In my research on lightweight high-temperature structural materials, TiAl-based intermetallic alloys have emerged as a focal point due to their exceptional specific strength, specific modulus, and retained properties at elevated temperatures, alongside commendable creep and oxidation resistance. These attributes herald their widespread adoption in aerospace and automotive applications, such as turbine blades and engine valves. However, the inherent brittleness and complex solidification behavior of TiAl alloys pose significant challenges for near-net-shape casting, with casting defects being a primary concern that directly impacts component integrity and performance. This article delves into the intricate relationship between casting methods and the formation and distribution of casting defects, specifically macro-shrinkage and micro-porosity, in TiAl alloy shaft-like castings. Through experimental investigation and process simulation, I aim to elucidate how gravitational and centrifugal casting techniques influence these casting defects, offering insights for process optimization.
The alloy system under investigation had a nominal composition (in wt.%) of 48% Al, 2% Cr, 2% Nb, with the balance being Ti. Melting was conducted in a water-cooled copper crucible vacuum induction furnace to minimize contamination. After complete melting and homogenization via superheating, the melt temperature was carefully adjusted downward before pouring to mitigate reaction with the mold and control solidification shrinkage. Two fundamental casting approaches were employed: gravity casting and centrifugal casting. For gravity casting, both top-pouring and horizontal side-gating configurations were used. In centrifugal casting, the mold was rotated about a vertical axis at speeds between 260 and 300 rpm, with a horizontal side-gating system. The shaft casting geometry was a simple rod, facilitating analysis of defect distribution across its cross-section.
A critical aspect of TiAl casting is its substantial volumetric change during phase transformation. My experiments involved measuring the linear solidification contraction of simple cylindrical castings (15 mm diameter × 18.5 mm height) poured at various temperatures. The data revealed a pronounced dependency of contraction on melt superheat. This relationship can be effectively captured by a quadratic polynomial, as presented below:
$$ s = \left(74.2096 – 0.09458t + 3.06575 \times 10^{-5} t^2\right)\% $$
Here, \( s \) represents the percentage linear shrinkage, and \( t \) is the melt temperature in degrees Celsius. This equation is valid for the studied temperature range of approximately 1595°C to 1685°C. The shrinkage magnitude reaches 1.4% to 1.8%, which is significant for precision casting. This macro-shrinkage is intrinsically linked to the development of micro-porosity, a pervasive casting defect in these alloys. The microstructure invariably shows dispersed micro-voids or porosity, especially in regions that solidify last.
The distribution of these casting defects, however, is not merely a function of alloy properties but is profoundly influenced by the casting process itself. To quantify this, I conducted metallographic analysis on transverse sections taken from various locations along the length of shaft castings produced by the three different processes. The location of porosity clusters, indicative of the last-to-freeze zones, was mapped. The results are summarized in Table 1, which contrasts the characteristic porosity distribution for each method.
| Casting Method | Gating Configuration | Characteristic Porosity Location (Relative to Casting Axis) | Relative Porosity Severity | Remarks on Post-HIP Behavior |
|---|---|---|---|---|
| Gravity Casting | Top-Pouring | Concentrated along the central axis (symmetric) | High | Symmetrical shrinkage during HIP minimizes distortion. |
| Gravity Casting | Horizontal Side-Gating | Shifted upwards from the axis (vertical asymmetry) | High | Moderate risk of distortion due to vertical asymmetry. |
| Centrifugal Casting | Horizontal Side-Gating | Deviated horizontally towards the “back-flow” side (significant radial asymmetry) | Moderate (reduced volume) | High risk of bending after HIP due to pronounced asymmetric shrinkage. |
As Table 1 illustrates, gravity casting tends to centralize these casting defects along the thermal centerline, albeit with a slight upward shift when side-gating is used due to thermal asymmetry. Centrifugal casting, while successfully reducing the overall volume of porosity—a common benefit from enhanced feeding pressure—introduces a severe radial asymmetry in defect distribution. This asymmetric distribution of casting defects is the root cause of a critical manufacturing problem: shaft castings undergoing Hot Isostatic Pressing (HIP) to heal porosity often warp or bend, leading to rejection. The bending is most severe near the ingate, correlating with the maximum deviation of the porous zone from the geometric axis.
To unravel the mechanism behind this asymmetric defect formation in centrifugal casting, I performed a fluid flow simulation using wax as an analog for molten TiAl. The setup involved a transparent tube rotating vertically, simulating the mold cavity. The filling sequence was captured and analyzed. The observations revealed a non-axisymmetric filling pattern: the melt initially advances along the leading wall of the tube (the side facing the direction of rotation), creating a thin, forward-moving stream. The cross-sectional area of this stream decreases initially before stabilizing. The trailing wall is filled subsequently by a back-flow after the melt front reaches the far end. This results in a distinct thermal history across any given transverse section. The region near the leading wall experiences earlier cooling and solidification onset, while the back-flow region remains liquid longer. Consequently, the last point to solidify, and hence the location where shrinkage porosity concentrates, is displaced from the geometric center towards the trailing “back-flow” side. This deviation is not constant along the length; it is most pronounced near the ingate where thermal gradients are steepest during initial filling.
This phenomenon can be described conceptually using a simplified model for solidification time. If we consider a cylindrical coordinate system (\(r\), \(\theta\), \(z\)) attached to the rotating mold, the local solidification time \(\tau\) at a point is a function of heat extraction, which is influenced by the local filling time \(t_f(r, \theta)\). A first-order approximation for the shift \(\Delta r\) of the last-to-freeze zone from the axis can be related to the difference in filling times between the leading (\(\theta=0\)) and trailing (\(\theta=\pi\)) sides:
$$ \Delta r \propto \left[ t_f(\pi) – t_f(0) \right] \cdot \frac{\partial T}{\partial t}_{solid} $$
Where \(\frac{\partial T}{\partial t}_{solid}\) is the average cooling rate during solidification. The filling time difference is maximized at the ingate (\(z \approx 0\)), explaining the observed location of maximum distortion. The presence of these process-induced casting defects underscores the need for precise control.
The quantitative relationship between process parameters and shrinkage further complicates the picture. Beyond the basic shrinkage formula, the effective feeding distance and pressure gradient in centrifugal casting can be modeled. The centrifugal pressure \(P_c\) at a radius \(r\) from the rotation axis is given by:
$$ P_c = \frac{1}{2} \rho \omega^2 (r^2 – r_0^2) $$
Here, \(\rho\) is the melt density, \(\omega\) is the angular velocity, and \(r_0\) is the radius of the free surface. While this pressure aids in feeding and reduces overall porosity volume, it does not eliminate the thermal asymmetry caused by the filling sequence. In fact, the rapid initial flow along the leading wall can enhance heat loss there, exacerbating the asymmetry. A comparative summary of the thermal and pressure conditions for the two main processes is provided in Table 2.
| Process Factor | Gravity Casting (Top-Pour) | Centrifugal Casting | Impact on Casting Defects |
|---|---|---|---|
| Primary Driving Force | Gravitational head pressure | Centrifugal pressure (~\(\rho \omega^2 r\)) | Centrifugal provides higher feeding pressure, reducing total porosity volume. |
| Filling Pattern | Generally axisymmetric, depending on gating | Highly non-axisymmetric; leading wall first | Non-axisymmetric filling causes asymmetric thermal fields and porosity distribution. |
| Thermal Gradient Direction | Primarily radial, from mold wall to center | Radial but skewed towards trailing side | Skewed gradient shifts the thermal center and last-to-freeze zone. |
| Final Porosity Location | At or near geometric center (thermal center) | Displaced from center towards trailing side | Asymmetric location leads to non-uniform densification and bending during HIP. |
| Susceptibility to Microporosity | High due to limited feeding pressure | Moderate due to enhanced feeding | Casting defects are less severe in volume but critically mislocated in centrifugal. |
The HIP process, intended to eliminate casting defects, applies isostatic pressure at high temperature to collapse and diffuse porosity. For a shaft with axisymmetric porosity, the volumetric shrinkage during HIP is uniform, preserving geometry. However, for a shaft with off-axis porosity, the material densification is non-uniform. The region with higher initial porosity (the off-axis zone) undergoes greater volumetric strain (\(\epsilon_v\)) during HIP compared to the denser axis region. This differential strain \(\Delta \epsilon_v\) over the cross-section induces bending curvature \(\kappa\), which can be approximated for small deflections by a simple beam theory relation:
$$ \kappa \approx \frac{\Delta \epsilon_v}{h} $$
Where \(h\) is the distance between the centroid of the porous zone and the neutral axis of the shaft. This directly links the asymmetry of casting defects to post-process distortion.
Given these findings, mitigating the problem of asymmetric casting defects in centrifugal casting is paramount for components requiring HIP. My analysis suggests several strategic approaches:
1. Increase Centrifugal Rotational Speed: Higher rotational speed (\(\omega\)) increases the centrifugal pressure, which can potentially reduce the cross-sectional area of the initial forward stream. A thinner stream may lose heat more rapidly, but the increased pressure might also improve back-filling dynamics. The goal is to reduce the time lag \(\Delta t_f = t_f(\pi) – t_f(0)\). An optimal speed may exist that minimizes asymmetry without causing other issues like turbulence.
2. Elevate Mold Preheating Temperature: A hotter mold reduces the initial heat extraction rate from the forward stream, slowing its solidification. This allows more time for temperature equalization across the section before the onset of significant solidification, thereby reducing the thermal asymmetry. The mold temperature \(T_{mold}\) becomes a critical control variable to manage the thermal gradient \(\nabla T\).
3. Modify Gating Design: Implementing a different gating system that promotes more symmetrical filling could be explored. For instance, a symmetrical bottom-filling system in centrifugal casting might help, though it introduces complexity.
4. Process Selection Based on Post-Processing: For shaft castings destined for HIP, gravity casting—despite its higher overall porosity—might be preferable if the porosity remains symmetric. The symmetric casting defects can be healed by HIP without causing distortion. Conversely, for as-cast applications where minimal porosity is critical and distortion is less concern, centrifugal casting is superior.
To further quantify the potential of the first two mitigation strategies, consider a simplified thermal model. The temperature field \(T(r,\theta,t)\) in the mold can be described by the heat conduction equation with a convective boundary condition representing the melt flow. The filling asymmetry introduces an initial condition where \(T(r,0,t_{fill}) > T(r,\pi,t_{fill})\). Preheating raises the initial mold temperature \(T_0\), reducing the initial heat flux \(q” = h_{int}(T_{melt} – T_0)\), where \(h_{int}\) is the interface heat transfer coefficient. The effect of rotational speed on filling time difference can be inferred from the balance of centrifugal and viscous forces. A dimensional analysis suggests the Reynolds number \(Re = \frac{\rho \omega R D}{\mu}\) (where \(R\) is mold radius, \(D\) is tube diameter, \(\mu\) is viscosity) influences the flow regime and filling profile.
An integrated approach to modeling these casting defects would involve coupling fluid flow, heat transfer, and stress analysis. The formation of microporosity itself is often modeled using criteria functions based on local thermal parameters, such as the Niyama criterion adapted for TiAl alloys. The local porosity fraction \(f_p\) might be related to the local solidification time \(t_{SL}\) and temperature gradient \(G\):
$$ f_p \propto \frac{1}{G \sqrt{t_{SL}}} $$
In centrifugal casting, both \(G\) and \(t_{SL}\) become functions of \(\theta\), leading to \(f_p(\theta)\), explaining the angular variation in casting defects.
In conclusion, my investigation underscores a fundamental trade-off in casting TiAl alloy shafts. Gravity casting yields symmetric casting defects along the axis, which, while voluminous, are manageable for post-HIP treatment. Centrifugal casting reduces the overall severity of casting defects but introduces a dangerous radial asymmetry in their distribution, leading to component warpage during subsequent HIP. This asymmetry is a direct consequence of the non-axisymmetric filling sequence inherent to horizontal centrifugal casting of elongated sections. The most severe manifestation of these casting defects occurs near the ingate. Therefore, the choice of casting method must be intimately tied to the intended post-processing route. For applications requiring HIP, gravity casting or fundamentally redesigned centrifugal processes that ensure symmetrical thermal histories may be necessary. Future work should focus on advanced simulation and experimental validation of the proposed mitigation strategies, such as optimizing the combination of rotational speed and mold temperature to minimize the angular deviation of the last-to-freeze zone. Only through such a nuanced understanding of the process-property relationships governing casting defects can the full potential of TiAl alloy precision castings be realized.