ELSEVIER

Contents lists available at ScienceDirect

Results in Physics

journal homepage: www.journals.elsevier.com/results-in-physics

Ferromagnetism in CdOX (X = Mn and N) with and without intrinsic point defects: A density functional theory q

Z. Nabia, S. Amarib'*, S. Méçabihb, A. Zaouib, B. Abbarb, B. Bouhafsb, R. Ahujac

a Laboratory of Catalysis and Reactive Systems, Physics Department, University of Sidi Bel Abbes, 22000, Algeria

b Laboratoire de Modélisation et de Simulation en Sciences des Matériaux, Département de Physique Université Djillali Liabès, Sidi Bel-Abbes, Algeria c Condensed Matter Theory Group, Department of Physics, Uppsala University, Box 530, S-751 21 Uppsala, Sweden

ARTICLE INFO ABSTRACT

The purpose of this study is to further understanding of the structural, electronic, magnetic properties of CdO doped with transition metal (Mn) and non metal element (N). The calculations are performed by the developed full-potential augmented plane wave plus local orbitals method within the spin density functional theory. As exchange-correlation potential we used the generalized gradient approximation (GGA) form. Moreover, the electronic structure study for our compounds was performed with and without oxygen deficiency. We treated the ferromagnetic and antiferromagnetic states and we found that all compounds are stable in the ferromagnetic structure, and all doped materials CdO:Mn and CdO:N adopt the half metallic character. In addition, we notice that the oxygen vacancy destroyed the ferromagnetism in N doped CdO, while Mn doped CdO becomes semiconductor.

© 2013 The Authors. Published by Elsevier B.V. All rights reserved.

CrossMark

Article history: Received 9 July 2013 Accepted 22 September 2013 Available online 27 September 2013

Keywords: DFT

FPLAPW Intrinsic defect Magnetism Spintronic

1. Introduction

Spintronics devices, which utilize both the charge and the spin freedom of electrons to create new functionalities beyond conventional semiconductor devices, have attracted much attention recently [1]. The search for materials combining properties of the ferromagnet and the semiconductor has been challenging because of differences in crystal structure and chemical bonding [2]. Dilute magnetic semiconductors (DMSs), are thought to be ideal materials for this field, enabling versatility in doping and fabrication of various structures and simple integration with the dominant semiconductor technology. As a wide ranging band gap of 2.2-2.7 eV semiconductor with a direct gap, CdO draws the attention of scientists for its potential use in UV laser diodes and UV-blue light emitters [3]. Very recently, CdO has also been identified as a promising host material for the realization of DMSs due to the prediction of possible ferromagnetism in Cd1_xMnxO [4]. Computational methods based on density functional theory (DFT) also predict ferromagnetism in most 3d transition metal (TM) based CdO [5-7]. More recently, it has also been predicted that the ferromagnetism

q This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial-No Derivative Works License, which permits non-commercial use, distribution, and reproduction in any medium, provided the original author and source are credited.

* Corresponding author. Address: Faculté des Sciences, Université Djillali Liabes, BP 89, Sidi Bel Abbes 22000, Algeria. Tel./fax: +213 48 54 43 44.

E-mail address: siham_amari@yahoo.fr (S. Amari).

in class of oxides can be induced by native defects, bringing new challenges for understanding the origin of ferromagnetism in these oxide compounds. We could mention for example, Ca vacancy in CaO [8-12], Hf vacancy in HfO2 [13-15], In2O3 [12,16,17], TiO2 [13] and ZnO [6,18]. In this paper, we present our first principles studies of the structural, electronic and magnetic properties of Mn and N doped CdO. The aim of our study is to understand the role played by defects in the origin of ferromagnetism in such narrow band gap semiconductor. The paper is organized as follows: in Section 2, we present the method of calculation, Section 3 contains all results and discussions, and finally we finish with a summary and conclusion in Section 4.

2. Computational method

The calculations were performed by using the full potential linearized augmented plane wave plus local orbitals (FP-L/APW+ lo) method based on the spin-polarized DFT and implemented in Wien2k [19] code. The exchange and correlations electronic energy were calculated with the generalized gradient approximation (GGA) [20]. Further, a muffin-tin model for the crystal potential is assumed and the unit cell is divided into two regions, within and outside the muffin-tin sphere. Inside the muffin-tin sphere, the spherical harmonic expansion is used and outside the sphere the plane wave basis set is chosen. The valence part is treated within a potential expanded into spherical harmonics up to l = 4. The valence wave functions, inside the spheres, are expanded up to l = 10. In order to achieve the energy convergence, we have

2211-3797/$ - see front matter © 2013 The Authors. Published by Elsevier B.V. All rights reserved. http://dx.doi.Org/10.1016/j.rinp.2013.09.009

Fig. 1. Schematic structure diagram for 2 x 2 x 2 supercell of CdO, positions of the host atoms (Cd and O), dopant (N) and oxygen vacancy (vac) are labeled.

expanded the basis function up to RMT„KMAX = 7 (RMT is the average radius of the spheres muffin-tin and KMAX is the maximum value of the wave vector K = k + G). The charge density was expanded with Gmax= 14. The relativistic effects are taken into account by using scalar relativistic. We have adopted the values of 2.3, 2.3, 1.45 and 1.9 a.u. for Cd, Mn, O and N atoms, respectively. Our calculations were performed in the rocksalt ferromagnetic phase, at the concentration x = 0.125, where two Cd are substituted by two Mn atoms and two O by two N atoms. We have considered the ferromagnetic FM and antiferromagnetic AFM phases. All calculations were performed using unit cell parameters optimized by minimizing the total energy as function of volumes. The self-consistent calculations are considered to be converged only when the calculated total energy of the crystal converged to less than 1 mRy. Ferromagnetic stability is determined by the total energy difference (DE) of the supercell between antiferromagnetic (AFM) state and ferromagnetic (FM) state. A representative structure diagrams corresponding to all studied systems are shown in Figs. 1 and 2. In Fig. 1, the positions of host atoms (Cd and O) and dopant (N) along

Fig. 2. Schematic structure diagram for 2 x 2 x 2 supercell of CdO, positions of the host atoms (Cd and O), dopant (Mn) and oxygen vacancy (vac) are labeled.

Table 1

Calculated values for the lattice parameter (in A), bulk modulus (in Gpa) and AE (in mRy) for all configurations considered in the present study.

Systems a B AE

Cd,6N2O,4 9.4432 119.896 +2.05

Cd,6N2O,3 9.3702 118.662 -8.51

Cdi4Mn2Oi6 9.4836 i29.897 +0.73

Cd,4Mn2O,5 9.4217 125.565 +150.8

with oxygen vacancy are labeled. While Fig. 2, represents, the positions of Cd, O and Mn with oxygen vacancy. The lattice parameter a, based on optimization of the energy of the perfect lattice, was calculated to be a = 4.67 A and compared very well with the corresponding experimental value of 4.69232 A [21]. The oxygen vacancy is simulated by using Cd16N2O13 and Cd14Mn2O15 super cells and the optimization of all atomic coordinates was carried out using the same procedure as discussed above.

3. Results and discussion

The structural properties are obtained by minimization of the total energy depending on the volumes of CdNO and CdMnO in FM and AFM phases with and without oxygen vacancy using GGA. We computed the lattice constants, bulk moduli by fitting the total energy versus volume according to Murnaghan's equation of state [22]. The obtained results are listed in Table 1. Up to now, no experimental or theoretical lattice constant of the compound has been reported. However, we can see that the introduction of

40 20 0 -20 -40 0,5 0,0

D -0,5

-d -1,0

0 -1 -2 1 0 -1 -2

-10 -8 -6 -4 -2 0 2 4 6 Energy (eV)

Fig. 3. Total DOS and partial PDOS of Cd16N2O14 compound.

40 20 0 -20 -40 0,5 0,0 -0,5 f -10 ■e 0,75

° 0,00 -0,75 -1,50 1 0

Cd16N2013

spin-up

spin-dn

Cd Cd Cd

-2 _i_I_i_I_i_I_i_I_i

-10 -8 -6 -4 -2 0 2 4 6

Energy(eV)

Fig. 4. Total DOS and partial PDOS of Cd16N2O13 compound.

vacancies in Cd16N2O14 and Cd14Mn2O16 systems leads to a reduction in volume and bulk moduli.

Now, we move our discussion toward the Cd16N2O14 compound by substituting two O by two N atoms. Our calculations give positive values of the total energy difference DE = EAFM - EFM = +2.05 mRy, thus Cd16N2O14 is stable in the FM phase. Moreover, this value means that this present result is in line with the previous work predicting a ferromagnetic-half-metallic (FMHM) character by GGA for Cd32N2O30 [23]. Also, a similar FM coupling was found in N-doped MgO [24] and C-doped ZnO [25]. Fig. 3 presents the calculated total (TDOS) and partial (PDOS) densities of states of the N dopant and its nearest neighboring O and Cd atoms. The Fermi level EF is set to zero and represented by the vertical line. This figure indicates that the N-2p states significantly overlap with the O-2p near the Fermi level. In other words, this means that there has been a strong interaction between p states of N and p states of O in CdO:N compound. Therefore, new states originate from N-2p and O-2p orbitals appeared near the Fermi level, which are responsible for the half metallic character. That is to say, the valence electrons at the Fermi level are almost fully polarized which gives rise to ferromagnetism. Most of calculations [23-26] that are performed in the same context have shown that the hole mediated double exchange mechanism is mainly responsible for the ferromagnetism coupling in non-TM doped oxides. Our theoretical conclusion agrees with these works and we believe that the strong p-p interaction between Nitrogen and Oxygen atoms induces spin polarized holes which leads to ferromagnetism in CdO. Furthermore, to explore the effect of oxygen vacancy on the magnetism of Cd16N2O13, we did the same steps of calculations with removing of one oxygen atom. TDOS and PDOS of Cd16N2O13 are shown in Fig. 4. We notice a little change in the electronic structure

in particular in the energy regions near the band edge comparatively to those without O vacancy. From this figure, we observe a shift of N 2p and O 2p orbitals toward conduction band region. Also, we remark that the spin up and spin down states are identical and the total magnetic moment is equal to zero. In this case, we say that the system is antiferromagnetic. This result is confirmed by the value of DE (-8.51 mRy) shown in Table 1 which indicates that the system is antiferromagnetic. Consequently, the half metallic character has disappeared with the disappearance of ferromagne-tism. We therefore deduce that the introduction of the oxygen vacancy has destroyed the ferromagnetism of the system. Finally, to investigate the ferromagnetism induced by transition metal doping, we have studied the Mn substitution for Cd in CdO. The calculations are performed with and without oxygen vacancy of Cd14Mn2O16 using the same supercell (2 x 2 x 2). From Table 1, we remark that the FM is the most stable, the total energies are very close, but the FM is lower by 0.73 mRy than the AFM phase. This result is in line with the experimental work of Rajkumar et al., [27]. The total (TDOS) and partial (PDOS) densities of states are shown in Fig. 5 and the EF is set to zero and represented by the vertical line. We have obtained a half metallic character that originates from the valence electrons at the Fermi level that are almost fully polarized. It is well known that the hybridization between Mn dopant and its neighboring host atom results in the splitting of the energy levels near the EF, which shifts the majority spin (spin-up) and minority spin (spin-down) channel near EF, leading to magnetism in CdO. Indeed, the Mn 3d orbitals split into t2g and eg orbitals in a tetrahedral crystal field of CdO crystal structure and the O-p orbitals have t2 symmetry. Therefore, the hybridization between Mn-dt2g and O-p states is responsible for the half-metallic behavior. This suggests that the double exchange

40 20 0 -20 -40 0,5 0,0 -0,5 -1,0 2 0 -2 -4 0,5 0,0 -0,5 -1,0

-10 -8 -6 -4 -2 0 Energy (eV)

Fig. 5. Total DOS and partial PDOS of Cd14Mn2O16 compound.

f -1,0

Cd14Mn2O15

spin-up

Mn s Mn p Mn d

-10 -8

-6 -4-2 0 2 Energy (eV)

and produces small local magnetic moments on the non-magnetic Cd and O sites. The magnetic moments are mainly contributed by the Mn-3d orbital (4.4733 ip/Mn). The nearest-neighboring O atom contributes 0.0285 ip/O and next neighbor Cd atom provides 0.0015 ip/Cd. Here, unlike the non metals, the p-d hybridization between 2-p and 3-d states will induce a certain amount of spin polarized hole states which are responsible for ferromagnetism in transition metal doped semi-conductors or oxides.

4. Conclusion

In summary, we have performed first principles calculations based on the FP-L/APW + lo method using GGA to evaluate the electronic and magnetic properties of cubic CdMnO and CdNO with 25% of Mn and N in FM and AFM phases. The calculated total energies show that the total energy difference DE = EAFM - EFM is positive, thus; both compounds are stable in the FM phase. Structural parameters have been computed. The calculated densities of states presented in this study identify the half-metallic behavior. These results are in agreement with previous findings mentioned above. On the other hand, we have found that the introduction of the n-type defects as O vacancy to N doped CdO leads to vanishing of ferromagnetism. CdO:N became antiferromagnetic. This is radically different from other systems previously known to exhibit point defect ferromagnetism. Hence, we can argue that Mn and N are the best potential as a Cd substitutional dopant in CdO for producing DMS. However, the mechanism which leads to the observed exchange coupling is different for the two cases. For N doped CdO, the p-p interaction between N and neighboring oxygen is mainly the origin of FM coupling in this system. While for CdO:Mn, the strong hybridization between 3d and O 2p is found to be responsible for FM coupling.

Fig. 6. Total DOS and partial PDOS of Cd14Mn2O15 compound.

References

Table 2

Calculated total and local magnetic moments (in iB) within the muffin tin spheres and in the interstitial sites for Cd16N2O14, Cd16N2O13, Cd14Mn2O16 and Cd14Mn2O15.

Systems Tot Cd N Mn O Interstitial

Cdi6N2Oi4 1.96 0.00029 0.42 - 0.05 0.15

Cdi6N2Oi3 0.0 - - - - -

CdMMn2Oi6 9.92 0.00159 - 4.47 0.02 0.57

Cdi4Mn2Oi5 10.02 0.0019 - 4.48 0.033 0.65

mechanism is most likely responsible for the FM coupling in this system.

Fig. 6 shows the TDOS and PDOS of Cd14Mn2O15 compound. It can be seen from the DOS a little shift of Mn-3d and O-2p orbitals toward conduction band region. Consequently the structure has changed from a half-metallic to a nearly semiconductor with an energy gap of 0.47 eV. The Fermi level falls within a region of very small spin-up DOS.

In Table 2, we summarized the calculated total and local magnetic moments per atom within the muffin-tin spheres as well as in the interstitial sites for CdO:N and CdO:Mn. One can notice that for CdO:N the total magnetic moment comes from the N ion. For CdO:Mn, the unoccupied Mn d states produce permanent local magnetic moments. The results show that the total magnetic moments generally come from the N and Mn ions with a small contribution of Cd and O sites. The p-d hybridization reduces the total magnetic moments of the Mn atom from its free space charge value

[1] Wolf SA, Awschalom DD, Buhrman RA, Daughton JM, Von Molnar S, Roukes ML, Chtchelkanova AY, Treger DM. Science 2001;294:1488.

[2] Tanaka M. J Cryst Growth 1999;202:660.

[3] Jeong SH, Kim BS, Lee BT. Appl Phys Lett 2003;82:2625.

[4] Kusigerski V, Spasojevic V, Mitric M. J Phys: Condens Matter 1996;8:10581.

[5] Thangavel R, Rajagopalan M, Kumar J. Physica B 2008;403:2768.

[6] Amari S, Megabih S, Abbar B, Bouhafs B. J Magn Magn Mater 2013;327:76-8.

[7] Amari S, Megabih S, Abbar B, Bouhafs B. Am J Condens Matter Phys 2012;2:27-31.

[8] Kenmochi K, Seike M, Sato K, Yanase A, Katayama Yoshida Hi. Jpn J Appl Phys, Part 2 2004;43:L934.

[9] Sharma V, Lowther JE. J Nano-Electron Phys 2011;3(№1):453-9.

[10] Gupta A, Cao H, Parekh K, Rao KV, Raju AR, Waghmare UV. J Appl Phys 2007;101:09N513.

[11] Huang LM, Moyses Araujo C, Ahuja R. EPL 2009;87:27013.

[12] Elfimov IS, Yunoki S, Sawazky GA. Phys Rev Lett 2002;89:216403.

[13] Venkatesan M, Fitzgerand CB, Coye JMD. Nature (London) 2004;430:630.

[14] Coye JMD, Venkatesan M, Stamenov P, Fitzgerand CB, Dorneles LS. Phys Rev B 2005;72:024450.

[15] Pemmaraju CP, Sanvito S. Phys Rev Lett 2005;94:217205.

[16] Hong NH, Sakai J, Poirot N, Brize V. Phys Rev 2006;B73:132404.

[17] Sunaresan A, Bhargavi R, Rangarajan N, Siddesh U, Rao CNR. Phys Rev 2006;B74:161306(R).

[18] Hong NH, Sakai J, Poirot N, Brize V. J Phys: Condens Matter 2007;19:036219.

[19] Schwarz K, Blaha P, Masden GKH. Comput Phys Commun 2002;147:71.

[20] PerdewJP, Burke K, Ernzerhof M. Phys Rev Lett 1996;77:3865.

[21] Ferro R, Rodriguez JA, Solomko M, Bertrand P. Phys Status Solidi C 2005;2(№10):3548.

[22] Murnaghan FD. Proc Natl Acad Sci USA 1944;30:5390.

[23] Zhang Chang-Wen, Wang Pei-ji, Su Yan. Phys Lett A 2010;374:1889.

[24] Sharma V, Lowther JE. J Nano Elect Phys 2011;3(№1):453.

[25] Pan H, Shen JBYL, Wu RQ, YangJH, LinJY, Feng YP, Ding J, Yin LH. Phys Rev Lett 2007;99:17201.

[26] Zhang Chang-Wen, Yan Shi-Shen. Appl Phys Lett 2009;95:232108.

[27] Rajkumar N, Susila VM, Ramachandran K.J Exp Nanosci 2011;6:389.