A. Microstructural development

According to the Cu–Al binary diagram and Cu-rich side of Cu–Al–Fe and Cu–Al–Ni ternary diagrams, the equilibrium microstructure of Cu–10Al–4Fe–4Ni alloy should be α + (α + γ₂) + K. The microstructure of NAB mainly probably consists of α, β, K_I, K_II, K_III, β′, and γ₂ phases. Wherein, α phase with an fcc crystal structure is based on copper-rich solution; β phase is based on Cu₃Al solution and possesses a bcc crystal structure; and among four kinds of K phases, the numbers are determined by the order of precipitation. K_I precipitates from the original liquid first, rich in iron and based on AlFe₃ mostly, possessing a bulk shape or a rosette shape and being more likely to be rosette shaped; K_II precipitates at a relatively high temperature, also rich in iron and based on AlFe₃, possessing a morphology similar to K_I, but the size is smaller and is more likely to be quasisphere shaped; K_III is lamellar or globular eutectoidal decomposition product, lying among the grain boundaries and being based on NiAl. As the temperature decreases, K_IV shows up in α phase, which contains high iron content, possesses a relatively small size and is based on AlFe₃. It is widely accepted that the K_I, K_II, and K_IV precipitates are rich in iron and based on AlFe with the DO₃ structure. The K_III phase is rich in aluminum and nickel and based on NiAl with the B2 structure. The morphology, the size, the location, and the composition are the main key factors to judge the kinds of K phases. The β′ is essentially martensite, and γ₂ is based on Al₄Cu₉ with a complex crystal structure.^{Reference Recarte, Perez-Saez, San Juan, Bocanegra and No10–Reference Chen, Liang, Zhang, Zhang, Luo and Li16} The allotropic transformation from β to β′ takes place as the temperature drops from a high point to a relatively low range rapidly.

Typical stages of microstructural development of alloys 1 and 2 continuously cooled from 1000 °C are exhibited in Figs. 1 and 2, respectively. With the aid of energy spectrum analysis (EDS) [Figs. 1(e) and 1(f)], the microstructure of alloy 1 was investigated. And it is confirmed to be made up of β and a few K_II particles at initial high temperature 1000 °C [Fig. 1(a)]. The quantity of K_II increases gradually and continuously as the temperature decreases. At about 850 °C, light-etching α precipitated from original β region, but its subsequent coarsening is impeded by K_II as shown in Fig. 1(c). On further cooling, the eutectoid decomposition of β into α + K_III initiates at α/β boundaries at about 750 °C. However, the reaction cannot be completed under normal cooling condition, and finally, the remnant β transforms to martensite β′. The microstructure of the specimen quenched from 450 °C is extremely similar to as-cast alloy, named as quasicast. Apparently, only α, K_II, K_III, and β′ phases present in the quasicast microstructure of alloy 1, and there is no K_I and K_IV phases present compared with the results investigated by Jahanafrooz et al.^{Reference Jahanafrooz, Hasan, Lorimer and Ridley7} on a set of NAB with a range of Al content from 9.0% to 9.4%.

FIG. 1.

Typical stages of microstructural development of alloy 1 prepared via SC at (a) 1000 °C, (b) 950 °C, (c) 850 °C, and (d) 750 °C and the EDS of (e) K_II and (f) K_III.

FIG. 2.

Typical stages of microstructural development of alloy 2 prepared via CC at (a) 1000 °C, (b) 950 °C, (c) 850 °C, (d) 800 °C, (e) 750 °C, and (f) 450 °C.

As illustrated in Fig. 2, α phase, most of which are in lamellar shape, and K_I particles are observed in alloy 2 at 1000 °C. K_I particles can act as the nucleation center of α in the subsequent solidification. When the temperature dropped to 950 °C, K_II particles precipitate from the original β region and hinder the growth of α phase, similar to alloy 1. K_IV starts to appear in the center of α at 850 °C. Previous literature^{Reference Jahanafrooz, Hasan, Lorimer and Ridley7} revealed that the region in β, where α first got observed, was always to be the area that K_IV chose to appear mainly due to the high amount of iron, which makes it the first region to reach the solubility limit of iron as the temperature continuously decreases. Figure 2(e) illustrates that the eutectoid (α + K_III) seems to be in lamellar or globular shape, and the reaction initiates at the α/β interface and grows into the β regions. In addition, a few initially formed particles of K_III are globular, whereas most of the subsequent K_III in eutectoid (α + K_III) are lamellar. In a word, the microstructure of quasicast alloy 2 consists of α, β′, K_I, K_II, K_III, and K_IV, which is extremely similar to the results investigated by Jahanafrooz et al.,^{Reference Jahanafrooz, Hasan, Lorimer and Ridley7} Culpan and Barnby,^{Reference Culpan and Barnby8} etc.

As shown in Fig. 3, the morphology of phases mentioned above was determined by TEM, and the orientation relationship between α and K_III was studied by electron diffraction. According to the thermodynamic principles of phase transformation and considering the driving force required,^{Reference van der Heide, Stam, Giraud, Lovato, Akdut, Clarysse, Caenen and Heikilla12} the phase interface/grain boundary is always chosen to be the habit precipitation point of phases. Hence, β′ precipitated from the phases interface/grain boundary and then extend into β as shown in Fig. 3(a); in this way, the interface energy can be reduced. The solid solubility of Fe was so sensitive to temperature variation that a sharply quench would lead to tremendous decline of the solubility, resulting in K_II particles [Fig. 3(b)] precipitated from β when the temperature dropped from a high point.

FIG. 3.

TEM micrographs taken along [011]α zone axis showing the morphology of (a) β′ martensite (b) K_II (c) α + K_III eutectoid, and (d, e) electron diffraction pattern taken from α + K_III eutectoid exhibits the K–S orientation relationship between α and K_III: ($00\overline 2$)α‖($00\overline 2$)K_III; [$1\overline 1 0$]α‖[$1\overline 1 0$]K_III (●K_III hkl, ○αhkl).

A fine lamellar aligned product (α + K_III) was formed from the eutectoid reaction β → α + K_III as shown in Fig. 3(c). The electron diffraction pattern [Figs. 3(d) and 3(e)] of the eutectoid which is composed of fcc α and bcc K_III shows a K–S orientation relationship: [$1\overline 1 0$]α‖[$1\overline 1 0$]K_III; ($00\overline 2$)α‖($002$)K_III. With the increase of the distance away from the interface, the nucleation work of eutectoid transformation needs to be raised, which means that it is more and more difficult for the eutectoid reaction to initiate. Due to the principle of minimum energy,^{Reference van der Heide, Stam, Giraud, Lovato, Akdut, Clarysse, Caenen and Heikilla12} the fine lamellar α + K_III eutectoid will roughen immediately to reduce the free energy of the whole system as soon as it precipitates.

The schematic representations of the precipitation sequence of alloys 1 and 2 are shown in Fig. 4, and the precipitation temperatures of various phases are listed in Table II. It is evident that the types, precipitation sequences, and initial precipitation temperatures of phases are quite different in the final alloys prepared via SC and CC processes. And the typical characteristics of phase transformation for alloy 1 are quite different from the original investigation by Jahanafrooz et al.,^{Reference Jahanafrooz, Hasan, Lorimer and Ridley7} etc.

FIG. 4.

Schematic representations of the sequence of microstructural development of (a) alloy 1 and (b) alloy 2.

TABLE II.

Initial precipitation temperature of various phases.

K_I can be observed in alloy 2, but unfortunately, it cannot be found in alloy 1 and Cu–9.6Al–4.2Fe–4.1Ni–1.2Mn alloy prepared by Zhang et al.,^{Reference Zhang, Chen, Zhang and Luo3} which has been confirmed to be related to the Fe/Ni ratio according to the researches completed by Jahanafrooz et al.^{Reference Jahanafrooz, Hasan, Lorimer and Ridley7} For further details, only when Fe/Ni ratio becomes higher than 1, will K_I turn up in the matrix, and the higher this value becomes, the more possible for K_I to be observed. Actually, alloy 2 has an eligible Fe/Ni ratio exactly listed in Table I. In metallographic photos, K_I is so easy to be corroded that it always presents as dark etched. In the bending test of as-cast nickel–aluminum examined by Culpan and Barnby,^{Reference Culpan and Barnby8} it was discovered that cracks always started from microinterface in the microstructure and then developed along or through K_I. As a consequence, K_I particles would be separated into two parts existing in each side of the broken pieces. The existence of K_I had an enormous negative influence on the mechanical and corrosion-resistance properties. So it is vital to control the content of Fe and the ratio of Fe/Ni properly. Furthermore, K_IV cannot be observed in alloy 1 possibly due to the reasons as follows: at first, the range of the precipitation temperature of K_IV is too narrow and the cooling rate is too fast for the particles to precipitate. Moreover, it is probably also related to the Fe content and the Fe/Ni ratio. During the microstructural development of alloy 1, there were some phenomena that do not coincide with those available results in previous literatures, such as K_II appearing prior to α and the delayed appearance of α phase, thus requiring further research in the future.

B. Trends of microhardness

The trends of microhardness at every stage during continuous cooling are briefly described as curves shown in Fig. 5. Both curves show a rapid decrease at 900 °C and reach a smooth platform finally. The decrease of hardness is due to the raise of the amount of soft α-phase as the temperature decreases. In terms of alloy 1, unexpectedly, when the temperature is higher than 900 °C, the value of the microhardness is lower than that of 900 °C. The data of 1000 °C could not be collected owing to the microcrack observed during the hardness test. Obviously, alloy 1 is extremely frangible between 950 and 1000 °C due to the present of β′ martensite and a small amount of K-phase [Figs. 1(a) and 1(b)] in this temperature range, both of which would be the crack source and paths for cracks propagation. From the figures, we can also get some information concluded as follows: when the temperature gets low enough, the as-quenched microstructures of both alloy 1 and alloy 2 stay stabilized. Additionally, the critical points should be 750 and 800 °C for alloy 1 and alloy 2, respectively, which is coincident with the precipitation temperature of K_III phase. Obviously, β → α + K_III eutectoid reaction has no significant contribution to the microharness.

FIG. 5.

Trends of microhardness of (a) alloy 1 and (b) alloy 2 at every stage during continuous cooling.

In brief, SC and CC processes are quite similar in the terms of solidification of the melt, whereas, the small distinction between them refers to the cooling rate. The original reason resulting in immense distinction in the precipitation temperatures of the phases between the two kinds of alloys is confirmed to be the differences of Fe contents, which basically due to the various casting mechanisms involved in these casting methods.