- Email: [email protected]
An insight to origin of ferromagnetism in ZnO and N implanted ZnO thin films: Experimental and DFT approach
Accepted Manuscript An insight to origin of ferromagnetism in ZnO and N implanted ZnO thin films: Experimental and DFT approach Parmod Kumar, Hitendra K. Malik, Anima Ghosh, R. Thangavel, K. Asokan PII:
S0925-8388(18)32597-0
DOI:
10.1016/j.jallcom.2018.07.097
Reference:
JALCOM 46813
To appear in:
Journal of Alloys and Compounds
Received Date: 28 February 2018 Revised Date:
25 June 2018
Accepted Date: 9 July 2018
Please cite this article as: P. Kumar, H.K. Malik, A. Ghosh, R. Thangavel, K. Asokan, An insight to origin of ferromagnetism in ZnO and N implanted ZnO thin films: Experimental and DFT approach, Journal of Alloys and Compounds (2018), doi: 10.1016/j.jallcom.2018.07.097. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
An Insight to origin of ferromagnetism in ZnO and N implanted ZnO thin films: Experimental and DFT approach Parmod Kumar1,2,*, Hitendra K. Malik2,*, Anima Ghosh3, R. Thangavel3 and K. Asokan2 1
Materials Science Division, Inter University Accelerator Centre, New Delhi – 110067, India Department of Physics, Indian Institute of Technology Delhi, New Delhi – 110016, India 3 Department of Applied Physics, Indian Institute of Technology (Indian School of Mines), Dhanbad– 826004, India
RI PT
2
M AN U
SC
In the present study, we have elucidated an effective way to simultaneously tune the optical bandgap and magnetic properties of zinc oxide (ZnO). This can be achieved by the incorporation of nitrogen ions via any means in the host matrix. On a broader way, we have systematically investigated the influence of N ions in ZnO thin films through experimental techniques and density functional theory (DFT) calculations to understand the physical mechanism governing the observed magnetic behaviour and variation in bandgap. For this, RF sputtered ZnO thin films deposited over Si substrates were implanted with N ions by varying the fluences and studied for their optical and magnetic properties. The pristine ZnO films is having saturation magnetization of ~ 2.45 emu/cm3 which becomes almost twice compared for the fluence of 1×1017 ions/cm2. Furthermore, the optical bandgap is tuned from 3.27 eV to 3.04 eV with N ion fluences. Our work signifies the new insight to understand the basis of ferromagnetism in non-magnetic ions doped ZnO system.
Corresponding authors:
TE D
Keywords: Ion implantation, bandgap tuning, defects induced magnetism, first principle calculations.
*Email: [email protected], [email protected] Phone: +91-11-26893955, Fax: + 91-11-26893666
AC C
EP
Present Affiliation of Corresponding author: Department of Physics, Deenbandhu Chhotu Ram University of Science and Technology, Murthal, Sonepat-131039, India
1
ACCEPTED MANUSCRIPT
1. Introduction ZnO has remarkable properties such as large binding energy (60 meV) and wide band gap of 3.37 eV which makes it an excellent material for excitonic devices [1-3]. However, in recent years there has been significant interest in the magnetic behaviour of ZnO i.e. dilute magnetic semiconductors (DMSs) [4-6]. Ferromagnetism at room temperature in these kind
RI PT
of materials can be utilized for various spin based devices. In the case of DMSs, the cations are being partially replaced by magnetic/rare earth atoms. Sufficient experimental data and models are available on transition metals/rare earth doped ZnO ferromagnetic material as suitable DMS candidate for spintronics but the mechanism of magnetism is indecisive [7-10].
SC
Besides the issues of stability and reproducibility, it remains always controversial that whether the ferromagnetism is intrinsic or extrinsic origin such as metal clusters or
M AN U
precipitation of secondary phases [11,12]. The magnetism originating due to cluster formation and phase segregations i.e. due to extrinsic/non-intrinsic behaviour is not applicable for technological applications. These controversies led to the search for suitable DMS that are based on non–magnetic elements. The idea is to introduce ferromagnetism in host lattice via non–magnetic dopant so that the magnetic property is purely intrinsic [13-15]. Such an emergent ferromagnetism due to non–magnetic elements/ defects is often called ‘d0
TE D
ferromagnetism’ where the possibility of origin of magnetism due to partially filled 3d or 4f subshells can be neglected. The existing theories of DMSs cannot be applied on these systems because these were based on d and f shell electrons, but there is no such orbitals in the case of p-shell elements doped materials. Defects (such as oxygen vacancy and zinc
EP
interstitials) are believed to be responsible for initiating hybridization at the Fermi level and establishing a long-range ferromagnetism in such systems. Such kind of defects arises either
AC C
due to the preparation method or via introduction of some non-magnetic ions in the host lattice. Following this idea, room temperature ferromagnetic studies have been carried out using non-magnetic ions (such as Mg, C doped and N) doped ZnO systems. Recently, nitrogen (N) as a non–magnetic dopant has attracted great attention for
inducing ferromagnetic properties in ZnO. Although several theoretical and experimental observations on the ferromagnetism obtained in nitrogen-doped ZnO are available, there are still some debates. Based on theoretical calculations by Lee et al.[16], it is found that O vacancies are the main compensating donors for N acceptors using a normal N2 source at low N doping levels, while Zn antisites are the main compensating donors at high doping levels.
2
ACCEPTED MANUSCRIPT Experimentally, Jindal et al. [17] attributed the ferromagnetism in nitrogen–doped ZnO films to the 2–2 interaction between the incorporated N at the oxygen site and the neighbouring oxygen atoms. There are also the evidences that the kind of carrier (electrons/holes) mediates the magnetic interaction [13]. Another report by Kumar et al. [18] has also suggested that lower dose of N implantation in CeO2 thin films enhances the saturation magnetization while
RI PT
it is reduced for higher dose. Hariwal et al. [19] also studied the role of N ions on ZnO thin films by varying the implantation angles. They also observed the enhancement in saturation magnetization upto implantation angle of 60° and then a reduction in the magnetization for 90°. However, the saturation magnetization for 90° N implanted ZnO ions is still greater than
SC
pure ZnO. Based on these observations, it is evident that N ions play an important role in the modification in magnetic properties. Therefore, it is an interesting and physically rich problem to study the mechanism of magnetism for the nitrogen embedded ZnO thin films.
M AN U
The incorporation of nitrogen in host lattice can be achieved via various ways such as doping, nitrogen gas flow, ion implantation technique etc. as discussed above. However, it becomes important to select the proper technique that is capable of providing the uniform and controlled distribution. Using ion implantation technique, there is not only the possibility of controlled and uniform distribution of dopants (by selection the implantation dose) but it also
TE D
maximizes the creation of certain intrinsic defects which are quite useful for the enhancement in magnetic ordering in oxide semiconductors. Present study aims to identify the origin of ferromagnetism in pure and N implanted ZnO thin films deposited by RF sputtering technique. To fully understand the role of N ions on the origin of ferromagnetism,
EP
experimental magnetization results were analyzed along with the theoretical simulations based on density functional theory (DFT).
AC C
2.1 Experimental Details
ZnO thin films were deposited on Si (100) substrate using RF sputtering technique.
Sputtering was done in Ar gas environment at a substrate temperature of 500 °C and 150 W RF power for 30 min. These films were then implanted with 60 keV N ions at Inter University Accelerator Centre, New Delhi by varying ion fluences (5×1016 and 1×1017 ions/cm2). The crystal structure and morphology were determined by Bruker D8 X-ray diffractometer and Nanoscope-IIIa atomic force microscope (AFM). The optical properties of the
films
were
measured
by
UV-Vis
spectrophotometer
(HITACHI-3300)
and
Photoluminescence (PL) spectroscopy measurements. Further, SQUID magnetometer was used for the magnetic measurements. 3
ACCEPTED MANUSCRIPT
2.2 Simulation Details Apart from above experimental investigation, first principle calculations for density of states (DOS), optical properties (absorption coefficient) and to understand the origin of magnetism within pure and N implanted ZnO thin films were performed using full potential linear augmented plane wave method (FP-LAPW) as implemented in WIEN2k code[20].
RI PT
The exchange-correlation functional has been treated within the generalized gradient approximation (GGA) and parameterized by Perdewe Burkee Ernzerhofer (PBE) potential [21]. The Tran–Blaha-modified Beck and Johnson (TB-mBJ) potential has been used to improve the band gap values of the thin films [22]. In tetrahedral method [23] for the
SC
Brillouin zone integrations 2×2×2 divisions for k-point sampling have used. The cut-off energy, which defines the separation between the core and valence states, was set to -6.0 Ry
M AN U
and the total energy convergence considered approximate 0.00001 Ry. The details about the simulation method of the electronic structure calculation and absorption coefficients have been reported elsewhere [24]. A supercell of 40 atoms was generated for performing this N implanted ZnO.
3.1 X-ray Diffraction
TE D
X-ray diffraction measurements were performed to investigate the effect of N ion implantation on the crystal structure or nature of crystallinity of ZnO thin films. Fig. 1 shows the XRD patterns of ZnO and N implanted ZnO which indicate the formation of highly (002)
EP
oriented films with hexagonal wurtzite structure [25]. No additional diffraction peaks corresponding to other phases are detected. This is attributed to the fact that N ions are incorporated completely into the host lattice site instead of occupying some interstitial
AC C
position. The diffraction pattern shows that there is reduction in the FWHM due to the increase in crystallite sizes. The crystallite size was estimated using the Scherrer’s formula; D = 0.9λ/βCosθ; where D is the crystallite size, λ is the wavelength and β is the FWHM [3]. The variation in crystallite size with N incorporation is shown in Table 1. This might be due to incorporation of N ions into the host lattice which has increased the nucleation of particles which in turn contribute towards the growth of crystallites. Further, it has been observed that the (002) diffraction peak shifts towards lower scattering angle. The systematic shift in the diffraction peak led to change in the c-axis lattice parameter. For pristine ZnO, the c-axis lattice parameter was found to be 5.228 Å and increases (5.244Å for the highest fluence) with the enhancement in N concentration which is consistent with the earlier report on ZnO [17]. This 4
ACCEPTED MANUSCRIPT is attributed to the larger ionic radii of N (1.71 Å) ions as compared to that of O (1.40 Å),
M AN U
SC
RI PT
resulting in the expansion of unit cell.
Fig. 1: XRD Pattern of pristine, N ion implanted ZnO films with fluencies of 5×1016 and 1×1017ions/cm–2. Inset shows the closer view of peak (002). Table 1: Physical parameters obtained from XRD, UV–Vis and magnetic measurements of N implanted ZnO thin films.
c-axis lattice parameter (Å) 5.228
Stress (GPa)
Pristine
Crystall ite Size (nm) 11.91
5×1016
14.17
1×1017
14.43
----
Saturation Magnetizatio n (emu/cm3) 2.445
Coercivit Retentivity y (Hc) (emu/cm3) Oe 126 0.26
5.239
– 0.780
3.07
4.028
193
0.75
5.244
– 0.953
3.04
5.235
52
0.97
EP
Ions/cm2
Band Gap (eV) 3.27
TE D
Ion fluence
To investigate the effect of the N implantation on the stress of ZnO thin films, we
AC C
have used the following relation [10].
c − c0 GPa c0
σ = −453.6
Where c0 (5.228 Å) corresponds to the c-axis lattice constant of ZnO and c is the c-
axis lattice constant for N implanted ZnO thin films estimated from XRD pattern. As shown in Table 1, the induced stress in the films is found to be increasing with implantation and compressive in nature. The compressive stress in the present system and its increment with N content gives supports that N atoms are incorporated into the ZnO lattice by substituting the O sites.
3.2 Atomic Force Microscopy 5
ACCEPTED MANUSCRIPT AFM image estimates the surface roughness and to study the topography of thin films. Fig. 2 shows the AFM images of ZnO and N implanted ZnO thin films over an area of 2×2 µm2. All these films exhibit dense microstructure with nm sized grains [25]. The grain size increases with the N ion fluences. However, the grain size of the films obtained from AFM is greater than that one determined from XRD. This is probably due to the fact that the grain
RI PT
size measured from AFM is the surface morphology of coalesced grains. On the other hand, the increase in N concentration results in the increase of the surface roughness which might
M AN U
SC
be related with enhanced stress as observed through XRD measurements.
3.3 Bandgap Tuning
TE D
Fig. 2: AFM images of (a) ZnO and N implanted ZnO thin films, (b) 5×1016 and (c) 1×1017 ions/cm–2.
The optical reflectance spectra of pristine and N implanted ZnO thin films were taken over the spectral range 300-600 nm. The Tauc’s procedure was employed to evaluate the
EP
optical band gap from (αhν)2 versus the photon energy (hν) curves as shown in Fig. 3(a) [25]. The band gap of the films was estimated by extrapolating the linear portion along the x-axis.
AC C
The optical bandgap for pristine ZnO film was found to be ~ 3.27 eV and it decreases to ~ 3.07 eV, ~ 3.04 eV respectively for N implanted ZnO thin films with 5×1016 and 1×1017ions/cm–2. The rapid decrease in the bandgap with the introduction of nitrogen in ZnO indicates that N ions play a significant role in modulating the bandgap. The shrinkage of optical bandgap is probably related with the increase in lattice constant and the difference in ionicity. The increase in lattice constant generates a stress with the incorporation of N ions in host lattice. It has been reported that expansion in lattice constant provides a narrow band gap due to the increased repulsion among O 2p and Zn 4s bands and vice-versa [26]. Further, it is well known that N (3.0) ions exhibits smaller the electronegativity as compared to that of O (3.5) ions, so the Zn–N bond has smaller ionicity as compared to Zn–O bond [27]. The 6
ACCEPTED MANUSCRIPT mixing of shallower N 2p states with the valence band of ZnO also leads to a reduction in the band gap [16,28]. Further increasing the N ion fluence i.e. 1×1017ions/cm–2 in ZnO host lattice, the bandgap shifts to a lower value of about ~ 3.04 eV, which shows that more O sites
SC
RI PT
are substituted by the N atoms to form Zn–N bonds.
M AN U
Figure 3: (a) Experimental Tauc’s plots for pristine, 5×1016 and 1×1017 ions/cm–2 N implanted ZnO thin films, (b) Theoretical Tauc’s plots of pristine and 1×1017 ions/cm–2 N implanted ZnO thin films, (c) Photoluminescence spectroscopy measurement for pure and implanted thin films. To compare the experimental band gap values, the WIEN2k (FP-LAPW) code has been used to simulate the band gap of pure ZnO and N implanted ZnO thin films. The
TE D
simulated Tauc’s plot from the absorption coefficient is shown in Fig. 3(b). These simulation data confirm the experimental values of optical bandgap. Theoretically, the estimated band gap values are 2.94 eV and 2.33 eV for ZnO and N implanted ZnO thin films respectively with scissor correction of values 0.35 eV, which is closer to the experimental band gap value
EP
and in good agreement with previous reported results [20,21]. It is known that the density functional theory (DFT) based calculations with generalized gradient approximation (GGA)
AC C
in Perdew-Burke-Ernzerhof (PBE) function overestimate the values of lattice constants and underestimate the values of electronic band gap compared with the experimental values [23, 24]. Since we also made use of GGA-PBE, the observed values of theoretical band gap estimated from Tauc’s plots are found to have slight variation in comparison with the experimental values. However, the trend is found to be similar to the experimental observations [23]. In order to further confirm the variation in band gap and understanding the role of defects towards magnetization, photoluminescence measurements were performed and are shown in Fig. 3(c). PL measurements infer that there is no significant variation in defect concentration after N-implantation, but band gap reduces and the values are similar to that observed through UV-Vis spectroscopy. In fact, Hariwal et al. [19] reported that addition of 7
ACCEPTED MANUSCRIPT N ions induces Zn-N bonds. Because of this, the band edge comes closer to each other and the fermi level (Ef) shifts towards the valence band (VB) which in turn causes the decrease in bandgap. In fact, Hariwal et al. [19] reported that addition of N ions induces Zn-N bonds and because of this the band edge comes closer to each other and the fermi level (Ef) shifts towards the valence band (VB) which in turn causes the decrease in bandgap. Recently, based
RI PT
on photoluminescence and magnetic hysteresis loops in the case of C implanted ZnO nanowires, Wang et al. [31] have suggested that the implantation of C reduces the number of intrinsic surface defects present and increases the saturation magnetization of ZnO nanowires. In another report by Jindal et al. [32], authors were able to achieve ferromagnetism for the
SC
highest current density for which the grown films have highest crystallinity (XRD) and minimal defects (PL measurements); they argued that Zn-N bonding plays dominant role compared to defects after incorporation of N ions into ZnO lattice. Therefore based on the
M AN U
existing literature and cited reports, it is quite clear that defects and the site occupancy by the dopants play significant role in these system.
3.4 Magnetic Properties
Magnetic properties of pristine and N implanted ZnO thin films are shown in Fig. 4.
TE D
Pristine film, ZnO, exhibits a weak ferromagnetic signal. The N ion implantation in ZnO leads to an enhancement in ferromagnetic behaviour which further increases with the ion fluences. For ZnO thin films, the value of saturation magnetization is estimated to be ~ 2.45 emu/cm3 whereas the N implanted ZnO thin film with fluence of 5×1016 ions/cm–2 exhibits the
EP
saturation magnetization of ~ 4.02×102 emu/cm3 which is almost 1.6 times higher than ZnO film. By further increasing the N ion implantation i.e. for 1×1017 ions/cm–2, the saturation
AC C
magnetization enhances to ~ 5.24 emu/cm3 indicating that the ferromagnetism has strong correlation with the N ion fluence. The observed magnetic behaviour is consistent with the earlier reports on ferromagnetism in N doped ZnO thin films [13]. It has been proposed by Jindal et al. [32] that doping with N at the O sites in host ZnO lattice is essential to induce the room temperature ferromagnetism in ZnO in contrast to the presence of intrinsic structural defects and crystallinity. They have been able to achieve maximum ferromagnetism for the highest current density for which the grown films have highest crystallinity (XRD) and minimal defects (PL measurements) [32]. Banerjee et al. [33] have reported that ferromagnetism increases and is attributed to the formation of oxygen vacancy clusters upon thermal annealing in ZnO. Gao et al. [34] have also reported that the ferromagnetism of ZnO 8
ACCEPTED MANUSCRIPT nanoparticles increases after annealing in vacuum condition and decreases after annealing in rich-oxygen atmosphere. Therefore, based on the existing literature and cited reports, it can be argued that defects and the site occupancy by the dopants in the host lattice plays a significant role in these system. In the present case, the saturation magnetisation has also been observed in pure ZnO thin films. There are number of reports in the literature on the
RI PT
existence of room temperature ferromagnetism in pure ZnO nanoparticles, nanostructures as well as in thin films and attributed to the various kind of defects (Zn vacancies, O vacancies, Zn interstitials, etc.) present in the lattice [35,36]. Khalid et al. [35] have reported the magnetism in pure ZnO thin films have been associated with the Zn vacancies present in the
SC
host ZnO. Our group has also previously reported the presence of magnetism in pure ZnO thin films and was found to be attributed to the presence of oxygen vacancies [9,37]. Based on the PL results, it can be understood that saturation magnetisation in pure ZnO thin films in
M AN U
the present case is associated with the presence of oxygen vacancies. The ferromagnetic behaviour in N implanted ZnO thin films further can be associated either with the presence of various kinds of defects, site occupation by N ions or secondary phase formation. The possibility of extrinsic ferromagnetism due to secondary phase can be excluded on the basis of XRD results. It has been experimentally proved that nitrogen can occupy three different
TE D
sites in the host lattice: it can occupy interstitial site, Zn and or oxygen sites. Earlier reports have shown that nitrogen at the interstitial and Zn sites did not contribute in the ferromagnetic behaviour [38]. On the other hand, the N substitution on O site favours the spin
AC C
EP
polarized state and results in the enhancement of ferromagnetism in ZnO.
Fig. 4: Magnetization curves of pristine, N ion implanted ZnO thin films: 5×1016 ions/cm2 and 1×1017 ions/cm2.
9
ACCEPTED MANUSCRIPT Total energy calculations of ZnO and N implanted (≈ 4.5 %) ZnO were performed with spin polarization using GGA (PBE) and TB-mBJ potential. The total and partial density of states (PDOS) for both majority and minority spin states of ZnO and N implanted (≈ 4.5 %) ZnO between −8 eV and + 8 eV is shown in the Fig. 5 (a–d) [20-23]. In this the dotted line represents the zero line Fermi energy (EF). In general, the electronic properties of spin
RI PT
polarized ZnO are influenced by the presence of O vacancy (VO) and Zn vacancy (VZn) in the system, resulting in weak magnetization [39,40]. It has been reported that addition of N ions increases the 2p-states in the valence band as a result of increase in the Zn–N bonds that leads to the reduction in VO in the host lattice [19]. Another report has shown that the doping
SC
of nitrogen at oxygen site in ZnO lattice is essential for the enhancement in room temperature magnetism in contrast to the presence of intrinsic structural defects [32]. Furthermore, Ye et al. [13] proposed a plausible mechanism for explaining the magnetic behaviour via doping of
M AN U
2p shell electrons (i.e. C atoms). It has been suggested that partial substitution of C ions over O site causes bonding with Zn ions Zn3d-C2p hybridization. Due to this, there is transfer of electrons of Zn ions (changes Zn d10 (completely filled) to d9 (incompletely filled) state) towards p-orbital of C ions, resulting in the enhancement in the ferromagnetic character with C doping. Our group has also reported that the dilute concentration of C ions on O site into
TE D
ZnO lattice increases ferromagnetism. Based on these observations, it can be concluded that the resultant magnetism increased when N ions are substituted in the place of O ions only in ZnO lattice [41]. The theoretical calculation demonstrates that N incorporation at Zn sites or at the interstitial sites (NI) did not enhance any ferromagnetism in ZnO. The effect of N
EP
substitution at O sites in ZnO super cell lattice favored spin polarization state and increased the magnetic moment. As the N-2p states are more delocalized compared to 3d or 4f states of
AC C
transition metal and also have long range coupling capacity, so below the Fermi level N-2p hybridizes with O-2p that results in the redistribution of spin [27]. Total spin magnetic moment per supercell in the spin polarization calculation is 4.008 µ B for ZnO and it increases to 5.012 µ B for N implanted (≈4.5 %) ZnO lattice which are in good agreement with reported simulation results [39,41]. From the Fig.5, it can be observed that near the Fermi level, N-2p overlap fully with those of O-2p, indicating a strong interaction between them. The spin up orbitals opening the band gap and spin down are partially filled. Furthermore, it also found that in the range from −6.0 eV to −4.8 eV, Zn-3d strong spin polarized and hybridizes mainly with O-2p states. Addition of N3- ions on O2- site of ZnO lattice create impurity moments and additional hole via the following relation N3- +e- → O2-[17]. These holes in the valence band of O 2p states act as main charge carriers which interact with N 2p spins resulting in 2p-2p 10
ACCEPTED MANUSCRIPT interaction. The spin density near each anion impurity tries to follow the impurity ion magnetic moments via 2p-2p interaction. The presence of large number of spin polarized charge carriers lead to indirect long range ferromagnetic ordering between dopants [27]. Similar to 2p-3d interaction, 2p-2p interactions also lead to the creation of additional mixed states in forbidden band gap. So, an enhancement of magnetic moment in ZnO:N is achieved
TE D
M AN U
SC
and O−2p states at the valence band maximum.
RI PT
through 2p−3d hybridization-like 2p−2p coupling interaction from anion sites of N−2p states
EP
Fig. 5 (a-b) Total and partial density of states of spin polarised ZnO super cell respectively; (c-d) total and partial density of states of spin polarised N implanted ZnO super cell respectively.
AC C
A comparison of optical bandgap and saturation magnetization as a function of N ion fluences is investigated. The incorporation of N ion in host lattice site results in the reduction of bandgap whereas there is an enhancement in saturation magnetization. It is expected that the bandgap decreases due to an increase in lattice constant (c-axis) by the substitution of N ions at O site which has been confirmed by X-ray diffraction technique. The increase in lattice constant decreases the coulomb interaction among the atoms, giving rise to decrease in bandgap. Further, a narrow bandgap enhances the ferromagnetic character in host ZnO through N incorporation which in turn is related with the exchange interaction. When the bandgap reduces, the impurity band shifts toward the lower energy side and overlap with the 3d states that result in the enhancement of magnetic exchange interaction.
11
ACCEPTED MANUSCRIPT
4. Conclusions In conclusion, this study demonstrated that N ions implantation plays an important role in the enhancement of ferromagnetic properties of ZnO thin films. XRD diffraction pattern shows the increase in the c-axis lattice parameter with the incorporation of N into
RI PT
ZnO host lattice. The mixing of shallower N 2p states with the valence band of ZnO leads to a reduced band gap which is in accordance with the experimental results obtained in the present work using UV-Vis spectroscopy. Further, the reduction in the band gap results in the enhancement of ferromagnetic behaviour in the present system which is mainly because of
SC
the induced hybridization of N ions at the O lattice sites.
M AN U
Acknowledgement
Author (PK) would like to acknowledge Department of Science and Technology for financial support via DST-INSPIRE Faculty Scheme [No. DST/INSPIRE/04/2015/003149].
References [1]
Ü. Özgür, Y.I. Alivov, C. Liu, A. Teke, M.A. Reshchikov, S. Doğan, V. Avrutin, S.-J. Cho, H. Morkoç, Journal of Applied Physics 98 (2005) 041301. P. Kumar, J.P. Singh, Y. Kumar, A. Gaur, H.K. Malik, K. Asokan, Current Applied
TE D
[2]
Physics 12 (2012) 1166. [3]
P. Kumar, V. Singh, V. Sharma, G. Rana, H.K. Malik, K. Asokan, Ceramics
EP
International 41 (2015) 6734. [4]
J.M.D. Coey, Solid State Sciences 7 (2005) 660.
[5]
H. Peng, H.J. Xiang, S.-H. Wei, S.-S. Li, J.-B. Xia, J. Li, Physical Review Letters 102
[6] [7]
AC C
(2009) 017201.
J.M.D. Coey, M. Venkatesan, C.B. Fitzgerald, Nat Mater 4 (2005) 173. P. Kumar, Y. Kumar, H.K. Malik, S. Annapoorni, S. Gautam, K.H. Chae, K. Asokan,
Applied Physics A 114 (2014) 453.
[8]
K. Ueda, H. Tabata, T. Kawai, Applied Physics Letters 79 (2001) 988.
[9]
K. Parmod, K.M. Hitendra, K. Asokan, EPL (Europhysics Letters) 110 (2015) 67006.
[10]
P. Kumar, V. Sharma, A. Sarwa, A. Kumar, Surbhi, R. Goyal, K. Sachdev, S. Annapoorni, K. Asokan, D. Kanjilal, RSC Advances 6 (2016) 89242.
[11]
J.H. Park, M.G. Kim, H.M. Jang, S. Ryu, Y.M. Kim, Applied Physics Letters 84 (2004) 1338. 12
ACCEPTED MANUSCRIPT [12]
T.C. Kaspar, T. Droubay, S.M. Heald, M.H. Engelhard, P. Nachimuthu, S.A. Chambers, Physical Review B 77 (2008) 201303.
[13]
X.J. Ye, H.A. Song, W. Zhong, M.H. Xu, X.S. Qi, C.Q. Jin, Z.X. Yang, C.T. Au, Y.W. Du, Journal of Physics D: Applied Physics 41 (2008) 155005.
[14]
S. Zhou, Q. Xu, K. Potzger, G. Talut, R. Grötzschel, J. Fassbender, M. Vinnichenko,
RI PT
J. Grenzer, M. Helm, H. Hochmuth, M. Lorenz, M. Grundmann, H. Schmidt, Applied Physics Letters 93 (2008) 232507. [15]
R.P. Borges, R.C.d. Silva, S. Magalhães, M.M. Cruz, M. Godinho, Journal of Physics: Condensed Matter 19 (2007) 476207.
E.-C. Lee, Y.S. Kim, Y.G. Jin, K.J. Chang, Physical Review B 64 (2001) 085120.
[17]
K. Jindal, M. Tomar, R.S. Katiyar, V. Gupta, Journal of Applied Physics 111 (2012)
SC
[16]
102805.
P. Kumar, P. Kumar, A. Kumar, I. Sulania, F. Chand, K. Asokan, RSC Advances 7
M AN U
[18]
(2017) 9160. [19]
R. Hariwal, H.K. Malik, A. Negi, A. Kandasami, RSC Advances 8 (2018) 6278.
[20]
P. Blaha, K. Schwarz, G. Madsen, D. Kvasnicka, J. Luitz, An augmented plane wave+ local orbitals program for calculating crystal properties (2001). G. Kresse, D. Joubert, Physical Review B 59 (1999) 1758.
[22]
S.L. Dudarev, G.A. Botton, S.Y. Savrasov, C.J. Humphreys, A.P. Sutton, Physical
TE D
[21]
Review B 57 (1998) 1505.
H.J. Monkhorst, J.D. Pack, Physical Review B 13 (1976) 5188.
[24]
A. Ghosh, R. Thangavel, M. Rajagopalan, Journal of Materials Science 48 (2013) 8259.
P. Kumar, H.K. Malik, A. Ghosh, R. Thangavel, K. Asokan, Applied Physics Letters
AC C
[25]
EP
[23]
102 (2013) 221903.
[26] [27]
M. Hirai, A. Kumar, Journal of Vacuum Science & Technology A 25 (2007) 1534.
L. Shen, R.Q. Wu, H. Pan, G.W. Peng, M. Yang, Z.D. Sha, Y.P. Feng, Physical
Review B 78 (2008) 073306.
[28]
H.T. Chang, G.J. Chen, Thin Solid Films 618 (2016) 84.
[29]
P. Haas, F. Tran, P. Blaha, Physical Review B 79 (2009) 209902.
[30]
P. Hao, Y. Fang, J. Sun, G.I. Csonka, P.H.T. Philipsen, J.P. Perdew, Physical Review B 85 (2012) 014111.
[31]
Y.F. Wang, Y.C. Shao, S.H. Hsieh, Y.K. Chang, P.H. Yeh, H.C. Hsueh, J.W. Chiou, H.T. Wang, S.C. Ray, H.M. Tsai, C.W. Pao, C.H. Chen, H.J. Lin, J.F. Lee, C.T. Wu, 13
ACCEPTED MANUSCRIPT J.J. Wu, Y.M. Chang, K. Asokan, K.H. Chae, T. Ohigashi, Y. Takagi, T. Yokoyama, N. Kosugi, W.F. Pong, Scientific reports 8 (2018) 7758. [32]
K. Jindal, M. Tomar, R.S. Katiyar, V. Gupta, AIP Advances 5 (2015) 027117.
[33]
S. Banerjee, M. Mandal, N. Gayathri, M. Sardar, Applied Physics Letters 91 (2007) 182501. D. Gao, Z. Zhang, J. Fu, Y. Xu, J. Qi, D. Xue, Journal of Applied Physics 105 (2009)
RI PT
[34]
113928. [35]
M. Khalid, M. Ziese, A. Setzer, P. Esquinazi, M. Lorenz, H. Hochmuth, M. Grundmann, D. Spemann, T. Butz, G. Brauer, W. Anwand, G. Fischer, W.A.
[36]
SC
Adeagbo, W. Hergert, A. Ernst, Physical Review B 80 (2009) 035331.
D. Wang, Z.Q. Chen, D.D. Wang, N. Qi, J. Gong, C.Y. Cao, Z. Tang, Journal of Applied Physics 107 (2010) 023524.
P. Kumar, H.K. Malik, S. Gautam, K.H. Chae, K. Asokan, D. Kanjilal, Journal of
M AN U
[37]
Alloys and Compounds 710 (2017) 831. [38]
T.S. Herng, S.P. Lau, S.F. Yu, H.Y. Yang, X.H. Ji, J.S. Chen, N. Yasui, H. Inaba, Journal of Applied Physics 99 (2006) 086101.
[39]
I.S. Elfimov, A. Rusydi, S.I. Csiszar, Z. Hu, H.H. Hsieh, H.J. Lin, C.T. Chen, R.
TE D
Liang, G.A. Sawatzky, Physical Review Letters 98 (2007) 137202. [40]
A. Zunger, S. Lany, H. Raebiger, Physics 3 (2010) 53.
[41]
J. Kumar, S. Ramasubramanian, R. Thangavel, M. Rajagopalan, in: R.M.S. Rao, T. Okada (Eds.), ZnO Nanocrystals and Allied Materials, Springer India, New Delhi,
AC C
EP
2014, p. 309.
14
ACCEPTED MANUSCRIPT
Highlights 1. Influence of N implantation on optical and magnetic properties of ZnO thin films.
3. Increase in saturation magnetization with N content.
AC C
EP
TE D
M AN U
SC
4. Decrease in optical bandgap with increasing N conc.
RI PT
2. Experimental & DFT calculations have been performed to understand the mechanism.
S0925-8388(18)32597-0
DOI:
10.1016/j.jallcom.2018.07.097
Reference:
JALCOM 46813
To appear in:
Journal of Alloys and Compounds
Received Date: 28 February 2018 Revised Date:
25 June 2018
Accepted Date: 9 July 2018
Please cite this article as: P. Kumar, H.K. Malik, A. Ghosh, R. Thangavel, K. Asokan, An insight to origin of ferromagnetism in ZnO and N implanted ZnO thin films: Experimental and DFT approach, Journal of Alloys and Compounds (2018), doi: 10.1016/j.jallcom.2018.07.097. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
An Insight to origin of ferromagnetism in ZnO and N implanted ZnO thin films: Experimental and DFT approach Parmod Kumar1,2,*, Hitendra K. Malik2,*, Anima Ghosh3, R. Thangavel3 and K. Asokan2 1
Materials Science Division, Inter University Accelerator Centre, New Delhi – 110067, India Department of Physics, Indian Institute of Technology Delhi, New Delhi – 110016, India 3 Department of Applied Physics, Indian Institute of Technology (Indian School of Mines), Dhanbad– 826004, India
RI PT
2
M AN U
SC
In the present study, we have elucidated an effective way to simultaneously tune the optical bandgap and magnetic properties of zinc oxide (ZnO). This can be achieved by the incorporation of nitrogen ions via any means in the host matrix. On a broader way, we have systematically investigated the influence of N ions in ZnO thin films through experimental techniques and density functional theory (DFT) calculations to understand the physical mechanism governing the observed magnetic behaviour and variation in bandgap. For this, RF sputtered ZnO thin films deposited over Si substrates were implanted with N ions by varying the fluences and studied for their optical and magnetic properties. The pristine ZnO films is having saturation magnetization of ~ 2.45 emu/cm3 which becomes almost twice compared for the fluence of 1×1017 ions/cm2. Furthermore, the optical bandgap is tuned from 3.27 eV to 3.04 eV with N ion fluences. Our work signifies the new insight to understand the basis of ferromagnetism in non-magnetic ions doped ZnO system.
Corresponding authors:
TE D
Keywords: Ion implantation, bandgap tuning, defects induced magnetism, first principle calculations.
*Email: [email protected], [email protected] Phone: +91-11-26893955, Fax: + 91-11-26893666
AC C
EP
Present Affiliation of Corresponding author: Department of Physics, Deenbandhu Chhotu Ram University of Science and Technology, Murthal, Sonepat-131039, India
1
ACCEPTED MANUSCRIPT
1. Introduction ZnO has remarkable properties such as large binding energy (60 meV) and wide band gap of 3.37 eV which makes it an excellent material for excitonic devices [1-3]. However, in recent years there has been significant interest in the magnetic behaviour of ZnO i.e. dilute magnetic semiconductors (DMSs) [4-6]. Ferromagnetism at room temperature in these kind
RI PT
of materials can be utilized for various spin based devices. In the case of DMSs, the cations are being partially replaced by magnetic/rare earth atoms. Sufficient experimental data and models are available on transition metals/rare earth doped ZnO ferromagnetic material as suitable DMS candidate for spintronics but the mechanism of magnetism is indecisive [7-10].
SC
Besides the issues of stability and reproducibility, it remains always controversial that whether the ferromagnetism is intrinsic or extrinsic origin such as metal clusters or
M AN U
precipitation of secondary phases [11,12]. The magnetism originating due to cluster formation and phase segregations i.e. due to extrinsic/non-intrinsic behaviour is not applicable for technological applications. These controversies led to the search for suitable DMS that are based on non–magnetic elements. The idea is to introduce ferromagnetism in host lattice via non–magnetic dopant so that the magnetic property is purely intrinsic [13-15]. Such an emergent ferromagnetism due to non–magnetic elements/ defects is often called ‘d0
TE D
ferromagnetism’ where the possibility of origin of magnetism due to partially filled 3d or 4f subshells can be neglected. The existing theories of DMSs cannot be applied on these systems because these were based on d and f shell electrons, but there is no such orbitals in the case of p-shell elements doped materials. Defects (such as oxygen vacancy and zinc
EP
interstitials) are believed to be responsible for initiating hybridization at the Fermi level and establishing a long-range ferromagnetism in such systems. Such kind of defects arises either
AC C
due to the preparation method or via introduction of some non-magnetic ions in the host lattice. Following this idea, room temperature ferromagnetic studies have been carried out using non-magnetic ions (such as Mg, C doped and N) doped ZnO systems. Recently, nitrogen (N) as a non–magnetic dopant has attracted great attention for
inducing ferromagnetic properties in ZnO. Although several theoretical and experimental observations on the ferromagnetism obtained in nitrogen-doped ZnO are available, there are still some debates. Based on theoretical calculations by Lee et al.[16], it is found that O vacancies are the main compensating donors for N acceptors using a normal N2 source at low N doping levels, while Zn antisites are the main compensating donors at high doping levels.
2
ACCEPTED MANUSCRIPT Experimentally, Jindal et al. [17] attributed the ferromagnetism in nitrogen–doped ZnO films to the 2–2 interaction between the incorporated N at the oxygen site and the neighbouring oxygen atoms. There are also the evidences that the kind of carrier (electrons/holes) mediates the magnetic interaction [13]. Another report by Kumar et al. [18] has also suggested that lower dose of N implantation in CeO2 thin films enhances the saturation magnetization while
RI PT
it is reduced for higher dose. Hariwal et al. [19] also studied the role of N ions on ZnO thin films by varying the implantation angles. They also observed the enhancement in saturation magnetization upto implantation angle of 60° and then a reduction in the magnetization for 90°. However, the saturation magnetization for 90° N implanted ZnO ions is still greater than
SC
pure ZnO. Based on these observations, it is evident that N ions play an important role in the modification in magnetic properties. Therefore, it is an interesting and physically rich problem to study the mechanism of magnetism for the nitrogen embedded ZnO thin films.
M AN U
The incorporation of nitrogen in host lattice can be achieved via various ways such as doping, nitrogen gas flow, ion implantation technique etc. as discussed above. However, it becomes important to select the proper technique that is capable of providing the uniform and controlled distribution. Using ion implantation technique, there is not only the possibility of controlled and uniform distribution of dopants (by selection the implantation dose) but it also
TE D
maximizes the creation of certain intrinsic defects which are quite useful for the enhancement in magnetic ordering in oxide semiconductors. Present study aims to identify the origin of ferromagnetism in pure and N implanted ZnO thin films deposited by RF sputtering technique. To fully understand the role of N ions on the origin of ferromagnetism,
EP
experimental magnetization results were analyzed along with the theoretical simulations based on density functional theory (DFT).
AC C
2.1 Experimental Details
ZnO thin films were deposited on Si (100) substrate using RF sputtering technique.
Sputtering was done in Ar gas environment at a substrate temperature of 500 °C and 150 W RF power for 30 min. These films were then implanted with 60 keV N ions at Inter University Accelerator Centre, New Delhi by varying ion fluences (5×1016 and 1×1017 ions/cm2). The crystal structure and morphology were determined by Bruker D8 X-ray diffractometer and Nanoscope-IIIa atomic force microscope (AFM). The optical properties of the
films
were
measured
by
UV-Vis
spectrophotometer
(HITACHI-3300)
and
Photoluminescence (PL) spectroscopy measurements. Further, SQUID magnetometer was used for the magnetic measurements. 3
ACCEPTED MANUSCRIPT
2.2 Simulation Details Apart from above experimental investigation, first principle calculations for density of states (DOS), optical properties (absorption coefficient) and to understand the origin of magnetism within pure and N implanted ZnO thin films were performed using full potential linear augmented plane wave method (FP-LAPW) as implemented in WIEN2k code[20].
RI PT
The exchange-correlation functional has been treated within the generalized gradient approximation (GGA) and parameterized by Perdewe Burkee Ernzerhofer (PBE) potential [21]. The Tran–Blaha-modified Beck and Johnson (TB-mBJ) potential has been used to improve the band gap values of the thin films [22]. In tetrahedral method [23] for the
SC
Brillouin zone integrations 2×2×2 divisions for k-point sampling have used. The cut-off energy, which defines the separation between the core and valence states, was set to -6.0 Ry
M AN U
and the total energy convergence considered approximate 0.00001 Ry. The details about the simulation method of the electronic structure calculation and absorption coefficients have been reported elsewhere [24]. A supercell of 40 atoms was generated for performing this N implanted ZnO.
3.1 X-ray Diffraction
TE D
X-ray diffraction measurements were performed to investigate the effect of N ion implantation on the crystal structure or nature of crystallinity of ZnO thin films. Fig. 1 shows the XRD patterns of ZnO and N implanted ZnO which indicate the formation of highly (002)
EP
oriented films with hexagonal wurtzite structure [25]. No additional diffraction peaks corresponding to other phases are detected. This is attributed to the fact that N ions are incorporated completely into the host lattice site instead of occupying some interstitial
AC C
position. The diffraction pattern shows that there is reduction in the FWHM due to the increase in crystallite sizes. The crystallite size was estimated using the Scherrer’s formula; D = 0.9λ/βCosθ; where D is the crystallite size, λ is the wavelength and β is the FWHM [3]. The variation in crystallite size with N incorporation is shown in Table 1. This might be due to incorporation of N ions into the host lattice which has increased the nucleation of particles which in turn contribute towards the growth of crystallites. Further, it has been observed that the (002) diffraction peak shifts towards lower scattering angle. The systematic shift in the diffraction peak led to change in the c-axis lattice parameter. For pristine ZnO, the c-axis lattice parameter was found to be 5.228 Å and increases (5.244Å for the highest fluence) with the enhancement in N concentration which is consistent with the earlier report on ZnO [17]. This 4
ACCEPTED MANUSCRIPT is attributed to the larger ionic radii of N (1.71 Å) ions as compared to that of O (1.40 Å),
M AN U
SC
RI PT
resulting in the expansion of unit cell.
Fig. 1: XRD Pattern of pristine, N ion implanted ZnO films with fluencies of 5×1016 and 1×1017ions/cm–2. Inset shows the closer view of peak (002). Table 1: Physical parameters obtained from XRD, UV–Vis and magnetic measurements of N implanted ZnO thin films.
c-axis lattice parameter (Å) 5.228
Stress (GPa)
Pristine
Crystall ite Size (nm) 11.91
5×1016
14.17
1×1017
14.43
----
Saturation Magnetizatio n (emu/cm3) 2.445
Coercivit Retentivity y (Hc) (emu/cm3) Oe 126 0.26
5.239
– 0.780
3.07
4.028
193
0.75
5.244
– 0.953
3.04
5.235
52
0.97
EP
Ions/cm2
Band Gap (eV) 3.27
TE D
Ion fluence
To investigate the effect of the N implantation on the stress of ZnO thin films, we
AC C
have used the following relation [10].
c − c0 GPa c0
σ = −453.6
Where c0 (5.228 Å) corresponds to the c-axis lattice constant of ZnO and c is the c-
axis lattice constant for N implanted ZnO thin films estimated from XRD pattern. As shown in Table 1, the induced stress in the films is found to be increasing with implantation and compressive in nature. The compressive stress in the present system and its increment with N content gives supports that N atoms are incorporated into the ZnO lattice by substituting the O sites.
3.2 Atomic Force Microscopy 5
ACCEPTED MANUSCRIPT AFM image estimates the surface roughness and to study the topography of thin films. Fig. 2 shows the AFM images of ZnO and N implanted ZnO thin films over an area of 2×2 µm2. All these films exhibit dense microstructure with nm sized grains [25]. The grain size increases with the N ion fluences. However, the grain size of the films obtained from AFM is greater than that one determined from XRD. This is probably due to the fact that the grain
RI PT
size measured from AFM is the surface morphology of coalesced grains. On the other hand, the increase in N concentration results in the increase of the surface roughness which might
M AN U
SC
be related with enhanced stress as observed through XRD measurements.
3.3 Bandgap Tuning
TE D
Fig. 2: AFM images of (a) ZnO and N implanted ZnO thin films, (b) 5×1016 and (c) 1×1017 ions/cm–2.
The optical reflectance spectra of pristine and N implanted ZnO thin films were taken over the spectral range 300-600 nm. The Tauc’s procedure was employed to evaluate the
EP
optical band gap from (αhν)2 versus the photon energy (hν) curves as shown in Fig. 3(a) [25]. The band gap of the films was estimated by extrapolating the linear portion along the x-axis.
AC C
The optical bandgap for pristine ZnO film was found to be ~ 3.27 eV and it decreases to ~ 3.07 eV, ~ 3.04 eV respectively for N implanted ZnO thin films with 5×1016 and 1×1017ions/cm–2. The rapid decrease in the bandgap with the introduction of nitrogen in ZnO indicates that N ions play a significant role in modulating the bandgap. The shrinkage of optical bandgap is probably related with the increase in lattice constant and the difference in ionicity. The increase in lattice constant generates a stress with the incorporation of N ions in host lattice. It has been reported that expansion in lattice constant provides a narrow band gap due to the increased repulsion among O 2p and Zn 4s bands and vice-versa [26]. Further, it is well known that N (3.0) ions exhibits smaller the electronegativity as compared to that of O (3.5) ions, so the Zn–N bond has smaller ionicity as compared to Zn–O bond [27]. The 6
ACCEPTED MANUSCRIPT mixing of shallower N 2p states with the valence band of ZnO also leads to a reduction in the band gap [16,28]. Further increasing the N ion fluence i.e. 1×1017ions/cm–2 in ZnO host lattice, the bandgap shifts to a lower value of about ~ 3.04 eV, which shows that more O sites
SC
RI PT
are substituted by the N atoms to form Zn–N bonds.
M AN U
Figure 3: (a) Experimental Tauc’s plots for pristine, 5×1016 and 1×1017 ions/cm–2 N implanted ZnO thin films, (b) Theoretical Tauc’s plots of pristine and 1×1017 ions/cm–2 N implanted ZnO thin films, (c) Photoluminescence spectroscopy measurement for pure and implanted thin films. To compare the experimental band gap values, the WIEN2k (FP-LAPW) code has been used to simulate the band gap of pure ZnO and N implanted ZnO thin films. The
TE D
simulated Tauc’s plot from the absorption coefficient is shown in Fig. 3(b). These simulation data confirm the experimental values of optical bandgap. Theoretically, the estimated band gap values are 2.94 eV and 2.33 eV for ZnO and N implanted ZnO thin films respectively with scissor correction of values 0.35 eV, which is closer to the experimental band gap value
EP
and in good agreement with previous reported results [20,21]. It is known that the density functional theory (DFT) based calculations with generalized gradient approximation (GGA)
AC C
in Perdew-Burke-Ernzerhof (PBE) function overestimate the values of lattice constants and underestimate the values of electronic band gap compared with the experimental values [23, 24]. Since we also made use of GGA-PBE, the observed values of theoretical band gap estimated from Tauc’s plots are found to have slight variation in comparison with the experimental values. However, the trend is found to be similar to the experimental observations [23]. In order to further confirm the variation in band gap and understanding the role of defects towards magnetization, photoluminescence measurements were performed and are shown in Fig. 3(c). PL measurements infer that there is no significant variation in defect concentration after N-implantation, but band gap reduces and the values are similar to that observed through UV-Vis spectroscopy. In fact, Hariwal et al. [19] reported that addition of 7
ACCEPTED MANUSCRIPT N ions induces Zn-N bonds. Because of this, the band edge comes closer to each other and the fermi level (Ef) shifts towards the valence band (VB) which in turn causes the decrease in bandgap. In fact, Hariwal et al. [19] reported that addition of N ions induces Zn-N bonds and because of this the band edge comes closer to each other and the fermi level (Ef) shifts towards the valence band (VB) which in turn causes the decrease in bandgap. Recently, based
RI PT
on photoluminescence and magnetic hysteresis loops in the case of C implanted ZnO nanowires, Wang et al. [31] have suggested that the implantation of C reduces the number of intrinsic surface defects present and increases the saturation magnetization of ZnO nanowires. In another report by Jindal et al. [32], authors were able to achieve ferromagnetism for the
SC
highest current density for which the grown films have highest crystallinity (XRD) and minimal defects (PL measurements); they argued that Zn-N bonding plays dominant role compared to defects after incorporation of N ions into ZnO lattice. Therefore based on the
M AN U
existing literature and cited reports, it is quite clear that defects and the site occupancy by the dopants play significant role in these system.
3.4 Magnetic Properties
Magnetic properties of pristine and N implanted ZnO thin films are shown in Fig. 4.
TE D
Pristine film, ZnO, exhibits a weak ferromagnetic signal. The N ion implantation in ZnO leads to an enhancement in ferromagnetic behaviour which further increases with the ion fluences. For ZnO thin films, the value of saturation magnetization is estimated to be ~ 2.45 emu/cm3 whereas the N implanted ZnO thin film with fluence of 5×1016 ions/cm–2 exhibits the
EP
saturation magnetization of ~ 4.02×102 emu/cm3 which is almost 1.6 times higher than ZnO film. By further increasing the N ion implantation i.e. for 1×1017 ions/cm–2, the saturation
AC C
magnetization enhances to ~ 5.24 emu/cm3 indicating that the ferromagnetism has strong correlation with the N ion fluence. The observed magnetic behaviour is consistent with the earlier reports on ferromagnetism in N doped ZnO thin films [13]. It has been proposed by Jindal et al. [32] that doping with N at the O sites in host ZnO lattice is essential to induce the room temperature ferromagnetism in ZnO in contrast to the presence of intrinsic structural defects and crystallinity. They have been able to achieve maximum ferromagnetism for the highest current density for which the grown films have highest crystallinity (XRD) and minimal defects (PL measurements) [32]. Banerjee et al. [33] have reported that ferromagnetism increases and is attributed to the formation of oxygen vacancy clusters upon thermal annealing in ZnO. Gao et al. [34] have also reported that the ferromagnetism of ZnO 8
ACCEPTED MANUSCRIPT nanoparticles increases after annealing in vacuum condition and decreases after annealing in rich-oxygen atmosphere. Therefore, based on the existing literature and cited reports, it can be argued that defects and the site occupancy by the dopants in the host lattice plays a significant role in these system. In the present case, the saturation magnetisation has also been observed in pure ZnO thin films. There are number of reports in the literature on the
RI PT
existence of room temperature ferromagnetism in pure ZnO nanoparticles, nanostructures as well as in thin films and attributed to the various kind of defects (Zn vacancies, O vacancies, Zn interstitials, etc.) present in the lattice [35,36]. Khalid et al. [35] have reported the magnetism in pure ZnO thin films have been associated with the Zn vacancies present in the
SC
host ZnO. Our group has also previously reported the presence of magnetism in pure ZnO thin films and was found to be attributed to the presence of oxygen vacancies [9,37]. Based on the PL results, it can be understood that saturation magnetisation in pure ZnO thin films in
M AN U
the present case is associated with the presence of oxygen vacancies. The ferromagnetic behaviour in N implanted ZnO thin films further can be associated either with the presence of various kinds of defects, site occupation by N ions or secondary phase formation. The possibility of extrinsic ferromagnetism due to secondary phase can be excluded on the basis of XRD results. It has been experimentally proved that nitrogen can occupy three different
TE D
sites in the host lattice: it can occupy interstitial site, Zn and or oxygen sites. Earlier reports have shown that nitrogen at the interstitial and Zn sites did not contribute in the ferromagnetic behaviour [38]. On the other hand, the N substitution on O site favours the spin
AC C
EP
polarized state and results in the enhancement of ferromagnetism in ZnO.
Fig. 4: Magnetization curves of pristine, N ion implanted ZnO thin films: 5×1016 ions/cm2 and 1×1017 ions/cm2.
9
ACCEPTED MANUSCRIPT Total energy calculations of ZnO and N implanted (≈ 4.5 %) ZnO were performed with spin polarization using GGA (PBE) and TB-mBJ potential. The total and partial density of states (PDOS) for both majority and minority spin states of ZnO and N implanted (≈ 4.5 %) ZnO between −8 eV and + 8 eV is shown in the Fig. 5 (a–d) [20-23]. In this the dotted line represents the zero line Fermi energy (EF). In general, the electronic properties of spin
RI PT
polarized ZnO are influenced by the presence of O vacancy (VO) and Zn vacancy (VZn) in the system, resulting in weak magnetization [39,40]. It has been reported that addition of N ions increases the 2p-states in the valence band as a result of increase in the Zn–N bonds that leads to the reduction in VO in the host lattice [19]. Another report has shown that the doping
SC
of nitrogen at oxygen site in ZnO lattice is essential for the enhancement in room temperature magnetism in contrast to the presence of intrinsic structural defects [32]. Furthermore, Ye et al. [13] proposed a plausible mechanism for explaining the magnetic behaviour via doping of
M AN U
2p shell electrons (i.e. C atoms). It has been suggested that partial substitution of C ions over O site causes bonding with Zn ions Zn3d-C2p hybridization. Due to this, there is transfer of electrons of Zn ions (changes Zn d10 (completely filled) to d9 (incompletely filled) state) towards p-orbital of C ions, resulting in the enhancement in the ferromagnetic character with C doping. Our group has also reported that the dilute concentration of C ions on O site into
TE D
ZnO lattice increases ferromagnetism. Based on these observations, it can be concluded that the resultant magnetism increased when N ions are substituted in the place of O ions only in ZnO lattice [41]. The theoretical calculation demonstrates that N incorporation at Zn sites or at the interstitial sites (NI) did not enhance any ferromagnetism in ZnO. The effect of N
EP
substitution at O sites in ZnO super cell lattice favored spin polarization state and increased the magnetic moment. As the N-2p states are more delocalized compared to 3d or 4f states of
AC C
transition metal and also have long range coupling capacity, so below the Fermi level N-2p hybridizes with O-2p that results in the redistribution of spin [27]. Total spin magnetic moment per supercell in the spin polarization calculation is 4.008 µ B for ZnO and it increases to 5.012 µ B for N implanted (≈4.5 %) ZnO lattice which are in good agreement with reported simulation results [39,41]. From the Fig.5, it can be observed that near the Fermi level, N-2p overlap fully with those of O-2p, indicating a strong interaction between them. The spin up orbitals opening the band gap and spin down are partially filled. Furthermore, it also found that in the range from −6.0 eV to −4.8 eV, Zn-3d strong spin polarized and hybridizes mainly with O-2p states. Addition of N3- ions on O2- site of ZnO lattice create impurity moments and additional hole via the following relation N3- +e- → O2-[17]. These holes in the valence band of O 2p states act as main charge carriers which interact with N 2p spins resulting in 2p-2p 10
ACCEPTED MANUSCRIPT interaction. The spin density near each anion impurity tries to follow the impurity ion magnetic moments via 2p-2p interaction. The presence of large number of spin polarized charge carriers lead to indirect long range ferromagnetic ordering between dopants [27]. Similar to 2p-3d interaction, 2p-2p interactions also lead to the creation of additional mixed states in forbidden band gap. So, an enhancement of magnetic moment in ZnO:N is achieved
TE D
M AN U
SC
and O−2p states at the valence band maximum.
RI PT
through 2p−3d hybridization-like 2p−2p coupling interaction from anion sites of N−2p states
EP
Fig. 5 (a-b) Total and partial density of states of spin polarised ZnO super cell respectively; (c-d) total and partial density of states of spin polarised N implanted ZnO super cell respectively.
AC C
A comparison of optical bandgap and saturation magnetization as a function of N ion fluences is investigated. The incorporation of N ion in host lattice site results in the reduction of bandgap whereas there is an enhancement in saturation magnetization. It is expected that the bandgap decreases due to an increase in lattice constant (c-axis) by the substitution of N ions at O site which has been confirmed by X-ray diffraction technique. The increase in lattice constant decreases the coulomb interaction among the atoms, giving rise to decrease in bandgap. Further, a narrow bandgap enhances the ferromagnetic character in host ZnO through N incorporation which in turn is related with the exchange interaction. When the bandgap reduces, the impurity band shifts toward the lower energy side and overlap with the 3d states that result in the enhancement of magnetic exchange interaction.
11
ACCEPTED MANUSCRIPT
4. Conclusions In conclusion, this study demonstrated that N ions implantation plays an important role in the enhancement of ferromagnetic properties of ZnO thin films. XRD diffraction pattern shows the increase in the c-axis lattice parameter with the incorporation of N into
RI PT
ZnO host lattice. The mixing of shallower N 2p states with the valence band of ZnO leads to a reduced band gap which is in accordance with the experimental results obtained in the present work using UV-Vis spectroscopy. Further, the reduction in the band gap results in the enhancement of ferromagnetic behaviour in the present system which is mainly because of
SC
the induced hybridization of N ions at the O lattice sites.
M AN U
Acknowledgement
Author (PK) would like to acknowledge Department of Science and Technology for financial support via DST-INSPIRE Faculty Scheme [No. DST/INSPIRE/04/2015/003149].
References [1]
Ü. Özgür, Y.I. Alivov, C. Liu, A. Teke, M.A. Reshchikov, S. Doğan, V. Avrutin, S.-J. Cho, H. Morkoç, Journal of Applied Physics 98 (2005) 041301. P. Kumar, J.P. Singh, Y. Kumar, A. Gaur, H.K. Malik, K. Asokan, Current Applied
TE D
[2]
Physics 12 (2012) 1166. [3]
P. Kumar, V. Singh, V. Sharma, G. Rana, H.K. Malik, K. Asokan, Ceramics
EP
International 41 (2015) 6734. [4]
J.M.D. Coey, Solid State Sciences 7 (2005) 660.
[5]
H. Peng, H.J. Xiang, S.-H. Wei, S.-S. Li, J.-B. Xia, J. Li, Physical Review Letters 102
[6] [7]
AC C
(2009) 017201.
J.M.D. Coey, M. Venkatesan, C.B. Fitzgerald, Nat Mater 4 (2005) 173. P. Kumar, Y. Kumar, H.K. Malik, S. Annapoorni, S. Gautam, K.H. Chae, K. Asokan,
Applied Physics A 114 (2014) 453.
[8]
K. Ueda, H. Tabata, T. Kawai, Applied Physics Letters 79 (2001) 988.
[9]
K. Parmod, K.M. Hitendra, K. Asokan, EPL (Europhysics Letters) 110 (2015) 67006.
[10]
P. Kumar, V. Sharma, A. Sarwa, A. Kumar, Surbhi, R. Goyal, K. Sachdev, S. Annapoorni, K. Asokan, D. Kanjilal, RSC Advances 6 (2016) 89242.
[11]
J.H. Park, M.G. Kim, H.M. Jang, S. Ryu, Y.M. Kim, Applied Physics Letters 84 (2004) 1338. 12
ACCEPTED MANUSCRIPT [12]
T.C. Kaspar, T. Droubay, S.M. Heald, M.H. Engelhard, P. Nachimuthu, S.A. Chambers, Physical Review B 77 (2008) 201303.
[13]
X.J. Ye, H.A. Song, W. Zhong, M.H. Xu, X.S. Qi, C.Q. Jin, Z.X. Yang, C.T. Au, Y.W. Du, Journal of Physics D: Applied Physics 41 (2008) 155005.
[14]
S. Zhou, Q. Xu, K. Potzger, G. Talut, R. Grötzschel, J. Fassbender, M. Vinnichenko,
RI PT
J. Grenzer, M. Helm, H. Hochmuth, M. Lorenz, M. Grundmann, H. Schmidt, Applied Physics Letters 93 (2008) 232507. [15]
R.P. Borges, R.C.d. Silva, S. Magalhães, M.M. Cruz, M. Godinho, Journal of Physics: Condensed Matter 19 (2007) 476207.
E.-C. Lee, Y.S. Kim, Y.G. Jin, K.J. Chang, Physical Review B 64 (2001) 085120.
[17]
K. Jindal, M. Tomar, R.S. Katiyar, V. Gupta, Journal of Applied Physics 111 (2012)
SC
[16]
102805.
P. Kumar, P. Kumar, A. Kumar, I. Sulania, F. Chand, K. Asokan, RSC Advances 7
M AN U
[18]
(2017) 9160. [19]
R. Hariwal, H.K. Malik, A. Negi, A. Kandasami, RSC Advances 8 (2018) 6278.
[20]
P. Blaha, K. Schwarz, G. Madsen, D. Kvasnicka, J. Luitz, An augmented plane wave+ local orbitals program for calculating crystal properties (2001). G. Kresse, D. Joubert, Physical Review B 59 (1999) 1758.
[22]
S.L. Dudarev, G.A. Botton, S.Y. Savrasov, C.J. Humphreys, A.P. Sutton, Physical
TE D
[21]
Review B 57 (1998) 1505.
H.J. Monkhorst, J.D. Pack, Physical Review B 13 (1976) 5188.
[24]
A. Ghosh, R. Thangavel, M. Rajagopalan, Journal of Materials Science 48 (2013) 8259.
P. Kumar, H.K. Malik, A. Ghosh, R. Thangavel, K. Asokan, Applied Physics Letters
AC C
[25]
EP
[23]
102 (2013) 221903.
[26] [27]
M. Hirai, A. Kumar, Journal of Vacuum Science & Technology A 25 (2007) 1534.
L. Shen, R.Q. Wu, H. Pan, G.W. Peng, M. Yang, Z.D. Sha, Y.P. Feng, Physical
Review B 78 (2008) 073306.
[28]
H.T. Chang, G.J. Chen, Thin Solid Films 618 (2016) 84.
[29]
P. Haas, F. Tran, P. Blaha, Physical Review B 79 (2009) 209902.
[30]
P. Hao, Y. Fang, J. Sun, G.I. Csonka, P.H.T. Philipsen, J.P. Perdew, Physical Review B 85 (2012) 014111.
[31]
Y.F. Wang, Y.C. Shao, S.H. Hsieh, Y.K. Chang, P.H. Yeh, H.C. Hsueh, J.W. Chiou, H.T. Wang, S.C. Ray, H.M. Tsai, C.W. Pao, C.H. Chen, H.J. Lin, J.F. Lee, C.T. Wu, 13
ACCEPTED MANUSCRIPT J.J. Wu, Y.M. Chang, K. Asokan, K.H. Chae, T. Ohigashi, Y. Takagi, T. Yokoyama, N. Kosugi, W.F. Pong, Scientific reports 8 (2018) 7758. [32]
K. Jindal, M. Tomar, R.S. Katiyar, V. Gupta, AIP Advances 5 (2015) 027117.
[33]
S. Banerjee, M. Mandal, N. Gayathri, M. Sardar, Applied Physics Letters 91 (2007) 182501. D. Gao, Z. Zhang, J. Fu, Y. Xu, J. Qi, D. Xue, Journal of Applied Physics 105 (2009)
RI PT
[34]
113928. [35]
M. Khalid, M. Ziese, A. Setzer, P. Esquinazi, M. Lorenz, H. Hochmuth, M. Grundmann, D. Spemann, T. Butz, G. Brauer, W. Anwand, G. Fischer, W.A.
[36]
SC
Adeagbo, W. Hergert, A. Ernst, Physical Review B 80 (2009) 035331.
D. Wang, Z.Q. Chen, D.D. Wang, N. Qi, J. Gong, C.Y. Cao, Z. Tang, Journal of Applied Physics 107 (2010) 023524.
P. Kumar, H.K. Malik, S. Gautam, K.H. Chae, K. Asokan, D. Kanjilal, Journal of
M AN U
[37]
Alloys and Compounds 710 (2017) 831. [38]
T.S. Herng, S.P. Lau, S.F. Yu, H.Y. Yang, X.H. Ji, J.S. Chen, N. Yasui, H. Inaba, Journal of Applied Physics 99 (2006) 086101.
[39]
I.S. Elfimov, A. Rusydi, S.I. Csiszar, Z. Hu, H.H. Hsieh, H.J. Lin, C.T. Chen, R.
TE D
Liang, G.A. Sawatzky, Physical Review Letters 98 (2007) 137202. [40]
A. Zunger, S. Lany, H. Raebiger, Physics 3 (2010) 53.
[41]
J. Kumar, S. Ramasubramanian, R. Thangavel, M. Rajagopalan, in: R.M.S. Rao, T. Okada (Eds.), ZnO Nanocrystals and Allied Materials, Springer India, New Delhi,
AC C
EP
2014, p. 309.
14
ACCEPTED MANUSCRIPT
Highlights 1. Influence of N implantation on optical and magnetic properties of ZnO thin films.
3. Increase in saturation magnetization with N content.
AC C
EP
TE D
M AN U
SC
4. Decrease in optical bandgap with increasing N conc.
RI PT
2. Experimental & DFT calculations have been performed to understand the mechanism.