The internal friction spectra of high purity aluminium (99.999%) has been t
horoughly studied but, at the present time, there is no agreement about the
number of peaks at medium temperatures and the number of components of eac
h peak. This paper discusses the reasons for that controversy and summarize
s the relaxation processes found in the internal friction spectra. Some aut
hors have assumed that the P-2 peak (also called k (e) over cap peak) is fo
rmed by two components, one associated to a mechanism of dislocation mobili
ty and the other to grain boundary sliding. However, this assertion is base
d on weak experimental evidence, which is critically reviewed in this work.
The fitting methods used to decompose the internal friction spectra have a
lso been discussed, since some of them can generate spurious components and
lead to equivocal conclusions. Another point of great interest that has re
mained unclear concerns the shifts of the temperature of the peaks observed
by many authors, which are explained in this paper with the help of some m
echanisms of dislocation mobility. Several specific experimental measuremen
ts in 99.999% aluminium and 99.9999% aluminium doped with 10 ppm of copper
are presented to clarify these questions. It is concluded that PZ is compos
ed of just one component and is associated to a mechanism of dislocation mo
bility. In addition, we have determined that there are only three peaks in
the medium temperature range, namely P-1, P-2 and P-3. (C) 2000 Elsevier Sc
ience S.A. All rights reserved.