Open Access
Add to BibSonomyAbstract
Large Faraday rotations, when achieved simultaneously with low optical losses, lead to obtaining high magneto-optic (MO) figures of merit in bismuth-substituted garnet-type material systems. Demonstrating high MO figures of merit typically requires the synthesis of garnet materials with high bismuth substitution levels (close to 3 Bi atoms per stoichiometric formula unit). In our previous experiments, garnet layers sputtered from a target of nominal stoichiometry Bi3Fe5O12 in pure argon atmosphere showed negligible amounts of specific Faraday rotation after annealing, in contrast with results reported typically for pulsed laser deposition of this material in plasma chemistries containing oxygen. We co-sputter Bi3Fe5O12 together with Dy2O3 in pure argon plasma, and obtain the garnet-type composite thin films on glass substrates possessing a specific Faraday rotation in garnet-Dy2O3 composite films in excess of 14°/µm at 532 nm and a coercive force as low as 100 Oe.
© 2014 Optical Society of America
1. Introduction
Bismuth-substituted metal-doped iron garnets of different composition types are widely used functional materials in applied magneto-optics, since they possess very attractive optical and MO properties in the near-infrared and visible spectral regions [1–5]. MO garnet materials are suitable for use in various technological applications such as magnetic memory, magneto-plasmonic devices, magneto-optic sensors, lightwave polarization controllers, and MO spatial or temporal light modulators [4, 6–11]. In applied magneto-optics, the development of highly Bi-substituted ferrimagnetic iron garnet thin films containing a high volumetric fraction of the garnet phase and possessing good surface quality and microstructure still remains a challenging task. This is due to the fact that high-quality garnet layers exhibiting giant Faraday rotation require experimental optimization of the multiple process parameters related to garnet synthesis, and the optimum deposition and annealing processes need to be determined for each new garnet composition. The optical, magnetic and MO properties (eg specific Faraday rotation) of magnetic garnet films are strongly dependent on the level of bismuth substitution as well as on the deposition and/or annealing process parameters, and these films are often prepared by liquid-phase epitaxy (LPE), RF or ion-beam-assisted sputtering, or pulsed laser deposition [2, 12]. The principal goals of our on-going garnet material development research include the identification of new garnet stoichiometry types that can provide new functional thin-film materials demonstrating useful combinations of giant Faraday rotation simultaneously with other attractive features, e.g., low optical absorption, low coercive force values for easy switching of magnetization states, and possessing either perpendicular or in-plane magnetization vector direction. The wide and continuously expanding range of applications that require high-performance garnet films, from magnetic photonic crystals to magneto-plasmonics [4–13], requires the development of new and previously unexplored material types with different combinations of optimized functional properties. A particularly challenging research goal is related to simultaneously maintaining the control over the different material property groups, such as magnetic parameters, switching behaviour, anisotropy type and especially the optical absorption spectra. Giant specific Faraday rotation values in the visible spectral region (reaching 34.5 °/μm near a peak at 530 nm) have been achieved recently in pulsed laser deposited (PLD) thin films of Bi3Fe5O12 (BIG) [14]. Films of Bi3Fe4Ga1O12 also deposited by PLD onto garnet substrates have been shown to possess improved visible-range transparency compared to BIG, simultaneously with very large specific Faraday rotation reaching 17 °/μm at 532 nm [15].
We so far haven’t been able to obtain good MO performance (in terms of large specific Faraday rotation) in sputtered layers of highly-substituted bismuth iron garnet deposited from oxide-mix-based garnet-stoichiometry targets containing more than four iron atoms per formula unit (such as Bi3Fe5O12, Bi3Fe4.6Mn0.4O12 and Bi2.5Sm0.5Fe5O12). However we have found experimentally that co-sputtering of these materials with other garnets or some oxides (like Dy2O3 or Bi2O3) using two different sputtering targets in oxygen-free, pure-argon process gas atmosphere can lead (after careful optimization of the annealing regimes) to achieving improved Faraday rotation in films of good surface quality and transparency. We therefore selected the co-sputtering approach as our principal way of synthesizing a garnet composition stoichiometrically as close as possible to Bi3Fe5O12 which still contains some gallium (Ga) or dysprosium (Dy) dilution. We have successfully synthesized the garnet-garnet mixed-stoichiometry material of type (Bi3Fe5O12-Bi2Dy1Fe4Ga1O12) to increase the number of bismuth atoms per formula unit and reported achieving a high-performance nanocomposite MO material, which exhibited simultaneously a high Faraday rotation, good MO quality, strong uniaxial magnetic anisotropy, and also possessed a notable “red-shift” in the spectral features of its magnetic circular dichroism (MCD) [13]. The flexibility of the co-sputtering approach also allows fine-tuning of stoichiometry variations within nano-composites.
In this paper, we report on the successful synthesis of co-sputtered garnet-oxide composite layers prepared using a ceramic sputtering target of fully bismuth-substituted iron garnet and a separate sputtering target of dysprosium oxide. We prepare several batches of Bi3Fe5O12: Dy2O3 composite films containing different amounts of added dysprosium oxide and optimize the annealing regimes (in terms of both the process temperatures and time durations) to obtain new materials with excellent combinations of the optical and magneto-optical properties. The material types developed are expected to be of importance for various MO and non-reciprocal integrated-optics applications including magneto-photonic and magneto-plasmonic devices for controlling light at the nanoscale and enabling ultra-fast modulation of light intensity in microdevices.
2. Bi3Fe5O12: Dy2O3 co-sputtered layers - growth and characteristics
Several batches of garnet-dysprosium oxide composite films (of thicknesses between 200 and 320 nm) were deposited onto glass and Gd3Ga5O12 (GGG) substrates (cleaned chemically inside an ultrasonic bath) by using RF magnetron co-sputtering processes. The main process-related parameters and conditions used to prepare the films are summarized in Table 1.A small volumetric fraction (from 2.7 to 20 vol. %) of Dy2O3 as measured using partial deposition rates data was added from a separate sputtering target during each of the several co-sputtering deposition runs of Bi3Fe5O12:Dy2O3. The partial deposition rates for each target were controlled by the adjustments of applied RF power and characterised separately. The film thicknesses of co-sputtered layers were monitored using quartz crystal microbalance sensor calibrated by using the transmission spectrum curve-fitting using custom-made thin film thickness measurement software; the methodology details of the thickness fitting procedure can be found in [4].
Table 1. Summary of the sputtering process parameters and conditions used to deposit garnet-oxide composite thin film layers onto GGG and glass substrates.
The garnet-Dy2O3 films were in an amorphous phase just after the deposition process and did not possess any measurable Faraday rotation, similarly to all other as-deposited RF magnetron sputtered Bi-substituted iron-garnet thin films deposited onto cold substrates. All films were subjected to annealing crystallization by using conventional high-temperature oven heat treatments. It was crucial to find the optimized annealing regimes for each batch of samples containing different oxide dilution levels, and several annealing trial runs were performed with each batch to find the annealing regimes that result in the best possible material quality in terms of specific Faraday rotation, sample transparency and surface roughness. Though the optimum annealing regimes usually depend very strongly on the layer composition and substrate type for most garnets, we found that rather similar annealing regimes were suitable for most Bi3Fe5O12:Dy2O3 batches having between 11 and 20 vol.% of added Dy2O3 content, for samples deposited on either the GGG or glass substrates. The noticeable annealing regime variations were only observed in two batches of Bi3Fe5O12:Dy2O3films having 2.7 and 5 vol. % of added dysprosium oxide.
Figure 1 shows the summary of air-atmosphere oven annealing regimes identified to be most suitable for the crystallisation of samples from all batches of garnet-oxide films co-sputtered onto GGG and glass substrates during this study.
Fig. 1 The crystallisation temperature boundaries (thermal windows) used to anneal the Bi3Fe5O12-Dy2O3 thin film composites deposited onto GGG and glass (Corning Eagle XG) substrates.
The high-quality annealed nanocomposite films (as judged by the surface quality of the films after the annealing process) were subjected to specific Faraday rotation and magnetic hysteresis loop measurements using plane-polarised 532 nm and 635 nm laser sources, an electromagnet and Thorlabs PAX polarimeter system. The measured hysteresis loops of Faraday rotation revealed important details related to the magnetic switching behaviour, anisotropy type and magnetization vector direction for the materials synthesised. A Beckman Coulter DU 640B UV/Visible spectrophotometer was used to measure the transmission spectra of all samples (as-deposited and annealed). Using the transmission spectra obtained from the samples, the physical thickness of each sample was fitted again with the help of software. X-ray diffraction (XRD) measurements were carried out using a Siemens D 5000 diffractometer to confirm the presence of crystallized garnet phase in annealed garnet-oxide composite thin films.
3. Properties of Bi3Fe5O12-Dy2O3 thin films, analysis and discussion
XRD measurements were performed by using detector-arm scanning technique in theta-theta diffractometer configuration, at near-grazing x-ray radiation incidence. Cu Kα1 radiation line was used, and measurements were performed using a diffracted beam collimator in the range of 2θ angles between 25° and 70°. The x-ray diffraction data sets obtained from several samples and pattern indexing information are shown in Fig. 2.The angular positions of the x-ray diffraction peaks were determined by using the peak-listing options using Jade 9 (MDI Corp.) software package. The data obtained from all of the annealed samples showed x-ray diffraction peaks at the sets of angles characteristic of the body-centered cubic lattice structure of garnets, and revealed their nanocrystalline microstructure. The lattice constants for garnet-dysprosium oxide composites synthesized have been calculated using identifiable diffraction lines in Fig. 2 and the standard methods explained in [16] and [17]. The calculated (averaged using the data from several indexed peaks) lattice constants of the film layer materials formed as a result of crystallising the layers originating from sputtering the targets of Bi3Fe5O12 and the co-sputtering of Bi3Fe5O12: Dy2O3 (11 and 15 vol. %) on glass substrates were found to be 12.48 Å, 12.52 Å and 12.49 Å, respectively.
Fig. 2 X-ray diffraction data sets obtained from a garnet layer fromed as a result of sputtering a target of stoichiometry Bi3Fe5O12 and from two garnet-oxide composite layers having different vol. % of added Dy2O3 content deposited onto glass substrates. The specific Faraday rotation values measured at 532 nm are also shown in square brackets.
The relatively small (0.13°/μm) specific Faraday rotation measured at 532 nm in a crystallized layer formed as a result of sputtering deposition from Bi3Fe5O12 target, together with a lattice parameter of only 12.48 Å (compared to 12.63 Å reported for epitaxially-grown Bi3Fe5O12 layers [12]) can be explained by a combination of significant bismuth content loss occurring during the deposition (our previous experiments and film composition measurements also support this hypothesis [17]), and the relatively small volumetric fraction of garnet phase formed during the annealing process – the smaller number and the relatively smaller intensities of the x-ray diffraction peaks observed (compared to Bi3Fe5O12: Dy2O3 data sets also shown) can indirectly confirm this possibility. The co-sputtering addition of Dy2O3 content appears to stimulate garnet phase formation, as is evident from the stronger x-ray diffraction peaks recorded and from the much larger specific Faraday rotations obtained. Dysprosium substitution into garnet’s dodecahedral sublattice sites competes with bismuth substitution and also leads to stronger uniaxial magnetic anisotropy, thus raising the film’s magnetization vector from its in-plane orientation characteristic of Bi3Fe5O12.
Specific Faraday rotations (defined as the ratio of the one-way rotation angle to the film thickness) in the visible spectral region (at 532 nm and 635 nm) were measured in all annealed films. All of the annealed nanocomposite-type films exhibited a high specific Faraday rotation, and we found that the specific Faraday rotation values were strongly dependent on the excess Dy2O3 content added during the deposition processes. Figure 3 shows the specific Faraday rotation data points measured at 532 nm for films prepared on GGG (red line) and glass (green line) substrates. The best obtained Faraday rotation per unit film thickness (of more than 14 °/µm at 532 nm) was measured in annealed garnet-dysprosium oxide composite films (Bi3Fe5O12 with 11 vol. % of co-sputtered Dy2O3 content) which were prepared on glass substrates. This giant Faraday rotation suggests the successful crystallization of a garnet composition having a large bismuth substitution level (garnet films of similar compositions with 2 or 2.2 Bi atoms per formula unit usually show specific Faraday rotation at 532 nm not exceeding 8-9 °/µm [17]).
Fig. 3 Specific Faraday rotation values measured in annealed Bi3Fe5O12:Dy2O3 composite thin films prepared on GGG (red dotted line) and glass (green dotted line) substrates. The specific Faraday rotation measurements were performed using a plane-polarised light source of wavelength 532 nm.
The specific Faraday rotation spectra of several annealed composite samples prepared on GGG substrates were studied as a function of the estimated volumetric fraction of Dy2O3 content added to the garnet system. The summary of these results is presented in Fig. 4.We observed a very small specific Faraday rotation (of only 0.15 degrees at 532 nm) in an annealed garnet layer sputtered from the target of nominal stoichiometry Bi3Fe5O12. It has been found that the specific Faraday rotation at 473 nm and 532 nm measured in composite-type films increased with the increasing Dy2O3 content up to only 5 vol. %, while at 635 nm, the best specific Faraday rotation peaks at the oxide fraction of about 11 vol. %.
Fig. 4 Spectral data points of specific Faraday rotation measured in Bi3Fe5O12:Dy2O3 composite-type garnet films deposited onto GGG substrates at 473 nm (blue lines) 532 nm (green lines) and 635 nm (red lines) as a function of the estimated volumetric content of Dy2O3 added by co-sputtering. Here, zero percent of added Dy2O3 indicates a garnet layer sputtered from a target of nominal stoichiometry Bi3Fe5O12.
The absorption spectra of garnet-oxide composite layers were derived using a simple formula A = 1-T-R (%), where A is the absorbed optical power fraction, T is the optical power transmission coefficient, and R is optical power reflectivity. Both the transmission and reflection spectra of film layers were measured using a UV/Visible spectrophotometer. The reflectivity measurements were performed using a non-polarizing beam-splitter cube, and a silver mirror having a spectrally-flat 97% reflectance in the spectrophotometer using the measurement method described in [18]. Figure 5 shows the plots of transmission and absorption spectra measured in a co-sputtered annealed layer of (Bi3Fe5O12 + 11 vol. % Dy2O3) prepared on a glass substrate. We also calculated the MO figure of merit (FOM) for all co-sputtered composite thin films using the measured data on their specific Faraday rotation and transmission.
Fig. 5 Transmission and absorption spectra of a layer of composition type Bi3Fe5O12: 11 vol. % Dy2O3 sputtered onto a glass substrate which was post-deposition annealed using an optimized annealing regime (1h at 630 °C).
The figures of merit (FOM) for the Bi3Fe5O12-Dy2O3 composite films were calculated using a simple formula FOM = FR/ln(T), where FR is the Faraday rotation angle in radians and T is optical transmission (this definition follows that reported in [15]. The Faraday rotation measurements were performed using three polarized light sources of wavelengths 473 nm, 532 nm and 635 nm. Figure 6 shows the scaled FOM data points achieved in the garnet-oxide composite films prepared on GGG substrates. High FOM values (at 532 nm) were obtained in the films of composition types Bi3Fe5O12:Dy2O3 (5 and 11 vol. %). The highest obtained FOM value in all films was almost equal to the FOM value obtained in an annealed high-performance garnet layer of composition type Bi2.1Dy0.9Fe3.9Ga1.1O12, which was synthesized recently by our group at Electron Science Research Institute, Edith Cowan University, Australia. The material properties of Bi2.1Dy0.9Fe3.9Ga1.1O12 garnet thin films will be reported elsewhere.
Fig. 6 Scaled figure of merit (FOM) values measured in Bi3Fe5O12:Dy2O3 (2.7-20 Vol. %) composite-type thin films prepared on GGG substrates.
Magnetic hysteresis curves of the Bi3Fe5O12:Dy2O3 films on GGG and glass substrates were measured at room temperature using a custom-made electromagnet in conjunction with Thorlabs PAX polarimeter. In all cases, the magnetic field was applied in the direction normal to the film plane. The linearity of these hysteresis curves of specific Faraday rotation (Fig. 7 and Fig. 8) indicates that the films possess an almost-in-plane magnetization vector direction, which is of interest for the development of nano-engineered artificial magnetic media, e.g., magneto-plasmonic crystals [6, 7] and micro-devices based on these, for ultra-fast on-chip modulation of light.
Fig. 7 Measured hysteresis loops of specific Faraday rotation in (Bi3Fe5O12: 11 vol. % Dy2O3) nanocomposite thin film prepared on (a) glass and (b) GGG substrates.
Fig. 8 Measured hysteresis loop of Faraday rotation in Bi3Fe5O12:Dy2O3 (2.7-15 Vol. %) composite thin films prepared on GGG substrates.
Figure 8 shows the measured hysteresis loops of specific Faraday rotation of different Bi3Fe5O12:Dy2O3 films prepared on GGG substrates. Composition-dependent coercive force and saturation magnetization values were observed, which confirms the feasibility of engineering the magnetic properties by controlling the material stoichiometry. Table 2 summarises the coercive force and saturation magnetization values measured in different material types.
Table 2. Coercive force and saturation magnetization values measured in co-sputtered garnet-oxide composite thin film layers of different stoichiometry types.
For optical applications, the reduction of the optical absorption of highly-substituted garnet compositions possessing giant Faraday rotation is of principal importance, and we found that this problem is best tackled using the approach of forming oxide-mixed nanocomposites [4]. In summary, we experimentally optimized the dysprosium oxide content to be added to Bi3Fe5O12 by co-sputtering, to adjust the optical, magnetic and MO properties in this new and previously-unexplored (to the best of our group’s knowledge) garnet-oxide material system type. The developed films showed extra-ordinarily high Faraday rotation per unit thickness, low switching fields, very good crystalline quality and good MO figures of merit; however their optical absorption in the visible spectral region was still relatively high. Further research work will be undertaken to reduce the optical losses in this type of garnet materials by way of synthesizing a range of possible nanocomposite materials based on Bi3Fe5O12, Dy2O3, Gd2O3, Sm2O3 and other oxides. We believe that one of the ways in which we can further improve the co-sputtered material system transparency is through the selection of a rare-earth oxide material type suitable for co-sputtering with highly bismuth-substituted garnets, and possessing a minimum of oxygen loss occurring during the RF sputtering deposition.
5. Conclusion
We have prepared Bi3Fe5O12: Dy2O3 (between 2.7 and 20 vol. % of added dysprosium oxide content) garnet-type nanocomposite thin films using RF magnetron co-sputtering in an attempt to synthesize a garnet material with bismuth substitution level approaching 3 formula units. Experimental results have shown that our newly-synthesized MO materials exhibit simultaneously a very high Faraday rotation, good MO quality with almost in-plane magnetization, making them very attractive for several MO device development applications including magneto-plasmonic chips for transparency modulation. Experimental results have also shown that the RF co-sputtering approach is very fruitful for the development of new, high-quality and cost-effective nanocomposite garnet-type materials. In summary, we have successfully developed a new and previously-unexplored MO material system with high Bi-substitution level, giant Faraday rotation and low coercivity.
Acknowledgments
This research is supported by the Electron Science Research Institute, Faculty of Health, Engineering and Science (FHES), Edith Cowan University, Australia.
References and links
1. C. F. Buhrer, “Faraday rotation and dichroism of bismuth calcium vanadium iron garnet,” J. Appl. Phys. 40(11), 4500 (1969). [CrossRef]
2. A. K. Zvezdin and V. A. Kotov, in Modern Magnetooptics and Magnetooptical Materials (Institute of Physics Publishing, 1997).
3. G. B. Scott and D. E. Lacklison, “Magnetooptic properties and applications of Bismuth substituted iron garnets,” IEEE Trans. Magn. 12(4), 292–311 (1976). [CrossRef]
4. M. Vasiliev, M. N.-E. Alam, V. A. Kotov, K. Alameh, V. I. Belotelov, V. I. Burkov, and A. K. Zvezdin, “RF magnetron sputtered (BiDy)3(FeGa)5O12:Bi2O3 composite materials possessing record magneto-optic quality in the visible spectral region,” Opt. Express 17(22), 19519–19535 (2009).
5. T. Okuda, T. Katayama, H. Kobayashi, N. Kobayashi, K. Satoh, and H. Yamamoto, “Magnetic properties of Bi3Fe5O12 garnet,” J. Appl. Phys. 67(9), 4944–4946 (1990). [CrossRef]
6. V. I. Belotelov, L. E. Kreilkamp, I. A. Akimov, A. N. Kalish, D. A. Bykov, S. Kasture, V. J. Yallapragada, A. Venu Gopal, A. M. Grishin, S. I. Khartsev, M. N.-E. Alam, M. Vasiliev, L. L. Doskolovich, D. R. Yakovlev, K. Alameh, A. K. Zvezdin, and M. Bayer, “Plasmon-mediated magneto-optical transparency,” Nat Commun 4, 2128 (2013). [CrossRef] [PubMed]
7. M. Pohl, L. E. Kreilkamp, V. I. Belotelov, I. A. Akimov, A. N. Kalish, N. E. Khokhlov, V. J. Yallapragada, A. V. Gopal, M. N-E. Alam, M. Vasiliev, D. R. Yakovlev, K. Alameh, A. K. Zvezdin, and M. Bayer, “Tuning of the transverse magneto-optical Kerr effect in magneto-plasmonic crystals,” New J. Phys. 15(7), 075024 (2013).
8. M. Vasiliev, V. A. Kotov, K. Alameh, V. I. Belotelov, and A. K. Zvezdin, “Novel magnetic photonic crystal structures for magnetic field sensors and visualizers,” IEEE Trans. Magn. 44(3), 323–328 (2008). [CrossRef]
9. M. N-E. Alam, M. Vasiliev, K. Alameh, and C. Valli, “Magneto-optical visualisation for high-resolution forensic data recovery using advanced thin film nano-materials,” in Proc. International Cyber Resilience Conference, Perth, Australia (2010).
10. M. Vasiliev, K. E. Alameh, V. I. Belotelov, V. Kotov, and A. K. Zvezdin, “Magnetic photonic crystals: 1-D Optimization and Applications for the Integrated Optics Devices,” IEEE/OSA. J. Lightwave Technol. 24(5), 2156–2162 (2006). [CrossRef]
11. M. N-E. Alam, M. Vasiliev, and K. Alameh, “Nano-structured magnetic photonic crystals for magneto-optic polarization controllers at the communication-band wavelengths,” Opt. Quantum Electron. 41(9), 661–669 (2009).
12. N. Adachi, V. P. Denysenkov, S. I. Khartsev, A. M. Grishin, and T. Okuda, “Epitaxial Bi3Fe5O12 (001) films grown by pulsed laser deposition and reactive ion beam sputtering techniques,” J. Appl. Phys. 88(1), 2734–2739 (2000). [CrossRef]
13. M. N-E. Alam, M. Vasiliev, and K. Alameh, “New Class of Garnet Nanocomposites for Use in Magnetic Photonic Crystals Prepared by RF Magnetron Co-sputtering,” In Proc. Int. Conf. on High-Capacity Optical Networks and Enabling Technologies (HONET 2012), Istanbul, Turkey (2012).
14. M. Deb, E. Popova, A. Fouchet, and N. Keller, “Magneto-optical Faraday spectroscopy of completely bismuth-substituted Bi3Fe5O12 garnet thin films,” J. Phys. D Appl. Phys. 45(45), 455001 (2012). [CrossRef]
15. S. Kang, S. Yin, V. Adyam, Q. Li, and Y. Zhu, “Bi3Fe4Ga1O12 garnet properties and its application to ultrafast switching in the visible spectrum,” IEEE Trans. Magn. 43(9), 3656–3660 (2007). [CrossRef]
16. A. H. Eschenfelder, Magnetic Bubble Technology (Springer-Verlag, 1980).
17. M. Vasiliev, M. N-E. Alam, P. Perumal, V. A. Kotov, K. Alameh, Y. T. Lee, and Y. P. Lee, “Annealing behavior and crystal structure of RF-sputtered Bi-substituted dysprosium iron garnet films having excess co-sputtered Bi-oxide content,” J. Phys. D Appl. Phys. 44, 075002 (2011).
18. M. N-E. Alam, M. Vasiliev, K. Alameh, and V. A. Kotov, “Garnet multilayer thin film structure with magnetostatically-altered and improved magnetic properties prepared by RF magnetron sputtering,” In Proc. Int. Conf. on High-capacity Optical Networks and Enabling Technologies Conference (HONET 2011), pp- 177–181, Riadh, Saudi Arabia (2011).
