remained untransformed, although the surrounding liquid exhibited continuity. The same sandwich
subsequently annealed at 930 °C for 18 h only contained γ+L. Therefore, the transition temperature is
probably close to 915 °C and it is conditionally accepted in the updated phase diagram (see Fig. 8).
Considering these results, DTA experiments with more complex thermal procedures were carried
out on Al80Pt20 containing initially ξ+sL. For example, after heating to 850-880 °C at 50 K/min
(maximal rate allowed by the DTA device used) the samples were kept at the existing temperatures of
the ξ-phase for up to 2 h and then further heated to 1200-1300 °C at 10-15 K/min. Examples of the
recorded DTA plots are shown in Fig. 9. In Fig. 9a in the first heating step the ξ-phase decomposed
with the formation of λ+γ (exothermal reaction starting from 530 °C), and then recombined starting
from 813 °C. In the second heating step the transformation of the recovered ξ-phase was already
revealed (see Fig. 9b,c, weak endothermic effects starting from 998 °C) prior to the peritectic melting
of γ, but the transition temperature was much higher than 915 °C estimated from the above-mentioned
annealing experiments. In Fig. 9c the plot of a sample of a lower Pt concentration additionally exhibits
eutectic melting.
The observation of the above-mentioned duplets (not always resolved) in the range of 800-836 °C
is consistent with the existence of two reactions at close temperatures: the lower-temperature thermal
effect could correspond to ξ ↔ γ + λ and the higher to L + ξ ↔ λ. The recombination of the ξ-phase
from γ + λ by heating could be more difficult if samples have a coarser structure, so in DTA
experiments higher transition temperatures are expected to be observed. Samples annealed at 780 °C
initially contained a coarser γ + λ mixture than those obtained by a decomposition of the initial ξ-
phase. Thus, the lowest temperature of 801 °C, observed in an Al76Pt24 sample annealed at and
quenched from 880 °C, can be associated with the reaction ξ ↔ γ + λ. The direct temperature
determination of the L + ξ ↔ λ reaction is complicated in these samples due to its small difference
from that of ξ ↔ γ + λ , while in Pt-poorer samples overheating is an obstacle.
In order to specify the temperature of the L + ξ ↔ λ reaction, an Al84.7Pt15.3 alloy was annealed in a
series of experiments for an hour in the DTA device at different temperatures between 812 and 837 °C
and then cooled down at 10 K/min. The samples cooled from 837 or 822 °C exhibited a well-resolved
endothermic effect associated with the formation of the λ-phase and then the effect of the eutectic
solidification. The former was very weak if the sample was cooled down from 817 °C, while the
sample cooled down from 812 °C exhibited only eutectic solidification. Therefore, 812 °C is already
below and 817 °C is close to the L + ξ ↔ λ transition. Due to some temperature variations around the
latter temperature the ξ-phase appeared and disappeared, so finally only a small fraction of this phase
8