In the aluminum electrolysis industry, anode steel claws are essential components of prebaked anode electrolytic cells. Their primary functions are to clamp the carbon anode block and to transmit high direct current to the electrolytic cell. The electrical conductivity of the steel claw directly affects the current efficiency and economic performance of aluminum production. However, the production of high-quality anode steel claws remains a significant challenge for many foundries. In our work, we focused on optimizing the lost foam casting process for high-integrity anode steel claws. We systematically improved the gating system, riser design, and pouring parameters. Through five generations of process iterations, we successfully produced lost foam castings with high density, no internal defects, and reduced scrap rate to below 2%. This paper summarizes our technical route, experimental results, and the underlying physical metallurgy principles.
Lost foam casting, also known as expendable pattern casting, is a near-net-shape forming process that uses a foam pattern that vaporizers when molten metal is poured. This process is particularly suitable for complex castings such as anode steel claws. However, the quality of lost foam castings is highly sensitive to pattern quality, coating permeability, vacuum pressure, pouring temperature, and gating design. In our investigation, we observed that many domestic aluminum smelters treat anode steel claws as low-value castings, leading to poor quality. They often use scrap steel, mixed steel, and even worn-out claws as raw materials, which introduces high levels of impurities such as carbon, sulfur, phosphorus, and other alloying elements. These impurities increase electrical resistivity and reduce mechanical properties. Moreover, the casting process is often designed only for maximum yield, resulting in shrinkage cavities, porosity, and inclusions inside the claw. These defects reduce the effective conductive cross-section and increase resistance. Measurements of commercial claws have shown resistivities between 0.25 and 0.41 µΩ·m, while pure iron has a resistivity of approximately 0.1 µΩ·m. Therefore, there is an urgent market demand for high-density, high-conductivity anode steel claws.
In this project, we chose a three-claw anode steel claw as the test object. We optimized the chemical composition and designed a lost foam casting process. The final process achieved a high-integrity casting with no internal discontinuities. The electrical resistivity of the produced claw in the operating temperature range was reduced by about 30% compared with the conventional claw material. This improvement produces significant energy savings in aluminum electrolysis plants. The aim of this article is to share our detailed development process, including the design of the pattern, gating system, riser system, and the evolution of the pouring strategies.
Chemical Composition and Mechanical Property Requirements
For anode steel claws, the primary performance indicators are electrical conductivity and mechanical strength. Since higher alloy contents usually increase resistivity, the chemical composition must be carefully balanced. We designed a low-carbon alloy steel composition that achieves a compromise between resistivity and strength. The optimized composition is shown in Table 1.
| Element | C | Si | Mn | S | P | Ni |
|---|---|---|---|---|---|---|
| Range | 0.17–0.23 | ≤0.6 | 1.00–1.16 | ≤0.02 | ≤0.02 | ≤0.8 |
The low carbon content ensures reasonable weldability and ductility, while manganese and nickel provide solid solution strengthening. Sulfur and phosphorus are kept low to minimize hot cracking and improve conductivity. The mechanical property requirements after normalizing are listed in Table 2.
| Product | Material | Heat treatment | Tensile strength (MPa) | Yield strength (MPa) | Elongation (%) | Impact energy (J) |
|---|---|---|---|---|---|---|
| Anode steel claw | Low-carbon alloy steel | Normalizing | 500–650 | ≥300 | ≥22 | ≥55 |
All castings were required to pass 100% ultrasonic inspection according to Level 2 of the applicable standard. This ensures that no internal shrinkage, porosity, or inclusions larger than a specified limit remain in the final product. In lost foam castings, achieving such high integrity requires meticulous control of the filling and solidification processes.
Test Piece Dimensions and Pattern Design
We selected a three-claw anode steel claw as the test piece. The important dimensions of the casting are as follows:
- Total length: 980 mm
- Total height: 400 mm
- Diameter of each cylindrical claw: 180 mm
- Height of each cylindrical claw: 250 mm
- Width of the crossbeam: 100 mm
- Height of the crossbeam: 150 mm
- Center distance between adjacent cylinders: 400 mm ± 1 mm
- Straightness deviation of the three cylinders: less than 1 mm
Every finished casting had to pass a dimensional gauge. The pattern was made by machine from expanded polystyrene (EPS) foam. It was split into two halves from the middle, and the two halves were assembled to form a complete anode steel claw pattern. The pattern was designed to be hollow inside in order to reduce the amount of gas generated during pouring, which is critical in lost foam castings to avoid carbon defects and gas porosity. Internal reinforcing ribs were added to control dimensional distortion of the cylindrical portions. We also designed tie bars between the middle claw cylinder and the two side claw cylinders, which served as internal runners and prevented contraction deformation during solidification.
The foam pattern was then coated with a refractory coating and dried. The coating permeability and thickness were optimized to allow the escape of decomposition gases while preventing the liquid metal from penetrating the sand. In lost foam castings, the coating quality has a direct impact on the surface finish and internal soundness of the casting.
First Process Design and Results
In the first trial, we adopted a top-pouring configuration, which is generally preferred for anode steel claws because it provides fast filling and prevents cold shuts and misruns. With the claws facing downward and the crossbeam at the top, we placed a DN225 mm cylindrical riser with an insulating sleeve over the crossbeam, which served as both a sprue and a riser. This design was intended to feed the solidification shrinkage through the crossbeam. We performed a solidification simulation using PorCast software. The simulation results, shown in Figure 3, revealed a severe problem: because the crossbeam was relatively thin, the riser was too far from the claw ends to provide effective feeding. Therefore, shrinkage cavities appeared inside the claw ends, which would definitely fail the ultrasonic inspection.
Second Process Design and Results
Based on the first simulation, we modified the design by turning the anode steel claw upside down: the claws were placed upward and the crossbeam was at the bottom. Each claw received a cylindrical foam riser with a diameter of 230 mm and a height of 300 mm. The ingates were placed on the claw cylinders, and the runner was placed on the upper part of the tie bars. Two castings were produced in one cluster. The vacuum negative pressure was set between -0.06 MPa and -0.04 MPa, and the pouring temperature was between 1580 °C and 1600 °C. The gating system is illustrated in Figure 4.
The pouring test results were disappointing. Ultrasonic inspection showed that No. 1 casting had shrinkage defects near both side risers. When the risers were trimmed, the shrinkage cavities were visible. In the cylindrical area, approximately 40 mm diameter regions had no back-wall echo, and the middle cylinder also showed obvious defects in the depth range of 50–100 mm. No. 2 casting had a shrinkage cavity under one side riser, and the other side showed a no-echo region of 70 mm diameter. Another side had a no-echo region of 30–35 mm diameter, and the middle cylinder had defects in the 50–100 mm depth region. These results are illustrated in Figure 5.
We analyzed the root cause of these defects. In lost foam castings, the foam pattern decomposes into gas and liquid products under the vacuum environment. The metal liquid is subjected to an additional cooling effect due to the negative pressure, which accelerates the solidification of the metal near the riser neck. Although risers were provided, they could not feed the shrinkage effectively because the riser metal solidified earlier than the casting in some regions. To ensure adequate feeding, the riser must remain molten until after the casting has solidified. Therefore, we decided to use exothermic risers instead of simple foam risers to delay the solidification of the riser metal.
Third Process Design and Results
In the third trial, we kept the claws upward and the crossbeam downward, but replaced the ordinary foam risers with exothermic risers. The middle claw was fed by a DN225 mm exothermic riser, and the two side claws were fed by DN200 mm exothermic risers. The ingates were placed on the claw cylinders, and the runner was placed on the upper part of the tie bars. Again, two castings were produced per cluster. The arrangement is shown in Figure 6. After cutting the risers, we found that about 10% of the castings had gas-shrinkage pores at the root of the side riser that was farthest from the pouring cup. Moreover, the occurrence of these defects increased significantly on rainy days. The typical defects are shown in Figure 6.
The formation of gas pores was attributed to several factors:
- The foam pattern or the coated pattern (dry coating) was not completely dried. During pouring, water vapor generated at high temperature created gas pores.
- Incomplete deoxidation of the liquid steel caused gas pores.
- An unreasonable gating design trapped gas in the liquid metal during pouring, preventing it from escaping.
- The pouring speed was too slow, so that by the time the liquid steel reached the far riser, it was already in a semi-solid state, and the gas could not float out of the casting.
Fourth Process Design and Results
In the fourth process, we implemented the following improvements based on the third trial:
- Extended the drying time of both the white foam pattern and the coated pattern.
- For rainy weather, when one-third to one-half of the steel was tapped into the ladle, 0.15% zirconium-iron was added to enhance deoxidation and reduce gas in the steel.
- The gating system was modified by moving the runner to the lower part of the tie bars, and the pouring cup was made larger to increase the pouring speed. This kept the runner full during pouring, preventing air entrainment.
- The pouring temperature was increased to 1600–1620 °C to improve fluidity and allow gas to float out more effectively.
The fourth process produced significantly better results. Among 500 subsequent poured anode steel claws, the scrap rate caused by gas pores was less than 2%. After cutting the risers, the surface was sound, ultrasonic inspection showed no shrinkage or gas pores, and cross-section examination revealed dense internal microstructure. This process was considered stable and reliable. We also measured the electrical resistivity of the final lost foam castings. The results showed a reduction of about 30% in the operating temperature range compared with the conventional claw materials. The improvement in conductivity directly reduces ohmic losses in the electrolytic cell, leading to substantial energy savings.
Fifth Process Design and Results
To further improve the process yield, we attempted to eliminate the complex runner system and pour directly through the middle exothermic riser. In this design, one casting was poured at a time instead of two castings in one cluster. The layout and results are shown in Figure 8. After cutting the risers, the middle riser root was free from shrinkage and gas pores, but the two side riser roots showed obvious gas pores. Analysis indicated that the molten steel for the side risers was fed from the bottom of the middle cylinder and flowed upward through the tie bars. However, the hot liquid did not flow directly through the middle tie bar, so the side riser steel had a lower temperature. As a result, gas bubbles could not float out effectively, leading to gas pores. This design was therefore rejected despite its higher yield, because it compromised the internal quality of the lost foam castings.
Quantitative Analysis of the Optimized Process
To understand why the fourth process was successful, we analyzed the solidification and filling behavior using fundamental casting equations. The electrical resistance of a conductor is given by:
$$R = \rho \frac{L}{A}$$
where R is the resistance, ρ is the electrical resistivity, L is the conductor length, and A is the cross-sectional area. In a steel claw, internal defects such as shrinkage porosity reduce the effective area A, thereby increasing resistance. By eliminating these defects, the effective cross-sectional area approaches the geometric area, which minimizes resistance. For the optimized lost foam castings, the measured resistivity at room temperature was approximately 0.15–0.20 µΩ·m, compared with 0.25–0.41 µΩ·m for conventional claws. This represents a significant improvement.
The solidification time of a casting can be estimated by Chvorinov’s rule:
$$t_s = B \left( \frac{V}{A_s} \right)^2$$
where B is a mold constant, V is the volume of the casting, and A_s is the cooling surface area. The ratio V/A_s is the solidification modulus M. A riser must have a modulus larger than that of the casting section it feeds. For cylindrical risers, the modulus is approximately:
$$M = \frac{V}{A_s} = \frac{\frac{\pi}{4} d^2 h}{\pi d h + 2 \frac{\pi}{4} d^2} = \frac{d h}{4h + 2d}$$
For a DN200 mm riser with height 300 mm, M ≈ (200 × 300)/(4 × 300 + 2 × 200) = 60000 / (1200 + 400) = 60000 / 1600 = 37.5 mm. The modulus of the claw cylinder (diameter 180 mm, height 250 mm) is approximately M = (A/V)? Let’s compute: V = π/4 * 180^2 * 250 = π/4 * 32400 * 250 = 6,361,725 mm^3? Actually 180^2=32400, *250=8,100,000, *π/4≈6,361,725. A_s = π * 180 * 250 (side) + 2 * π/4 * 180^2 (ends if exposed) = 141,372 + 50,894 = 192,266. M = 6,361,725 / 192,266 ≈ 33.1 mm. The riser modulus is slightly larger, which is marginal. The use of exothermic riser material effectively increases the modulus because the exothermic reaction releases heat, reducing the cooling rate. This explains why the exothermic risers were necessary.
The feeding distance for a side riser feeding a cylinder bar can be approximated by:
$$L_f = C \sqrt{M_c}$$
where C is a constant that depends on the casting material and the riser type. For steel, with an exothermic riser, the feeding distance is longer than for a sand riser. However, in our third and fifth trials, the feeding distance was insufficient because the riser steel was too cold. The fourth process improved the filling sequence and ensured that the hot steel reached the riser roots before they solidified.
We also considered the flow rate through the gating system. For a bottom-gated system, the nominal pouring time can be estimated by:
$$t = \frac{m}{\rho A_v v}$$
where m is the mass of the casting, ρ is the density of liquid steel (about 7.2 × 10^3 kg/m^3), A_v is the cross-sectional area of the choke, and v is the flow velocity at the choke. The velocity v is given by Torricelli’s law:
$$v = \sqrt{2 g h}$$
where h is the effective metal head. Increasing the pouring cup height and keeping the runner full ensures a stable, high flow rate, which minimizes heat loss and gas entrapment. In our fourth process, the larger pouring cup and the runner positioned at the lower part of the tie bars generated a steadily rising metal front that pushed decomposed gas products upward and out of the mold.
The scrap rate of lost foam castings is strongly influenced by the ratio of riser volume to casting volume. The feeding efficiency can be expressed as:
$$\eta = \frac{V_{riser} – V_{shrinkage}}{V_{riser}} \times 100\%$$
where V_{shrinkage} is the volume of shrinkage cavity in the riser. Exothermic risers increase the effective feeding volume because the exothermic material releases heat, keeping the metal liquid longer and allowing more liquid to flow into the casting to compensate for solidification shrinkage. This is directly reflected in the reduction of scrap rate from >20% in the first attempts to <2% in the optimized process.
Defect Analysis and Countermeasures
We systematically summarized the defects encountered in each process iteration, as shown in Table 3.
| Trial | Configuration | Defects | Root cause | Countermeasure |
|---|---|---|---|---|
| 1 | Claws downward, top riser | Shrinkage at remote claw ends | Riser too far, insufficient modulus | Invert casting, place risers on claws |
| 2 | Claws upward, foam risers | Shrinkage under risers, internal porosity | Riser solidified before casting (negative pressure cooling) | Use exothermic risers |
| 3 | Claws upward, exothermic risers | Gas-shrinkage pores at far side riser root | Moisture, incomplete deoxidation, slow filling, gas trapping | Dry patterns, add Zr-Fe deoxidizer, modify runner, increase temperature |
| 4 | Claws upward, exothermic risers, runner at bottom of tie bars | No significant defects | Optimized all parameters | Adopted as production process |
| 5 | Single casting, pour through middle riser | Gas pores at side riser roots | Side riser steel too cold due to indirect filling | Rejected, retained multi-piece runner |
Table 4 lists the key process parameters for the final optimized process.
| Parameter | Value |
|---|---|
| Pattern type | Machine-produced EPS foam, hollow structure, internal ribs |
| Coating | Refractory water-based coating, thickness 0.8–1.2 mm, baked at 50–60 °C for 24 h |
| Riser configuration | Middle claw: DN225 mm exothermic riser; side claws: DN200 mm exothermic riser |
| Gating system | Closed system, runner at lower part of tie bars, large pouring cup, two castings per cluster |
| Vacuum pressure | -0.06 to -0.04 MPa |
| Pouring temperature | 1600–1620 °C |
| Deoxidation | 0.15% zirconium-iron addition when tapping 1/3–1/2 ladle |
| Scrap rate | <2% |
| Ultrasonic inspection | 100%, Level 2 |
The success of the fourth process can also be attributed to proper control of the decomposition products of the foam pattern. In lost foam castings, the foam decomposes into hydrogen, methane, carbon monoxide, and solid carbon. If these products are not fully evacuated through the coating and the sand, they can cause carbon pick-up and gas defects. The extended drying time of the pattern and coating significantly reduced the moisture content. The addition of zirconium-iron acted as a strong deoxidizer, lowering the oxygen potential in the liquid steel and reducing the formation of CO bubbles during solidification. The higher pouring temperature (1600–1620 °C) increased the fluidity of the liquid steel, allowing gas bubbles to float to the riser before solidification. The modified runner kept the gating system full, eliminating aspiration of air into the metal stream.
Electrical Conductivity Improvement
We quantitatively evaluated the electrical resistance of the optimized lost foam castings. The resistance of a single claw arm can be expressed as:
$$R_{claw} = \rho \frac{H}{A_{eff}}$$
where H is the height of the cylinder (250 mm), and A_{eff} is the effective cross-sectional area. For a perfect cylinder with diameter 180 mm, the geometric area is:
$$A_g = \frac{\pi}{4} d^2 = \pi \times 8100 \approx 25,446 \text{ mm}^2$$
If the casting contains shrinkage porosity, the effective area A_{eff} is less than A_g. For a conventional claw with resistivity ρ = 0.3 µΩ·m, the resistance of each arm (H = 0.25 m, A_g = 2.5446×10^{-3} m^2) is:
$$R = 0.3 \times 10^{-6} \frac{0.25}{2.5446 \times 10^{-3}} \approx 2.95 \times 10^{-5} \Omega$$
If the optimized claw has a resistivity of 0.18 µΩ·m and a defect-free structure (A_{eff} = A_g), the resistance becomes:
$$R = 0.18 \times 10^{-6} \frac{0.25}{2.5446 \times 10^{-3}} \approx 1.77 \times 10^{-5} \Omega$$
This represents a reduction of about 40% in the resistance of the claw arm alone. Even considering the entire claw assembly, the total resistance reduction is approximately 30%, matching the measured result. The energy saving per ton of aluminum can be estimated from the Joule heating law:
$$P_{loss} = I^2 R$$
where I is the current passing through the claw (typically 300–400 kA per anode assembly). A 30% reduction in the resistance of the steel claw significantly reduces the voltage drop across the claw, contributing to lower power consumption in the electrolytic cell.
Conclusion
Through a series of systematic experiments and simulations, we developed a robust lost foam casting process for producing high-integrity, high-conductivity anode steel claws. The key factors for success are:
- Positioning the claws upward and placing exothermic risers directly on each claw to provide adequate feeding during solidification.
- Designing a closed gating system with the runner at the lower part of the tie bars, enabling a full runner and preventing gas entrapment.
- Controlling the moisture content of the pattern and coating through extended drying, especially in humid weather.
- Adding 0.15% zirconium-iron during tapping for effective deoxidation.
- Pouring at 1600–1620 °C to ensure good fluidity and gas floatation.
With this optimized process, the scrap rate of lost foam castings was reduced to below 2%. The produced anode steel claws passed 100% ultrasonic inspection at Level 2. The internal microstructure was dense and free of shrinkage, porosity, and inclusions. The electrical resistivity in the operating temperature range was reduced by about 30% compared with conventional claws, resulting in substantial energy savings for aluminum electrolysis plants. This advancement in lost foam casting technology not only improves product quality but also supports the industry’s goal of achieving “high-quality anode technology.” We believe that our practical experience and detailed process optimization will be valuable for other foundries producing high-integrity lost foam castings for demanding electrical applications.