Room temperature ferromagnetism in Mn doped dilute ZnO semiconductor; an electronic structure study

ARTICLE IN PRESS Physica B 404 (2009) 3275–3280

Contents lists available at ScienceDirect

Physica B journal homepage: www.elsevier.com/locate/physb

Room temperature ferromagnetism in Mn doped dilute ZnO semiconductor; an electronic structure study R.K. Singhal a,, M. Dhawan a, S. Kumar b, S.N. Dolia a, Y.T. Xing c, Elisa Saitovitch c a

Department of Physics, Rajasthan University, Jaipur 302055, India Department of Physics, M. L. Sukhadia University, Udaipur 313 002, India c CBPF, Rua Xavier Siguad 150, Urca, Rio de Janeiro, RJ, Brazil b

a r t i c l e in fo

Keywords: Diluted magnetic semiconductors Room temperature ferromagnetism X-ray photoemission spectroscopy X-ray absorption spectroscopy SQUID measurements

abstract Confronting reports of room temperature ferromagnetic ordering in bulk as well as thin films of dilute doped semiconductors motivated us to prepare and study some dilute Mn-doped (2% and 4%) ZnO polycrystalline pellets. Our aim was to find out how the changes in the electronic structure can correlate to the observed magnetic properties in such diluted magnetic semiconductor materials. The 2% Mn sample shows above room temperature magnetic ordering with further enhancement in magnetic properties upon cooling. The magnetic ordering gets completely quenched for 4% Mn doping at room temperature and no evidence of ordering found even down to 5 K. The electronic structure of these two samples has been investigated using Mn L23 X-ray absorption spectroscopy, Zn 2p and 3p, Mn 3p and O 1s X-ray photoemission spectroscopy. Our results show that overall the manganese shifts toward the higher valence state upon Mn doping while there is no change in the zinc and oxygen valence. The concentration of divalent Mn state is found to be much higher in the ferromagnetic sample (Mn ¼ 2%). For the non-ferromagnetic sample a larger contribution of higher oxidation Mn states is present and also the oxygen content is found to be higher. These two factors have been correlated to the suppressed ferromagnetism but it is difficult to predict their relative weights. & 2009 Elsevier B.V. All rights reserved.

1. Introduction The occurrence of room temperature ferromagnetic (FM) ordering in Mn-doped ZnO and other semiconductors such as GaAs, GaN, when doped with much diluted transition metals, as predicted theoretically by Dietl et al. [1], has attracted intense research interest, as the host materials are a well known highly stable wide-gap semiconductor which is readily available. If the prediction were right, it could be applied with high technological impact as a very useful transparent diluted magnetic semiconductor (DMS) for spintronic and magnetooptical devices. This explains the large number of subsequent experimental works [2–18] carried out to try and verify or clarify the theoretical prediction. Unfortunately there are many controversial reports especially on Mn-doped ZnO so far. The diluted magnetic semiconductors (DMSs) based on III–V semiconductors show magnetic ordering only at very low temperatures, the oxidebased DMSs may be ferromagnetic at 300 K or at even higher temperatures. Many doubts have been expressed that even

 Corresponding author. Tel.: +91141 2545931.

E-mail address: [email protected] (R.K. Singhal). 0921-4526/$ - see front matter & 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.physb.2009.07.100

though the Mn clusters are expected to give an anti-ferromagnetic state, rather than a ferromagnetic (FM) order, very small clusters can be ferromagnetic [2]. However, the first reports of room temperature FM ordering in Mn-doped ZnO powder, bulk pellets as well as 2–3 mm laser-ablated transparent films with Curie temperatures far above 300 K came from Sharma et al. [19,20]. Blythe et al. [21] confirmed this, who made a detailed study of dependence of the FM ordering on the processing temperature. On the contrary, Kane et al. [22] claimed no evidence of an FM ordering in ZnxMn1xO, ZnxCo1xO bulk single crystals. Instead, they claimed a paramagnetic behavior in case of the former and hysteresis in the latter owing to superparamagnetic Co clusters embedded in a diamagnetic ZnO matrix. Kittilstved et al. [23] have, on the other hand, not only observed high-Tc ferromagnetism in ZnMnO and ZnCoO systems but also show how their Tc can be manipulated through in situ as well as ex situ chemical manipulation. Long et al. [24] studied the effects of annealing atmosphere on FM behavior for the ZnCoO bulk samples. They show that the air-annealed samples have weak FM ordering, but the samples annealed in vacuum or Ar/H2 show a strong FM behavior, indicating that the strong ferromagnetism is associated with high oxygen vacancy density. The enhanced room temperature FM behavior is also found in the samples with high carrier concentration controlled by doping interstitials Zn.

ARTICLE IN PRESS 3276

R.K. Singhal et al. / Physica B 404 (2009) 3275–3280

Many attempts have been made to explain the ferromagnetism in DMSs, theoretically. It has been proposed that the ferromagnetism in ZnO system can be obtained by the mediation of the shallow donors or acceptors [23] or the holes in the valence band [1]. Some workers suggested [6] that the ferromagnetism in this system should be associated with the co-existence of Mn3+ and Mn4+ via a double-exchange mechanism. Based on X-raymagnetic-circular dichroism studies of diluted Zn1xCoxO films, it is shown that Co ions substituted to Zn in the ZnO matrix are not responsible for the origin of FM ordering [6–8] but the anion sublattice may be instead responsible for the ferromagnetism [8]. Some reports [25–28] on ferromagnetism in various undoped oxides claimed that magnetism is certainly possible in undoped semiconducting and insulating oxide thin films, which originate due to oxygen vacancies. However, the origin of ferromagnetism in semiconductors remains an issue of debate, despite a lot of theoretical and experimental attempts. When assigning the origin of ferromagnetism in Mn-doped ZnO DMS, it is important to distinguish the contributions of the transition metal with different valence bonds, providing evidence to explore whether the magnetization originates from the magnetic precipitates or not. In this paper we have studied the electronic structure of two single phase polycrystalline ZnMnO samples with 2% and 4% Mn concentration which exhibit different magnetic behavior. The 2% sample displayed clear FM ordering at 300 K, the 4% sample did not show any ordering even upon cooling. The electronic structure of these two samples has been investigated using X-ray absorption and photoemission spectroscopy with an aim to find out how the changes in the electronic structure can correlate to the observed magnetic properties.

were then sintered at (50075) 1C for 18 h. The hysteresis loops were measured at different temperatures by a dc-SQUID magnetometer (MPMS-XL, Quantum Design) at CBPF, Brazil. The X-ray diffraction (XRD) patterns of the samples were collected on a Rigaku X-ray diffractometer equipped with CuKa radiation. The 2y scan was from 101 to 901 with a step size of 0.051. Rietveld profile refinements of the XRD patterns were carried out using the FULLPROF Program [29]. In the structure refinement the parameters varied were: two cell parameters, one atomic position, site occupancies, overall isotropic temperature factor in addition to the instrumental parameters such as zero angle, three half width parameters and a scale factor. The Mn L23 edge X-ray absorption spectra were recorded at the BACH beamline of Elettra in Trieste, Italy. The spectra were recorded in total electron yield mode at 300 K at pressure better than 5  1010 mbar. The measurements were performed on Zn0.98Mn0.02O and Zn0.96Mn0.04O pellets (referred to as the 2% sample and 4% sample, respectively, in the following) and on some reference samples, i.e. MnO, La1.2Sr1.65Ca0.15Mn2O7 bilayer manganite, all scraped in UHV before each measurement. The core level X-ray photoemission spectra were measured using an X-ray photoelectron spectrometer at IUC, Indore, India, using the Al Ka radiation. The samples were scraped uniformly with diamond files before carrying out the measurements. Final spectra were taken after minimizing the feature coming from carbon contamination of the surface (C 1s peak). The vacuum in the chamber was 4.4  1010 Torr. We kept scraping the samples in situ to get uncontaminated surface throughout the measurements. No shift was observed due to charging of the samples.

2. Experimental procedure

(a) XRD data. The XRD patterns of pure ZnO with Wurtzite hexagonal structure and no impurity peaks are displayed in Fig. 1a. The cell parameters were refined with the help of a least square refinement program that are in good agreement with the reported values. Fig. 1b shows the fitted XRD patterns of the 2% Mn doped ZnO sample recorded at 300 K (the XRD patterns for the 4% sample are not shown here). The patterns show the single phase ZnS type Wurtzite hexagonal symmetry with no signature of any peak from any other phase. We used the FULLPROF Program [29] for Rietveld profile refinements of the XRD patterns. The

3. Results and discussion

2000

(101)

Appropriate quantities of heat treated ZnO and MnO2 powders were mixed and ground thoroughly for nearly 10–12 h to ensure a homogeneous mixture. The ground mixture was calcined at (45075) 1C for nearly 14 h raising the temperature very slowly above 100 1C in a microprocessor controlled furnace and similarly cooling it very slowly after completion of the calcinations. The resulting lumps were ground for 5 h and pellets of 10 mm diameter using a hydraulic pressure of nearly 7 tons. The pellets

9000 7500 Counts (a.u.) (203)

(113), (104)

(103) (200) (112) (201) (004) (202)

400

(110)

800

(102)

1200

(100) (002)

Counts

1600

0

6000 4500 3000 1500 0

10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 2θ (deg.)

10

20

30

40 50 60 2θ (degrees)

70

80

90

Fig. 1. (a) The XRD pattern of ZnO powder at 300 K. (b) The fitted XRD pattern of Zn0.98Mn0.02O sample. Observed (calculated) profiles are shown by dotted curves. The short vertical marks represent Bragg reflections. The lower curve is the difference plot.

ARTICLE IN PRESS R.K. Singhal et al. / Physica B 404 (2009) 3275–3280

3277

Table 1 Crystallographic data for Zn1xMnxO samples at 300 K obtained using the Rietveld refinement of the powder XRD patterns. Sample

Cell parameters

ZnO Zn0.98Mn0.02O Zn0.96Mn0.04O

Discrepancy factors

˚ a (A)

˚ c (A)

V (A˚ 3)

O (positional parameter z)

w2

Rp (%)

Rwp (%)

3.249 3.252(3) 3.253(2)

5.207 5.212(5) 5.213(4)

47.69 47.73 47.76

0.612 0.628 0.637

1.76 1.86

6.15 6.34

7.92 8.63

Space group is P63mc (no. 186, Z ¼ 2). Each atom in the Wurtzite hexagonal structure resides on the 2b Wyckoff position with Zn atoms at (1/3,2/3,0) and O atoms at (1/3,2/3,z).

0.004

0.003

Zn0.98Mn0.02O at 300 K

0.003

Mn=2% 300 K

0.000

Moment (emu/g)

0.001

0.0032

-0.001 -0.002

0.0024

Mn=4%

0.0016

300 K

0.0008 0.0000

-100

0.000 -0.001 -0.002

-0.0016 -0.0024 -0.0032 -300

-200

0.001

-0.0008

-200

-100 0 100 Magnetic field (Oe)

200

300

-0.003 -300

As recorded data FM Contribution PM Contribtion PM + FM Contribution

0.002

Magnetic Moment (emu/g)

Magnetic Moment (emu/g)

0.002

0 Field (Oe)

100

200

300

-0.003 -0.004 -600

-400

-200

0 Field (Oe)

200

400

600

Fig. 2. (a) Magnetization vs. field curve for Zn0.98Mn0.02O at 300 K. The inset shows the curve for Zn0.96Mn0.04O at 300K. (b) Magnetization vs. field curve for Zn0.98Mn0.02O at 300 K along with the paramagnetic and FM contributions.

50 K

0.003 Zn0.98Mn0.02O 0.002 Moment (emu/g)

Rietveld refinement was carried out using the P63mc space group (no. 186, Z ¼ 2), with each atom in the Wurtzite hexagonal structure residing on the 2b Wyckoff position with the Zn atoms at (1/3,2/3,0) and the O atoms at (1/3,2/3,z) co-ordinates. In the refinement process the Mn occupancy was varied for the two Zn and O sites to locate the exact site of Mn. The best fit was obtained when Mn atoms occupy the Zn site with total preference while the Mn atoms occupying the O site gave very poor fits. In fact, the atomic scattering lengths of the three atoms i.e. Zn (Z ¼ 30), Mn (Z ¼ 25) and O (Z ¼ 8) are quite different as these are far away in the periodic table, hence all the three atoms can be easily distinguished in the profile refinement. Our Rietveld analysis shows that Mn atoms occupy the Zn site with total preference. The cell parameters are presented in Table 1. The refined parameters presented in the table further confirm the stoichiometry and the single phase nature of the samples. (b) SQUID data. Fig. 2a shows the magnetization vs. field (hysteresis) curves at 300 K for Zn0.98Mn0.02O and Zn0.96Mn0.04O samples. For each point, the SQUID did four scans and took the average of the data. Also the measurements were repeated several times with the time interval of several weeks to minimize the experimental error. The curves clearly reveal a room temperature FM ordering of the 2% sample but no such signature for the 4% sample. In fact, the 4% sample shows a purely a paramagnetic state (see inset). The shape of the loop does not rule out the presence of some paramagnetic (PM) contribution even for the 2% sample. Fig. 2b displays the separated PM and the FM contributions along with the experimental curve and the fit made

300 K

0.001 0.000 -0.001 -0.002 -0.003 -1000

-500

0

500

1000

Field (Oe) Fig. 3. Comparison of hysteresis curves for Zn0.98Mn0.02O sample at 50 and 300 K.

to it. The saturation magnetization and coercivity estimated come out to be 0.0019 emu/g and 19.8 Oe, respectively. Fig. 3 displays the comparison of the curves at 300 and 50 K for the FM sample. The sample shows an appreciable enhancement in saturation magnetization as well as the coercivity upon cooling (0.0029 emu/g and 27.9 Oe, respectively at 50 K). The sample

ARTICLE IN PRESS 3278

R.K. Singhal et al. / Physica B 404 (2009) 3275–3280

Mn L23 XAS 2% sample Intensity (arb. units)

MnO

La1.2Sr1.65Ca0.15Mn2O7 Difference 2% sample

Bilayer manganite

MnO 640

650 Photon Energy (eV)

660

638

640 642 644 Photon Energy (eV)

646

Fig. 4. (a) Mn L23 XAS spectra of La1.2Sr1.65Ca0.15Mn2O7 bilayer manganite, Zn0.98Mn0.02O, MnO systems. (b) Mn L3 XAS of the Zn0.98Mn0.02O compared to Mn L3 XAS of MnO, difference spectrum between the Mn L3 XAS of MnO and the Mn L3 XAS of Zn0.98Mn0.02O, Mn L3 XAS of the bilayer manganite shown for comparison.

3p3/2

1016 1018 1020 1022 1024 1026 1028 1030 1032 Binding Energy (eV)

x=0.00 x=0.02 x=0.04 3p1/2

Intensity (a. u.)

Intensity (a. u.)

x=0.00 x=0.02 x=0.04

80

82

84

86

88

90

92

94

96

98 100

Binding Energy (eV)

Fig. 5. (a) The Zn 2p3/2 XPS spectra for the Zn1xMnxO [x ¼ 0, 0.02, 0.04] samples along with the Gaussian fits. (b) The Zn 3p XPS spectra along with the Gaussian fits.

with 4% Mn did not show any FM ordering even upon cooling down to 5 K (not shown here). (c) XAS data. The Mn L23 edge X-ray absorption spectrum measured for the 2% sample is shown in Fig. 4a. Spectra obtained for MnO (Mn2+ reference system) and the La1.2Sr1.65Ca0.15Mn2O7 bilayer manganite (reference system of a mixed valent Mn3+/Mn4+ compound) are also shown for a comparison. The spectrum of the 2% magnetic sample is quite different from that of the bilayer manganite, which has the main peak at nearly 2.5 eV higher energy than the main peak of the 2% sample (at 640 eV). This indicates a significant change in the charge state and crystal field of Mn in the two systems. The spectrum of the 2% sample is also different from the spectrum of Mn metal (see Ref. [10]), which does not display multiplet structures. This rules out the possibility of Mn impurities getting segregated in metallic clusters form. In fact, the spectrum from of the 2% sample shows multiplet structures qualitatively similar to those of MnO, suggesting that Mn ions in the sample Zn0.98Mn0.02O are nearly divalent. However, the Mn L3 edge reveals several differences between the two spectra, if we look at Fig. 4b carefully. Firstly, the lower energy multiplet lines at 639 eV do not appear for Zn0.98Mn0.02O.

Secondly, there is an increased spectral weight around 642 eV in Zn0.98Mn0.02O. These differences indicate that the local environments of Mn in Zn0.98Mn0.02O and MnO are different. The absence of the multiplet lines at 639 eV can be assigned to a lower absolute value of the crystal field parameter 10Dq for Zn0.98Mn0.02O with respect to the corresponding value for MnO. In fact the difference spectrum between the Zn0.98Mn0.02O and MnO spectra, displayed in Fig. 4b, does show similarities with the main Mn L3 XAS peak of the bilayer manganite. (d) Zn 2p and 3p XPS data. The Zn 2p3/2 XPS spectra for the pure ZnO and the Mn doped samples are displayed in Fig. 5a. All the three spectra show a symmetric featured single peak at 1023 eV. A single Gaussian was nicely fit to all the spectra and no change in the peak positions was noticed upon Mn doping. The single Gaussian rules out the possibility of multiple components of Zn in any of these samples. Fig. 5b shows the Zn 3p core level XPS spectra. The spectra show the 3p3/2 and 3p1/2 bands located at 88.6 and 91.6 eV, respectively. The two Gaussians were fit to the Zn 3p spectra but again no shift was observed in the positions of two bands (3p3/2, 3p1/2), which reconfirms the absence of any other multiple Zn components in these samples. The area under the peaks

ARTICLE IN PRESS R.K. Singhal et al. / Physica B 404 (2009) 3275–3280

of 2p3/2, 3p3/2 and 3p1/2 progressively decreases as the concentration of Mn at Zn site increases which was natural. Thus no electronic structural change is observed at the Zn site by Mn doping. (e) Mn 2p3/2 XPS data. The Mn 2p3/2 XPS spectra are shown in Fig. 6. The three Gaussian peaks located at 640.0, 642.6 and 645.0 eV can be attributed to Mn2+, Mn3+ and Mn4+ valence states, respectively. To make a quantitative estimation of atomic percent of Mn2+ and the other valence states, the area of the different valence bonds of Mn was calculated and the results are summarized in Table 2. It has been found that the largest Mn2+ atomic concentration appeared in the FM 2% sample. In the nonFM 4% sample the Mn2+ percent declines and the higher valence states Mn3+ and Mn4+ dominate. (f) Oxygen 1s data. Asymmetric oxygen 1s spectra, shown in Fig. 7, indicate that multi-component oxygen species are present

x=0.04

Mn4+ Mn3+ Mn2+

3279

in the near-surface region of these samples. The high energy peaks located at 532.7 and 533.5 eV are due to the chemisorbed oxygen of the surface hydroxyl, –CO3, absorbed H2O, absorbed O2 or the surface contamination. But the most intense first peak at 531.2 eV is due to the actual O 1s contribution from the samples. Area under the peak O 1s progressively increases by the doping of Mn. The first Gaussian peak that is corresponding to oxygen 1s feature has been highlighted in the inset of Fig. 7. This enhancement of oxygen and the increase of Mn valence state with increase in Mn concentration can be attributed to the charge neutrality of the samples similar to the reports of magnetic ordering in ZnO thin films [30] and may be causing the suppression of ferromagnetism. This is in agreement to reports [24] that have shown that the air-annealed sample has weak FM ordering, but the sample annealed in vacuum or Ar/H2 shows a strong FM behavior, indicating that the strong magnetic ordering might be associated with high oxygen vacancy density. Doubts have been expressed that the FM ordering in DMSs, especially the Mn doped ZnO system, could be due to the presence of some secondary phases or clustering effect. In our samples the possible secondary phases could be manganese oxide phases, such as MnO, MnO2 and Mn3O4. Of these manganese oxides, the MnO and the MnO2 may be safely ruled out as they are antiferromagnetic. Ferromagnetic Mn3O4 has a Curie temperature at 43 K that is far below all these samples [31]. In addition, another possible phase ZnMnO3 is established to show a spin–glass behavior [32,33]. If the formation of a secondary Mn-related

x=0.00

O 1s Intensity (arb. unit)

Intensity (a.u.)

x=0.02 x=0.04

Intensity (a. u.)

x=0.02

527 528 529 530 531 532 533 534 535 Binding Energy (eV)

Surface contamination

630

632 634

636

638

640

642

644 646

648

650

526

528

530

Binding Energy (eV) Fig. 6. Gaussian fits in the Mn 2p3/2 XPS for the Zn1xMnxO [x ¼ 0.02, 0.04] samples.

532

534

536

538

540

Binding Energy (eV) Fig. 7. The O 1s XPS spectra for the two samples along with the Gaussian fits. The Gaussian corresponding to oxygen 1s feature is shown in the inset for clarity.

Table 2 Area of the peaks under the Mn2+, Mn3+ and Mn4+ Gaussians and the absolute quantity of Mn2+ in the 2 and 4% doped samples. Sample

Zn0.98Mn0.02O Zn0.96Mn0.04O

Area of peaks (in arb. Units) 640.0 eV Mn2+

642.6 eV Mn3+

645.0 eV Mn4+

3672.1 1011.8

2521.3 4073.9

1681.3 3494.2

Quantity (%) of Mn2+ in all Mn valence bonds

Quantity (%) of Mn2+ in Zn1xMnxO sample

46.63 11.79

0.93 0.47

ARTICLE IN PRESS 3280

R.K. Singhal et al. / Physica B 404 (2009) 3275–3280

phase were responsible for the FM behavior, an increase in Mn concentration would presumably increase the secondary phase volume fraction and related magnetization signature, instead, the opposite behavior is observed that by increasing the Mn content from 2 to 4 at% caused a complete destruction of FM order. This rules out the magnetization due to any precipitating secondary phase in our samples. Various mechanisms have been put forward to understand the origin of FM ordering in these DMSs. For example some are based upon mediation of the charge carriers [23] or the holes in the valence band [1], others are based upon the co-existence of Mn3+ and Mn4+ via a double-exchange mechanism. Also it has been argued, on the basis of experimental data, that the FM ordering in the ZnO system should not originate from oxygen vacancies as in the case of TiO2 and HfO2 films, but from defects on Zn sites [30,34]. A theoretical work stating that for ZnO, oxygen vacancies or Zn interstitials should result in anti-ferromagnetism, while ZnO with Zn vacancies has more chance to be ferromagnetic [34], supports this argument. It was also reported that magnetization of films 10–50 nm thick is much larger than that of the thicker films. This reveals that defects must be located mostly at the surface and/or the interface between the film and the substrate, or possibly nanometer-size layers at the interface/surface. According to Ruderman–Kittel–Kasuya–Yoshida theory [35,36] the ferromagnetism is a carrier-induced mechanism. This arises due to the exchange interaction between local spin-polarized electrons like the electrons of Mn2+ ions, and conductive electrons. This interaction leads to the spin polarization of conductive electrons. Subsequently, the spin-polarized conductive electrons perform an exchange interaction with local spin-polarized electrons of other Mn2+ ions. Thus, after the long-range exchange interaction, almost all Mn2+ ions exhibit the same spin direction. The conductive electrons are regarded as a media to contact all Mn2+ ions. As a result, the material exhibits FM ordering. More local spinpolarized electrons are produced with increasing Mn2+ concentration. Hence the best FM properties should occur for the low percent Mn sample as seen in our case. This appears to be a much satisfactory prediction of the FM ordering in Mn doped ZnO.

4. Conclusion The electronic structure of two polycrystalline ZnMnO pellets doped with diluted Mn concentration (2% and 4%) has been studied. The XRD patterns confirmed the ZnO lattice with Wurtzite hexagonal symmetry with no detectable impurities. The SQUID data show that the samples exhibit different magnetic properties. The 2% sample shows a room temperature FM ordering, the 4% sample did not show any ordering down to 5 K. The electronic structure of these samples was investigated using X-ray absorption and X-ray photoemission spectroscopy. The XPS, XAS results show that most of the Mn ions of the ferromagnetic sample are in the divalent state. For the non-magnetic sample, a larger contribution of higher oxidation Mn states (3+, 4+) is present and the oxygen content also shows an increment. Our studies reveal that the suppression of ferromagnetism can be associated with higher oxidation Mn state and the higher oxygen vacancy density. More studies are, however, required to conclude that which of these two factors weighs more.

Acknowledgments One of us (R.K.S) is grateful to the TWAS, Italy, and CNPq Brazil for the associate membership and CBPF Brazil for providing the hospitality and all the experimental facilities during the SQUID measurements. Part of this work was performed at UGC-DAE Consortium for Scientific Research, Indore, India. We are thankful to DST, New Delhi, and ELETTRA Sincrotrone, Trieste, Italy, for funding this work under the Indo-Italian POC in S&T. References [1] T. Dietl, H. Ohno, F. Matsukura, J. Cibert, D. Ferrand, Science 287 (2000) 12019. [2] P. Bobadova-Parvanova, K.A. Jackson, S. Srinivas, M. Horoi, J. Chem. Phys. 122 (2005) 014310. [3] K. Rode, R. Mattana, A. Anane, V. Cros, E. Jacquet, J.-P. Contour, F. Petroff, A. Fert, Appl. Phys. Lett. 92 (2008) 012509. [4] K. Ueda, H. Tabata, T. Kawai, Appl. Phys. Lett. 79 (2001) 988. [5] S.A. Chambers, T. Droubay, C.M. Wang, A.S. Lea, R.F.C. Farrow, L. Folks, V. Deline, S. Anders, Appl. Phys. Lett. 82 (2003) 1257. ¨ [6] M.A. Garc´ıa, M.L. Ruiz-Gonza lez, A. Quesada, J.L. Costa-Kramer, J.F. Ferna ndez, S.J. Khatib, A. Wennberg, A.C. Caballero, M.S. Mart´ın-Gonza lez, M. Villegas, F. Briones, J.M. Gonza lez-Calbet, A. Hernando, Phys. Rev. Lett. 94 (2005) 217206. [7] M. Gacic, G. Jakob, C. Herbort, H. Adrian, T. Tietze, S. Brueck, E. Goering, Phys. Rev. B 75 (2007) 205206. [8] A. Barla, G. Schmerber, E. Beaurepaire, A. Dinia, H. Bieber, S. Colis, F. Scheurer, ¨ J.P. Kappler, P. Imperia, F. Nolting, F. Wilhelm, A. Rogalev, D. Muller, J.J. Grob, Phys. Rev. B 76 (2007) 125201. [9] Y.D. Park, A.T. Hanbicki, S.C. Erwin, C.S. Hellberg, J.M. Sullivan, J.E. Mattson, T.F. Ambrose, A. Wilson, G. Spanos, B.T. Jonker, Science 295 (2002) 651. [10] F. Bondino, K.B. Garg, E. Magnano, E. Carleschi, M. Heinonen, R.K. Singhal, S.K. GaurF. Parmigiani, J. Phys. Condens. Matter 20 (2008) 275205. [11] T. Fukumora, Y. Yamada, H. Toyosaki, T. Hasegawa, H. Koinuma, M. Kawasaki, Appl. Surface Sci. 223 (2004) 62. [12] S.W. Jung, S.J. An, G.-C. Yi, C.U. Jung, S.-i. Lee, S. Choe, Appl. Phys. Lett. 80 (2002) 4561. [13] R.K. Singhal, S.N. Dolia, M.S. Dhawan, S.K. Gaur, S. Kumar, Y.T. Xing, E. Saitovitch, K.B. Garg, Internat. J. Mod. Phys. B (2009) 2029. [14] N.H. Hong, J. Magn. Magn. Mater. 303 (2006) 338. [15] S. Chambers, Mater. Today (2002) 34. [16] H. Ohno, F. Matsukura, Y. Ohno, JSAP Internat. 5 (2002) 4. [17] S.A. Wolf, D.D. Awschalom, R.A. Buhrman, J.M. Daughton, S. von Molnar, M.L. Roukes, A.Y. Chtchelkanova, D.M. Treger, Science 294 (2001) 1488. [18] D. Chiba, M. Yamanouchi, F. Matsukara, H. Ohno, Science 301 (2003) 943. [19] P. Sharma, A. Gupta, K.V. Rao, F.J. Owens, R. Sharma, R. Ahuja, J.M. Osro Guillen, B. Johansson, G.A. Gehring, Nature Mater. 2 (2003) 673. [20] P. Sharma, A. Gupta, F.J. Owens, A. Inoue, K.V. Rao, J. Magn. Magn. Mater. 282 (2004) 115. [21] H.J. Blythe, R.M. Ibrahim, G.A. Gehring, J.R. Neal, A.M. Fox, J. Magn. Magn. Mater. 283 (2004) 117. [22] M.H. Kane, K. Shalini, C.J. Summers, R. Vartharajan, J. Nause, C.R. Vestal, Z.J. Zhang, I.T. Ferguson, J. Appl. Phys. 97 (2005) 23906. [23] K.R. Kittilstved, N.S. Norberg, D.R. Gamelin, Phys. Rev. Lett. 94 (2005) 147209. [24] P. Long, Z. H-Wu, W. Qi-Ye, S. Yuan-Qiang, S. Hua, J.Q. Xiao, Chinese Phys. Lett. 25 (2008) 1438. [25] M. Venkatesan, C.B. Fitzgerald, J.M.D. Coey, Nature 430 (2004) 630. [26] D.P. Pemmaraju, S. Sanvito, Phys. Rev. Lett. 94 (2005) 217205. [27] N.H. Hong, J. Sakai, N. Poirot, V. Brize , Phys. Rev. B 73 (2006) 132404. [28] A. Sundaresan, R. Bhagravi, N. Rangarajan, U. Siddesh, C.N.R. Rao, Phys. Rev. B 74 (2006) 161306. [29] J. Rodriguez-Carvajal, FULLPROF version 3.0.0, Laboratorie Leon Brillouin, CEA-CNRS, 2003. [30] N.H. Hong, J. Sakai, V. Brize , J. Phys. Condens. Matter 19 (2007) 036219. [31] S.J. Pearton, D.P. Norton, K. Ip, Prog. Mater. Sci. 50 (2005) 293. [32] S. Kolesnik, B. Dabrowski, J. Mais, J. Super. 15 (2002) 251. [33] C.J. Cong, K.L. Zhang, Phys. Stat. Sol. (b) 243 (12) (2006) 2764. [34] D. Iusan, B. Sanyal, O. Eriksson, Phys. Rev. B 74 (2006) 235209. [35] M.A. Ruderman, C. Kittel, Phys. Rev. 96 (1954) 99. [36] K. Yosida, Phys. Rev. 106 (1957) 893; K. Yosida, Phys. Rev. 112 (1958) 1649; K. Yosida, Phys. Rev. 126 (1962) 1470.