Stable metal–insulator transition in epitaxial SmNiO3 thin films

Journal of Solid State Chemistry 190 (2012) 233–237

Contents lists available at SciVerse ScienceDirect

Journal of Solid State Chemistry journal homepage: www.elsevier.com/locate/jssc

Stable metal–insulator transition in epitaxial SmNiO3 thin films Sieu D. Ha n, Miho Otaki, R. Jaramillo, Adrian Podpirka, Shriram Ramanathan School of Engineering and Applied Sciences, Harvard University, Cambridge, MA 02138, USA

a r t i c l e i n f o

abstract

Article history: Received 14 January 2012 Received in revised form 16 February 2012 Accepted 17 February 2012 Available online 25 February 2012

Samarium nickelate (SmNiO3) is a correlated oxide that exhibits a metal–insulator transition (MIT) above room temperature and is of interest for advanced electronics and optoelectronics. However, studies on SmNiO3 thin films have been limited to date, in part due to well-known difficulties in stabilizing the Ni3 þ valence state during growth, which are manifested in non-reproducible electrical characteristics. In this work, we show that stable epitaxial SmNiO3 thin films can be grown by rf magnetron sputtering without extreme post-deposition annealing conditions using relatively high growth pressure (4 200 mTorr). At low growth pressure, SmNiO3 is insulating and undergoes an irreversible MIT at  430 K. As pressure is increased, films become metallic across a large temperature range from 100 to 420 K. At high pressure, films are insulating again but with a reversible and stable MIT at  400 K. Phase transition properties can be continuously tuned by control of the sputtering pressure. & 2012 Elsevier Inc. All rights reserved.

Keywords: Samarium nickelate SmNiO3 Metal–insulator transition High pressure synthesis

1. Introduction There is growing interest in investigating complex transition metal oxides for advanced electronics, such as neuromorphic circuits, nanoelectronics, and solid-state energy conversion. Among these materials, correlated oxides that exhibit a temperature-induced metal–insulator phase transition (MIT) are currently being studied. The family of rare-earth nickelates (RNiO3, R¼La, Nd, Gd, etc.) displays MI transitions for all RaLa, where the MIT temperature (TMIT) increases with increasing R atomic number (decreasing cation radius) [1]. The distorted perovskite-structure RNiO3 is insulating below and metallic above TMIT, with a change in resistivity that can be several orders of magnitude around TMIT. Recent work suggests that the MIT is caused by charge disproportionation (2Ni3þ Ni3 þ d þNi3 d) below TMIT, which opens a charge transport gap [2,3]. In the RNiO3 series, SmNiO3 (SNO) is the first material to have TMIT above room temperature (400 K in the bulk) and, therefore, has potential for incorporation into existing semiconductor platforms. It is known that the entire RNiO3 series is difficult to synthesize, with high pressure oxygen annealing (4100 bar, 41000 K) typically required to achieve stoichiometric powders [1]. The difficulty arises because the stoichiometric Ni3 þ valence state in RNiO3 is metastable, with Ni2 þ more likely to form. Pure RNiO3, synthesized under high oxygen pressure, decomposes in air into R2O3 and NiO at temperatures that decrease with decreasing R cation radius (1123 K for SNO) [4]. Thin film growth by rf sputtering is also challenging, with very high sputtering background pressure ( 200 mTorr) or

n

Corresponding author. E-mail address: [email protected] (S.D. Ha).

0022-4596/$ - see front matter & 2012 Elsevier Inc. All rights reserved. doi:10.1016/j.jssc.2012.02.047

high pressure oxygen post-annealing required to stabilize the MIT, for example, in the related compound NdNiO3 [5,6]. In contrast, sputtering pressures for most materials are generally in the tens of mTorr range. Stabilized SNO thin films grown by sputtering have not been demonstrated to date, to the best of our knowledge. Low pressures typically used for sputtering growth lead to an irreversible MIT in SNO, in which resistivity increases continuously for temperatures above TMIT due to oxygen loss [7]. This limits interest for many applications that require recovery of the initial resistance state and further complicates fundamental physical studies. A reversible phase transition is needed for many devices, such as switches, sensors, oscillators, and photonic applications [8]. Studying the detailed physics of the MIT in thin films requires stable film properties as well. Here, we disclose high pressure (4200 mTorr) growth of epitaxial SNO thin films with stable MIT properties. We show that modifying the deposition pressure can alter the electrical characteristics of SNO from insulating at low pressure to metallic at moderate pressure to insulating again at high pressure. In addition, we demonstrate that stable MIT characteristics of SNO, such as TMIT and resistivity ratio, can be widely tuned at high pressures. Synthesis of stable SNO with straight-forward deposition procedures may contribute to greater research of fundamental physics in correlated oxides and advance exploration towards room-temperature correlated electronics.

2. Experimental SmNiO3 thin films were grown by rf magnetron sputtering from a sintered ceramic target in a flow of Ar and O2 gas. The deposition temperature and target rf power were maintained at

S.D. Ha et al. / Journal of Solid State Chemistry 190 (2012) 233–237

923 K and 200 W respectively for all deposition runs. The sputtering gas background pressure was varied from 10–400 mTorr. Film thicknesses varied from 10–100 nm, depending on sputtering pressure. Films were cooled in ambient, and no post-deposition annealing was performed. All films were grown on (001)-oriented single crystal LaAlO3 (LAO) substrates purchased from MTI Corporation. LAO was chosen because SNO is known to grow epitaxially on LAO, thereby reducing experimental uncertainty from polycrystalline phases [9]. X-ray diffraction (XRD) was performed on a Bruker D8 Discover diffractometer system with Cu Ka radiation. X-ray reflectivity (XRR) was used to calibrate film thickness. Atomic force microscopy (AFM) images were measured on an MFP-3D system from Asylum Research. Resistivity measurements were made in the van der Pauw geometry using an MMR Variable Temperature Hall System. Sputtered Pt was photolithographically patterned to define square electrical contacts.

SmNiO3

LaAlO3 (002)

(002)

(Kα1, Kα2)

log(counts) (arb. u.)

234

10 mTorr

190 mTorr

250 mTorr

44

46

48

3. Results and discussion

Resistivity (Ω-cm)

3.1. SmNiO3 vs. sputtering pressure We have measured 2y–y XRD patterns of SNO films grown at three different sputtering pressures from 10 to 250 mTorr. Only features related to SNO or LAO were observed from 2y ¼201 to 801, indicating phase purity. In Fig. 1a, we focus on the (002) pseudocubic reflection of SNO, which appears near the (002) pseudocubic reflection of LAO. In the bulk, both reflections occur at 48.01, but due to in-plane compressive strain (   0.15%) in thin films, the SNO peak appears at lower 2y angle, corresponding to higher out-of-plane lattice constant [10,11]. As a point of reference, the bulk pseudocubic c-axis lattice constant of SNO is ˚ It is clear that as deposition pressure is increased, the 3.784 A. out-of-plane lattice constant decreases (see Table 1). Aside from epitaxial strain, the lattice constant is also affected by stoichiometry, which is a function of sputtering pressure. As oxygen content is reduced, the valence state of Ni is modified from 3 þ to 2þ, corresponding to a larger Ni ionic radius and expansion of the unit cell. Indeed, it has been reported in NdNiO3  d bulk samples that all unit cell parameters increase with increasing d [12]. Therefore, we hypothesize that the reduction in out-of-plane lattice parameter with increasing sputtering pressure is due to an increase in oxygen concentration and increase in ratio of Ni3 þ to Ni2 þ . The stabilization of the Ni3 þ valence state at high sputtering pressures agrees well with the requirement of high oxygen pressure annealing for bulk powder synthesis [1]. Resistivity for each sample is plotted as a function of temperature in Fig. 1b. It can be seen that the sample grown at 10 mTorr is insulating at low temperatures and appears to exhibit a phase transition with increasing temperature at 430 K. However, upon cooling, the resistivity increases monotonically and does not follow the heating curve. This behavior cannot be characterized as hysteretic because the final resistivity after cooling is larger than the initial resistivity and is stable for days. The initial resistivity is irrecoverable. We have shown previously that this change in resistivity is due to irreversible oxygen loss induced by the MIT and that it occurs because of low oxygen concentration in films grown at 10 mTorr [7]. If the temperature is held above TMIT, oxygen reduction will continue for some time and resistivity will continuously increase. In this regard, the metallic phase cannot be regarded as stable. From Fig. 1b, films grown at 190 mTorr display markedly different electrical characteristics than films grown at 10 mTorr. Over the temperature range 200–420 K, the r–T plot is metallic and reversible in nature with no signature of a phase transition.

50

52

2θ (°)

6x10-3

0.05

4x10-3

0.045

2x10-3

250 mTorr

400

10 mTorr

420

440

0.2

0.1 0.08 0.06 0.04

10-3 8x10-4 6x10-4

Heat Cool

190 mTorr

0.02

4x10-4 200

250

300

350

400

450

Temperature (K) Fig. 1. (A) XRD patterns of the (002) pseudocubic reflection of SmNiO3 films as a function of rf sputtering pressure. Plots offset for clarity. Two peaks from the (002) pseudocubic reflection of LaAlO3 are due to contributions from Cu Ka1 and Ka2 lines. (B) r–T measurements of SmNiO3 films as a function of sputtering pressure. The topmost curve corresponds to films grown at 10 mTorr and is referenced to the right hand ordinate axis. The 190 and 250 mTorr films are referenced to the left hand ordinate axis. Solid lines are data recorded during heating and dashed lines are data recorded during cooling.

Table 1 Structural and electrical properties of SmNiO3 films as a function of sputtering pressure. Lattice constant in reference to pseudocubic structure. Transition temperature is defined as point at which dr/dT changes sign. Growth pressure (mTorr)

Oxygen partial pressure (mTorr)

Out-of-plane lattice Electrical characteristics ˚ constant (A)

10

5

4.00

190

38

3.86

250

50

3.81

Irreversible MIT at 430 K Metallic from 100 to 420 K Reversible MIT at 405 K

The films remain metallic down to 100 K from electrical transport measurements (see Fig. 4). The room temperature resistivity of the film grown at 190 mTorr (8.35  10  4 O-cm) is about 2 orders

S.D. Ha et al. / Journal of Solid State Chemistry 190 (2012) 233–237

expansion above TMIT favors the larger Ni2 þ radius, causing growth of reduced Ni2 þ regions.

3.2. High pressure sputtered SmNiO3 Epitaxial growth of SmNiO3 at 250 mTorr sputtering pressure is confirmed in the (113) pseudocubic reflection j-scan shown in Fig. 2a. Four equally spaced peaks indicate that the film is indeed pseudocubic. Moreover, when plotted with a j-scan of the (111) pseudocubic reflection of the LAO substrate, it can be seen that the film is epitaxially indexed to the substrate. This implies cubeon-cube growth of the film, as expected [18]. An AFM micrograph of the SNO film grown at 250 mTorr is given in Fig. 2b. The film appears relatively smooth, with an RMS rough˚ There are occasional peaks in the film (2 such bright ness of  5 A. spots in Fig. 2b) that are above 3 nm in height, but otherwise the range of the height image is from  1.3 to 2.0 nm. This is well within the 10 nm thickness of the film as determined by XRR. Therefore, within the resolution of the AFM image, the SNO films grown at 250 mTorr are contiguous and not significantly rough, which is an encouraging result for high pressure growth. A r–T plot of the 250 mTorr SNO film is shown in Fig. 3 over a wider range than in Fig. 1b. A reversible MIT is evident in the figure with a change in resistivity of over two orders of magnitude from the insulating phase at 100 K to the metallic phase. The MIT temperature in RNiO3 is typically reported in the literature in one of two ways [19]. In one definition (TMIT(i)), the MIT temperature is given empirically as the temperature at which there is a peak in d(ln(r))/dT, which corresponds to the point of steepest resistivity change. In the alternate, more fundamental definition (TMIT(ii)), the MIT occurs at the temperature at which dr/dT changes sign from negative (insulating phase) to positive (metallic phase). This is the definition used in Table 1. In bulk SNO, both of these values are within 10–15 K [15]. In thin films, the sharpness of the resistivity transition depends on growth conditions, e.g. through oxygen

log(counts) (arb. u.)

of magnitude lower than that of the film grown at 10 mTorr (8.29  10  2 O-cm). The r–T here is approximately linear in the 200–420 K range, suggesting that electron–phonon scattering is the primary source of temperature dependence. The extracted temperature coefficient dr/dT is 8.4  10  7 O-cm/K, well within the range of bulk stoichiometric SmNiO3 in the metallic phase (2.7  10  6 O-cm/K) [13]. SNO films grown at 250 mTorr most closely resemble bulk samples with respect to r–T characteristics, as seen in Fig. 1b. These films exhibit an MIT at  400 K upon heating that is completely reversible and stable, unlike films grown at 10 mTorr. The precise overlap of heating and cooling curves naturally depends on ramp rates, but given sufficient time, the final resistivity will match the initial resistivity. We have found negligible change in r–T after more than 10 heating and cooling cycles. Further details of the r–T behavior are discussed in the subsequent section. From Fig. 1b, it is clear that sputtering in high background pressure can well stabilize the phase transition of SNO. Properties of films grown at 10, 190, and 250 mTorr are given in Table 1. There is a wide variation in SNO electrical properties as a function of sputtering pressure. This is despite no signature of extraneous oxide phases in XRD experiments, only contraction/expansion of the out-of-plane lattice constant, which implies varying levels of oxygen concentration. At 250 mTorr, the SNO MIT is stabilized and the resistivity behavior is in good agreement with the literature [14]. We can therefore assume that, of all three samples, the stoichiometry of SNO is most optimized in the 250 mTorr film. When the growth pressure is decreased to 190 mTorr, the MIT appears suppressed, at least to below 100 K. We cannot rule out the possibility of a phase transition occurring below 100 K [15], which is the lower temperature limit of our measurements. Lowering of TMIT and/or suppression of the phase transition in RNiO3 has been observed before by Cadoping in SNO films and oxygen reduction in bulk NdNiO3 [12,14]. The phase transition is usually suppressed entirely for high defect concentrations. Thus, suppression of the MIT in SNO films grown at 190 mTorr can be understood based on oxygen reduction, which introduces donor-like defects in the band gap and lowers TMIT [16]. In addition, expansion of the unit cell with respect to the film grown at 250 mTorr may reduce the Ni–O octahedra distortion, which also serves to lower TMIT, as in RNiO3 with larger rare-earth cations (e.g. R¼La, Pr, Nd). By adjusting the sputtering pressure, it could therefore be possible to tune TMIT over a broad temperature range, which we demonstrate and discuss further in Section 3.3. The transition from metallicity in the 190 mTorr film back to an insulating phase in the 10 mTorr film is non-intuitive. However, precisely the same behavior has been reported in Co-doped SNO (i.e. SmNi1  xCoxO3) [15]. In that work, as x is increased from 0 to 0.20, TMIT monotonically decreases from 400 to 25 K, significantly widening the temperature range of the metallic phase. For xZ0.3, Co-doped SNO is insulating from 10 to 500 K and metallic above 500–600 K. The resistivity of these samples can be fit to a constant activation energy over a wide temperature range, similar to SNO films grown by sputtering at 10 mTorr [17]. Although the SNO films grown in our work are not intentionally doped, the diminished oxygen concentration at low sputtering pressures seems to have a similar effect as substitutional Ni-site doping. It has been suggested that local structural or magnetic disorder may be responsible for driving the film into an insulating state [15]. Although our results agree qualitatively well with the Co-doped SNO work, we observe oxygen loss above TMIT in the 10 mTorr film that is not reported in the referenced work, in which samples are doped but have near-stoichiometric oxygen content. As we have previously suggested [7], it may be possible that low initial oxygen content in films grown at 10 mTorr is not sufficient to stabilize the Ni3 þ valence state, and that the structural phase change and corresponding unit cell volume

235

Sm NiO3

LaAlO3

-200 -150 -100 -50

ϕ (°)

0

50

100 150

Fig. 2. (A) j-scan of the (111) pseudocubic reflection of the LaAlO3 substrate and the (113) pseudocubic reflection of SmNiO3 film sputtered at 250 mTorr. Plots offset for clarity. (B) AFM micrograph (1.0  1.0 mm2) of SmNiO3 film sputtered at 250 mTorr.

236

S.D. Ha et al. / Journal of Solid State Chemistry 190 (2012) 233–237

103

450

400 MIT(ii) (K)

d(ln(ρ))/dT

10-2

T

325

290

310

Normalized resistivity

Resistivity (Ω-cm)

TMIT(i)

330

10

400

350

2 300 220

280

340

400

p

sputter (mTorr)

250

101

T ↑ T ↓ 225

10-3

TMIT(ii) Heating Cooling

100

150

200

100 190

250

300

350

400

450

100

150

stoichiometry and epitaxial strain, and TMIT(i) may be considerably lower than TMIT(ii). From Fig. 3, we can extract TMIT(i) ¼316 K (see inset) and TMIT(ii) ¼405 K. The value for TMIT(ii) agrees very well with the MIT temperature of bulk SNO, while TMIT(i) is lower. For both definitions of TMIT, we observe negligible hysteresis (0–2 K) between heating and cooling curves (1 K/min ramp rate, 1 min soak), in agreement with films produced by metal-organic chemical vapor deposition and pulsed-laser deposition [20,21], but in contrast to bulk samples, which show  20 K hysteresis around TMIT [13]. The reduction in hysteresis in thin films with respect to bulk samples has been reported previously for NdNiO3 [22]. The authors state that inhomogeneities in local strain due to mechanical and chemical defects in thin films broaden the phase transition and reduce the temperature range over which metallic and insulating phases can co-exist, thereby suppressing the hysteresis window. This effect may occur in SNO thin films as well, independent of deposition technique. Another feature of Fig. 3 is the kink in the resistivity curve that is observed at  190 K (see arrow), which is somewhat close to the Ne´el temperature (TN) of bulk SNO (220 K). This feature is generally observed in bulk and thin film SNO and it has been suggested that it may be related to the antiferromagnetic ordering at TN [14]. Independent measurements of TN and r–T as TMIT is modified by doping show that the kink in r–T occurs at TN for low doping concentration [15]. Therefore, the feature we observe at 190 K in r–T measurements of the 250 mTorr films may be related to antiferromagnetic ordering.

250

300

350

400

450

Temperature (K)

Temperature (K) Fig. 3. Wide range r–T measurement of SmNiO3 film sputtered at 250 mTorr. TMIT(i) and TMIT(ii), as defined in the text, are noted. Solid lines are data recorded during heating and dashed lines are data recorded during cooling. Arrow at  190 K highlights kink that may be related to antiferromagnetic ordering. Inset: d(ln(r))/dT plotted over narrow T range to illustrate TMIT(i) and low hysteresis.

200

Fig. 4. Normalized r–T characteristics of SmNiO3 films grown over a wide sputtering pressure range. Curves are normalized to high temperature resistivity. Numerals on left side of figure are sputtering pressures in units of mTorr. Solid lines are data recorded during heating and dashed lines are data recorded during cooling. Inset: MIT temperature TMIT(ii) as a function of sputtering pressure.

tuned as a function of sputtering pressure. As noted above, films grown at 190 mTorr show metallic r–T characteristics over the entire measurement range from 100 to 420 K. As sputtering pressure is increased slightly to 225 mTorr, SNO films become semiconducting at low temperatures with a reversible metal–insulator phase transition near room temperature at  315 K. The ratio between resistivity at 100 K and TMIT is relatively low, with Dr(225 mTorr) (defined here as r(100 K)/r(TMIT)) 3. At 250 mTorr, films are semiconducting over a wide temperature range and exhibit a clear MIT at 405 K, as previously discussed. The resistivity change is significantly larger for films grown at 250 mTorr as opposed to 225 mTorr, with Dr(250 mTorr) 100. For films grown at sputtering pressures above 250 mTorr, the r–T properties are qualitatively similar, but with appreciable changes in TMIT(ii) and Dr. The resistivity ratio increases to Dr(325 mTorr) 600 and Dr(400 mTorr)1000. In addition, TMIT(ii) shifts to higher temperature with increasing pressure, to approximately 410 K at 325 mTorr and 440 K at 400 mTorr. The change in phase transition temperature as a function of sputtering pressure is plotted in the inset of Fig. 4. For all SNO films deposited at high pressure, we find minimal hysteresis between heating and cooling cycles. It is evident from Fig. 4 that the electrical properties of SNO thin films can be widely tuned by varying the background pressure during rf sputtering, which likely affects the oxygen content and Ni valence state. Room temperature resistivity varies by more than one order of magnitude and Dr by nearly three orders of magnitude across the examined pressure range. In addition, the phase transition temperature can be altered by well over 100 K above room temperature.

3.3. SmNiO3 phase transition tunability Resistivity as a function of temperature for films sputtered in background pressures from 190 to 400 mTorr is plotted in Fig. 4. In the figure, each curve is normalized to the respective high temperature resistivity value to facilitate comparison. It can be seen that the electrical MIT properties of SNO can be strongly and continuously

4. Conclusions We have studied structural and electrical properties of epitaxial SmNiO3 grown on LaAlO3 by rf magnetron sputtering as a function of background pressure. As pressure increases, the out-of-plane

S.D. Ha et al. / Journal of Solid State Chemistry 190 (2012) 233–237

lattice parameter decreases, consistent with greater oxygen incorporation. This serves to stabilize the metal–insulator phase transition in SmNiO3 films grown at 250 mTorr. Films grown at 190 mTorr are metallic in the measured range from 100 to 420 K, with oxygen vacancies either shifting the transition temperature below 100 K or suppressing the phase transition entirely. Films grown at 10 mTorr are insulating from 200 to 430 K, at which point the phase transition induces an irreversible oxygen reduction in the film, increasing the resistivity. High pressure growth does not cause large roughness or porosity in SmNiO3 films, and epitaxial growth is maintained. Stable phase transition properties can be widely tuned as pressure is increased from 190 to 400 mTorr. In agreement with other thin film growth techniques, we do not observe appreciable hysteresis in the high pressure films. Stabilization of the phase transition in SmNiO3 using a standard deposition technique without complicated annealing procedures may allow for incorporation of SmNiO3 and related materials into advanced electronic platforms.

Acknowledgments We gratefully acknowledge the ARO MURI (Grant no. W911NF-09-1-0398), NSF Grant DMR-0952794, and the Focus Center Research Program in the Materials Structures and Devices Focus Center for financial support. This work was performed in part at the Center for Nanoscale Systems at Harvard University, which is supported under NSF award ECS-0335765. References [1] G. Catalan, Phase Transitions 81 (2008) 729–749.

237

[2] M. Medarde, C. Dallera, M. Grioni, B. Delley, F. Vernay, J. Mesot, M. Sikora, J.A. Alonso, M.J. Martı´nez-Lope, Phys. Rev. B 80 (2009) 245105. [3] I.I. Mazin, D.I. Khomskii, R. Lengsdorf, J.A. Alonso, W.G. Marshall, R.M. Ibberson, A. Podlesnyak, M.J. Martı´nez-Lope, M.M. Abd-Elmeguid, Phys. Rev. Lett. 98 (2007) 176406. [4] M.J. Martı´nez-Lope, J.A. Alonso, Eur. J. Solid State Inorg. Chem. 32 (1995) 361–371. [5] R. Scherwitzl, P. Zubko, I.G. Lezama, S. Ono, A.F. Morpurgo, G. Catalan, J.M. Triscone, Adv. Mater. 22 (2010) 5517–5520. [6] P. Laffez, M. Zaghrioui, R. Retoux, P. Lacorre, J. Magn. Magn. Mater. 211 (2000) 111–117. [7] S.D. Ha, G.H. Aydogdu, S. Ramanathan, Appl. Phys. Lett. 98 (2011). (012105-3). [8] Z. Yang, C. Ko, S. Ramanathan, Annu. Rev. Mater. Res. 41 (2011) 337. [9] G.H. Aydogdu, S.D. Ha, B. Viswanath, S. Ramanathan, J. Appl. Phys. 109 (2011). (124110-6). [10] Y.-D. Yu, S.-S. Xie, H. Boysen, Acta Phys. Sin. 42 (1993) 605. [11] P. Lacorre, J.B. Torrance, J. Pannetier, A.I. Nazzal, P.W. Wang, T.C. Huang, J. Solid State Chem. 91 (1991) 225–237. [12] I.V. Nikulin, M.A. Novojilov, A.R. Kaul, S.N. Mudretsova, S.V. Kondrashov, Mater. Res. Bull. 39 (2004) 775–791. [13] J. Pe´rez-Cacho, J. Blasco, J. Garcı´a, M. Castro, J. Stankiewicz, J. Phys.: Condens. Matter 11 (1999) 405. [14] P.H. Xiang, S. Asanuma, H. Yamada, I.H. Inoue, H. Akoh, A. Sawa, Appl. Phys. Lett. 97 (2010). (032114-3). [15] J. Pe´rez-Cacho, J. Blasco, J. Garcı´a, J. Stankiewicz, Phys. Rev. B 59 (1999) 14424. [16] J.L. Garcı´a-Mun˜oz, M. Suaaidi, M.J. Martı´nez-Lope, J.A. Alonso, Phys. Rev. B 52 (1995) 13563. [17] S.D. Ha, G.H. Aydogdu, S. Ramanathan, J. Appl. Phys. 110 (2011) 094102–094105. [18] N. Ihzaz, M. Oumezzine, J. Kreisel, H. Vincent, S. Pignard, Chem. Vapor Deposition 14 (2008) 111–114. [19] F. Capon, P. Laffez, J.F. Bardeau, P. Simon, P. Lacorre, M. Zaghrioui, Appl. Phys. Lett. 81 (2002) 619–621. [20] C. Girardot, J. Kreisel, S. Pignard, N. Caillault, F. Weiss, Phys. Rev. B 78 (2008) 104101. [21] A. Ambrosini, J.-F. Hamet, Appl. Phys. Lett. 82 (2003) 727–729. [22] D. Kumar, K.P. Rajeev, A.K. Kushwaha, R.C. Budhani, J. Appl. Phys. 108 (2010) 063503–063507.