My investigation into the formation mechanisms of underside shrinkage in sand-cast ZA27 alloy reveals a complex interplay between solidification characteristics, density segregation, and feeding dynamics. This defect, which manifests as irregular upward depressions on the bottom surface of castings, represents a significant challenge in the production of high-aluminum zinc-based alloys. Through systematic experiments and theoretical analysis, I have elucidated the fundamental processes responsible for this particular sand casting defect.
Experimental Methodology
Alloy Preparation and Melting
I prepared the ZA27 alloy using a graphite crucible in an electric resistance furnace. The process involved first melting an Al-50%Cu master alloy, which I then cast into ingots for subsequent use. The raw materials consisted of zinc ingots, aluminum ingots, magnesium ingots, and the Al-50%Cu master alloy. To account for elemental losses during melting, I controlled the final chemical composition within the following ranges: 25-28% aluminum, 2.0-2.5% copper, 0.01-0.015% magnesium, with the balance being zinc.
Test Specimen Configuration
The experimental setup utilized step-shaped test specimens designed to evaluate the solidification behavior and defect formation in ZA27 alloy. I employed clay-bonded green sand molds for all casting trials. The investigation compared two distinct process variants: one without any chilling material at the casting bottom, and another incorporating a 60mm thick aluminum chill plate at the bottom surface.
To monitor the solidification sequence, I positioned NiCr-NiAl thermocouples at various heights along the centerline of the test specimen. Temperature data were continuously recorded using an XWT-664 six-channel automatic balancing recorder throughout the entire solidification process.
Analytical Methods
Following solidification, I conducted comprehensive examinations including surface observation of the castings, metallographic analysis at different cross-sections, chemical composition analysis at various heights, and density measurements. For the density analysis, I calculated the effective density \( \rho_e \) using the relationship:
$$ \rho_e = \frac{M}{\rho_m} \cdot \rho_p $$
Where: M represents the density coefficient calculated from the average alloy composition (using densities of 2.7 g/cm³ for Al, 8.9 g/cm³ for Zn, and 7.1 g/cm³ for Cu), giving M = 5.976 g/cm³ for the nominal ZA27 composition; \( \rho_m \) denotes the density coefficient calculated from the actual local composition; and \( \rho_p \) represents the measured density value.
Results and Observations
Thermal Analysis
The cooling curves recorded at different heights of the test specimen reveal distinct solidification patterns between the two process variants. Table 1 summarizes the key temperature data and solidification parameters derived from these measurements.
| Position Height (mm) | Without Chill (T_l, °C) | Without Chill (T_s, °C) | With Chill (T_l, °C) | With Chill (T_s, °C) |
|---|---|---|---|---|
| 475 | 488 | 455 | 482 | 445 |
| 430 | 490 | 458 | 488 | 453 |
| 308 | 500 | 467 | 493 | 459 |
| 130 | 503 | 470 | 500 | 467 |
Chemical Composition and Density Distribution
The chemical analysis results, presented in Table 2, demonstrate substantial compositional segregation along the height of the casting, particularly in the absence of chilling.
| Height (mm) | Al (wt%) – No Chill | Al (wt%) – With Chill | Density (g/cm³) – No Chill | Density (g/cm³) – With Chill |
|---|---|---|---|---|
| 521 | 29.0 | 27.5 | 4.74 | 4.81 |
| 473 | 29.2 | 27.8 | 4.80 | 4.82 |
| 418 | 27.4 | 26.9 | 4.81 | 4.78 |
| 237 | 26.8 | 26.5 | 5.12 | 4.98 |
| 173 | 25.2 | 26.1 | 5.16 | 5.04 |
| 87 | 22.4 | 25.8 | 5.33 | 5.11 |
The data reveal a pronounced inverse segregation pattern: aluminum content increases toward the top of the casting while zinc concentrates at the bottom. Copper shows minimal segregation, with only a slight enrichment in the lower regions. This compositional gradient is significantly reduced when chilling is applied.
Formation Mechanism Analysis
Density Segregation and Primary Phase Behavior
The key to understanding the formation of underside shrinkage lies in the solidification behavior of ZA27 alloy. The alloy exhibits a wide crystallization temperature range, spanning from approximately 375°C to 487°C. During solidification, the primarily formed α-dendrites have a composition significantly different from the bulk liquid. Energy dispersive spectroscopy analysis conducted on quenched samples revealed that the primary α-phase contains approximately 42.47% Al, 5.6% Zn, and 1.87% Cu. This composition yields a density approximately 85% of the liquid alloy, making these primary crystals buoyant within the melt.
The partition coefficient \( k \) for zinc between the α-phase and liquid can be approximated as:
$$ k = \frac{C_s}{C_l} = 0.865 $$
Using this value, the calculated zinc content in the primary α-phase would be:
$$ C_{Zn}^{\alpha} = 71\% \times 0.865 = 61.4\% $$
This corresponds to an aluminum content of approximately 36.6%, consistent with the direct measurement results.
Inverse Solidification Sequence
The compositional segregation induces significant variations in the equilibrium liquidus temperature \( T_l \) and solidus temperature \( T_s \) along the casting height. The relationship between aluminum content and liquidus temperature follows the Zn-Al binary phase diagram, with higher aluminum content corresponding to higher liquidus temperatures. The observed compositional gradient creates a situation where the upper portions of the casting (enriched in aluminum) have higher liquidus temperatures than the lower portions (depleted in aluminum).
For the casting produced without chilling, the liquidus temperature difference between top and bottom exceeds 25°C. However, during solidification, the actual temperature difference between these locations remains below 20°C. This results in the counterintuitive situation where the bottom of the casting begins solidifying later than the top, establishing an inverse solidification sequence from top to bottom.
The solid fraction evolution can be described using the Scheil equation for non-equilibrium solidification:
$$ f_s = 1 – \left(\frac{T_m – T}{T_m – T_l}\right)^{1/(k-1)} $$
Where: \( f_s \) represents the solid volume fraction; \( T_m \) denotes the melting point of the solvent (660.3°C for pure aluminum); \( T_l \) is the liquidus temperature; \( T \) is the current temperature; and \( k \) is the partition coefficient (0.865 for this system).
Macroscopic Feeding Termination Point
Previous research has established that macroscopic feeding ceases when approximately 75% solid fraction is achieved. Using the Scheil equation with \( f_s = 0.75 \) and \( k = 0.865 \), I calculated the temperature at which bulk feeding stops: T_f. The results, presented in Table 3, vary significantly with position due to the compositional differences.
| Position Height (mm) | T_f, No Chill (°C) | T_f, With Chill (°C) |
|---|---|---|
| 521 | 437 | 445 |
| 473 | 453 | 448 |
| 418 | 455 | 453 |
| 237 | 459 | 459 |
| 173 | 467 | 463 |
| 87 | 470 | 467 |
Anomalous Feeding Direction
By correlating the cooling curves with the calculated feeding termination temperatures, I determined the temporal sequence of feeding cessation across the casting height. For the casting without chilling, position 1 (bottom) reaches its feeding termination temperature at time t₁. At this same instant, position 2 (slightly higher) has already cooled below its own feeding termination temperature, indicating that the upper region has lost its feeding capability before the lower region.
This analysis reveals a critical finding: during the middle and late stages of solidification, the upper portions of the casting, which have already ceased bulk feeding, require additional liquid to compensate for solidification shrinkage. However, the liquid available for feeding exists primarily in the still-molten lower regions. This establishes an anomalous feeding direction from bottom to top, contrary to conventional gravity-assisted feeding.
The driving force for this reverse feeding consists of two components: the vacuum created by solidification shrinkage in the upper regions, and the surface tension of interdendritic liquid. Mathematically, the net feeding pressure can be expressed as:
$$ \Delta P = \Delta P_s + \frac{2\sigma}{r} – \rho g h $$
Where: \( \Delta P_s \) represents the pressure differential due to solidification shrinkage; \( \sigma \) is the surface tension of the liquid; \( r \) denotes the characteristic radius of interdendritic channels; \( \rho \) is the liquid density; \( g \) is gravitational acceleration; and \( h \) represents the vertical feeding distance.
For the conditions prevailing in sand casting without chilling, the combination of shrinkage-induced vacuum and capillary forces must overcome the gravitational head to sustain upward feeding. When the feeding demand exceeds the available pressure gradient, the underside depression forms as the liquid attempts to feed upward against gravity.
Effect of Chilling on Defect Prevention
The application of a 60mm thick aluminum chill at the casting bottom fundamentally alters the solidification sequence and feeding behavior. The enhanced cooling rate at the bottom creates a steeper temperature gradient along the casting height, promoting bottom-up solidification. Analysis of the cooling curves with chilling shows that position 1 (bottom) reaches its feeding termination temperature at time t₁’. At this moment, position 2 remains significantly above its feeding termination temperature, demonstrating that the upper region maintains feeding capability throughout the bottom solidification process.
The thermal gradient enhancement can be quantified by examining the temperature differences between adjacent positions. Without chilling, the maximum temperature difference between positions 1 and 4 during solidification remains below 20°C. With chilling, this difference increases substantially, ensuring that the liquidus isotherms propagate from bottom to top.
The aluminum chill alters the local solidification time at the bottom according to:
$$ \tau_c = \tau_0 \cdot \left(1 + \frac{B_i}{2}\right)^{-2} $$
Where \( \tau_c \) represents the solidification time with chilling, \( \tau_0 \) is the solidification time without chilling, and \( B_i \) is the Biot number characterizing the heat transfer at the mold-metal interface.
Interdendritic Feeding and Rough Surface Morphology
The characteristic rough, honeycomb-like appearance of the underside depression surface results from the nature of interdendritic feeding during the final stages of solidification. When bulk feeding becomes impossible, the remaining liquid must traverse increasingly narrow interdendritic channels to reach the areas requiring compensation. The surface tension-driven flow through these channels creates multiple small feeding paths, each leaving a characteristic mark on the depression surface.
The morphology of the underside depression reflects the competition between capillary-driven feeding and gravitational resistance. The depression typically exhibits a central pit surrounded by shallower honeycomb-like areas, corresponding to regions where the feeding distance and channel geometry create varying levels of resistance.
Critical Feeding Parameter Analysis
To quantify the conditions leading to underside shrinkage, I developed a feeding effectiveness parameter F that considers the combined effects of solidification sequence and alloy properties:
$$ F = \frac{\Delta T_{gradient}}{\Delta T_{liquidus}} \cdot \frac{\rho_{solid}}{\rho_{liquid}} \cdot \frac{t_{local}}{t_{total}} $$
Where: \( \Delta T_{gradient} \) represents the actual temperature gradient along the casting height; \( \Delta T_{liquidus} \) denotes the liquidus temperature variation due to segregation; \( \rho_{solid} \) and \( \rho_{liquid} \) are the solid and liquid densities respectively; \( t_{local} \) is the local solidification time; and \( t_{total} \) represents the total solidification time of the casting.
When F exceeds a critical value, the feeding direction becomes normal (top-down) and underside shrinkage is prevented. Conversely, when F falls below this threshold, the anomalous bottom-up feeding predominates, leading to defect formation. The experimental results correlate well with this parameter, showing that chilling increases F above the critical threshold for defect prevention.
Threshold Chill Thickness Considerations
My experiments also explored the effect of chill thickness on defect formation. When insufficient chill thickness was employed, a different defect morphology appeared. Although the bottom surface remained free of underside shrinkage, extensive shrinkage porosity developed in the middle-to-lower regions of the casting cross-section.
This transitional behavior occurs because thin chills provide only limited cooling enhancement. The thermal influence penetrates only to a specific height, creating a bifurcated solidification pattern: normal bottom-up solidification near the bottom, but inverse top-down solidification in the upper regions. The interface between these two solidification fronts becomes a zone of high porosity as the two feeding systems compete.
The critical chill thickness for complete defect prevention depends on the casting geometry and thermal properties, following the relationship:
$$ d_{crit} = \alpha \cdot \sqrt{\frac{\lambda \cdot H \cdot \tau_{solidification}}{c_p \cdot \rho \cdot \Delta T}} $$
Where \( d_{crit} \) is the required chill thickness; \( \lambda \) is the thermal conductivity; \( H \) is the casting height; \( \tau_{solidification} \) is the solidification time; \( c_p \) is the specific heat capacity; \( \rho \) is the density; \( \Delta T \) is the superheat; and \( \alpha \) is a geometric factor.
Microstructural Evidence
Metallographic examination provides supporting evidence for the proposed mechanism. Microstructures from the upper portions of castings produced without chilling exhibit more rounded and smoother dendrite boundaries compared to those from lower portions. This morphology suggests that the α-dendrites in the upper regions were subjected to washing and dissolution by the upward-flowing liquid during the anomalous feeding process. The lower portions contain more fragmented and irregular dendrites, consistent with a region that served as a liquid source rather than a sink.
The upper microstructure also shows fewer secondary dendrite arms, indicating coarsening due to extended contact with liquid at elevated temperatures. This microstructural gradient provides physical evidence of the upward liquid flow that characterizes the anomalous feeding behavior.
Summary of Defect Formation Sequence
Based on my comprehensive investigation, the sequence of underside shrinkage formation in sand-cast ZA27 alloy proceeds as follows:
Stage 1: During the early stages of solidification, primary α-dendrites nucleate throughout the casting. Due to their lower density relative to the liquid, these dendrites begin to float upward, initiating the compositional segregation process.
Stage 2: The upward migration of aluminum-rich dendrites and downward displacement of zinc-rich liquid intensify the compositional gradient. This establishes the inverse liquidus temperature distribution, with higher liquidus temperatures at the top and lower liquidus temperatures at the bottom.
Stage 3: Solidification begins at the top of the casting due to its higher liquidus temperature, while the bottom remains partially liquid. This creates an inverse solidification front traveling from top to bottom.
Stage 4: As the upper regions solidify and undergo volumetric shrinkage, they require liquid feeding. However, the available liquid resides predominantly in the lower regions, forcing feeding to occur upward against gravity.
Stage 5: When bulk feeding becomes impossible (at approximately 75% solid fraction), interdendritic feeding predominates. The capillary forces driving this feeding must overcome the gravitational head, creating the characteristic rough depression surface.
Stage 6: The final solidification of the bottom region traps the depression morphology, leaving a permanent sand casting defect in the form of underside shrinkage.
Preventive Measures and Implications
The understanding of this formation mechanism provides clear guidance for preventing underside shrinkage in sand-cast ZA27 alloy. The most effective approach involves manipulating the thermal conditions to establish a normal bottom-up solidification sequence. My experiments demonstrate that proper chilling can achieve this objective while maintaining the compositional advantages of the alloy.
For foundry practice, the critical parameters to control include: cooling rate at the casting bottom, thermal gradient along the casting height, and the relationship between local liquidus temperature and actual temperature distribution. The implementation of these findings can significantly reduce the incidence of this sand casting defect in production environments.
Conclusions
Through this detailed investigation, I have established that underside shrinkage in sand-cast ZA27 alloy represents a unique sand casting defect resulting from the specific solidification characteristics of high-aluminum zinc-based alloys. The defect originates from the combined effects of mushy-zone solidification, density-driven segregation, and the resulting inverse solidification sequence.
The primary conclusions from this work are:
The underside depression forms during the late stages of solidification, resulting from both bulk feeding and interdendritic liquid feeding in the anomalous upward direction. The driving force for this reverse feeding must overcome gravitational resistance, with the available pressure differential consisting of solidification shrinkage vacuum and interfacial tension forces.
The formation of this defect originates from two fundamental solidification characteristics of high-aluminum zinc alloys during sand casting: the mushy solidification mode and the density segregation between zinc and aluminum. This segregation creates the top-down solidification sequence that ultimately drives the bottom-up feeding phenomenon.
The upward feeding necessary to prevent shrinkage in the upper regions must work against gravity, with the motivating forces being solidification contraction vacuum and the surface tension of interdendritic liquid channels.
Proper application of chilling, which accelerates cooling at the casting bottom and establishes a normal bottom-up solidification sequence, effectively prevents the formation of this sand casting defect. The chill thickness must be sufficient to influence the entire casting height to avoid creating a transition zone of porosity in the middle regions.