ARTICLE IN PRESS
C. Frandsen, S. M^rup / Journal of Magnetism and Magnetic Materials 266 (2003) 36–48
43
of 10ꢁ9–10ꢁ8 s and these particles will give
contributions with broad lines [19]. However, if
250 K. The spectra are dominated by sextets with
broad lines far above the temperature where the
spectra of the non-interacting particles have
collapsed to a doublet. The magnetic energy of a
ꢁ10
t0o10
s, the exponential dependence of the
relaxation time in conjunction with the particle
size distribution results in a very broad distribu-
tion of relaxation times such that only a small
fraction of the particles have relaxation times in
this range. The contribution of these particles to
the spectra is therefore almost negligible. This is
discussed in more detail elsewhere [20].
particle, i; with volume V ; which interacts with its
i
neighbours, j; may be written [1,2]
Ei ¼ KVi sin y ꢁ Mi ꢃ
X KijMj:
ð4Þ
2
j
Here Mi and Mj represent the (sublattice)
magnetisation of the particles i and j; respectively.
Kij represents the effective exchange coupling
constant originating from exchange coupling
between surface atoms of the particles i and j: If
the anisotropy term (the first term) in Eq. (4) is
predominant, superparamagnetic relaxation of the
magnetisation may take place between orienta-
tions of the sublattice magnetisation along the easy
axes close to y ¼ 0ꢀ and 180 . In this case the
distribution of the values of the energy barriers
Below the superparamagnetic blocking tempera-
ture, the (sublattice) magnetisation vectors fluc-
tuate in directions close to the easy directions of
magnetisation. These fluctuations (collective mag-
netic excitations) are usually fast compared to the
time scale of M o. ssbauer spectroscopy and result in
a reduction in the magnetic splitting of the spectra
given by [21,22]
ꢀ
BobsCB f1 ꢁ k T=ð2 KVÞg;
ð3Þ
0
B
where B0 is the magnetic hyperfine field in the
absence of these fluctuations. The maximum
reduction will normally not exceed 10–15%,
because further increase of the temperature or
decrease of the particle size results in fast super-
paramagnetic relaxation [22], i.e. the spectrum
collapses to a doublet. Collective magnetic excita-
tions in conjunction with the particle size distribu-
tion results in some line broadening, but because
the reduction of the magnetic splitting is less than
(KV ) in a typical sample results in a broad
i
distribution of relaxation times leading to
M o. ssbauer spectra like those in Fig. 1a. If the
interaction term (the second term) is large, it may,
below a critical temperature, result in formation of
an ordered state of otherwise superparamagnetic
particles [1,2,8,9]. If the interaction energy is
predominant compared to the anisotropy energy,
there will be only one energy minimum, defined by
the effective interaction field. At finite tempera-
tures, the (sublattice) magnetisation vectors may
then fluctuate in directions close to that corre-
sponding to the energy minimum. The magnetic
properties of such samples have been described by
a simple mean field model, in which the summa-
tion in the last term of Eq. (4) is replaced by the
average value [1,2]. The degree of ordering can be
described by an order parameter, which is unity at
T ¼ 0 K and decreases with increasing tempera-
tures until it vanishes at the transition temperature
at which the sample becomes superparamagnetic.
The temperature dependence of the order para-
meter can be calculated by using Boltzmann
statistics [1,2]. Well below the ordering tempera-
ture there is no energy barrier, which limits the
frequency of the fluctuations of the sublattice
magnetisation directions. Therefore, it is likely
that the fluctuations are fast compared to the time
1
narrow at all temperatures. In the present case
0–15%, the lines in the sextet remain relatively
ꢁ
11
with t0E10 s, one finds from Eq. (2) that
ꢁ9
tE5 ꢂ 10 s for KV=kBTE6: Therefore, the
magnetic splitting collapses when the reduction
of the magnetic hyperfine field approaches 8%. A
substantial line broadening, like that seen in, for
example, the high-temperature spectra in Figs. 1b
and 4a, cannot be explained by collective magnetic
excitations, because these spectra contain sextets
with hyperfine fields, which are reduced by much
more than 10%.
The evolution with temperature of the spectra of
the interacting particles (Fig. 1b) is completely
different from that of the non-interacting particles.
The lines of the a-Fe O sextet become increas-
2
3
ingly broadened with increasing temperature, but
there is no clearly visible doublet at least up to