An Article from the January 2002 JOM-e: A Web-Only Supplement to JOM

The authors of this article are with the Laboratoire de Metallurgie Physique at Ecole Polytechnique Federale de Lausanne.
Exploring traditional, innovative, and revolutionary issues in the minerals, metals, and materials fields.





















Casting Process Simulation and Visualization: Overview

Hot Tear Formation and Coalescence Observations in Organic Alloys

P.-D. Grasso, J.-M. Drezet, and M. Rappaz


Hot tear formation has been observed during the solidification of a succinonitrile-acetone (SCN-acetone) alloy by pulling the columnar dendrites in the transverse direction with a stick. The cracking of the mushy zone (hot tears) always occurs at grain boundaries. At low volume fraction of solid, the opening can be compensated for by leaner-solute interdendritic liquid (i.e., "healed" hot tears). At higher volume fraction of solid, hot tears directly nucleate in the interdendritic liquid or develop from pre-existing micropores or air bubbles induced by solidification shrinkage. Moreover, coalescence/bridging of dendrite arms has been carefully observed and the temperature at which this occurs has been measured to determine the corresponding solid fraction. It is observed that coalescence between columnar dendrites inside a grain (intragranular coalescence) occurs at a higher temperature/lower solid fraction than coalescence of dendrites. Located across a grain boundary (intergranular coalescence), these results shed new light on the formation of hot tears in metallic alloys.


Hot tears are cracks that initiate during solidification (i.e. at non-zero solid fraction). They represent a major defect commonly encountered during the casting of large freezing range alloys and can lead to catastrophic cracking of the cast parts. In Figure 1a, a cracked rolling sheet ingot of very large dimensions is shown. In this example, two hot tears that were initiated during the start-up phase degenerated in two long cracks all along the ingot. In extrusion billets, hot tears are located in the center of the cast part, as shown in Figure 1b. This phenomenon leads to a loss in productivity as high as ten percent for some very sensitive alloys. The study of hot tear formation has gained a new interest during the last few years, in particular with the derivation of predictive criteria1,3,5,6 that can be implemented in finite element method models of casting.

Hot tears originate from a lack of liquid feeding of the mushy zone,1 especially at the end of solidification, and, more precisely, as highlighted by Campbell, when grains start to impinge and finally touch one another, but are still surrounded by a continuous liquid film.3

Most hot tearing criteria neglect the importance of thermomechanical aspects and simply consider the solidification interval of the alloy:3 the larger the freezing range, the more susceptible is the alloy to hot tearing. Clyne and Davies4 defined a criterion in which the time interval spent by the mushy zone in a vulnerable stage appears. As an alternative, Feurer5 focused on the liquid present between the grains and argued that a hot tear will nucleate as a pore if the liquid is no longer able to fill the intergranular openings. Rappaz et al.1 extended Feurer's approach in order to take into account the feeding associated with both solidification shrinkage and tensile deformation of the solidifying material. Recently, Farup and Mo6 formulated a two-phase model of a deforming, solidifying mushy zone where both interdendritic liquid flow and thermally-induced deformation of the solid phase were taken into account.

Although the investigations on as-cracked surfaces flourish, in situ observations of hot tear formation are rare. Nevertheless, transparent organic alloys offer an interesting alternative to observe the formation of hot cracks. Succinonitrile with acetone (SCN-acetone alloy) was selected by Farup et al.7 to induce hot tearing by mechanical pulling of the mushy zone during directional solidification. The same device has been used in the present investigation with two improvements: the use of a second solute element, a dye, which helps in distinguishing the liquid, the solid, and the voids, and the control of the pulling speed by an electrical motor. The main purpose of the present work is to visualize in situ hot tear formation during solidification, and, in particular, to study the nucleation of hot tears and the coalescence of grains.

Figure 1a
Figure 1b
Figure 2

Figure 1. (a-left) Hot tears in an aluminium slab that led to a complete cracking of the ingot; (b-right) Typical hot tear in extrusion billets.2(Click on the figures to enlarge them).     Figure 2. The schematics of the experimental device used to observe in-situ the formation of hot tears in succinonitrile-acetone alloys (Click on the figure to enlarge it).



The set-up used to induce hot tears in SCN-acetone is essentially the same as that used by Farup et al.7 As depicted in Figure 2, the succinonitrile-acetone alloy contained in a thin glass cell is solidified in a constant temperature gradient (Bridgman-type experiment) by moving it away from a heat source with an electrical motor at a constant velocity (typically 10 mm/s). In Farup's experiments, a "puller" was used to manually pull apart the growing dendrites in the transverse direction, while the current experiments achieved the same results, albeit with better control of the applied strain rate, using a supplementary electrical motor. A small quantity of Fuchsin was added to the transparent alloy (typically 0.05 wt.%) to enhance the contrast between the different phases of the system (liquid, solid, air). Temperature was measured every second by placing a type-K thermocouple, 50 mm in diameter, in the cell. This measurement was coupled with video recording of the growing dendrites in order to deduce the temperature at which dendrites coalesce or bridge.


The observation of hot tears is performed on time by a color video camera mounted on an optical microscope positioned on top of the set-up. As already reported in Reference 7, hot tears always appear between two grains; the grain boundary being the last part of the system to solidify. Moreover, two situations can be distinguished from these observations. First, when the stick is pulled at too high liquid fractions, liquid can feed the opening because of the high permeability of the mush. Therefore, the crack is healed by solute-enriched liquid and no defect results from this event, as shown in Figure 3, except maybe some enhanced segregation.

Figure 3
  Animation 1

Figure 3. A healed hot crack (Click on the figure to enlarge it).
Animation 1. The formation of a "healed" hot crack. To best experience this presentation, you should employ the latest version of RealPlayer.


Animation 1 shows how, by using a stick to induce a tensile opening in the organic alloy, deformation concentrates at the grain boundary. The dotted lines indicate the position of grain boundaries. In this case, the opening takes place between grains 2 and 3 (i.e., closest to the stick). Moreover, the solid fraction at this point of solidification is still low enough to allow liquid metal to feed the opening owing to the relatively high permeability of the mush. This leads to the formation of a healed crack avoiding any formation of hot tears.

It can be observed in Animation 1 that the tensile deformation induced by the stick is compensated by some liquid flow. The direction of this flow is revealed by the movement of some equiaxed grains that have nucleated in the solute-enriched liquid and which are moving to the right. During the opening, some dendritic arms grow in the liquid spacing between the two grains, as shown in Figure 4.

Figure 4
  Figure 5

Figure 4. An opening fed by some solute-enriched liquid (Click on the figure to enlarge it).
Figure 5. Two hot cracks growing in intergranular liquid film regions (Click on the figure to enlarge it).


The second situation observed in hot tears is that when the stick is pulled at a sufficiently high solid fraction (i.e., deep in the mushy zone where the permeability is low), some liquid cannot feed the opening from the tip of the dendrites and an intergranular crack appears, as shown in Figure 5 and Animation 2. Moreover, it can be seen from the deformation of the dendrites during pulling that they are particularly ductile at this temperature. Please note that two cracks are growing in parallel in this figure. The air bubble with the larger tip radius (top) is slightly ahead of the sharper one (bottom) as a result of the associated curvature depression (Laplace term).

The pressure in the intergranular liquid associated with both the solidification shrinkage and the tensile deformation increases owing to the inability of the liquid to feed the opening and reaches the cavitation pressure at which an air bubble appears. One can clearly see in Animation 2 air bubbles propagating into the opening thus creating a hot crack.

Animation 2
  Animation 3

Animation 2. The formation of a hot tear by propagation of air bubbles in the opening. To best experience this presentation, you should employ the latest version of RealPlayer.
Animation 3. A hot tear growing in intergranular liquid regions. To best experience this presentation, you should employ the latest version of RealPlayer.


Animation 3 shows more precisely the flow of air bubbles in the opening created by the pulling of the stick in the SCN-acetone alloy. It explains clearly how the entrapped porosity contributes to the formation of a hot tear. Moreover, it highlights the unstable equilibrium due to the associated curvature depression between the opening of the crack, the tip radius of the bubble and the movement of the bubble.


Figure 6a shows a typical cooling curve obtained with the thermocouple inserted in the cell. The cooling rate is constant during the experiment (-0.03 K/s) until heating is switched off at lower temperature to speed up solidification (slope changes at 1,100 s).

By associating temperature measurements with video observations of the solidifying transparent alloy, the liquidus temperature, Tliq, has been measured (see Figure 6b). In the same manner, the coalescence temperature of dendrites across a grain boundary (intergranular coalescence temperature), Tcg, and within a grain (intragranular coalescence temperature), Tcd, have been evaluated, as shown in Figures 7a and 7b and Figures 8a and 8b, respectively. Coalescence, or bridging of dendrite arms, is established when the liquid film between two adjacent dendrites or two grains is no more continuous (see Figure 7b and Figure 8b). Intergranular and intragranular coalescence does not occur simultaneously: Figures 7a and 7b were taken about 400 s after Figures 8a and 8b. These coalescence temperatures are highly dependent on the experimental conditions and on the disorientation of the grains. For this reason, the values of Tcd and Tcg reported in Figure 6a are mean values estimated from several measurements. Tcd is around 43.8°C whereas Tcg is much lower, 27.6°C. Indeed, for dendrites belonging to the same grain, no grain boundary energy has to be overcome to establish bridges.

Figure 6a
Figure 5b
Figure 8a
Figure 6b
Figure 7b
Figure 8b

Figure 6. (a-top) A typical temperature history during the growth of columnar dendrites in the SCN-acetone system; (b-bottom) a micrograph of the dendrites when the tips reach the thermocouple TC (determination of the liquidus temperature of the alloy, Tliq) (Click on the figures to enlarge them).   Figure 7. Pictures used to determine the temperature of coalescence (a-top) of dendrites across a grain boundary (Tcg); (b-bottom) magnification showing spots where coalescence is complete (B) and where some liquid films are still present (A). The upper part of the picture shows some completely solidified region, where only segregation of Fuchsin can be observed (Click on the figures to enlarge them).   Figure 8. Pictures used to determine the temperature of coalescence (a-top) of dendrites belonging to the same grain (Tcd); (b-bottom) magnification showing spots where coalescence is complete (B) and where some liquid films are still present (A) (Click on the figures to enlarge them).


Using the back-diffusion model of Clyne and Kurz8 and the thermophysical properties of Glicksman et al.,9,10 Grasso et al.11 estimated the solid fraction at intergranular and intragranular coalescence: fs(Tcg) = 0.99 and fs(Tcd) = 0.95.


Coalescence of dendrite arms has also been observed by catching the very moment when liquid films are no longer continuous. Intragranular coalescence is found to occur at a solid fraction around 95%, whereas intergranular coalescence takes place much deeper in the mushy zone at a higher solid fraction, around 99%. Further work should be carried out in order to determine more carefully when bridging of dendrites occur as a function of the grain boundary energy. Indeed, coalescence is a key parameter for the understanding of the mechanical property of the mush and of the hot tearing phenomenon. The theoretical foundation of coalescence within a grain and at grain boundary is also being established.12


This research was carried out as part of the Fifth Framework Competitive and Sustainable Growth program project GRD1-1999-10921 VIRCAST (Contract No. G5RD-CT-2000-00153). It included the partners: Alusuisse Technology & Management Ltd., Switzerland, Calcom SA, Switzerland, Elkem Aluminum ANS, Norway, Ecole Polytechnique Federale de Lausanne, Switzerland, Hoogovens Corporate Services, The Netherlands, Hydro Aluminum AS, Norway, Institute National Polytechnique de Grenoble, France, Institute National Polytechnique de Lorraine, France, Norwegian University of Science and Technology, Norway, Pechiney S.A., France, VAW aluminum AG, Germany, and IFE, Norway and SINTEF, Norway, as major subcontractors. Funding by the European Community and by the Office Federal de l'education et de la Science (Bern) for the Swiss partners is gratefully acknowledged.


1. M. Rappaz, J.-M. Drezet, and M. Gremaud, Met. and Mater. Trans. A, 30A (February 1999), p. 449.
2. I. Farup, "Thermally Induced Deformations and Hot Tearing During Direct Chill Casting of Aluminum" (Ph.D. thesis, University of Oslo, Norway, 2000).
3. J. Campbell, Castings (Oxford, U.K.: Butterman-Heinemann Ltd., 1991).
4. T.W. Clyne and G.J. Davies, J. Brit. Foundry, 74 (1981), p. 65.
5. U. Feurer, Giessereiforschung, Heft 2 (Neuhausen am Rheinfall, Schweiz, 1976), p. 75.
6. I. Farup and A. Mo, Met. and Mater. Trans. A, 31A (May 2000), p. 1461.
7. I. Farup, J.-M. Drezet, and M. Rappaz, Acta Materialia 49 (2001), pp. 1261-1269.
8. T.W. Clyne and W. Kurz, Metall. Trans. A, 12A (1981), p. 965.
9. M.E. Glicksman, R.J. Scjaeffer, and J.D. Ayers, Metall. Trans., 7A (1976), p. 1747.
10. M.E. Glicksman, P.W. Voorhees, and R. Setzko, "The Triple-Point Equilibria of Succinonitrile: Its Assessment as a Temperature Standard", TEMPERATURE, Its measurement and Control in Science and Industry, Vol. 5, Part I (Amer. Inst. of Physics, 1982), p. 321.
11. P.-D. Grasso, J.-M. Drezet, I. Farup, and M. Rappaz (Paper presented at EUROMAT Conference, Rimini, 2001).
12. M. Rappaz, A. Jacot, and W. Boettinger, to be submitted to Met. Mater. Trans.

For more information, contact P.-D. Grasso, Ecole Polytechnique Federale de Lausanne, Laboratoire de Metallurgie Physique, MX-G CH-1015, Lausanne, Switzerland;;

Copyright held by The Minerals, Metals & Materials Society, 2002

Direct questions about this or any other JOM page to

If you would like to comment on the January 2002 issue of JOM, simply complete the JOM on-line critique form
Search TMS Document Center Subscriptions Other Hypertext Articles JOM TMS OnLine