9 Magnetism of Individual Nanoparticles Probed by X-Ray …
233
is not sufficient to induce magnetically blocked states at room temperature [59, 60].
Instead, the interface contribution leads to superparamagnetism with preferred inplane magnetization as discussed in [59]. Also, in case of iron nanoparticles in
contact with NiO(110), one could expect to find magnetically blocked states due to
a strong exchange interaction with the antiferromagnetic substrate as observed in
case of iron films grown on NiO [62, 63]. However, the absence of magnetically
blocked states in iron nanoparticles on NiO(110) suggests a reduced interaction via
the finite or imperfect contact interface when compared to the thin film systems.
Imperfect interfaces are indeed likely for iron nanoparticles on all four investigated
substrates, since RHEED experiments show that nanoparticles in the present size
range are randomly oriented upon deposition on different substrates and under sample
preparation conditions similar to the XPEEM works discussed here [5, 64]. For
instance, in single crystalline fcc systems, the deposition of nanoparticles can lead
to the formation of defects at the interface such as twin boundaries between epitaxial
and non-epitaxial parts of the nanoparticles [65].
While for the iron nanoparticles on Cu(001) no particularly strong interface
anisotropy is anticipated, enhanced magnetic energy barriers have been observed in
ensemble measurements and in XPEEM investigations [28, 45, 60]. In [28, 45] the
enhanced magnetic energy barriers of iron nanoparticles on copper substrates were
attributed to indirect, surface state mediated interactions between the nanoparticles,
which are expected to occur over short distance (few nm) and should only be present
in dense samples. In the XPEEM experiments, the nanoparticles are separated by a
few hundred nm and short range interactions can be excluded. Thus, the XPEEM
data demonstrate that short range interactions are not necessary to establish enhanced
magnetic energy barriers in iron nanoparticles on Cu(001), but suggest instead that
they are intrinsic to the nanoparticles. Likewise, for the iron nanoparticles on the
silicon substrates, no strong interface anisotropy or substrate induced magnetic interactions are expected, but magnetically blocked nanoparticles are observed at room
temperature. Also this observation hints at an intrinsic rather then interface-induced
origin of the enhanced magnetic energy barriers of the magnetically blocked nanoparticles. This conclusion is further supported by the fact that the magnetic moments
of the magnetically blocked nanoparticles are randomly oriented on the silicon substrates [38], while a dominant interface-induced magnetic anisotropy would prefer
either in-plane or out-of-plane oriented magnetic moments.
These considerations show that the magnetically blocked states of iron nanoparticles are not induced by interface effects, such as magnetic interface anisotropy,
exchange interaction or surface-mediated magnetic interactions. Further, since the
magnetically blocked states of iron nanoparticles can spontaneously relax towards
a superparamagnetic state, as found in case of the silicon substrates, one is led to
conclude that the enhanced magnetic energy barriers are an intrinsic, but metastable
property of iron nanoparticles, possibly originating from the nanoparticle growth
process [31]. This enhanced magnetic anisotropy is then preserved upon deposition
on silicon wafers and on the Cu(001) single crystal surface, but lost upon deposition on NiO(110) and W(110). Such behavior could be assigned to the free surface energy E s of the investigated substrates: SiO x (E s ∼ 0.2 J/m
2 ) and Cu(001)
233
is not sufficient to induce magnetically blocked states at room temperature [59, 60].
Instead, the interface contribution leads to superparamagnetism with preferred inplane magnetization as discussed in [59]. Also, in case of iron nanoparticles in
contact with NiO(110), one could expect to find magnetically blocked states due to
a strong exchange interaction with the antiferromagnetic substrate as observed in
case of iron films grown on NiO [62, 63]. However, the absence of magnetically
blocked states in iron nanoparticles on NiO(110) suggests a reduced interaction via
the finite or imperfect contact interface when compared to the thin film systems.
Imperfect interfaces are indeed likely for iron nanoparticles on all four investigated
substrates, since RHEED experiments show that nanoparticles in the present size
range are randomly oriented upon deposition on different substrates and under sample
preparation conditions similar to the XPEEM works discussed here [5, 64]. For
instance, in single crystalline fcc systems, the deposition of nanoparticles can lead
to the formation of defects at the interface such as twin boundaries between epitaxial
and non-epitaxial parts of the nanoparticles [65].
While for the iron nanoparticles on Cu(001) no particularly strong interface
anisotropy is anticipated, enhanced magnetic energy barriers have been observed in
ensemble measurements and in XPEEM investigations [28, 45, 60]. In [28, 45] the
enhanced magnetic energy barriers of iron nanoparticles on copper substrates were
attributed to indirect, surface state mediated interactions between the nanoparticles,
which are expected to occur over short distance (few nm) and should only be present
in dense samples. In the XPEEM experiments, the nanoparticles are separated by a
few hundred nm and short range interactions can be excluded. Thus, the XPEEM
data demonstrate that short range interactions are not necessary to establish enhanced
magnetic energy barriers in iron nanoparticles on Cu(001), but suggest instead that
they are intrinsic to the nanoparticles. Likewise, for the iron nanoparticles on the
silicon substrates, no strong interface anisotropy or substrate induced magnetic interactions are expected, but magnetically blocked nanoparticles are observed at room
temperature. Also this observation hints at an intrinsic rather then interface-induced
origin of the enhanced magnetic energy barriers of the magnetically blocked nanoparticles. This conclusion is further supported by the fact that the magnetic moments
of the magnetically blocked nanoparticles are randomly oriented on the silicon substrates [38], while a dominant interface-induced magnetic anisotropy would prefer
either in-plane or out-of-plane oriented magnetic moments.
These considerations show that the magnetically blocked states of iron nanoparticles are not induced by interface effects, such as magnetic interface anisotropy,
exchange interaction or surface-mediated magnetic interactions. Further, since the
magnetically blocked states of iron nanoparticles can spontaneously relax towards
a superparamagnetic state, as found in case of the silicon substrates, one is led to
conclude that the enhanced magnetic energy barriers are an intrinsic, but metastable
property of iron nanoparticles, possibly originating from the nanoparticle growth
process [31]. This enhanced magnetic anisotropy is then preserved upon deposition
on silicon wafers and on the Cu(001) single crystal surface, but lost upon deposition on NiO(110) and W(110). Such behavior could be assigned to the free surface energy E s of the investigated substrates: SiO x (E s ∼ 0.2 J/m
2 ) and Cu(001)