You are on page 1of 5

660 IEEE TRANSACTIONS ON MAGNETICS, VOL. 41, NO.

2, FEBRUARY 2005

Chemically Synthesized L10-Type FePt Nanoparticles


and Nanoparticle Arrays via Template-Assisted
Self-Assembly
Yuichi Sasaki, M. Mizuno, A. C. C. Yu, T. Miyauchi, D. Hasegawa, T. Ogawa, M. Takahashi, B. Jeyadevan, K. Tohji,
K. Sato, and S. Hisano

Abstract—Chemically ordered L10 -type FePt nanoparticle ag- behavior. To obtain ordered structure, high temperature an-
glomerates were synthesized directly by the co-reduction of Fe(III) nealing is required. A self-assembled array of such particles
and Pt(II) acetylacetonates in tetraethylene glycol at 300 C in the shows well-aligned close-packed nanostructures on a substrate,
absence of surfactants. These nanoparticles could be dispersed in
n-hexane by coating with oleic acid and oleylamine. However, the however, each particle strongly coalesces during annealing [10],
dispersed particles exhibited only chemically disordered fcc phase which results in undesirable increase in size and its distribu-
and superparamagnetic behavior. The FePt nanoparticle film tion. Though doping with a third element (Ag) was effective in
composed of dispersed particles and stabilized using amino-silane reducing the transformation temperature to 350 C, it was im-
began to structurally transform to ordered L10 phase at 600 C, possible to prevent particles from coalescing completely [11].
which is lower compared to that prepared by the hot soap method.
Rotational hysteresis loss measurement suggested that the or- We recently reported that -type FePt nanoparticles could
dering was incomplete at 600 C and the nanoparticle film had be obtained directly in solution phase at 300 C by using the
the distribution of magnetocrystalline anisotropy field values. The “modified polyol process”, which opened up the possibility for
FePt nanoparticle array was fabricated using the template-assisted the fabrication of the ultrahigh density nanoparticle thin film
self-assembly technique. To produce periodic dots on a substrate, media without heating process [12]. As the nanoparticle syn-
positive-biased pulse voltage was applied to the substrate coated
with octadecyltrichlorosilane monolayer by using a conducting thesis was carried out without using any stabilizers such as sur-
cantilever used in a scanning probe microscope. This process factants in solution, these particles agglomerated strongly so
induced electrochemical modification of CH3 groups into polar that the coating of the nanoparticles on a substrate was diffi-
ones. The resulting template had well-aligned sub-100-nm dot cult. In the first part of this paper, the dispersion of the directly
arrays with sub-100-nm periodicity. The FePt nanoparticles were synthesized -type FePt nanoparticles in an organic solvent
fixed on the patterned areas selectively.
is investigated. The unique morphological and crystallographic
Index Terms—Dispersion, electrochemical modification, FePt characteristics are examined and compared with those synthe-
nanoparticle, L10 structure, magnetic anisotropy field, polycrys-
talline, polyol process, scanning probe lithography, template-as- sized by the hot-soap method [13]. The investigation of their
sisted self-assembly. magnetic properties focusing on the magnetic anisotropy is also
discussed.
The ultimate goal in the development of ultrahigh density
I. INTRODUCTION recording media is to store one bit per one magnetic unit. To
achieve this, the establishment of periodic arrays is necessary.
T HE -type alloys such as FePt and CoPt have great
potential as the promising candidates for ultrahigh den-
sity recording media because of their high magnetocrystalline
Various lithographic techniques have been used to fabricate
the magnetic arrays [14], [15], however, periodic assembly of
anisotropy (FePt: , CoPt: nanoparticles system with an extended area has not been accom-
) [1]. Chemically synthesized FePt and CoPt plished. We demonstrated the application of amino-functional
nanoparticles have attracted much attention due to narrow size silane as a coupling layer for FePt nanoparticles dispersion,
distribution as well as the dimensions with a few nanometer resulting in the formation of a monolayered film on a rigid
scale compared to those fabricated by the existing sputtering substrate [16]. This fabrication method offered the possibility
technique [2]–[9]. As-grown FePt and CoPt nanoparticles are of a long-range array of the nanoparticles, however, no periodic
mostly face-centered cubic (fcc) and show superparamagnetic array was formed. We have recently proposed the template as-
sisted self-assembly technology, which incorporated scanning
Manuscript received August 20, 2004.
probe lithography as a promising tool for the establishment of
Y. Sasaki, M. Mizuno, A. C. C. Yu, and T. Miyauchi are with Sony Cor- such arrays [17]. In the second part of this paper, successful
poration Sendai Technology Center, Tagajo, Miyagi 985-0842, Japan (e-mail: production of well-aligned FePt nanoparticles array is demon-
Yuichi.Sasaki@jp.sony.com).
D. Hasegawa, T. Ogawa, and M. Takahashi are with Graduate School of En-
strated.
gineering, Tohoku University, Sendai, Miyagi 980-8579, Japan.
B. Jeyadevan and K. Tohji are with Graduate School of Environmental
Studies, Tohoku University, Sendai, Miyagi 980-8579, Japan. II. EXPERIMENT
K. Sato and S. Hisano are with Dowa Mining Corporation, Honjo, Saitama
367-0002, Japan. FePt nanoparticles with structure were synthesized by
Digital Object Identifier 10.1109/TMAG.2004.838040 co-reduction of Fe(III) and Pt(II) acetylacetonates in tetraethy-
0018-9464/$20.00 © 2005 IEEE
SASAKI et al.: FePt NANOPARTICLES AND NANOPARTICLE ARRAYS VIA TEMPLATE-ASSISTED SELF-ASSEMBLY 661

Fig. 1. Powder XRD pattern of FePt nanoparticles agglomerates synthesized


by modified polyol process at 300 C.

lene glycol (TEG) at 300 C [9]. Contrary to the hot-soap


method, the above method did not involve the stabilization of
particles by surfactants. Thus the nanoparticles formed agglom-
erates and precipitated out. Dispersion of the nanoparticles in
a nonpolar organic solvent was then carried out as follows:
Oleic acid and oleylamine were injected into the solution after
cooling down to room temperature, and a designated amount
of n-hexane was added. The mixture was transferred into an
airtight container, and agitated vigorously using a shaker. After
several hours, the suspension separated into TEG and n-hexane
phases. The n-hexane phase turned from colorless to dark
brown, which is an indication of the presence of nanoparticles.
Meanwhile, part of the nanoparticles remained at the interface
between the two solvents. Thus this process was repeated to
bring more particles into the n-hexane phase.
The FePt nanoparticles monolayered films were formed on a
thermally oxidized Si wafer using a coupling layer of amino-
silane [16]. Then, the monolayered film was annealed under a
vacuum of Torr at 600 C for 30 min to evaluate the crys-
tallographic features and magnetic properties.
The crystallographic textures and average grain sizes of
the FePt nanoparticle monolayered films were investigated
by in-plane X-ray diffraction (XRD) measurements with the
X-ray incidence angle set at 0.25 from the plane. The in-plane
XRD measurements were done using a RIGAKU ATX-G with
Cu radiation. On the other
hand, the as-prepared and agglomerated FePt nanoperticles Fig. 2. HR-TEM images of the (a) dispersed portion of FePt nanoparticles
were examined using powder XRD. The particle morphology synthesized by modified polyol process and (b) those prepared by the hot soap
method.
evaluation was carried out using high-resolution transmis-
sion electron microscopy (HR-TEM). Chemical compositions
were determined by X-ray microanalysis (XMA). Magnetic of (001) and (110) superlattice reflections suggested the exis-
properties were measured using a superconducting quantum tence of structure in the as-prepared FePt nanoparticles.
interference device and high-sensitivity torque magnetometer. The ratio calculated using the (111) and (200) peaks was
0.972 and was slightly larger compared to that of the ideal bulk
III. MAGNETIC PROPERTIES AND CRYSTALLOGRAPHIC value which is only 0.964 [18]. This suggested the presence of
STRUCTURES OF DIRECTLY SYNTHESIZED -TYPE partially ordered phase. As previously reported [12], the
FePt NANOPARTICLES sample consisted of agglomerates of nanoparticles with diame-
ters ranging between 5 and 10 nm.
A. Morphology and Crystallographic Structures The HR-TEM image of the FePt nanoparticles dispersed in
Fig. 1 shows the powder XRD pattern of the as-prepared n-hexane is shown in Fig. 2(a). Mismatched lattice planes ob-
FePt nanoparticles before dispersion. Their average chemical served in almost all the particles confirmed the polycrystalline
composition analyzed using XMA was . The presence nature of the particles. This morphology was characteristic
662 IEEE TRANSACTIONS ON MAGNETICS, VOL. 41, NO. 2, FEBRUARY 2005

Fig. 4. In-plane hysteresis loops for the 600 C annealed film of the dispersed
FePt nanoparticles synthesized by modified polyol process measured at RT and
5 K.

reaction kinetics of the two preparation techniques, this phe-


nomenon could be due to the following additional reasons, that
is, the difference in morphological features of the nanoparticles
prepared by both procedures and size dependence of ordering.
As noted above, FePt nanoparticles synthesized by the mod-
Fig. 3. In-plane XRD patterns for the as-coated and 600 C annealed FePt
nanoparticle monolayered films. (a) and (b) denote the dispersed portion of the ified polyol process were polycrystalline, while those by the
nanoparticles synthesized by modified polyol process and those prepared by hot-soap method were single crystalline. With a polycrystalline
hot-soap method, respectively. structure, ordering progresses due to interdiffusion of atoms
within grains and at grain boundaries. The latter is expected
of the FePt nanoparticles synthesized by the modified polyol to possess lower activation energy barrier, contributing to the
process (polycrystalline), and was different from that obtained reduction of the phase transformation temperature. Another
by the hot-soap method (single crystalline) [13] as indicated in possible reason is the dependence of phase transformation on
Fig. 2(b). The mean diameter of the particles was 7.6 nm with a the particle size. Recently, Takahashi et al. reported that smaller
standard deviation of 1.4 nm (18%). From XMA, the chemical FePt particles were harder to chemically transform by applying
composition of the dispersed FePt nanoparticles was found to the diffuse-interface theory [19]. The average grain diameter
be . The reason for this phenomenon is not clear yet, of the as-synthesized nanoparticles prepared by the hot-soap
however, dissolution by oleic acid could be considered as a method, which was calculated by the Scherrer’s equation using
possible reason. the full width at half maximum of the (111) peak, was 3.4 nm,
Fig. 3(a) shows the in-plane XRD patterns for the as-coated which was slightly smaller than that by the modified polyol
and 600 C-annealed FePt monolayered films. The nanopar- process of 4.1 nm. The former kept constant after annealing at
ticles used here correspond to the dispersed portion of the 600 C, while the latter increased up to 5.6 nm. Consequently,
particles synthesized by the modified polyol process. The the ordering in nanoparticles prepared by the hot-soap method
profiles for the monolayered films of the FePt nanoparticles seems to be more difficult to progress. The grain diameter of
prepared by the hot-soap method are shown in Fig. 3(b) for the 600 C-annealed nanoparticles was smaller than the average
comparison. Before annealing, only the chemically disordered particle diameter measured by TEM observation, suggesting
fcc (111) and (200) peaks were observed for the film of the that the interparticle coalescence did not occur.
nanoparticles synthesized by the modified polyol process.
This meant that the as-coated nanoparticles did not possess B. Magnetic Properties
the structure. The reason for this structural difference The as-synthesized FePt nanoparticle agglomerates showed
between the as-synthesized particles and the dispersed particles ferromagnetic behavior with a coercivity of 2.2 kOe at
is unclear. A chemically ordered (110) peak appeared for room temperature (RT). On the other hand, the as-coated film
the 600 C-annealed film of the nanoparticles produced by the of the dispersed nanoparticles showed superparamagnetic be-
modified polyol process and had a ratio of 0.986 compared havior at RT, which was consistent with the crystallographic fcc
with 0.964 for the bulk [18]. Therefore, the sample has to be structure with low magnetic anisotropy. Fig. 4 shows in-plane
annealed at higher temperatures than 600 C for more complete hysteresis loops of the 600 C-annealed film measured at RT and
ordering. Conversely, no superlattice peak was confirmed 5 K. It exhibited small value of 290 Oe at RT, and 4.6 kOe
in the XRD pattern for the film prepared using the particles at 5 K, which arises from the complicated contribution of in-
synthesized by the hot-soap method even after annealing at complete ordering, slightly Pt-rich composition and small grain
600 C. From these results, we concluded that the structural size of around 5 nm. The rotational hysteresis loss for
transformation of the FePt nanoparticles synthesized by the the 600 C-annealed film of FePt nanoparticles synthesized by
modified polyol process occurred at lower temperatures than the modified polyol process is plotted as a function of the ap-
those by the hot-soap method. Besides the differences in the plied field up to 25 kOe in Fig. 5. As a reference, curve
SASAKI et al.: FePt NANOPARTICLES AND NANOPARTICLE ARRAYS VIA TEMPLATE-ASSISTED SELF-ASSEMBLY 663

Fig. 6. Schematic diagram for electrochemical patterning of the periodic small


0
dot arays on a CH , terminated substrate.

Fig. 5. Rotational hysteresis loss curves for FePt nanoparticles prepared by


the modified polyol process measured at RT. Open and closed circles denote the
as-synthesized agglomerates and the dispersed nanoparticle monolayered film
afler annealing at 600 C, respectively.

for the as-synthesized FePt nanoparticle agglomerates by the


modified polyol process is shown. The curve for the 600 C-an- 0
Fig. 7. LFM images of (a) CH , terminated (before patterned) and
nealed film of FePt nanoparticles synthesized by the modified (b) electrochemically patterned surfaces on a Si substrate.
polyol process had two maximum fields of about 7 and
15 kOe, leading to two different uniaxial magnetic anisotropy surface was electrochemically modified using a conducting
fields of at least 14 and 30, respectively, according to cantilever used in scanning probe microscopes. Fig. 6 shows
classical single domain theory. However, these estimated the schematic diagram for patterning periodic small dot arrays
values were larger than double of the at 5 K with minimal on the substrate. The conducting cantilever was scanned on the
thermal effects. Namely, of the 600 C-annealed film substrate with the contact mode, applying a positive pulse bias
indicated smaller value compared to the theoretical one of non- of 9.5 V with frequency of 20 Hz and 10 ms per pulse. The
interacting, magnetically isolated Stoner–Wohlfarth-type grains scanning rate was 2 m/s. Finally, the modified substrate was
of 0.479 (for random orientation) [20]. We consider this reason immersed in n-hexane based FePt nanoparticle dispersion, then
to be mainly due to the distribution of of nanoparticles. The taken out, and rinsed by n-hexane.
magnetization reversal which is characteristic of the polycrys- The patterned surface was observed using a scanning probe
talline particles could not be discussed here because of the exis- microscope with two different modes, namely, atomic force mi-
tence of this distribution of . Interaction between and within croscope (AFM) and lateral force microscope (LFM). The ob-
the particles seems not to be dominant in this case because the servation of the resulting nanoparticle array was carried out
interparticle coalescence was not likely to occur after annealing using a field emission scanning electron microscope (FE-SEM).
and the particles were separated enough. Nevertheless detailed
investigation based on some magnetic measurements should be B. Analysis of Patterned Surface and Nanoparticle Array
made for further understanding. On the other hand, a single peak Fig. 7 indicates LFM images of (a) CH terminated (before
was observed for the film composed of as-synthesized nanopar- patterned) and (b) electrochemically patterned surfaces on a Si
ticles by the modified polyol process. It had extremely small substrate. The difference in frictional force is reflected in the
value of 200 Oe compared to the estimated of nearly 30 kOe, contrast as shown in the figure. There was almost no contrast
which is thought to be due to the distribution in the switching in the image before patterning. Conversely, periodic dark spots,
field caused by that in the degree of ordering and the exchange which corresponded to the periodicity of the modification condi-
coupling between the nanoparticles. The estimated one is con- tion, appeared in the image after patterning. Therefore, we rec-
sidered to be the maximum value in the film. ognize these spots as the modified areas. The average diameter
of the spots was about 70 nm and they were well aligned with a
IV. FePt NANOPARTICLE ARRAYS VIA TEMPLATE-ASSISTED separation of about 80 nm between centers. Each dark area had
SELF-ASSEMBLY relatively higher friction, which meant that CH terminated
groups were changed into some polar ones by the electrochem-
A. Fabrication Method of Nanoparticle Array ical reaction. Fig. 8 shows the AFM images of (a) CH ter-
The periodic array of FePt nanoparticles was fabricated using minated (before patterned) and (b) electrochemically patterned
the template-assisted self-assembly technique, which involved surfaces. Topographic change after patterning was not observed
scanning probe lithography. The procedure was composed of from the AFM results. Thus, the electrochemical patterning was
the following three steps. First, the monolayered octadecyl- carried out without any change in roughness of the substrate.
trichlorosilane (OTS) was formed on a P-type silicon wafer by The FE-SEM image of the FePt nanoparicle arrays coated on
dipping it in the solution, which allowed the termination of the the patterned substrate is shown in Fig. 9. It could be found that
methyl group on the substrate. Second, the CH terminated FePt nanoparticles were fixed on the patterned areas selectively.
664 IEEE TRANSACTIONS ON MAGNETICS, VOL. 41, NO. 2, FEBRUARY 2005

The periodic array of FePt nanoparticles was fabricated using


the template-assisted self-assembly technique. For making peri-
odic dots on a substrate, a positive-biased pulse voltage was ap-
plied to the CH terminated surface by a conducting cantilever
of a scanning probe microscope, which induced electrochem-
ical modification into polar groups. The resulting template had
well-aligned sub-100-nm dot arrays with sub-100-nm period-
icity. We also succeeded in selectively fixing the FePt nanopar-
0
Fig. 8. AFM images of (a) CH terminated (before patterned) and (b) ticles on the patterned areas.
electrochemically patterned surfaces on a Si substrate.
REFERENCES
[1] D. Weller and A. Moser, “Thermal effect limits in ultrahigh-density mag-
netic recording,” IEEE Trans. Magn., vol. 35, no. 6, pp. 4423–4439, Nov.
1999.
[2] S. Sun, C. B. Murray, D. Weller, L. Folks, and A. Moser, “Monodisperse
FePt nanoparticles and ferromagnetic FePt nanocrystal superlattices,”
Science, vol. 287, pp. 1989–1992, 2000.
[3] S. Sun, E. E. Fullerton, D. Weller, and C. B. Murray, “Compositionally
controlled FePt nanoparticle materials,” IEEE Trans. Magn., vol. 37, no.
4, pp. 1239–1243, Jul. 2001.
[4] A. C. C. Yu, M. Mizuno, Y. Sasaki, H. Kondo, and K. Hiraga, “Structural
characteristics and magnetic properties of chemically synthesized CoPt
nanoparticles,” Appl. Phys. Lett., vol. 81, pp. 3768–3770, 2002.
[5] S. Sun, S. Anders, T. Thomson, J. E. E. Baglin, M. F. Toney, H. F.
Hamann, C. B. Murray, and B. D. Terris, “Controlled synthesis and
assembly of FePt nanoparticles,” J. Phys. Chem. B, vol. 107, pp.
5419–5425, 2003.
[6] M. Chen and D. E. Nikles, “Synthesis, self-assembly, and magnetic
properties of Fe Co Pt nanoparticles,” Nano Lett., vol. 2, pp.
211–214, 2002.
Fig. 9. FE-SEM image of the FePt nanoparticle array coated on the patterned [7] C. Chen, O. Kitakami, S. Okamoto, and Y. Shimoda, “Large coercivity
substrate. White dots correspond to the FePt nanoparticles. and granular structure of CoPt=SiO films,” IEEE Trans. Magn., vol.
35, no. 5, pp. 3466–3468, Sep. 1999.
[8] M. Yu, Y. Liu, A. Moser, D. Weller, and D. J. Sellmyer, “Nanocomposite
The modified polar sites are thought to work as templates for an- CoPt:C films for extremely high-density recording,” Appl. Phys. Lett.,
vol. 20, pp. 3992–3994, 1999.
choring the nanoparticles effectively. The number of nanoparti- [9] M. Watanabe, M. Masumoto, D. H. Ping, and K. Hono, “Microstrucutre
cles per dot was around 33 with a standard deviation of 4.8 and and magnetic properties of FePt-Al-O granular thin films,” Appl. Phys.
the dot density corresponded to 0.1 Tparticles/in . Ideally, this Lett., vol. 76, pp. 3971–3973, 2000.
[10] Z. R. Dal, S. Sun, and Z. L. Wang, “Phase transformation, coalescence,
should be decreased to one particle per dot when the application and twinning of monodisperse FePt nanocrystals,” Nano Lett., vol. 1, pp.
to recording media is taken into consideration. The possibility of 443–446, 2001.
the same is being investigated now. Thus, the template-assisted [11] S. Wang, S. S. Kang, D. E. Nikles, J. W. Harrell, and X. W. Wu, “Mag-
netic properties of self-organized L1 FePtAg nanoparticle arrays,” J.
self-assembly technique to fabricate the periodic nanoparticle Magn. Magn. Mater., vol. 266, pp. 49–56, 2003.
array was successfully demonstrated and can be the promising [12] B. Jeyadevan, K. Urakawa, A. Hobo, N. Chinnasamy, K. Shinoda, K.
tool for developing future ultrahigh-density recording media. Tohji, D. J. Djayaprawira, M. Tsunoda, and M. Takahashi, “Direct syn-
thesis of fct-nanoparticles by chemical route,” Jpn. J. Appl. Phys., vol.
42, pp. L350–L352, 2003.
V. CONCLUSION [13] Y. Sasaki, M. Mizuno, A. C. C. Yu, M. Inoue, K. Yazawa, I. Ohta, M.
Takahashi, B. Jeyadevan, and K. Tohji, “Crystallographic structures and
Chemically ordered -type FePt nanoparticles were syn- magnetic properties of L1 -type FePt nanoparticle monolayered films
stabilized on functionalized surfaces,” J. Magn. Magn. Mater., vol. 282,
thesized directly without surfactants in the solution phase pp. 122–126, 2004.
through the co-reduction of Fe(III) and Pt(II) acetylacetonates [14] C. A. Ross, “Patterned magnetic recording media,” Annu. Rev. Mater.
at 300 C. These nanoparticles could be dispersed in a nonpolar Res., vol. 31, pp. 203–235, 2001.
[15] S. Anders, S. Sun, C. B. Murray, C. T. Rettner, M. E. Best, T. Thomson,
solvent such as n-hexane by coating the particles with surfac- M. Albrecht, J.-U. Thiele, E. E. Fullerton, and B. D. Terris, “Lithography
tants with long alkyl chains. However, the dispersed particles and self-assembly for nanometer scale magnetism,” Microelectron. Eng.,
were of only chemically disordered fcc phase and exhibited vol. 61–62, pp. 569–575, 2002.
[16] A. C. C. Yu, M. Mizuno, Y. Sasaki, M. Inoue, H. Kondo, I. Ohta, D.
superparamagnetic behavior. This may be due to the deficiency Djayaprawira, and M. Takahashi, “Fabrication of monodispersive FePt
of the Fe content caused by the dispersion treatment. An alter- nanoparticle films stabilized on rigid substrates,” Appl. Phys. Lett., vol.
native dispersion technique is being developed. The dispersed 82, pp. 4532–4534, 2003.
[17] M. Mizuno, Y. Sasaki, and M. Inoue, “Fabrication of FePt nanoparticle
FePt nanoparticle film stabilized using amino-silane began arrays via template-assisted self-assembly,” in Extended Abst. 51th Spr.
to structurally transform to the ordered phase at 600 C. Meeting Jpn. Soc. Appl. Phys., 2004, p. 194.
[18] JCPDS-International Centre for Diffraction Data, 1999.
This phenomenon could be explained by the characteristic [19] Y. K. Takahashi, T. Koyama, M. Ohnuma, T. Ohkubo, and K. Hono,
polycrystalline nature and the relatively large crystal size of the “Size dependence of ordering in FePt nanoparticles,” J. Appl. Phys., vol.
particles synthesized by the modified polyol process. Rotational 95, no. 5, pp. 2690–2696, 2004.
[20] E. C. Stoner and E. P. Wohlfarth, “A mechanism of magnetic hys-
hysteresis loss measurement suggested that the nanoparticle teresis in heterogeneous alloys,” Trans. R. Soc. London, vol. A240, pp.
film annealed at 600 C was not completely ordered. 599–642, 1948.