Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2021 Jun 21;13(25):29770–29779. doi: 10.1021/acsami.1c00403

Stoner Ferromagnetism in Hole-Doped CuMIIIAO2 with MIIIA = Al, Ga, and In

Konstantina Iordanidou 1,*, Clas Persson 1
PMCID: PMC8289245  PMID: 34152119

Abstract

graphic file with name am1c00403_0010.jpg

Using density functional theory calculations, we examine the effect of hole doping on the magnetic and electronic properties of CuMIIIAO2, with MIIIA = Al, Ga, and In. CuMIIIAO2 nonmagnetic semiconductors switch to ferromagnetic half-metals upon hole doping. For CuAlO2, the nonmagnetic-to-ferromagnetic transition occurs for hole densities of ∼7 × 1019/cm3. Ferromagnetism arises from an exchange splitting of the electronic states at the valence band edge, and it can be attributed to the high-lying Cu-d states. Hole doping induced by cation vacancies and substitutional divalent dopants is also investigated. Interestingly, both vacancies and nonmagnetic divalent dopants result in the emergence of ferromagnetism.

Keywords: CuAlO2, hole doping, defects, ferromagnetism, first-principles

1. Introduction

Ferromagnets such as iron, cobalt, and nickel, are typically metals, whereas materials with both ferromagnetic and semiconducting properties are scarce. Ferromagnetic semiconductors are attractive candidates for future spintronic devices, owing to their easy integration into the existing semiconductor devices.1 Over the past few decades, diluted magnetic semiconductors (DMS), i.e., nonmagnetic semiconductors doped with magnetic ions, have triggered intense interest.2−5 Magnetic moments mainly arise from partially filled 3d and 4f orbitals of transition metals and rare earth metals, and examples of DMS include (Ga,Mn)As, (Zn,Cr)Se, and (Pb,Eu)Te. Although Ga1–xMnxAs has been one of the most widely studied DMS, its low Curie temperature hinders its utilization in practical applications. Contrary to GaAs-based DMS, room-temperature ferromagnetism has been observed in diluted magnetic oxides such as TiO2, ZnO, and SnO.6−10 Intriguingly, high-temperature ferromagnetism in the absence of magnetic ions has been additionally reported.11−21 For instance, Kenmochi et al. found that carbon-doped oxides such as MgO, SrO, BaO, and CaO are ferromagnetic without any transition-metal impurities, and based on the mean-field approximation, Curie temperatures higher than room temperature have been predicted.19−21

Two types of diluted magnetic semiconductors with different properties have been widely studied. First, for DMS like Mn-doped GaSb with localized majority d-states deep in the valence band, ferromagnetism arises from the Zener’s p–d exchange mechanism, resulting in holes in the majority p-valence band. Such interaction is relatively weak and long-range, and the Curie temperature scales linearly with the impurity concentration. On the other hand, for DMS like Mn-doped GaN with impurity bands within the band gap, ferromagnetism arises from the Zener’s double-exchange mechanism. The magnetic couple is strong and short-ranged, and the Curie temperature scales proportionally to the square root of the impurity concentration.22,23

Doping of CuAlO2 with magnetic ions such as Mn, Fe, and Co, to form oxide-based DMS, has been intensively investigated.24−27 For instance, first-principles calculations revealed that (Cu,Fe)AlO2 and (Cu,Co)AlO2 are attractive candidates as high-Tc ferromagnets. Overall, CuAlO2 is an attractive candidate for various applications owing to its p-type conductivity (without intentional doping) and its visible light transparency.28−30 In general, the design of p-type transparent conductive oxides (TCO) is found to be difficult, and the difficulty arises from the nature of the VB edge, dominated by localized O-2p states.31,32 Holes are typically trapped at O sites and cannot migrate through the oxide, leading to poor p-type behavior. To overcome this limitation, modulation of the VB edge through hybridization of the O-2p orbitals with metal orbitals has been proposed.31 The discovery of p-type conductivity in CuAlO2 has paved the way to the discovery of other p-type TCO like CuGaO2 and CuInO2.

In this paper, using density functional theory (DFT) calculations, we examine the effect of hole doping on the magnetic properties of CuMIIIAO2 with MIIIA = Al, Ga, and In, having CuAlO2 as a prototype material. Despite that these materials are intrinsically nonmagnetic, stable ferromagnetic phases appear for a wide range of hole densities. In addition the possibility of inducing hole doping and also ferromagnetism by cation vacancies and substitutional divalent dopants is investigated. Contrary to previous studies, we focus on doping with nonmagnetic ions. Magnetic ions typically lead to localized magnetic moments instead of ferromagnetic phases, and materials with localized magnetic moments are likely to present low spin polarization of the charge carriers, hindering their utilization in future spintronic devices. In our work, we find that nonmagnetic divalent dopants result in the emergence of ferromagnetism, and ferromagnetism can be attributed to the high-lying Cu-d states, dominating over the O-p states at the VB edge.

2. Methods and Computational Details

We perform spin-polarized DFT calculations using the Vienna Ab initio Simulation Package (VASP).33,34 For the exchange correlation functional, the generalized gradient approximation introduced by Perdew, Burke, and Ernzerhof (PBE) is used.35 The electron–ion interaction is described by the projector augmented wave (PAW) method.36 For the unit cells, both the atomic positions and the volume are optimized, through the conjugate gradient method, with 10–8 eV energy convergence criteria and 10–2 eV/Å force convergence criteria. The kinetic energy cutoff is set to 600 eV, and the Brillouin zone is sampled by a 16 × 16 × 16 k-point grid for the atomic relaxations, whereas a 24 × 24 × 24 k-point grid is used for the density of states (DOS) calculations.

To achieve hole doping, we change the number of electrons of the unit cell and we add a compensating opposite charge background. Since the magnetic instabilities sensitively depend on the DOS near the Fermi level, a very dense grid of 48 × 48 × 48 k-points was used for the calculations of the magnetic moments. Hybrid functional calculations were also performed, using the Heyd–Scuseria–Ernzerhof (HSE) functional and the standard range-separation and mixing (0.25 Fock exchange) parameters.37 Due to computational issues, for the hybrid functional calculations, an 8 × 8 × 8 k-point grid was adopted.

Finally, for the defective systems, we used supercells consisting of 192 atoms in total. The Brillouin zone was sampled by a 2 × 2 × 2 k-point grid for the atomic relaxations, whereas a 4 × 4 × 4 k-point grid was used for the DOS calculations.

3. Results

3.1. Ferromagnetism in Hole-Doped CuMIIIAO2

CuAlO2 crystallizes in the delafossite structure (space group: R3̅m), where each Cu atom is bonded to two O atoms forming O–Cu–O dumbbells, whereas each Al atom is bonded to six O atoms forming distorted octahedra. The Cu–O and Al–O bond lengths are found to be 1.885 and 1.925 Å, respectively, whereas for the optimized nonprimitive unit cell, the in-plane and out-of-plane lattice constants are a = b = 2.880 and c = 17.119 Å, in excellent agreement with experimental observations.38 A summary of the structural parameters of CuMIIIAO2 with MIIIA = Al, Ga, and In, in the delafossite structures is given in Table S1 in the Supporting Information. As shown in Figure 1, CuAlO2 is an indirect band gap semiconductor with Eg = 1.79 eV, in agreement with previously reported DFT-PBE calculations.39 Using HSE calculations, a gap opening of about 1.8 eV is found. The valence band (VB) top and the conduction band (CB) bottom are located at the F and Γ points, respectively, and the VB top is mainly derived from Cu-d orbitals along with a smaller contribution of O-p orbitals, as observed in Figure 2. Notably, the dispersion of the topmost VB looks like a slightly asymmetric inverted Mexican hat, where the band extrema are located at the F and L points, and the energy difference between the F and L points is 85 meV. This peculiar VB dispersion leads to a sharp high peak of the DOS near the VB edge. By shifting the Fermi level close to this peak, the Stoner criterion is fulfilled and the system becomes ferromagnetic.40−47 Hence, hole-doped CuAlO2 could be a ferromagnetic material.

Figure 1.

Figure 1

Open in a new tab

Two-dimensional band structure of CuAlO2 unit cell, and three-dimensional band structure for the topmost valence band and the bottommost conduction band. The three-dimensional band structure is obtained using the VASPKIT code.56 The energies refer to the valence band maximum. The high-symmetry k-points are Γ (0, 0, 0), F (0.5, 0.5, 0), L (0.5, 0, 0), and Z (0.5, 0.5, 0.5). The topmost VB is highlighted in red.

Figure 2.

Figure 2

Open in a new tab

Total and projected valence band DOS of CuMIIIAO2 with MIIIA = Al, Ga, and In (first-row panels), CuAlY2 with Y = S, Se, and Te (second-row panels), AgMIIIAO2 (third-row panels), and AuAlO2 (fourth-row panel). A Lorentzian broadening of 0.02 eV is used. The energies refer to the valence band maximum. The nonparticipating MIIIA states are not displayed.

To verify this assumption, we consider charged systems. Upon hole doping, a very slight shortening of the Cu–O bond lengths is observed, whereas the Al–O bond lengths present the opposite trend (see Supporting Information Figure S1). The magnetic moments and the spin polarization energies as a function of the injected holes are computed. The spin polarization energies are obtained through the equation ESP = ENM – EFM, where ENM and EFM are the energies of the nonmagnetic and ferromagnetic phases, respectively. Using this definition, positive spin polarization energies indicate stable ferromagnetic phases. As shown in Figure 3, a ferromagnetic state spontaneously emerges for hole densities of ∼7 × 1019/cm3. The magnetic moment per carrier rapidly reaches the saturation value of 1 μB, and it remains unchanged for hole densities up to ∼1 × 1022/cm3. Further increase of the hole density leads to the reduction of the magnetic moment. On the other hand, the spin polarization energy per carrier strongly depends on the doping level. At the relatively low doping level, it increases by increasing the density. For p = 6 × 1021/cm3, it reaches the maximum value of ∼46 meV/carrier, which is larger compared to the corresponding values in similar investigations,44,45 whereas for larger hole densities, it gradually decreases until it becomes zero. Contrary to the magnetic moment per carrier, no plateau region is observed in the spin polarization energy per carrier.

Figure 3.

Figure 3

Open in a new tab

Carrier-dependent magnetic moment and spin polarization energy for CuAlO2.

Owing to the self-interaction error, semilocal functionals tend to overdelocalize electrons, and for systems with d electrons, this can be particularly problematic. Hartree–Fock and semilocal functionals typically present an opposite self-interaction error. Thus, hybrid functionals, which use a fraction of the Hartree–Fock and a fraction of the semilocal functional, tend to reduce the self-interaction error.48 In our work, hybrid functional calculations have been additionally performed for the hole-doped systems. Notably, using HSE calculations, the peculiar Mexican-hat-like valence band dispersion is retained and the energy difference between the F and L points is 91 meV, i.e., slightly larger compared to the PBE calculations. The magnetic properties of CuAlO2 at different hole densities in the range of ∼1019–1022/cm3 are examined, and similar to our PBE calculations, hole-doped CuAlO2 is found to be ferromagnetic. It is worth noting that the so-called pseudopotential self-interaction method (pseudo-SIC) has been widely used to eliminate self-interaction errors in standard DFT calculations.49−51 For instance, the self-interaction-corrected electronic structures of ZnO-based DMS have been reported. Compared to those computed using standard DFT, a larger band gap and a qualitatively different description of the transition-metal d states have been observed.

As a next step, we compute the electronic band structures and the spin density plots of the charged systems. The spin density plots are obtained by the spin-up and spin-down charge density differences. As shown in Figure 4, an exchange splitting of the electronic states near the Fermi level is observed at p = 5.0 × 1020/cm3. The energy difference between the spin-up and spin-down valence band maxima significantly increases, by increasing the hole density. On the other hand, we observe a shift of the Fermi level toward the lower energies. This results in a half-metallic behavior, i.e., the spin-up and spin-down states present semiconducting and metallic nature, respectively, allowing a fully polarized spin transport. The magnetization mainly arises from Cu-d orbitals along with O-p orbitals. For instance, for p = 1 × 1021/cm3, the magnetization of Cu, Al, and O atoms is 86, 0, and 14% of the total magnetization, and very similar results are found for other hole densities. These results are consistent with the observation that the magnetic behavior is originated from the VB edge, which mainly consists of Cu-d states with a smaller contribution of O-p states.

Figure 4.

Figure 4

Open in a new tab

Electronic band structures (left) and spin density plots (right) of CuAlO2 unit cells at different hole densities: (a, b) 5.0 × 1020/cm3, (c, d) 1.0 × 1021/cm3, and (e, f) 5.0 × 1021/cm3. The blue and black lines correspond to spin-up-like and spin-down-like states. The high-symmetry k-points are Γ (0, 0, 0), F (0.5, 0.5, 0), L (0.5, 0, 0), and Z (0.5, 0.5, 0.5). The gray, light blue, and red spheres correspond to Cu, Al, and O atoms. The isosurface is set to 0.0005 electrons per Bohr3.

Finally, we perform calculations to estimate the Curie temperature of CuAlO2 at various hole densities. Our calculations are based on the mean-field evaluation of the magnetic moment by minimizing the electronic free energy of the system at specific temperatures.52 To introduce the finite temperature, we modify the smearing factor (σ = kBT) of the Fermi–Dirac distribution. At the relatively low doping level, the Curie temperature is below room temperature. By increasing the hole density, the Curie temperature increases, and at the high doping level, it becomes much larger than room temperature (see Supporting Information Figure S2). It should be noted that the Curie temperatures can be significantly overestimated using this approach, and the real values can be lower compared to the predicted values by as much as a factor of 2.53

Overall, our findings revealed that CuAlO2 could be used for future spintronic devices like the spin field effect transistors (spin FETs) and the magnetic tunnel junctions (MTJs). We found that the injection of holes in CuAlO2 could control/manipulate its electronic and magnetic properties. Control of magnetism through an applied voltage offers various advantages like ultralow power dissipation and reversibility, and electrically controlled ferromagnetism is highly promising for achieving optimal spintronic devices.

Besides CuAlO2, other group-IIIA Cu-based oxides are examined, namely, CuGaO2 and CuInO2. Both CuGaO2 and CuInO2 present an asymmetric Mexican-hat-like VB dispersion, and the energy difference between the band extrema located at the F and L points are 149 and 83 meV, respectively (see Supporting Information Figures S3 and S4). Upon hole doping, a slight shortening of the Cu–O bond lengths along with a slight elongation of the Ga–O and In–O bond lengths is observed (see Supporting Information Figure S5). A nonmagnetic-to-ferromagnetic transition occurs at p = 1 × 1020/cm3 and p = 3 × 1019/cm3 for Ga- and In-based compounds, respectively, and half-metallicity has been additionally observed (see Supporting Information Figures S6–S8). Both CuGaO2 and CuInO2 return to the nonmagnetic state for a hole density of ∼4 × 1022/cm3.

Furthermore, we examine if the corresponding chalcogenides CuAlY2, where Y = S, Se, and Te, in the delafossite structures, are also promising for realizing a ferromagnetic phase upon hole doping. First, we introduce a measure for the hybridization strength between the projected orbital l of atom α and the projected orbital l′ of atom α′, using the normalized cross-correlation of DOS g(E)

3.1. 1

In the entire VB region, the hybridization strengths between Cu-d and Y-p orbitals are found to be 0.50, 0.50, and 0.41 for CuAlS2, CuAlSe2, and CuAlTe2, respectively, whereas the corresponding value for CuAlO2 is only 0.21. In addition, contrary to the oxides where the VB edge is dominated by Cu-d states, for the sulfides and selenides, a strong hybridization of Cu-d and Y-p states is observed, whereas for tellurides, the VB edge is dominated by Te-p states, as shown in Figure 2. Thus, going from oxygen to tellurium compounds, the Cu-d character of the VB edge is found to decrease. The different nature of the VB edge in chalcogenides compared to the oxides leads to the absence of the peculiar Mexican-hat-like VB dispersion and the subsequent absence of the sharp high peak of the DOS at the VB edge. Thus, CuAlY2, with Y = S, Se, and Te is expected to be nonmagnetic upon hole doping. To verify this assumption, additional calculations were performed and we found that hole doping in the range of 1 × 1019–1 × 1022/cm3 results in no magnetization in CuAlY2.

AgMIIIAO2, where MIIIA = Al, Ga, and In, as well as AuAlO2 are also discussed. The hybridization strengths between Ag-d and O-p orbitals, in the entire VB region, are 0.83, 0.79, and 0.78 for MIIIA = Al, Ga, and In, respectively, whereas the corresponding values for their Cu-based counterparts are only 0.21, 0.22, and 0.20. In addition, going from Cu- to Ag-based compounds, the metal-d character of the VB edge significantly decreases, leading to a less sharp peak of the DOS at the VB edge, as shown in Figure 2. Accordingly, for Ag-based compounds, the asymmetry in the Mexican-hat-like VB dispersion significantly increases and the energy differences between the band extremum located at the F and L points are 258, 428, and 340 meV for AgAlO2, AgGaO2, and AgInO2, respectively. This should lead to a much less promising behavior of hole-doped AgMIIIAO2 compounds compared to CuMIIIAO2. Indeed, our calculations reveal that for AgMIIIAO2, no ferromagnetism is observed for hole densities up to ∼1 × 1022/cm3.

Interestingly, AuAlO2 presents an intermediate behavior between CuAlO2 and AgAlO2. The hybridization strength between Au-d and O-p orbitals in the entire VB region is 0.76, whereas the VB edge has a weaker (stronger) d-character compared to CuAlO2 (AgAlO2), as shown in Figure 2. Furthermore, for AuAlO2, the asymmetry in the Mexican-hat VB is larger (smaller) compared to CuAlO2 (AgAlO2) and the energy difference between the F and L points is 204 meV. We find that AuAlO2 becomes ferromagnetic for only a narrow range of hole densities, i.e., ∼5 × 1021 < p < 9 × 1021/cm3, and for the optimal hole density, the magnetic moment is less than 1 μB/carrier. A summary of the key electronic- and magnetic-related characteristics for all studied systems is shown in Table 1. A measure of the d-state strength in the VB edge is introduced using the formula

3.1. 2

and for this measure, we use an integration region of 0.3 eV below the VB top.

Table 1. Electronic- and Magnetic-Related Properties for All Studied Systemsa.

system ΔEFL (meV) γ Δd pNM-FM (cm–3)
CuAlO2 85 0.21 0.81 7 × 1019
CuGaO2 149 0.22 0.84 1 × 1020
CuInO2 83 0.20 0.85 3 × 1019
AgAlO2 258 0.83 0.50  
AgGaO2 428 0.79 0.49  
AgInO2 340 0.78 0.52  
AuAlO2 204 0.76 0.68 5 × 1021
CuAlS2   0.50 0.42  
CuAlSe2   0.50 0.32  
CuAlTe2   0.41 0.21  
Open in a new tab
a

ΔEFL refers to the energy difference between the F and L high-symmetry k-points, γ is the hybridization strength between the metal-d and oxygen/chalcogen-p states in the VB, Δd is the d-state strength in the VB edge, and pNM-FM corresponds to the hole density, where a nonmagnetic-to-ferromagnetic transition occurs. For the estimation of pNM-FM, calculations for hole densities up to ∼1 × 1022/cm3 are performed.

3.2. Hole Doping Induced by Intrinsic Defects in CuAlO2

As a next step, we examine the possibility of inducing hole doping and also ferromagnetism by intrinsic defects. In our study, we focus on cation vacancies (VCu and VAl) in the neutral charge state. Regarding the structural properties, by removing a Cu atom, two undercoordinated O atoms are found. These atoms exhibit a slight outward relaxation by 2.1%, leading to shorter Al–O bond lengths of 1.852 Å. Accordingly, Cu atoms surrounding the missing lattice site display an outward relaxation of ∼1.3%. Contrary to Cu vacancy, which causes only a minor distortion of the lattice, Al vacancy induces a larger deformation. In particular, threefold-coordinated O atoms, neighboring the vacant site, relax outward by 10.5%, resulting in shorter Cu–O bond lengths of 1.805 Å.

We compute the defect formation energies through the equation Efor = Etot (CuAlO2:VA) – Etot (CuAlO2) + μA, where Etot (CuAlO2:VA) and Etot (CuAlO2) are the energies of the defective and defect-free structures, respectively, whereas μA is the chemical potential of the removed cation atom, which refers to the corresponding solid structure. Cu vacancy formation energy is found to be 0.9 eV, whereas the formation energy of Al vacancy is as high as 8.4 eV, in agreement with previously reported theoretical calculations.54

When a Cu atom is removed, one hole is introduced in the system, whereas the removal of an Al atom corresponds to three induced holes in the system. The corresponding hole densities are about 5 × 1020 and 15 × 1020/cm3 for Cu- and Al-deficient systems, respectively. Cu vacancy leads to a total magnetization of 1 μB, whereas Al vacancy results in a total magnetic moment of 3 μB, which is equivalent to the number of holes. For the case of Cu vacancy, the local magnetic moments of the nearest-neighboring Cu sites are ∼0.05 μB, whereas all other Cu sites present nonzero magnetic moments up to 0.04 μB. In addition, the total Cu, Al, and O magnetizations are found to be ∼0.9, 0.0, and 0.1 μB, respectively. At the same hole concentration, for the hole-injected CuAlO2, the Cu, Al, and O magnetizations are also 90, 0, and 10% of the total magnetization, respectively. Next, the relaxation effect on the magnetic properties is investigated. In agreement with the previous results, the unrelaxed defective structure has a total magnetic moment of 1 μB and the distribution of the local magnetic moments is very similar. Figure 5 shows the spin density plots for both vacancy defects. The spin polarization energies of Cu- and Al-deficient systems are found to be ∼7 and 93 meV, respectively, and the positive values reveal that the ferromagnetic phases are favorable compared to the nonmagnetic phases. Test calculations for singly negatively charged Cu vacancies and triply negatively charged Al vacancies are also performed. The charged defective systems are found to be nonmagnetic, confirming that the induced holes are responsible for the emergence of ferromagnetism.

Figure 5.

Figure 5

Open in a new tab

Spin density plots of (a) Cu-deficient and (b) Al-deficient CuAlO2. The isosurface is 0.001 electrons per Bohr3.

Regarding the electronic properties, Figure 6 shows the DOS of the defective CuAlO2. For both cation vacancies, no deep gap states are observed. Due to the induced holes, the Fermi level is shifted within the VB, leading to an asymmetry between the spin-up and spin-down states, and both systems are found to be half-metals. Obviously, the Fermi-level shift for the Al-deficient system is larger compared to the Cu-deficient one.

Figure 6.

Figure 6

Open in a new tab

DOS of (a) Cu-deficient and (b) Al-deficient CuAlO2. A Lorentzian broadening of 0.02 eV is used. The blue and black lines correspond to spin-up-like and spin-down-like DOS.

3.3. Hole Doping Induced by Extrinsic Defects in CuAlO2

Next, we study the substitution of Al atoms by group-IIA and group-IIB atoms, namely, X = Mg, Ca, Zn, and Cd. Note that Al atoms have three valence electrons, whereas group-IIA and group-IIB have only two valence electrons. Thus, these substitutional dopants are attractive alternatives for inducing hole doping and also ferromagnetism in CuAlO2. Concerning the structural properties, oxygens surrounding the Mg, Ca, Zn, and Cd dopants display an outward relaxation of ∼6.1, 14.1, 8.0, and 14.8%, resulting in elongated Mg–O, Ca–O, Zn–O, and Cd–O bond lengths of 2.04, 2.20, 2.08, and 2.21 Å, respectively. We compute the dopant formation energies through the equation Efor = Etot (CuAlO2:X) – Etot (CuAlO2) + μAl – μX, where Etot (CuAlO2:X) and Etot (CuAlO2) are the energies of the doped and undoped structures, whereas μAl and μX are the chemical potentials of Al and dopant atoms, respectively, which refer to the corresponding solid structures. The formation energies are found to be 1.7, 2.8, 4.8, and 6.7 eV for Mg-, Ca-, Zn-, and Cd-doped systems, respectively.

Taking Mg-doped CuAlO2 as an example, the total magnetic moment is found to be 1 μB, i.e., equivalent to the number of holes. The 32 Cu atoms, which sandwich the dopant-containing Al plane, present local magnetic moments of about 0.02–0.04μB, whereas the remaining 16 Cu atoms present negligible magnetic moments. Magnetization is found to be less localized in plane, compared to the out-of-plane direction. The total Cu and O magnetizations are found to be ∼ 0.9 and 0.1 μB, respectively, whereas the dopant atom and all Al atoms have zero magnetic moments. Remarkably, our findings are different from those reported in similar investigations, where the magnetic moments are found to be strongly localized in the impurity atoms.44 Our theoretical calculations are also in good agreement with recent experimental observations, where CuAl1–xMgxO2 with x = 0.05 is found to be ferromagnetic at room temperature.55 Similar results are also obtained for the magnetic moments for the other group-IIA- and group-IIB-doped systems, and their spin density plots are shown in Figure 7. For all doped systems, positive spin polarization energies are found and their values are 7.2, 6.3, 7.3, and 7.2 meV for Mg, Ca, Zn, and Cd dopants, respectively. Concerning the electronic properties, no deep gap states are observed, and an obvious shift of the Fermi level within the VB along with half-metallicity is observed, as shown in Figure 8.

Figure 7.

Figure 7

Open in a new tab

Spin density plots of doped CuAlO2 (a) Mg-on-Al site, (b) Ca-on-Al site, (c) Zn-on-Al site, and (d) Cd-on-Al site. The orange, purple, brown, and pink spheres correspond to Mg, Ca, Zn, and Cd atoms. The isosurface is set to 0.001 electrons per Bohr3.

Figure 8.

Figure 8

Open in a new tab

DOS of doped CuAlO2 (a) Mg-on-Al site, (b) Ca-on-Al site, (c) Zn-on-Al site, and (d) Cd-on-Al site. A Lorentzian broadening of 0.02 eV is used. The blue and black lines correspond to spin-up-like and spin-down-like DOS.

4. Conclusions

Using density functional theory calculations, we examined the electronic and magnetic properties of CuMIIIAO2 with MIIIA = Al, Ga, and In, upon hole doping. Our simulations revealed a nonmagnetic-to-ferromagnetic transition for the hole-doped systems. Ferromagnetism arises from an exchange splitting of the electronic states at the valence band edge, and it can be attributed to the high-lying Cu-d states dominating over the O-p states. Besides ferromagnetism, half-metallicity has been additionally observed, i.e., the spin-up and spin-down states present semiconducting and metallic behavior, respectively, allowing a fully polarized spin transport. The possibility of inducing hole doping and a subsequent ferromagnetic order by cation vacancies was also investigated. We found that Cu vacancy presents a formation energy lower than 1 eV and leads to a total magnetization of 1 μB. The local magnetic moments of the nearest-neighboring Cu sites are ∼0.05 μB, whereas all other Cu sites present nonzero magnetic moments up to 0.04 μB. Besides cation vacancies, group-IIA and group-IIB atoms replacing Al atoms were examined. Interestingly, these nonmagnetic divalent dopants also result in the emergence of ferromagnetism.

Acknowledgments

Fruitful discussions with Prof. Michel Houssa (KU Leuven) are gratefully acknowledged. This work was financially supported by the Research Council of Norway (ToppForsk project: 251131). Part of this work was supported by computational time at NSC, Linköping University, granted from the Swedish National Infrastructure for Computing (SNIC). The authors acknowledge access to HPC resources provided by UNINETT Sigma2—the National Infrastructure for High Performance Computing and Data Storage in Norway.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsami.1c00403.

  • Bond length variations for hole-doped CuMIIIAO2; temperature-dependent magnetic moment for CuAlO2; band structures of CuGaO2 and CuInO2; carrier dependent magnetic moment and spin polarization energy for CuGaO2 and CuInO2; band structures of hole-doped CuGaO2 and CuInO2; and structural properties of CuMIIIAO2 (PDF)

The authors declare no competing financial interest.

Supplementary Material

am1c00403_si_001.pdf (1.2MB, pdf)

References

  1. Wolf S. A.; Awschalom D. D.; Buhrman R. A.; Daughton J. M.; von Molnár V. S.; Roukes M. L.; Chtchelkanova A. Y.; Treger D. M. Spintronics: A Spin-Based Electronics Vision for the Future. Science 2001, 294, 1488–1495. 10.1126/science.1065389. [DOI] [PubMed] [Google Scholar]
  2. Furdyna J. K. Diluted Magnetic Semiconductors. J. Appl. Phys. 1988, 64, R29–R64. 10.1063/1.341700. [DOI] [Google Scholar]
  3. De Jonge W. J. M.; Swagten H. J. M. Magnetic Properties of Diluted Magnetic Semiconductors. J. Magn. Magn. Mater. 1991, 100, 322–345. 10.1016/0304-8853(91)90827-W. [DOI] [Google Scholar]
  4. Sarma S. D.; Hwang E. H.; Kaminski A. Temperature-Dependent Magnetization in Diluted Magnetic Semiconductors. Phys. Rev. B 2003, 67, 155201 10.1103/PhysRevB.67.155201. [DOI] [Google Scholar]
  5. Sato K.; Bergqvist L.; Kudrnovský J.; Dederichs P. H.; Eriksson O.; Turek I.; Sanyal B.; Bouzerar G.; Katayama-Yoshida H.; Dinh V. A.; Fukushima T.; et al. First-Principles Theory of Dilute Magnetic Semiconductors. Rev. Mod. Phys. 2010, 82, 1633. 10.1103/RevModPhys.82.1633. [DOI] [Google Scholar]
  6. Matsumoto Y.; Murakami M.; Shono T.; Hasegawa T.; Fukumura T.; Kawasaki M.; Ahmet P.; Chikyow T.; Koshihara S. Y.; Koinuma H. Room-Temperature Ferromagnetism in Transparent Transition Metal-Doped Titanium Dioxide. Science 2001, 291, 854–856. 10.1126/science.1056186. [DOI] [PubMed] [Google Scholar]
  7. Matsumoto Y.; Takahashi R.; Murakami M.; Koida T.; Fan X. J.; Hasegawa T.; Fukumura T.; Kawasaki M.; Koshihara S. Y.; Koinuma H. Ferromagnetism in Co-Doped TiO2 Rutile Thin Films Grown by Laser Molecular Beam Epitaxy. Jpn. J. Appl. Phys. 2001, 40, L1204–L1206. 10.1143/JJAP.40.L1204. [DOI] [Google Scholar]
  8. Wang Z.; Wang W.; Tang J.; Tung L. D.; Spinu L.; Zhou W. Extraordinary Hall Effect and Ferromagnetism in Fe-Doped Reduced Rutile. Appl. Phys. Lett. 2003, 83, 518–520. 10.1063/1.1593825. [DOI] [Google Scholar]
  9. Ogale S. B.; Choudhary R. J.; Buban J. P.; Lofland S. E.; Shinde S. R.; Kale S. N.; Kulkarni V. N.; Higgins J.; Lanci C.; Simpson J. R.; Browning N. D.; et al. High Temperature Ferromagnetism with a Giant Magnetic Moment in Transparent Co-doped SnO2-δ. Phys. Rev. Lett. 2003, 91, 077205 10.1103/PhysRevLett.91.077205. [DOI] [PubMed] [Google Scholar]
  10. Buchholz D. B.; Changm R. P. H.; Song J. H.; Ketterson J. B. Room-Temperature Ferromagnetism in Cu-Doped ZnO Thin Films. Appl. Phys. Lett. 2005, 87, 082504 10.1063/1.2032588. [DOI] [Google Scholar]
  11. Venkatesan M.; Fitzgerald C. B.; Coey J. M. D. Unexpected Magnetism in a Dielectric Oxide. Nature 2004, 430, 630 10.1038/430630a. [DOI] [PubMed] [Google Scholar]
  12. Ghosh S.; Khan G. G.; Mandal K.; Samanta A.; Nambissan P. M. G. Evolution of Vacancy-Type Defects, Phase Transition, and Intrinsic Ferromagnetism During Annealing of Nanocrystalline TiO2 Studied by Positron Annihilation Spectroscopy. J. Phys. Chem. C 2013, 117, 8458–8467. 10.1021/jp3115836. [DOI] [Google Scholar]
  13. Yi J. B.; Lim C. C.; Xing G. Z.; Fan H. M.; Van L. H.; Huang S. L.; Yang K. S.; Huang X. L.; Qin X. B.; Wang B. Y.; Wu T.; Wang L.; Zhang H. T.; Gao X. Y.; Liu T.; Wee A. T. S.; Feng Y. P.; Ding J. Ferromagnetism in Dilute Magnetic Semiconductors through Defect Engineering: Li-Doped ZnO. Phys. Rev. Lett. 2010, 104, 137201 10.1103/PhysRevLett.104.137201. [DOI] [PubMed] [Google Scholar]
  14. Martínez-Boubeta C.; Beltrán J. I.; Balcells L.; Konstantinović Z.; Valencia S.; Schmitz D.; Arbiol J.; Estrade S.; Cornil J.; Martínez B. Ferromagnetism in Transparent Thin Films of MgO. Phys. Rev. B 2010, 82, 024405 10.1103/PhysRevB.82.024405. [DOI] [Google Scholar]
  15. Srivastava S. K.; Lejay P.; Barbara B.; Pailhès S.; Madigou V.; Bouzerar G. Possible Room-Temperature Ferromagnetism in K-Doped SnO2: X-Ray Diffraction and High-Resolution Transmission Electron Microscopy Study. Phys. Rev. B 2010, 82, 193203 10.1103/PhysRevB.82.193203. [DOI] [Google Scholar]
  16. Yang K.; Dai Y.; Huang B.; Whangbo M. H. On the Possibility of Ferromagnetism in Carbon-Doped Anatase TiO2. Appl. Phys. Lett. 2008, 93, 132507 10.1063/1.2996024. [DOI] [Google Scholar]
  17. Yang K.; Dai Y.; Huang B.; Whangbo M. H. Density Functional Studies of the Magnetic Properties in Nitrogen Doped TiO2. Chem. Phys. Lett. 2009, 481, 99–102. 10.1016/j.cplett.2009.09.050. [DOI] [Google Scholar]
  18. Yang K.; Wu R.; Shen L.; Feng Y. P.; Dai Y.; Huang B. Origin of d0 Magnetism in II-VI and III-V Semiconductors by Substitutional Doping at Anion Site. Phys. Rev. B 2010, 81, 125211 10.1103/PhysRevB.81.125211. [DOI] [Google Scholar]
  19. Kenmochi K.; Ann Dinh V.; Sato K.; Yanase A.; Katayama-Yoshida H. Materials Design of Transparent and Half-Metallic Ferromagnets of MgO, SrO and BaO without Magnetic Elements. J. Phys. Soc. Jpn. 2004, 73, 2952–2954. 10.1143/JPSJ.73.2952. [DOI] [Google Scholar]
  20. Kenmochi K.; Seike M.; Sato K.; Yanase A.; Katayama-Yoshida H. New Class of Diluted Ferromagnetic Semiconductors Based on CaO without Transition Metal Elements. Jpn. J. Appl. Phys. 2004, 43, L934. 10.1143/JJAP.43.L934. [DOI] [Google Scholar]
  21. Kemmochi K.; Seike M.; Sato K.; Yanase A.; Katayama-Yoshida H. New Class of High-Tc Diluted Ferromagnetic Semiconductors Based on CaO without Transition Metal Elements. J. Supercond. 2005, 18, 37–40. 10.1007/s10948-005-2147-7. [DOI] [Google Scholar]
  22. Sato K.; Bergqvist L.; Kudrnovský J.; Dederichs P. H.; Eriksson O.; Turek I.; Sanyal B.; Bouzerar G.; Katayama-Yoshida H.; Dinh V. A.; Fukushima T.; et al. First-Principles Theory of Dilute Magnetic Semiconductors. Rev. Mod. Phys. 2010, 82, 1633. 10.1103/RevModPhys.82.1633. [DOI] [Google Scholar]
  23. Sato K.; Katayama-Yoshida H.; Dederichs P. H. Dilute Magnetic Semiconductors. Psi-k Newsl. 2005, 70, 93–110. [Google Scholar]
  24. Kizaki H.; Sato K.; Yanase A.; Katayama-Yoshida H. Ab Initio Calculations of CuAlO2-Based Dilute Magnetic Semiconductor. Phys. B 2006, 376–377, 812–815. 10.1016/j.physb.2005.12.203. [DOI] [Google Scholar]
  25. Chen C.; Dong C.; Wang B.; Huang J.; Wang Y. Synthesis and Room Temperature Ferromagnetism in Fe-Doped CuAlO2 Semiconductor. J. Wuhan Univ. Technol., Mater. Sci. Ed. 2013, 28, 500–503. 10.1007/s11595-013-0720-7. [DOI] [Google Scholar]
  26. Dong C. J.; Yu W. X.; Xu M.; Cao J. J.; Zhang Y.; Chuai Y. H.; Wang Y. D. Evidence of Room Temperature Ferromagnetism in Co-Doped Transparent CuAlO2 Semiconductor. J. Alloys Compd. 2012, 512, 195–198. 10.1016/j.jallcom.2011.09.062. [DOI] [Google Scholar]
  27. Kizaki H.; Sato K.; Yanase A.; Katayama-Yoshida H. First-Principles Materials Design of CuAlO2 Based Dilute Magnetic Semiconducting Oxide. Jpn. J. Appl. Phys. 2005, 44, L1187–L1189. 10.1143/JJAP.44.L1187. [DOI] [Google Scholar]
  28. Bandara J.; Yasomanee J. P. P-Type Oxide Semiconductors as Hole Collectors in Dye-Sensitized Solid-State Solar Cells. Semicond. Sci. Technol. 2007, 22, 20–24. 10.1088/0268-1242/22/2/004. [DOI] [Google Scholar]
  29. Zheng X. G.; Taniguchi K.; Takahashi A.; Liu Y.; Xu C. N. Room Temperature Sensing of Ozone by Transparent p-Type Semiconductor CuAlO2. Appl. Phys. Lett. 2004, 85, 1728–1729. 10.1063/1.1784888. [DOI] [Google Scholar]
  30. Ahmed J.; Blakely C. K.; Prakash J.; Bruno S. R.; Yu M.; Wu Y.; Poltavets V. V. Scalable Synthesis of Delafossite CuAlO2 Nanoparticles for p-Type Dye-Sensitized Solar Cells Applications. J. Alloys Compd. 2014, 591, 275–279. 10.1016/j.jallcom.2013.12.199. [DOI] [Google Scholar]
  31. Kawazoe H.; Yasukawa M.; Hyodo H.; Kurita M.; Yanagi H.; Hosono H. P-Type Electrical Conduction in Transparent Thin Films of CuAlO2. Nature 1997, 389, 939–942. 10.1038/40087. [DOI] [Google Scholar]
  32. Zhang K. H.; Xi K.; Blamire M. G.; Egdell R. G. P-Type Transparent Conducting Oxides. J. Phys.: Condens. Matter 2016, 28, 383002 10.1088/0953-8984/28/38/383002. [DOI] [PubMed] [Google Scholar]
  33. Kresse G.; Furthmüller J. Efficient Iterative Schemes for Ab Initio Total-Energy Calculations Using a Plane-Wave Basis Set. Phys. Rev. B 1996, 54, 11169–11186. 10.1103/PhysRevB.54.11169. [DOI] [PubMed] [Google Scholar]
  34. Kresse G.; Furthmüller J. Efficiency of Ab-Initio Total Energy Calculations for Metals and Semiconductors Using a Plane-Wave Basis Set. Comput. Mater. Sci. 1996, 6, 15–50. 10.1016/0927-0256(96)00008-0. [DOI] [PubMed] [Google Scholar]
  35. Perdew J. P.; Burke K.; Ernzerhof M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865–3868. 10.1103/PhysRevLett.77.3865. [DOI] [PubMed] [Google Scholar]
  36. Blöchl P. E. Projector Augmented-Wave Method. Phys. Rev. B 1994, 50, 17953–17979. 10.1103/PhysRevB.50.17953. [DOI] [PubMed] [Google Scholar]
  37. Heyd J.; Scuseria G. E.; Ernzerhof M. Hybrid Functionals Based on a Screened Coulomb Potential. J. Chem. Phys. 2003, 118, 8207–8215. 10.1063/1.1564060. [DOI] [Google Scholar]
  38. Ishiguro T.; Kitazawa A.; Mizutani N.; Kato M. Single-Crystal Growth and Crystal Structure Refinement of CuAlO2. J. Solid State Chem. 1981, 40, 170–174. 10.1016/0022-4596(81)90377-7. [DOI] [Google Scholar]
  39. Yanagi H.; Inoue S. I.; Ueda K.; Kawazoe H.; Hosono H.; Hamada N. Electronic Structure and Optoelectronic Properties of Transparent p-Type Conducting CuAlO2. J. Appl. Phys. 2000, 88, 4159–4163. 10.1063/1.1308103. [DOI] [Google Scholar]
  40. Meng R.; Houssa M.; Iordanidou K.; Pourtois G.; Afanas’ev V. V.; Stesmans A. Two-Dimensional Gallium and Indium Oxides from Global Structure Searching: Ferromagnetism and Half Metallicity via Hole Doping. J. Appl. Phys. 2020, 128, 034304 10.1063/5.0012103. [DOI] [Google Scholar]
  41. Meng R.; Houssa M.; Iordanidou K.; Pourtois G.; Afanas’ev V. V.; Stesmans A. Ferromagnetism and Half-Metallicity in Two-Dimensional MO (M=Ga,In) Monolayers Induced by Hole Doping. Phys. Rev. Mater. 2020, 4, 074001 10.1103/PhysRevMaterials.4.074001. [DOI] [Google Scholar]
  42. Houssa M.; Meng R.; Iordanidou K.; Pourtois G.; Afanas’ev V. V.; Stesmans A. Doping-Induced Ferromagnetism in InSe and SnO Monolayers. J. Comput. Electron. 2021, 20, 88–94. 10.1007/s10825-020-01535-0. [DOI] [Google Scholar]
  43. Houssa M.; Meng R.; Iordanidou K.; Pourtois G.; Afanas’ev V. V.; Stesmans A. Stoner Ferromagnetism in Two-Dimensional Materials. ECS Trans. 2019, 92, 35–41. 10.1149/09201.0035ecst. [DOI] [Google Scholar]
  44. Iordanidou K.; Houssa M.; Persson C. Carrier-Mediated Ferromagnetism in Two-Dimensional PtS2. RSC Adv. 2020, 10, 952–957. 10.1039/C9RA09756B. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Iordanidou K.; Houssa M.; Kioseoglou J.; Afanas’ev V. V.; Stesmans A.; Persson C. Hole-Doped 2D InSe for Spintronic Applications. ACS Appl. Nano Mater. 2018, 1, 6656–6665. 10.1021/acsanm.8b01476. [DOI] [Google Scholar]
  46. Houssa M.; Iordanidou K.; Pourtois G.; Afanas’ev V. V.; Stesmans A. Ferromagnetism in two-dimensional hole-doped SnO. AIP Adv. 2018, 8, 055010 10.1063/1.5025272. [DOI] [Google Scholar]
  47. Houssa M.; Iordanidou K.; Pourtois G.; Afanas’ev V. V.; Stesmans A. Hole-Doping Induced Ferromagnetism in Monolayer SnO: A First-Principles Study. ECS Trans. 2017, 80, 339–345. 10.1149/08001.0339ecst. [DOI] [Google Scholar]
  48. Janesko B. G.; Henderson T. M.; Scuseria G. E. Screened Hybrid Density Functionals for Solid-State Chemistry and Physics. Phys. Chem. Chem. Phys. 2009, 11, 443–454. 10.1039/B812838C. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Toyoda M.; Akai H.; Sato K.; Katayama-Yoshida H. Electronic Structures of (Zn, TM) O (TM: V, Cr, Mn, Fe, Co, and Ni) in the Self-Interaction-Corrected Calculations. Phys. B 2006, 376–377, 647–650. 10.1016/j.physb.2005.12.163. [DOI] [Google Scholar]
  50. Toyoda M.; Akai H.; Sato K.; Katayama-Yoshida H. Curie Temperature of GaMnN and GaMnAs from LDA-SIC Electronic Structure Calculations. Phys. Status Solidi C 2006, 3, 4155–4159. 10.1002/pssc.200672846. [DOI] [Google Scholar]
  51. An Dinh V.; Toyoda M.; Sato K.; Katayama-Yoshida H. Exchange Interaction and T C in Alkaline-Earth-Metal-Oxide-Based DMS without Magnetic Impurities: First Principle Pseudo-SIC and Monte Carlo Calculation. J. Phys. Soc. Jpn. 2006, 75, 093705 10.1143/JPSJ.75.093705. [DOI] [Google Scholar]
  52. Cao T.; Li Z.; Louie S. G. Tunable Magnetism and Half-Metallicity in Hole-Doped Monolayer GaSe. Phys. Rev. Lett. 2015, 114, 236602 10.1103/PhysRevLett.114.236602. [DOI] [PubMed] [Google Scholar]
  53. Ashcroft N. W.; Mermin N. D.. Solid State Physics; Holt Rinehart/Winston: New York, 1976. [Google Scholar]
  54. Kumar M.; Zhao H.; Persson C. Study of Band-Structure, Optical Properties and Native Defects in AIBIIIO2 (AI = Cu or Ag, BIII = Al, Ga or In) Delafossites. Semicond. Sci. Technol. 2013, 28, 065003 10.1088/0268-1242/28/6/065003. [DOI] [Google Scholar]
  55. Agrawal S.; Parveen A.; Azam A. Influence of Mg on Structural, Electrical and Magnetic Properties of CuAlO2 Nanoparticles. Mater. Lett. 2016, 168, 125–128. 10.1016/j.matlet.2016.01.046. [DOI] [Google Scholar]
  56. Wang V.; Xu N.; Liu J. C.; Tang G.; Geng W. T.. Vaspkit: A User-Friendly Interface Facilitating High-Throughput Computing and Analysis Using Vasp Code, arXiv:1908.08269. arXiv.org e-Print archive. https://arxiv.org/abs/1908.08269 (accessed Aug 22, 2019).

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

am1c00403_si_001.pdf (1.2MB, pdf)

Articles from ACS Applied Materials & Interfaces are provided here courtesy of American Chemical Society

RESOURCES