226
A. Kleibert
ration and characterization techniques were needed until the first systematic and
detailed investigations of mono-disperse iron nanoparticle samples became possible. However, despite considerable achievements, an unambiguous understanding of
the magnetic properties of iron nanoparticles has still not been achieved. For instance,
extensive size- and temperature-dependent investigations of iron nanoparticles were
reported by Gangopadhyay et al. [23]. By combining TEM, with integral SQUID and
Mössbauer spectroscopy, the authors could explain the frequent reports of a reduced
magnetization in iron nanoparticles when compared to bulk iron by the formation
of a thin native oxide shell around a metallic iron core with bulk-like magnetic
moments. The authors also noticed a significantly enhanced magnetic anisotropy in
the iron nanoparticles when compared to the magneto-crystalline anisotropy of bulk
iron, which they assigned to the core–shell-type particle morphology and a strong
interaction between the oxide shell with the iron core in addition to surface effects.
Indeed, a subsequent Mössbauer spectroscopy study of metallic iron nanoparticles
without oxide shell, which were prepared by reduction of ferric nitrate at 600 K
in hydrogen atmosphere, revealed mostly bulk-like properties including a smaller,
size-dependent surface contribution to the magnetic anisotropy which the authors
assigned to smaller deviations from spherical symmetry [24]. However, high magnetic anisotropies were later found in another work on metallic iron nanoparticles
deposited on copper surfaces under UHV conditions [28, 45]. Uniaxial and large
anisotropy was also found by means of vibrating sample magnetometry of silver- or
SiO 2 -capped iron nanoparticles [25]. Similarly, microSQUID investigations of individual iron nanoparticles embedded in a niobium matrix revealed a dominant uniaxial
anisotropy and a reduction of the magnetic moment due to the interaction with the
matrix. The uniaxial anisotropy was assigned to surface- and shape anisotropy contributions originating from shape deviation from a perfect truncated dodecahedron
[27]. Bulk-like magnetic anisotropy was later found in wet-chemically prepared iron
nanocubes in [29]. As a last example, we may mention iron nanoparticles prepared by
a high pressure magnetron sputtering technique, which revealed enhanced magnetic
anisotropies with decreasing size [30].
These examples illustrate the large scatter of reported properties, and in particular
the difficulty to disentangle the impact of the sample preparation technique, chemical state and the matrix material on the magnetic anisotropy of iron nanoparticles
from ensemble measurements. In the following we will show that, by combining in
situ XPEEM with SEM and AFM investigations, it is possible to directly correlate
the magnetic properties of a large number of individual nanoparticles in extended
ensembles with their actual size, shape, and orientation. Such approach overcomes
the major uncertainties in averaging ensemble measurements, namely the impact
of the inherent distribution of particle sizes, shapes, orientations, and the possible
contribution of clusters of interacting nanoparticles to the measured mean values,
and can therefore resolve previously controversial observations. Moreover, in situ
XPEEM allows one to study the pristine state of the nanoparticles on various supports
without the additional influence of organic or oxide shells or a matrix material.
A. Kleibert
ration and characterization techniques were needed until the first systematic and
detailed investigations of mono-disperse iron nanoparticle samples became possible. However, despite considerable achievements, an unambiguous understanding of
the magnetic properties of iron nanoparticles has still not been achieved. For instance,
extensive size- and temperature-dependent investigations of iron nanoparticles were
reported by Gangopadhyay et al. [23]. By combining TEM, with integral SQUID and
Mössbauer spectroscopy, the authors could explain the frequent reports of a reduced
magnetization in iron nanoparticles when compared to bulk iron by the formation
of a thin native oxide shell around a metallic iron core with bulk-like magnetic
moments. The authors also noticed a significantly enhanced magnetic anisotropy in
the iron nanoparticles when compared to the magneto-crystalline anisotropy of bulk
iron, which they assigned to the core–shell-type particle morphology and a strong
interaction between the oxide shell with the iron core in addition to surface effects.
Indeed, a subsequent Mössbauer spectroscopy study of metallic iron nanoparticles
without oxide shell, which were prepared by reduction of ferric nitrate at 600 K
in hydrogen atmosphere, revealed mostly bulk-like properties including a smaller,
size-dependent surface contribution to the magnetic anisotropy which the authors
assigned to smaller deviations from spherical symmetry [24]. However, high magnetic anisotropies were later found in another work on metallic iron nanoparticles
deposited on copper surfaces under UHV conditions [28, 45]. Uniaxial and large
anisotropy was also found by means of vibrating sample magnetometry of silver- or
SiO 2 -capped iron nanoparticles [25]. Similarly, microSQUID investigations of individual iron nanoparticles embedded in a niobium matrix revealed a dominant uniaxial
anisotropy and a reduction of the magnetic moment due to the interaction with the
matrix. The uniaxial anisotropy was assigned to surface- and shape anisotropy contributions originating from shape deviation from a perfect truncated dodecahedron
[27]. Bulk-like magnetic anisotropy was later found in wet-chemically prepared iron
nanocubes in [29]. As a last example, we may mention iron nanoparticles prepared by
a high pressure magnetron sputtering technique, which revealed enhanced magnetic
anisotropies with decreasing size [30].
These examples illustrate the large scatter of reported properties, and in particular
the difficulty to disentangle the impact of the sample preparation technique, chemical state and the matrix material on the magnetic anisotropy of iron nanoparticles
from ensemble measurements. In the following we will show that, by combining in
situ XPEEM with SEM and AFM investigations, it is possible to directly correlate
the magnetic properties of a large number of individual nanoparticles in extended
ensembles with their actual size, shape, and orientation. Such approach overcomes
the major uncertainties in averaging ensemble measurements, namely the impact
of the inherent distribution of particle sizes, shapes, orientations, and the possible
contribution of clusters of interacting nanoparticles to the measured mean values,
and can therefore resolve previously controversial observations. Moreover, in situ
XPEEM allows one to study the pristine state of the nanoparticles on various supports
without the additional influence of organic or oxide shells or a matrix material.