Halo Formation of Zn-Al Alloys underConventional Solidification and Intensive Convection Solidification

WenchaoYanga,b, HasseFredrikssonc, Shouxun Jib,*

aState Key Laboratory of Solidification Processing, NorthwesternPolytechnical University, Xi’an 710072, China

b Brunel Centre for Advanced Solidification Technology, Brunel University London, Uxbridge, UB8 3PH, UK

cMaterials Science and Engineering, Royal Institute of Technology, Stockholm, Sweden

*Corresponding author. Tel: +44(0)1895 266663, Fax: +44(0)1895 269758, Email:

Abstract

A halo occurred usually as an envelope of one phase around a primary phase in many alloysafter solidification. Itsformationmechanism was investigated for hypoeutectic, eutectic and hypereutectic compositions of Zn-Al alloys under conventional solidification and under intensive convection solidification. It was found that the Zn-rich halos occurred in the surroundings of the Al-rich primary phase for the hypereutectic Zn-Alalloys at Al>5wt.% and no halos occurred for the hypoeutectic and eutecticZn-Al alloys at Al≤5wt.% under conventional solidification. However, the Zn-rich halos were completely absent from the Al-rich phase because of the uniform temperature distribution and enhanced mass transport under intensive convection solidification. Once the intensive convection was interrupted during solidification for the solid-liquid co-existing melt, a halo was formed on the surface of the existing Al-rich phase created either during the primary solidification or the eutectic solidification. Therefore, it was concluded that the halo formation should be a growth-dominant phenomenon not a nucleation-dominant phenomenon. And, the interaction among the solid phases and the liquid phase was responsible for the halo formation, in which the difference in the elasticity modulus and the density of the different phases resulted in the variation of strain energy in the individual phase.

Key words: Solidification;Microstructure;Crystal growth; Alloys; HaloFormation

1.Introduction

The phenomenon of primary phase surrounded by a halo of second phase is a common occurrence during the solidification of a variety of eutectic alloys. Some of these alloys are commercially very important and have been widely used in industry, such as Al-Si [[1],[2],[3]], Sn-Pb[[4]], Fe-C [[5],[6]], Zn-Al[[7]] and many others [[8],[9]].Halo formation is generally characterized by its uniqueness of an envelope of non-faceted phase around a faceted primary phase, not vice versa. By controlling the morphology of halo, the alloys with required mechanicalproperties may be developed, such as the deformationbehavior of ductile iron.This unique feature is important to affect other performance such as corrosion resistant and the heat treatment process of casting alloys. Therefore, a number of investigations have been the subject for elucidating the origin and characteristics of halos. They are mainly grouped into two categories according to the mechanism of solidification: nucleation-dominant or growth-dominant.

Sundquist et al. [[10]] and later supported by Salli et al.[[11]] interpreted the halo formation in eutectic alloys on the basis of nonreciprocal nucleation of alloy pairs, i.e. if phase  nucleated phase  with little or no undercooling; was a poor nucleant of . The liquid composition would follow the metastable liquidus line to a temperature significantly below the eutectic temperature before the nucleation of phase. The supersaturated phase in the remaining liquid would form halos around the primary  phase in order to bring the system back to eutectic composition for lamellar growth. Therefore, the size of the halo should be proportional to the undercooling for the formation of the second phase, and the halo phase is the phase with relatively weak nucleation ability. Bluni et al. [7] provided the evidence from Zn-Al alloy to support nucleation-dominant theory.The undercooling was 0.25°C for the hypereutectic Zn-Alalloys with Al>5wt.% where Al wasa primary phase and 4°C for the hypoeutectic Zn-Alalloys with Al<5wt.% where Zn wasa primary phase. The 0.25°C extension of the Al liquidus line below the eutectic point represented a very small undercooling for the nucleation of Zn by Al, and the halo formation was therefore not necessary to bring the liquid composition back to eutectic point. The halo was thought to result from the fact that Zn was a poorer nucleant of Al rather than vice versa.Therefore,the formation of Zn halos around the Al phase is mainly due to the relatively large undercooling required for the Zn phase to nucleate on the Al phase below the eutectic temperature. The Zn halo will grow at such an undercooling before the Al phase nucleates and grows up steadily.

Contrarily to the nucleation theory, Kofler [[12]] suggested that the reasonof halo formationwas the coupled zone argument rather than undercooling considerations. In a binary system, the coupled zone defined the alloy compositions and interface undercooling. The primary  phase continued to grow up until nucleation of  phase occurred at a undercooling. If the undercooling for nucleation of  phase was negligible, or the coupled zone was sufficiently skewed towards the  phase liquidus extension, coupled eutectic growth occurred directly from the primary  phase and no halos formed in the eventual microstructure. If the undercooling for the nucleation of the  phase was significant and the coupled zone did not encompass the liquidus extension of the  phase, halos of the  phase formed around the primary  dendrites. As the halos grown, the composition of the liquid at each halo/liquid interface moved towards the coupled zone and coupled growth commenced once it reached the boundary of the coupled-zone. Gigliotti et al. [[13]] and Suk et al. [9,[14]] developed some models for explaining not only the general phenomenon, but also a few exceptional cases often observed: non-faceted halo surrounded non-faceted primary phase, and faceted halo surrounded faceted primary phase.

The growth-dominant theory was further developed extensively. Yilmaz et al. [3] gave the experimental evidence in Al-Si alloys, in which -Al halos surrounded the primary Si. Barclay et al. [[15]] developed a model for the formation of non-faceted halos around a faceted primary phase in directional solidification, in which the liquidus of the faceted phase was assumed to have a significant depression due to kinetic undercooling. Once the composition of the liquid adjacent to the primary phase reached the liquidus of the non-faceted phase, the growth of the non-faceted  halo caused the composition of the liquid at the halo–liquid interface to follow the liquidus of the non-faceted  phase until the undercooling for coupled eutectic growth was reached. Suk et al. [[16],[17]] developed a model of halo formation that assumed competitive growth between the halo phase and coupled eutectic in liquid with a composition equal to the coupled zone boundary on the primary phase side. The analysis by Sharp and Flemings [[18]] assumed that solute diffusion in the liquid state was sufficiently rapid, and therefore any concentration gradients perpendicular to the growth direction were negligible. Recently, Nave et al. [[19]] and Li et al. [[20]]developed models to describe the halo formation in terms of competitive growth between the halo phase and coupled eutectics in liquid having a nominal composition that follows the primary phase liquidus extension with decreasing temperature. Liet al.[20] found that no halo structure could be formed in directional solidification of Sn-Pb eutectic alloy, but α-Al halo structure existed at the composition between 12.6wt.% and 25wt.% Si in Al-Si alloys.

On the basis of the investigations summarised above, the characteristics and conditions for the halo formation under conventional solidification in both equiaxed solidification and directional solidification is still not fully understood. There are still many open questions surrounding the mechanism for the halo formation. In conventional solidification, if the growth process can be isolated during solidification, the halo formation can be subsequently identified. The halo formation is the essential result of simultaneous interaction of three phases: the primary α phase, the halo β phase and the melt encompassing the solid phases. There is a still lacking of information explaining the phase interaction and the resultant effect on halo formation. Meanwhile, the observed halos are always in the surroundings of the primary phase formed above eutectic temperature. There is no evidence to show the halo characteristics on solid phase when the solidification is under intensive convection which leads to a significant alternation of the solidification process, especially for the alloys at eutectic and near eutectic composition. It is therefore desirable to investigate the effect of intensive convection on the halo characteristics as it can completely redistribute the solute concentration at solid/liquid interface during solidification.

Therefore, this study aimed to explore the halo formation inZn-Al alloys solidified under different solidification conditions. By using the recently developed twin-screw shearing [[21],[22]], the effect of intensive convection on the halo formation was fully investigated on three alloys with hypoeutectic, eutectic and hypereutectic compositions. Simultaneously, the halo formation of these alloys was also studied under conventional solidification for comparison. Finally, the reason was discussed based on the effect of intensive convection on the halo characteristics and the effect the interaction of different phases, especially the strain energy created at the interface during solidification on the halo formation.

2.Experimental

Pure elemental raw materials Zn (99.95wt.% purity) and Al (99.995wt.% purity) were used to make Zn-Al alloy ingots containing 3-7wt.%Al. The crucible was heated up to 700oC and then Al was added and melted, before adding Zn into the crucible whilst reducing the temperature of furnace to 250oC. To complete melting and ensure solute homogenisation, each alloy was held at 450-500oC for 30min before casting into a copper mould. The chemical composition was analysed by the average value of 5 burnings on a cross section of a 40mm specially made casting by optical mass spectroscopy. One of these castings was metallographically examined in order to observe the resultant microstructures to confirm alloy homogeneity. The ingots were re-melted in an electric-heated furnace to a temperature 50oC above their liquidus and held for about 20-30 min before use.

In the experiments without melt shearing and under conventional solidification, the melt were poured into a wedge-shape copper mould at a fixed superheat and solidification occurred subsequently without further disturbing. The casting was cut at different thickness for microstructural examination.

To examine the effect of intensive convection, a 16mm twin-screw shearing devicewas used as the mechanism to produce shear for the current investigation. The twin-screw shearing mechanism is provided by a pair of self-wiping and full meshing co-rotating screws that can be rotated at variable speed up to 1500 rpm. The diagram of experimental setup is shown in Figure 1. The detail description for the twin screw shearing mechanismhas been described in ref. [21,[23]]. The shear rate used in this paper referred to that developed between the inner surface of barrel and the tip of screw flight and is calculated by the equation [23]:

(1)

Where ω is the rotation speed of screw, Dis the outer diameter of screw and δis the gap between the tip of screw flight and the inner surface of barrel. The shear rate is used as a measure of both the intensity of shear and turbulence and a measure of the forced convection.

In the experiments with melt shearing, the temperature control was achieved by control the barrel temperature with an accuracy of 1oC. The extruder was heated up to a temperature above the liquidus of alloys and run at a given rotation speed. The Zn-Al melt was poured into the twin-screw extruder at a temperature 10oC above its liquidus. The melt immediately experienced shearing whilst passing through the gap between the screws and barrel. The melt was isothermally sheared for about 30s before being cooled down continuously with a controlled cooling rate. When the sheared melt reached the sampling points, the melt temperature was recorded by a data logger that had an accurate of 0.1oC at an interval of 0.2s and the melt was tapped into a cold-water tank for quenching. The samples were then examined individually by microscopy. The temperature, where one kind of solid phase was observed in the water-quenched specimens, was defined as nucleation temperature of primary phase. The temperature, where two kinds of solid phaseswere observed in the water-quenched specimens, was defined as nucleation temperature of eutectics.

The microstructure examination of each alloy was carried out using an optical microscope equipped with Zeiss KS300 3.0 image analysis system with quantitative metallography and a JEOL JXA-840A scanning electron microscope (SEM) (JEOL Ltd., Tokyo, Japan). The halo thickness was measured by the image analysis system based on the mean intercept length of at least of 20 segments for each experimental data.

3.Results

3.1. Halo characteristics under conventional solidification

In the conventional solidification experiments, the alloys were poured into a wedge shape copper mould at 10oC above its liquidus temperature and solidified subsequently without disturbing. Figure 2showed that Zn-rich halos were formed around the primary Al-rich phase. It was obvious that three types of alloys exhibited different characteristics. In the hypereutecticZn-6.3wt.%Al alloy, the primary Al-rich phase was surrounded by a halo of Zn-rich phase, andthe eutectic microstructure was clearly a lamellar morphology as shown in Figure 2a. In the eutectic Zn-5wt.%Al alloy, the eutectic microstructure showed lamellar eutectic cell morphology all over the matrix (Figure 2b). There was no halo in this microstructure. In the hypoeutectic Zn-4.1wt.%Al alloy, the primary Zn-rich phase showedno halo in the surroundings of Zn-rich dendrites as displayed in Figure 2c. The results confirmed that halos only occurred on the primary Al-rich phase with Zn-rich halo at the hypereutectic Zn-Al alloys with Al>5wt.%, and no halos were formed for the hypoeutectic and eutectic Zn-Al alloys at Al ≤5wt.%.

Furthermore, the thickness of the halos was measured by mean intercept length at different cooling rates in the Zn-6.3 wt.%Al alloy. Figure 3shows the relationship between the cooling rate and the thickness of the halos in the Zn-6.3wt.%Al alloy solidified in a wedge-shape copper mould. The results obtained by Bluni et al. [7] with Zn-6wt.%Al alloy from a differential scanning calorimetry (DSC) testing samplewas also shown in the figure for comparison. The cooling rate of wedge casting was estimated with the aid of measurements of the cooling curve shown in Figure 4. It was in very good agreement with similar measurements by Pryds et al. [[24]], Curiotto et al. [[25]] and Hildal et al. [[26]]. The results confirmed that the halos became thinner with increasing cooling rates. It was notable in Figure 3 that the halo thickness from this work was larger than those observed by Bluni et al. at the same cooling rate [7]. The difference might be related to the difference in the experimental methods. It was well known that the physical volume of samples for DSC testing wassmall and DSC tests offered a more precise control of the cooling rate. But for a wedge shape casting, the quantity of melt was more than thousands times of the quantity used for DSC testing. Thus accurate measurement of temperature within the wedge casting was difficult, especially in the very thin section. Meanwhile, the interaction in the bulk melt of the wedge casting might be significantly higher than that in DSC testing, leading to the variation of halo thickness between the experiments. The decrease of the halo thickness with increasing the cooling rate was a result of a larger undercooling during solidification at the higher cooling rate. The higher cooling rate could result in a finer primary phase in the casting, which was corresponded to thinner halos accordingly. This observation was contrary to the prediction by Sundquist et al. [10] and Salli et al. [11]. One reason for the difference might be caused by their ignorance of the microstructural refinement at high cooling rate.

3.2. Halo characteristics underintensive convection solidification

In order to describe the whole solidification process under intensive convection, the solidification process was deliberately described in three steps for convenience in explanation. The detailed description was shown in Figure 5. During the solidification under intensive convection, the solidification of hypoeutectic (Al<5wt.%) and hypereutectic (Al>5wt.%) alloys to form one solid phase in the water quenched sample was defined as primary solidification. In the case of eutectic reaction, two solid phases formed simultaneously and the solidification in the water quenched sample was defined as eutectic solidification. Post solidification was defined as the solidification that occurred from the exit of the barrel of melt shearing to the fully solidified sample within the water tank used for quenching. As the sheared melt was directly discharged from the shearing device into the water tank for quenching, the post solidification would happen within a very small time gap.

In order to determine the nucleation and growth temperature of the primary phase and the eutectic reaction, a number of specimens were metallographically investigated. The temperatures were measured by the thermocouple in direct contact with the melt.A typical cooling curve of Zn-5wt.%Al alloy under intensive shearing at 4082s-1 and a cooling rate of extruder at 1oC/min, and the measured nucleation temperatures for eutectic and near-eutectic Zn-Al alloys have been published in previous reference [23]. They indicated that no apparent recalescence was found in the experimental conditions and the nucleation temperatures of Zn-Al alloys exposed to intensive convection were very close to the temperature shown in the equilibrium phase diagram.For hypoeutectic and hypereutectic Zn-Al alloys, the primary phase emerged within 0.5oC below the equilibrium liquidus temperature. Both Al-rich phase and Zn-rich phase were observed with similar behaviour during the primary solidification. And, the eutectic solidification occurred at 380.6oC for the Zn-5wt.%Al eutectic alloy and 380.5oC for the hypereutectic Zn-6.3wt.%Al alloy and the hypoeutectic Zn-4.1wt.%Al alloy, respectively.The variation of composition did not alter the behaviour during the eutectic solidification. The repeatable results produced by the twin-screw shearing mechanism confirmed that the solidification during intensive convection required small undercooling for the nucleation of both the primary phase and the eutectic phases. This implies that the solidification under intensive convection was close to equilibrium solidification. This observation did not agree with the observations during conventional solidification by Bluni et al. [7]. They found a significant variation of the undercooling of hypoeutectic and hypereutectic Zn-Al alloys during eutectic solidification without convection.