3 About the effect of pressure on the solidifying melt

Overcoming the shrinkage porosity caused by contraction of an alloy upon solidification has always been a major challenge for foundry men wishing to produce high-quality castings. According to [Gho00], Chernov was the first to suggest, in 1878, the application of external steam pressure on solidifying molten metal to enhance feeding.

The feeding of the remaining liquid is improved by increasing the pressure on the melt. The feeding velocity is controlled by Darcy`s law, which states that it is a function of the applied pressure gradient and the feeding resistance, which depends on the permeability of the already-solidified phases.

v0 = -K/? · ?P          [3-1]

In this equation K is the permeability, m the viscosity of the melt (e.g. eutectic liquid) and DP the pressure drop.

It is known that permeability is a function of solid fraction. Feeding stops suddenly when the permeability is too high and the pressure difference is insufficient to force metal into the interdendritic space. The process of feeding, similar to the infiltration of fiber preforms by molten light metals, can be described relatively easily for the isothermal case of gas pressure infiltration [Kauf92, Kauf93], but becomes very complicated when the permeability is a function of time due to solidification.

As just stated, solidification time and therefore the available feeding time limits the success of feeding due to externally-applied pressure. In thin-walled High Pressure Die Casting the external pressure cannot support feeding as much as it can in thick-walled squeeze casting, where runners and gating systems are designed for improved feeding. However, applied pressure not only helps reduce shrinkage porosity due to improved feeding: it may also alter the microstructure due to better heat transfer into the surrounding die material, and due to a possible grain refinement effect following extensive undercooling of the melt. This effect can be attributed to a change in the phase diagram following the Clausius-Clapeyron equation

?Tf/?P = Tf(V1-Vs)/?Hf         [3-2]

where Tf is the equilibrium freezing temperature, Vl and Vs are the specific volumes of the liquid and the solid respectively, and DHf is the latent heat of fusion [Gho00]. Substituting the appropriate thermodynamic equation for volume (by taking the liquid metals as an ideal gas), the effect of pressure on the freezing point may be estimated roughly as follows:

P = P0exp(-?Hf/RTf)         [3-3]

P0, ?Hf and R are constant, and Tf increases with increasing pressure during solidification. Equation 3-3 leads to a change of the liquidus and solidus lines in a binary phase diagram, as shown in Figure 3-1.

Figure 3-1: Change of liquidus and solidus lines in the binary Al-Si phase diagram following rapid solidification under high pressure [Gho00].

During the filling phase the melt experiences no, or practically no, external pressure above atmospheric pressure, but it may still drop in temperature close to or even just below liquidus. Immediately after die filling is completed, external pressure is applied. The latter can reach up to 150 MPa in today’s squeeze casting machines. In High Pressure Die Casting machines roughly 100 MPa is the upper limit. For binary Al-Si alloys an increase in the liquidus temperature by about 9°C was shown when solidification took place under 150 MPa [Gho00].

With this external pressure on the melt – which may be slightly undercooled at the moment of pressurization – at a given temperature, the phase diagram changes according to equations 3-2 and 3-3 and the melt is suddenly in a much more undercooled state than moments previously. This thermal undercooling stimulates existing nuclei in the melt to commence spontaneous heterogeneous solidification, resulting in much finer grain structures than in the same melt solidified under atmospheric pressure. Yong and Clegg observed a reduction in cell size of the magnesium alloy RZ5DF (4.2% Zn, 1 % Rare Earth) from 127 mm for gravity and 21 mm for direct squeeze casting [Yon04], leading to an improvement in strength of about 25%.

In this context the question of additional grain refinement and modification of light metal alloys must be discussed. Here it is difficult to elaborate a definite rule. The geometry of the casting and especially the wall thickness and solidification time must be taken into consideration. The heat transfer coefficient changes when internal metallostatic pressure from the feeding system suppresses the gap formation during the solidification of thin-walled castings [Cam03], and therefore external pressure alone will generate an improvement in the microstructure. However, further improvements may also result from altering the alloys. This subject is dealt with in Section 7.3, which covers grain refinement and modification.

From the microstructural point of view there is also one negative aspect, often attributed to the application of high external pressures on solidifying melts: macrosegregation. There seem to be two types, distinguishable according to the time when the separation of phases occurs. Primary phase and eutectic can separate during filling if pre-solidified primary phase is present, or during feeding under pressure, when a relative movement of solid and liquid phases (due either to shrinkage gap formation or to incomplete filling) is possible.

Recently published research work investigated this phenomenon mostly in the direct Squeeze Casting process [Bri03, Hon00], but it is also a familiar issue in indirect Squeeze Casting and Semi-solid Casting. Macrosegregation after filling may be less of a problem in conventional high pressure casting of thin-walled parts, but phase separation during filling is very prevalent.

St.John et al. report an increased volume fraction of eutectic Mg17Al12 at the skin layer on Mg-HPDC parts and attribute this to normal fluid flow phenomena during flow of semi-solid slurries (assuming that pre-solidified particles have been transferred from the sleeve into the die) [StJ03]. The formation of such a layer can significantly alter the corrosion behaviour of a part in machined or as-cast areas [Küh02].

In the authors´ opinion the basic requirement for the formation of macrosegregation after filling of the die is the potential for relative movement of solid and liquid phase. Sufficient amounts of liquid and sufficient time for movement are also essential. While the former may be present in thin-walled HPDC parts, there the solidification time is so short and wall-thickness so thin that the permeability of the solid is too high for relative movement during the very short time after pressurization. In Squeeze Casting and Semi-solid Casting the situation is quite different, since the cross-sections of the parts, the runners and the gating system are usually thicker, and solidification takes longer.

A simple example of relative movement of solid and liquid is seen when the solidifying melt begins to contract. Gap formation between the die and a first, weak solid shell is a phenomenon familiar to all foundry workers. Associated with a drop of heat transfer, it can generate huge gaps in large castings. Campbell [Cam03] describes cases of large castings of about 1 meter in length where gaps of 10mm on opposing sides of the die can develop. There is, however, a significant pressure drop between the gap area, which is under atmospheric pressure (it pulls air through ejector and venting channels), and the melt, which is subject to sufficient external pressure. This pressure drop can enable the liquid to flow back into the gap. It does not necessarily mean that the applied pressure ruptures the solid shell, which is in any case under tensile stress upon contraction. The thin shell may even remelt locally due to the temperature increase after the drop of heat transfer coefficient at first contraction.

It is mentioned above that extensive undercooling causes solidification to start throughout the melt once high pressure is applied. This means that primary solid crystals and eutectic liquid also coexist in the pool of melt inside the solidified shell. The permeability of the already-solidified layer may be so low, and the available channels so narrow that only the eutectic liquid can flow into the gap, while the primary crystals are basically filtered by the solidified shell. This type of macrosegregation is shown in Figure 3-2. A casting was made by New Rheocasting (NRC) with the alloy AlSi7Mg at about 50% solid fraction. The die filling of an end section was incomplete prior to pressure intensification. Upon pressurization the oxide shell ruptured, the solid phase was immobile and the remaining liquid filled the gap.

Figure 3-2: Macrosegregation in a New Rheocasting section of the alloy AlSi7Mg at roughly 50% solid fraction due to gap formation between the die and the part prior to pressurisation. The oxide skin ruptured and mainly liquid fraction moved into the available space.

This can only be avoided if large relative movements of liquid and solid phases can be kept to a minimum. This requires complete die filling (sufficient wall thickness, large radii) and low levels of gas entrapment. Collapsing gas bubbles can also be a source of macrosegregation, as...