Experimental and Numerical Investigation of a Solidification-Based Aluminum-Cooled Finger Refinement Process From Micro to Macro-Scale

ORIGINAL RESEARCH ARTICLE Experimental and Numerical Investigation of a Solidification-Based Aluminum-Cooled Finger Refinement Process From Micro to Macro-Scale DANILO CURTOLO, CHRISTIAN SCHUBERT, ALEXANDRE VIARDIN, SEMIRAMIS FRIEDRICH, MORITZ EICKHOFF, BERND BÖTTGER, BERND FRIEDRICH, HERBERT PFEIFER, and MARKUS APEL The interest in ultra-pure metals is steadily growing due to the increasing demand for these materials in modern technology. To be able to meet the increasing demand in the future, it is necessary to implement more efficient and productive processes. As a fractional crystallization method in this application area, the cooled finger method exhibits higher productivity and lower energy requirements when compared to industry well-established methods like zone melting. In this study, the mechanisms and relevant phenomena crucial for a successful implementation of a cooled finger process were investigated using a multidisciplinary approach. With carefully selected process parameters, we present here an experimental setup with a purification potential of approximately 80 pct. Additional micro- and macro-scale simulations demonstrate that the process is sensitive to parameters such as rotation rate, cooling rate, and temperature gradient within the melt, which explains the difficulty in optimizing this process in practice. An analysis and description of various phenomena that characterize the behavior of the cooled finger process are presented within this multi-scale approach. As a result, these approaches can also be transferred to the description of processes for other metals, opening application areas outside of the purification of aluminum. https://doi.org/10.1007/s11661-023-07147-0 Ó The Author(s) 2023 I. INTRODUCTION ALUMINUM in its primary form—from Hall– Héroult process—has a purity range of 2N7 to 3N.[1] While such purity is normally sufficient for the majority of industrial applications and alloying, the need for higher purity levels grows together with the advances of modern technology. Fields such as semiconductor, electronics, automotive, aerospace, high precision instruments, batteries, and superconducting are some examples of high-purity and ultra-high-purity applications of aluminum.[2] DANILO CURTOLO, SEMIRAMIS FRIEDRICH, and BERND FRIEDRICH are with the IME - Institute of Process Metallurgy and Metal Recycling, RWTH Aachen University, Intzestraße 3, 52072 Aachen, Germany. Contact e-mail: CHRISTIAN SCHUBERT, MORITZ EICKHOFF, and HERBERT PFEIFER are with the IOB - Department for Industrial Furnaces and Heat Engineering, RWTH Aachen University, Kopernikusstraße 10, 52074 Aachen, Germany. ALEXANDRE VIARDIN, BERND BÖTTGER, and MARKUS APEL are with the ACCESS e.V., Intzestraße 5, 52072 Aachen, Germany. Manuscript submitted March 1, 2023; accepted July 18, 2023. Article published online August 3, 2023 3988—VOLUME 54A, OCTOBER 2023 The high-volume and industrial production of ultrapure aluminum follows two main routes: three-layer electrolysis and fractional crystallization. Both processes can be performed separately to achieve purities of up to 4N8 or in series for higher purification levels reaching 7N.[3,4] Other alternative routes such as vacuum distillation are reported in the literature and can be used for low volume production and/or highly specialized applications.[5] A. Fractional Crystallization of Aluminum The production of ultra-pure aluminum via fractional crystallization (a.k.a. segregation) has been since many decades performed as an alternative to the cost- and investment-intensive three-layer electrolysis. The fractional crystallization mechanism promotes the expelling of the impurities from the crystallization interface. In fractional crystallization, the difference between the solubility of the impurity elements in both solid and liquid phases of the base metal is explored to crystallize a solid with lower impurities content than the initial molten phase.[6,7] The ratio between the concentration of an impurity element in the solid (CS ) and the liquid (CL ) phase is called distribution coefficient (k). This METALLURGICAL AND MATERIALS TRANSACTIONS A coefficient indicates the maximum theoretical impurity segregation that can be achieved in one step of purification.[8] In Table I, the literature values of the k coefficient of most impurities present in aluminum are compiled.[4,9–12] The values of CS and CL can be easily taken from a binary phase diagram of any element dissolved in the metal. The value of k can vary from less than 103 to greater than 10. Impurities that have a k value lower than unity consist of the majority of impurities present in metals and can be segregated to the liquid phase during crystallization.[6,13] On the other hand, impurities with k higher than unity tend to be incorporated into the forming solid during crystallization. For Al, the elements V, Ti, Cr, and Zr belong to this category, and can be removed prior crystallization with the stoichiometric addition of boron to form non-soluble borides. Any eventual excess of B k<1 will be removed during the crystallization. Some impurities like Pb, Bi, Cd, present themselves with a monotectic crystallization behavior, Table I. Literature Values of Distribution Coefficients of Impurities in Aluminum[4,9–12] Elements Distribution Coefficient k Elements Distribution Coefficient k Fe Cu Ag Au Zn Ni Mn Mg Ca Cr 0.018 to 0.053 0.15 to 0.153 0.2 to 0.3 0.18 0.35 to 0.47 0.004 to 0.09 0.55 to 0.9 0.29 to 0.5 0.006 to 0.08 1.8 Ti Si K Zr Pb P Sc Sb V Na 7 to 11 0.082 to 0.12 0.56 2.3 to 3 0.0007 to 0.093 < 0.01 0.9 0.09 3.3 to 4.3 0.013 i.e., the formation of a solid phase and a secondary liquid phase upon crystallization. For these cases, two stages of crystallization can occur depending on whether the impurity concentration is higher than its solid solubility in the metal matrix. For concentrations lower than the solid solubility, the impurity behaves as any other impurity with k<1 with no secondary liquid phase formed.[13] B. Growth Mechanisms The fractional crystallization process aims for a low crystal growth rate, with a moving velocity of the crystallization front usually not higher than 2:5 mm min1 :[6] This growth rate is necessarily low to allow enough time for the impurities segregated during the crystallization to diffuse into the liquid phase. In addition, the lower growth rates enable the formation and stabilization of cellular or even planar growth interface. Under high growth rate conditions or high concentration of solutes, a dendritic growth interface will be formed. This ultimately leads to a poor refining performance due to the excess solute entrapped between the dendrite arms.[14] 1. Growth interface morphology The growth interface morphology is mostly driven by the growth conditions, such as the constitutional supercooling, growth rate, and temperature gradient at the solid–liquid interface. Supercooling during the growth induces different interface mor (...truncated)