Metallurgical and Materials Transactions A

, Volume 42, Issue 8, pp 2456–2465

Favorable Environment for a Nondendritic Morphology in Controlled Diffusion Solidification


  • Abbas A. Khalaf
    • The Light Metal Casting Research Centre (LMCRC), Department of Mechanical EngineeringMcMaster University
    • The Light Metal Casting Research Centre (LMCRC), Department of Mechanical EngineeringMcMaster University

DOI: 10.1007/s11661-011-0641-z

Cite this article as:
Khalaf, A.A. & Shankar, S. Metall and Mat Trans A (2011) 42: 2456. doi:10.1007/s11661-011-0641-z


The novelty of the controlled diffusion solidification (CDS) process is the mixing of two precursor alloys with different thermal masses to obtain the resultant desired alloy, which is subsequently cast into a near-net-shaped product. The critical event in the CDS process is the ability to generate a favorable environment during the mixing of the two precursor alloys to enable a well-distributed and copious nucleation event of the primary Al phase leading to a nondendritic morphology in the cast part. The turbulence dissipation energy coupled with the undercooling of the precursor alloy with the higher temperature enables the copious nucleation events, which are well distributed in the resultant mixture.


Alloy 1

precursor alloy with higher thermal mass (higher temperature and higher mass)

Alloy 2

precursor alloy with lower thermal mass

Alloy 3

resultant desired alloy

C1 and C2

average solute concentration of Alloys 1 and 2, respectively


transient solute concentration (Cu)




characteristic length

m1, m2, and mT

mass of Alloy 1, Alloy 2, and total mass, respectively


Weber number


solid-liquid interface

T1 and T2

temperatures of Alloys 1 and 2, respectively


actual temperature


temperature of point B

TL1, TL2, TL3

liquidus temperatures of Alloys 1, 2, and 3, respectively


melting temperature


quenching temperature




transient instantaneous position




dynamic viscosity


surface tension

1 Introduction

A nondendritic morphology of the primary phase in a solidified binary alloy could be achieved by controlling the solute redistribution and thermal fields in the solidifying phases both individually and together. The first such efforts were in the rapid solidification process wherein a nondendritic (cellular) microstructure could be formed in the solidified part, because the extraordinarily high growth rates experienced by the solid-liquid interface lead to a stable growth of the same.[1] In the latter part of the twentieth century, the thixoforming and rheocasting processes were developed as casting methods wherein a nondendritic morphology of the primary phase could be obtained in the cast part with a binary eutectic alloy along with the benefits in improved mechanical properties and performance. Unfortunately, these processes have proven prohibitive in capital and manufacturing costs.[27] Aluminum alloys have been the most significantly researched material to obtain a nondendritic cast microstructure, and near-net-shaped casting of Al wrought alloys along with their superior properties and performance have been a challenge for conventional casting routes due to the main disadvantage of hot tearing or hot cracking during solidification as caused by the presence of a pronounced dendritic network,[8] which renders the cast components ineffective. To circumvent the disadvantages of thixoforming and rheocasting processes and to enable a cost-effective near-net-shape casting of Al wrought alloys, the controlled diffusion solidification (CDS) process was innovated to enable casting aluminum alloys with a nondendritic morphology of the primary Al phase in the resultant cast microstructure by circumventing the problem of hot tearing and obtaining a product with improved mechanical properties.[9,10]

The CDS is a simple process involving the mixing of two precursor alloys of different thermal masses (temperature and solute content) and, subsequently, casting the resultant mixture as a near-net-shaped cast product. The process lends itself to easy commercialization with a marginal capital cost required for set up of an additional melt holding furnace. Further, the CDS process would prove itself to be unique in its ability to cast Al-based wrought alloys into near-net-shaped components without additional processes and cost.[11,12] The success of the CDS process in yielding a nondendritic primary Al phase morphology stems from the initial favorable environment for such an event created during the mixing of the two precursor alloys: copious nucleation and forced convection. In this article, evidence is presented to elaborate the specific thermal and solute conditions occurring during the mixing process that leads to such a favorable environment. As an example alloy, the Al-4.7 wt pct Cu alloy was solidified by CDS in a laboratory experiment scale in this study. Figure 1 shows the binary phase diagram of the Al-Cu system as simulated by the thermodynamic software, PANDAT1, with the two precursor alloys marked as Alloys 1 and 2, respectively, and the desired resultant alloy as Alloy 3.
Fig. 1

Binary phase diagram of Al-Cu system

In Figure 1, Alloy 1 would be pure Al, Alloy 2 would be Al-33 wt pct Cu, and Alloy 3 would be the desired Al-4.7 wt pct Cu alloys.

2 Experiments

Controlled CDS experiments were carried out in a laboratory setting such that two precursor alloys could be mixed and allowed to solidify in a crucible. Two rates of solidification were carried out: a slower rate wherein the mixed resultant alloy was allowed to solidify in a crucible placed on a table and a higher rate of solidification wherein the mixture of resultant alloy was quenched on a rotating copper wheel in ribbon casting equipment so as to interrupt the solidification process at a predetermined time-step. The purpose of using an extraordinarily high growth rate with the ribbon casting equipment was to effect a nearly complete solute trapping, leading to a diffusionless transformation, thereby arresting the solidification process at the predetermined time-steps to investigate nucleation and growth of the primary Al phase.[13,14]

Figures 2(a) and (b) show a schematic and photograph of the experimental apparatus for the laboratory CDS process, respectively, wherein Alloy 1 was taken in a crucible with a hole (9-mm diameter) at the bottom to which a funnel (9-mm diameter) was attached with a stopper blocking the hole from the inside, and Alloy 2 was taken in a second crucible, which was placed directly under the first crucible. The temperatures of Alloys 1 and 2 were controlled and monitored continuously by three exposed K-type thermocouples (0.62-mm diameter) connected to a data acquisition system SCXI-11002. When the desirable temperatures of T1 and T2 were attained, the stopper in the top crucible was lifted to enable Alloy 1 to mix into Alloy 2 in the bottom crucible.
Fig. 2

Setup of laboratory experiments for the CDS process: (a) schematic and (b) photograph. Also shown are the locations of three thermocouples to control and monitor the temperatures during the process

Two electric resistance furnaces were used to melt Alloys 1 and 2. Alloy 2 was taken out from the furnace at a superheat of about 323.15 K (50 °C) above its liquidus temperature, and the two thermocouples Tu and TL were inserted in their respective locations. The funnel and stopper were heated along with Alloy 1 as well. The funnel and stopper were removed from the furnace and installed in their respective locations and the thermocouple T1 was fitted as well, as shown in Figures 2(a) and (b). Alloy 1 was taken out of its furnace and poured in the top crucible, and consequently, the stopper was lifted when temperature T1 was reached for Alloy 1 to mix into Alloy 2. The temperatures were continuously recorded until the end of the process. The experiments were repeated several times to ensure repeatability. The composition of Alloy 3 was designed to be 4.71 wt pct Cu, and the average mixing rate was found experimentally to be 167 g/s. Table I presents the design of experiments including the independent parameters and constants. The dependent parameter in these experiments was the morphology of the primary Al phase in the solidified microstructure. The mixture in these experiments was allowed to solidify in the bottom crucible without disturbing the experimental setup. In order to compare the difference in the microstructure between the CDS process and conventional solidification process, the experiment designated as CONV1 in Table I was carried out by remelting a portion of the solidified casting from experiment CDS1 and poured at 936.15 K (663 °C) through the same funnel into the empty crucible, which was already preheated to 828.15 K (555 °C) using the same apparatus as shown in Figure 2(a). In the experiment designated as CONV2 in Table I, Al-4.7 wt pct Cu was melt in a crucible at 953.15 K (680 °C) and subsequently allowed to solidify in still air.
Table I

Experiment Design with Independent Parameters and Constants for the Laboratory CDS Experiments


Alloy 1

Alloy 2

Rate of Mixing (g/s)


m1 (g)

T[K (°C)]

TL[K (°C)]

m2 (g)

T[K (°C)]

TL[K (°C)]



935.15 (662)

933.1 (660)


828.55 (555.4)

827.15 (545)





938.15 (665)


828.55 (555.4)




943.15 (670)


827.15 (554)




948.15 (675)


828.15 (555)




956.15 (683)


826.15 (553)



remelt Alloy 3 to 936.15 K (663 °C) from CDS2 and pour into empty crucible at 828.15 K (555 °C) to solidify


solidify Al-4.7 wt pct Cu alloy from 953.15 K (680 °C) in a crucible

The solidified samples were suitably sectioned and mounted using black phenolic (Black Phenolic, MetLab Corporation, Niagara Falls, Canada) and were ground and polished in automatic polishing machines. Etching was carried out on some samples with freshly prepared Keller’s reagent (1 mL HF, 1.5 mL HCL, 2.5 mL HNO3, and remainder H2O). The microstructure analysis was carried out using the stereo, light optical, and scanning electron microscopes. A stereo microscope type NIKON AZ 100 M3 was used along with image analysis software.

To identify the nucleation events during the mixing process, the resultant mixtures of the CDS2, CDS3, and CONV2 experiments shown in Table I were quenched at about 104 °C/s in ribbon casting equipment.[15] The ribbon casting process equipment shown in Figure 2 was used in this study to quench the CDS mixture at a very high quenching rate so as to study the nucleation of the primary Al phase during various stages of the CDS process.

A few modifications to the existing ribbon casting equipment were carried out to enable the quenching of a mixture from the CDS process. The experimental setups with the modifications are explained subsequently.
  1. (1)
    The copper wheel shown in Figure 3(b) is used to quench the mixture. The wheel was rotated at 1100 rpm in all the quenching experiments.
    Fig. 3

    Photographs of equipment setup and solidified samples for the interrupted quench experiments with the ribbon casting process: (a) ribbon casting equipment with the two electric furnaces to melt Alloys 1 and 2; (b) Cu wheel, induction coil, and steel tube; (c) steel tube and ceramic funnel; and (d) typical solidified ribbon samples

  2. (2)

    A long steel tube (423 mm) of 21.12-mm inner diameter was fixed with one end at about 3 mm above the copper wheel and the other end coming out at the top of the chamber, as shown in Figure 3(c). The end of the tube closest to the Cu wheel was closed with steel, and a 5-mm-diameter orifice was drilled in the center to enable the flow of the mixture onto the wheel. A ceramic funnel with a 7-mm diameter of outlet orifice was fitted on the top end of the steel tube, as shown in Figure 3(c), and the mixture was poured into this funnel. The total time between pouring the mixture into the funnel to the first liquid in contact with the wheel was about 0.3 seconds.

  3. (3)

    An induction copper coil shown in Figure 3(b) was used to heat the bottom of the steel tube to 973.15 K (700 °C) to prevent any appreciable heat loss from the melt mixture.

Two electric furnaces were used to melt Alloys 1 and 2. The funnel was heated with Alloy 1, and it was first removed from the furnace and fitted in its respective location shown in Figure 3(c). Alloys 1 and 2 were taken out of their respective furnaces, and Alloy 1 was mixed into Alloy 2; consequently, the mixture was poured into the funnel after 2.8 seconds of mixing. Table II presents the design of the experiments along with the independent parameters and constants for all the interrupted quenching experiments with the ribbon casting process. The solidified samples in the form of small broken ribbons were collected in the long sealed chamber shown in Figure 3(a). Further, the experimental designation CONVQ in Table II was the Al-4.7 wt pct alloy quenched at a 278.15 K (5 °C) superheat temperature above the liquidus temperature. Figure 3(d) shows the solidified ribbon samples produced from the ribbon casting process. The ribbons were mounted using black phenolic and ground and polished in automatic polishing machines. The scanning electron microscope (SEM) analysis was carried out on mounted samples with a JEOL4 JSM-7000F equipped with an energy dispersive X-ray (EDX) analysis system.
Table II

Experimental Design with Designation, Independent Parameters, and Constants for the Interrupted Quenching Experiments with the Ribbon Casting Process


Alloy 1

Alloy 2

Time After Mixing (s)

m1 (g)

T[K ( °C)]

TL[K ( °C)]

m2  (g)

T[K ( °C)]

TL[K ( °C)]



938.15 (665)

933.15 (660)


828.1 (555)

818.15 (545)




942.15 (669)






Cu (wt pct)






929.15 (656)




922.15 (649)


3 Results and Discussion

Figure 4(a) shows the typical thermal curve obtained during a successful CDS experiment, as proposed by Khalaf et al.[16,17] In Figure 4(a), there are several critical thermal events demarcated by the three distinct segments: AB, BCD, and DE. The segment AB is when the two alloys mix, causing the favorable environment of copious nucleation and forced convection. The segment BCD is when the unique redistribution of the thermal and solute occur along with the cells in the liquid, setting the stage for nondendritic growth morphology of the primary Al phase.[18] Point D is when the final nucleation event occurs and leads to a nearly stable growth of the primary Al phase through the segment DE. A detailed description of the events in the various stages is presented by Khalaf et al.[16,17] The favorable environment for the copious and well-distributed nucleation events of the primary Al phase occurs in the segment AB. Figure 4(b) shows the actual thermal data obtained from thermocouple TL in Figure 2(a) wherein point B changes with the superheat temperature of Alloy 1.
Fig. 4

Details of the CDS process: (a) schematic of the typical thermal data obtained during the CDS process and (b) thermal data from the CDS experiments and conventional experiment

Figure 5(a) shows the transient temperature data obtained from the mixture during the mixing process as a function of the solute (Cu) concentration in the resultant mixture that changes with the time of mixing. Alloy 1 with a liquidus temperature of 933.15 K (660 °C) would initially be introduced to Alloy 2, which is at a significantly lower temperature of 828.15 K (555 °C), resulting in an undercooled environment. The Weber number, NWe, is described as the ratio between the inertial forces to the surface tension of the liquid,[19] and this number described the propensity of the breakup of a flowing stream into smaller pockets of liquid. The higher the value of NWe, the higher is the propensity of breakup of the stream. Equation [1] presents the equation to evaluate NWe wherein ρ is 2375 kg·m−3,[20] L is 0.009 m, u is 1.1 m·s−1, and σ is 0.868 N·m−1.[11]
$$ {\text{N}}_{\text{We}} = {\frac{{\rho {\text{Lu}}^{2} }}{\sigma }} $$
Fig. 5

Thermal data from CDS laboratory experiments: (a) transient temperature data during mixing as a function of Cu concentration and (b) TB as a function of T1

The value of NWe was evaluated as 30, which is far greater than unity, suggesting that the inertial forces during mixing are far greater than the surface tension forces, which would result in a significant breakdown of the stream of Alloy 1 entering the mixture and distributing the Alloy 1 as small pockets of liquid.[21] Figure 5(a) shows that at any time during the mixing process, the temperature of the resultant mixture would be significantly lower than the liquidus temperature of Alloy 1; further, since the time of mixing is merely a couple of seconds long, the inter-diffusion of the species between Alloys 1 and 2 would be insignificant. Hence, the small pockets of Alloy 1 would experience a large undercooled environment leading to a nucleation event of the Al phase from the various pockets of liquid Alloy 1. Since these pockets of liquid are numerous and well distributed in the resultant mixture, there would be copious nucleation events in the resultant mixture during mixing and this would lead to a favorable environment for the stable growth of the nuclei leading to a nondendritic morphology of the primary Al phase in the resultant casting.

Figure 5(b) shows the temperature of point B, TB, as a function of the T1 wherein with increasing values of T1, TB initially decreases and subsequently shows a gradual increase. The value of TB depends on four parameters: rate of mixing, heat added into the resultant mixture from the enthalpy of Alloy 1, heat added to the mixture by the heat of fusion resulting from the nucleation events of primary Al during mixing, and heat loss to the environment from the resultant mixture. The rate of mixing and heat loss to the environment would be nearly constant for a specific experimental design. Hence, the variation in the value of TB would primarily be due to the variation in the heat added to the mixture during mixing from the enthalpy of Alloy 1 and the nucleation reaction, alike. If there were no nucleation events of the primary Al phase from Alloy 1 in the resultant mixture during the segment AB, TB would continuously increase with increasing values of T1. If there are nucleation events in the resultant mixture during mixing, then enthalpy of fusion would be released into the mixture from these events and, thereby, increase the heat content of the mixture, leading to an increase in the value of TB. The higher the number of nucleation events, the higher would be TB; and the amount of nucleation events would increase when the temperature T1 decreases for a constant value of T2, because the lower superheat of Alloy 1 would result in a higher number of nucleation events for the same rate of heat extraction by Alloy 2 during mixing. In Figure 5(b), the value of TB decreases between the superheat temperatures of 275.15 K and 278.15 K (2 °C and 5 °C) because of the copious nucleation events at the lower superheat, adding heat of fusion to the mixture to result in a higher value of TB. After the superheat of 278.15 K (5 °C), the value of TB steadily increases due to the high enthalpy of Alloy 1 with increasing superheat temperatures and nominal nucleation events.

Numerical simulations were carried out with the computational fluid dynamic software, Flow 3D (Flow Science Inc., Santa Fe, NM), to evaluate the transient density and velocity distribution in the resultant mixture during the mixing process. The simulation domain and boundary conditions were similar to the experiment setup shown in Figure 4(a). Figure 6(a) shows a snapshot image after 1 second of the mixing process, showing the transient density distribution in the mixture during the mixing process wherein the density of the resultant mixture is fairly uniform except in the stream of Alloy 1 entering the mixture. Figure 6(a) shows that mixing in the CDS process results in a nearly homogenous composition of the resultant mixture almost instantaneously; however, this homogeneity would only be in the macroscale (~1 mm). Figure 6(b) shows the transient velocity distribution in the mixture wherein a significant gradient in velocity is observed throughout the resultant mixture caused by the large turbulent dissipation energy generated during the mixing process. This would result in a significant breakdown of the stream of Alloy 1 entering the mixture. A mixture of water with blue ink was mixed through the same experimental setup as in Figure 4(a) into pure water to visualize the stream of liquid entering the mixture during the CDS process. Water was used because of similarities in the fluidity with Al[22] and the transparency of the liquid. Figure 6(c) shows snapshot photographs obtained from the digital video recording obtained during the mixing of the different waters wherein the breakup of the stream of water with blue ink (shown by circles in Figure 6(c)), as it enters the resultant mixture, and the distribution of the broken pockets of blue liquid in the resultant mixture could be easily observed. The broken pockets of the blue liquid are similar to that from Alloy 1 in the CDS experiments, which would lead to the favorable environment of copious and well-distributed nucleation events in the resultant mixing of water with blue ink mixed into water.
Fig. 6

Snapshot photographs (after 1 s) taken during simulation of the mixing process: (a) transient density distribution, (b) transient velocity distribution, and (c) mixing of water with blue ink mixed into water

Figures 7(a) through (c) show typical low-magnification microstructure images of the surface of the quenched ribbons for the CDSQ1, CDSQ2, and CONVQ experiments, respectively; wherein the nucleation of the Al phase during the mixing process highlighted by the arrows was clearly discernable in the CDSQ1 and CDSQ2 experiments and was absent in the CONVQ experiment showing evidence of the copious nucleation events in the CDS process.
Fig. 7

Typical low-magnification microstructure images obtained from the surface of the rapidly quenched sample: (a) CDSQ1, (b) CDSQ2, and (c) CONVQ. The arrows in (a) and (b) show the nucleated primary Al phase during mixing

The samples from the interrupted quenching experiments were further analyzed in a SEM equipped with an EDX analyzer. Figure 8(a) shows the secondary electron image microstructure of the typical Al-phase particle shown in Figure 7(a) that nucleated during the mixing process of the CDS. Figure 8(b) shows the line scan profile obtained using the EDX analysis for the line drawn across the microstructure in Figure 7(a). The composition of Cu at any location in the blocky Al-phase particle is nearly zero, thus confirming that these phase particles nucleated from Alloy 1 (pure Al). If they were to nucleate from Alloy 2 or the resultant mixture, there would be a segregation of Cu across the primary Al phase and the average composition would reflect the respective Cu concentration, as dictated by the phase diagram in Figure 1.
Fig. 8

Typical high-magnification SEM images obtained from the surface of the rapidly quenched sample: (a) magnified image of the Al phase in the surface of the CDSQ2 sample and (b) line scan profile of Cu concentration for the line in Fig. 7(a)

The nucleation during the mixing process would be that of the primary Al phase from the pockets of liquid Alloy 1 due to the large undercooling experienced by them. The environment faced by these nuclei would be unique and is shown by the schematic of the transient temperature and Cu concentration profiles ahead of the growing S/L interface in Figure 9 for progressive time-steps during growth, respectively. In Figure 9, the progressive time-steps are shown by the numbers 1, 2, and 3 for the respective profiles of TL, Tactual, and CCu; and an undercooling of the actual temperature below the respective liquidus temperature exists between the S/L interface and the point denoted by ξ in the domain ahead of the growing S/L interface. In Figure 9, at any time-step, the CCu in the liquid increases ahead of the boundary of the liquid Al droplet, of pure Al concentration, into the liquid due to the high Cu concentration in the resultant alloy mixture. This Cu profile would define the profile of TL, and these two profiles would change with time due to the transfer of the Cu solute toward the growing nuclei caused by diffusion and advection, alike, resulting in a gradual decrease of the pure Al boundary in the liquid pocket. The profile of Tactual would also change with increasing time-steps due to the heat extraction from the growing nuclei caused by the low temperature of the mixture. The intersection point of the TL and Tactual ahead of the S/L interface denoted by ξ in Figure 9 denotes the extent of undercooling of liquid below the liquidus temperature. The interplay between the solute transport toward and heat flux away from the S/L interface would result in relative movement of the point ξ toward the S/L interface, causing a decreasing undercooled environment of the liquid ahead of the interface and resulting in an increasingly stable growth of the S/L interface. The nuclei would initially face a liquid at a lower temperature than the liquidus temperature, which would initially be the melting point of pure Al (Tm). The gradient of the actual liquid temperature ahead of the S/L interface would continue to decrease, but the liquidus temperature would still be at Tm due to the slower mass transport than temperature transport. There would be an initial perturbation of the S/L interface during growth showing tendencies of instability in the growth. However, the decreasing temperature gradient in the liquid ahead of the S/L interface will decrease the extent of instability of this interface, because the extent of undercooling decreases, as shown by point ξ in the time-steps 1 and 2 in Figure 9. Hence, the growth of primary Al phase would be nearly stable with a marginally perturbed interface, and the perturbation increases with increasing rates of heat extraction from the S/L interface (increasing cooling rate of the casting). At increasing time-steps, in Figure 9(a), the point ξ would move relatively closer to the S/L interface; and when these two intersect, the growth would cease. In Figure 9, the point of intersection between the S/L interface and point ξ is denoted by the profiles for time-step f, assuming the worst case scenarios that the actual temperature of the liquid would equalize prior to the cessation of the growth. In Figures 8(a) and (b), the profiles show that the Cu concentration is nearly zero in the middle of the grain due to the nucleation from the pure Al in Alloy 1 and growth in the droplets of pure Al; and toward the outer regions of the grain, the Cu concentration gradually increases until the end of solidification of the grain effected by the loss of undercooling in the liquid ahead of the growing S/L interface, as explained previously with Figure 9. The Cu profile shown in the schematic of Figure 9 for the time-step f from the S/L interface to the boundary defined by the point ξ for this time-step would match the one in the line scan profile obtained from the primary Al phase in Figure 8(b). The growth of the nuclei formed during the mixing process in the microstructure is unique in that the S/L interface grows in the direction of the heat flux; further, the solute transport is toward the S/L interface and the heat flux is away from the same into the liquid ahead of the interface. Such thermal and solute redistribution directions are exactly opposite to those observed in conventional solidification processes.[23]
Fig. 9

Schematic of the proposed transient thermal and solute fields during the growth of the primary Al nuclei formed during mixing in the CDS process

Figures 10(a) through (g) show the typical microstructure from an optical microscope of the final solidified samples for CDS and conventional experiments presented in Table I, revealing the nondendritic morphology of the primary Al phase forming in the CDS samples that mixed at lower Alloy 1 temperatures (935.15 K, 938.15 K, and 943.15 K (662 °C, 665 °C, and 670 °C)), as shown in Figures 10(a), (b), and (c), respectively, whereas the microstructure of CDS samples changes to a rosette morphology mixed with dendritic and to fully dendritic as the temperature of Alloy 1 increases, as shown in Figures 10(d) and (e), respectively. The morphology of the Al phase in samples from CONV1 and CONV2 were dendritic, as shown in Figures 10(f) and (g), respectively. The nondendritic morphology of the primary Al phase obtained from the CDS process is different than that from a rapid solidification process. In the latter, the favorable morphology (cellular) would be obtained by causing a nearly complete solute trapping scenario during the high growth rate of the primary phase, leading to a diffusionless (solute) transformation with a stable solid-liquid interface, whereas in the CDS process, the controlled and favorable redistribution of the solute and thermal field in the resultant mixture of the Alloy is critical to effect a nondendritic morphology of the primary Al phase.
Fig. 10

Typical microstructure for samples made via CDS process and conventional solidification presented in Table I: (a) CDS1 (nondendritic), (b) CDS2 (nondendritic), (c) CDS3 (nondendritic/rosette), (d) CDS4 (rosette), (e) CDS5 (dendritic), (f) CONV1 (dendritic), and (g) CONV2 (dendritic)

4 Summary

The results shown in Figures 4 through 10 demonstrated the existence of a unique and favorable environment for copious and well-distributed nucleation events in the resultant mixture of the desired alloy composition caused by mixing an alloy with a higher thermal mass into one with a lower thermal mass. The copious nucleation formed during and after the mixing process along with the thermal and solute redistribution would make a favorable environment around the growing nuclei to increase the stability of the solid liquid interface and result in a predominantly nondendritic microstructure, which is favorable to improve the mechanical properties due to the decrease of the hot tearing and porosity defects associated with dendritic morphology in conventiontory experiments. The critical condition for copious nucleation taking place is by setting up the first precursor alloy (Alloy 1) temperature closer to its respective liquidus temperature.

A simple analogous example could be that of making coffee wherein if the higher thermal mass of black coffee is mixed into a lower thermal mass of cream at nearly room temperature, one would observe a nearly uniform mixture of coffee without the aid of stirring, whereas if the colder cream is mixed into the hotter coffee, stirring of the coffee is mandatory to obtain a nearly uniform mixture.

The observations of this study could also be used to develop novel methods of obtaining semisolid slurries in slurry ready semisolid metal casting processes such as the new rheocasting process, continuous rheoconversion process, subliquidus casting process, semisolid rheocasting process, and rheo die-casting process.[24] In all these semisolid casting processes, a large amount of capital and processing costs are required to obtain nearly homogeneous thermal and solute fields in the semisolid slurry mixture prior to the casting process. CDS would enable such a mixture with minimal cost and use the nature of liquid mixing to obtain a favorable semisolid slurry material.


PANDAT 8.1 is a trademark of CompuThermLLC, Madison, WI.


SCXI-1100 is a trademark of National Instruments, Vaudreuil-Dorion, PQ, Canada.


NIKON AZ 100M is a trademark of Eberbach Corporation, Ann Arbor, MI.


JEOL is a trademark of Japan Electron Optics Ltd., Tokyo.



The authors acknowledge the financial support provided by the Natural Science and Engineering Research Council (NSERC) of Canada through their Discovery Grant Program.

Copyright information

© The Minerals, Metals & Materials Society and ASM International 2011