J Mater Sci: Mater Electron (2008) 19:509–513

DOI 10.1007/s10854-007-9372-8

Effect of erbium oxide on synthesis and magnetic properties

of yttrium-iron garnet nanoparticles in organic medium
Haitao Xu Æ Hua Yang Æ Lei Lu

Received: 30 March 2007 / Accepted: 24 July 2007 / Published online: 10 August 2007
Ó Springer Science+Business Media, LLC 2007

Abstract Yttrium iron garnet doped with the erbium oxide structure can be represented by writing the garnet formula
using a conventional complex agent (citric acid) in organic as {R3}[Fe2](Fe3)O12, { }, [ ], ( ) representing 24c
medium (CH3COOH) was synthesized by sol-gel technique. (dodecahedral), 16a (octahedral) and 24d (tetrahedral),
The obtained samples were characterized by X-ray powder respectively [5].
diffraction superconducting quantum interference device, In recent years sol-gel methods have attracted much
scanning probe microscopy and transmission electron attention due to the lower synthesis temperatures and finer
microscope. The influences of the pH value,quantities of and more homogeneous particles produced. By sol-gel
citric acid and the initial concentration of the solution on the route, gelation is frequently attained from the solution of
final product were studied. At the same time, the crystallite inorganic metal salts with a complexing agent [6, 7]. The
size and magnetic properties were observed and studied in concentration of such solutions will produce a homoge-
detail. neous gel if most of the metal ions are forming chelate
complexes. Therefore, the selection of a suitable chelating
agent is important, as well as the pH and the concentrations
in the precursor solution. In addition, it is quite important
1 Introduction to choose an appropriate medium.
For a magnetic material to be applied in microwave
Iron-containing oxide phases with A3B5O12 garnet struc- devices and electrochemical devices, the most important
ture are significant for numerous applications, including static magnetic properties are the saturation magnetization
magnetic materials, laser, phosphorescent sources, micro- (Ms), remanent magnetization (Mrr), anisotropy constants,
wave and electrochemical devices [1–4]. Ferrimagnetic Neel temperature (TN), coercivity (Hc).
garnets are as signed to cubic structure (space group Ia3d), In the present work, yttrium-iron garnet nanoparticles
every cell contains eight R3+ 3+
3 Fe5 O12 molecules. R
3+
ion doped with the erbium oxide were synthesized by a
cannot occupy the octahedral and tetrahedral sites because organic sol–gel technique and their structure was char-
of its large ion radius, so R3+ ion can only occupy acterized. The magnetic properties of the samples were
dodecahedral sites which have larger space. In the case of measured with an emphasis on coercivity and magneti-
ferrimagnetic garnet R3Fe5O12, the ion distribution zation as a function of crystallite size and doped
quantities of erbium oxide in YIG. The microstructural
evolution and magnetic features of synthesized com-
pounds were also discussed.
H. Xu
H. Yang (&)
College of Chemistry, Jilin University, Changchun 130012,
P.R. China
e-mail: [email protected] 2 Experimental
L. Lu
College of Physics, Jilin University, Changchun 130012,
Yttrium iron garnet nanoparticles doped with the erbium
P.R. China oxide were synthesized by sol–gel method using a

123
510 J Mater Sci: Mater Electron (2008) 19:509–513

conventional complex agent (citric acid, Kanto Chemical corresponding to the cubic garnet phase, so the Y3Fe5O12
Inc, Japan, 99.5% purity). The preparation of the samples is with a single garnet phase was prepared at 850 °C. The
detailed below: The yttrium oxide and erbium oxide were crystallization of samples was more complete as the cal-
first dissolved in CH3COOH (containing a little of nitric cination temperature was increased. According to the
acid) (CH3COOH and nitric acid, Hecheng Chemical Inc, Scherrer’s relationship, the crystalline sizes for samples
China, 80.0 and 75.0% purity, respectively). The clear Y3Fe5O12 were calculated. The crystalline sizes increase as
solution of yttrium acetate and erbium acetate were the treatment temperature is increased from 850 to
obtained after stirring for 10 h at 100 °C in a beaker cov- 1,000 °C
ered with a watching glass cover. Iron nitrate (Hecheng In sol–gel methods, the preparation of a homogeneous
Chemical Inc, China, 99.0% purity) was dissolved in gel with respect to the distribution of cations was very
CH3COOH and the solution was stirred for 1 h at 100 °C. important to achieve the composition homogeneity of final
Er(CH3COO)3, Y(CH3COO)3 (Er(CH3COO)3, Y(CH3COO)3 , oxide powder [8]. In organic methods, the role of citric acid
Nanjing Chemical Inc, China, 99.5% purity)and citric acid is to form a complex with metal ions and is expected to
were added to the above solution and then nitric acid was reduce segregation of cations. In the experiment, the ratio
added to adjust pH = 2. The mixture solution was heated at of citric acid to the Y3Fe5O12 (w/w) are 10 in initial
120 °C in order to obtain the gel. To obtain the Y3-xErx- solution. The single garnet phase was not obtained when
Fe5O12 samples, gels were dried initially at 130 °C for 36 h the ratio was below 10. This probably is due to the fact that
and further heat treated in air between 800 and 1,000 °C the amount of citric acid is not sufficient to form a com-
for different time. plete metal ion–citric acid complex, which leads to Y3+ and
The structure and crystallite sizes were determined by Fe3+ not to be dispersed equably. In organic sol–gel
X-ray diffractometer (XRD) in the 2h range 25–651 using methods, the complex formation reactions depend on the
Cu–Ka radiation (k = 0.15405 nm) and scanning probe pH value, and at such low value (pH < 1.5) metal ions can
microscopy (SPM, Nanoscope IIIA). Samples have been be free in solution, no complexing reaction occurring,
crushed flakes at 20 MPa. Voltage in between the probe whereas the preparation of citric acid gels at higher pH
and sample within 50 mV–100 mV. The electricity of value (pH > 2.5) was not possible due to precipitation
tunnel control within 0.2 nA–1.0 nA. The type of XRD is occurring. Thus nitric acid was added to adjust pH = 2 in
SHIMADZU Co, Tokyo Japan. The database of the Joint the initial solution. The initial concentration of the solution
Committee on Powder Diffraction Data was used for the has no significant influence on the final product. In the
interpretation of XRD spectra. The crystalline sizes were experiment, the concentration of yttrium ions, erbium ions
calculated using Scherrer’s relationship: D = kk/B cos h, and iron ions were 0.1 M, 0.1 M and 0.2 M, respectively.
where D is the average diameter in nm, k is the shape Figure 2 shows X-ray diffraction patterns of samples with
factor, and B is the half intensity width of the relevant different Er contents treated at 1,000 °C for 3 h. The X-ray
diffraction peak and k is the X-ray wavelength and h is the diffraction patterns reveal that all samples have only a
Bragg’s diffraction angle. Transmission electron micro-
scope (TEM), Hitachi H-800, was used to characterize the
morphology and microstructure of the samples. Magnetic
measurements were carried out in a superconducting
quantum interference device (SQUID) with a field 1T.

3 Results and discussion

Figure 1 shows the X-ray diffraction patterns for the

Y3Fe5O12 powders calcined at different temperatures. The
patterns of the samples treated at 800 °C contain mainly
YIG phase, perovskite YFeO3 phase, maghemite (a-Fe2O3)
phase and magnetite Fe3O4 phase. With larger quantities of
citric acid in the initial solution, the percentage of perov-
skite, maghemite and magnetite decreases, but even in the
prepared samples with the largest quantity of citric acid,
perovskite, maghemite and magnetite still present. The X-
ray patterns at 850 °C shows a number of diffraction lines Fig. 1 X-ray diffraction patterns of samples calcined at different
(400), (420), (422), (521), (611), (444), (640), (642) temperatures

123
J Mater Sci: Mater Electron (2008) 19:509–513 511

regularity. In addition, the crystallite sizes for samples with

the same Er concentration (x) and calcination temperature
increase as the calcination time is increased. When calci-
nation temperature is at 1,000 °C, the calcination time are
3, 9 and 15 hour, and accordingly, the mean sizes are 60,
150 and 200 nm, respectively. Figure 3a shows the photos
of transmission electron micrograph and electron diffrac-
tion patterns for the Y2.8Er0.2Fe5O12 at 1,000 °C for 15 h.
From the microstructure it is clear that the nanoparticles
aggregate and exhibit irregular rounded shapes, generally
lengthened, with a crystallite size of about 200 nm. Due to
the aggregation of the particles and their irregular shapes, it
is impossible to obtain accurate values of crystallite size.
The interplanar spacing dhkl values were computed from
the diameters of the diffraction rings (1: (211), 2: (400), 3:
(420), 4: (422), 5: (611), 6: (444), 7: (642)). Because the
Fig. 2 X-ray diffraction of samples with Er concentration (x) from experimental values are agreement with standard ASTM
x = 0.2–1.0 treated at 1,000 °C for 3 h
data, the presence of only garnet phase is confirmed. The
single phase structure of garnet. According to the Scher- SPM patterns of Y2.8Er0.2Fe5O12 at 1,000 °C for 15 h are
rer’s relationship, we calculated the crystalline size for shown in Fig. 3b. The sample size is about 150–200 nm,
samples with the same Er concentration (x) but treated at the phase analysis picture of the sample is uniform and
different temperature. The crystalline sizes increase as the exist no obvious other colors, the above phenomenon
treatment temperature is increased from 850 to 1,000 °C. shows that the sample is single phase. The particles size
Furthermore, we also observed that for the same calcina- and structure of the phase observed for the Y2.8Er0.2Fe5O12
tion temperature, the crystallite sizes of Y3-xErxFe5O12 sample analyzed by SPM were in good agreement with
(0.2  x  1.0) vary a little , but not showing apparent these data from TEM.

Fig. 3 (a) TEM and ED

patterns of Y2.8Er0.2Fe5O12
treated at 1,000 °C for 15 h
(b) SPM patterns of
Y2.8Er0.2Fe5O12 treated at
1,000 °C for 15 h

123
512 J Mater Sci: Mater Electron (2008) 19:509–513

In this experiment, we substituted some Er3+ for Y3+. The variation of coercivity (Hc) with the Er concentra-
Because Er3+ ions in Y3-xErxFe5O12 are magnetic, whose tion (x) is shows in Fig 5. It can be seen from the figure
magnetic moment is 9.6 A
m2 (Y3+ ions is non-magnetic that the coercivity (Hc) increases with increasing the Er
and it’s magnetic moment is 0 A m2). According to Neel’s concentration (x), and this can be related to the coercivity
theory [9], there exist three magnetic sub-lattices in YIG: of nanoparticles, which is controlled by irreversible domain
one (c) formed by the Er3+ ions occupying the dodecahe- rotation, defined by the equation Hc = ± 2 K1/l0Ms,
dral sites, another (a) formed by Fe3+ ions occupying the where K1 is the magnetocrystalline anisotropy constant, l0
octahedral sites and the third (d) formed by the Fe3+ ions is vacuum susceptibility.
occupying the tetrahedral sites. The two iron sub-lattices In general, the rare-earth ions have stronger s-l coupling
are coupled antiferro-magnetically by the super exchange and weaker crystal field, so they have stronger magneto-
interaction acting via the intervening O2 ions. The c sub- crystalline anisotropy. Besides, saturation magnetization of
lattice is coupled antiferro-magnetically with the two iron YIG sample is higher than that of rare-earth ions doped
sub-lattices. At room temperature, the three sub-lattices sample. According to the above relationship, we can
moments align along the [111] direction. The net magnetic explain why Y2.0Er1.0Fe5O12 sample have a considerable
moments is M = jMd  Maj  Mc. The variation of sat- high coercivity [10]. The magnetic hysteresis loop of
uration magnetization (Ms) with the Er concentration (x) is
shown in Fig. 4. The samples were treated at 1,000 °C for
3 h. From the figure, it can be observed that the saturation
magnetization (Ms) decreases with increasing the Er con-
centration (x), and this can be related to the fact that the
magnetic moments of Er3+ ions align oppositely to the
effective moments formed by Fe3+ ions.
For sample Y2.8Er0.2Fe5O12, the crystallite sizes are
45.2, 52.8 and 55.9 nm (the crystalline sizes were calcu-
lated using Scherrer’s relationship D = kk/B cosh) at 850,
900 and 1,000 °C, and accordingly, the saturation magne-
tization (Ms) are 21.1, 23.2 and 24.5 emu/g, respectively.
One can conclude that the saturation magnetization (Ms)
rises with increasing the particle’s size. This rise can be
related to the higher surface-to-volume ratio in the smaller
particles, which results in the existence of non-magneti-
zation surface layer; and the saturation magnetization (Ms)
of the particles decreases as the particle’s sizes reduce.
Fig. 5 The variation of coercivity (Hc) with the Er concentration (x)

Fig. 4 The variation of saturation magnetization (Ms) with the Er Fig. 6 Hysteresis loop patterns of Y2.0Er1.0Fe5O12 samples at room
concentration (x) temperature

123
J Mater Sci: Mater Electron (2008) 19:509–513 513

be formed through such reaction pathways: The gelatin is

first partially transformed to YFeO3 with a mixture of
Fe3O4 and a-Fe2O3, which transforms to YIG by further
heat treatment in a diffusion-controlled process. The sam-
ple Y3-xErxFe5O12 (0  x  1.0) with a single garnet
phase was prepared at 850 °C. The crystallite size increases
with increasing the treating temperature and time. The
samples exhibit irregular rounded shapes due to the
aggregation, and have obvious boundary, being different
from synthetical samples in water medium. The saturation
magnetization is reduced but coercivity increases, and at
the meantime remanent magnetization is no changed with
increasing the Er concentration (x). The saturation mag-
netization is increased but remanent magnetization is
reduced with increasing the crystallite size.

Acknowledgement This work is supported by the National Natural

Fig. 7 The variation of remanent magnetization with the tempera- Science Foundation of China (NSFC) (Grant No. 50572033).
tures. The inset shows the variation of remanent magnetization with
the Er concentration (x)

Y2.0Er1.0Fe5O12 at room temperature is shown in Fig. 6. References

The remanent magnetization of samples is not changed
1. V.V. Kharton, A.L. Shaula, E.N. Naumovich, J. Elec. Soc. 150, J
obviously with increasing the Er concentration (x). More- 33 (2003)
over, we also observed that for the same Er concentration 2. J.W. Ostrander, N.A. Kotov, J. Am. Chem. Soc. 123, 1101–1110
(x), the remanent magnetization decreases with increasing (2001)
the temperature (Fig. 7). 3. T.C. Mao, J.C. Chen, J. Magn. Magn. Mater. 302, 74–81 (2006)
4. M. Rajendran, S. Deka, J. Magn. Magn. Mater. 301, 212–219
(2006)
5. S. Mandal, S. Phadtare, Curr. Appl. Phys. 5, 118–127 (2005)
4 Conclusions 6. P. Vaqueiro, M.A. Lopez-Quintela, Chem. Mater. 9, 2836–2841
(1997)
In conclusion, we have synthesized Y3-xErxFe5O12 (0  7. E. Garskaite, K. Gibson, Chem. Phys. 323, 204–210 (2006)
8. X.Z. Guo, B.G. Ravi, J. Magn. Magn. Mater. 295, 145–154
x  1.0) with a single garnet phase sample in organic (2005)
medium (CH3COOH) by a sol–gel method. The experi- 9. D. Rodic, M. Mitric, J. Magn. Magn. Mater. 191, 137–145 (1999)
ments conducted in this study suggest that YIG phase can 10. A.T. Ngo, P. Bonville, J. Appl. Phys. 89, 3370 (2001)

123