A large range of Bi-substituted iron garnet (Bi:IG) compounds with
compositions described by the generic formula (Bi, RE[).sub.3](Fe,
M[).sub.5][O.sub.12], where RE stands for a rare-earth atom and M is a
metal atom like Ga or Al, have been synthesized since the 1960s using
various physical [1-7] and chemical [8-11] methods. Bi:IGs have
initially found widespread application in magnetic recording
technologies, after the discovery of giant Faraday effect in Bi:IG
garnet materials in 1969 . In recent decades, significant research
attention had been devoted to the synthesis of highly Bi-substituted
iron garnets by multiple groups working worldwide in diverse fields
ranging from magnetic data recording to photonics and quantum-optical
information processing [12-16]. Highly Bi:IG materials in which the Bi
content approaches its theoretical maximum of 3 formula units are very
promising for optics and photonics applications due to possessing
record-high Faraday rotation which, in turn, allows control over the
polarization states of the transmitted light signals on a
nanosecond-range time scale. The optical, magnetic, and magneto-optical
(MO) properties of all garnet materials depend significantly on the Bi
substitution levels (within the dodecahedral sublattice of the garnet
structure) as well as on all other substitution atoms and dopants
content . More recently, many new and exciting technologies which
rely on using Bi:IG compounds have started to appear, ranging from the
synthesis of magnetic-garnet nanoparticles  to the development of
magnetic photonic crystals  for a wide range of applications.
Various physical approaches including crystal growth from melts and from
vapor phase (liquidphase epitaxy and vapor-phase epitaxy techniques) and
physical vapor deposition techniques (such as radio-frequency, RF,
magnetron sputtering and pulsed laser deposition) have made the growth
of highquality, impurity-phase-free garnet-phase layers on various
substrate types feasible [4-7]. In most cases, the use of physical vapor
deposition techniques leads initially to growing the amorphous-phase
oxide mixes (for all substrate temperatures during the deposition
process below the crystallization temperature of any given garnet type),
which do not possess any ferrimagnetic properties characteristic of
Bi:IG. For high substrate temperatures (in excess of 500-600
[degrees]C), the deposition processes can lead to growing the in situ
crystallized or (most often) the partially crystallized (poly- or
nanocrystalline) layers with the required garnet-type volume-averaged
stoichiometry, which then still require suitably optimized
post-deposition annealing processes for garnet phase formation. Finely
optimized deposition processes using hot monocrystalline garnet
substrates which are very closely lattice-matched to the growing garnet
layers have led to obtaining epitaxial-quality garnet layer growth .
Despite the fact that all physical vapor deposition pathways toward the
Bi:IG synthesis require complex and multi-parameter process
optimizations at both the deposition and annealing stages, this approach
to garnet synthesis has the strongest potential for the practical use of
this class of functional materials in photonic integrated circuits,
optical fiber components, and other technologies linked to using
microfabrication processes. We believe that RF sputtering deposition
followed by the oven post-annealing in air atmosphere is the most
flexible approach to MO garnet materials synthesis, which is also most
compatible with other material types and with various modern
microfabrication technologies. Most importantly, this approach allows
the most efficient engineering of the garnet material properties through
conveniently implementing the required variations in the chemical
composition without having to redesign any principal synthesis pathways.
In this paper, we report on the synthesis of several important
types of high-performance nanocrystalline ferrimagnetic garnet materials
suitable for use in various photonics applications and also in MO
imaging, as well as document a number of technological process details
and parameters relevant to the synthesis of high-quality thin-film
garnets. For the first time, we report on our estimates of the kinetics
parameters of the crystallization processes relevant to the synthesis of
MO garnet nanocomposites, which have recently been demonstrated to
possess world-record MO figures of merit in the visible spectral range
FILM SYNTHESIS AND CHARACTERIZATION
Multiple thin-film batches of Bi-substituted dysprosium-gallium
iron garnet, Bi-substituted lutetium-aluminum iron garnet, and also
several types of composite garnet-[Bi.sub.2][O.sub.3] films were
prepared on glass (Corning 1737) and monocrystalline GGG (111)-oriented
substrates by RF magnetron sputtering in low-pressure (1-2 mTorr) argon
plasma under the process conditions specified in Table 1. The
as-deposited films were amorphous and were later subjected to
high-temperature oven annealing processes in air atmosphere to
synthesize garnet-phase layers as a result of crystallization. As was
expected, the optical and magnetic properties of the garnet materials
synthesized depended significantly on their composition type and the
degree of crystallization. The nanocomposite (co-sputtered)
garnet-Bi-oxide layers had a significantly improved optical transparency
compared to the same garnet types deposited without co-sputtering
[Bi.sub.2][O.sub.3], which is illustrated in Fig. 1.
[FIGURE 1 OMITTED]
The optimum annealing regimes suitable for the crystallization of
all garnet layers and (especially) the garnet-[Bi.sub.2][O.sub.3]
composite films were found to be extremely composition-dependent. We
experimentally optimized the annealing temperatures and process
durations for the crystallization of a range of our garnet material
types, in order to avoid the garnet decomposition processes and film
surface degradation caused by the thermal over-exposure. The
crystallized garnet material layers were characterized comprehensively
using spectrophotometry, X-ray diffractometry (XRD), polarization
microscopy, microstructure imaging, and MO testing.
The nominal stoichiometries of the sputtering targets
and [Bi.sub.2][O.sub.3] were used to deposit a large batch of composite
garnet-oxide films of about 1100 nm thickness containing an estimated
(17 [+ or -] 2) vol % of added bismuth oxide, and the crystallization
process kinetics of samples from this batch was studied in some detail
for three different annealing temperatures (550, 560, and 570
[degrees]C) to generate an estimate for the activation energy of the
crystallization process. Multiple reports on the optical, magnetic, and
MO properties of various MO garnet material types have been published to
date, as well as some reports on the thermal processing techniques
suitable for crystallizing Bi:IG films and also on the annealing
behaviors of different MO garnet compounds [17-22]. Several studies have
also been devoted in particular to the optimization of rapid thermal
annealing regimes suitable for the crystallization of garnet films
[18,19]. However, to the best of our knowledge, no reports have yet been
published on the crystallization kinetics parameters of oven-annealed
highly Bi-substituted nanocomposite iron-garnet-[Bi.sub.2][O.sub.3]
materials. The only report describing the crystallization kinetics and
providing an estimate of the activation energy of crystallization for
oven-annealed epitaxial garnet films could be found in . We analyzed
the time- and temperature dependencies of the extent of crystallization
of our garnet-oxide materials and found these (as was expected) to
satisfy the Avrami equation  and Arrhenius law for isothermal
crystallization [24-26]. The activation energy of isothermal
crystallization of our materials was found to be rather large (similarly
to the results presented in ), which helps explain the substantial
differences in the optimum annealing durations for processes run at
different temperatures in this material type.
RESULTS AND DISCUSSION
Two important types of highly-Bi-substituted MO doped-iron-garnet
compounds have been synthesized repeatably as high-quality thin films on
optical substrates (Corning 1737 glass and polished monocrystalline
[Gd.sub.3][Ga.sub.5][O.sub.12] (GGG)) and later characterized
comprehensively. The first material type was described by the formula
[(Bi,Dy).sub.3][(Fe,Ga).sub.5][O.sub.12] and its nanocomposite
derivatives of type [(Bi,Dy).sub.3][(Fe,Ga).sub.5][O.sub.12]:[Bi.sub.2][O.sub.3]. We have previously reported on the optical and MO properties
of this type of materials in . The second material type tested was
[(Bi,Lu).sub.3][(Fe,Al).sub.5][O.sub.12] and its nanocomposite
[(Bi,Lu).sub.3][(Fe,Al).sub.5][O.sub.12]:[Bi.sub.2][O.sub.3]. The two
material types possessed similar optical properties and crystallization
behavior, yet had very different magnetic and MO properties. To the best
of our knowledge, the synthesis of garnet materials of the second type
with high Bi content using any physical vapor deposition methods have
not yet been reported in the literature.
Bi-substituted Ga-doped dysprosium iron garnets
synthesized on glass substrates have been studied using X-ray
diffraction analysis and microstructure imaging (transmission electron
microscopy, TEM). The diffraction datasets obtained with Panalytical
XPert Pro X-ray diffractometer configured for near-grazing-incidence
measurements using CuK[[alpha].sub.1] line ([lambda] = 0.15406 nm) are
shown in Fig. 2, together with a cross-section microstructure image of a
garnet layer obtained using high-resolution TEM. The diffraction
patterns obtained showed peaks at a set of angles characteristic of the
body-centered cubic lattice of garnets. Interestingly, the only impurity
phase detected reliably was iron oxide [Fe.sub.3][O.sub.4] (identified
through a search of the standard JCPDS XRD database limited by all
chemical elements possibly present in the film structure). The source of
excess [Fe.sub.3][O.sub.4] was the sputtering target of stoichiometry
[Bi.sub.2][Dy.sub.1][Fe.sub.4][Ga.sub.1][O.sub.12] sintered from a mix
of oxides, and all samples tested (as-deposited and crystallized) showed
XRD peaks at angles (2[THETA]) near 44.5[degrees] and 65[degrees]. The
diffraction datasets of the films synthesized by co-sputtering
deposition with [Bi.sub.2][O.sub.3] showed significantly reduced
iron-oxide peak intensities, suggesting that less iron oxide was present
outside garnet nanocrystallites. [Bi.sub.2][O.sub.3] still remained in
its amorphous phase after annealing and therefore did not generate any
[FIGURE 2 OMITTED]
The optical properties of
[(BiDy).sub.3][(FeGa).sub.5][O.sub.12]:[Bi.sub.2][O.sub.3] films as well
as the oven-annealing regimes found suitable for the crystallization of
this type of garnets have been reported in detail in .
Bi-substituted Al-doped lutetium iron garnets
Our motivation for the synthesis of
[(Bi,Lu).sub.3][(Fe,Al).sub.5][O.sub.12] films on optical substrates
(glass and GGG) has been to establish the technology for the manufacture
of magnetically soft garnet layers with high Bi substitution,
magnetization vector direction close to being in the film's plane,
and possessing high MO figures of merit in the visible and near-infrared
ranges. The latter figure is defined as the doubled ratio of the
specific Faraday rotation to the absorption coefficient at any given
wavelength (Q = 2[[THETA].sub.F]/A) and determines the suitability of MO
materials for most practical applications.
[(Bi,Lu).sub.3][(Fe,Al).sub.5][O.sub.12] films have only been
synthesized previously by using liquid-phase epitaxy, but that technique
does not allow synthesis of garnet films with high Bi substitution
(approaching 2.0 formula units and above). We selected a sputtering
target of stoichiometry
[Bi.sub.1.8][Lu.sub.1.2][Fe.sub.3.6][Al.sub.1.4][O.sub.12] as deposition
material source and deposited several batches of garnet and also
garnet-[Bi.sub.2][O.sub.3] films using essentially the same deposition
process parameters as for Bi-substituted dysprosium iron garnets. The
result was the development of new thin-film garnet material with optical
and MO properties useful for a range of applications in MO sensors and
magnetic field imagers and possessing very high MO figure of merit. As
was expected, the synthesis of co-deposited nanocomposites of type
[(Bi,Lu).sub.3][(Fe,Al).sub.5][O.sub.12]:[Bi.sub.2][O.sub.3] has led to
films with improved optical transparency and therefore with better MO
quality. The effect of excess [Bi.sub.2][O.sub.3] concentration on the
magnetic properties of these materials is currently under investigation.
The initial results of the optical, magnetic, and MO characterization of
this type of MO films are presented in Figs. 3 and 4.
[FIGURE 3 OMITTED]
[FIGURE 4 OMITTED]
Kinetics of garnet nanocrystals formation within the
Our analysis of the annealing behavior of garnet-oxide composites
was limited to the conventional oven-annealing processes, in which
isothermal crystallization was achieved at the "annealing
temperatures" as specified, and the temperature-ramp processes were
run at a constant rate of 5 [degrees]C/min. Since we estimated the
absolute accuracy of temperature control achievable in our oven to be
about [+ or -]5 [degrees]C, we defined the "annealing process
duration" in our analysis to be equal
to the duration of the isothermal crystallization process plus two
minutes, in order to account for the effects of the last minute of the
temperature ramp-up process and the first minute of the temperature
ramp-down process. We performed a series of annealing experiments with a
large batch of garnet-oxide nanocomposite films of composition type
[Bi.sub.2][Dy.sub.1][Fe.sub.4.3][Ga.sub.0.7][O.sub.12]:[Bi.sub.2][O.sub.3] (17 vol %) using a conventional box- furnace-type oven system
(Sentrotech, Inc., USA) at the process temperatures of 550, 560, and 570
[degrees]C, for a number of different annealing durations, and studied
the evolution of the optical and MO properties to characterize the
garnet crystallization kinetics of this material. We observed
significantly shorter optimum annealing durations which led to maximized
specific Faraday rotation (about 4-6 min) for processes run at 570
[degrees]C compared to these run at 550 [degrees]C (about 45-60 min).
The amorphous (as-deposited) to nanocrystalline phase
transformation kinetics was studied using the samples with high-quality
surfaces (except in some over-annealed films), and possessing good
optical and MO properties. Since the conventional oven-annealing
processes induce the temperature-activated crystallization of amorphous
(as-deposited) garnet layers, it is of interest to quantify the
activation energy of these crystallization processes. We followed the
approach described in [23-26] for this analysis. It is well known that
the kinetics of thermally activated processes (including the case of
garnet crystallization being considered) follows a dependence on
temperature described by the Arrhenius law (eq. 1)
K = [K.sub.0] * exp( - [E.sub.c] / kT) (1)
where K is the rate constant of the crystallization process,
[K.sub.0] is termed the pre-exponential factor, [E.sub.c] is the
activation energy of crystallization, and k is the Boltzmann constant.
During the isothermal phase change, the extent of crystallization A of a
material is described by Avrami's equation (eq. 2) 
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)
where t is the process duration and n is the order parameter
dependent on the crystal growth mechanism.
The extent of crystallization was quantified as the ratio of the
specific Faraday rotation measured in films after running any given
annealing process, to the maximum specific Faraday rotation at the same
wavelength (we used 532- and 635-nm light) achieved in any given
material/substrate system after running the process at the same
annealing temperature with the "best-known" parameters. All
specific Faraday rotation data obtained from film samples deposited onto
GGG substrates and possessing rectangular hysteresis loops were measured
in their remnant magnetization states, in order to exclude the
paramagnetic effects of 0.5-mm-thick GGG substrates which led to
measuring slightly increased Faraday rotation angles when placed in the
electromagnet's field of up to 2.5 kOe. For samples deposited onto
Corning 1737 glass substrates, Faraday rotation data was taken at the
saturation magnetization, even though all films maintained about 90 % of
Faraday rotation after being removed from the electromagnet. The data
analysis strategy used to derive the Arrhenius equation parameters of
our garnet materials followed the approach reported in [24-26].
Figure 5 illustrates a typical relationship between the measured
extent of crystallization and the annealing time (a sigmoid-shaped
crystallization curve) and some data used to derive the estimate for the
activation energy. It is important to note that some data points in the
dependencies of the extent of crystallization on process duration were
considered "outliers" since these were observed outside the
expected linear trends in the logarithmic plot used for linear
regression fitting. The latter was likely due to the variations in the
phase content of film samples (the presence of some residual iron oxide
and also variations in excess vol % of [Bi.sub.2][O.sub.3]).
Avrami plots of ln[ln(1/(1-A))] vs. ln(t) were studied to reveal
the values of crystallization order parameters and also these of ln(K).
The "outlier" or "over-annealed" data points not
used in the regression fitting are also shown within dashed circles. All
time durations were measured in minutes. The slopes and the intercepts
of the plots of ln[ln(1/(1-A))] vs. ln(t) as shown in Fig. 5b revealed
the (approximate) values of the crystallization order parameter and also
these of ln(K) at each process temperature. According to the Arrhenius
expression (eq. 1), the crystallization rate constant is a function of
temperature, and it is well known that it also depends on both the
nucleation and growth rates of the new phase. Annealing at higher
temperatures increases the growth rate substantially, but studies of
crystallization at temperatures above 570 [degrees]C could not be
carried out due to the oven temperature controller's limitations.
On the other hand, annealing at low temperatures (in our case, below 550
[degrees]C) was not productive due to the very slow crystallization
rates, which could mean that weeks or even years of annealing would be
required to transform the amorphous-phase materials into polycrystalline
[FIGURE 5 OMITTED]
For thermally activated isothermal crystallization processes, the
Arrhenius equation shows that the crystallization rate constant is
significantly temperature-dependent, and the obtained linear-regression
straight-line fits of the ln(K) vs. 1000/T data points confirm the
applicability of Arrhenius law to the process considered. The slopes of
fitted lines shown in the Arrhenius plots of Figs. 5c,d are defined by
the activation energy of crystallization, whilst the intercepts can be
used to reveal the value of the pre-exponential factor of the
crystallization process (ln([K.sub.o])). Only three temperature data
points were used due to the narrowness of the thermal processing window
suitable for the material type selected, and because we utilized only (a
limited number of) material samples from a single deposition batch.
Therefore, any material composition uncertainties arising out of
batch-to-batch repeatability were excluded. Using eq. 1 and the fitted
regression slope values shown in Figs. 5c,d, the following estimates of
the activation energy of garnet nanocrystals formation were obtained:
14.99 eV for films deposited onto GGG substrates and 15.6 eV for films
deposited onto Corning 1737 glass substrates; both values were estimated
to about [+ or -]10 % accuracy. It is important to note that the
activation energy values obtained from our estimates were much larger
than those expected for simple diffusion-driven processes. This might
indicate that diffusion and also nucleation/growth mechanisms as well as
interface processes may affect the overall kinetics of the rather
complex crystallization process, as described by the authors of .
The optimization of annealing regimes used for this material type
with est. (17 [+ or -] 2) % of excess bismuth oxide coming from a
different target) enabled some improvement in the MO figures of merit
(as measured in the visible range) compared with the results previously
achieved for this material type and published in . The chemical
composition, crystal structure, and the microstructural properties of
our range of garnet-oxide nanocomposite films annealed using different
regimes are being investigated and will be reported elsewhere.
We have studied the synthesis of several important types of Bi:IG
compounds using RF sputtering deposition technology and high-temperature
post-deposition annealing. The properties of MO garnet thin films of
composition types [(BiDy).sub.3][(FeGa).sub.5][O.sub.12],
[(BiLu).sub.3][(FeAl).sub.5][O.sub.12]:[Bi.sub.2][O.sub.3] have been
discussed. We have shown that these materials can possess very high MO
performance, nanocrystalline densely packed microstructure with very
fine grains, and very attractive magnetic properties, making them useful
for diverse applications in nanophotonics, ultrafast optical devices,
and integrated optoelectronics. A study of crystallization kinetics
undertaken with a sample batch of garnet-[Bi.sub.2][O.sub.3] composite
films has yielded the estimates of the activation energy of isothermal
crystallization for this type of material. The results presented can be
used as a guide for obtaining high-performance garnet films and for the
design of optimized thermal processing regimes suitable for the
synthesis of highly Bi-substituted garnets using physical vapor
This research is supported by the Faculty of Computing, Health and
Science, Edith Cowan University. We also acknowledge the support
provided by the Department of Nanobio Materials and Electronics, Gwangju
Institute of Science and Technology (Republic of Korea).
[1.] G. B. Scott, D. E. Lacklison. IEEE Trans. Magn. 12, 292
[2.] A. K. Zvezdin, V. A. Kotov. In: Modern Magnetooptics and
Magnetooptical Materials, Institute of Physics Publishing, Bristol and
[3.] C. F. Buhrer. J. Appl. Phys. 40, 4500 (1969).
[4.] T. Okuda, N. Koshizuka, K. Hayashi, T. Takahashi, H. Kotani,
H. Yamamoto. Advances in Magneto-Optics, Proc. Int. Symp.
Magneto-Optics, J. Magn. Soc. Jpn. 11, Supplement S1, 179 (1987).
[5.] Y. H. Kim, J. S. Kim, S. I. Kim, M. Levy. J. Korean Phys. Soc.
43, 400 (2003).
[6.] S.-H. Wee, H.-S. Hong, Y.-H. Kim, S.-I. Yoo, J. Kang. Metals
Mater. Int. 9, 507 (2003).
[7.] M. Vasiliev, M. Nur-E-Alam, V. A. Kotov, K. Alameh, V. I.
Belotelov, V. I. Burkov, A. K. Zvezdin. Opt. Express 17, 19519 (2009).
[8.] X. Ma, S. Zhang, F. Li, D. Que. J. Mater. Sci.: Mater.
Electron. 9, 347 (1998).
[9.] N. Dubnikova, E. Garskaite, J. Pinkas, P. Bezdicka, A.
Beganskiene, A. Kareiva. J. Sol-Gel Sci. Technol. 55, 213 (2010).
[10.] Y. Zhang, X. Wang, H. Xia, D. Shen, F. Gan, S. Wang. J.
Mater. Sci. Technol. 20, 66 (2004).
[11.] T. Kim, S. Nasu, M. Shima. J. Nanoparticle Res. 9, 737
[12.] K. Nakagawa, S. Kurashina, A. Itoh. J. Appl. Phys. 75, 7096
[13.] S. Kang, S. Yin, V. Adyam, Q. Li, Y. Zhu. IEEE Trans. Magn.
43, 3656 (2007).
[14.] M. Nur-E-Alam, M. Vasiliev, K. Alameh. Opt. Quant. Electron.
41, 661 (2010).
[15.] A. Abdelrahman, M. Vasiliev, K. Alameh, P. Hannaford. Phys.
Rev. A 82, 012320 (2010).
[16.] I. L. Lyubchanskii, N. N. Dadoenkova, M. I. Lyubchanskii, E.
A. Shapovalov, Th. Rasing. J. Phys. D: Appl. Phys. 36, R277 (2003).
[17.] W. T. Stacy, M. A. H. Huyberts, R. Metselaar, A. B. Voermans.
J. Appl. Phys. 48, 4766 (1977).
[18.] T. Suzuki, G. Zaharchuk, G. Gorman, F. Sequeda, P. Labun.
IEEE Trans. Magn. 26, 1927 (1990).
[19.] T. Suzuki. J. Appl. Phys. 69, 4756 (1991).
[20.] F. B. Hagedorn. J. Appl. Phys. 45, 3123 (1974).
[21.] A. Sztaniszlav, M. Farkas-Jahnke, M. Balla. J. Magn. Magn.
Mater. 215-216, 188 (2000).
[22.] Q. Yang, Z. Huaiwu, L. Yingli, W. Qiye. IEEE Trans. Magn. 43,
[23.] M. Avrami. J. Chem. Phys. 8, 212 (1940).
[24.] N. Mehta, A. Kumar. J. Mater Sci. 42, 490 (2007).
[25.] G. H. Adel, A. S. Abdrabo. J. Ovonic Res. 5, 51 (2009).
[26.] R. P. Mathur, D. Akhtar. Bull. Mater. Sci. 9, 267 (1987).
Mohammad Nur-E-Alam (1)([double dagger]), Mikhail Vasiliev (1),
Kamal Alameh (1,2)([double dagger]), and Viacheslav Kotov (3)
(1) Electron Science Research Institute, Edith Cowan University,
270 Joondalup Dr., Joondalup, WA 6027, Australia; (2) Department of
Nanobio Materials and Electronics, Gwangju Institute of Science and
Technology, Korea; (3) Institute of Radio Engineering and Electronics,
Russian Academy of Sciences, 11 Mohovaya St., Moscow 125009, Russia
* Paper based on a presentation made at the International
Conference on Nanomaterials and Nanotechnology (NANO-2010),
Tiruchengode, India, 13-16 December 2010. Other presentations are
published in this issue, pp. 1971-2113.
([double dagger]) Corresponding authors: E-mail:
firstname.lastname@example.org and email@example.com
Table 1 Typical RF sputtering deposition
conditions and annealing process parameters
used during the synthesis
of high-quality MO garnet films.
Process parameters Values/comments
sputtering target [Ga.sub.1][O.sub.12],[Bi.sub.2]
target was only used in
Sputter gas Ar, P(total) = 1 mTorr. No oxygen input.
Base pressure P(base) < 1-2E-06 Torr (high vacuum)
RF power densities 3.3-7 W/cm2 (150-320 W, 3" targets)
- garnets; 0.44-0.88 W/[cm.sub.2]
(20-40 W - co-sputtered [Bi.sub.2]
Deposition rates [O.sub.3]) 3.5-8.7 nm/min (garnets)
1.2-5 nm/min ([Bi.sub.2][O.sub.3]
Substrate temperatures 250 [degrees]C (typ.)
during deposition 30-40
rotation rate (rpm)
Target-to-substrate 18 [+ or -] 2 cm
Sputtering system model KVS-T4065 (Korea Vacuum
and description Technology, Ltd), down-sputtering
type, three RF guns for 3" targets.
Crystallization process 1 h @ 700 [degrees]C
temperatures and optimum ([Bi.sub.2][Dy.sub.1][Fe.sub.4]
annealing duration [Ga.sub.1][O.sub.12] and [Bi.sub.2]
(typical values) [Dy.sub.1][Fe.sub.4.3][Ga.sub.0.7]
[O.sub.12]) 3 h @ 630 [degrees]C
[Al.sub.1.4][O.sub.12]) 30 min @
580 [degrees]C ([Bi.sub.2][Dy.sub.1]
[Bi.sub.2][O.sub.3], 25 vol %
excess [Bi.sub.2][O.sub.3]); 30 min @
560 [degrees]C ([Bi.sub.2][Dy.sub.1]
[Bi.sub.2][O.sub.3], 17 vol %
excess [Bi.sub.2][O.sub.3]); 10 h @
610 [degrees]C ([Bi.sub.1.8]
4.5 vol % excess [Bi.sub.2][O.sub.3]);
annealing regimes for all
garnet-oxide composite formulations.