Modern aero-engine fuel adapters are highly integrated castings that combine a thin-walled shell, complex internal cavities, long slender oil passages, and strict pressure-tightness requirements. The adapter studied in my work operates at approximately 80–120 ℃, experiences severe vibration, and must withstand a 12 MPa aviation kerosene leak test for 10 min after machining. Because the component contains about 23 groups of oil pipes, 12 of which are formed directly by solidification, the casting process must produce a dense microstructure with minimal porosity. In this article I describe my research on improving the mechanical performance of ZL105A alloy fuel adapters by alloy selection, grain refinement, modification, hot isostatic pressing, deep cryogenic treatment, and optimized casting technique. I place special emphasis on the influence of cores and molds, because the slender internal passages are particularly sensitive to a family of defects commonly classified as sand foundry defect.
Keywords: ZL105A; fuel adapter; sand foundry defect; grain refinement; hot isostatic pressing; deep cryogenic treatment.
1. Background and Research Objectives
Fuel adapters in aero-engine architectures have evolved from long separate pipes to complex integrated castings. The integrated design reduces weight, shortens assembly time, and improves structural reliability. However, the resulting casting contains curved thin-walled channels, irregular cavities, and small-bore oil passages that are difficult to form with conventional sand cores. The main technical challenge is to ensure complete filling of long cores, avoid core distortion, and produce a casting with high density and gas tightness.
Aluminum-silicon casting alloys are widely used because of their good castability, corrosion resistance, and strength-to-weight ratio. The original design specified AK7Ch, which is close to ZL101. I compared AK7Ch with ZL105A and found that ZL105A provides higher strength, better elongation, fewer pinholes, and less micro-shrinkage. Therefore the work focused on ZL105A.
The key objectives of my thesis were:
- to select a suitable core material and core-manufacturing process for the complex internal oil passages;
- to optimize the ZL105A alloy composition, grain refiner addition, and modification treatment;
- to use numerical simulation for the low-pressure casting process and verify the results experimentally;
- to evaluate hot isostatic pressing and deep cryogenic treatment as means of improving mechanical properties;
- to identify and reduce the various forms of sand foundry defect that can affect long internal channels.
2. Materials and Experimental Procedure
The base materials used in the experiments were commercial-purity aluminum, Al-50Cu master alloy, Al-10Mn master alloy, pure Mg, Al-5Ti-B grain refiner, and Al-Ti-B wire. Refining was carried out with C2Cl6, and additional refining was performed with high-purity argon. Sb was added as a modifier in the form of a master alloy, and Be was added for oxidation control.
The nominal composition of ZL105A used in this work is given in Table 1.
| Element | Si | Cu | Mg | Mn | Zn | Fe | Al |
|---|---|---|---|---|---|---|---|
| Content (wt %) | 4.5–5.5 | 1.0–1.5 | 0.40–0.55 | ≤0.10 | ≤0.10 | ≤0.20 | Balance |
Melting was performed in graphite crucibles placed in an electric resistance furnace. The charge was preheated to about 200 ℃ before melting. After Al and Al-Si master alloy were melted, pure Mg was added with a bell jar at about 710 ℃. The Al-Ti-B grain refiner was added at 730 ℃, and the melt was skimmed, degassed, and refined. Tensile specimens were machined from separately cast test bars or from coupons attached to the casting. The gauge diameter was 5 mm and the gauge length was 25 mm. Engineering stress and strain were calculated as:
$$ \sigma = \frac{F}{A_0}, \quad \varepsilon = \frac{\Delta L}{L_0} \times 100\% $$
Microstructural examination was performed by optical microscopy, scanning electron microscopy, and electron backscatter diffraction. For grain-size observation, polished specimens were anodized and examined under polarized light. X-ray radiography and fluorescent penetrant inspection were used to detect internal defects. Relative density was evaluated from Archimedes measurements, and porosity can be quantified by the relation:
$$ D = \frac{\rho_{\text{casting}}}{\rho_{\text{theoretical}}} \times 100\% $$
3. Alloy Selection and Optimization
3.1 Comparison of AK7Ch and ZL105A
AK7Ch is an Al-Si-Mg alloy with low nominal strength and good castability. ZL105A contains about 1.0–1.5 wt % Cu and 0.4–0.55 wt % Mg, which produce Al2Cu and Mg2Si strengthening phases. I cast identical fuel adapters with both alloys and compared the properties. The results are summarized in Table 2.
| Alloy | UTS (MPa) | Elongation (%) | Pinhole severity | Micro-shrinkage |
|---|---|---|---|---|
| AK7Ch | 266–283 | 2.5–5.9 | Severe | Many |
| ZL105A | 347–351 | 5.4–6.4 | Mild | Few |
The better performance of ZL105A is related to the low iron content, the strengthening contribution of Cu and Mg, and the formation of Al2Cu in the final stages of solidification. This low-melting-point phase can partially feed micro-shrinkage cavities, reducing the tendency for micro-porosity. Therefore ZL105A was selected as the casting alloy for the fuel adapter.
The main strengthening phases in ZL105A can be represented by the formation reactions:
$$ \text{Al} + \text{Cu} \rightarrow \text{Al}_2\text{Cu} $$
$$ 2\text{Mg} + \text{Si} \rightarrow \text{Mg}_2\text{Si} $$
The copper-to-magnesium ratio affects the relative amount of room-temperature strengthening and high-temperature stability. For this engine fuel adapter, the working temperature of 80–120 ℃ is moderate, and Al2Cu provides a useful contribution to both strength and creep resistance.
3.2 Effect of Ti Content on ZL105A
Grain refinement is essential because coarse columnar grains reduce strength, elongation, and pressure tightness. I studied Ti additions of 0, 0.1%, 0.15%, and 0.2% using Al-5Ti-B master alloy. The Ti-rich particles Al3Ti and TiB2 act as heterogeneous nucleation substrates for α-Al. The relationship between yield strength and grain size follows the Hall-Petch equation:
$$ \sigma_y = \sigma_0 + k_y d^{-1/2} $$
where \(\sigma_0\) is the lattice friction stress, \(k_y\) is a constant, and \(d\) is the average grain diameter. A reduction in grain size therefore increases yield strength and also improves ductility by distributing strain more uniformly.
For multicomponent aluminum alloys, the growth restriction factor can be expressed as:
$$ Q = \sum_i m_i C_{0,i} \left(k_i – 1\right) $$
where \(m_i\) is the liquidus slope, \(C_{0,i}\) is the initial solute concentration, and \(k_i\) is the equilibrium partition coefficient. A larger Q value restricts grain growth and promotes a finer equiaxed structure.
The measured mechanical properties for different Ti contents after T6 heat treatment are listed in Table 3.
| Ti addition (wt %) | UTS (MPa) | YS (MPa) | Elongation (%) |
|---|---|---|---|
| 0 | 310 | 262 | 1.2 |
| 0.1 | 316 | 260 | 2.2 |
| 0.15 | 319 | 243 | 4.8 |
| 0.2 | 308 | 242 | 3.6 |
When the Ti content was increased from zero to 0.15%, the microstructure became finer and the eutectic silicon was more uniformly distributed. At 0.15% Ti, the best combination of strength and elongation was obtained. At 0.2% Ti, the refinement effect decreased due to agglomeration and settling of Ti-rich particles, which can also become a source of sand foundry defect when large particles are pushed into the liquid during filling.
3.3 Effect of Ti Addition Schedule
The addition schedule of the grain refiner strongly affects refinement efficiency. I compared a single addition of 0.1% Ti with a split addition of 0.05% Ti early in melting and another 0.05% Ti added within 40 min before pouring. The split addition schedule reduced the time during which Ti-rich particles remain in the liquid, limiting particle coarsening and settling. The results are given in Table 4.
| Addition mode | UTS (MPa) | YS (MPa) | Elongation (%) |
|---|---|---|---|
| Single addition | 312 | 266 | 2.3 |
| Split addition | 337 | 262 | 4.2 |
Grain-size analysis showed that the split addition reduced the equiaxed grain diameter from roughly 300–790 µm to 100–490 µm. The improvement is mainly due to the presence of a larger number of effective nuclei at the moment of pouring. This practice also reduced the tendency for macrosegregation and shrinkage porosity, which are common forms of sand foundry defect in large castings.
3.4 Sb Modification of Eutectic Silicon
Unmodified ZL105A contains coarse plate-like eutectic silicon that can crack at low stress. I used Sb as a modifier because Sb provides long-lasting modification and good pressure tightness. Different Sb contents from 0 to 0.3 wt % were tested. The mechanical results are summarized in Table 5.
| Sb content (wt %) | YS (MPa) | UTS (MPa) | Elongation (%) | Hardness (HV) |
|---|---|---|---|---|
| 0 | — | — | — | 69.65 |
| 0.1 | 86.24 | 107.84 | 1.2 | 77.98 |
| 0.15 | 92.93 | 164.03 | 4.64 | 80.37 |
| 0.2 | 108.64 | 205.47 | 4.72 | 81.45 |
| 0.25 | 91.58 | 130.62 | 2.08 | 80.95 |
| 0.3 | 99.64 | 138.16 | 2.0 | 78.07 |
At 0.2% Sb, the eutectic silicon became fine and fibrous, and both tensile strength and elongation reached their maximum values. Above 0.2% Sb, the microstructure showed signs of over-modification, with coarser eutectic silicon and reduced ductility. Sb also reacts with Mg to form Mg2Sb, which reduces the amount of Mg2Si and affects the balance between strength and ductility.
4. Hot Isostatic Pressing and Deep Cryogenic Treatment
4.1 Hot Isostatic Pressing
Even with optimized melting and casting, small gas pores and micro-shrinkage cavities can remain in the interior of a complex casting. These defects are extremely dangerous for a pressure-tight fuel adapter. Hot isostatic pressing applies high inert-gas pressure at elevated temperature, causing the material to creep and collapse internal voids. The densification rate can be represented by a simplified creep law:
$$ \frac{dD}{dt} = A \frac{\sigma_{\text{eff}}^n}{(1-D)^m} \exp\left(-\frac{Q}{RT}\right) $$
where \(D\) is relative density, \(\sigma_{\text{eff}}\) is the effective stress in the porous body, \(Q\) is the activation energy for creep, \(R\) is the gas constant, \(T\) is temperature, and \(A\), \(n\), and \(m\) are material constants.
I compared ZL105A specimens with and without hot isostatic pressing. The mechanical properties are listed in Table 6.
| Condition | UTS (MPa) | YS (MPa) | Elongation (%) |
|---|---|---|---|
| Without HIP | 319 | 219 | 7.6 |
| With HIP | 334 | 227 | 11.4 |
HIP clearly improved tensile strength and, more significantly, ductility. Fracture observation showed that the HIP-treated samples had many deep dimples, while the untreated samples contained visible gas pores and inclusions. Thus HIP reduces internal porosity and eliminates a critical sand foundry defect that is often invisible on the surface but becomes apparent during pressure testing.
4.2 Deep Cryogenic Treatment
Deep cryogenic treatment involves cooling the material to about -120 ℃, holding it until the temperature is uniform, and then reheating it to room temperature or directly to artificial-aging temperature. The thermal contraction and expansion can produce beneficial residual stresses and refine the precipitation state. For a constrained plate, the residual compressive stress induced by a temperature change can be approximated by:
$$ \sigma_c \approx \frac{E \alpha \Delta T}{1-\nu} $$
where \(E\) is Young’s modulus, \(\alpha\) is the thermal expansion coefficient, \(\Delta T\) is the temperature change, and \(\nu\) is Poisson’s ratio.
In my experiments, deep cryogenic treatment applied to T6-treated ZL105A samples increased the elongation from 11.2% to 16.5%, while the tensile strength increased slightly from 315 to 321 MPa. The fracture surfaces of the cryogenically treated specimens contained more numerous and deeper dimples than the untreated specimens. When deep cryogenic treatment was combined with hot isostatic pressing and optimized heat treatment, the final alloy reached an ultimate tensile strength of about 358 MPa and an elongation of 13.2%. This combination satisfied the design requirements for the fuel adapter.
Microstructural observation showed that deep cryogenic treatment did not change the overall morphology of the α-Al matrix or the eutectic silicon dramatically, but it altered the distribution of fine precipitates and reduced residual stresses. This treatment also helps reduce distortion during final machining, which is an important consideration for thin-walled channels.
5. Casting Process Development
5.1 Core Materials and Core Manufacturing
The fuel adapter contains long, small-diameter oil passages that are impossible to produce by conventional core-box methods. I therefore evaluated three core manufacturing routes: copper tube forming and welding, phenolic resin laser sintering, and furan resin 3D printing. Copper tubes have high strength but poor dimensional consistency, long production time, and very difficult removal after casting. Copper dissolution in nitric acid took about one week and was considered unacceptable.
The two 3D-printed sand cores were then compared in actual casting trials. The main results are shown in Table 7.
| Core type | Strength | Gas evolution | Collapsibility | Cleaning | Leakage before HIP |
|---|---|---|---|---|---|
| Phenolic laser-sintered core | High but variable | Light gas porosity | Poor; micro-cracks were found | Baking plus high-pressure water washing, twice | 3 of 12 tested channels leaked |
| Furan 3D-printed core | Sufficient and uniform | Heavier gas porosity if venting was poor | Good | High-pressure water washing only | None leaked |
The furan 3D-printed core had lower hot residual strength and better thermal decomposition behavior. This improved collapsibility and made core removal easier. Although furan cores generated more gas during pouring, I was able to control gas porosity by designing hollow cores, adding venting slots, and optimizing the gating system. The phenolic laser-sintered core had a tendency to re-harden after cooling, causing a severe sand foundry defect in the form of hot tears and poor surface quality. Therefore furan 3D printing was selected as the core manufacturing process.
5.2 Core Assembly and Positioning
The complex geometry of the fuel adapter required many individual cores. Some slender cores had very small core-print areas, so I designed additional core prints and process holes. In one region, seven small oil pipes were grouped close together; I combined them into one larger core to ensure dimensional accuracy and stable positioning. All large channels were made hollow to reduce gas pressure and improve collapsibility.
The key concern in core assembly is that a small positioning error can cause an eccentric oil passage and reduce wall thickness. Such location errors can also promote a sand foundry defect because the metal velocity varies around the core and causes localized erosion. The 3D-printed furan cores had good dimensional repeatability, and the assembly procedure was verified by metrology before pouring.
5.3 Melt Treatment and Oxidation Control
ZL105A is an Al-Si alloy that easily absorbs hydrogen in the molten state. Hydrogen porosity is one of the most important forms of sand foundry defect in aluminum castings. The solubility of hydrogen in liquid aluminum follows Sieverts law:
$$ [\mathrm{H}] = S(T) \, p_{\mathrm{H}_2}^{1/2} $$
where \(p_{\mathrm{H}2}\) is the partial pressure of hydrogen in equilibrium with the melt and \(S(T)\) is the Sieverts constant at temperature \(T\). Reduced moisture and efficient degassing are therefore essential.
I used a combination of C2Cl6 tablets and argon gas injection. Increasing the C2Cl6 amount by 30% and supplementing it with argon reduced the pinhole size and number substantially. The C2Cl6 produced large bubbles that were efficient at removing oxides, while the argon produced fine bubbles that were more effective at removing hydrogen. The combined treatment produced cleaner and denser alloy. The measured hydrogen content dropped from about 0.17 μg/g before degassing to below 0.05 μg/g after 10 min of inert-gas refining.
In addition, I added Be to the alloy. Beryllium oxidizes preferentially and forms a protective BeO layer that heals cracks in the surface oxide. The relevant oxidation reaction is:
$$ 2\mathrm{Be} + \mathrm{O}_2 \rightarrow 2\mathrm{BeO} $$
This reduced the amount of oxidized inclusions and improved the internal cleanliness of the casting. The microstructural examination confirmed that the addition of 0.2% Be decreased the number and size of oxide films and associated porosity.
5.4 Low-Pressure Casting Process and Simulation
For a thin-walled casting with long internal channels, I selected low-pressure counter-gravity casting. Low-pressure filling allows the mold cavity to fill slowly and reduces turbulence, which minimizes oxide entrapment and core erosion. I performed finite-element simulations using ProCAST to evaluate the filling sequence, temperature distribution, and solid fraction field. The shrinkage porosity tendency can be related to Chvorinov’s rule:
$$ t_s = B \left(\frac{V}{A}\right)^n $$
where \(t_s\) is the local solidification time, \(V/A\) is the local modulus, and \(B\) and \(n\) are mold constants. A proper temperature gradient is required to feed shrinkage and avoid micro-porosity.
The optimized pouring parameters are listed in Table 8.
| Parameter | Value |
|---|---|
| Pouring temperature | 693–698 ℃ |
| Filling speed | 35–40 mm/s |
| Pressurization during solidification | 0.04 MPa |
| Pressure increase rate | 0.002 MPa/s |
The simulation showed that the designed gating system produced a stable filling front and a favorable temperature gradient. I then used the same parameters for verification castings. X-ray radiography of the first castings showed acceptable porosity, but some φ6 mm cores were broken when the filling speed was too high. Reducing the filling speed to 35 mm/s eliminated core breakage. Fluorescent penetrant inspection confirmed that no surface cracks or misruns were present. The casting was also covered with iron sand on the upper surface, which accelerated cooling and eliminated surface micro-shrinkage.
During the casting trials, I observed that even the best core material can cause problems if the venting design is inadequate. Gas generated by the furan binder can be trapped in an enclosed pocket and create a classic gas-related sand foundry defect. To solve this, I added vent grooves, hollow cores, and a small number of process vent holes. These measures were sufficient to eliminate the gas-induced pores.
The automatic pouring station used for the verification trials is shown in the image below.
5.5 Inspection and Pressure Testing
After casting, each fuel adapter was inspected by X-ray radiography, fluorescent penetrant inspection, and hydrostatic pressure testing. The oil passages that were formed by the furan 3D-printed cores showed excellent internal quality. The combination of optimized melting, split Ti addition, Sb modification, low-pressure casting, hot isostatic pressing, and deep cryogenic treatment produced a casting with the following properties:
- Ultimate tensile strength: about 358 MPa.
- Elongation: about 13.2%.
- No visible pinholes or micro-shrinkage in critical sections.
- Successful 12 MPa / 10 min aviation kerosene leak test.
- Stable performance at 80–120 ℃ under severe vibration.
The final low-magnification inspection of the machined castings showed a dense structure in thin-walled pipes, thick sections, and complex junction regions. The microstructure was fine, with uniformly distributed eutectic silicon and almost no casting porosity. This demonstrated that the process was capable of producing fuel adapters that meet the original design requirements.
6. Discussion
The results of this work show that the improvement of a complex casting involves more than selecting a good alloy. Each step affects the final quality in a different way:
First, the choice of ZL105A instead of AK7Ch improved the strength and reduced the tendency for micro-shrinkage because of the presence of low-melting-point Al2Cu. The lower iron content also reduced the amount of brittle needle-shaped iron intermetallics.
Second, grain refinement with Al-5Ti-B must be carefully controlled. An excessive addition of Ti does not produce a continuous refinement benefit because Ti-rich particles agglomerate and settle. Splitting the addition reduces this problem and produces a more uniform equiaxed structure. A fine grain size is beneficial not only for strength and elongation but also for feeding and pressure tightness.
Third, Sb modification is important for the eutectic silicon phase. With 0.2% Sb, the eutectic silicon is transformed from coarse plates to fine fibers, and this greatly improves ductility without sacrificing strength. Sb is an attractive alternative to Na and Sr because its modification effect lasts longer and does not fade easily.
Fourth, hot isostatic pressing is an indispensable step for pressure-tight castings. It closes internal voids and improves the reliability of the oil passages. Deep cryogenic treatment adds a further benefit by improving the fracture resistance and residual stress state. Together, these two post-casting processes reduce the risk of leakage under pressure.
Fifth, the casting process itself must be designed around the core structure. The furan 3D-printed core provides good collapsibility and cleanability, but its higher gas evolution requires a robust venting design. If the venting is not adequate, the result will be a gas-related sand foundry defect. The low-pressure casting process with multiple gates and a low filling speed prevents core erosion and ensures complete filling of thin walls.
Another important example of sand foundry defect is metal penetration, which can occur when the sand core has insufficient strength or when the metal velocity is too high. My trials showed that reducing the filling speed from 40 mm/s to 35 mm/s prevented core deformation and the resulting metal penetration in the slender passages. The use of iron sand on the upper surface also reduced the local solidification time and suppressed micro-shrinkage.
The interaction between core binder and molten alloy is also a possible source of sand foundry defect. Phenolic resin cores were more prone to micro-cracking because of their higher residual strength and re-hardening behavior. Furan cores decomposed more completely and produced a friable shell that allowed the casting to shrink freely. This improved dimensional accuracy and reduced hot tearing.
From an engineering perspective, the results confirm that a robust casting process must be calibrated not only for the alloy and heat treatment, but also for the geometric complexity of the component. In particular, the long internal channels of a fuel adapter are highly sensitive to any variations in core quality, gas evolution, and metal flow. Therefore the control of sand foundry defect should be considered as a core design requirement from the early stages of process development.
7. Conclusions
Based on my research on the mechanical performance enhancement of ZL105A fuel adapter castings, the following conclusions can be drawn:
- ZL105A alloy is more suitable for the fuel adapter than AK7Ch. It provides higher tensile strength, better elongation, fewer pinholes, and less micro-shrinkage under the same casting conditions.
- Furan resin 3D-printed cores are the preferred core material for slender oil passages. They offer good dimensional accuracy, excellent collapsibility, easy cleaning, and acceptable gas evolution when the venting system is properly designed.
- The addition of 0.15% Ti as a grain refiner gives the best balance of strength and ductility. A split addition schedule, in which part of the refiner is added early in melting and part is added shortly before pouring, produces a finer and more uniform grain structure than a single addition.
- Sb modification with 0.2% Sb refines the eutectic silicon and improves the tensile properties of ZL105A. Excessive Sb causes over-modification and reduces mechanical performance.
- Hot isostatic pressing significantly increases the density and ductility of ZL105A, increasing the elongation from 7.6% to 11.4%. Deep cryogenic treatment further improves the elongation and residual stress state; the combined optimized route achieved about 358 MPa ultimate tensile strength and 13.2% elongation.
- Low-pressure casting with a filling speed of 35 mm/s, pouring temperature of 693–698 ℃, and suitable pressurization produced a sound fuel adapter. X-ray and fluorescent inspection confirmed that the casting was free of critical internal defects, and the oil passages passed the 12 MPa kerosene leak test.
- The elimination of various types of sand foundry defect was central to the success of the project. Core selection, venting design, gating design, and cooling layout all had to be optimized together with the metallurgical treatment to achieve the required mechanical and pressure-tightness properties.
In conclusion, the combination of ZL105A alloy, furan 3D-printed cores, optimized grain refinement and modification, low-pressure casting, hot isostatic pressing, and deep cryogenic treatment provides a reliable route for producing complex aluminum alloy fuel adapters for aero-engine applications. The process developed in this work can serve as a basis for transferring this technology to batch production.