COST 507 - Repositório Aberto da Universidade do Porto

1 Introduction
During the last two years new experimental measurements on the Al-Mn and Al-Fe-Mn
systems have been carried out at Alean International Ltd (Banbury Laboratory) by
Thornton and Evans, and at University of Manchester/UMIST by Hayes et al,
specifically to elucidate the problems, following recommendations made COST 507
progress meetings. Work carried out as part of the COST 507 Action and under the
present project, together with supplementary work has incorporated these new data, and
(a) has provided a revised and improved dataset for the Al-Mn binary system which
is consistent with our current understanding of experimental data for this system, and
(b) has provided an updated dataset for the Al-Fe-Mn ternary system which is
considered very satisfactory for present purposes.
The effort has been considerable, and the problems, now largely resolved, have
understandably delayed the development of the Al-Fe-Mg-Mn-Si database, also the
addition of Cu. Although further work is required on the Al-Mn-Si system, which
suffers related experimental uncertainties, especially the data of [43Phi], the current
situation is good, very much improved compared with that of a year ago. The results
of experimental measurements on the Al-Mn and Al-Fe-Mn systems carried out at
Alean International during the course of the project, have proven essential to our
understanding of the experimental uncertainties and to the development of the database.
2 Data Assessment
2.1 Al-Mn system
The calculated phase diagram for this system is shown in Fig 1. The phase diagram
data for this binary system, crucial to the present project, has caused great concern due
to experimental inadequacies. Much painstaking work and careful analyses of the
available information has been carried out both in the present work and by others in
the field. The experimental data, particularly for equilibria involving Al 4 Mn, are fraught
with problems due to the appearance of metastable phases, rather than the stable phase,
see eg, [87Mur/McA, 87 McA/Mur, 92Jan, 96Wei/Rog]. This leads to low, nonequilibrium
liquidus temperatures, which, however, are consistently reproduced by
different experimental studies. For the present project the most important part of the
system is the range 0 to 4 or 6 wt % Mn, although for multicomponent calculations it
is important to have a good overall representation of the system. For the Al-rich region
the prime data source is due to Phillips [43Phi], coupled with recent data from
University of Manchester/UMIST [95Hay/Rob], now more generally available
[96Wei/Rog], carried out using a sophisticated calorimetrie technique, and data of
Thornton [96Tho4]. In addition Sigli [95Sig] has redetermined the solid solubilities of
Mn in fcc Al.
The current analysis takes into account a revision of the original [43Phi] thermal
analysis data by Phillips himself, reported obscurely in an evaluation of the Al-Mn-Si
system [59Phi] (see also [61Phi]), and includes information obtained on heating, rather
than cooling, by [33Dix/Fin, 87McA/Mur], which avoids the problem of undercooling.
The data of [96Wei/Rog] and [96Tho4] have been incorporated. Figs 2-4 show
calculated phase equilibria for the Al-rich part of the system. Fig 2 shows the
undercooling effects observed over the composition range 10 - 30 wt % Mn, almost
- 258 -