Figure 3 – uploaded by Arne Dahle

See full PDF downloadDownload figure

Figure 3 generates the thermal gradient required for constitutional undercooling. Figure 3: Illustration of the principles for constitutional undercooling |"*! in constrained growth. Solute rejection at the crystal interface and simultaneous release of latent heat of fusion generates a constitutionally undercooled zone. For an equiaxed dendrite the principal difference between alloy systems is the growth

Related Figures (93)

Figure 1: Illustration of cold chamber high pressure die casting. Reproduced from [% 7), A production cycle in HPDC consists of: 1) metering of metal into the shot sleeve; 2) plunger movement; 3) rapid die filling. The steel die, typically 200-300°C, dissipates the latent heat, and during solidification the casting is 4) pressurized hydraulically by the plunger to feed the solidification shrinkage. Locking forces up to 4000 tons are commercially available to withstand the large pressures. Eventually, 5) the die is opened and 6) the casting is ejected. The hydraulic energy is provided by a computerized system that permits control of metal position, velocity and plunger acceleration to optimize the flow and the pressure during filling and solidification.

Table 1. Effects of elements in the A356 and the AM6OB alloys. The standard composition ranges and the major effects of the elements in the A356 and AMO60B alloys are shown in Table 1. The magnesium alloys are designated by an alphanumerical code; the first letter of the two major alloy constituents and their respective amounts describes the alloys: AluminiumManganese6%0.5%. The character B, (as in AM60B), refers to the composition ranges, and unless otherwise stated, AM60=AM60B in the thesis.

Figure 2: Nucleation is feasible in the thermally undercooled region adjacent to the die wall when the melt undercooling, ' T,, exceeds the required nucleation undercooling ' T,. 1) The melt becomes thermally undercooled adjacent to the mould wall, as schematically illustrated in Figure 2. The thermal undercooling, ' T,, can be sufficient to overcome the required nucleation undercooling, 'T,, which depends on the substrates that are present. Furthermore, the thermal gradient, G, (which depends on the melt superheat and the mould wall temperature), can alter the amount of thermal undercooling. This is the principal requirement for the separation theory presented by Ohno !'7! as well as the theory on free chill crystals by Chalmers '"'*!, Additionally, the showering theory by Southin ''”!, presumes that the heat loss by means of radiation at the melt surface generates a thermal undercooling which is sufficiently large for nucleation of crystals. the melt surface generates a thermal undercooling which is sufficiently large for showering theory by Southin ©”, presumes that the heat loss by means of radiation at

Table 2. Physical properties of the A356 alloy 7”), and the AM60B alloy”*), "Trig is reported in the literature as either 615°C or 619°C, the latter used in this work.

Figure 4: The low-solute parts of the Al-Si, (black lines), and Mg-Al, (blue lines), binary phase diagrams. Liquidus-lines assumed linear. Thermosolutal transition temperatures indicated with Tyrans. For a further introductory treatment, it is assumed that: 1) the melt is superheated, composition change is termed the thermosolutal transition temperature, (Tyans) ~~’, anc

Figure 5: Measurements of the shear strength of the material as a function of fraction solid, (f,), for the two extremes in crystal morphology; large dendritic crystals and small globular crystals °°. During stirring of a melt which is cooled, the shear strength can be measured as a function of the fraction solid. Measurements obtained with this method are shown in Figure 5. The shear strength of the material pass through several stages, and the coherency point is identified by changes in the shear strength of the mushy material, i.e. when the material starts to build up strength. The fraction solid at which this occurs depends on the crystal morphology. The coherency point, (f,"), is displaced to larger fraction solid when the crystals change from large dendritic to small globular. With increasing fraction solid, a maximum packing fraction is reached where the shear strength of the dendrite network is further increased.

Figure 6: a) Simulation of the filling of an A380 aluminium alloy 7). The scale bar indicates metal velocity; b) Simulation of particle position during filling. Reproduced with data from °7), There have been many investigations on the metal behavior during shot sleeve filling and a relatively large amount of literature is therefore available on the subject. Numerical modeling and experimental work generally reveal a similar constitutive metal behavior. Shot sleeve filling simulations are shown in Figure 6 °” **!, and the major flow characteristics are: 1) the metal impinges on the shot sleeve bottom, near the plunger; 2) it flows to the end of the shot sleeve, and then backwards; 3) a surge wave moves backwards towards the pouring hole; 4) additionally, metal continuously washes upon the plunger and partly accumulates there, (note the velocity scale bar at the top of figure 6a). The surge wave is similar to a hydraulic jump ““!, where the metal flow is characterized by the Froude number °°, Fr = v/(gh)'”. It is important to consider where the metal becomes positioned during filling, as shown in Figure 6b ®*!: the green particles are immersed in the melt early, and primarily become located near the die. The “older” red and yellow particles remain near the plunger.

Figure 7: Simulation of the melt temperature evolution in the shot sleeve, (longitudinal cross section), for an A356 aluminium alloy, (details in APPENDIX B). Total f,; when injection starts is ~10%. The heat transfer coefficient was determined from experimental measurements.

Figure 8: a) Cross sectional micrograph from the biscuit. Interfaces between solid surface layers and melt solidified later indicated with arrows; b) Magnification of the microstructure near the plunger. AM60B alloy.

Figure 9: Cross sectional micrograph from an A380 aluminium alloy die casting “! which shows large cold flakes embedded in the microstructure. 2.2.2 Externally Solidified Crystals (ESCs)

Figure 10: Optical micrograph from an industrially produced HPDC AM60 instrument panel. Etched microstructure: The large white particles are ESCs, whilst the gray structure is the fine grained matrix. The ESCs are a striking feature in the microstructure, because they are significantly larger than the grains formed from the metal that solidified in-situ in the die cavity. The ESCs are more frequently observed in the central regions of cross sections in the castings and a traditional misconception has therefore been that these coarser crystals are a result of the slower solidification encountered here. At different locations in a casting, the ESCs are observed to occupy up to 40% of the volume in magnesium alloy die castings and up to 20% in aluminium alloy die castings. An example of ESCs in an industrially produced AM60 instrument panel is shown in Figure 10. casting, the ESCs are observed to occupy up to 40% of the volume in magnesium alloy

A comprehension of each HPDC casting can be achieved by inspection of the shot diagram, as shown in Figure 11. The diagram provides a connection between the plunger movement and the pressure build-up as a function of time. Significant discrepancies between the measured and the programmed values can give indications of the quality of the casting. The plunger movement is usually separated into three different phases: I) the slow shot phase (1-7s); I) the die filling phase (0.01-0.3s), and; III) the pressurization phase (400-1200bar for 5-20s) °**!, Figure 11: Shot profile from an AM60 magnesium alloy die casting. The diagram shows plunger stroke, velocity and pressure as a function of time. I): slow shot phase (1-7s); ID) die filling phase (0.01-0.3s); III): pressurization phase (5-20s).

The first analytical model on plunger movement was based on the Bernoulli equation’, and was developed by Garber ©°!. This model does not include surface tension, viscosity variations and is only applicable in cases where the shot sleeve fill ratio is ~0.5. Figure 12 shows how the surge wave appears, and provides the basis for Garber’s model. Figure 12: Surge wave, with velocity v,,, ahead of plunger tip, with velocity vy.

Figure 13: Simulation of slow shot plunger movement and simultaneous metal flow and fraction solid distribution, indicated by the color scale bar 2] The hottest melt, (blue colour), enters runner first. part of the melt enters the runner first, cf. the scale bar. For the purpose of visualization, an example from the simulation of plunger movement is shown in Figure 13 ©!, When the plunger moves, the metal near the plunger is lifted up, and fills the cross section. With further plunger movement, a wave is formed ahead of the plunger. The metal adjacent to the die in the shot sleeve remains quiescent at this stage. Eventually, the quiescent metal is forced upwards, filling the residual empty part of the shot sleeve and, furthermore, flows into the runner system. Note that the hottest part of the melt enters the runner first, cf. the scale bar.

Figure 14: At approximately 70% plunger movement photographs of: a) The fragmented flow front; b) A location where the metal has frozen through the thickness during die filling, (marked with arrow). Principal die filling direction is vertical, parallel with acceleration of gravity. There are different interpretations to the metal flow mode in the die, i.a. atomized and turbulent flow. However, it is widely accepted that the die filling becomes complicated by rib structures, bends and variable metal velocities. It may therefore be inappropriate to generalize the flow mode. The apparently simple filling of a box-shaped die cavity can be difficult to predict and analyze. In an interrupted plunger movement, the plunger can be stopped during the die filling stage. The metal freezes quickly, approximately in-situ, providing an understanding of the metal flow pattern in the die cavity. Some features are shown in Figure 14, (from the HPDC castings in the doctorate work). The photograph in Figure 14a, indicates the extent of fragmentation at the flow front, and the arrow in Figure 14b shows a location where two metal flows met and the metal froze through the thickness during die filling.

Figure 15: a) Segregation band in an A356 die casting; b) A porosity band in an AM60 die casting. Bands delineated with black lines and marked with arrows. Many defects that occur in HPDC are related to the craftsmanship in the casting process, as described in (1 Not all defects, however, are included in the HPDC textbooks. Striking features commonly observed in HPDC microstructures are the defect bands. The bands typically follow the contour of the casting and consist of segregated eutectic and/or porosity. An example of a segregation band is shown in Figure 15a, whilst a porosity band is shown in Figure 15b.

$Figure 16: Illustration of: a) The fraction solid profile; b) The corresponding strength profile; c) The resulting microstructure where the central region facilitates deformation by slip along the segregated shear [30] plane *. Two decades later, a theoretical framework was put forward by Dahle et al. °°! to explain the banded defects: The defect bands form by a combination of a fraction solid gradient away from the die walls, (because of heat transfer into the relatively cold die), and the migration of ESCs to the centre as a result of lift forces acting on the solid. The solute enriched liquid concentrates next to the die walls resulting in the formation of the casting skin, leaving a region of solute enriched liquid between the semi-solid inner core and the outer skin. Deformation therefore concentrates in the lower fraction solid region, which then forms a shear plane. The nature of the defect bands can depend on the local f, in the shear plane during flow, it is argued, and if f, < f,° for much of the filling, bands of segregation appears. Contrarily, if larger f; is present in the plane, the band can be made up of pores or tears. The relationship between the strength, fraction solid, and the developing microstructure is shown in Figure 16. The central region moves in plug-type flow, whilst the liquid segregates to the low f, regions forming a hand moves in plug-type flow, whilst the liquid segregates to the low f, regions forming a$

Figure 16: Illustration of: a) The fraction solid profile; b) The corresponding strength profile; c) The resulting microstructure where the central region facilitates deformation by slip along the segregated shear [30] plane *. Two decades later, a theoretical framework was put forward by Dahle et al. °°! to explain the banded defects: The defect bands form by a combination of a fraction solid gradient away from the die walls, (because of heat transfer into the relatively cold die), and the migration of ESCs to the centre as a result of lift forces acting on the solid. The solute enriched liquid concentrates next to the die walls resulting in the formation of the casting skin, leaving a region of solute enriched liquid between the semi-solid inner core and the outer skin. Deformation therefore concentrates in the lower fraction solid region, which then forms a shear plane. The nature of the defect bands can depend on the local f, in the shear plane during flow, it is argued, and if f, < f,° for much of the filling, bands of segregation appears. Contrarily, if larger f; is present in the plane, the band can be made up of pores or tears. The relationship between the strength, fraction solid, and the developing microstructure is shown in Figure 16. The central region moves in plug-type flow, whilst the liquid segregates to the low f, regions forming a hand moves in plug-type flow, whilst the liquid segregates to the low f, regions forming a

segregation bands are aligned parallel with the surface of the casting, and few ESCs are

Figure 18: Measurements of microhardness (VHN), (10g indentation load), from the surface to the centre in half of the cross sections of thin walled, (2.5mm), AM60 magnesium alloy die castings: a) Large fraction of ESCs ~evenly distributed, with an average VHN 63; b) The ESCs are only located in the central region. Average VHN: 79. Measurements of the hardness are occasionally used for characterization of the variation in strength. In Figure 18, the measurements of the microhardness, (VHN), in two AM60 magnesium alloy castings are shown. The microstructure consists of a large fraction of ESCs distributed over the cross section in Figure 18a. In Figure 18b the ESCs are only present in the central region, and the hardness is observed to decrease significantly towards the centre.

Figure 19: Cross sectional micrographs from tensile tested specimens taken from AM60 magnesium alloy die castings: a) Near the gate, typical elongation: 3-10%; b) Farther from the gate, typical elongation: 15- 25%.

Figure 20: Schematic illustration of the plunger movement and corresponding qualitative description of fraction solid distribution. Metal with increasing f,, (ESCs), corresponding to locations I-III], enters the runner with further plunger movement.

Figure 21: The solidification and metal flow in the shot sleeve results in microstructures with different thermal histories in the casting. The microstructures correspondingly possess different fractions of ESCs. The locations marked II, II and I, correspond to position A, B and C, respectively, in the box shaped die castings in article #1-4. Thickness of the casting: 2.5mm. This results in different fractions of ESCs in different locations in the castings. Figure 21 shows the microstructures in the box shaped die castings that result from metal with thermal histories approximately corresponding to the different isotherms in Figure 20. The distance between the locations in the casting is indicated, (position HI is 25mm from the gate in the casting). The fraction of ESCs decreases with distance from the gate. from the gate in the casting). The fraction of ESCs decreases with distance from the

Table 3. Summary of the effects of variations in the shot sleeve thermal and constitutional conditions on the ESC formation mechanisms and the grain structures in the die castings. the AITiS5B1 grain refiner additions generated a finer grain size in the castings.

The ESCs were observed to segregate/migrate towards the central region, and the distribution of the ESCs was affected by the die temperature and by the initial f,. The ESCs were observed to reach a maximum packing in the centre. When further packing was not possible, the ESCs occupied a larger area of the cross section. Thus, at higher f,, the distribution of ESCs generally became wider and was spread towards the surface. This is schematically illustrated in Figure 22. The central ESC-rich region becomes larger by an amount S' r(2r+' r), when the fraction solid, (f,°°°), is increased. Figure 22: Schematic illustration of the packing of ESCs in the central region. At a specific fraction solid, further packing is restricted and the central region with ESCs becomes larger by an amount S' r(2r+' r).

Table 1. The main alloying elements in the AlSi7Mg alloy [wt%]. The boxes investigated were at a production stage of steady state with the following process parameters:

Figure 1 - a) The box shaped casting (Length: 400mm, height: 40mm, width: 75mm) with runner and gating system and overflows; b) Interrupted filling of the die at 80% plunger stroke. Note the bridge of solid material midway between the gates.

Figure 2 - Upper left: AlSi7Mg die casting with investigated positions indicated, A, B and C. Lower left: Optical micrograph taken from position A. Melt flow vertically. Floating crystals all over the cross section. Lower right: Optical micrograph taken from position B. Melt flow horizontally. Floating crystals mainly in the interior. Upper right: Optical micrograph taken from position C. Melt flow horizontally. Few floating crystals. approximately half the section thickness. In position C few coarse grains are found. thickness of the layer of coarse grains is largest near the gate and covers nearly the

Figure 3 - Upper left: Subsurface area grain map, position B. Lower left: Center area grain map, position B. Upper right: Subsurface area grain map, position C. Lower right: Center area grain map, position C.

Figure 4 - EPMA-detected variation in Si concentration crossing coarse and fine grains. Microprobe analysis performed in as polished condition. Optical micrograph in lightly etched condition. The variation in chemical composition along a line crossing fine and coarse grains as measured by EPMA, i.e. electron probe microanalysis, is depicted in figure 4. The specimen was analyzed in as polished condition and the image was obtained after lightly etching the specimen. With a 10 Pm spot size and 20Pm step size, each point on the micrograph corresponds to the volume analyzed. A significantly lower Si concentration is found in the coarse grain compared with the small grains. the micrograph corresponds to the volume analyzed. A significantly lower Si

Table 1 Process parameters. Inside diameter 60 mm, length 340 mm. The die castings, with 2.5 mm wall thickness, were produced in a Bithler SC42D 420 ton cold chamber die casting machine. The AM60B magnesium alloy was cast with the nominal composition Mg—6AI—0.4Mn(<0.2%Zn,<0.05%Si,<0.008%Cu,<0.001%Ni, <0.004%Fe). This alloy has a liquidus temperature of 619°C as determined by the equilibrium binary Mg—Al phase diagram.'* Table 1 lists the process parameters used for this work. A typical shot profile can be seen in Fig. 1. " Corrected from print in the journal.

1 measured plunger stroke; 2 measured plunger velocity; 3 measured pressure; 4 programmed pressure 1 Typical shot profile with filling phase plunger velocity of 5.5 ms"! and pressure 600bar. 1 measured plunger stroke; 2 measured plunger velocity; 3 measured pressure; 4 programmed pressure 1 Typical shot profile with filling phase plunger velocity of 5.5 ms”' and pressure 600bar.

2 AM60 die casting with investigated through thickness cross-sections designated A, B and C. Examination of the microstructure in several positions within the die casting revealed three distinct grain structures. The through thickness microstructures of various boxes were investigated at two positions, A and B, in front of one of the gates and in an area, C, midway between the gates in the bottom of the box (see Fig. 2). For investigations in the optical microscope the samples were ground, mechanically polished and finally etched in an acetic acid glycol solution.'®

3 Optical micrograph of etched AM60 microstructure, showing coarse dendrite surrounded by fine equiaxed grains, as occasionally observed in position A and centre of cross-section at position B.

the section thickness, (see Fig. 5). The area fraction of coarse crystals is approximately

4 Optical micrograph of etched microstructure in position A (melt flow vertical, T,,=680°C).

7 Grain size distribution maps: upper left: position B, subsurface, grain size ~11 Pm; lower left: position B, centre, coarse and fine grain sizes 20-75 Pm and 5-10 Pm respectively; upper right: position C, subsurface, grain size ~11 Pm; /ower right: position C, centre, grain size ~10 Pm. The grain maps in Fig. 7 were taken from the centre and subsurface areas, i.e. approximately 300 Pm from the surface, in both positions B and C. A fine grained microstructure with few coarse grains is observed in the subsurface maps of positions B (upper left) and C (upper right). In the centre of position B (lower left), many coarse grains are surrounded by fine grains, in contrast to the microstructure in the centre of position C (lower right) where only a few coarse grains are present.

8 Effect of superheat on grain size: upper left: position C, T,,=640°C; lower left: position B, T»=640°C; upper right: position C, T,=710°C; lower right: position B, T,=710°C. contrast to position C at 710°C (upper right micrograph), where a thin layer is observed

9 Fill tests and corresponding runner microstructures: top: interrupted movement at 200 mm plunger stroke; bottom: filled cavity at 320 mm plunger stroke.

Table 1. Process parameters used for producing AM60B die castings. (Tiiquidus Of the AM60 alloy is 619°CH 8), Samples were prepared for optical microscopy |"! taken from two positions in front of one of the gates, A) close to the gate and B) farther away, and C) in an area midway between the gates in the bottom of the box '*!, Area fraction measurements of the floating crystals were performed with an equivalent circle diameter (ECD) threshold of 10Pm, derived from image analysis ”°! Approximately 15 boxes were cast to obtain a quasi-steady state temperature in the shot sleeve. These initial castings were thus produced with constant melt temperatures at the initiation of filling, at qualitatively increasing shot sleeve temperatures. At steady state a minimum of 30 castings were produced for each melt temperature with process parameters as listed in Table 1 (details in '*!). The steady state shot sleeve temperature has been estimated to be 250°C directly below the pouring hole and 300°C at the end!*!,

distributed from the center to the surface. A more random distribution and fewer coarse Figure 1: Optical micrographs of etched cross sections (2.5mm thickness) of a casting produced with low superheat T,,=640°C. Micrograph A: flow vertical. B: flow horizontal. C: flow horizontal.

position A (640°C), the floating crystals covers 35% of the area in 50% of the cross section thickness. At 680°C this drops to 30% in 30% of the thickness. At 710°C the area fraction of floating crystals is approximately 10% in 10% of the section thickness. A maximum in area fraction is present in the center of position B for all temperatures. At 640°C the area fraction is approximately 40%, at 680°C it is 35% and at 710°C 13%. A slight increase in area fraction is observed closer to the surface at 640°C. In position C a maximum in area fraction is observed at 640°C, i.e. 25%. At 680°C and 710°C the area fraction in position C is approximately 2-3%. Figure 2: Area fraction of floating crystals in A, B and C for T,,=640°C (upper left), 680°C (upper right’ and 710°C (lower right).

““ a a — The castings produced at a low superheat (640°C) revealed two distinct crystal morphologies. The floating crystals are observed to be coarse dendritic or rosette like, as observed in Figure 4 a), or globular in shape, Figure 4 b). The latter dominates in position C. At higher melt temperatures (680 and 710°C) the crystal morphology is generally elongated dendrite trunks or arms (Figure 4 c) or branched dendrites (Figure 4 d). The branched crystals dominate in the center of the cross sections and closer to the gate Figure 4: Floating crystal morphologies; a) Coarse dendritic/rosette like, b) globular (T,,=640°C), c) elongated trunks/arms, d) branched dendritic (T,,=680 and 710°C). 4 d). The branched crystals dominate in the center of the cross sections and closer to

Figure 5: Micrograph from position A in a casting produced with coated shot sleeve. The coating procedure interrupted steady state and hence generated a_ slightly lower shot sleeve temperature. Initial T,,=680°C. ite) despite that a longer cycle time was inevitably generated during coating.

Figure 6: Microstructure of gate-section with dendrite fragments present.

Table 1. Alloy compositions based on average batch contents [wt%]. ‘The higher Ti and B contents are in the form of particles, (added as AITiSB1 grain refiner). Three variants of the A356 aluminium alloy were utilised: (1) A356 with a relatively low Ti content (0.007); (2) A356+AITi5B1, (same batch as (1), but AITi5B1 added as rod to the melt in a ratio of 10g to Ikg less than 60 seconds prior to HPDC to avoid fading); (3) A356+Ti, with a relatively high Ti content (0.13), in solution, (added during ingot production). This was intended to study the principal effects of Ti in solution and the effects of additions of grain refiner particles. The high iron content is usually added to prevent soldering in the die.

Figure 1: Complete box shaped casting with biscuit (B), runners (R) and overflows (O). Samples fot microstructural investigations were taken from the locations marked A, B and C. Wall thickness: 2.5mm.

Figure 2: Temperature as function of fraction solid for the A356 and A356+Ti alloys as predicted by the computer model Alstruc [4] Tsothermal solidification at the eutectic temperature, T.,=565°C. The temperature as a function of fraction solid-curves for the A356 and A356+Ti alloys, as determined by phase diagram calculations using the computer model Alstruc (41 are shown in Fig.2. A discrepancy at low fractions solid is present: the liquidus temperature of the A356 alloy (Tiig8609°C) is approximately 4K less than T)g8613°C for the A356+Ti alloy. However, the estimated variations in the melt temperature measurements is ~5K, and furthermore, the calculations are also disposed to some uncertainty due to the inherent complexity of calculating the solidification behaviour of multi-component alloy systems. It has therefore been assumed that equal melt temperatures induce approximately similar superheats. Additionally, it was assumed that the addition of AITiSB1 to the A356 alloy, (which gave the A356+AITi5B1 alloy - A356 with grain refiner particles), generated a negligible effect on the Tyg, and possessed a similar solidification behaviour.

Three different ESC morphologies are typically observed in the castings: 1) globular crystals, 2) elongated trunks and 3) branched dendrites. The globular crystals (Fig.3a) are associated with a low superheat. They are isolated, near-spherical crystals that are generally present farther from the gate as well as in intermediate positions in the cross section, (i.e. near the surface). The elongated trunks (Fig.3 b) are observed at all superheats and are typically present near the surface. The branched dendritic crystals (Fig.3 c), are coarser at low superheats, and are observed in the centre of cross sections and dominate near the gate. Figure 3: Different ESC morphologies; a) Globular; b) Elongated trunks; c) Branched dendritic. The ESCs are surrounded by a fine grained microstructure. Alloy: A356+T1.

Figure 4: Grain orientation maps obtained from EBSD measurements in SEM. a) Elongated trunk (A356); b) Coarse agglomerate-like crystal (A356+Ti). The white spots are non-indexed grid points. The arrows marks the lines separating the dendrite branches. To gain further information on the ESC-morphologies the microstructures were investigated in the SEM with the EBSD technique. This technique allows the crystallographic orientation of the grains to be determined by scanning the sample surface with the electron beam. The different crystallographic orientations can be processed to appear as grain maps, where a specific colour indicates a grain. Grains with misorientations greater than a specific value, e.g. 15°, are separated by grain boundaries, (delineated as black lines). Fig.4a shows an elongated trunk and a few coarse ESCs positioned in a fine grained matrix. An agglomerate-like ESC, (greyish colour with the circumference marked as a red line), which possesses dendrite arms with approximately similar colour, i.e. crystallographic orientation, is shown in Fig.4b. The vague, dark lines mark the borders between the dendrite branches, whilst the white areas are non-indexed grid points which can result from particles or a non-uniform ion _ Oage areas are non-indexed grid points which can result from particles or a non-uniform ion

Figure 5: Cross sectional micrographs from an A356+Ti casting (~25K superheat) at locations a) A, b) B and c) C. Flow directions: vertically (A); horizontally (B and C), parallel to the page. d) Measurements of the average area fraction of ESCs, (white particles in the micrographs), as a function of position fot different superheats in the A356+Ti castings.

Figure 6: Measurements of the area fraction of ESCs as a function of the thickness in positions a) A, b) B and c) C in A356+Ti castings produced with superheats of 25K, 65K and 95K. superheat a defined peak is present in the centre with an area fraction of ESCs of

and red, respectively. The measurements of the Ti content are included in the diagram in Fig.7 c. The measurements were done from the surface to the centre of the cross sections in defocus mode, (i.e. 50Pm probe size) and represent the average of 3 parallel lines. The amount of Ti increases towards the central region in the A356+Ti casting, (line marked A356+Ti_Ti), and furthermore, the Si content, (Fig.7 c), decreases markedly from the surface to the ESC-rich central region (A356+Ti_Si). This is less pronounced for the A356 casting, (A356 Si), which possesses a lower fraction of ESCs. The scatter in the measurements of the Si content is inherently related to whether the probe was positioned in Al-Si eutectic or a primary crystal. An example of an EPMA map is shown in Fig.7 d, where the Al and Si contents are qualitatively illustrated with dark and red, respectively. Figure 7: Micrographs from position A, (from the surface to the centre of the cross sections), in castings produced with a superheat of ~65K: a) A356 alloy; b) A356+Ti alloy. Flow direction: vertically and parallel to the page; c) Measurements of the area fraction of ESCs and EPMA-measurements of the Si and Ti contents; d) EPMA map from position A in an A356 casting produced with 65K superheat. Primary Al appears dark and AI-Si eutectic appears red. (Probe size: 1Pm; step size: 0.5Pm; acc.voltage: 15kV; probe current: 30*10° Amp). shown in Fig.7 d, where the Al and Si contents are qualitatively illustrated with dark

Figure 8: Micrographs from location B, produced with 65K superheat: a) A356+AITiSB1 and b) A356 castings. Flow direction horizontal, parallel to the page. c) Diagram of the area fraction of ESCs as a function of the thickness. Micrographs from location B in A356 castings produced with 65K superheat are shown in Fig.8. A relatively large fraction of the ESCs are segregated to the central region in Fig.8 a, (A356+AITi5B1), whilst a relatively dispersedly distributed and significantly lower fraction of ESCs is present in Fig.8 b, (A356). The measurements of the area fraction of ESCs as a function of the thickness are shown in the diagram in Fig.8 c. The A356+Ti castings possess a similar ESC-distribution as the A356+AITiSBI castings with a maximum area fraction of approximately 18%, whilst the A356 casting possesses a significantly lower area fraction of ESCs (~4%). Additionally, the ESCs in the A356+AITi5B1 casting are more centred. However, a few ESCs are occasionally observed near the surface, which is a result of some ESCs being deposited in the surface layer that forms at the wall during filling ''*.

Figure 9: Micrographs from location B in castings produced with superheats of 65K: a) A356 and b) A356+AITiSB1; and 95K: c) A356 and d) A356+AITi5B1. A significantly finer grain size is present in b and d.

Figure 10: Micrographs from location B in A356+Ti castings produced with 95K superheat and a) shot delay time 2 seconds and b) conventionally cast (i.e. no shot delay time).

therefore established, as shown in Fig.11. It has been shown that the die temperature and the fraction of ESCs affects the position of defect bands in A356 die castings, and the responsible mechanisms are described in detail in ''”, The ESC distributions can therefore be important for the interpretation of the mechanisms that controls the formation of defect bands as well as the segregation of crystals to the central region. For later work, a polynomial curve fitting was therefore established, as shown in Fig.11. Figure 11: Diagram with area fraction of ESCs as function of thickness. Experimental data for A356+Ti casting (25K superheat) and polynomial curve fitting.

Figure 1: Typical HPDC microstructures obtained from box-shaped die castings containing externally solidified crystals (ESCs): a) Straight section, flow direction to the right and parallel to the page; b) Bend, flow direction vertical; c) Magnification of the microstructure in the bend.

Figure 2: Schematic outline of crystal migration towards the centreline, CL. Initially: Uniform distribution of crystals; At a later stage: Particles migrated towards the CL. FD: Flow direction.

Figure 3: Illustration of surface layer formation and ESC distribution. Melt with a low fraction of ESCs solidifies at the die walls forming a surface layer. Later, the metal which flows in the same cross section is consequently forced to the centre. Reproduced with data from !"7!, Secondly, it has been proposed that the distribution of ESCs in a die casting results from the solidification conditions present in the shot sleeve and in the die cavity in the HPDC process '”!, An inhomogeneous temperature distribution is inherently present ir the shot sleeve resulting in non-uniform solidification conditions. This results in < variable distribution of ESCs in the castings, with a higher fraction of ESCs near the gate and lower fractions farther from the gate. Additionally, a solidifying surface layet forms, which constricts the metal to flow in the centre. This is schematically illustratec in Fig.3. The metal in position II, far from the gate, possesses a thermal history originating from the hotter shot chamber end and therefore contains few ESCs. The metal in position I (closer to the gate) originates from closer to the plunger anc contains a larger fraction of ESCs (as shown in the micrograph, far left in Fig.3). The surface layer, made up of the first metal to enter the die, possesses few ESCs anc forces later metal, with a larger fraction of ESCs, to the central region. surface layer, made up of the first metal to enter the die, possesses few ESCs and

Table 1. Principal alloy constituents [wt%], according to the A356 batch composition. The alloy was first heated in a 11 graphite crucible to a superheat of 90K. The crucible was then removed from the furnace and allowed to cool in air whilst being vigorously stirred using a boron nitride coated and preheated impeller. A rotation rate of 600rpm was used as it was found to be high enough to cause significant dendrite fragmentation and some spheroidisation so that the material had a low enough viscosity for pouring. At a lower stirring rate the dendrites became large and branched and were found to impede flow. The superheat dissipated in approximately 3 minutes and the cooling rate in the solidification range was approximately 3.8Kmin’'. When the desired fraction solid (f,) was obtained, stirring was stopped and the metal was poured into a steel die within 3 seconds.

dashed lines. The solidification behaviour of the alloy, (approximated as Al-7wt%Si), was estimated by the Scheil-equation. The temperature as a function of f, is shown in Fig.4. The different melt temperatures that were used and the corresponding f,, are indicated with dashed lines. Figure 4: Solidification behaviour of the A356 alloy as estimated by the Scheil-equation. Isothermal solidification at the eutectic temperature, T.,=577°C. Solidification completed at T,..=553°C. different melt temperatures that were used and the corresponding f,, are indicated with

Figure 5: Transverse drawing of the die cavity and drawing of complete die. The elbow-shaped steel die (Fig.5) was first pre-heated in an air-circulating furnace where the temperature was controlled with a thermocouple probe to +/-5K. The die has an open end (Fig. 5) and, during casting, the metal therefore flowed continuously through the cavity until flow arrested due to solidification. To further assist in the movement of the slurry, the die was tilted 8°. The cooling rate of the die was approximately 8Kmin’' in the ambient temperature range (T,i;=21°C at 33% humidity) and the heat dissipation from the die was therefore negligible during casting.

Table 2. Castings produced with f,~0.16 and a range of die temperatures. The five different die temperatures that were used are listed in Table 2. A constant fraction of ESCs of 0.16 was used in these experiments to simulate the f, injected in cold chamber HPDC. These conditions induced a significant amount of flow through the die, which is assumed to be essential to observe pronounced effects. The die temperatures used were chosen based on initial casting trials that generated observable differences in the microstructure.

Figure 6: a) Macrographs of a longitudinal section of an A356 casting produced with Tg;.=270°C and f,~0.16. The ESCs appear as large primary dendrites (white). The fine grains and the Al-Si eutectic comprising the matrix appears as dark grey; b) A photograph of the casting. Microstructural samples were taken from the plane marked with an arrow.

Figure 7: Optical micrographs of the microstructures obtained from different positions in the cross section. a) In the central region: large agglomerated ESCs and AI-Si eutectic; b) In the transition between the central region and the surface layer: smaller ESCs surrounded by fine dendrites; c) Near the surface: a fine dendritic microstructure.

Microstructures of samples produced with a variable Tyie and a constant f,~0.16 are u shown in Fig.8. At a low die temperature of 180°C (Fig.8 a) the ESCs are positioned in the centre of the cross section. A casting produced with a die temperature of 270°C is shown in Fig.8 b. The ESCs are centred with a few ESCs positioned near the upper edge. At a die temperature of 350°C (Fig.8 c) the ESCs are principally located in the pper part of the cross section but the ESCs are still segregated away from the surface. At higher die temperatures (e.g. 450°C) the metal did not fill the entire cross section, as shown in Fig.8 d, and few ESCs are present.

The effect of variable f, on the distribution of ESCs is shown in Fig.9. The ESCs are distinctly centred in the cross section at f, = 0.07 (Fig.9 a) and the ESCs were found to be more centred at lower fractions of ESCs. At f; = 0.16 (Fig.9 b), the central ESC containing region covers a larger area of the cross section. In the castings produced with f, = 0.22, the ESCs are more dispersed (Fig.9 c), and when f, = 0.3, the metal did not fill the entire cross section as shown in Fig.9 d. However, a relatively thin surface layer free from ESCs is present adjacent to the surface in the bottom part of the casting. layer free from ESCs is present adjacent to the surface in the bottom part of the casting. Effect of Variable f, on ESC Distribution The micrographs indicate differences in the local area fraction of ESCs in the central region. Measurements of the area fraction of ESCs conducted at higher magnification, indicate that the local area fraction in the centre is typically ~0.35-0.4. No significant variations in the ESC packing fraction in the centre occur irrespective of the bulk ESC fraction used. The micrographs indicate differences in the local area fraction of ESCs in the central

0.07 sample is a random effect and can be attributed to the local variations in f, as a result of large pools of eutectic in the regions analysed. To determine the variations in distance from the edge to the ESCs, a line following the apparent interface was drawn between the ESC-free surface layer and the ESC-rich centre region, as shown in Fig. 10 b. The average distance from the ESCs to the surface was measured, (Fig.10 a), and found to decrease from 3.5mm to ~l1mm with an increase in the fraction of ESCs from 0.07 to 0.22. Figure 10: a) Measurements of the local area fraction of ESCs as a function of the diameter (D) from the edge of the samples (from left to right). The vertical lines indicate the average distance from the edge to the central ESCs. b) Micrographs of the samples with superimposed lines to distinguish the surface layer and the centre region. The upper micrograph is from a casting with f,~0.07. The bottom micrograph is from a casting with f,~0.22. Die temperature ~150°C.

Figure 11: Macrograph from a longitudinal section in a binary Mg-6wt%Al alloy casting. An increased fraction of ESCs is observed towards the centreline analogous to the A356 castings. Casting produced with an initial f, of ~0.16 and a die temperature of 270°C. The magnified area reveals relatively fine ESCs. Figure 11: Macrograph from a longitudinal section in a binary Mg-6wt%Al alloy casting. An increased fraction of ESCs is observed towards the centreline analogous to the A356 castings. Casting produced with initial f, of ~0.16 and a die temperature of 270°C. The magnified area reveals relatively fine ESCs.

Figure 12: a) Microstructure of metal that contains ESCs quenched in a steel crucible produced with an initial f,~0.16 and a die temperature of 250°C; b) From central region of casting produced with similar parameters. To test the hypothesis that an approximately uniform mixture of liquid and ESCs was poured during the casting operation, the first metal that flowed through the die was quenched in a crucible. The resulting microstructure in the quenched material is shown in Fig.12 a. This metal contains a qualitatively similar area fraction of ESCs as originally aimed at. A micrograph from the central region of a casting produced with similar f, and die temperature as in a, is shown in Fig.12 b. A slightly higher area fraction of ESCs is present in b.

Figure 13: Qualitative flow front appearance with variations in die temperature. a) Low die temperature, e.g. 130°C; b) Higher die temperature, e.g. 350°C; c) Very high die temperature, e.g. 450°C. The microstructure in Fig.8d shows that at a die temperature of 450°C, there is no upper wall constraint at all. In this case, there is only a migration force from the bottom die wall forming a layer which contains few ESCs at the bottom wall. On top of this mushy surface layer, the slurry, containing the majority of the ESCs, continuously flows through the die. Flow stops after the whole crucible has been poured into the die and few ESCs exist in the final microstructure (Fig.8 d). This is illustrated in Fig. 13 c.

Figure 14: Spherical particle, of diameter D, in fluid flow with a linear velocity profile du,/dy adjacent to -, ee | the particle and fluid (V-V,,) and the induced Saffman lift force (F,) per unit mass is a

Figure 3 – uploaded by Arne Dahle

Related Figures (93)

Connect with 287M+ leading minds in your field