Content uploaded by Michael Farle
Author content
All content in this area was uploaded by Michael Farle on Dec 05, 2019
Content may be subject to copyright.
Content uploaded by Kai Fauth
Author content
All content in this area was uploaded by Kai Fauth
Content may be subject to copyright.
Enhanced Orbital Magnetism in Fe50 Pt50 Nanoparticles
C. Antoniak, J. Lindner, M. Spasova, D. Sudfeld, M. Acet, and M. Farle
Experimentalphysik —AG Farle, Fachbereich Physik, Universita
¨t Duisburg-Essen, Lotharstrasse 1, 47048 Duisburg, Germany
K. Fauth
Experimentelle Physik IV, Universita
¨tWu
¨rzburg, Am Hubland, 97074 Wu
¨rzburg, Germany,
and MPI fu
¨r Metallforschung, Heisenbergstrasse 3, 70569 Stuttgart, Germany
U. Wiedwald, H.-G. Boyen, and P. Ziemann
Abteilung Festko
¨rperphysik, Universita
¨t Ulm, Albert-Einstein-Allee 11, 89069 Ulm, Germany
F. Wilhelm and A. Rogalev
European Synchrotron Radiation Facility (ESRF), 6 Rue Jules Horowitz, BP 220, 38043 Grenoble Cedex, France
Shouheng Sun
Department of Chemistry, Brown University, Providence, Rhode Island 02912, USA
(Received 28 September 2005; published 13 September 2006)
X-ray absorption and magnetic circular dichroism spectra at both the Fe and Pt L3;2edges were
measured on wet-chemically synthesized monodisperse Fe50 Pt50 particles with a mean diameter of 6.3 nm
before and after complete removal of the organic ligands and the oxide shell covering the particles by soft
hydrogen plasma resulting in a pure metallic state. After thermal treatment of the metallic particles, the
coercive field increased by a factor of 6, the orbital magnetic moment at the Fe site increased by 330% and
is reduced at the Pt site by 30%, while the effective spin moments did not change. A decrease of the
frequency of oscillations in the extended x-ray absorption fine structure at the Pt L3;2edges provides
evidence for crystallographic changes towards the L10phase.
DOI: 10.1103/PhysRevLett.97.117201 PACS numbers: 75.75.+a, 61.10.Ht, 61.46.w, 78.70.Dm
The magnetic properties of nanoparticles with diameters
<7nmmay differ from those of the corresponding bulk
materials due to finite size effects, different crystal struc-
tures, and large surface contributions. Microscopic quan-
tities like the orbital and spin magnetic moments are
expected to reflect changes in crystal and electronic struc-
ture in a very sensitive way. In the case of bimetallic alloys
like FePt, also the differences in composition at the surface
and in the volume may lead to characteristic changes of the
magnetic moments in Fe and Pt in a nanoparticle. From the
technological perspective, FePt has become one of the
most intensely researched nanostructured materials (see,
e.g., [1–6]), since its large magnetic anisotropy of Keff
6106J=m3[7,8] and associated large coercivity makes
it the prime candidate for new ultrahigh density magnetic
storage media or biomedical applications [6]. Organo-
metallic synthesis of FePt nanoparticles with subsequent
heat treatment has been tried as an inexpensive route to
obtain L10ordered FePt particles with diameters around
4nm[7]. Later analysis, however, has unambiguously
shown that the thermal treatment led to agglomerated large
crystals [9]. Here, we focus on the determination of the Fe
and Pt magnetic moments in nonagglomerated wet-
chemically synthesized FePt particles. The changes from
the as prepared ligand covered state, over the pure metallic
chemically disordered fcc state without a ligand shell, to a
partially chemically ordered fct state after annealing are
examined. At all stages the size distribution of the particles
was monitored by high-resolution scanning electron mi-
croscopy (SEM). Specifically, we demonstrate the follow-
ing: (i) pure metallic 6 nm Fe50 Pt50 particles self-
assembled on a Si wafer can be derived from the wet-
chemical synthesis after a soft hydrogen plasma treatment;
(ii) subsequent annealing at 600 Cyields structurally
modified nanoparticles and a dramatic enhancement
(330%) of the orbital magnetic moment lat the Fe site
while lat the Pt site is slightly reduced and the effective
spin magnetic moments remain unchanged after annealing;
(iii) the structural transformation after annealing is evi-
denced by the lower frequency of the extended x-ray
absorption fine structure (EXAFS) oscillations. In addition,
the increase of the coercive field after thermal treatment
indicates the partial formation of the chemically ordered
L10state.
The as prepared sample consists of monodisperse nano-
particles with a mean diameter of d6:3nmand a stan-
dard deviation of 0:9nmof a chemically disordered
Fe50Pt50 alloy surrounded by a ligand shell of oleyl amine
and oleic acid. The particles, suspended in hexane, were
deposited by spin coating onto a naturally oxidized Si
substrate to form a monolayer. SEM images show a cover-
age of 10% on a 510 mm2area and the formation of
PRL 97, 117201 (2006) PHYSICAL REVIEW LETTERS week ending
15 SEPTEMBER 2006
0031-9007=06=97(11)=117201(4) 117201-1 ©2006 The American Physical Society
monolayer islands of about 2104nm2consisting of
hexagonally arranged particles with a mean center-to-
center distance dcc 9nm.
For the determination of the element specific magnetic
moments, x-ray absorption spectroscopy (XAS) spectra of
the identical sample of photon energy were taken at the Fe
L3;2and Pt L3;2edges in the total electron yield (TEY)
mode at the PM-3 bending magnet beam line of the
BESSY II synchrotron facility in Berlin, Germany, in fields
of 2:8Tand in the hard x-ray regime in the fluorescence
yield mode at the undulator beam line ID12 of the ESRF in
Grenoble, France, in fields of 0:6T. After each scan,
either the magnetic field or the polarization of the x rays
was flipped.
Figure 1shows the normalized XAS, the corresponding
x-ray magnetic circular dichroism (XMCD), and the inte-
grated XMCD signal at the Fe and Pt L3;2edges in the
hydrogen plasma treated and annealed states of the parti-
cles. To separate the electron excitations to unoccupied d
states from those to the continuum, a standard two-step-
like function is subtracted in the case of the spectra at
the Fe L3;2edges [10]. Since the error in performing such
a separation for the Pt L3;2edges is large due to the not
well pronounced absorption peaks, the adjusted isotropic
absorption spectra of Au were subtracted instead [11].
The XAS of the sample in the as prepared state at the
Fe L3;2edges showed that the ligand shell does not pro-
tect the surface of the particles from oxidation, confirm-
ing previous results [12,13]. After the sample was exposed
to a soft hydrogen plasma at room temperature and a
pressure of 5 Pa, pure metallic XAS at the Fe L3;2edges
was obtained [Fig. 1(a)]. At the carbon Kedge, no absorp-
tion peaks were observed, since the hydrogen plasma does
not only reduce Fe oxides but also removes the organic
ligands as is also known for reactive oxygen plasma
[14,15].
SEM images reveal that the number and the arrangement
of particles have not changed within the error bars with
respect to the as prepared state. In order to achieve the
chemically ordered phase, the plasma treated particles
have been annealed in hydrogen atmosphere of 5 Pa for
30 min at 600 C. At this temperature the disorder-order
transformation in FePt nanostructures takes place (see,
e.g., [7,16–19]). Since the hard magnetic, chemically or-
dered L10phase is known to have a large coercivity
[13,20,21], element specific magnetization curves [22]at
the Fe L3edge were measured to provide evidence for the
formation of the L10state. The normalized dichroic signal
at T15 K as a function of the external magnetic field is
shown for perpendicular and grazing incidence, i.e., for
magnetization parallel to the normal of the sample plane
(0) and 75(Fig. 2). The signals are normalized
to the white line absorption. Before thermal treatment, the
coercive field is 0Hc36 5mT, and the magnetiza-
tion is favored in plane due to weak dipolar interactions.
After annealing, the coercive field is not only enhanced,
but also found to depend slightly on the angle :
0Hc0292 8mT,0Hc75228 8mT.
Note that the sample may still not be fully saturated at
0H2:8Tfor both geometries [Fig. 2(b)]. At 300 K, a
clear hysteresis was observed with a coercive field of
0Hc35 5mT. The values of the coercivity are in
agreement to reported data for well-separated chemically
ordered FePt particles of the same size with a random
distribution of anisotropy axes [23] and are much smaller
than the one for cosputtered FePt pancakelike particle
layers for which a coercive field of 2 T at room temperature
has been reported [8]. The ratio of remanence-to-saturation
magnetization is 0.5, as predicted for randomly oriented,
noninteracting particles by the Stoner-Wohlfarth theory,
confirming that the dipolar coupling of the particles is
negligible with respect to the large magnetocrystalline
anisotropy of every single particle after annealing.
The size distribution obtained from a statistical analysis
of many SEM images confirms that the median value of the
diameter distribution does not change within the error bar
Fe L3,2 Pt L3,2
(a) (b)
(c) (d)
FIG. 1. Isotropic XAS, XMCD, and
integrated XMCD spectra at 15 K for
Fe50Pt50 nanoparticles after Hplasma
treatment (a),(b) and after annealing in
Hplasma (c),(d) measured at the Fe L3;2
edges (left) and at the Pt L3;2edges
(right). The thin lines in panels (b) and
(d) are Au reference spectra with a
shifted and stretched energy scale.
XMCD spectra at the Pt L3;2edges are
scaled up by a factor of 4.
PRL 97, 117201 (2006) PHYSICAL REVIEW LETTERS week ending
15 SEPTEMBER 2006
117201-2
and that more than 80% of the annealed sample consist of
well-separated nanoparticles.
To address the question of whether structural distortions
associated with the L10phase occur, we analyzed the
difference of the frequency of the EXAFS oscillations at
the Pt L3;2edges of the metallic particles before and after
annealing [Figs. 1(b) and 1(d)]. Using a Au reference
spectrum, we determined a distance of nearest neighbor
atoms consistent with the lattice constant of 3:85 0:02
A
of the fcc disordered particles before annealing and a
reduced one by 21%after annealing in agreement to
the corresponding decrease known in FePt bulk alloy
[13,24].
By applying the XMCD sum rules [25] it is possible to
determine the effective spin magnetic moment eff
sand the
orbital magnetic moment l.eff
ss7tconsists of
the spin magnetic moment sand the magnetic dipole
moment twhich accounts for the asphericity of the spin
moment distribution and might not cancel out especially in
the case of nanostructured materials. Since the spin-orbit
coupling in FePt is large, tcannot be eliminated by
angular dependent XMCD measurements [26], and even
for randomly oriented crystallographic axes in the nano-
particle ensemble twill not average out [27]. Hence, we
can quote effective spin moments eff
sonly.
Values of land eff
sobtained for Fe and Pt, which
include a TEY saturation correction [28,29] of 3% for eff
s
and 30% for lin the case of Fe moments, are listed in
Table I. Absolute values were calculated using the theo-
retical numbers of dholes ~
nhnPt
hnAu
hwith nPt
h
1:73,nAu
h0:75, and nFe
h3:705 [30] for both the dis-
ordered and the ordered states. Note that the different peak
intensities in the XAS reveal a small decrease (<10%)of
nFe
hafter annealing, indicating a change in the electronic
structure. The Pt moments of the annealed particles have
been obtained by extrapolation to the saturation moments,
since in the applied magnetic field of 0.6 T the XMCD is
only about 74% of the one in the magnetically saturated
state. For a better comparison, we also present in Table I
the theoretical values (b) obtained from band structure
calculations for a L10structure and the values (a) derived
from calculated XMCD spectra by application of the sum
rules [30]. The later values include t.
First, we discuss the spin moment of Fe and Pt. We
found no significant change of eff
sat both the Fe and Pt
sites before and after annealing (Table I), in agreement to
theory [31]. However, the absolute values of the spin mo-
ments at the Fe sites are about 15% smaller than the one
obtained from band structure calculations (2:87B,
[30,31]) for bulk materials. A reduction of the effective
spin magnetic moment caused by a noncollinear spin
structure at the particles surface [14] is assumed to be
unlikely due to the dominating exchange energy between
all spins in such small particles. One may note that for Fe
nanoclusters a decrease of eff
swith decreasing size was
associated with a negative contribution of t[32], which
may be present in the FePt nanoparticles as well. Another
reasonable explanation is the formation of small Fe clus-
ters of a few atoms in the nonannealed particles resulting in
a low-spin Fe state. In the case of Fe chains epitaxially
grown at the steps of Pt(997) such a low-spin state was
identified for a coverage above the monochain limit re-
flecting a strong dependence of eff
son the Fe-Fe coordi-
nation [33]. Additionally, in the case of the annealed
particles the magnetic moments measured in 2.8 T may
not be fully saturated (Fig. 2).
Now we turn to the discussion of the orbital contribu-
tions of the magnetic moments. At the Pt site, lin the
annealed particles is enhanced by about 100% compared to
the calculated one for bulk. This is most likely due to the
contribution of the increased orbital magnetic moments at
the surface [34]. On the other hand, it is reduced by 30%
with respect to the plasma treated (disordered) state. This
may indicate local composition variations within the par-
TABLE I. Element specific effective spin and orbital magnetic
moments in B. Theoretical values derived from band structure
calculations (b) and from calculated XMCD spectra (a) for the
L10structure are also listed [30].
Fe Pt
eff
sleff
sl
Plasma treatment 2.48(25) 0.056(10) 0.41(2) 0.054(6)
Annealed 2.59(26) 0.240(18) 0.41(3) 0.042(6)
Theory (a) 2.50 0.064 0.41 0.020
Theory (b) 2.87 0.072 0.33 0.046
100nm
(b)
100nm
(a)
FIG. 2. Hysteresis loops of the XMCD for Fe50 Pt50 nanopar-
ticles at 15 K measured at the Fe L3edge for perpendicular (
0, solid lines) and grazing (75, dotted lines) incidence
(a) after Hplasma treatment and (b) after 30 min annealing at
600 C. The insets show SEM images of the sample before and
after annealing.
PRL 97, 117201 (2006) PHYSICAL REVIEW LETTERS week ending
15 SEPTEMBER 2006
117201-3
ticle. One may argue that in the chemically disordered state
a local segregation of a few atoms of Pt polarized by the Fe
may effectively acquire a larger ldue to the larger spin-
orbit coupling in Pt while sremains the same. This effect
may even be pronounced if the Pt enrichment occurs at the
surface of the disordered particles. At the Fe site the orbital
magnetic moment is strongly enhanced after annealing,
namely, 330% with respect to the nonannealed state. The
same dramatic increase is found when we consider just the
ratio of orbital-to-spin magnetic moments, which is inde-
pendent of the number of unoccupied dstates. At the Fe
site the ratio increases from 2% to 9% after annealing and
is nearly identical to the ratio at the Pt site (10%) This
indicates a noncubic environment of the Fe and Pt atoms as
expected for the L10structure.
In conclusion, we determined the element specific mag-
netic moments, land eff
s, in pure metallic 6.3 nm
Fe50Pt50 particles and found a strongly enhanced value of
lat the Fe site after annealing for 30 min at 600 C.
Evidence for local composition variations within the par-
ticles based on a comparison of all contributions to the
magnetic moments is discussed. Additionally, changes of
the EXAFS oscillations and the enhanced coercivity after
thermal treatment indicate the partial formation of the L10
structure.
We thank A. Trounova, A. Schlachter, N. Friedenberger,
and S. Stienen for assistance in the measurements. This
work was supported by the BMBF (05 ES3XBA/5), the
ESRF, the DFG (SFB 445, SFB 569, Zi 21-1), and the
EU under Contract No. MRTN-CT-2004-005567
(SyntOrbMag).
[1] B. Rellinghaus, S. Stappert, M. Acet, and E. F.
Wassermann, J. Magn. Magn. Mater. 266, 142 (2003).
[2] R. W. Chantrell, D. Weller, T. J. Klemmer, S. Sun, and
E. E. Fullerton, J. Appl. Phys. 91, 6866 (2002).
[3] M. Ulmeanu, C. Antoniak, U. Wiedwald, M. Farle,
Z. Frait, and S. Sun, Phys. Rev. B 69, 054417 (2004).
[4] V. Sagueirin
˜o-Maceira, L. M. Liz-Marza
´n, and M. Farle,
Langmuir 20, 6946 (2004).
[5] C. Antoniak, J. Lindner, and M. Farle, Europhys. Lett. 70,
250 (2005).
[6] D. K. Kim, D. Kan, T. Veres, F. Normadin, J. K. Liao,
H. H. Kim, S.-H. Lee, M. Zahn, and M. Muhammed,
J. Appl. Phys. 97, 10Q918 (2005).
[7] S. Sun, C. B. Murray, D. Weller, L. Folks, and A. Moser,
Science 287, 1989 (2000).
[8] T. Shima, K. Takanashi, Y.K. Takahashi, and K. Hono,
Appl. Phys. Lett. 85, 2571 (2004).
[9] H. Zeng, S. Sun, T. S. Vedanton, J. P. Liu, Z.-R. Dai, and
Z.-L. Wang, Appl. Phys. Lett. 80, 2583 (2002).
[10] C. T. Chen, Y. U. Idzerda, H.-J. Lin, N.V. Smith, G. Meigs,
E. Chaban, G. H. Ho, E. Pellegrin, and F. Sette, Phys.
Rev. Lett. 75, 152 (1995).
[11] W. Grange, M. Maret, J.-P. Kappler, J. Vogel, A. Fontaine,
F. Petroff, G. Krill, A. Rogalev, J. Goulon, M. Finazzi, and
N. B. Brookes, Phys. Rev. B 58, 6298 (1998).
[12] S. Anders, M. F. Toney, T. Thomson, J.-U. Thiele, B. D.
Terris, S. Sun, and C. B. Murray, J. Appl. Phys. 93, 7343
(2003).
[13] C. Liu, T. J. Klemmer, N. Shukla, X. Wu, D. Weller,
M. Tanase, and D. Laughlin, J. Magn. Magn. Mater.
266, 96 (2003).
[14] H.-G. Boyen, K. Fauth, B. Stahl, P. Ziemann, G. Ka
¨stle,
F. Weigl, F. Banhart, M. Heßler, G. Schu
¨tz, N. S. Gajbhije,
J. Ellrich, H. Hahn, M. Bu
¨ttner, M. G. Garnier, and
P. Oelhafen, Adv. Mater. 17, 574 (2005).
[15] U. Wiedwald, K. Fauth, M. Heßler, H.-G. Boyen, F. Weigl,
M. Hilgendorff, M. Giersig, G. Schu
¨tz, P. Ziemann, and
M. Farle, Chem. Phys. Chem. 6, 2522 (2005).
[16] Z. Jia, S. Kang, S. Shi, D. E. Nikles, and J. W. Harrell,
J. Appl. Phys. 97, 10J310 (2005).
[17] K. E. Elkins, T. S. Vedantam, J. P. Liu, H. Zeng, S. Sun,
Y. Ding, and Z. L. Wang, Nano Lett. 3, 1647 (2003).
[18] B. Bian, K. Sato, Y. Hirotsu, and A. Makino, Appl. Phys.
Lett. 75, 3686 (1999).
[19] Y.K. Takahashi, T. Koyama, M. Ohnuma, T. Ohkubo, and
K. Hono, J. Appl. Phys. 95, 2690 (2004).
[20] J. W. Harrell, S. Wang, D. E. Nikles, and M. Chen, Appl.
Phys. Lett. 79, 4393 (2001).
[21] B. Stahl, J. Ellrich, R. Theissmann, M. Ghafari,
S. Bhattacharya, H. Hahn, N. S. Gajbhiye, D. Kramer,
R. N. Viswanath, J. Weissmu
¨ller, and H. Gleiter, Phys.
Rev. B 67, 014422 (2003).
[22] E. Goering, A. Fuss, W. Weber, J. Will, and G. Schu
¨tz,
J. Appl. Phys. 88, 5920 (2000).
[23] J.-M. Qiu, J. H. Judy, D. Weller, and J.-P. Wang, J. Appl.
Phys. 97, 10J319 (2005).
[24] M. R. Visokay and R. Sinclair, Appl. Phys. Lett. 66, 1692
(1995), and references therein.
[25] B. T. Thole, P. Carra, F. Sette, and G. van der Laan, Phys.
Rev. Lett. 68, 1943 (1992); P. Carra, B. T. Thole,
M. Altarelli, and X. Wang, Phys. Rev. Lett. 70, 694
(1993).
[26] J. Sto
¨hr and H. Ko
¨nig, Phys. Rev. Lett. 75, 3748
(1995).
[27] C. Ederer, M. Komelj, and M. Fa
¨hnle, Phys. Rev. B 68,
052402 (2003).
[28] R. Nakajima, J. Sto
¨hr, and Y. U. Idzerda, Phys. Rev. B 59,
6421 (1999).
[29] K. Fauth, Appl. Phys. Lett. 85, 3271 (2004).
[30] I. Galanakis, M. Alouani, and H. Dreysse
´, J. Magn. Magn.
Mater. 242, 27 (2002).
[31] Ya. Perlov, H. Ebert, A. N. Yaresko, V. N. Antonov, and
D. Weller, Solid State Commun. 105, 273 (1998).
[32] K. W. Edmonds, C. Binns, S. H. Baker, S. C. Thornton,
C. Norris, J. B. Goedkoop, M. Finazzi, and N. B. Brookes,
Phys. Rev. B 60, 472 (1999).
[33] A. Enders (private communication).
[34] O. S
ˇipr, M. Kos
ˇuth, and H. Ebert, Phys. Rev. B 70, 174423
(2004).
PRL 97, 117201 (2006) PHYSICAL REVIEW LETTERS week ending
15 SEPTEMBER 2006
117201-4