Chemistry Reference
In-Depth Information
medium) [
68
]. It also explains the disappearance of charge ordering by cooling and the
Verwey transition, which is not strictly correlated to confinement effects but to the
transformation of magnetite into maghemite according to an oxidation process [
69
,
70
].
In the case of other ferrites, the situation might be more complex because the
lack of resolution of the hyperfine structure yet observed on bulk systems is worth
because of the occurrence of relaxation phenomena: the single method is to apply
an external magnetic field to overcome the superparamagnetic relaxation phe-
nomena. In such a case, the Mössbauer spectrum consists in two well resolved
components allowing thus the octahedral and tetrahedral Fe proportions, the mean
effective fields of each Fe species and the mean direction of the Fe moments under
the external field to be accurately estimated [
71
-
76
]. When the ferrite nanopar-
ticles contain two cationic species, one can estimate the degree of cationic
inversion to be compared with that established from either Extended X-ray
Absorption Fine Structure (EXAFS) or X-ray magnetic circular dichroism
(XMCD) [
71
-
79
]. With more cationic species, EXAFS experiments have to be
performed at different edges in order to determine the cationic populations at each
site. The magnetization corresponding to the final composition can be thus cal-
culated and compared to that observed by magnetic measurements: the disagree-
ments, if any, have to be correlated to dynamics and confinement effects [
71
-
86
].
It is important to note the case of Zn ferrite which behaves as an antiferromagnet
below 10 K [
87
] in the microcrystalline state becomes a ferrimagnet with high
magnetic ordering temperature in the nanocrystalline state [
72
,
83
-
86
].
The profile of the in-field Mössbauer lines has to be carefully analyzed and two main
different scenarios can be distinguished. When the lines are quite narrow suggesting an
homogeneous Fe environment, the fitting model consists in describing each compo-
nent with a single effective field and a single angle between Fe moment and c-beam
direction, allowing the hyperfine field to be calculated using Eq.
4.1
. It is now possible
to simulate the zero-field Mössbauer spectrum (putting together isomer shift, quad-
rupolar shift, hyperfine field values and proportions of each species) and to compare to
the experimental one: a fair agreement would consolidate the homogeneity of the
nanoparticles while a disagreement would suggest some remaining relaxation phe-
nomena. On the contrary, components with broadened lines (suggesting different
atomic environments of Fe sites) have to be described by means of discrete distribution
of effective fields P(B
eff
)
A,B
assuming the same orientation of Fe magnetic moments
respect to the applied field: it is thus easy to evaluate the hyperfine field distributions in
order to reconstruct the zero-field Mössbauer spectrum. A disagreement with the
experimental spectrum leads to improve the ''naive'' fitting model into the introduction
of joint distribution functions P(B
eff,
h)
A,B
of the effective field and the canting angle,
which is consistent with a non chemical homogeneity within the nanoparticles [
64
].
Figure
4.9
illustrates the in-field Mössbauer spectrum of 6 nm Co ferrite nanoparticles
prepared by thermal decomposition route: one observes clearly broadened lines and a
decrease of the intensities of the intermediate lines [
73
]. The corresponding distribu-
tions of effective field and canting angles are reported in Fig.
4.10
for both tetrahedral
and octahedral Fe sites and finally the estimated hyperfine field distributions which
well describe the zero-field spectrum [
73
].
