Twinned dendrite growth in binary aluminum alloys

Abstract

The formation of twinned dendrites (feathery grains) in binary Al-Zn, Al-Mg, Al-Cu and Al-Ni alloys has been studied in specimens directionally solidified under identical thermal conditions, i.e. G % 100 K cm À1, v % 1 mm s À1, and with slight natural convection in the melt. The influence of the solute element nature and content has been found to be of less importance than previously reported since feathery grains were formed in all four alloys, regardless whether the alloying elements are hexagonal close packed (Zn and Mg) or face-centered cubic with a high (Ni) or low (Cu) stacking fault energy. A detailed analysis confirmed that twinned dendrites grow along h1 1 0i directions in all four cases, with a complex branch morphology made of up to six to nine arms. Surprisingly, at high Zn or Mg compositions for which regular dendrites grow along h1 1 0i instead of h1 0 0i, [Gonzales F, Rappaz M. Metall Trans A 2006; 37: 2797. [1]] no twinned dendrites could be formed. In terms of both the growth kinetics advantage of twinned dendrites over regular ones and the associated tip shape, some experimental evidence seems to contradict the doublon conjecture suggested by Henry [Henry S. PhD thesis, Ecole Polytechnique Fédéral de Lausanne, 1999. [21]], at least for the solute compositions studied in the present work.

Study of the twinned dendrite tip shape II: Experimental assessment

“…The first argument of Henry is thus put into question. The third question is also unclear, since the present authors [6] did not find any increase in segregation near the twin plane in Al-30 wt.% Zn twinned dendrites. However, this was done on a transverse section of an specimen using wavelength-dispersive X-ray spectroscopy (WDS) chemical analysis in a scanning electron microscope, and not by HRTEM, and thus the resolution was possibly insufficient.…”

“…This might explain why Henry observed in these alloys cells in sections nearly parallel to the twin plane and dendrites perpendicular to it, with the consequence that k // ( k \ . In previous works, Salgado-Ordorica and Rappaz [6,7] showed that, in binary Al alloys with relatively high solute-content, k // is comparable to k \ , even though dendrites trunks were aligned along the twin planes. The first argument of Henry is thus put into question.…”

“…Doublons at the center of dendrite trunks, all aligned along a (1 1 1) twin plane, would definitely affect solute diffusion parallel and perpendicular to the twin plane, and thus the primary trunk spacing; (ii) besides the twin relationship between successive lamellae, the dendritic structure in these lamellae exhibits a gradual and systematic misorientation which is responsible for the fan-shape appearance of feathery grains [2,[6][7][8] solute gradient perpendicular to the twin planes and associated with doublons at the center of twinned dendrites can induce a gradual rotation of successive twin planes during branching when the solute element has a radius different from that of the solvent; and (iii) using highresolution transmission electron microscopy (HRTEM) in a region near to a coherent twin plane, Henry measured a composition of the solid close to the nominal composition C 0 of the alloy, instead of a value close to k 0 C 0 , where k 0 is the partition coefficient [2]. However, these measurements were made in industrial alloys of fairly low composition (Al-4.3 wt.% Cu-0.3 wt.% mg).…”

Influence of Cr on the nucleation of primary Al and formation of twinned dendrites in Al–Zn–Cr alloys: Can icosahedral solid clusters play a role?

“…It is known in industry that some aluminum alloys are more prone to feathery grains formation than others, but the role of the nature and amount of alloying elements is still not understood. For example, Salgado-Ordorica and Rappaz [7] could produce twinned dendrites in different binary Al alloys with solute elements having a low/high stacking fault energy (Cu and Ni, respectively) or solidifying with a hexagonal close packed structure (Mg and Zn), providing the other conditions were met. Surprisingly, these authors could not produce twinned dendrites in AlZn for compositions higher than 40 wt.%, for which "regular" dendrites (i.e., untwinned) are known to grow not along h1 0 0i directions, as do most cubic-symmetry metals, but along directions which are closer to h1 1 0i [10,11].…”

“…Up to now, the formation of twinned dendrites was shown to occur under specific solidification conditions which can be summarized as follows [6][7][8][9]: (i) poor or no inoculation of the alloy (in order to prevent the formation of equiaxed grains); (ii) high thermal gradient (typically 100 K cm -1 ) and cooling rate (typically 10 K s -1 ) (i.e., growth rate of typically 1 mm s -1 ); (iii) strong convection of hot liquid (e.g. associated with pouring) insofar as it promotes high gradient at the front, as convection induced by stirring drastically decreases the thermal gradient in the liquid and does not promote the formation of twinned dendrites.…”