Thermodynamic modelling of the Al–Co–Fe system

Abstract

In the present work, the Al–Co–Fe ternary system is thermodynamically modelled using the Calphad method. Experimental data such as liquidus, tie lines, phase boundaries, magnetic transition and order–disorder data points are included and critically examined. The data that has a better quality has been chosen to optimize the system. An order–disorder model has been used to describe the bcc and B2 phases. Experimental bcc/B2 transition data points were carefully examined and inconsistent data points were weighted less. A four-stage optimization was employed to fit the magnetic and bcc/B2 transitions and phase boundaries. The thermodynamic models of Al₅Fe₂, Al₅Co₂, Al₂Fe, and Al₉Co₂ are adjusted to include the third element to reflect the solubility of this element in the ternary system. Ternary interaction parameters for bcc and fcc were optimized, using all the relevant experimental data in the literature. The calculation of isothermal and vertical sections are performed using the optimized model parameters and compared with the experimental data. A comparison between modelling and experimental measurements showed a good agreement between the present results and experiments.

Introduction

The Al–Co–Fe ternary system is one of the systems that describes Fe- and Ni-based superalloys, and high-entropy alloys. Al–Co–Fe has been extensively studied. In this ternary system, B2 as an ordered structure of disordered bcc phase dominates the centre of the phase diagram and extends from the AlCo to AlFe side with a miscibility gap from low to medium temperature. The disordered bcc solid solution is stable in the Fe-rich side, where it has bcc/B2 and magnetic transitions. Fcc is located in the Co-rich side of the phase diagram and propagates toward the Fe-rich side in medium temperature with a limited solubility range of Al. The solubility of hcp in the ternary phase diagram has not been investigated but is assumed to be very limited in the Co-rich side at low temperature. All the intermetallic phases are located in the Al-rich part. In this region, M-Al₁₃(Co,Fe)₄ is stable from Al–Co to Al–Fe binary. Al₅Co₂ extends in the ternary system toward Al–Fe binary. Al₉Co₂, Al₅Fe₂, and Al₂Fe have a restricted range of solubility of the third element. All the intermetallic phases have a very limited homogeneity range of Al in the ternary system. It should be noted that Al₅Co₂ and Al₅Fe₂ have different types of crystallographic structure (see Table 1). The propagation of Al₈Fe₅ in the ternary system is unknown. Other intermetallic phases have no solubility in the ternary system.

Al–Co–Fe has been experimentally studied by Köster [1], Edwards [2], Raynor and Waldron [3], Miyazaki et al. [4], Ackermann [5], Kozakai and Miyazaki [6], Kozakai et al. [7], Kamiya et al. [8], Grushko et al. [9], and very recently by Zhu et al. [10]. Ostrowska and Cacciamani [11] have published a Calphad assessment of this ternary system.

In a recent assessment of the Al–Co–Mn system [12], we modified the bcc and B2 phases in the binary Al–Co system in order to improve the fit of the bcc/B2 second-order transition in the ternary system. If this modified Al–Co assessment is used together with the Al–Co–Fe assessment from Ostrowska and Cacciamani [11], the ternary phase diagram is not reproduced. The new experimental data from Zhu et al. [12], which were not available to Ostrowska and Cacciamani [11], are in good agreement with previous data, but additional data for bcc-fcc equilibria for Fe-rich compositions at 1000, 1100 and 1200 °C, show that the calculated Al solubility in fcc from Ostrowska and Cacciamani [11] is too low, in particular at 900 °C. Furthermore, they used rather large ternary interaction parameters and, in particular, the bcc interaction has a large temperature dependence. With this they could get a reasonably good representation of experimental data, both at relatively low temperature (650 °C) and at higher temperature, but extrapolations to higher order systems could be problematic. The bcc/B2 transition is second-order at 800 °C and above, but in Ostrowska and Cacciamani [11] it is shown as first-order to a considerable extent up to 1000 °C. For these reasons we decided to make a complete new assessment of the Al–Co–Fe system using all available experimental data.