, R. J. Cava, S. G. M. A. McGuire, D. Cobden, W. Yao, X. di, P. Jarillo- Louie, J. Xia, and X. Zhang, Nature 546, 265 (2017). 10 D. J. O'Hara, T. Zhu, A. H. Trout, A. S. Ahmed, Y. K. G. K. H. Madsen, R. M. Nieminen, J. K. Nrskov, Luo, C. H. Lee, M. R. Brenner, S. Rajan, J. A. Gupta, M. Puska, T. T. Rantala, J. Schitz, K. S. Thygesen, and D. W. McComb, and R. K. Kawakami, Nano Lett. 18, K. W. Jacobsen, J. Phys. Condens. Matter 22, 253202 3125 (2018). (2010). 5 26 M. Bonilla, S. Kolekar, Y. Ma, H. C. Diaz, V. Kalappattil, A. Hjorth Larsen, J. Jrgen Mortensen, J. Blomqvist, R. Das, T. Eggers, H. R. Gutierrez, M.-H. Phan, and I. E. Castelli, R. Christensen, M. Du lak, J. Friis, M. Batzill, Nat. Nanotechnol. 13, 289 (2018). M. N. Groves, B. Hammer, C. Hargus, E. D. Hermes, J. M. Kosterlitz and D. J. Thouless, J. Phys. C Solid State P. C. Jennings, P. Bjerre Jensen, J. Kermode, J. R. Phys. 6, 1181 (1973). Kitchin, E. Leonhard Kolsbjerg, J. Kubal, K. Kaasb- P. C. W. Holdsworth and S. T. Bramwell, Phys. Rev. B jerg, S. Lysgaard, J. Bergmann Maronsson, T. Max- 49, 8811 (1994). son, T. Olsen, L. Pastewka, A. Peterson, C. Rostgaard, Z. Fei, B. Huang, P. Malinowski, W. Wang, T. Song, J. Schitz, O. Schutt,  M. Strange, K. S. Thygesen, J. Sanchez, W. Yao, D. Xiao, X. Zhu, A. F. May, W. Wu, T. Vegge, L. Vilhelmsen, M. Walter, Z. Zeng, and K. W. D. H. Cobden, J.-H. Chu, and X. Xu, Nat. Mater. 17, 778 Jacobsen, J. Phys. Condens. Matter 29, 273002 (2017). (2018). J. E. Saal, S. Kirklin, M. Aykol, B. Meredig, and C. Gong and X. Zhang, Science 363, eaav4450 (2019). C. Wolverton, JOM 65, 1501 (2013). 10 28 J. Liu, M. Shi, J. Lu, and M. P. Anantram, Phys. Rev. B T. Olsen, Phys. Rev. B 94, 235106 (2016). 97, 8 (2018). C. Xu, J. Feng, H. Xiang, and L. Bellaiche, npj Compu- S. Jiang, J. Shan, and K. F. Mak, Nat. Mater. 17, 406 tational Materials 4, 57 (2018). (2018). T. Olsen, Phys. Rev. B 96, 125143 (2017). 12 31 B. Huang, G. Clark, D. R. Klein, D. MacNeill, E. Navarro- E. Torun, H. Sahin, S. K. Singh, and F. M. Moratalla, K. L. Seyler, N. Wilson, M. A. McGuire, D. H. Peeters, Applied Physics Letters 106, 192404 (2015), Cobden, D. Xiao, W. Yao, P. Jarillo-Herrero, and X. Xu, https://doi.org/10.1063/1.4921096. Nat. Nanotechnol. 13, 544 (2018). Y. Feng, X. Wu, J. Han, and G. Gao, Journal of Materials C. Cardoso, D. Soriano, N. A. Garc a-Mart nez, and Chemistry C 6 (2018), 10.1039/C8TC00443A. J. Fern andez-Rossier, Phys. Rev. Lett. 121, 67701 (2018). M. S. Dresselhaus and G. Dresselhaus, Advances in Physics T. Taniguchi, K.-M. C. Fu, R. Cheng, X. Linpeng, X. Xu, 51, 1 (2002), https://doi.org/10.1080/00018730110113644. D. Zhong, K. L. Seyler, D. Xiao, B. Huang, M. A. McGuire, X. Qi, J. Qu, H.-B. Zhang, D. Yang, Y. Yu, C. Chi, and K. Watanabe, W. Yao, N. Sivadas, and E. Schmidgall, Sci. Z.-Z. Yu, J. Mater. Chem. A 3, 15498 (2015). Adv. 3, e1603113 (2017). N. Kim, K. S. Kim, N. Jung, L. Brus, and K. Burch, D. Mandrus, and J.-G. Park, Nature 563 P. Kim, Nano Lett. 11, 860 (2011), pMID: 21268591, (2018), 10.1038/s41586-018-0631-z. https://doi.org/10.1021/nl104228f. 16 36 Z. Wang, I. Guti errez-Lezama, N. Ubrig, M. Kroner, T. H. Bointon, I. Khrapach, R. Yakimova, A. V. Shytov, M. Gibertini, T. Taniguchi, K. Watanabe, A. Imamo glu, M. F. Craciun, and S. Russo, Nano Lett. 14, 1751 (2014). E. Giannini, and A. F. Morpurgo, Nat. Comm. 9, 2516 M. A. McGuire, Crystals 7 (2017), 10.3390/cryst7050121. (2018). Y. Omomo, T. Sasaki, Wang, and M. Watanabe, J. Am. T. Bj orkman, A. Gulans, A. V. Krasheninnikov, and R. M. Chem. Soc. 125, 3568 (2003). Nieminen, Phys. Rev. Lett. 108, 235502 (2012). M. Kan, J. Zhou, Q. Sun, Y. Kawazoe, and P. Jena, J. M. Ashton, J. Paul, S. B. Sinnott, and R. G. Hennig, Chem. Phys. Lett. 4, 3382 (2013). Phys. Rev. Lett. 118, 106101 (2017). D. Weber, L. M. Schoop, V. Duppel, J. M. Lippmann, N. Mounet, M. Gibertini, P. Schwaller, D. Campi, J. Nuss, and B. V. Lotsch, Nano Lett. 16, 3578 (2016). A. Merkys, A. Marrazzo, T. Sohier, I. E. Castelli, A. Ce- S. Sarikurt, Y. Kadioglu, F. Ersan, E. Vatansever, O. U. pellotti, G. Pizzi, and N. Marzari, Nat. Nanotechnol. 13, Akturk, Y. Yuksel, U. Aknc, and E. Akturk, Phys. Chem. 246 (2018). Chem. Phys. 20, 997 (2018). 20 42 S. Haastrup, M. Strange, M. Pandey, T. Deilmann, P. S. C. Huang, J. Zhou, H. Wu, K. Deng, P. Jena, and E. Kan, Schmidt, N. F. Hinsche, M. N. Gjerding, D. Torelli, P. M. Phys. Rev. B 95, 045113 (2017). Larsen, A. C. Riis-Jensen, J. Gath, K. W. Jacobsen, J. J. F. Ersan, E. Vatansever, S. Sarkurt, Y. Yuksel, Y. Ka- Mortensen, T. Olsen, and K. S. Thygesen, 2D Materials dioglu, H. Ozaydin, O. uzengi akturk, U. Aknc, and 5, 042002 (2018). A. Ethem, Journal of Magnetism and Magnetic Materials T. Olsen, E. Andersen, T. Okugawa, D. Torelli, T. Deil- 476 (2018), 10.1016/j.jmmm.2018.12.032. mann, and K. S. Thygesen, Phys. Rev. Mater. 3, 024005 N. Miao, B. Xu, L. Zhu, J. Zhou, and Z. Sun, J. Am. (2019), 1812.06666. Chem. Soc. 140, 2417 (2018). 22 45 K. Malarz, M. Zborek, and B. Wrobel, Task Q. 9, 475 L. Wang, T. Maxisch, and G. Ceder, Phys. Rev. B 73, (2005). 195107 (2006). 23 46 M. Kan, S. Adhikari, and Q. Sun, Phys. Chem. Chem. V. V. Kulish and W. Huang, J. Mater. Chem. C 5, 8734 Phys. 16, 4990 (2014). (2017). 24 47 D. Torelli and T. Olsen, 2D Materials 6, 015028 (2018). J. W. Cable, M. K. Wilkinson, E. O. Wollan, and W. C. J. Enkovaara, C. Rostgaard, J. J. Mortensen, J. Chen, Koehler, Phys. Rev. 125, 1860 (1962). M. Du lak, L. Ferrighi, J. Gavnholt, C. Glinsvad, E. O. Wollan, W. C. Koehler, and M. K. Wilkinson, Phys. V. Haikola, H. a. Hansen, H. H. Kristo ersen, M. Kuisma, Rev. 110, 638 (1958). a. H. Larsen, L. Lehtovaara, M. Ljungberg, O. Lopez- A. Berkowitz and K. Takano, Journal of Magnetism and Acevedo, P. G. Moses, J. Ojanen, T. Olsen, V. Petzold, Magnetic Materials 200, 552 (1999). N. a. Romero, J. Stausholm-Mller, M. Strange, G. a. Trit- M. Winiarski, J. Zasada, and M. Samsel-Czeka la, Com- saris, M. Vanin, M. Walter, B. Hammer, H. H akkinen, putational Materials Science 142, 372 (2018). http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Condensed Matter arXiv (Cornell University)

High throughput computational screening for 2D ferromagnetic materials: the critical role of anisotropy and local correlations

Condensed Matter , Volume 2019 (1903) – Mar 27, 2019

Loading next page...
 
/lp/arxiv-cornell-university/high-throughput-computational-screening-for-2d-ferromagnetic-materials-PmtTsTHaL3
ISSN
2053-1583
eISSN
ARCH-3331
DOI
10.1088/2053-1583/ab2c43
Publisher site
See Article on Publisher Site

Abstract

High throughput computational screening for 2D ferromagnetic materials: the critical role of anisotropy and local correlations 1 1, 2 1 Daniele Torelli, Kristian S. Thygesen, and Thomas Olsen Computational Atomic-scale Materials Design (CAMD), Department of Physics, Technical University of Denmark, DK-2800 Kgs. Lyngby, Denmark Center for Nanostructured Graphene (CNG), Department of Physics, Technical University of Denmark, DK-2800 Kongens Lyngby, Denmark (Dated: February 18, 2020) The recent observation of ferromagnetic order in two-dimensional (2D) materials has initiated a booming interest in the subject of 2D magnetism. In contrast to bulk materials, 2D materials can only exhibit magnetic order in the presence of magnetic anisotropy. In the present work we have used the Computational 2D Materials Database (C2DB) to search for new ferromagnetic 2D materials using the spinwave gap as a simple descriptor that accounts for the role of magnetic anisotropy. In addition to known compounds we nd 12 novel insulating materials that exhibit magnetic order at nite temperatures. For these we evaluate the critical temperatures from classical Monte Carlo simulations of a Heisenberg model with exchange and anisotropy parameters obtained from rst principles. Starting from 150 stable ferromagnetic 2D materials we nd ve candidates that are predicted to have critical temperatures exceeding that of CrI . We also study the e ect of Hubbard corrections in the framework of DFT+U and nd that the value of U can have a crucial in uence on the prediction of magnetic properties. Our work provides new insight into 2D magnetism and identi es a new set of promising monolayers for experimental investigation. I. INTRODUCTION rst principles calculations, but in most cases the predic- tions have not yet been con rmed by experiments and estimates of the critical temperatures are often unjusti- The nature of magnetic order in two-dimensional (2D) ed or very crude. materials is fundamentally di erent from the three- Two-dimensional CrI has proven to comprise a highly dimensional case. In 3D, magnetic order arises from 3 versatile material. For example, an applied electric spontaneously broken symmetry of the magnetization eld can induce Dzyaloshinskii-Moriya interactions, direction and the magnetic anisotropy only plays a 11,12 and switch the magnetic state in bilayer samples. In marginal role. In 2D, however, the Mermin-Wagner addition, it has been demonstrated that one can obtain theorem prohibits a broken symmetry phase at nite control of in-plane conductivity and valley polarization temperatures and the spin rotational symmetry has to by constructing heterostructures of CrI /graphene and be broken explicitly by magnetic anisotropy. 3 CrI /WSe respectively. van der Waals heterostruc- 3 2 In 2017, two examples of 2D ferromagnetic insulators tures of 2D materials involving magnetic layers thus con- were discovered experimentally. 1) A monolayer of CrI stitute a highly exible platform for designing spin tunnel that exhibits magnetic order below 45 K. 2) A bilayer junctions and could provide new ways to build nanos- of Cr Ge Te with a Curie temperature of 20 K. In 2 2 6 9,15,16 tructured spintronics devices. However, in order to the case of CrI the the magnetic order is driven by a make 2D magnetism technologically relevant there is a strong out-of-plane magnetic anisotropy - a case that is pressing need to nd new 2D materials that exhibit mag- often referred to as Ising-type ferromagnet. In contrast, netic order at higher temperatures. Cr Ge Te has a rather weak magnetic anisotropy and 2 2 6 the magnetic order is maintained in bilayer structures A theoretical search for materials with particular prop- by interlayer exchange couplings. Subsequently, ferro- erties may be based on either experimental databases magnetic order at room temperature has been reported such as the Inorganic Crystal Structure Database (ICSD) 4 5 in monolayers of MnSe and VSe . Both of these are or computational databases where rst principles sim- 2 2 itinerant (metallic) ferromagnets and the origin of mag- ulations are employed to predict new stable materials. netism in these materials is still not completely clari ed. The former approach has been applied to predict new In particular, VSe has an easy plane, which implies lack 2D materials rooted in exfoliation energies of 3D materi- 17{19 of magnetic order by virtue of the Mermin-Wagner theo- als in the ICSD and several materials was found to rem. However, such a two-dimensional spin system may have a magnetically ordered ground state (at T = 0 K). comprise an example of a Kosterlitz-Thouless phase, An example of the latter approach is the Computational 20,21 which is known to display magnetic order due to nite 2D Materials Database (C2DB); presently contain- 7 8 size e ects. More recently, Fe GeTe was reported to ing 3712 2D materials of which 20 % are predicted to be 3 2 host itinerant ferromagnetic order below 130 K, which stable. One advantage of using a theoretical database is originates from strong out-of-plane magnetic anisotropy. that the search is not restricted to materials that are Several other 2D materials have been predicted to exhibit experimentally known in a 3D parent van der Waals either ferromagnetic or anti-ferromagnetic order based on structure. However, materials predicted from theoreti- arXiv:1903.11466v2 [cond-mat.mtrl-sci] 17 Feb 2020 2 cal databases may pose severe challenges with respect to II. METHOD AND COMPUTATIONAL DETAILS synthetization and experimental characterization; even if they are predicted to be stable by rst principles calcu- lations. The materials in C2DB have been found by perform- ing rst principles calculations of hypothetical 2D mate- rials in the framework of density functional theory (DFT) with the Perdew-Burke-Ernzerhof (PBE) exchange- Regarding the magnetic properties of materials, a ma- correlation functional using the electronic structure pack- jor diculty stems from the fact that standard rst prin- 25,26 age GPAW. The geometry of all materials are fully ciples methods can only predict whether or not mag- optimized and the dynamical stability is obtained based netic order is present at T = 0 K. For 2D materials on phonon frequencies at the the center and corners of the Mermin-Wagner theorem implies that magnetic or- the Brillouin zone. The heat of formation is calculated der at T = 0 vanishes at any nite temperature in the with respect to standard references and a material is absence of magnetic anisotropy. A rst principles pre- regarded as thermodynamically stable if it situated less diction of magnetic order at T = 0 is therefore irrele- than 0.2 eV above the convex hull de ned by the 2807 vant unless other properties of the material are taken most stable binary bulk compounds from the OQMD. into account. The question then arises: how to calculate We nd that more than 700 materials in the C2DB are the critical temperature for magnetic order given a set predicted to have a ferromagnetic ground state and 150 of exchange and anisotropy parameters for a particular of these are thermodynamically and dynamically stable. material. It is clear that the Mermin-Wagner theorem The DFT calculations show whether or not the mate- disquali es any standard mean- eld approach because rials have a ferromagnetic ground state at T = 0 and such methods neglect the uctuations that are respon- for insulators the critical temperature can then be ob- sible for deteriorating magnetic order at nite tempera- tained from the descriptor derived in Ref. 24 or Monte tures in the absence of magnetic anisotropy. On the other Carlo simulations. The procedure requires knowledge of hand, the importance of having magnetic anisotropy and exchange and anisotropy parameters, which can be ob- an easy axis for the magnetization (as opposed to an tained from an energy mapping analysis including spin- easy plane) has led many authors to derive the mag- orbit coupling non-selfconsistently. We will brie y out- netic properties from an Ising model for which the critical 22,23 line the approach below. temperatures are known for all Archimedian lattices. The magnetic properties of a system of localized spins However, the Ising model only provides a good magnetic are commonly analyzed in terms of the Heisenberg model. model in the limit of in nite single-ion anisotropy and The most basic ingredient in the model is the isotropic ex- simply provides an upper bound for the critical tem- change interactions arising between neighboring spins as perature in general. For example, in the case of CrI , a consequence of Coulomb repulsion and Pauli exclusion. which is regarded as an Ising-type ferromagnet, the Ising In addition, spin-orbit coupling may lead to magnetic model overestimates the critical temperature by a factor anisotropy, which manifests itself through anisotropic ex- of three. The e ect of nite anisotropy was analyzed in change interactions as well as single-ion anisotropy. 2D Ref. 24, where Monte Carlo simulations and renormal- materials often exhibit (nearly) isotropic magnetic inter- ized spin-wave theory were applied to obtain a simple actions in the plane of the materials and in the following expression for the critical temperature of Ising-type fer- we will restrict ourselves to models of the form romagnets. The expression only depends on the number of nearest neighbors, the nearest neighbor exchange inter- X X X 1 1 z 2 z z actions, and two anisotropy parameters. In the present H = J S S A (S ) B S S ; (1) ij i j i ij i i j 2 2 work we have applied this expression to search the C2DB i6=j i i6=j for ferromagnetic materials with nite critical tempera- where the sums over i and j run over all magnetic sites. tures. For some materials in the C2DB, the magnetic J denotes the isotropic exchange between spins at site i structure is not well approximated by an Ising-type fer- ij and j, B is the anisotropic exchange for spins pointing romagnet and we have performed full Monte Carlo sim- ij out of the plane (here assumed to be the z-direction), ulations to obtain the critical temperatures of these ma- and A is the single-ion anisotropy. We will also assume terials. i that the model is composed of a single kind of magnetic atom, which is characterized by a half-integer S, yielding the maximum possible eigenvalue of S for any site. The The paper is organized as follows. In Sec. II we sum- most general form of the exchange interaction between marize the computational details and discuss the Heisen- sites i and j can be written as S J S , where i ij j berg model, which forms the basis for calculations of crit- ical temperatures in the present work. In Sec. III we J is a 3 3 tensor for a given pair of i and j. This in- ij present the magnetic materials found by searching the cludes the Dzyaloshinkii-Moriya interactions as the anti- C2DB and discuss and compare the calculated critical symmetric part as well as symmetric o -diagonal compo- temperatures with previous works. Sec. IV contains a nents that give rise to Kitaev interactions. Such terms conclusion and outlook. are neglected in the present work since, we are mainly in- 3 a) b) energy di erences between in-plane and out-of-plane spin con gurations for ferromagnetic(anti-ferromagnetic) structures and N is the number of nearest FM(AFM) neighbors with aligned(anti-aligned) spins in the anti- ferromagnetic con guration. For bipartite lattices c) d) N = 0 and N is simply the number of nearest FM AFM neighbors, but the triangular magnetic lattices (for ex- ample the MoS crystal structure) have no natural anti- ferromagnetic con gurations and one has to consider a frustrated con guration where each atom has aligned as FIG. 1. Examples of spin con gurations for the calculation of well as anti-aligned nearest neighbors. ? ? Heisenberg parameters A, B and J : (a) E , (b) E , (c) AFM FM Once the parameters have been determined the Curie k k E and (d) E . AFM FM temperatures can be calculated from an expression ob- tained in Ref. 24, which was derived by tting to classi- cal Monte Carlo (MC) simulations of the model (1) com- terested in critical temperatures which is dominated by bined with renormalized spinwave theory. This simpli- the terms included in Eq. (1). However, we emphasize es the procedure signi cantly compared to running MC that the neglect of such terms as well as the assumption calculations for all materials and still assures a good re- of in-plane magnetic isotropy is an approximation we will 24 liability compared to available experimental data. The make to reduce the set of parameters needed for the iden- expression for the Curie temperature is. ti cation of promising candidates. Below we discuss a Ising few important exceptions exempli ed by materials with S JT T = f (5) large critical temperatures that are not well described by k J (2S 1) the model (1). In order to obtain the magnetic properties of a given where material based on the model (1), one needs to extract the 1=4 parameters J , A and B . In the case of a single mag- ij i ij f (x) = tanh log(1 + x) (6) nn netic element we have A = A and restricting ourselves to nearest neighbor interactions we take J = J and ij is a tted function with = 0:033 and N the num- nn B = B if i; j are nearest neighbors and J = B = 0 Ising ij ij ij ber of nearest neighbors. T is the critical tem- otherwise. The parameters can then be obtained by map- perature of the corresponding Ising model (in units of ping the model to rst principles calculations based on JS =k ), which has been tabulated for all the Archime- density functional theory. In particular, the three pa- dian lattices, and are given by 1.52, 2.27, and 3.64 for rameters can be obtained from the total energies of the honeycomb, square and quadratic lattices respectively. ?(k) ?(k) four spin con gurations E and E , where E FM FM AFM is the energy of a fully ferromagnetic con guration and = A(2S 1) + BSN (7) nn E is an anti-ferromagnetic state that involves anti- AFM is the spinwave gap derived from spinwave theory. It fol- parallel spin alignment. The superscripts ? and k in- lows from the Mermin-Wagner theorem that a positive dicates whether the spinors are lying in the plane of the spinwave gap is a minimal requirement for magnetic or- materials or perpendicular to the plane. A ferromagnetic der in 2D, and  thus provides a crucial parameter that material with E < E , will thus have an out-of-plane FM FM can be used for a rough screening for materials that ex- easy axis. The four con gurations are illustrated in Fig. hibit magnetic order at nite temperature. It should be 1. All energies are evaluated with the geometry obtained noted that for S = 1=2, the single-ion anisotropy alone from the relaxed ferromagnetic ground state. For materi- cannot open a gap in the spinwave spectrum and mag- als with a single magnetic atom in the unit cell, we have netic order thus requires anisotropic exchange. We note doubled the unit cell in order to accommodate the anti- that the present approach can lead to situations where a ferromagnetic con guration. Comparing with Eq. (1) material with an out-of-plane easy axis (E > 0) has FM and approximating the spin operators by classical vec- a negative spinwave gap indicating that the ground state tors we can obtain the parameters as is unstable. For example, for a honeycomb lattice with N N FM FM N = N = 3 and S = 1 Eqs. (2)-(4) one obtains E (1 ) + E (1 + ) nn AFM FM AFM N N AFM AFM A = ; (2) < 0 if E > 3E . This is due to the factor AFM FM 2S of 2S 1, which replaces a factor of 2S when quantum E E FM AFM B = ; (3) corrections to the anisotropy terms are taken into ac- N S AFM 3,24 count in renormalized spinwave theory. In principle k k E E this is inconsistent with the energy mapping approach, AFM FM J = ; (4) which is based on a classical treatment of the Heisenberg N S AFM model. However, a full quantum mechanical energy map- where E = E E are the ping analysis is beyond the scope of the present work. In FM(AFM) FM(AFM) FM(AFM) 4 0.2 Metals Insulators 0.1 0.0 0.1 0.05 0.00 0.05 0 20 40 60 T [K] Δ [eV] FIG. 2. MC calculations of the magnetic moment per atom FIG. 4. Distribution of the calculated parameters J and  for and heat capacity (dE=dT ) calculated as a function of tem- 87 metallic and 270 insulating materials obtained with PBE. perature for CrI . The dashed vertical line at T = 31 K, in- dicates the predicted critical temperature obtained from Eq. (5). have an out-of-plane ferromagnetic magnetization and a negative spinwave gap thus implies that the ground state a) b) c) d) must have in-plane magnetization. The transition metal halogen chalcogen (TMHC) com- prises another crystal structure that deserves an addi- tional comment here. These materials display an atomic structure that resembles a distorted hexagonal magnetic lattice arranged over two layers. Although at least two comparable - but distinct - exchange paths are identi- able, MC calculations show that we can obtain rough estimates of the critical temperatures from the model (5) FIG. 3. Top and side view of a (a) square, (b) honeycomb, by treating it as an hexagonal lattice with a single ef- (c) triangular, (d) TMHC crystal structures. Magnetic atoms fective nearest neighbour coupling obtained from the en- are in blue. ergy mapping analysis. For example, for CrIS we obtain T = 118 K from the model (5), which is in decent agree- Fig. 2 we compare the magnetization and heat capac- ment with the MC results of 140 K including both nearest and next-nearest neighbour exchange interactions. ity obtained from MC calculations of CrI as well as the model result from Eq. (5). The critical temperature can be obtained from the position of the peak in the heat III. RESULTS AND DISCUSSION capacity. The parameters A, B, and J (Eqs. (2)-(4)) and criti- cal temperatures (Eqs. (5)-(7)) have been calculated for In Fig. 5 we show the exchange coupling J and spin- the nearly 550 materials listed in the C2DB database, wave gap  for the stable ferromagnetic materials with which display honeycomb, square or triangular magnetic  > 0. We have performed the calculations for insulat- lattices, including stable as well as unstable materials. ing as well as metallic materials. For metals, the value The calculations were performed with the same plane of S is ill-de ned and here we have simply used the mag- wave cuto and k-point sampling as used for the mag- netic moment localized on the magnetic atoms, which is netic anisotropy calculations in the database. Examples obtained by integrating the magnetization density over of such structures are shown in Fig. 3 and includes the the PAW spheres. Moreover, the Heisenberg model is transition metal dichalcogenides (TMD) in the 1T phase not a reliable starting point for itinerant magnets and to and in the 2H phase (triangular magnetic lattice), com- our knowledge there is no simple method to obtain crit- pounds adopting the FeSe crystal structure (square mag- ical temperatures for metallic ferromagnetic materials. netic lattice), and transition metal trihalides such as CrI For this reason we will not discuss metallic materials any (honeycomb magnetic lattice). In Fig. 4 we show all the further in the present work, but simply note that large calculated parameters J and  for insulators and metals anisotropies and exchange couplings indicate that metal- with triangular, square or honeycomb lattice. The spin- lic compounds such as CoBr , VBr , NiI , and NiBr 3 3 3 3 wave gap was calculated by taking the ground state to could potentially exhibit very high critical temperatures. M [μ ] z B dE/dTk J [eV] 5 FIG. 5. PBE calculations of exchange coupling J (triangles) and spinwave gap  (squares) of stable ferromagnetic materials with  > 0. Green background indicates insulating materials. Formula Structure J [meV]  [meV] S [~] T [K] main reason for the high value of T is the large magnetic FeCl MoS 15.2 0.056 2.0 208 2 2 moment of 4  per Fe atom and an exchange coupling CuCl BiI 15.3 0.058 1.0 37 3 3 of J = 15 meV, which is one of the largest values found CrI BiI 2.3 0.96 1.5 35 3 3 in the present study. Previous ab initio calculations have CoCl CdI 2.0 0.058 1.5 31 2 2 reported that FeCl in the CdI crystal structure (which 2 2 CrBr BiI 2.0 0.23 1.5 23 3 3 is metallic) is more stable compared to the MoS crys- MnO CdI 0.54 0.31 1.5 19 2 2 31,32 tal structure and the Curie temperature was esti- NiCl CdI 7.2 0.001 1.0 14 2 2 mated to 17 K based on mean- eld theory. Our cal- CrCl BiI 1.4 0.033 1.5 13 3 3 culations con rm the stability hierarchy and predict an RuCl MoS 18.7 2.3 2.0 606 2 2 even more stable prototype GeS (formation energy re- RuBr MoS 16.1 1.77 2.0 509 2 2 duced by  30 meV/atom compared to the MoS phase). However, we do not expect out-of-plane long-range fer- TABLE I. List of 2D magnetic insulating materials with posi- romagnetic order in either the CdI or the GeS crystal 2 2 tive exchange coupling J and positive spinwave gap . Struc- structures, since the spinwave gaps are negative in both ture denotes the prototypical crystal structure and S is the cases. Interestingly, FeCl in the CdI crystal structure 2 2 spin carried by each magnetic atom. The critical tempera- has positive single-ion anisotropy (A), which could in- ture T is obtained from Eq. (5). The top part of the table dicate magnetic order. However, a negative anisotropic contains dynamically and thermodynamically stable materi- exchange coupling (B) yields an overall negative spin- als. The lower part of the table contains materials that are wave gap and the material thus serves as a good example not expected to be stable in their pristine form but exhibit of a case where the single-ion anisotropy is not a good high critical temperatures. indicator of magnetic order. To our knowledge there is no experimental reports of isolated 2D FeCl com- pounds. However, FeCl in the BiI crystal structure has 3 3 In general we observe that most compounds contain tran- been intercalated in bulk graphite exhibiting a ferromag- sition metal atoms with 3d valence electrons. In partic- netic transition at temperature T = 8:5 K. . Recently ular Cr, Mn, Fe, Ni and Co, which are all well-known it has also been employed as functional intercalation in elements in magnetic materials. In addition, most of the few-layer graphene compounds to weaken restacking of compounds contain halides, albeit with a few important 34 35 graphene sheets and bilayer graphene compounds to exceptions (for example MnO ). promote magnetic order in graphene. Nevertheless, ac- In Tab. I we show a list of the all the insulating fer- cording to our calculations, the FeCl crystal structure romagnetic materials and the calculated critical temper- is less stable than the FeCl ones and is not expected to atures. The top part of the table contains the stable exhibit ferromagnetic order as free standing layers due materials and in the lower part we have included a few to a negative value of the spinwave gap. In bulk form examples of materials that exhibit very high critical tem- FeCl is known in the CdI crystal structure with in- 2 2 peratures but which are predicted to be unstable in their plane ferromagnetic order, but the long range order is pristine form. stabilized by interlayer anti-ferromagnetic exchange cou- pling, which supports our assertion that exfoliated layers of this type will not exhibit magnetic order. Bulk FeCl FeCl has also been reported to form di erent stacking poly- mormphs of the BiI crystal structure, but the magnetic The largest Curie temperature is found for FeCl in the properties of these materials are not known. MoS crystal structure where we obtain T = 202 K. The 2 C 6 MnO XCl 2 2 Monolayers of MnO in the CdI crystal structure have Bulk CoCl and NiCl are both known to display anti- 2 2 2 2 been exfoliated in 2003, but the magnetic properties ferromagnetic interlayer coupling and in-plane ferromag- have not yet been thoroughly investigated experimen- netic order . As seen in Tab. I, our calculations predict tally. Our calculations con rm a ferromagnetic ground the materials to exhibit out-of-plane order. For NiCl , in agreement with previous calculations , where a criti- however, one should be a bit cautious due to the ex- cal temperature of 140 K was predicted. However, that tremely small value of the spinwave gap  and more result were obtained from energy mapping analysis using accurate calculations could lead to a ground state with PBE+U DFT calculations (U = 3.9 eV) and MC cal- in-plane magnetic order. Experimental measurements on culations based on the Ising model. From simulations bulk samples indicate anomalies in the heat capacity re- of the Heisenberg model - explicitly including the nite lated to magnetic phase transitions at 24 K and 52 K. anisotropy - we obtain a T ) of 63 K using Heisenberg While the rst result is in good agreement with our pre- parameters from a pure PBE calculations. The e ect of dicted properties, the second one is signi cantly higher Hubbard correction will be discussed in the next section. and could be related to an additional phase transition in the 3D structure. YCl Metastable high-T compounds A critical temperature of T = 55 K is found for YCl C 2 in the MoS prototype. There is neither experimental The lower part of Tab. I shows two materials that or theoretical reports on this material and it could pose we do not predict to be stable, but may be of interest an interesting new 2D magnetic compound. Our calcula- due to the large predicted critical temperatures. Here tions indicate that it is highly stable in the ferromagnetic we comment brie y on the case of RuCl in the MoS 2 2 con guration with magnetic moment of 1  per Y atom. crystal structure, which we predict to be a dynamically However, since the material comprises a spin-1/2 system stable insulator with a critical temperature of 598K. It the classical MC calculations of the critical temperature is, however, situated 0.5 eV above the convex hull, which may not be very accurate. will mostly likely pose an obstacle to experimental syn- thesis. Nevertheless, the calculations show that very high values of critical temperatures are indeed possible CrX in 2D materials with realistic atomic-scale parameters. It should be mentioned that monolayers of RuCl in the As reported in a previous study employing the same BiI crystal structure have been exfoliated and character- 24 40 41 ized experimentally. Moreover, in a recent study the method we predict CrI in the BiI structure to have 3 3 T = 31 K, while the similar compounds and CrCl and critical temperature of monolayer RuCl was calculated C 3 using DFT and MC simulations based on the Heisenberg CrBr have T of 9 K and 19 K respectively. We note 3 C that our calculated critical temperature for CrI is some- model and found T = 14:21 K. However, a Hubbard term is required to open a gap and RuCl in the BiI what lower than the experimental value of 45 K. This is 3 3 42,43 mainly due to the fact that PBE tends to underestimate crystal structure is metallic within PBE, which is why we do not include it in Tab. I. the exchange coupling and can be improved by using a PBE+U scheme as discussed below. CrCl and CrBr 3 3 have not previously been described in their 2D form, but In-plane anisotropy are known as ferromagnetic bulk compounds consisting of layers in the BiI crystal structure with out-of-plane magnetization. The experimental Curie temperatures As mentioned above, materials with the TMDH crys- of bulk CrCl , CrBr , and CrI are 27 K, 47 K, and 70 K 3 3 3 tal structure have been considered as e ective triangular respectively. Our calculated values show the same hierar- magnetic lattices with a single nearest-neighbour cou- chy, but are reduced compared to the bulk values due to pling. However, this model can only be used for a rough the lack of stabilization by interlayer exchange coupling. screening of materials. For example, CrIS exhibits a strong in-plane anisotropy and the axis of magnetiza- tion are ordered (from the hardest axis to the easiest) CuCl as: x, z and y. In Eq. (1), in-plane anisotropy is not considered and we thus extend the model with the full For CuCl in the BiI crystal structure we nd a criti- set of anisotropy parameters A , A , B , and B that 3 3 x y x y cal temperature of 33 K, which is similar to the calculated measures the single-ion anisotropy and anisotropic ex- value of CrI . The material does, however, lie above the change with respect to both x and y directions (relative convex hull by 0.15 eV per atom, which could complicate to the z-direction). These parameters can be found by experimental synthesis and characterization. generalizing the energy mapping analysis Eqs. (2)-(3) to 7 J J A A B B Easy axis T 1 2 x y x y C CrIS 5.71 4.85 0.084 -0.223 0.025 0.033 x,z,y 140 J MnClN 2.66 5.76 0.023 0.044 0.022 0.012 x,y,z 75 CrClO 1.08 0.74 -0.010 0.034 0.004 0.001 y,x,z 15 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 0.1 0.0 TABLE II. Ferromagnetic materials in the TMDH crystal 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 0.3 structure. The rst two columns show nearest neighbour and B 0.2 next-nearest neighbour exchange coupling constants in meV. Columns three to six display anisotropy parameters calculated 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 with respect to the two in-plane directions x and y in meV. 1.25 In the second last column we state the crystallographic direc- 1.00 tions of magnetization listed from the hardest to the easiest 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 axis. The last column shows the critical temperature in K obtained from MC calculations with these parameters. 0 1 2 3 4 U [eV] include di erent in-plane directions. We then run MC calculations using the full set of parameters to nd the FIG. 6. Calculated magnetic parameters of CrI as a func- critical temperatures, including nearest and next-nearest 3 tion of U. J , A and B,  are in units of meV and critical neighbours couplings J and J . We nd three insulat- 1 2 temperatures T are in units of K. ing materials in this crystal structure that shows ferro- magnetic order. The results are shown in Tab. II. In Element Fe Mn Cr Co Ni V Cu particular, CrIS is predicted to have a critical tempera- U [eV] 4.0 3.8 3.5 3.3 6.4 3.1 4.0 ture of 140 K. We also note that we obtain a Curie tem- perature of 15 K for CrClO, which has previously been predicted to have a Curie temperature of 160 K based TABLE III. Hubbard parameters employed in the PBE+U on an Ising model approach. Again, this comparison calculations. emphasizes that the magnetic anisotropy cannot simply be regarded as a mechanism that xes the magnetization to the out-of-plane direction: approximating magnetic material the structure was relaxed with the given value properties by the Ising model may yield a critical tem- of U, but the stability analysis was based on bare PBE. perature that is wrong by an order of magnitude. The inclusion of U can have a rather dramatic e ect on the results. For example, the magnetic con guration of the ground state or the magnetic moment localized on A. Hubbard U the transition metal ion may change. The results for the stable materials are shown in Fig. 7, while the insulating systems are listed in Tab. IV. Including a Hubbard term Almost half of the materials present in C2DB contain in the entire work ow (from the relaxation step onward) at least one element with a partially lled d-shell. Local a ects quantitatively and in some cases also qualitatively and semi-local xc-functionals such as PBE are known to the ground-state. This means that the magnetic moment, overestimate delocalization of correlated electrons, due to the energy gap or the sign of  may change, making the the uncompensated Coulomb self-interaction of the elec- comparison with Tab. I meaningful only for a subset of tron. In the Hubbard model a term is introduced that materials. Compounds that are also present in Tab. I acts as an e ective electronic on-site repulsion and pro- are shown in bold face for comparison. vides a penalty to delocalization. In order to determine the in uence of the Hubbard correction we have recalcu- lated exchange and anisotropy parameters for CrI for a E ect of U on T range of U values in the PBE+U scheme. The structure was fully relaxed for each value of U and the results are shown in Fig. 6. We observe that an increasing value of For MnO , the main e ect of adding a Hubbard correc- U leads to an overall increase of both  and J , which tion is to increase the exchange parameter J by a factor result in higher critical temperatures. The dependence of two. Interestingly the anisotropy parameters A and B of T on U is roughly linear with T increasing by 5 K decrease and the spinwave gap  becomes less than half C C per eV that U is increased. the value obtained with PBE. Nevertheless, the overall In order to gain more insight into the general in u- e ect is an increase of the critical temperature from 63 ence of U for the calculations of magnetic properties, we K to 82 K. This number can be compared to the result have recalculated the magnetic parameters and critical in Ref. 39 where a critical temperature of 140 K was temperature for all magnetic materials in the C2DB con- estimated from the Ising model. taining 3d valence electrons. We used the optimal values For CrI we predict a critical temperature of 50 K with determined in Ref. 45 and listed in Tab. III. For each PBE+U. This is signi cantly closer to the experimental 8 FIG. 7. PBE+U calculations of exchange coupling J (triangles) and spinwave gap  (squares) of stable ferromagnetic materials with  > 0. Green background indicates insulating materials. 37,47 Formula Structure J [meV]  [meV] S [~] T [K] moments being aligned in the individual planes. This MnO CdI 6.43 0.125 1.5 82 could indicate that PBE+U provides a more accurate 2 2 CoCl CdI 3.21 0.249 1.5 57 2 2 description than PBE, which does not predict magnetic CrI BiI 3.95 1.280 1.5 50 3 3 order for MnI . For MnBr in the CdI crystal struc- 2 2 2 CrBr BiI 2.82 0.185 1.5 24 3 3 ture, neutron scattering experiments on the bulk par- MnI CdI 0.40 0.081 2.5 21 2 2 ent structure revealed an anti-ferromagnetic order below MnBr CdI 0.41 0.024 2.5 16 2 2 T = 2:16 K with magnetic moments lying in-plane. CrCl BiI 2.19 0.016 1.5 10 3 3 4 However, this is not necessarily in contradiction with our NiCl CdI 5.69  10 1.0 7 2 2 calculations since the observed anti-ferromagnetic con g- FeBr BiI 0.04 0.124 2.5 2 3 3 uration is "double-striped", a con guration that has not CoO FeSe 106.54 0.199 1.5 520 been considered in the present study. For FeBr the Hub- FeS FeSe 28.99 0.591 2.0 413 3 bard term makes the spin jump from ~=2 to 5~=2per Fe atom and opens a spinwave gap. A previous investigation TABLE IV. List of 2D non-metal materials with positive of this material showed that it is predicted to be a quan- exchange coupling J and spin wave gap , obtained from tum spin Hall insulator with PBE while PBE+U predicts PBE+U calculations. Structure denotes the prototypical a trivial insulator above a critical value of U= 0:18 eV. crystal structure and S is the spin carried by each magnetic atom. The structures in bold are present also in I for com- parison. The top part contains stable materials, whereas the lower part contains materials with large critical temperatures that may be unstable in their pristine form. Metastable high-T compounds The lower part of Tab. IV lists materials, which are not value than the 31 K obtained with bare PBE. Similarly predicted to be completely stable in their pristine form T increases from 19 K to 24 K for CrBr while the T C 3 C according to PBE calculations (we have not performed of CrCl is increased from 9 K to 10 K. The critical tem- a full stability analysis with PBE+U). Bulk CoO has an perature of CoCl is almost una ected. These results anti-ferromagnetic rock-salt structure with a critical tem- indicate that it is non-trivial to predict how the inclu- perature of 293 K . According to PBE calculations the sion of a Hubbard U in uences the calculated critical most stable 2D phase is a metallic CdI crystal structure temperatures in general. (parameters J and  are shown in Fig. 5). In the FeSe For the compounds MnI , MnBr , and FeBr , which 2 2 3 crystal structure, CoO has a low dynamic stability but all have large magnetic moment of 5 per magnetic we report it here due to the very high critical tempera- atom we obtain rather low critical temperatures of 21, ture of 520 K originating from the extraordinarily large 16 and 2 K respectively. This is mainly due to small exchange coupling predicted by PBE+U. values of exchange coupling J for these materials. The inclusion of U in MnI and MnBr increases the elec- FeS in the FeSe crystal structure has a non-magnetic 2 2 tronic gap as well as the spinwave gap. But most impor- ground state with PBE, but is predicted to be highly tantly, it yields a ferromagnetic ground state, while the stable and is situated on the convex hull. With PBE+U ground state is anti-ferromagnetic without the inclusion the ground state becomes ferromagnetic and we predict a of U. For MnI The result appears to be in qualita- high critical temperature of 413 K. According to previous tive agreement with neutron scattering experiments on calculations, however, the true ground state is a striped the bulk compounds, which reports a helical magnetic anti-ferromagnetic con guration, which is not taken into structure below a critical temperature of 3.4 K, with the account in this work. 9 J J A A B B Easy axis T 1 2 x y x y C correlated systems. We have thus tested how the re- CrBrO 1.12 0.78 0.043 -0.010 0.001 0.001 x,z,y 35 sults are modi ed if the parameters are evaluated with CrIO 0.49 -1.46 0.586 -0.123 -0.008 -0.003 x,z,y 25 PBE+U instead and we nd that the predictions do in- CrClO 1.38 1.27 0.007 0.016 0.001  10 y,x,z 20 deed change in a non-systematic way. For the hexagonal and honeycomb systems three materials that were pre- dicted to be ferromagnetic (at nite temperature) are TABLE V. Parameters and results for TMHC structures ob- no longer predicted to show magnetic order when the tained from PBE+U calculations. Symbols and units are the PBE+U scheme is employed and three materials that same as in Tab. II. Materials in bold are the ones listed in were not magnetic with PBE become magnetic with both tables. PBE+U. For the ve materials that are magnetic with both PBE and PBE+U the critical temperatures are slightly di erent in the two approximations. The biggest In-plane anisotropy di erence is seen for CrI where inclusion of U increases the critical temperature from 31 K to 50 K, which is In Tab. V we list Heisenberg parameters and crit- closer to the experimental value of 45 K. ical temperatures for TMHC structures obtained from In the present work we have mainly focused on insu- PBE+U calculations and MC calculations following the lators. This restriction is rooted in the simple fact that same procedure as in the previous section where no Hub- we do not have a reliable way to estimate Curie temper- bard correction was included. Comparing the results atures of metallic 2D magnetic materials. Metallic ferro- with Tab. II, it is noted that MnClN and CrIS are not magnetism in 2D is, however, a highly interesting subject predicted to be ferromagnetic insulators with PBE+U. In and we note that room-temperature magnetism has re- particular, MnClN is predicted to be a metal and CrIS 5 4 cently been reported in the 2D metals VSe and MnSe . 2 2 exhibits a negative spinwave gap. On the other hand, two Moreover, Figs. 4 and 5 indicate that in the C2DB the new materials - CrIO and CrBrO - are predicted to ex- largest values of both spinwave gaps and exchange cou- hibit ferromagnetic order at 25 K and 35 K respectively. plings are found in metallic materials. Clearly, there is pressing need for theoretical developments of 2D itinerant magnetism that can be applied in conjunction with rst principles simulations to provide accurate predictions of IV. CONCLUSIONS the magnetic properties of 2D metallic materials. Finally, we have restricted ourselves to ferromag- We have presented a high throughput computational netic order. Nevertheless, the C2DB contains 241 anti- screening for magnetic insulators based on the Compu- ferromagnetic entries - 50 of which are predicted to be tational 2D Materials Database. In contrast to several stable. The prediction of a novel 2D anti-ferromagnetic previous studies of magnetism in 2D, we have empha- compound would certainly comprise an important step sized the crucial role of magnetic anisotropy and used the forward in the study of 2D magnetism, but the theoretical spinwave gap as a basic descriptor that must necessarily treatment is complicated by the possibility of several or- be positive in order for magnetism to persist at nite dered structures that may coexist at a given temperature temperatures. This criterion severely reduces the num- - in particular for non-bipartite lattice such as the trian- ber of relevant candidates and we end up with 12 stable gular one. We will leave the study of anti-ferromagnetism candidate materials for which the critical temperatures in 2D to future work. were calculated from classical MC simulations. Seven of the materials were predicted to have Curie temperatures V. ACKNOWLEDGEMENTS exceeding that of CrI . The classical MC simulations appear to comprise an accurate method for obtaining the critical temperatures T. Olsen and D. Torelli acknowledges supported from for insulating materials with S > 1=2. However, the the Independent Research Fund Denmark, grant number Heisenberg parameters that enter the simulations may be 6108-00464A. K. Thygesen acknowledges support from sensitive to the approximations used to calculate them. the European Research Council (ERC) under the Euro- In the C2DB all calculations are performed with the PBE pean Union's Horizon 2020 research and innovation pro- functional, which may have shortcomings for strongly gram (Grant No. 773122, LIMA). N. D. Mermin and H. Wagner, Phys. Rev. Lett. 17, 1133 Herrero, and X. Xu, Nature 546 (2017), 10.1038/na- (1966). ture22391. 2 3 B. Huang, G. Clark, E. Navarro-Moratalla, D. Klein, C. Gong, L. Li, Z. Li, H. Ji, A. Stern, Y. Xia, T. Cao, R. Cheng, K. Seyler, D. Zhong, E. Schmidgall, W. Bao, C. Wang, Y. Wang, Z. Q. Qiu, R. J. Cava, S. G. M. A. McGuire, D. Cobden, W. Yao, X. di, P. Jarillo- Louie, J. Xia, and X. Zhang, Nature 546, 265 (2017). 10 D. J. O'Hara, T. Zhu, A. H. Trout, A. S. Ahmed, Y. K. G. K. H. Madsen, R. M. Nieminen, J. K. Nrskov, Luo, C. H. Lee, M. R. Brenner, S. Rajan, J. A. Gupta, M. Puska, T. T. Rantala, J. Schitz, K. S. Thygesen, and D. W. McComb, and R. K. Kawakami, Nano Lett. 18, K. W. Jacobsen, J. Phys. Condens. Matter 22, 253202 3125 (2018). (2010). 5 26 M. Bonilla, S. Kolekar, Y. Ma, H. C. Diaz, V. Kalappattil, A. Hjorth Larsen, J. Jrgen Mortensen, J. Blomqvist, R. Das, T. Eggers, H. R. Gutierrez, M.-H. Phan, and I. E. Castelli, R. Christensen, M. Du lak, J. Friis, M. Batzill, Nat. Nanotechnol. 13, 289 (2018). M. N. Groves, B. Hammer, C. Hargus, E. D. Hermes, J. M. Kosterlitz and D. J. Thouless, J. Phys. C Solid State P. C. Jennings, P. Bjerre Jensen, J. Kermode, J. R. Phys. 6, 1181 (1973). Kitchin, E. Leonhard Kolsbjerg, J. Kubal, K. Kaasb- P. C. W. Holdsworth and S. T. Bramwell, Phys. Rev. B jerg, S. Lysgaard, J. Bergmann Maronsson, T. Max- 49, 8811 (1994). son, T. Olsen, L. Pastewka, A. Peterson, C. Rostgaard, Z. Fei, B. Huang, P. Malinowski, W. Wang, T. Song, J. Schitz, O. Schutt,  M. Strange, K. S. Thygesen, J. Sanchez, W. Yao, D. Xiao, X. Zhu, A. F. May, W. Wu, T. Vegge, L. Vilhelmsen, M. Walter, Z. Zeng, and K. W. D. H. Cobden, J.-H. Chu, and X. Xu, Nat. Mater. 17, 778 Jacobsen, J. Phys. Condens. Matter 29, 273002 (2017). (2018). J. E. Saal, S. Kirklin, M. Aykol, B. Meredig, and C. Gong and X. Zhang, Science 363, eaav4450 (2019). C. Wolverton, JOM 65, 1501 (2013). 10 28 J. Liu, M. Shi, J. Lu, and M. P. Anantram, Phys. Rev. B T. Olsen, Phys. Rev. B 94, 235106 (2016). 97, 8 (2018). C. Xu, J. Feng, H. Xiang, and L. Bellaiche, npj Compu- S. Jiang, J. Shan, and K. F. Mak, Nat. Mater. 17, 406 tational Materials 4, 57 (2018). (2018). T. Olsen, Phys. Rev. B 96, 125143 (2017). 12 31 B. Huang, G. Clark, D. R. Klein, D. MacNeill, E. Navarro- E. Torun, H. Sahin, S. K. Singh, and F. M. Moratalla, K. L. Seyler, N. Wilson, M. A. McGuire, D. H. Peeters, Applied Physics Letters 106, 192404 (2015), Cobden, D. Xiao, W. Yao, P. Jarillo-Herrero, and X. Xu, https://doi.org/10.1063/1.4921096. Nat. Nanotechnol. 13, 544 (2018). Y. Feng, X. Wu, J. Han, and G. Gao, Journal of Materials C. Cardoso, D. Soriano, N. A. Garc a-Mart nez, and Chemistry C 6 (2018), 10.1039/C8TC00443A. J. Fern andez-Rossier, Phys. Rev. Lett. 121, 67701 (2018). M. S. Dresselhaus and G. Dresselhaus, Advances in Physics T. Taniguchi, K.-M. C. Fu, R. Cheng, X. Linpeng, X. Xu, 51, 1 (2002), https://doi.org/10.1080/00018730110113644. D. Zhong, K. L. Seyler, D. Xiao, B. Huang, M. A. McGuire, X. Qi, J. Qu, H.-B. Zhang, D. Yang, Y. Yu, C. Chi, and K. Watanabe, W. Yao, N. Sivadas, and E. Schmidgall, Sci. Z.-Z. Yu, J. Mater. Chem. A 3, 15498 (2015). Adv. 3, e1603113 (2017). N. Kim, K. S. Kim, N. Jung, L. Brus, and K. Burch, D. Mandrus, and J.-G. Park, Nature 563 P. Kim, Nano Lett. 11, 860 (2011), pMID: 21268591, (2018), 10.1038/s41586-018-0631-z. https://doi.org/10.1021/nl104228f. 16 36 Z. Wang, I. Guti errez-Lezama, N. Ubrig, M. Kroner, T. H. Bointon, I. Khrapach, R. Yakimova, A. V. Shytov, M. Gibertini, T. Taniguchi, K. Watanabe, A. Imamo glu, M. F. Craciun, and S. Russo, Nano Lett. 14, 1751 (2014). E. Giannini, and A. F. Morpurgo, Nat. Comm. 9, 2516 M. A. McGuire, Crystals 7 (2017), 10.3390/cryst7050121. (2018). Y. Omomo, T. Sasaki, Wang, and M. Watanabe, J. Am. T. Bj orkman, A. Gulans, A. V. Krasheninnikov, and R. M. Chem. Soc. 125, 3568 (2003). Nieminen, Phys. Rev. Lett. 108, 235502 (2012). M. Kan, J. Zhou, Q. Sun, Y. Kawazoe, and P. Jena, J. M. Ashton, J. Paul, S. B. Sinnott, and R. G. Hennig, Chem. Phys. Lett. 4, 3382 (2013). Phys. Rev. Lett. 118, 106101 (2017). D. Weber, L. M. Schoop, V. Duppel, J. M. Lippmann, N. Mounet, M. Gibertini, P. Schwaller, D. Campi, J. Nuss, and B. V. Lotsch, Nano Lett. 16, 3578 (2016). A. Merkys, A. Marrazzo, T. Sohier, I. E. Castelli, A. Ce- S. Sarikurt, Y. Kadioglu, F. Ersan, E. Vatansever, O. U. pellotti, G. Pizzi, and N. Marzari, Nat. Nanotechnol. 13, Akturk, Y. Yuksel, U. Aknc, and E. Akturk, Phys. Chem. 246 (2018). Chem. Phys. 20, 997 (2018). 20 42 S. Haastrup, M. Strange, M. Pandey, T. Deilmann, P. S. C. Huang, J. Zhou, H. Wu, K. Deng, P. Jena, and E. Kan, Schmidt, N. F. Hinsche, M. N. Gjerding, D. Torelli, P. M. Phys. Rev. B 95, 045113 (2017). Larsen, A. C. Riis-Jensen, J. Gath, K. W. Jacobsen, J. J. F. Ersan, E. Vatansever, S. Sarkurt, Y. Yuksel, Y. Ka- Mortensen, T. Olsen, and K. S. Thygesen, 2D Materials dioglu, H. Ozaydin, O. uzengi akturk, U. Aknc, and 5, 042002 (2018). A. Ethem, Journal of Magnetism and Magnetic Materials T. Olsen, E. Andersen, T. Okugawa, D. Torelli, T. Deil- 476 (2018), 10.1016/j.jmmm.2018.12.032. mann, and K. S. Thygesen, Phys. Rev. Mater. 3, 024005 N. Miao, B. Xu, L. Zhu, J. Zhou, and Z. Sun, J. Am. (2019), 1812.06666. Chem. Soc. 140, 2417 (2018). 22 45 K. Malarz, M. Zborek, and B. Wrobel, Task Q. 9, 475 L. Wang, T. Maxisch, and G. Ceder, Phys. Rev. B 73, (2005). 195107 (2006). 23 46 M. Kan, S. Adhikari, and Q. Sun, Phys. Chem. Chem. V. V. Kulish and W. Huang, J. Mater. Chem. C 5, 8734 Phys. 16, 4990 (2014). (2017). 24 47 D. Torelli and T. Olsen, 2D Materials 6, 015028 (2018). J. W. Cable, M. K. Wilkinson, E. O. Wollan, and W. C. J. Enkovaara, C. Rostgaard, J. J. Mortensen, J. Chen, Koehler, Phys. Rev. 125, 1860 (1962). M. Du lak, L. Ferrighi, J. Gavnholt, C. Glinsvad, E. O. Wollan, W. C. Koehler, and M. K. Wilkinson, Phys. V. Haikola, H. a. Hansen, H. H. Kristo ersen, M. Kuisma, Rev. 110, 638 (1958). a. H. Larsen, L. Lehtovaara, M. Ljungberg, O. Lopez- A. Berkowitz and K. Takano, Journal of Magnetism and Acevedo, P. G. Moses, J. Ojanen, T. Olsen, V. Petzold, Magnetic Materials 200, 552 (1999). N. a. Romero, J. Stausholm-Mller, M. Strange, G. a. Trit- M. Winiarski, J. Zasada, and M. Samsel-Czeka la, Com- saris, M. Vanin, M. Walter, B. Hammer, H. H akkinen, putational Materials Science 142, 372 (2018).

Journal

Condensed MatterarXiv (Cornell University)

Published: Mar 27, 2019

There are no references for this article.