Chemistry Reference
In-Depth Information
subnanometer precision, the site location of impurity atoms in a single crystal
lattice. The minimum impurity concentration needed for such studies is substan-
tially higher than what can be studied in Mössbauer spectroscopy experiments.
Fe in Si shows a complex behaviour which has challenged Mössbauer physicists
for almost 50 years and even today still is not fully understood. As will be discussed
in
Sect. 6.5
of this chapter, it is still a challenge for semiconductor industry.
In a review paper on ''The Mössbauer search for Fe in Si'' it is stated [
18
] that
among the close to one thousand papers that so far had been published on Mössbauer
spectroscopy studies in semiconductors, involving twenty different elements, the
Mössbauer work on
57
Fe work in Si had been particularly combersome. In
early Mössbauer experiments, as early as 1962 [
19
,
20
], shortly after the discovery
of the Mössbauer effect in 1958, the role of precipitate formation during diffusion
was insufficiently realized, and the interpretation of the data was not very reliable.
After ion implantation the Mössbauer spectra of
57
Fe in Si, both in emission
Mössbauer spectroscopy, after implanting radioactive
57
Co [
21
], as in absorption
Mössbauer spectroscopy, after implanting
57
Fe [
22
], were dominated by two single
lines. Also in early in-beam experiments by the Stanford group, starting from
the
57m
Fe parent state, a technique discussed in
Sect. 6.3
, two lines were observed
[
13
,
23
] with asymmetric intensities.
A controversy arose about the interpretation of these two lines. As can be seen
in Fig.
6.10
,in
57
Co/
57
Fe emission experiments the line intensities varied as a
function of implantation fluence. In
57
Fe(Si) absorber experiments, after implan-
tation at fluences higher than to the highest fluences used in Fig.
6.10
, a fairly
symmetric doublet was observed. In the in-beam Mössbauer experiments, at
implantation fluences lower than the lowest fluences shown in Fig.
6.10
,an
asymmetric doublet was observed. Was there a common interpretation possible?
It was thought that the two single lines might be representative of two
implantation sites, presumed to be substitutional and interstitial Fe in Si, but the
observed dose dependence could not be accounted for. By applying an external
magnetic field it could unambiguously be shown [
24
] that the high fluence
57
Fe(Si)
spectrum was a quadrupole interaction doublet, with parameters very similar to
those of amorphous Fe
x
Si
1-x
films [
22
] and hence associated with Fe atoms in an
amorphized surrounding.
The origin of the asymmetry at lower fluencies was first accounted for by new
emission Mössbauer experiments whereby
57
Co was implanted into Si held at the
temperature of 50 K [
25
]. This cold implantation completely reversed the asymmetry
(Fig.
6.11
) in the spectrum compared to room temperature implanted samples. The
model put forward was that at low temperatures, where vacancies are not mobile,
57
Co/
57
Fe atoms land up in interstitial sites in Si, a site preferred by transition metals in
Si. Upon increasing the temperature to room temperature, vacancies become mobile at
about 100 K and are trapped by the interstitial Co atoms that become substitutional.
New experiments on Fe in Si with the in-beam Mössbauer spectroscopy tech-
nique were performed at the Hahn-Meitner institute in Berlin. The technique is
outlined in
Sect. 6.3
. The results of these Coulomb excitation recoil implantation
studies [
26
] confirmed this picture and led to an unambiguous identification of
