Shape-controlled synthesis of Fe nanoparticles with high-Ms via modest-temperature reduction method

Article abstract of DOI:10.1002/pssa.200777380

To minimize the oxidation of Fe nanoparticles, 3 to 12 nm Fe nanoparticles have been synthesized by the modest-temperature reduction of FeCl₂ in a non-polar solvent. Trioctylphosphine oxide and oleylamine were used to avoid additional oxidation introduced by surfactants in comparison with oleic acid. The magnetic properties of Fe nanoparticles with different sizes reveal that the saturation magnetization (M_s) of larger than 8.0 nm Fe nanoparticles is over 180 emu/g at 300 K.

Full text of DOI:10.1002/pssa.200777380

Original
Paper
phys. stat. sol. (a) 204, No. 12, 4013–4016 (2007) / DOI 10.1002/pssa.200777380
Shape-controlled synthesis of Fe nanoparticles with high-Ms
via modest-temperature reduction method
1
*, 1
H. T. Yang , T. Ogawa , D. Hasegawa , and M. Takahashi
2
2
1
New Industry Creation Hatchery Center, Tohoku University, Aoba-yama 10, Sendai 980-8579, Japan
Department of Electronic Engineering, Graduate School of Engineering, Tohoku University,
Aoba-yama 05, Sendai 980-8579, Japan
2
Received 2 May 2007, revised 1 August 2007, accepted 20 September 2007
Published online 10 December 2007
PACS 75.50.Bb, 75.75.+a, 81.07.Bc, 81.16.Be
To minimize the oxidation of Fe nanoparticles, 3 to 12 nm Fe nanoparticles have been synthesized by the
modest-temperature reduction of FeCl2 in a non-polar solvent. Trioctylphosphine oxide and oleylamine
were used to avoid additional oxidation introduced by surfactants in comparison with oleic acid. The
magnetic properties of Fe nanoparticles with different sizes reveal that the saturation magnetization (Ms)
of larger than 8.0 nm Fe nanoparticles is over 180 emu/g at 300 K.
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1 Introduction
Iron-based nanoparticles (NPs), such as FePt alloy and iron oxides, have been successfully synthesized
by wet chemistry, and have been studied for a wide range of applications including magnetic recording
media and biomedical field [1–3]. Pure Fe NPs have a greater specific magnetization than either of iron
alloys or oxides, including a high value of saturation magnetization (Ms), a sufficiently high Curie tem-
perature (Tc) for practical applications, and a high permeability in a high-frequency field. Several syn-
thetic approaches, including thermal decomposition of organometallic precursors [4–7], and the reduc-
tion of iron salts inside reverse micelles [8] have been applied for the fabrication of Fe NPs. However, Fe
NPs are highly sensitive to oxidation in air, which has traditionally made the study of Fe NPs difficult
and inconvenient. In the microemulsion method, an environment rich in oxygen and water will accelerate
the formation of iron oxide and hydroxide [8]. The decomposition of Fe(CO)5 is frequently used largely
due to the ease of use and the lack of byproducts. In order to minimize the extent of oxidation, a short-
chain surfactant [9] and heterogeneous nucleation synthesis [10] have been applied in the decomposition
process of Fe(CO)5 to improve the stability of Fe NPs. However, a range of intermediate iron carbonyls
formed in the decomposition process usually introduces carbon atoms into Fe NPs, which results in an
obvious decrease of Ms as compared to 218 emu/g of bulk iron due to the mixture of α-Fe and various
iron carbide phases [11]. Moreover, the decomposition of Fe(CO)5 cannot provide a reductive environ-
ment during the synthesis to avoid oxidation from surfactants with a polar group, typically oleic acid
(OA) with a carboxylic acid group.
In this paper, we report a synthesis method involving modest temperature reduction of an iron salt in
an organic solvent by superhydride (LiBEt3H). A combination of surfactants, alkylphosphine oxide and
oleyl amine, are employed to control particle growth and to avoid the oxidization from surfactants. The
sizes and shape of the Fe NPs are adjusted mainly by varying the molar ratio of the iron salt to the sur-
*
Corresponding author: e-mail: htyang@ecei.tohoku.ac.jp, Phone: +81 22 795 7134, Fax: +81 22 263 9402
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

4014
H. T. Yang et al.: Shape-controlled synthesis of Fe nanoparticles with high-Ms
factants. The magnetic properties of the Fe NPs, including saturation magnetization at different tempera-
tures and sizes, were investigated.
2 Experimental
All commercial reagents were purchased from Aldrich and were purified by using a freeze-pump-flaw
technique prior to use. The synthesis processes were carried out in a glove box with 0.1-1 ppm of O2 and
H2O. In a typical synthesis of 8 nm Fe NPs, 1 mmol FeCl2, 1 mmol trioctylphosphine (TOPO), 1 mmol
oleyl amine (OY), and 20 mL phenyl ether were added into a flask and continuously heated to a tempera-
ture above 200 °C until a clear solution formed. Superhydride solution (2 mL, 1M in THF) was injected
with vigorous stirring. The black dispersion was vigorously stirred for 40 min and then was cooled to
room temperature. The black product was precipitated by adding ethanol to the dispersion. The particles
were re-dispersed in heptane with ~50 µl of oleyl amine. To investigate the influence of surfactants on
the size and shape of the Fe NPs, TOPO/OY was replaced in this synthesis procedure by different com-
binations of surfactants, TOPO/tributylphosphine (TBP) and OA/OY.
Transmission electron microscope (TEM) [Hitachi HF 2000 operating at 200 kV] was employed to de-
termine the morphology and size of the Fe NPs. Samples were formed by drying hexane dispersions of
particles on amorphous-carbon-coated copper grids, which were then transferred to the TEM chamber as
quickly as possible. X-ray fluorescence (XRF) using a Rigaku instrument was used to determine the total
amount of Fe from the integrated intensity of fluorescence of the Kα line of Fe at 6.40 keV. X-rays were
directed onto a fixed mass of the particle powder (by drying the solution of Fe NPs on Si substrate). The
energies of the emitted X-rays are specific to the elements in the sample, and the relative concentrations
of these elements were determined by calibrating the intensities with Fe elemental standard films (purity
of 99.99%) fabricated by an ultra-high vacuum sputtering method. Magnetic measurements were con-
ducted using a Quantum Design MPMS-5 superconducting interference (SQUID) instrument. The
SQUID samples were prepared in a glove box by adding the Fe NPs suspension solution of known vol-
ume fraction (typically 0.5 vol%) into a quartz tube (∅ 5 mm×15 mm). The tube was then sealed in
glove box with Araldite to form an ampule.
3 Results and discussion
Superhydride was specifically chosen for this reduction process since it is easily dissolved in an organic
ether solvent, facilitating homogeneous reduction of the metal salt, which has been previously applied for
the preparation of Co NPs by Sun et al. [12]. It is well known that the –COOH group of OA has high
electronegativity and can lead to the formation of iron carboxylate (-COO-Fe) on the metal NPs which
can prevent further oxidation of the metal NPs in air. Since our synthesis and magnetic measurements
were carried out in an Ar atmosphere without exposure to air, such a protective layer that would decrease
the Ms of the Fe NPs is not necessary. TOPO, which is less reactive as compared to OA, is used along
with OY to avoid formation of an oxide layer as well as to provide steric hindrance.
Figure 1 shows TEM images of Fe NPs synthesized with a reaction time of 40 min and different molar
ratio of paired surfactants (including TOPO/OY, TOPO/TBP, and OA/OY) to FeCl2. At a molar ratio of
0.5, 12.0 nm Fe nanocubes (in Fig. 1a), 15 nm Fe nanopolyhedra (in Fig. 1b), and 11 nm Fe NPs (in
Fig.1c) were observed in the TOPO/OY, TOPO/TBP, and OA/OY, respectively. In Fig. 1a, some small
Fe NPs appear around relatively large Fe nanocubes. This suggests that the formation of Fe nanocubes is
due to Oswald ripening and selective adsorption of TOPO on the surfaces of the Fe NPs, since the fast
injection of superhydride leads to a complete reduction of the Fe salt in a short time and small Fe nuclei
grow in an environment of low surfactant concentration. Shapes of Fe NPs were changed in the case of
replacing OY by TBP and replacing TOPO by OA, which is attributed to TBP providing a greater steric
hindrance than OY, and OA having stronger bonding strength than TOPO. At a high molar ratio (over 1),
Fe NPs were observed in all different paired surfactants serials due to a high surface coverage of surfac-
tants on Fe nuclei. In the TOPO/OY pair, 8.6, 7.2, 5.5 and 3.0 nm Fe NPs were synthesized (as shown in
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.pss-a.com

Original
Paper
phys. stat. sol. (a) 204, No. 12 (2007)
4015
Figs. 1d, e, f, and g, respectively) with increasing molar ratio of from 1 to 5. With the reaction time in-
creasing, the size of Fe NPs also increased. However, there was less influence on the size after the reac-
tion time over 20 min. It should be noted that the oxide layer on the surface of Fe NPs was not clearly
observed in the TOPO/OY pair though Fe NPs are possibly oxidized because of the brief exposure dur-
ing transfer to the microscope. However, in the OA/OY pair, the oxide shell with a thickness of 1-2 nm
can be clearly observed in Fig. 1-(h) for the 9.0 nm Fe NPs. This indicates that the surfaces of the Fe NPs
were oxidized by OA during the synthesis process. Electron diffraction patterns were used for phase
identification of the Fe NPs instead of XRD patterns due to their high sensitivity to oxidation in air. Elec-
tron diffraction patterns can be clearly observed from the larger than 7.2 nm Fe NPs. The typical pattern
of 8.6 nm Fe NPs consisted of four diffraction rings with some spots corresponding to the strongest bcc
Fe diffraction from the (110), (211), (200), and (220) planes, which reveals that the Fe NPs synthesized
by the modest-temperature reduction method had a good crystallinity.
Fig. 1 TEM images of (a) 12.0 nm Fe nanocubes, (b) 15.0 nm Fe nanopolyhedra, and (c) 11.0 nm Fe NPs at a
molar ratio 0.5 of TOPO/OY, TOPO/TBP, and OA/OY pair, respectively. With the molar ratio of TOPO/OY to
FeCl2 from 1 to 5, (d) 8.6 nm, (e) 7.2 nm, (f) 5.5 nm, (g) 3.0 nm Fe nanoparticles were observed. And (h) with the
molar ratio 1 of OA/OY to FeCl2, 9.0 nm Fe nanoparticles with an oxide shell was observed. The upper-right insert
in (d) is the electron diffraction pattern of the 8.6 nm Fe nanoparticles. All scale bars correspond to 50 nm.
The sample sealed in the quartz tube with Araldite showed no magnetic signal decay even after keep-
ing it in air for a week. In all measurements, the diamagnetic background from the quartz tube was sub-
tracted. Figure 2 shows the hysteresis loops of the Fe NPs with different sizes at 300 K and 5 K, and
Table І summarizes the magnetic properties of the Fe NPs. Since small NPs cannot saturate even over 50
kOe, we used the magnetization at 50 kOe, Mat50kOe, instead of the saturation magnetization, Ms, to de-
scribe the magnetization for the NPs. In fact, there is no great difference between Ms and Mat50kOe for
large NPs (larger than 7.0 nm). At 300 K, the samples with a size less than 12 nm are superparamagnetic
without any coercivity. The sample with a size of 12 nm has a coercivity of 30 Oe and shows ferromag-
netic behavior. The Mat50kOe of Fe NPs with a size greater than 8 nm is over 180 emu/g at room tempera-
ture, which is obviously higher than that of the Fe NPs prepared by decomposition of Fe(CO)5 [13, 14].
The relatively high magnetization observed in our experiments can be attributed to the good crystallinity
of Fe NPs with a minimum amount of C atom-diffusion and weak oxidization ability of TOPO/OY. The
behavior of the 3.0 nm NPs is almost similar to the paramagnetic behavior at the applied field between 1
kOe and 50 kOe, and the magnetization is not saturated even above 50 kOe. Therefore, the low magneti-
zation of the Fe NPs with a small size is due not only to the contribution from the surface oxide layer,
but also possibly due to the existence of a non-magnetic surface layer, or spin canting at the outer layer
of the particles. For the hysteresis loops at 5 K, there are two main changes with increasing particle size,
the magnetization saturates at lower fields and the coercivity decreases from 1020 Oe for the 5.5 nm
particles to 425 Oe for the 12.0 nm particles. With increasing size, the corresponding increasing interac-
tions between NPs will couple moments from different particles so that they switch together, which leads
to a lower field required for saturation and reduced Hc. However, for the 3.0 nm NPs, the coercivity
decreases to 260 Oe, which may be due to a finite size effect, and needs to be investigated further in the
future.
www.pss-a.com
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

4016
H. T. Yang et al.: Shape-controlled synthesis of Fe nanoparticles with high-Ms
200
100
0
200
(a)
(b)
12.0 nm
12.0 nm
8.6 nm
8.6 nm
7.2 nm
5.5 nm
3.0 nm
100
7.2 nm
5.5 nm
3.0 nm
0
200
100
0
-100
-100
-100
-4000 -2000
0
2000
4000
H
(Oe)
-200
-200
-40000 -20000
0
20000 40000
-3000-2000-1000
0
1000 2000 3000 4000
H (Oe)
H (Oe)
Fig. 2 Hysteresis loops for Fe nanoparticles with different sizes at (a) 5 K and (b) 300 K. The inset is the corre-
sponding high magnification part from -4 to 4 kOe.
Table 1 Comparison of magnetostatic properties with different nanoparticles size.
Size
Mat 50 kOe
Mat 50 kOe
TB
Hc
(nm) (5K, emu/g) (300K, emu/g) (K)
(5K,Oe)
12.0
8.6
7.2
5.5
3.0
195
189
170
130
90
190
180
144
115
78
225
70
55
35
10
400
490
570
1020
260
In summary, Fe NPs have been prepared by a modest-temperature solution-phase reduction method us-
ing superhydride and a combination of surfactants, TOPO and OY. Electron diffractions shows greater
crystallinity, and the Fe NPs are not significantly oxidized as-formed. The molar ratio of surfactants to
the Fe salt has a large influence on the size as well as shape. Magnetically, the properties of the Fe NPs
in our experiments are dominated by those of their Fe cores instead of iron oxide layers on the surface.
For the more than 8.0 nm Fe NPs, the saturation magnetization at 300 K is always over 180 emu/g,
which proves that our method could open up a new way to prepare Fe NPs with high magnetization.
Acknowledgements We thank Prof. Shouheng Sun for helpful discussions and suggestions. This work is sup-
ported by the New Energy and Industrial Technology Development Organization (NEDO) of Japan and the Japan
Society for the Promotion of Science (JSPS).
References
[1] S. Sun, C. B. Murray, D. Weller, L. Folks, and A. Moser, Science 287, 1989 (2000).
[2] D. W. Elliott and W. X. Zhang, Environ. Sci. Technol. 35, 4922 (2001).
[3] S. Sun, S. Anders, T. Thomson et.al., J. Phys. Chem. B 107, 5419 (2003).
[4] F. E. Luborsky, J. Phys Chem. 61, 1336 (1957).
[5] T. W. Smith, J. Phys. Chem. 84, 1621 (1980).
[6] K. S. Suslick, M. Fang, and T. Hyeon, J. Am. Chem. Soc. 118, 11960 (1996).
[7] F. Dumestre, B. Chaudret, C. Amiens, P. Renaud, and P. Fejes, Science 303, 821 (2004).
[8] F. Li, C. Vipulanandan, and K. Mohanty, Colloids Surf. A 223, 103 (2003).
[9] D. L. Huber, J. E. Martin, E. L. Venturini, P. P. Provencio, and R. J. Patel, J. Magn. Magn. Mater. 278, 311
(2004).
[10] D. Farrell, S. A. Majetich, and J. P. Wilcoxon, J. Phys. Chem. B 107, 11022 (2003).
[11] T. Smith and D. Wychick, J. Phys. Chem. 84, 1621 (1980).
[12] S. Sun and C. B. Murray, J. Appl. Phys. 85,4325 (1999).
[13] C. H. Griffiths, M. P. O’Horo, and T. W. Smith, J. Appl. Phys. 50, 7108 (1979).
[14] M. W. Grinstaff, M. B. Salamon, K. S. Suslick, Phys. Rev. B 48, 269 (1993).
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.pss-a.com