- Email: [email protected]
The effect of Al substitution on the structural and magnetic properties of epitaxial thin films of epsilon ferrite
Scripta Materialia 140 (2017) 63–66
Contents lists available at ScienceDirect
Scripta Materialia journal homepage: www.elsevier.com/locate/scriptamat
Regular article
The effect of Al substitution on the structural and magnetic properties of epitaxial thin films of epsilon ferrite Luca Corbellini a,⁎, Christian Lacroix b, David Ménard b, Alain Pignolet a a b
Centre Énergie, Matériaux et Télécommunications, INRS, 1650 Boulevard Lionel-Boulet, Varennes, Québec J3X 1S2, Canada Département de Génie Physique & Regroupement Québécois sur les Matériaux de Pointe (RQMP), Polytechnique Montréal, Montréal, Québec H3T 1J4, Canada
a r t i c l e
i n f o
Article history: Received 23 March 2017 Received in revised form 26 May 2017 Accepted 8 July 2017 Available online xxxx Keywords: Thin film Epitaxy Pulsed laser deposition Magnetism Ferromagnetic resonance frequency
a b s t r a c t Epitaxial growth of aluminum substituted epsilon ferrite by pulsed laser deposition on strontium titanate and yttrium-stabilized zirconium substrates is demonstrated for a wide range of aluminum substitution. Linear decrease of the out-of-plane lattice parameter with aluminum substitution correlates with the decrease of the coercive magnetic field, while the magnetization first increases and then decreases with increasing aluminum concentration. A characteristic inflexion in the hysteresis loops has been attributed to the presence of a small quantity of magnetite. Epitaxial growth on yttrium-stabilized zirconium is technologically significant due to its use as buffer layer for the growth of complex oxides on silicon. © 2017 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
Nanostructured iron oxides have recently drawn a lot of attention given their many potential applications, first based on their inherent magnetic properties, but also in various fields of medicine because of their well-known biochemical characteristics of nontoxicity, biocompatibility, and biodegradability [1–13]. Iron oxides are the most common iron compounds in nature and several of them, such as magnetite (Fe3O4), maghemite (γ-Fe2O3) and hematite (α-Fe2O3) occur naturally and are generally easy to synthesize. A notable exception is epsilon ferrite (ε-Fe2O3), a ferrimagnetic metastable phase intermediate between maghemite and hematite. The magnetization in εFe2O3 arises from a non-fully compensated magnetic moment on the anti-parallel aligned iron atoms (see Fig. 1), in which the tetrahedrally coordinated “D” sites are holding a lower magnetic moment than the other sites. Although being characterized by lower magnetization compared to γ-Fe2O3, ε-Fe2O3 has received notable attention lately due to its very strong magnetocrystalline anisotropy, which results in a gigantic coercive field [14–17], and also in a natural ferromagnetic resonance (FMR) frequency above 100 GHz, flirting with the low THz range (~0.1–100 THz) at room temperature [18]. This is of particular interest given its potential use in short-range wireless communications (e.g. 60 GHz WiFi) [19] and ultrafast non-volatile memories [20]. The origin of such strong magnetocrystalline anisotropy was investigated by Xray magnetic circular dichroism and found to be due to lattice distortion [21]. Later, first principle calculations supported those experimental findings, by predicting that the magnetic easy axis should lie along the ⁎ Corresponding author. E-mail address: [email protected] (L. Corbellini).
http://dx.doi.org/10.1016/j.scriptamat.2017.07.005 1359-6462/© 2017 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
crystal a axis direction [22]. Finally, given its crystal structure belonging to the polar space group Pna21 [23,24] (Fig. 1), ε-Fe2O3 is characterized by a spontaneous polarization that, if proved switchable at room temperature, would make it one of the few single phase, room temperature multiferroic materials. Up until recent years, due to its metastability in ambient condition, epsilon ferrite has mostly been synthesized in the form of nanostructure [15–17,25–27], with size confinement acting as the stabilizing factor. However, in order to further investigate their relatively unexplored magnetic properties and eventually integrate them into devices, synthesis of stable epsilon ferrite thin films and their compounds must be mastered. Recent reports showed that it is indeed possible to grow epsilon ferrite thin films epitaxially directly on single crystalline strontium titanate (SrTiO3) [28] and yttrium stabilized zirconium (YSZ) substrates by pulsed laser deposition (PLD) [29], and on sapphire by using a buffer layer of gallium ferrite (GaFeO3) [30]. Furthermore, presence of reversible spontaneous polarization was reported in layered structure of SrTiO3:Nb/AlFeO3/SrRuO3/ε-Fe2O3 [31]. Extensive studies have also been conducted on the effect of metal substitution into epsilon ferrite (ε-MxFe2 − xO3) nanoparticles, in particular regarding its effect on the FMR frequency. Ohkoshi et al. have reported that substitution of Fe3+ ions in ε-Fe2O3 with metals having a smaller ionic radius, such as Al3 + and Ga3 +, results in a lowering of the FMR frequency [18,19,32], while through substitution with larger ions like Rh3+, an increase of the FMR is observed [33]. It is interesting to note that AlFeO3 and GaFeO3, which are isostructural to ε-Fe2O3, are known room temperature ferroelectrics characterized by spontaneous magnetic ordering at cryogenic temperatures [34–36].
64
L. Corbellini et al. / Scripta Materialia 140 (2017) 63–66
Fig. 1. ε-Fe2O3 unit cell and configuration of the magnetic moments on the four different iron sites. It has to be noted that the magnetic moments lie in the a–b plane. According to recent works, they should be aligned along the a-axis of the crystal unit cell.
In this paper, we report the epitaxial growth by pulsed laser deposition of thin films of Al-substituted epsilon ferrite and the effect of Al substitution on the structural and magnetic properties. Thin films of εAlxFe2 − xO3 of thickness of about 50 nm were grown directly on single crystalline (111)-oriented strontium titanate (SrTiO3) and (100)-oriented Yttrium Stabilized Zirconia (YSZ) substrates (no buffer layer was used). While the different symmetry of the two substrates give different epitaxial match and result in different twins formation [29,30], the films grown on both type of substrates exhibited very similar structural and functional properties and the same trends with increasing aluminum substitution. The films were grown following the conditions reported in reference 27; the only peculiarity was that, in order to obtain films with different percentage of Al, the growth was performed by alternatively ablating a standard Fe2O3 target and a second target of AlFeO3. Different Al substitution percentages were obtained by changing the ratio of the number of pulses shot on each target. In order to achieve a homogenous distribution of aluminum, the films were grown roughly one monolayer at the time, alternating the targets in order to get the desired composition. Films with Al concentration ranging from pure ε-Fe2O3 to ε-Al0.5Fe1.5O3 (0 b x b 0.5) were then prepared. Thin films of pure AlFeO3 (i.e. x = 1) were also prepared (with the same deposition conditions as the epsilon ferrite films) to estimate its growth rate and for comparison with the ε-AlxFe2 − xO3 films. In order to confirm the correctness of our estimation of the composition, which is based on the number of laser pulses on each target and on the growth rate of both AlFeO3 and ε-Fe2O3, energy-dispersive X-ray spectroscopy (EDX) was performed on our films. The Al:Fe ratio obtained was in very good agreement with the Al:Fe ratio estimated from the number of pulses shot on each target.
Structural analysis was performed by X-ray diffractometry with a PANalytical X'Pert MRD PRO diffractometer using the Cu Kα radiation (λ = 1.54056 Å). Goniometer (θ/2θ) scans confirmed that ε-AlxFe2 − xO3 grows epitaxially both on SrTiO3 (111) and YSZ (100). The films are (001)-oriented as shown by the diffraction patterns of pure ε-Fe2O3, ε-Al0.1Fe1.9O3, and pure AlFeO3 grown on SrTiO3 (111) (Fig. 2a). A shift towards higher 2θ angles with increasing Al concentration is observed, which was expected given the smaller ionic radius of Al3+ (53.5 pm) compared to Fe3+ (64.5 pm). The out-of-plane lattice parameter “c” was calculated from the measured goniometer scans, averaging the values found utilizing the inter-planar distances corresponding to the two detectable 00l peaks (l = 2 and 6). A linear decrease in the out-of-plane lattice parameter with increasing aluminum content was found, as shown in the graph in Fig. 2b, in agreement with previous studies [19]. The magnetic properties of our films were measured using a Vibrating Sample Magnetometer (VSM - model EV9 from ADE Technologies). The hysteresis loops were recorded at room temperature for all the samples, with the magnetic field applied both parallel (in-plane) and perpendicular (out-of-plane) to the film surface. In order to obtain solely the signal arising from the thin films, the contribution of the vibrating rod holding the samples and of the substrates were measured and subtracted from the measured hysteresis loops. The resulting magnetic moment was divided by the volume of the thin film. The measured magnetic hysteresis loops of films with x = 0, x = 0.1, and x = 0.5 are shown in Fig. 3. We observe an inflexion at low applied fields in all loops, which we attribute to the presence of a secondary soft magnetic phase that was identified as magnetite (Fe3O4) by detailed XRD and magnetic analysis [29]. Presences of these inflections were also recorded
Fig. 2. (a) θ/2θ scans of (100)-oriented ε-Fe2O3, ε-Al0.1Fe1.9O3, and AlFeO3 thin films, revealing how all the three materials grow epitaxially on (111)-oriented SrTiO3. The line drawn in grey helps visualizing the shift in position of the (006) peak in the Al-substituted films in comparison to pure ε-Fe2O3. The peaks noted by asterisks belong to the SrTiO3 (111) substrates and its secondary lines. (b) Evolution of the lattice parameter “c” with increasing Al concentration “x”.
L. Corbellini et al. / Scripta Materialia 140 (2017) 63–66
65
Fig. 3. (a) In-plane and out-of-plane hysteresis loop measured for a pure ε-Fe2O3 (orange and green), (b) a ε-Al0.1Fe1.9O3 (blue and red), and (c) for a ε-Al0.5Fe1.5O3 epitaxial thin film (burgundy and violet), all grown on YSZ (100). In all cases, we can infer the high magnetic anisotropy of the samples. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
for pure epsilon ferrite thin films grown on SrTiO3 (111) and YSZ (100) by our group, as well as by other groups [28,31], who, however, proposed a different explanation for their origin. Comparison between the three loops revealed that (i) the aluminum-substituted film preserved their in-plane anisotropy (see the difference between the in-plane and the out-of-plane measurement), (ii) are characterized by a lower coercive field in comparison to pure epsilon ferrite films, and (iii) the saturation magnetization depends on the aluminum concentration. Similar trends were observed in Al-substituted ε-Fe2O3 nanoparticles, that is, a decrease of the coercivity expected from the non-magnetic nature of the Al3+ ions (completely filled 2p6 shell) along with an increase of the magnetization at saturation up to a concentration of about 40% of Al [37]. Such an effect has also been reported for gallium-substituted nanoparticles of epsilon ferrite (which have a behavior similar to that of Al-substituted ε-Fe2O3 nanoparticles), where the increase of the magnetization is attributed to the preferential substitution of the iron located in the tetragonal site “D” of the lattice (Fig. 1) by the doping metal ions (gallium in ref. 35, aluminum in our case), due to their ionic radius smaller than that of Fe3+. Such a preferential substitution results in a decrease of the “D”-site magnetic moment, hence to an increase of the total magnetic moment, up to a point where the substitution starts to occur on other iron-sites as well, resulting then in a decrease of magnetic moment (see Fig. 1) [37]. In Fig. 4a, the temperature dependence of the magnetization is shown for the ε-Al0.1Fe1.9O3 film under a field of 1 kOe. A paramagnetic-toferromagnetic phase transition is observed at T ≈ 475 K. No phase transition is observed close to 300 K, which is the Curie temperature of AlFeO3 [35]. This indicates that aluminum substitution was successfully achieved, and that the resulting film is homogeneous, i.e. it does not consist of a mixture of AlFeO3 and ε-Fe2O3. Moreover, we observe that the magnetization does not fall to zero at temperatures higher than the Curie temperature of ε-Fe2O3 (TC ≈ 475 K), which is consistent with the presence of a small amount of magnetite (TC ≈ 850 K).
In order to accurately quantify the effect of aluminum substitution on the magnetic properties of epsilon ferrite, we used the technique described in ref. 27 to separate the contributions of the two magnetic phases (epsilon ferrite and magnetite). Using the values of saturation magnetization found in literature for magnetite (309 emu/cm3), we estimated the volume ratio of each phase (magnetite volume/epsilon ferrite volume), which was found to vary between 0.05 and 0.25 (average ≈ 0.1). For example, in samples with higher percentage of magnetite, as the one shown in Fig. 3b, the higher content of magnetite has the effect of decreasing the coercive field and increasing the saturation magnetization of the whole film. This allowed us to quantify the in-plane coercive field and the magnetization at saturation for the sole ε-AlxFe2 − xO3 phase for different Al concentrations (Fig. 4b and c). We observe that the coercive field decreases linearly as the aluminum concentration increases, suggesting that the magnetocrystalline anisotropy decreases linearly with aluminum concentration [32]. We note that both the magnetocrystalline anisotropy and the out-of-plane lattice parameter “c” decrease linearly with increased aluminum concentration. We also observe that, as the Al concentration increases, the saturation magnetization increases first by a factor of 2 (for x = 0.33), and then decreases for higher Al concentration. We propose that, as discussed in Ref. [19], the decrease of the saturation magnetization is due to the fact that, while Al substitution for x b 0.33 occurs at the iron “D” tetragonal site, further Al substitution above the threshold of x = 0.33 occurs both at the iron “B” and “C” octahedral sites, thus decreasing the total magnetization. For sake of completeness, in order to compare the magnetic properties of epsilon ferrite with that of the whole film (epsilon ferrite + magnetite), we also present the evolution of the coercive field and the saturation magnetization as a function of aluminum concentration for films with a “normalized” magnetite/epsilon volume ratio equal to 0.1, which is the average value found in our films. In view of potential applications, it is important to report the characteristic properties of the
Fig. 4. (a) Dependence of the magnetization (M) with the temperature for a ε-Al0.1Fe1.9O3 thin film grown on YSZ (100), recorded under a field of 1 kOe applied in the sample plane. (b) Evolution of the in-plane coercive field HC and (c) magnetization at saturation MS depending on the aluminum concentration “x”.
66
L. Corbellini et al. / Scripta Materialia 140 (2017) 63–66
whole ε-AlxFe2 − xO3 system. We observe that the effect of magnetite, which is characterized by a very low coercive field and high magnetization compared to epsilon ferrite, is to reduce the coercive field and to increase the magnetization at saturation of the whole film (Fig. 4b and c). This demonstrates the importance of carrying out a complete structural and magnetic characterization for this kind of epitaxial grown ferromagnetic system in order to extract accurate quantitative results. However, as the two magnetic phases do not appear to be interacting (see the angular dependency of the magnetic properties in pure epitaxial films of epsilon ferrite) [29], we expect a linear decrease of the natural ferromagnetic resonance frequency for ε-AlxFe2 − xO3 with increased aluminum concentration, as reported for al-substituted nanoparticles [19,37], along with a second ferromagnetic resonance frequency at lower frequencies due to magnetite, this in spite of the “effective” coercive field remaining roughly unchanged with Al substitution. In summary, epitaxial thin films of ε-AlxFe2 − xO3 were grown on SrTiO3 (111) and on YSZ (100). This latter achievement may constitute a decisive step towards integration of such material into consumer electronics, in particular its ability of growing epitaxially on YSZ (100), given how such oxide is often used as a buffer layer for the growth of complex oxides on silicon (100). Substitution over a wide range (x = 0.05 to 0.5, as well as x = 1) was achieved. Aluminum substitution linearly reduces the coercive field, and most probably, the magnetocrystalline anisotropy of epsilon ferrite, with increasing aluminum concentration. Moreover, an increase of the saturation magnetization up to a threshold of x = 0.33 was shown, similar to results previously reported for aluminum and gallium-substituted epsilon-ferrite nanoparticles. Finally, a characteristic inflexion at low fields in the hysteresis loops, also observed by other groups, has been however attributed to the presence of a secondary phase of magnetite rather than to other effects, such as surface effects. This work was supported by grants from the Natural Sciences and Engineering Research Council of Canada (NSERC - Canada) (# 2616622013) and the Fonds de Recherche du Québec - Nature et Technologies (FRQNT - Québec) (# 2013-PR-167502). References [1] [2] [3] [4]
A.-H. Lu, E.L. Salabas, F. Schuth, Angew. Chem. Int. Ed. 46 (2007) 1222–1244. W. Wu, Q.G. He, C.Z. Jiang, Nanoscale Res. Lett. 3 (2008) 397. A.S. Teja, P.-Y. Koh, Prog. Cryst. Growth Charact. Mater. 55 (2009) 22–45. J. Gonzalez, O. Chubykalo, J.M. Gonzalez, Encyclopedia of Nanoscience and Nanotechnology, American Scientific Publishers, Stevenson Ranch, CA, 2004. [5] R.M. Cornell, U. Schwertmann, The Iron Oxides: Structure, Properties, Reactions, Occurrence and Uses, Wiley-VCH, Weinheim, Germany, 2003. [6] J.W.M. Bulte, D.L. Kraitchman, NMR Biomed. 17 (2004) 484–499.
[7] C.C. Berry, A.S.G. Curtis, J. Phys. D Appl. Phys. 36 (2003) R198–R206. [8] A.K. Gupta, M. Gupta, Biomaterials 26 (2005) 3995–4021. [9] S. Laurent, D. Forge, M. Port, A. Roch, C. Robic, L.V. Elst, R.N. Muller, Chem. Rev. 108 (2008) 2064. [10] S. Mornet, S. Vasseur, F. Grasset, E. Duguet, J. Mater. Chem. 14 (2004) 2161–2175. [11] Q.A. Pankhurst, J. Connolly, S.K. Jones, J. Dobson, J. Phys. D Appl. Phys. 36 (2003) R167–R181. [12] P. Tartaj, M. del Puerto Morales, S. Veintemillas-Verdaguer, T. Gonzalez-Carreno, C.J. Serna, J. Phys. D Appl. Phys. 36 (2003) R182–R197. [13] D.L.J. Thorek, A.K. Chen, J. Czupyrna, A. Tsourkas, Ann. Biomed. Eng. 34 (2006) 23–38. [14] S. Ohkoshi, S. Sakurai, J. Jin, K. Hashimoto, J. Appl. Phys. 97 (2005), 10K312. [15] C. Savii, M. Popovici, C. Enache, J. Subrt, D. Niznansky, S. Bakardzieva, C. Caizer, I. Hrianca, M.P. Cecilia, Savii Corina Enache, J. Subrt, D. Niznansky, Snejana Bakardzieva, C. Caizer, I. Hrianca, Solid State Ionics 151 (2002) 219–227. [16] M. Kurmoo, J.-L. Rehspringer, A. Hutlova, C. D'Orléans, S. Vilminot, C. Estournès, D. Niznansky, Chem. Mater. 17 (2005) 1106–1114. [17] J. Jin, S. Ohkoshi, K. Ashimoto, Adv. Mater. 16 (2004) 48–51. [18] S. Ohkoshi, S. Kuroki, S. Sakurai, K. Matsumoto, K. Sato, S. Sasaki, Angew. Chem. Int. Ed. Engl. 46 (2007) 8392–8395. [19] A. Namai, S. Sakurai, M. Nakajima, T. Suemoto, K. Matsumoto, M. Goto, S. Sasaki, S. Ohkoshi, J. Am. Chem. Soc. 131 (2009) 1170–1173. [20] M. Nakajima, A. Namai, S. Ohkoshi, T. Suemoto, Opt. Express 18 (2010) 18260–18268. [21] Y.-C. Tseng, N.M. Souza-Neto, D. Haskel, M. Gich, C. Frontera, A. Roig, M. van Veenendaal, J. Nogués, Phys. Rev. B 79 (2009) 94404. [22] M. Yoshikiyo, K. Yamada, A. Namai, S. Ohkoshi, J. Phys. Chem. C 116 (2012) 8688–8691. [23] R. Zboril, M. Mashlan, D. Petridis, Chem. Mater. 14 (2002) 969–982. [24] E. Tronc, C. Chanéac, J.P. Jolivet, J. Solid State Chem. 139 (1998) 93–104. [25] J. Jin, K. Hashimoto, S. Ohkoshi, J. Mater. Chem. 15 (2005) 1067. [26] G. Carraro, D. Barreca, C. Maccato, E. Bontempi, L.E. Depero, C. de Julian Fernandez, A. Caneschi, CrystEngComm 15 (2013) 1039–1042. [27] G. Carraro, C. Maccato, A. Gasparotto, T. Montini, S. Turner, O.I. Lebedev, V. Gombac, G. Adami, G. Van Tendeloo, D. Barreca, P. Fornasiero, Adv. Funct. Mater. 24 (2014) 372–378. [28] M. Gich, J. Gazquez, A. Roig, A. Crespi, J. Fontcuberta, J.C. Idrobo, S.J. Pennycook, M. Varela, V. Skumryev, M. Varela, Appl. Phys. Lett. 96 (2010) 112508. [29] L. Corbellini, C. Lacroix, C. Harnagea, A. Korinek, G.A. Botton, D. Ménard, A. Pignolet, (Accepted for publication) Scientific Reports 7 (2017), 3712. http://dx.doi.org/10. 1038/s41598-017-02742-9. [30] T.M.N. Thai, D.T. Nguyen, N.-S. Lee, J.-S. Rhyee, J. Song, H.-J. Kim, J. Appl. Phys. 120 (2016) 185304. [31] M. Gich, I. Fina, A. Morelli, F. Sánchez, M. Alexe, J. Gàzquez, J. Fontcuberta, A. Roig, Adv. Mater. 26 (2014) 4645–4652. [32] J. Tucek, R. Zboril, A. Namai, S. Ohkoshi, Chem. Mater. 22 (2010) 6483–6505. [33] A. Namai, M. Yoshikiyo, K. Yamada, S. Sakurai, T. Goto, T. Yoshida, T. Miyazaki, M. Nakajima, T. Suemoto, H. Tokoro, S. Ohkoshi, Nat. Commun. 3 (2012) 1035. [34] K. Sharma, V. Raghavendra Reddy, A. Gupta, R.J. Choudhary, D.M. Phase, V. Ganesan, Appl. Phys. Lett. 102 (2013) 212401. [35] Y. Hamasaki, T. Shimizu, H. Taniguchi, T. Taniyama, S. Yasui, M. Itoh, Appl. Phys. Lett. 104 (2014) 82906. [36] M. Trassin, N. Viart, G. Versini, J.L. Loison, J.P. Vola, G. Schmerber, O. Crégut, S. Barre, G. Pourroy, J.H. Lee, W. Jo, C. Mény, Appl. Phys. Lett. 91 (2007) 202504. [37] S. Ohkoshi, H. Tokoro, Bull. Chem. Soc. Jpn. 86 (2013) 897–907.
Contents lists available at ScienceDirect
Scripta Materialia journal homepage: www.elsevier.com/locate/scriptamat
Regular article
The effect of Al substitution on the structural and magnetic properties of epitaxial thin films of epsilon ferrite Luca Corbellini a,⁎, Christian Lacroix b, David Ménard b, Alain Pignolet a a b
Centre Énergie, Matériaux et Télécommunications, INRS, 1650 Boulevard Lionel-Boulet, Varennes, Québec J3X 1S2, Canada Département de Génie Physique & Regroupement Québécois sur les Matériaux de Pointe (RQMP), Polytechnique Montréal, Montréal, Québec H3T 1J4, Canada
a r t i c l e
i n f o
Article history: Received 23 March 2017 Received in revised form 26 May 2017 Accepted 8 July 2017 Available online xxxx Keywords: Thin film Epitaxy Pulsed laser deposition Magnetism Ferromagnetic resonance frequency
a b s t r a c t Epitaxial growth of aluminum substituted epsilon ferrite by pulsed laser deposition on strontium titanate and yttrium-stabilized zirconium substrates is demonstrated for a wide range of aluminum substitution. Linear decrease of the out-of-plane lattice parameter with aluminum substitution correlates with the decrease of the coercive magnetic field, while the magnetization first increases and then decreases with increasing aluminum concentration. A characteristic inflexion in the hysteresis loops has been attributed to the presence of a small quantity of magnetite. Epitaxial growth on yttrium-stabilized zirconium is technologically significant due to its use as buffer layer for the growth of complex oxides on silicon. © 2017 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
Nanostructured iron oxides have recently drawn a lot of attention given their many potential applications, first based on their inherent magnetic properties, but also in various fields of medicine because of their well-known biochemical characteristics of nontoxicity, biocompatibility, and biodegradability [1–13]. Iron oxides are the most common iron compounds in nature and several of them, such as magnetite (Fe3O4), maghemite (γ-Fe2O3) and hematite (α-Fe2O3) occur naturally and are generally easy to synthesize. A notable exception is epsilon ferrite (ε-Fe2O3), a ferrimagnetic metastable phase intermediate between maghemite and hematite. The magnetization in εFe2O3 arises from a non-fully compensated magnetic moment on the anti-parallel aligned iron atoms (see Fig. 1), in which the tetrahedrally coordinated “D” sites are holding a lower magnetic moment than the other sites. Although being characterized by lower magnetization compared to γ-Fe2O3, ε-Fe2O3 has received notable attention lately due to its very strong magnetocrystalline anisotropy, which results in a gigantic coercive field [14–17], and also in a natural ferromagnetic resonance (FMR) frequency above 100 GHz, flirting with the low THz range (~0.1–100 THz) at room temperature [18]. This is of particular interest given its potential use in short-range wireless communications (e.g. 60 GHz WiFi) [19] and ultrafast non-volatile memories [20]. The origin of such strong magnetocrystalline anisotropy was investigated by Xray magnetic circular dichroism and found to be due to lattice distortion [21]. Later, first principle calculations supported those experimental findings, by predicting that the magnetic easy axis should lie along the ⁎ Corresponding author. E-mail address: [email protected] (L. Corbellini).
http://dx.doi.org/10.1016/j.scriptamat.2017.07.005 1359-6462/© 2017 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
crystal a axis direction [22]. Finally, given its crystal structure belonging to the polar space group Pna21 [23,24] (Fig. 1), ε-Fe2O3 is characterized by a spontaneous polarization that, if proved switchable at room temperature, would make it one of the few single phase, room temperature multiferroic materials. Up until recent years, due to its metastability in ambient condition, epsilon ferrite has mostly been synthesized in the form of nanostructure [15–17,25–27], with size confinement acting as the stabilizing factor. However, in order to further investigate their relatively unexplored magnetic properties and eventually integrate them into devices, synthesis of stable epsilon ferrite thin films and their compounds must be mastered. Recent reports showed that it is indeed possible to grow epsilon ferrite thin films epitaxially directly on single crystalline strontium titanate (SrTiO3) [28] and yttrium stabilized zirconium (YSZ) substrates by pulsed laser deposition (PLD) [29], and on sapphire by using a buffer layer of gallium ferrite (GaFeO3) [30]. Furthermore, presence of reversible spontaneous polarization was reported in layered structure of SrTiO3:Nb/AlFeO3/SrRuO3/ε-Fe2O3 [31]. Extensive studies have also been conducted on the effect of metal substitution into epsilon ferrite (ε-MxFe2 − xO3) nanoparticles, in particular regarding its effect on the FMR frequency. Ohkoshi et al. have reported that substitution of Fe3+ ions in ε-Fe2O3 with metals having a smaller ionic radius, such as Al3 + and Ga3 +, results in a lowering of the FMR frequency [18,19,32], while through substitution with larger ions like Rh3+, an increase of the FMR is observed [33]. It is interesting to note that AlFeO3 and GaFeO3, which are isostructural to ε-Fe2O3, are known room temperature ferroelectrics characterized by spontaneous magnetic ordering at cryogenic temperatures [34–36].
64
L. Corbellini et al. / Scripta Materialia 140 (2017) 63–66
Fig. 1. ε-Fe2O3 unit cell and configuration of the magnetic moments on the four different iron sites. It has to be noted that the magnetic moments lie in the a–b plane. According to recent works, they should be aligned along the a-axis of the crystal unit cell.
In this paper, we report the epitaxial growth by pulsed laser deposition of thin films of Al-substituted epsilon ferrite and the effect of Al substitution on the structural and magnetic properties. Thin films of εAlxFe2 − xO3 of thickness of about 50 nm were grown directly on single crystalline (111)-oriented strontium titanate (SrTiO3) and (100)-oriented Yttrium Stabilized Zirconia (YSZ) substrates (no buffer layer was used). While the different symmetry of the two substrates give different epitaxial match and result in different twins formation [29,30], the films grown on both type of substrates exhibited very similar structural and functional properties and the same trends with increasing aluminum substitution. The films were grown following the conditions reported in reference 27; the only peculiarity was that, in order to obtain films with different percentage of Al, the growth was performed by alternatively ablating a standard Fe2O3 target and a second target of AlFeO3. Different Al substitution percentages were obtained by changing the ratio of the number of pulses shot on each target. In order to achieve a homogenous distribution of aluminum, the films were grown roughly one monolayer at the time, alternating the targets in order to get the desired composition. Films with Al concentration ranging from pure ε-Fe2O3 to ε-Al0.5Fe1.5O3 (0 b x b 0.5) were then prepared. Thin films of pure AlFeO3 (i.e. x = 1) were also prepared (with the same deposition conditions as the epsilon ferrite films) to estimate its growth rate and for comparison with the ε-AlxFe2 − xO3 films. In order to confirm the correctness of our estimation of the composition, which is based on the number of laser pulses on each target and on the growth rate of both AlFeO3 and ε-Fe2O3, energy-dispersive X-ray spectroscopy (EDX) was performed on our films. The Al:Fe ratio obtained was in very good agreement with the Al:Fe ratio estimated from the number of pulses shot on each target.
Structural analysis was performed by X-ray diffractometry with a PANalytical X'Pert MRD PRO diffractometer using the Cu Kα radiation (λ = 1.54056 Å). Goniometer (θ/2θ) scans confirmed that ε-AlxFe2 − xO3 grows epitaxially both on SrTiO3 (111) and YSZ (100). The films are (001)-oriented as shown by the diffraction patterns of pure ε-Fe2O3, ε-Al0.1Fe1.9O3, and pure AlFeO3 grown on SrTiO3 (111) (Fig. 2a). A shift towards higher 2θ angles with increasing Al concentration is observed, which was expected given the smaller ionic radius of Al3+ (53.5 pm) compared to Fe3+ (64.5 pm). The out-of-plane lattice parameter “c” was calculated from the measured goniometer scans, averaging the values found utilizing the inter-planar distances corresponding to the two detectable 00l peaks (l = 2 and 6). A linear decrease in the out-of-plane lattice parameter with increasing aluminum content was found, as shown in the graph in Fig. 2b, in agreement with previous studies [19]. The magnetic properties of our films were measured using a Vibrating Sample Magnetometer (VSM - model EV9 from ADE Technologies). The hysteresis loops were recorded at room temperature for all the samples, with the magnetic field applied both parallel (in-plane) and perpendicular (out-of-plane) to the film surface. In order to obtain solely the signal arising from the thin films, the contribution of the vibrating rod holding the samples and of the substrates were measured and subtracted from the measured hysteresis loops. The resulting magnetic moment was divided by the volume of the thin film. The measured magnetic hysteresis loops of films with x = 0, x = 0.1, and x = 0.5 are shown in Fig. 3. We observe an inflexion at low applied fields in all loops, which we attribute to the presence of a secondary soft magnetic phase that was identified as magnetite (Fe3O4) by detailed XRD and magnetic analysis [29]. Presences of these inflections were also recorded
Fig. 2. (a) θ/2θ scans of (100)-oriented ε-Fe2O3, ε-Al0.1Fe1.9O3, and AlFeO3 thin films, revealing how all the three materials grow epitaxially on (111)-oriented SrTiO3. The line drawn in grey helps visualizing the shift in position of the (006) peak in the Al-substituted films in comparison to pure ε-Fe2O3. The peaks noted by asterisks belong to the SrTiO3 (111) substrates and its secondary lines. (b) Evolution of the lattice parameter “c” with increasing Al concentration “x”.
L. Corbellini et al. / Scripta Materialia 140 (2017) 63–66
65
Fig. 3. (a) In-plane and out-of-plane hysteresis loop measured for a pure ε-Fe2O3 (orange and green), (b) a ε-Al0.1Fe1.9O3 (blue and red), and (c) for a ε-Al0.5Fe1.5O3 epitaxial thin film (burgundy and violet), all grown on YSZ (100). In all cases, we can infer the high magnetic anisotropy of the samples. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
for pure epsilon ferrite thin films grown on SrTiO3 (111) and YSZ (100) by our group, as well as by other groups [28,31], who, however, proposed a different explanation for their origin. Comparison between the three loops revealed that (i) the aluminum-substituted film preserved their in-plane anisotropy (see the difference between the in-plane and the out-of-plane measurement), (ii) are characterized by a lower coercive field in comparison to pure epsilon ferrite films, and (iii) the saturation magnetization depends on the aluminum concentration. Similar trends were observed in Al-substituted ε-Fe2O3 nanoparticles, that is, a decrease of the coercivity expected from the non-magnetic nature of the Al3+ ions (completely filled 2p6 shell) along with an increase of the magnetization at saturation up to a concentration of about 40% of Al [37]. Such an effect has also been reported for gallium-substituted nanoparticles of epsilon ferrite (which have a behavior similar to that of Al-substituted ε-Fe2O3 nanoparticles), where the increase of the magnetization is attributed to the preferential substitution of the iron located in the tetragonal site “D” of the lattice (Fig. 1) by the doping metal ions (gallium in ref. 35, aluminum in our case), due to their ionic radius smaller than that of Fe3+. Such a preferential substitution results in a decrease of the “D”-site magnetic moment, hence to an increase of the total magnetic moment, up to a point where the substitution starts to occur on other iron-sites as well, resulting then in a decrease of magnetic moment (see Fig. 1) [37]. In Fig. 4a, the temperature dependence of the magnetization is shown for the ε-Al0.1Fe1.9O3 film under a field of 1 kOe. A paramagnetic-toferromagnetic phase transition is observed at T ≈ 475 K. No phase transition is observed close to 300 K, which is the Curie temperature of AlFeO3 [35]. This indicates that aluminum substitution was successfully achieved, and that the resulting film is homogeneous, i.e. it does not consist of a mixture of AlFeO3 and ε-Fe2O3. Moreover, we observe that the magnetization does not fall to zero at temperatures higher than the Curie temperature of ε-Fe2O3 (TC ≈ 475 K), which is consistent with the presence of a small amount of magnetite (TC ≈ 850 K).
In order to accurately quantify the effect of aluminum substitution on the magnetic properties of epsilon ferrite, we used the technique described in ref. 27 to separate the contributions of the two magnetic phases (epsilon ferrite and magnetite). Using the values of saturation magnetization found in literature for magnetite (309 emu/cm3), we estimated the volume ratio of each phase (magnetite volume/epsilon ferrite volume), which was found to vary between 0.05 and 0.25 (average ≈ 0.1). For example, in samples with higher percentage of magnetite, as the one shown in Fig. 3b, the higher content of magnetite has the effect of decreasing the coercive field and increasing the saturation magnetization of the whole film. This allowed us to quantify the in-plane coercive field and the magnetization at saturation for the sole ε-AlxFe2 − xO3 phase for different Al concentrations (Fig. 4b and c). We observe that the coercive field decreases linearly as the aluminum concentration increases, suggesting that the magnetocrystalline anisotropy decreases linearly with aluminum concentration [32]. We note that both the magnetocrystalline anisotropy and the out-of-plane lattice parameter “c” decrease linearly with increased aluminum concentration. We also observe that, as the Al concentration increases, the saturation magnetization increases first by a factor of 2 (for x = 0.33), and then decreases for higher Al concentration. We propose that, as discussed in Ref. [19], the decrease of the saturation magnetization is due to the fact that, while Al substitution for x b 0.33 occurs at the iron “D” tetragonal site, further Al substitution above the threshold of x = 0.33 occurs both at the iron “B” and “C” octahedral sites, thus decreasing the total magnetization. For sake of completeness, in order to compare the magnetic properties of epsilon ferrite with that of the whole film (epsilon ferrite + magnetite), we also present the evolution of the coercive field and the saturation magnetization as a function of aluminum concentration for films with a “normalized” magnetite/epsilon volume ratio equal to 0.1, which is the average value found in our films. In view of potential applications, it is important to report the characteristic properties of the
Fig. 4. (a) Dependence of the magnetization (M) with the temperature for a ε-Al0.1Fe1.9O3 thin film grown on YSZ (100), recorded under a field of 1 kOe applied in the sample plane. (b) Evolution of the in-plane coercive field HC and (c) magnetization at saturation MS depending on the aluminum concentration “x”.
66
L. Corbellini et al. / Scripta Materialia 140 (2017) 63–66
whole ε-AlxFe2 − xO3 system. We observe that the effect of magnetite, which is characterized by a very low coercive field and high magnetization compared to epsilon ferrite, is to reduce the coercive field and to increase the magnetization at saturation of the whole film (Fig. 4b and c). This demonstrates the importance of carrying out a complete structural and magnetic characterization for this kind of epitaxial grown ferromagnetic system in order to extract accurate quantitative results. However, as the two magnetic phases do not appear to be interacting (see the angular dependency of the magnetic properties in pure epitaxial films of epsilon ferrite) [29], we expect a linear decrease of the natural ferromagnetic resonance frequency for ε-AlxFe2 − xO3 with increased aluminum concentration, as reported for al-substituted nanoparticles [19,37], along with a second ferromagnetic resonance frequency at lower frequencies due to magnetite, this in spite of the “effective” coercive field remaining roughly unchanged with Al substitution. In summary, epitaxial thin films of ε-AlxFe2 − xO3 were grown on SrTiO3 (111) and on YSZ (100). This latter achievement may constitute a decisive step towards integration of such material into consumer electronics, in particular its ability of growing epitaxially on YSZ (100), given how such oxide is often used as a buffer layer for the growth of complex oxides on silicon (100). Substitution over a wide range (x = 0.05 to 0.5, as well as x = 1) was achieved. Aluminum substitution linearly reduces the coercive field, and most probably, the magnetocrystalline anisotropy of epsilon ferrite, with increasing aluminum concentration. Moreover, an increase of the saturation magnetization up to a threshold of x = 0.33 was shown, similar to results previously reported for aluminum and gallium-substituted epsilon-ferrite nanoparticles. Finally, a characteristic inflexion at low fields in the hysteresis loops, also observed by other groups, has been however attributed to the presence of a secondary phase of magnetite rather than to other effects, such as surface effects. This work was supported by grants from the Natural Sciences and Engineering Research Council of Canada (NSERC - Canada) (# 2616622013) and the Fonds de Recherche du Québec - Nature et Technologies (FRQNT - Québec) (# 2013-PR-167502). References [1] [2] [3] [4]
A.-H. Lu, E.L. Salabas, F. Schuth, Angew. Chem. Int. Ed. 46 (2007) 1222–1244. W. Wu, Q.G. He, C.Z. Jiang, Nanoscale Res. Lett. 3 (2008) 397. A.S. Teja, P.-Y. Koh, Prog. Cryst. Growth Charact. Mater. 55 (2009) 22–45. J. Gonzalez, O. Chubykalo, J.M. Gonzalez, Encyclopedia of Nanoscience and Nanotechnology, American Scientific Publishers, Stevenson Ranch, CA, 2004. [5] R.M. Cornell, U. Schwertmann, The Iron Oxides: Structure, Properties, Reactions, Occurrence and Uses, Wiley-VCH, Weinheim, Germany, 2003. [6] J.W.M. Bulte, D.L. Kraitchman, NMR Biomed. 17 (2004) 484–499.
[7] C.C. Berry, A.S.G. Curtis, J. Phys. D Appl. Phys. 36 (2003) R198–R206. [8] A.K. Gupta, M. Gupta, Biomaterials 26 (2005) 3995–4021. [9] S. Laurent, D. Forge, M. Port, A. Roch, C. Robic, L.V. Elst, R.N. Muller, Chem. Rev. 108 (2008) 2064. [10] S. Mornet, S. Vasseur, F. Grasset, E. Duguet, J. Mater. Chem. 14 (2004) 2161–2175. [11] Q.A. Pankhurst, J. Connolly, S.K. Jones, J. Dobson, J. Phys. D Appl. Phys. 36 (2003) R167–R181. [12] P. Tartaj, M. del Puerto Morales, S. Veintemillas-Verdaguer, T. Gonzalez-Carreno, C.J. Serna, J. Phys. D Appl. Phys. 36 (2003) R182–R197. [13] D.L.J. Thorek, A.K. Chen, J. Czupyrna, A. Tsourkas, Ann. Biomed. Eng. 34 (2006) 23–38. [14] S. Ohkoshi, S. Sakurai, J. Jin, K. Hashimoto, J. Appl. Phys. 97 (2005), 10K312. [15] C. Savii, M. Popovici, C. Enache, J. Subrt, D. Niznansky, S. Bakardzieva, C. Caizer, I. Hrianca, M.P. Cecilia, Savii Corina Enache, J. Subrt, D. Niznansky, Snejana Bakardzieva, C. Caizer, I. Hrianca, Solid State Ionics 151 (2002) 219–227. [16] M. Kurmoo, J.-L. Rehspringer, A. Hutlova, C. D'Orléans, S. Vilminot, C. Estournès, D. Niznansky, Chem. Mater. 17 (2005) 1106–1114. [17] J. Jin, S. Ohkoshi, K. Ashimoto, Adv. Mater. 16 (2004) 48–51. [18] S. Ohkoshi, S. Kuroki, S. Sakurai, K. Matsumoto, K. Sato, S. Sasaki, Angew. Chem. Int. Ed. Engl. 46 (2007) 8392–8395. [19] A. Namai, S. Sakurai, M. Nakajima, T. Suemoto, K. Matsumoto, M. Goto, S. Sasaki, S. Ohkoshi, J. Am. Chem. Soc. 131 (2009) 1170–1173. [20] M. Nakajima, A. Namai, S. Ohkoshi, T. Suemoto, Opt. Express 18 (2010) 18260–18268. [21] Y.-C. Tseng, N.M. Souza-Neto, D. Haskel, M. Gich, C. Frontera, A. Roig, M. van Veenendaal, J. Nogués, Phys. Rev. B 79 (2009) 94404. [22] M. Yoshikiyo, K. Yamada, A. Namai, S. Ohkoshi, J. Phys. Chem. C 116 (2012) 8688–8691. [23] R. Zboril, M. Mashlan, D. Petridis, Chem. Mater. 14 (2002) 969–982. [24] E. Tronc, C. Chanéac, J.P. Jolivet, J. Solid State Chem. 139 (1998) 93–104. [25] J. Jin, K. Hashimoto, S. Ohkoshi, J. Mater. Chem. 15 (2005) 1067. [26] G. Carraro, D. Barreca, C. Maccato, E. Bontempi, L.E. Depero, C. de Julian Fernandez, A. Caneschi, CrystEngComm 15 (2013) 1039–1042. [27] G. Carraro, C. Maccato, A. Gasparotto, T. Montini, S. Turner, O.I. Lebedev, V. Gombac, G. Adami, G. Van Tendeloo, D. Barreca, P. Fornasiero, Adv. Funct. Mater. 24 (2014) 372–378. [28] M. Gich, J. Gazquez, A. Roig, A. Crespi, J. Fontcuberta, J.C. Idrobo, S.J. Pennycook, M. Varela, V. Skumryev, M. Varela, Appl. Phys. Lett. 96 (2010) 112508. [29] L. Corbellini, C. Lacroix, C. Harnagea, A. Korinek, G.A. Botton, D. Ménard, A. Pignolet, (Accepted for publication) Scientific Reports 7 (2017), 3712. http://dx.doi.org/10. 1038/s41598-017-02742-9. [30] T.M.N. Thai, D.T. Nguyen, N.-S. Lee, J.-S. Rhyee, J. Song, H.-J. Kim, J. Appl. Phys. 120 (2016) 185304. [31] M. Gich, I. Fina, A. Morelli, F. Sánchez, M. Alexe, J. Gàzquez, J. Fontcuberta, A. Roig, Adv. Mater. 26 (2014) 4645–4652. [32] J. Tucek, R. Zboril, A. Namai, S. Ohkoshi, Chem. Mater. 22 (2010) 6483–6505. [33] A. Namai, M. Yoshikiyo, K. Yamada, S. Sakurai, T. Goto, T. Yoshida, T. Miyazaki, M. Nakajima, T. Suemoto, H. Tokoro, S. Ohkoshi, Nat. Commun. 3 (2012) 1035. [34] K. Sharma, V. Raghavendra Reddy, A. Gupta, R.J. Choudhary, D.M. Phase, V. Ganesan, Appl. Phys. Lett. 102 (2013) 212401. [35] Y. Hamasaki, T. Shimizu, H. Taniguchi, T. Taniyama, S. Yasui, M. Itoh, Appl. Phys. Lett. 104 (2014) 82906. [36] M. Trassin, N. Viart, G. Versini, J.L. Loison, J.P. Vola, G. Schmerber, O. Crégut, S. Barre, G. Pourroy, J.H. Lee, W. Jo, C. Mény, Appl. Phys. Lett. 91 (2007) 202504. [37] S. Ohkoshi, H. Tokoro, Bull. Chem. Soc. Jpn. 86 (2013) 897–907.