Figure 2.1(a) is a selected-area diffraction pattern of the as-quenched alloy, exhibiting the superlattice reflection spots of the ordered L21 phase [14-16]. Figures 2.1(b) and (c) are (200) L21 (or, equivalently, (100) B2) and (111) L21 dark-field (DF) electron micrographs of the as-quenched alloy, showing the presence of the small B2 domains and fine D03
domains, respectively [12-13]. In Figures 2.1(b) and (c), it is seen that the sizes of both B2 and L21 domains are very small, indicating that these domains were formed during quenching [12-13]. In Figure 2.1(b), it is also seen that a high density of extremely fine disordered A2 phase (dark contrast) could be observed within the B2 domains; otherwise, there would be no contrast within these domains by using a (200) superlattice reflection [6]. Accordingly, the as-quenched microstructure of the alloy was a mixture of (A2+ L21) phases. This is similar to that observed by the present workers in the Fe-23 at.% Al-8.5 at.% Ti alloy quenched from 1373K [14].
Figure 2.2 is a (111) L21 DF electron micrograph of the alloy aged at 1173K for 3 h, clearly revealing that the L21 domains grew significantly and a/2<100> APBs were coated with the disordered A2 phase. However,
Figure 2.1(a)
Figure 2.1(b)
Figure 2.1(c)
Figure 2.1 Electron micrographs of the as-quenched alloy: (a) a
selected-area diffraction pattern. The foil normal is [011].
(hkl: disordered A2, hkl: L21 plane); (b) and (c) (200) and
(111) L21 DF, respectively.
Figure 2.2 (111) L21 DF electron micrograph of the alloy aged at 1173K
for 3h.
after prolonged aging at 1173K, some tiny particles started to form within the A2 phase. A typical example is shown in Figure 2.3. Figures 2.3(a) and (b) are (111) and (200) L21 DF electron micrographs of the alloy aged at 1173K for 6 h, showing that the (111) DF image and the (200) DF image are morphologically identical. Therefore, it is likely to conclude that the microstructure of the alloy aged at 1173K for 6 h was also L21 phase and the a/2<100> APBs were coated with the disordered A2 phase. However, the (111) L21 DF electron micrograph of the same area as Figure 2.3(a) with a higher magnification revealed that the a/2<100> APBs were fully dark in contrast, as shown in Figure 2.3(c); whereas the (200) L21 DF electron micrograph showed that some tiny particles (indicated with arrows) could be observed at the a/2<100> APBs, as illustrated in Figure 2.3(d). Therefore, it is reasonable to deduce that the tiny bright particles present in Figure 2.3(d) should be of B2 phase, since the (111) reflection spot comes from the L21 phase only; while the (200) reflection spot can come from both the L21 and B2 phases[12-13]. With continued aging at 1173K, the L21 domains continued to grow and a phase separation started to occur basically contiguous to a/2<100> APBs of the L21
domains. An example is shown in Figure 2.4. Figure 2.4(a) is a (111) L21
DF electron micrograph of the alloy aged at 1173K for 12 h, showing that
Figure 2.3(a)
Figure 2.3(b)
Figure 2.3(c)
Fig. 2.3(d)
Figure 2.3 Electron micrographs of the alloy aged at 1173 k for 6 h. (a)
and (b) (111) and (200) L21 DF, respectively. (c) and (d) (111)
and (200) L21 DF with a higher magnification of (a) and (b),
respectively.
the a/2<100> APBs broadened and well-grown L21 domains decomposed into fine L21 domains (designated as L21* phase to be distinguished from the original L21 phase) separated by dark layers. Figure 2.4(b) is a (200) L21 DF electron micrograph taken from the same area as Figure 2.4(a), clearly revealing that in addition to the presence of a few A2 particles, the whole region is bright in contrast. This indicates that the broadened dark lines on a/2<100> APBs and dark layers around the periphery of the L21* domains should be of the B2 phase. It is apparent that the B2 phase was formed at a/2<100> APBs and phase separation from L21 to (B2+ L21*) occurred basically contiguous to the a/2<100> APBs. Figure 2.5(a) is a (111) L21 DF electron micrograph of the alloy aged at 1173K for 24 h, indicating that with increasing aging time, the phase separation would proceed toward the inside of the L21 domains. Figure 2.5(b), (200) L21 DF electron micrograph taken from the same area as Figure 2.5(a), clearly reveals that only one a/2<100> APB and one A2 particle (indicated with arrows in Figures 2.5(a) and (b)) could be observed. Figures 2.6(a) and (b) are (111) and (200) L21 DF electron micrographs of the alloy aged at 1173K for 36 h, revealing that besides a little A2 phase, the well-grown L21 domains decomposed into the (B2+ L21*) phases completely.
Figure 2.4(a)
Figure 2.4(b)
Figure 2.4 Electron micrographs of the alloy aged at 1173 k for 12 h.
(a) and (b) (111) and (200) L21 DF, respectively.
Figure 2.5(a)
Figure 2.5(b)
Figure 2.5 Electron micrographs of the alloy aged at 1173 k for 24 h.
(a) and (b) (111) and (200) L21 DF, respectively.
Figure 2.6(a)
Figure 2.6(b)
Figure 2.6 Electron micrographs of the alloy aged at 1173 k for 36 h.
(a) and (b) (111) and (200) L21 DF, respectively.
Based on the above observations, some important experimental results are discussed below. When the present alloy was aged at 1173K for moderate times, the B2 phase was formed at a/2<100> APBs and phase separation from well-grown L21 to (B2+ L21*) occurred basically contiguous to the APBs. With increased aging time at 1173K, the phase separation would proceed toward the inside of the whole well-grown L21
domains. This finding is different from that observed by the present workers in the Fe-23 at.% Al-8.5 at.% Ti alloy[14], in which we have demonstrated that when the Fe-23 at.% Al-8.5 at.% Ti alloy was aged at 1173K, the mixture of the (B2+ L21) phases occurred at a/2<100> APBs and no evidence of the phase separation could be observed. In order to clarify the apparent difference, an STEM-EDS study was undertaken. The EDS analyses were taken from the areas of the L21 domains, APBs, B2 phase and L21* domains marked as “L”, “A”, “B” and “L*” in Figures. 2.2 through 2.6, respectively. The average concentrations of the alloying elements obtained by analyzing at least ten different EDS spectra of each phase are listed in Table 2.1. For comparison, the chemical composition of the as-quenched alloy is also listed in Table 2.1. It is clearly seen in Table 2.1 that when the alloy was aged at 1173K for 3h, the Al and Ti concentrations in the L21 domains were distinctly higher than those in the
Table 2.1 Chemical compositions of the phases revealed by EDS.
Heat treatment Phase Chemical composition (at.%)
Fe Al Ti
as-quenched alloy. This means that along with the growth of the L21
domains, the concentrations of both Al and Ti at a/2<100> APBs would be lacking. The insufficient concentrations of both Al (19.8 at.%) and Ti (3.8 at.%) would cause the disordered A2 phase to form at a/2<100> APBs, which is consistent with the previously established Fe-Al-Ti phase diagram as shown in Figure 2.7 [10]. According to the phase diagram, the chemical composition of Fe-19.8 at.% Al-3.8 at.% Ti is just located in the A2 phase region. EDS analyses indicated that when the alloy was aged at 1173K for 3-12 h, the Ti concentration in the L21 domains maintained to be about 7.8 at.% and the Al concentration gradually decreased with increasing the aging time. It is thus expected that the Al atom would proceed to diffuse toward the a/2<100> APBs during aging. In Table 2.1, it is seen that when the alloy was aged at 1173K for 6 h, the Al concentration at a/2<100> APBs increased to 22.1 at.%. The significant increase of the Al concentration would lead the tiny B2 particles to form at a/2<100> APBs, which is also consistent with the Fe-Al-Ti phase diagram in Figure 2.7 [10]. In the phase diagram, it is clearly seen that the chemical composition of Fe-22.1 at.% Al-3.9 at.% Ti is close to the A2/B2 transition boundary. With increasing the aging time, the L21 domains continued to grow. The quantitative analyses revealed that the Al
Fig. 2.7 Isothermal section of the Fe-Al-Ti system at 1173K.
concentration of the L21 domains in the alloy aged for 6 h was 25.7 at.%
while that for 12 h was 25.2 at.%, which gave a decrease of only 0.5 at.%.
However, along with the enlargement of the L21 domain size, the volume fraction of the a/2<100> APBs would be lessened considerably. This means that the slight decrease of the Al concentration in the L21 domains would cause the Al concentration at a/2<100> APBs to increase appreciably. Therefore, it is reasonable to believe that due to the appreciable increase of the Al concentration, the A2 phase existing at a/2<100> APBs would be transformed to B2 phase [10], as observed in Figure 2.4(a). TEM observations indicated that along with the formation of the B2 phase, the phase separation from well-grown L21 to (B2+ L21*) occurred basically contiguous to a/2<100> APBs. In the following description, we will attempt to discuss why the well-grown L21 domains underwent the phase separation. In the previous studies, it is well-known that the D03 phase could be found to exist in the Fe-Al binary alloys only at temperatures below 823K [12-13], and the Ti addition in the Fe-Al binary alloys would result in a particularly large increase of the D03
(L21)→B2 transition temperature about 60K/at.% [6-11]. Obviously, the Ti concentration would be the predominant factor for the stabilization of the L21 phase at high temperature. In our previous study of the Fe-23 at.%
Al-8.5 at.% Ti alloy aged at 1173K for 6-24 h [14], it was found that the Ti concentration in the well-grown L21 domains was up to about 11.1 at.%, therefore, the L21 phase exhibited more stable and no evidence of the phase separation could be detected. However, when the present alloy was aged at 1173K for 6-24 h, the Ti concentration in the well-grown L21
domains was found to be only about 7.8 at.%, which is noticeably lower than that detected in the previous alloy. Therefore, it is plausible to suggest that owing to the lower Ti concentration, the well-grown L21
domains seemed not very stable at 1173K. Consequently, the well-grown L21 domains would separate to the mixture of the Ti-riched L21* and Ti-lacked B2 phase, as observed in Figures 2.4 through 2.6.
Finally, three more features are worthwhile to note as follows: (1) In the previous study [14], it is clearly seen that the a/2<100> APBs of the well-grown L21 domains exhibited more pronounced anisotropy than those observed in the present alloy. The reason is possibly that the well-grown L21 domains in the previous alloy had significantly higher Ti concentration. (2) The chemical composition of the previous alloy had
higher Ti and lower Al contents [14]. When the previous alloy was aged at 1173K, the L21 domains grew rapidly and the Ti atom was the major element for diffusing into the a/2<100> APBs. This effect caused the tiny L21 particles to precipitate preferentially at a/2<100> APBs [14]. However, when the present alloy was aged at 1173K, the Ti concentration in the well-grown L21 domains maintained to be about 7.8 at.% and the Al atom played the dominant role to diffuse into the a/2<100> APBs. Therefore, the tiny B2 particles were found to form at the a/2<100> APBs, rather than L21 particles. (3) According to the Fe-Al-Ti phase diagrams [6-11], the A2 phase can not appear for the Fe-24.6 at.% Al-7.5 at.% Ti alloy. However, a little A2 phase was always observed in the present alloy aged at 1173K for 12-36 h. The reason for the difference is plausibly that the phase diagrams were determined by the Fe-Al-Ti alloys heat-treated at 1173K for a time period longer than 336 h, whereas the present alloy was aged for 36 h only. Actually, in the present study, it was found that when the as-quenched alloy was aged at 1173K for 3-36 h, the amount of the A2 phase was indeed decreased with increasing the aging time.
2-4 Conclusions
In the as-quenched condition, the microstructure of the Fe-24.6 at.%
Al-7.5 at.% Ti alloy was the mixture of (A2+ L21) phases. When the as-quenched alloy was aged at 1173K for 6 h, the L21 domains grew considerably and the B2 phase was formed at a/2<100> APBs as well as the phase separation from well-grown L21 to (B2+ L21*) occurred basically contiguous to the a/2<100> APBs. After prolonged aging at 1173K, the phase separation would proceed toward the inside of the whole well-grown L21 domains. Consequently, the microstructure of the alloy aged at 1173K for 36 h was essentially the mixture of (B2+ L21*) phases.
References
1. Y. Nishino, S. Asano, T. Ogawa, Mater. Sci. Eng. A 234–236 (1997) 271-274.
2. F. Stein, A. Schneider, G. Frommeyer, Intermetallics 11 (2003) 71-82.
3. M. Palm, Intermetallics 13 (2005) 1286-1295.
4. L. Anthony, B. Fultz, Acta Metall. Mater. 43 (1995) 3885-3891.
5. O. Ikeda, I. Ohnuma, R. Kainuma, K. Ishida, Intermetallics 9 (2001) 755-761.
6. M.G. Mendiratta, S.K. Ehlers, H.A. Lipsitt, Metall. Trans. A 18 (1987) 509-518.
7. I. Ohnuma, C.G. Schon, R. Kainuma, G. Inden, K. Ishida, Acta Mater.
46 (1998) 2083-2094.
8. S.M. Zhu, K. Sakamoto, M. Tamura, K. Iwasaki, Mater. Trans. JIM 42 (2001) 484-490.
9. M. Palm, G. Sauthoff, Intermetallics 12 (2004) 13451359.
10. M. Palm, J. Lacaze, Intermetallics 14 (2006) 1291-1303.
11. G. Ghosh, in: G. Effenberg (Ed.), Ternary Alloy Systems, Springer, Berlin, 2005, pp. 426-452.
12. P.R. Swann, W.R. Duff, R.M. Fisher, Metall. Trans. 3 (1972) 409-419.
13. S.M. Allen, J.W. Cahn, Acta Metall. 24 (1976) 425-436.
14. C.W. Su, C.G. Chao and T.F. Liu: Scr. Mater. 57 (2007) 917-920.
15. S.Y. Yang and T.F. Liu: Scr. Mater. 54 (2006) 931-935.
16. C.H. Chen, T.F. Liu, Metall. Trans. A 34 (2003) 503-509.