Mater. Res. Soc. Symp. Proc. Vol. 1577 © 2013 Materials Research SocietyDOI: 1 557/op 0130.1 l.2 .

Room Temperature Ferromagnetism and Band Gap Engineering in Mg Doped ZnO

RF/DC Sputtered Films

Sreekanth K. Mahadeva1,2

, Zhi-Yong Quan 1,3

, Jin-Cheng Fan1,4

, Hasan B. Albargi5, Gillian A

Gehring5 , Anastasia V. Riazanova

1, Lyubov M. Belova

1 and K. V. Rao

1

1. Department of Materials Science and Engineering, Royal Institute of Technology, Stockholm,

SE 100 44, Sweden

2. Department of Physics, Amrita Vishwa Vidyapeetham University, Amritapuri Campus,

Kollam 690 525, Kerala, India

3. Key Laboratory of Magnetic Molecules and Magnetic Information Materials of Ministry of

Education, Shanxi Normal University, Linfen 041004, China

4. School of Materials and Engineering, Anhui University of Technology, Maanshan, 243002,

China

5. Department of Physics and Astronomy, University of Sheffield, Sheffield S3 7RH, U. K.

ABSTRACT

Mg doped ZnO thin films were prepared by DC/RF magnetron co-sputtering in (Ar+O2)

ambient conditions using metallic Mg and Zn targets. We present a comprehensive study of the

effects of film thickness on the structural, optical and magnetic properties. Room temperature

ferromagnetism was observed in the films and the saturation magnetization (MS) increases at first

as the film’s thickness increases and then decreases. The MS value as high as ~15.76 emu/cm3

was achieved for the Mg-doped ZnO film of thickness 120 nm. The optical band gap of the films

determined to be in the range 3.42 to 3.52 eV.

INTRODUCTION

Many studies have reported room temperature ferromagnetism (RTFM) in undoped and

doped metal oxides especially including dilute magnetic semiconductors (DMSs) and dilute

metal oxides (DMOs). Conventionally RTFM in diamagnetic II-VI materials have been achieved

by the introduction of the atoms of magnetic materials into metal oxides host lattice. These

materials are the key for developing magneto-optic and spin electronics devices [1,2]. To realize

spintronics devices, materials that are ferromagnetic above room temperature are essential. With

wide direct band gap (Eg = 3.37 eV) and large exciton binding energy (~60 meV) at room

temperature, ZnO thin films were predicted to be a suitable host material to achieve RTFM [3-4].

Extensive studies show that defects and non-magnetic impurities are playing an important role in

inducing RTFM in ZnO [5-7]. The RTFM in non-transition metal doped and un-doped ZnO

films may be attributed to the different types of defects, such as oxygen vacancies (VO), zinc

vacancies (VZn), zinc interstitial (Zni), film thickness, intrinsic strain and chemisorbed oxygen

among others. Yi et al. reported RTFM in Li doped ZnO and suggested that the origin of

ferromagnetism was associated with VZn produced by the induction of Li doping [7]. Also,

RTFM has been reported in both pristine and doped MgO films [8,9]. MgO is almost an insulator

with a band gap of 7.8 eV, but with exciton energy comparable to that of ZnO. The wide

tunability of the band gap in Mg incorporated ZnO films opens the door for the realization of

novel optoelectronic devices especially short wavelength light emitters and photo detectors.

509

Tunability of the band gap from 3.3 eV to 7.8 eV covers the ultraviolet (UV) regions [10]. So

that, MgZnO alloy is one of the important barrier material for realizing the high efficient UV

light emission in the quantum well structure [11] and furthermore the spintronic functionalities

can be conceived if a DMSs barrier is used. Recently, Li et al. reported that tailoring the band

gap and engineering the defects were effective in tuning the RTFM in MgZnO alloy [12]. While

there is significant amount of study of the optical properties of band gap engineered MgZnO

relatively little has appeared on the magnetic properties in literature.

In this work, we report a comprehensive study of RTFM in RF/DC sputtered thin films of

Mg doped ZnO using metallic Mg and Zn as targets. We find that the saturation magnetization

value, MS, for a specific concentration of Mg/Zn increases with film thickness and then after a

maximum value decreases eventually reaching the bulk diamagnetism. Experimental

investigations show that intrinsic defects in the films deposited on silicon substrates play an

important role in the observed RTFM. Furthermore, we find that the band gap and RTFM in

Mg-doped ZnO films is tunable by varying the film thickness.

EXPERIMENTAL DETAILS

Mg-doped ZnO films were deposited on Si and glass substrates by co-sputtering pure Mg

(99.99%) and Zn (99.99%) targets in the Leybold-Heraeus sputtering system (Combivac IT230)

with DC power of 10W on Mg target and RF power of 50W on Zn target at RT. The vacuum

chamber was evacuated to ~10−6

mbar. The film depositions on Si (100) substrates were made at

a fixed oxygen partial pressure of 1.5 × 10−4

mbar in a mixture of Ar and O2 with total pressure

of 1.5 × 10−3

mbar, for different deposition times. The silicon substrates were etched with hydro

fluoric acid (HF) to remove surface native oxide (SiO2) and the etched Si substrates were cleaned

in the ultrasonic baths of acetone, isopropanol and de-ionized water, and blown dry in nitrogen.

Prior to the film growth, the targets were cleaned by sputtering them for 30min.

Our bulk target materials are diamagnetic and energy dispersive spectroscopy (EDS)

analyses for them do not show any form of transition metal contamination within the limit of

detection. The crystal structure of the films was characterized by x-ray diffraction (XRD,

Siemens D5000 and Bruker D2 Phaser Desktop) equipped with a parallel beam optical

system using Cu Kα (λ = 1.5405Å) irradiation. The surface morphologies and the thickness of

the thin films were analyzed by using a Dual-Beam scanning electron microscope

(SEM)/focused ion beam (FIB) system. The RT optical absorption measurements were

performed using an UV-visible-near infrared spectrophotometer. The magnetic properties of Mg-

doped ZnO films were measured at RT by means of SQUID magnetometer produced by

Quantum Design. In the evaluation of all the magnetic raw data the diamagnetic response arising

from the respective Si substrates was accounted for.

RESULTS AND DISCUSSION

The cross-section of Mg-doped ZnO films deposited on Si substrate was analyzed by using

FIB/SEM technique in our FEI Nova 600 Nanolab microscope. The thickness of the films was in

the range of 40-240 nm. Figure 1 (A) shows the typical cross-section of the films deposited on Si

substrate for 90 min with O2 content of 10%. Uniform thickness of ~120 nm was found at

different sections of the film. The deposition rate was estimated to be 1.33 nm/min. From this

value, the thickness of the thin films was calculated to be ~40, 80, 120, 160, 200 and0 240 nm

for the films deposited at 30, 60, 90, 120, 150 and 180 min, respectively.

EDS was used to analyze the compositions of Mg-doped ZnO films. Figure 1(B) shows a

typical EDS spectrum of as-grown films on Si substrate. Only Mg, Zn and O elements were

detected, indicating that there is no form of transition metal contamination in the films within the

detection limit of 0.1at.%. The inset of figure 1(B) shows the corresponding atomic % of Mg, Zn

and O. The atomic concentration of Mg is 6% (written as Mg0.06Zn0.94O).

Figure 1. (A) The typical FIB cross-section of the Mg-doped ZnO film deposited for 90 min from

Mg and Zn targets (view at a 52 tilt) and (B) is the EDS spectrum of the same sample deposited

on Si substrate.

The SEM images of as-grown Mg-doped ZnO films with thickness of 120 nm and 240 nm

on Si substrates are shown in figure 2(A&B), which indicates a smooth and dense structure

without any micro cracks, suggesting the homogeneous surface quality of the films. The uniform

dense granular structure of the film can be seen from the images. As thickness increases, the size

and structure of the grains in the films becomes more consistent. Figure 2(C) shows the XRD

patterns of Mg-doped ZnO thin films on Si substrates with different thickness. It is observed that

all the films exhibits distinctly multiple diffraction peaks except the shallow ones in the case of

films of thickness of 40 nm and 80 nm. The diffraction peaks corresponding to (100), (002) and

(103) were observed around at 30.64o, 33.98

o and 62.37

o, respectively. The peaks of (100) and

(103) appear in the film with thickness of 120 nm and become stronger in the films with

thickness more than 120 nm perhaps due to the changing defect and strain effects. The slight

change in the broadening of the peaks indicates the variation of the grain size. The average

crystal grain size (d) can be estimated using the Scherrer formula 0.9 180

cosd

where λ=0.15405 nm, is the wavelength of the Cu Kα radiation and β is the full width at half

maximum (FWHM) of the diffraction peak from the (hkl) crystal plane. The average crystallite

size for the samples was estimated to be 8.9, 7.7, 6.9, 7.6, 8.5 and 8.3 nm for the films with

approximate thickness of 40, 80, 120, 160, 200 and 240 nm according to the analyses based on

the (002) peaks.

Figure 2. SEM morphologies of Mg-doped ZnO thin films prepared with thickness of 120 nm (A)

and 240 nm (B) respectively. Fig.2 (C) is the XRD patterns of the films with different thickness

on Si substrate.

The band gap of the Mg0.06Zn0.94O films were evaluated using the relation:

( ) ( )gh h E where α is the absorption coefficient and hv is the photon energy. Figure 3

shows (αhv)2 plot of Mg0.06Zn0.94O thin films deposited on glass substrate as a function of photon

energy hv. Inset of the figure 3 shows that Eg as a function of film thickness. As the films

thickness increases the Eg decreases first from 3.48 to 3.42 eV and then increases up to 3.52 eV.

Figure 3. Optical absorption spectra of Mg0.06Zn0.94O thin films deposited on glass substrates

and the inset shows corresponding optical Eg as a function of thickness.

There have been many measurements that show that the band gap of pure ZnO depends on

the method of preparation which affects both the oxygen stoichiometry and also the grain size.

The commonly quoted values for the band gap of pure ZnO are ~3.3 [10], however it has

recently been shown that gap can vary between 3.13eV and 4.06eV for films deposited by

MOCVD as the temperature of the substrate is varied [13]. The band width of the mixed film

MgZnO is expected to increase smoothly as x increases so long as it remains in the wurzite

structure. A recent theoretical work predicted that Eg(x) = Eg(0) +2.03x [14], which compared

well with some experimental results [15], and in this present work the value of band gap is

almost in agreement with the reported results ranging from 3.42 to 3.52 eV. Our samples were

deposited on Corning® Glass which starts to absorb at around 4 eV and hence the measurements

of the band edge were reliable only for energies below that value. Mg doping increases the value

of band gap compared to that of pure ZnO (3.37 eV). The band gap value slightly varies with the

thickness of the films due to the ambient growth conditions related to defect, and strain among

other parameters. Thus, it is clear that the Eg of the Mg-doped ZnO films can be tuned by

changing thickness of the films and growth atmosphere. Also, we find the experimental results

from the optical band gap are consistent with the translation of phase to a crystalline character as

suggested by the observed XRD patterns. This may be attributed to that the concentration of

defects (such as zinc vacancy, and magnesium vacancy etc.) reaches maximum peak followed by

reduction due to the combination of changing nature and densities of the various defects when

the thickness is larger than 120 nm.

Systematic studies of the magnetic properties of Mg-doped ZnO films were carried out

using thee Quantum Design SQUID system. Figure 4(a) shows the corrected magnetic hysteresis

loops measured at room temperature for the films deposited on Si substrates for various

thicknesses. Ferromagnetism was observed in all the as-grown Mg-doped ZnO films. The MS

values as a function of film thickness, shown in figure 4(b), increases with the thickness then

decreases in the ranges of ~1.3-15.76 emu/cm3 with a maximum value for the film with 120 nm

thickness. The magnetism of the films may be due to the intrinsic complexities of cation and

other defects. Initially, doping Mg in films produces cation vacancies and the polarized oxygen

atoms around these defect sites induce ferromagnetism. The effective magnetization increases to

become a maximum when some sort of equilibrium is reached between the cation vacancy

concentration and the strain. Above such optimized thickness in the film the density of the

relative cation induced defect concentration begins to decrease with combined effect of other

defects (such as zinc interstitial) to decrease magnetism eventually the diamagnetism of the bulk.

Figure. 4 (A) the saturation magnetization (MS) as a function of magnetic field (H) of the as-

grown Mg-doped ZnO films deposited on Si substrate with different thickness and (B) MS as a

function of film thickness.

CONCLUSIONS

Mg doped ZnO films were prepared on Si substrates by co-sputtering Mg and Zn targets.

We observed the Eg of the films varies with thickness, which may be ascribed to the growth

condition and Mg doping. The RTFM was observed in all the Mg-doped ZnO films prepared on

Si. With the increase in film thickness, a transition from ferromagnetic to diamagnetic behavior

was observed which is attributed to the cation defect induced strain. The intrinsic defects in the

films play an important role in structural and magnetic properties in Mg-doped ZnO films.

Extensive studies of XPS, NEXAS, and XMCD on these films will be published in the near

future.

ACKNOWLEDGMENTS

The work was supported by the Swedish funding Agencies, VINNOVA, Hero-M Centre of

Excellence at KTH. Sreekanth K Mahadeva acknowledges a graduate study fellowship funded

by the India 4EU- Erasmus Mundus External Cooperation Window program. Jin-Cheng Fan and

Zhi-Yong Quan acknowledge the Carl Trygg's Foundation in Sweden for post-doctoral

Scholarships. We thank Dr. K S Sreelatha of Amrita University for her interest and comments.

REFERENCES

[1] J. K. Furdyna, J. Appl. Phys. 64, R29 (1988).

[2] S. D. Sarma , American Scientist 89, 516 (2001).

[3] T. Dietl, H. Ohno, F. Matsukura, J. Cibert, and D. Ferrand, Science 287, 1019 (2000).

[4] J. M. D. Coey, A. P. Douvalis, and C. B. Fitzgerald, Nature Mater. 4, 173 (2005).

[5] M. Kapilashrami, J. Xu, V. Ström, K. V. Rao, and L. Belova, Appl. Phys. Lett. 95, 033104

(2009).

[6] G. Xing, D. Wang, J. Yi, L. Yang, M. Gao, M. He, J. Yang, J. Ding, T. C. Sum, and T. Wu,

Appl. Phys. Lett. 96, 112511 (2010).

[7] J. B. Yi, C. C. Lim, G. Z. Xing, H. M. Fan, L. H. Van, S. L. Huang, K. S. Yang, X. L.

Huang, X. B. Qin, B. Y. Wang, T. Wu, L. Wang, H. T. Zhang, X. Y. Gao, T. Liu, A. T. S.

Wee, Y. P. Feng, and J. Ding, Phys. Rev. Lett. 104, 137201 (2010).

[8] C. M. Araujo, M. Kapilashrami, J. Xu, O. D. Jayakumar, S. Nagar, Y. Wu, C. Århammar, B.

Johansson, L. Belova, R. Ahuja, G. A. Gehring, K. V. Rao, Appl. Phys. Lett. 96, 232505

(2010).

[9] S. Nagar, O. D Jayakumar, L. Belova, and K.V. Rao, Materials Express 2, 233(2012).

[10] Ü. Özgür, Ya. I. Alivov, C. Liu, A. Teke, M. A. Reshchikov, S. Doğan, V. Avrutin, S.-J.

Cho, and H. Morkoçd, J. Appl. Phys. 98, 041301(2005).

[11] A. Ohtomo, K. Tamura, M. Kawasaki, T. Makino, Y. Segawa, Z. K. Tang, G. K. L. Wong,

Y. Matsumoto, and H. Koinuma, Appl. Phys. Lett. 77, 2204 (2000).

[12] Y. Li, R. Deng, B. Yao, G. Xing, D. Wang, and T. Wu, Appl. Phys. Lett. 97, 102506 (2010).

[13] S. T. Tan, B. J. Chen, X. W. Sun, W. J. Fan, H. S. Kwok, X. H. Zhang, and S. J. Chua, J.

Appl. Phys. 98, 013505(2005).

[14] C. Franz, M. Giar, M. Heinemann, M. Czerner, and C. Heiliger, MRS Proceedings, 1494,

mrsf12-1494-z04-32 doi:10.1557/opl.2012.1709 (2013).

[15] A. Ohtomo, M. Kawasaki, T. Koida, K. Masubuchi, H. Koinuma, Y. Sakurai, Y. Yoshida,

T. Yasuda, and Y. Segawa, Appl. Phys. Lett. 72, 2466(1998).