Chem. Mater. 2009, 21, 1375–1380
1375
Critical Limitations in the Fabrication of Biferroic BiFeO3-CoFe2O4 Columnar Nanocomposites Due to Bismuth Loss N. Dix,† R. Muralidharan,† B. Warot-Fonrose,‡ M. Varela,§ F. Sa´nchez,*,† and J. Fontcuberta† Institut de Cie`ncia de Materials de Barcelona-CSIC, Campus de la UAB, Bellaterra 08193, Barcelona, Spain, CEMES-CNRS, 29 rue Jeanne MarVig, BP 94347, Toulouse Cedex 4, France, and Departament de ` ptica and Institut de Nanocie`ncia i Nanotecnologia (IN2UB), UniVersitat de Fiı`sica Aplicada i O Barcelona, AVenida Diagonal 647, Barcelona 08028, Spain ReceiVed December 24, 2008. ReVised Manuscript ReceiVed February 10, 2009
The influence of the temperature during self-assembled growth of BiFeO3-CoFe2O4 columnar nanocomposites on (001) and (111) oriented SrTiO3 substrates has been investigated aiming to control pillar size and lateral ordering. The dramatic influence of the growth temperature on the nanoobject morphology is found to be basically due to the strong monotonic reduction in the bismuth content as the deposition temperature increases. There is a very narrow (tens of degrees) optimal window of substrate temperature, in which BiFeO3 could permit a small deficit of Bi without decomposition. The enhanced ferromagnetic response of samples deposited at higher temperature signals formation of secondary ferromagnetic phases caused by BiFeO3 decomposition. Therefore, the deposition temperature is not a suitable free parameter to control the nanoobjects topology in BiFeO3 nanocomposites.
Introduction The codeposition of immiscible ferromagnetic spinels and ferroelectric perovskite oxides results in the spontaneous formation of columnar nanocomposite films1 in which an intimate interface coupling could allow dual tuning of their ferroic properties. Indeed, demonstration of room temperature biferroicity (ferroelectric and ferromagnetic orders) as well as direct switching of the magnetization of ferromagnetic pillars by electrical poling has been reported in BiFeO3 (BFO)-CoFe2O4 (CFO) nanocomposites.2 These results have triggered a strong interest in biferroic nanocomposites. However, use of nanocomposites for some applications, such as memory elements, would probably require a detailed control of the morphology and long-range self-ordered growth of the constituent elements. This objective is far from present status of knowledge. The nanocomposites grow in a bottom-up process, by codeposition of two immiscible compounds, and the driving force promoting columnar growth is the surface energy anisotropy of the involved materials;3-7 the (111) planes in the spinel (CoFe2O4, NiFe2O4, etc.) and the (001) planes in the perovskite (BiFeO3, * Corresponding author. E-mail:
[email protected]. † Institut de Cie`ncia de Materials de Barcelona-CSIC. ‡ CEMES-CNRS. § Universitat de Barcelona.
(1) Zheng, H.; Wang, J.; Lofland, S. E.; Ma, Z.; Mohaddes-Ardabili, L.; Zhao, T.; Salamanca-Riba, L.; Shinde, S. R.; Ogale, S. B.; Bai, F.; Viehland, D.; Jia, Y.; Schlom, D. G.; Wuttig, M.; Roytburd, A.; Ramesh, R. Science 2004, 303, 661. (2) Zavaliche, F.; Zheng, H.; Mohaddes-Ardabili, L.; Yang, S. Y.; Zhan, Q.; Shafer, P.; Reilly, E.; Chopdekar, R.; Jia, Y.; Wright, P.; Schlom, D. G.; Suzuki, Y.; Ramesh, R. Nano Lett. 2005, 5, 1793. (3) Zheng, H.; Zhan, Q.; Zavaliche, F.; Sherburne, M.; Straub, F.; Cruz, M. P.; Chen, L. Q.; Dahmen, U.; Ramesh, R. Nano Lett. 2006, 6, 1401. (4) Zheng, H.; Straub, F.; Zhan, Q.; Yang, P. L.; Hsieh, W. K.; Zavaliche, F.; Chu, Y. H.; Dahmen, U.; Ramesh, R. AdV. Mater. 2006, 18, 2747.
BaTiO3, etc.) are the ones with lower specific surface energy. However, the microstructure of thin films results from a delicate balance between thermodynamic conditions (surface and interface energies) and kinetics (i.e., adatom diffusivity, etc.) and nucleation rate (i.e., supersaturation, etc.) that largely depends on the growth conditions.6,8 It has been reported that substrate temperature and growth rate can be used to tune the pillar size in BFO-CFO nanocomposites.3,4 However, the presence of highly volatile Bi in BiFeO3 (and similarly of Pb in the case of other usual ferroelectrics) could limit or make much more complex this possibility of controlling nanoobjects morphology. We have deposited BFO-CFO nanocomposites on (001) and (111) oriented SrTiO3 (STO) substrates to investigate the impact of growth temperature on crystal structure, morphology, and stoichiometry. It is concluded that the Bi content decreases monotonically with the growth temperature. There is only a very narrow (tens of degrees) window of temperatures in which nominal Bi1-xFeO3 can be grown without evidence of decomposition. Substrate temperature can then not be used as a free parameter to control the stoichiometry and morphology of nanoobjects in BFO-CFO nanocomposites. Experimental Section Nanocomposite thin films were prepared by pulsed laser deposition (KrF, 5 Hz) using a BFO-CFO target with molar ratio of 65: (5) Lu¨ders, U.; Sa´nchez, F.; Fontcuberta, J. Phys. ReV. B 2004, 70, 045403. (6) Sa´nchez, F.; Lu¨ders, U.; Herranz, G.; Infante, I. C.; Fontcuberta, J.; Garcı´a-Cuenca, M. V.; Ferrater, C.; Varela, M. Nanotechnology 2005, 16, S190. (7) Levin, I.; Li, J.; Slutsker, J.; Roytburd, A. L. AdV. Mater. 2006, 18, 2044. (8) Bachelet, R.; Sa´nchez, F.; Santiso, J.; Fontcuberta, J. Appl. Phys. Lett. 2008, 93, 151916.
10.1021/cm803480q CCC: $40.75 2009 American Chemical Society Published on Web 03/10/2009
1376
Chem. Mater., Vol. 21, No. 7, 2009
Dix et al.
35. The target, with 10% excess bismuth, was prepared by a conventional solid-state reaction process. STO (001) and (111) substrates were placed simultaneously, side by side, on the heater. The oxygen pressure during deposition was maintained at 0.1 mbar. Four couples of films were deposited at 600, 625, 650, and 675 °C (heater block temperature). The films, with a nominal thickness t of 100 nm, were grown at a rate of around 0.1 Å/pulse (0.5 Å/s). An additional couple of films, 200 nm thick, were deposited at 650 °C. At the end of the growth, the samples were cooled under 1 atm of oxygen. The crystal structure of the films was analyzed by X-ray diffractometry (XRD) with Cu KR radiation, measuring θ-2θ scans of symmetrical reflections and φ scans of asymmetrical reflections. Atomic force microscopy (AFM) in dynamic mode was used to investigate the surface morphology. Chemical composition was investigated by X-ray wavelength dispersive spectroscopy (WDS), energy-dispersive X-ray spectroscopy (EDX), and by backscattered electrons images from field emission secondary electron microscopy (FESEM). Magnetization hysteresis loops of (001) samples were measured by superconducting quantum interference device (SQUID) at room temperature, with in-plane magnetic field. The t ) 200 nm film on STO(001) was studied by cross-section high-resolution transmission electron microscopy (HRTEM) and electron energy loss spectroscopy (EELS).
Results and Discussion In a previous study,9 we observed from X-ray diffraction (XRD) measurements phase separation of BFO and CFO, and absence of other phases, in nanocomposites deposited at 600 or 650 °C. In contrast, samples prepared at 550 °C or above 700 °C presented spurious phases. Similarly, a narrow temperature window to grow epitaxial single phase nanocomposite films has been reported very recently for the BiFeO3-NiFe2O4 system.10 Here we present a detailed study of BFO-CFO nanocomposites using a small temperature step of 25 °C in the 600-675 °C range. In all the samples, both the perovskite BFO and the spinel CFO phases are epitaxial. To illustrate it, we show first in Figure 1 the XRD scans for the t ) 100 nm films grown at 650 °C on STO(001) and (111). Textured growth of CFO and BFO is observed on both substrates: CFO(001) and BFO(001) textures on STO(001) (Figure 1a), and CFO(111) and BFO(111) textures on STO(111) (Figure 1b). In the θ-2θ scans. the film reflections are indicated by vertical lines. Reflections from other orientations or additional phases are not detected within the instrumental sensibility. φ-2θ area scans around asymmetrical reflections were also collected to determine the inplane texture. The scans show that both spinel and perovskite phases grow epitaxially. In the case of the (001) samples (Figure 1c), there are four BFO(202) and four CFO(404) reflections occurring at the same φ angles that the STO(202) ones. It indicates cube-on-cube epitaxial relationships of both phases with the substrate: [100]BFO(001)//[100]STO(001) and [100]CFO(001)//[100]STO(001). In the case of the (111) samples (Figure 1e), there are three BFO(202) reflections located at the same φ position that the STO(202) ones, (9) Muralidharan, R.; Dix, N.; Skumryev, V.; Varela, M.; Sa´nchez, F.; Fontcuberta, J. J. Appl. Phys. 2008, 103, 07E301. (10) Crane, S. P.; Bihler, C.; Brandt, M. S.; Goennenwein, S. T. B.; Gajek, M.; Ramesh, R. J. Magn. Magn. Mater. 2009, 321, L5.
Figure 1. XRD θ-2θ scans for t ) 100 nm nanocomposite films grown at 650 °C on (a) STO(001) and (b) STO(111); (c) φ-2θ area scan of asymmetrical reflections on STO(001) and (d) sketch indicating the epitaxial relationship for (001) samples; (e) corresponding φ-2θ area scan and (f) sketch of the epitaxy for (111) samples.
whereas the three CFO(404) reflections are shifted by 60°. The epitaxial relationships are [1-10]BFO(111)// [1-10]STO(111) and [01-1]CFO(111)//[1-10]STO(111). Schematic drawings of the epitaxial relationships for (001) and (111) samples are plotted in sketches d and f in Figure 1, respectively. The θ-2θ scans of the samples prepared at different temperatures on STO(001) and STO(111) are compared in the amplifications shown in Figures 2(a) and 2(b), respectively. The vertical lines mark the position of the reflections for bulk samples. Out-of-plane parameters dCFO(hkl) of CFO on STO(001) and STO(111), dCFO(001) ) 0.836 nm and dCFO(111) ) 0.483 nm, are measured for all growth temperatures. It corresponds to a compressive strain ε ≈ -0.4% for both orientations. Concerning BFO reflections, there is a double peak in the case of (001) films, corresponding to strained (dBFO(001) ≈ 0.403 nm, ε ≈ 1.8%) and relaxed (dBFO(001) ) 0.396 nm) BFO. The two peaks are labeled S for strained BFO and R for relaxed BFO. The temperature dependences of the BFO(002)/CFO(004) peak intensity ratio are shown as the inset in Figure 2a. Appropriate deconvolution of the underlying R and S peaks (see Figure 2a) allows us to determine the relative contribution to the diffraction pattern. It turns out that the intensity ratio for the relaxed part decreases with the growth temperature, whereas the
Limitations in the Fabrication of BiFeO3-CoFe2O4
Chem. Mater., Vol. 21, No. 7, 2009 1377
hand, some triangular prisms are observed, likely to be (111) oriented CFO. The change in morphology with increasing temperature is dramatic in the 675 °C sample, where most of the islands are spinel prisms and only a small number of {100} faceted BFO pyramids are still visible. In the panel at the right, the shape of the (111) oriented BFO and CFO nanoobjects is sketched.
Figure 2. XRD θ-2θ scans amplifications for samples grown at different temperatures on (a) STO(001) and (b) STO(111). The vertical lines mark the position of bulk CFO and BFO reflections. In (a) labels S and R correspond to strained and relaxed BFO, respectively. Insets: intensity ratio between BFO and CFO reflections as a function of the substrate temperature. Intensities were calculated from fits (the corresponding ones to the 600 °C samples are shown in the main panels).
strained one remains constant (see inset in Figure 2a). Thus, there is a reduction of the measured total diffracted intensity from BFO in the samples prepared at high temperature. In the case of films on STO(111), only fully relaxed BFO reflections (dBFO(111) ) 0.229 nm) are observed (Figure 2b). As well as for the films grown on STO(001), a decrease in the BFO/CFO peaks intensity ratio with increasing growth temperature occurs (see the inset in Figure 2b). It will be shown below that the intensity reduction of the BFO peaks is due to the loss of Bi at high temperature. Before proceeding, it is worth emphasizing that the coexistence of relaxed and unrelaxed BFO contrasts with reports on fully relaxed BFO in BiFeO3-NiFe2O4 nanocomposite films on SrTiO3(001) substrates 10,11 The dependence of the morphology on the deposition temperature is shown in the topographic AFM images in Figure 3, for (001) films (top panels) and (111) films (bottom panels). Height profiles, along the lines marked, are presented below each image. Unless the sample was deposited at 675 °C, the other (001) films present the expected3 columnar growth of the CFO spinel with {111} faceted squared pyramidal objects, embedded in a flat BFO perovskite matrix, and oriented along [110] directions of the substrate. The morphology of both phases is sketched on the right side. The CFO pyramids protrude on BFO plates of varied height. Higher the deposition temperature, the pyramids stick out more on the BFO matrix (typical heights in the 650 and the 625 °C samples are around 50 and 20 nm, respectively). It causes the apparent lateral size of the pyramids to increase with the temperature. This is evident in the 675 °C sample, where the morphology is basically determined by pyramids. In the case of (111) samples, those deposited at 600 and 625 °C present a high density of {100} faceted triangular pyramidal BFO nanoobjects, as expected for this orientation and the dominant molar ratio of BFO respect to CFO. In the 650 °C sample, there is a lower density of BFO pyramids and they are smaller (height 60 nm, much lower than the typical 110 nm in height of the 625 °C sample). On the other (11) Zhan, Q.; Yu, R.; Crane, S. P.; Zheng, H.; Kisielowski, C.; Ramesh, R. Appl. Phys. Lett. 2006, 89, 172902.
The changes in morphology when growth temperature increases could be directly related to some Bi loss. Although XRD measurements revealed the presence of FexOy phases only in films deposited at temperatures above 700 °C, reflections from some of these phases could overlap with the CFO ones (Fe3O4 and γ-Fe2O3 present spinel structure as CFO and lattice parameters are similar: aCFO ) 8.3919 Å, aFe3O4 ) 8.3940 Å, and aγ-Fe2O3 ) 8.3515 Å). The strong reduction of the intensity of the BFO reflections at the highest temperatures could signal Bi loss that would decrease the amount of BFO in the nanocomposites thin film. To address this issue, we performed chemical analysis by WDS on t ) 100 nm (001) samples to quantify the composition. The dependence of the [Co]/[Fe] and [Bi]/([Co]+[Fe]) cationic ratios on the deposition temperature is plotted in Figure 4. The [Co]/[Fe] ratio of all the films fits the nominal value (0.26) well; in contrast, the [Bi]/([Co]+[Fe]) one is only around the nominal (0.38) in samples grown at 600 and 625 °C, with a substantial reduction in the Bi amount increasing the substrate temperature. The measured composition fits well with the data corresponding to other series with samples prepared with a broader temperature range (Figure 4, inset). There is a strong and monotonic reduction of the [Bi] compared with ([Co]+[Fe]) as temperature increases, and samples with TS > 700 °C are virtually Bi free. Clearly, the changes in the chemical composition causes the evolution of the morphology with the temperature observed by AFM. Indeed, the correlation between morphology and composition is directly observed in the FESEM backscattered electron images (Figure 4, bottom panel). The BFO phase is identified (bright regions) because the atomic number of Bi (Z ) 83) is much higher than that of Fe (Z ) 26) and Co (Z ) 27). Inspection of images in Figure 4 shows a monotonic reduction of the surface fraction of the matrix in the (001) samples and of pillars in the (111) samples, corresponding to the perovskite structure (BFO), which is directly correlated with the loss of Bi as substrate temperature is increased. To investigate the local chemical composition a t ) 200nm (001) nanocomposite was characterized by cross-section HRTEM and EELS. The sample was deposited at 650 °C, and thus, according to data in Figure 4, some Bi deficiency is expected. Low-magnification cross-section images (not presented here) show truncated and fully {111} faceted pyramids (columns) and a matrix with flat top surface. Chemical analysis, by EELS, was performed on selected areas (Figure 5). Measurements allow quantifying Co, Fe, and O, although the transition edges of Bi are difficult to acquire. The cross-section TEM image in Figure 5a includes matrix (left side) and a pyramid (right side). The edge of the pyramid forms 45° with the [100] crystal direction, thus confirming that (111) planes form the lateral faces of the pyramids. The EELS spectrum acquired in this region
1378
Chem. Mater., Vol. 21, No. 7, 2009
Dix et al.
Figure 3. AFM topographic images of nanocomposite films on STO(001) (top panels) and STO(111) (bottom panels), grown at different substrate temperature. Line scan are indicated as white lines and plotted below each image. Panels on the right: sketches of the shape of the spinel and perovskite nanostructures for both orientations.
confirms presence of Co and Fe (Figure 5b), and the calculated percentages of Co, Fe, and O (the estimated precision is 10%) are in good agreement with a CoFe2O4 composition: O ) 58%, Fe ) 30%, and Co ) 11% (the nominal percentages are 57, 28, and 14%, respectively). In contrast, in the matrix region, Co is not detected (Figure 5c). To distinguish between Fe3O4, Fe2O3, and BiFeO3, we have compared the shape of the O K-edge, which is very dependent on the O chemical environment, with references acquired in the same conditions. In Figure 5d, the comparison of the recorded spectra at the O K-edge with a BiFeO3 reference sample confirms the BiFeO3 phase. Whereas presence of FexOy phases in other regions of the sample cannot be excluded, some Bi deficiency in the BFO matrix could occur as a first option for stoichiometry accommodation, before formation of FexOy phases occurring at higher Bi deficiency. This is in agreement with recent suggestions of possibility of significant Bi deficiency in BFO single layers.12,13 The magnetic hysteresis loop corresponding to the (001) sample deposited at 625 °C is displayed in Figure 6a, with the magnetization divided by the CFO volume. The loop is (12) Ihlefeld, J. F.; Kumar, A.; Gopalan, V.; Schlom, D. G.; Chen, Y. B.; Pan, X. Q.; Heeg, T.; Schubert, J.; Ke, X.; Schiffer, P.; Orenstein, J.; Martin, L. W.; Chu, Y. H.; Ramesh, R. Appl. Phys. Lett. 2007, 91, 07922. (13) Ihlefeld, J. F.; Podraza, N. J.; Liu, Z. K.; Rai, R. C.; Xu, X.; Heeg, T.; Chen, Y. B.; Li, J.; Collins, R. W.; Musfeldt, J. L.; Pan, X. Q.; Schubert, J.; Ramesh, R.; Schlom, D. G. Appl. Phys. Lett. 2008, 92, 142908.
completely saturated above around 20 kOe, and the coercive field (see the inset) is around 1.5 kOe. The saturation magnetization of the samples is plotted in Figure 6b against the substrate temperature. We note that CFO volume is the same in all samples since they were deposited sequentially and using the same number of laser pulses. The samples deposited at the lower temperatures, 600 and 625 °C, have magnetization saturation around 440 emu/cm3, which closely fits the bulk magnetization (430 emu/cm3). Interestingly, there is an enhanced magnetization in the nanocomposites deposited at higher temperature, particularly in the 675 °C sample. The samples were clearly deficient in Bi and thus the increased ferromagnetic response would be caused by the concomitant formation of spurious FexOy phases. It is noted that the 650 and the 675 °C samples were deficient in Bi in around 35 and 52%, respectively, whereas their saturation magnetization, compared with the value of the stoichiometric samples, is increasing by around 12 and 35%, respectively. The a priori expected linear relation between Bi deficit and increased magnetization is thus not observed, which suggests that part of the Bi deficit could be accommodated in a Bi1-xFeO3 matrix without causing formation of secondary ferromagnetic phases. We observe ferromagnetic phases in the nanocomposites deposited at 650-675 °C and an oxygen pressure of 0.1 mbar. We notice that the presence of ferromagnetic Fe-O phases in the case of single (i.e, no nanocomposites) BFO films 14,15 was only observed at pressures below 1 × 10-3 to 1 × 10-2 mbar for similar growth temperatures. Therefore, we should conclude that the
Limitations in the Fabrication of BiFeO3-CoFe2O4
Figure 4. Cationic composition ratio, [Co]/[Fe] and [Bi]/([Co]+[Fe]), as a function of the substrate temperature. Composition was measured by WDS in t ) 100 nm samples deposited on STO(001), the inset shows additional [Bi]/([Co] + [Fe]) data, measured by EDX, in a series of samples prepared in the 550-800 °C range (open symbols) plotted as a function of the reciprocal of the temperature (bottom axis) and the temperature (top axis). Solid symbols are the data plotted in the main panel. Bottom panel: backscattered electron FESEM images for (001) and (111) samples deposited in the 600-675 °C range. Scale bar corresponds to 1 µm.
codeposition of CFO somehow favors the stabilization of magnetic Fe-O oxides at relatively higher oxygen pressures. The microscopic mechanism for such reinforced stabilization is at present unknown. However, it could be argued that the coexisting CFO crystallites, with spinel structure, may act as template and promote, via epitaxial stabilization, the growth of isomorphous Fe-O magnetic oxides (i.e., γ-Fe2O3 or Fe3O4) on its surface. Because of the expected similar shape of crystallites of γ-Fe2O3 or Fe3O4 and CFO, morphological inspection using field-emission SEM and AFM did not allow distinction among them. The instability of BFO during the growth of BFO-CFO nanocomposites thin films (as well as single phase BFO films) is not directly related to the volatility of Bi2O3 (it decomposes at pressures close to 1 × 10-6 mbar at 675 °C 16 ), but directly of sublimation of Bi since its vapor pressure at 675 °C is slightly lower than 1 × 10-2 mbar.17 Moreover, we stress that the reported vapor pressure are equilibrium data between a condensed phase and its vapor, whereas the heteroepitaxial growth of a two-phase nanocomposite is a more complex process, with the sticking probability of Bi (14) Be´a, H.; Bibes, M.; Barthe´le´my, A.; Bouzehouane, K.; Jacquet, E.; Khodan, A.; Contour, J. P.; Fusil, S.; Wyczisk, F.; Forget, A.; Lebeugle, D.; Colson, D.; Viret, M. Appl. Phys. Lett. 2005, 87, 072508. (15) Murakami, M.; Fujino, S.; Lim, S. H.; Salamanca-Riba, L. G.; Wuttig, M.; Takeuchi, I.; Varyghese, B.; Sugaya, H.; Hasegawa, T.; Lofland, S. E. Appl. Phys. Lett. 2006, 88, 112505. (16) (a) Weber, A. H.; Plantenberg, S. G. Phys. ReV. 1946, 69, 649. (b) Shoji, Y.; Matsui, T.; Nagasaki, T.; Kurata, M.; Inoue, T. Int. J. Thermophys. 2000, 21, 585. (17) Milch, A. Thin Solid Films 1973, 17, 231.
Chem. Mater., Vol. 21, No. 7, 2009 1379
Figure 5. (a) Cross-section TEM of a t ) 200 nm sample deposited at 650 °C on STO(001); (b, c): EELS spectra acquired in the pyramid island (white circle) and the matrix (red circle), respectively; (d) zoom of (c) around the O K-edge, compared with a BiFeO3 reference sample.
adsorbed on CFO islands expected to be low, particularly when increasing deposition temperature. On the other hand, we would like to mention that after completion of the present work, a detailed study of the influence of the temperature on the thermodynamic stability of bulk BiFeO3 pellets has been published.18 The authors have found decomposition of bulk BiFeO3 in Bi25FeO39 and Bi2Fe4O9, under ambient atmosphere in a temperature interval around 450-770 °C, whereas heating of the same samples at higher temperatures causes pure BiFeO3. In contrast, these secondary phases are not found in BiFeO3 nanocomposites or single thin films,14,15 in spite of the similar temperature during growth and the lower pressure. The higher lattice mismatch of the secondary phases with the substrate could prevent its formation and thus cause epitaxial stabilization of BFO on SrTiO3 or other structurally compatible substrates. Conclusion In conclusion, a small variation in the growth temperature of biferroic nanocomposites having BFO as ferroelectric phase results in dramatic effects in morphology and stoichiometry. Bi content decreases monotonically as temperature increases, and it appears that there is not an optimal window where Bi (18) Selbach, S. M.; Einarsrud, M. A.; Grande, T. Chem. Mater. 2009, 21, 169.
1380
Chem. Mater., Vol. 21, No. 7, 2009
Dix et al.
Figure 6. (a) Magnetization hysteresis loop (measured at room temperature, with the field applied in the plane) of a t ) 100 nm sample deposited at 625 °C on STO(001). A zoom of the low field region is in the inset. (b) Saturation magnetization divided by CoFe2O4 volume plotted against the deposition temperature.
stoichiometry is preserved. However, BFO could allow a certain Bi deficit before formation of FexOy spurious phases. Therefore, the deposition temperature can not be used as a free parameter to tune the pillar size. The absence of a well-defined temperature region where BiFeO3 stoichiometry is obtained suggests that in the case of small variations of growth conditions, the formed nanocomposites would be largely nonstoichiometric with severe consequences for dielectric losses.
Acknowledgment. Financial support by the Ministerio de Ciencia e Innovacio´n of the Spanish Government Projects: NAN2004-9094-C03, MAT2005-5656-C04, MAT2008-06761C03, Nanoselect CSD2007-00041, and Accio´n Integrada HF20070102, and by European Union [Project MaCoMuFi (FP6-03321), FEDER, and ESTEEM (FP6-0260019)] is acknowledged. CM803480Q
