As an engineer engaged in casting quality control, I have encountered a serious service failure in a batch of pin rails used in a coal mine scraper conveyor. The pin rails were manufactured by the lost foam casting process, also widely known as lost foam castings, and were made of ZG35CrMnSi low-alloy cast steel. After a short period in service, the components fractured in a brittle manner in large numbers, which posed a significant risk to production safety and equipment reliability. In this study, I conducted a systematic failure investigation combining macro fracture observation, microstructural examination, chemical composition analysis, and mechanical property testing. Based on the evidence, I identified the root causes and proposed practical prevention measures. This paper highlights the critical role of process control in lost foam castings, particularly regarding slag inclusion and carburization, and provides a reference for similar cast components.
The pin rail has a nominal outside dimension of 650 mm × 250 mm × 300 mm. The material specification is ZG35CrMnSi steel, which belongs to the family of low-alloy structural steels. After appropriate quenching and tempering, this steel is expected to exhibit a good combination of strength and toughness. However, during actual service, the component experienced unexpected brittle fracture. The fracture occurred not as a single event but repeatedly across a whole batch, indicating a systematic process deficiency rather than an accidental overload. To understand the failure mechanism, I carefully examined the fracture surfaces and the associated metallurgical quality of the material sampled from the same heat used to cast the pin rails.
1. Manufacturing Process Description
Before analyzing the failure, I first reviewed the production route employed for the lost foam castings. The melting practice involved deoxidation followed by holding the molten steel at 1600–1630 °C for about 3 to 5 minutes. The pouring temperature was maintained at 1590–1620 °C, which is somewhat higher than that used in conventional sand casting, in order to promote complete gasification of the foam pattern. The pattern was fabricated from EPS sheets by cutting with hot wire and then assembling with adhesive. The EPS density was 32 kg/m³, with a bead diameter of 0.5–0.6 mm. A brown fused alumina (corundum) refractory coating, with a particle size of 200 mesh, was applied to the pattern surface to a thickness of 1.0–1.2 mm. The coating was designed to provide adequate permeability for the evacuation of decomposition gases. A side-gated pouring system was adopted. During moulding, the pattern was placed in a flask, filled with dry sand, and subjected to vibration to achieve compaction. A vacuum was then applied using a water-ring vacuum pump, with the vacuum level maintained between -0.053 MPa and -0.067 MPa. After mould preparation, the molten steel was poured into the mould, allowing the foam pattern to gasify and be replaced by the liquid metal. This process is typical for production of large steel castings, but it also introduces specific risks such as gas entrapment, incomplete pattern gasification, and carbon pickup.
2. Fracture Failure Analysis
2.1 Macro Fracture Surface Observation
I took the tensile test bars that had been cast together with the pin rails, without any subsequent heat treatment, to examine the fracture morphology. The macro fracture surface displayed a honeycomb-like appearance with coarse features. Numerous visible inclusions or slag particles were scattered on the fracture surface. There was no obvious plastic deformation before fracture. The fracture facet had a granular and dull appearance, typical of brittle failure. This initial observation strongly suggested that the material had not achieved the expected ductility and that the fracture was likely initiated by severe internal discontinuities such as inclusions and slag films. The macro fracture observations are illustrated below.
2.2 Microscopic Fracture and Metallographic Examination
For microstructural examination, I prepared metallographic specimens from the fractured area. In the unetched condition, the polished surface revealed a large amount of non-metallic inclusions, some of which were accompanied by cavities and microcracks. These inclusions appeared as clusters or chains, which are known to act as stress raisers and preferential crack initiation sites. The presence of such inclusions is particularly detrimental in lost foam castings because the foam pattern decomposes into gases and liquid residues that may become entrapped in the solidifying metal if the coating permeability or vacuum conditions are inadequate.
After etching with a 4% nitric acid alcohol solution, the microstructure was observed under an optical microscope. The matrix was not uniform; it contained bainitic and martensitic-like features along with some acicular constituents. The local alignment of the microstructure indicated possible segregation or directional solidification effects. No spheroidized or tempered structure was observed, meaning that the as-cast material retained a brittle, hard microstructure. This is consistent with the fact that the test bars were used in the as-cast condition. However, the high hardenability of Cr-Mn-Si steel can produce martensite even in as-cast sections, leading to high hardness but very low toughness. This microstructural condition, combined with the abundant inclusions, is highly likely to cause premature brittle fracture under service loads.
2.3 Chemical Composition Analysis
I took samples from the fracture location to determine the chemical composition. The results are summarized in Table 1 together with the standard specification for ZG35CrMnSi steel.
| Element | C | Si | Mn | P | S | Cu | Cr |
|---|---|---|---|---|---|---|---|
| Measured (wt.%) | 0.42 | 1.43 | 1.34 | 0.037 | 0.034 | — | 0.96 |
| Standard (wt.%) | 0.32–0.39 | 1.10–1.40 | 0.80–1.10 | ≤0.040 | ≤0.040 | ≤0.030 | 1.10–1.40 |
The measured carbon content is 0.42%, which exceeds the upper specification limit of 0.39%. Silicon and manganese are also above the allowed maximum values, while chromium is slightly below the minimum limit. Phosphorus and sulfur are within the acceptable range. The excessive carbon content is particularly noteworthy. In lost foam castings, carbon pickup is a well-known phenomenon. When molten steel comes into contact with the decomposing polystyrene pattern, carbon from the pyrolysis products can dissolve into the liquid metal at the solidification front, increasing the carbon content in the surface and sometimes even throughout the section. Since carbon increases hardenability and reduces ductility, the elevated carbon level would deteriorate the toughness of the pin rail. Manganese also contributes to solid solution strengthening and increases hardenability, but excessive manganese promotes austenite grain growth and reduces impact toughness. Moreover, manganese can combine with sulfur to form MnS inclusions, which are detrimental to ductility if they are present in large quantities. Silicon is primarily added for deoxidation and strengthening, but too much silicon can cause grain boundary segregation and reduce the cohesive strength of the grain boundaries, thereby lowering the fracture toughness. Chromium is an important alloying element in this steel for improving hardenability and wear resistance, but its low content in the failed parts reduces the desired strength and hardenability without necessarily increasing toughness. Thus, the overall compositional deviation explains the low ductility and brittle fracture behavior.
To assess the combined effect of carbon and alloying elements, I calculated the carbon equivalent using the International Institute of Welding formula:
$$ CE = w(C) + \frac{w(Mn)}{6} + \frac{w(Cr) + w(Mo) + w(V)}{5} + \frac{w(Ni) + w(Cu)}{15} $$
Substituting the measured composition, the carbon equivalent is:
$$ CE = 0.42 + \frac{1.34}{6} + \frac{0.96}{5} \approx 0.42 + 0.223 + 0.192 = 0.835\% $$
This high carbon equivalent indicates that the steel has strong hardenability and a tendency to form hard, brittle microstructures such as martensite or bainite in air-cooled castings. The as-cast structure could therefore be extremely brittle, especially when combined with severe slag contamination.
2.4 Mechanical Property Testing
I conducted tensile tests on standard round specimens with a gauge diameter of 10 mm and a gauge length of 50 mm, according to GB/T 228-2002. The maximum tensile force was 70.58 kN, giving a tensile strength of 900 MPa. The elastic modulus was measured as 74.94 GPa. The force-displacement curve showed almost no yield plateau and no necking before fracture. The elongation after fracture was only 9.4%. The hardness was measured as 235 HBW, which is acceptable for this grade. The fracture surface of the tensile specimen was identical to that of the actual failed component, confirming that the test bar represented the true material condition. Table 2 summarizes the mechanical properties.
| Property | Value |
|---|---|
| Tensile strength σb (MPa) | 900 |
| Elastic modulus E (GPa) | 74.94 |
| Elongation δ (%) | 9.4 |
| Hardness (HBW) | 235 |
| Fracture behavior | Brittle, no necking |
The elongation of 9.4% is quite low for a 35CrMnSi cast steel, which would normally be expected to exhibit an elongation of at least 15% after proper heat treatment. The as-cast condition partially explains this, but even so, the high inclusion content and chemical deviations act synergistically to reduce the ductility. The fracture surface of the tensile specimen showed the same honeycomb-like appearance with a central region containing many inclusions. The fracture initiated from the internal slag particles and propagated rapidly in a predominantly cleavage mode. The absence of a flat, fibrous tear regions confirms that the material had very limited plastic deformation capability.
I further evaluated the reduction of area (RA), which is a more sensitive indicator of ductility than elongation. The elongation and reduction of area are related by:
$$ \psi = \frac{A_0 – A_f}{A_0} \times 100\% $$
Although I did not measure the final area precisely due to the brittle fracture, the visual inspection indicated a negligible reduction of area, far below the typical value of 30% for ductile steel. This confirms the brittle nature of the material.
3. Root Cause Discussion
The failure of the pin rail can be attributed to a combination of metallurgical defects and improper process design in the lost foam castings. The most prominent factor is the massive presence of slag inclusions. There are two primary sources for these inclusions. First, the deoxidation practice in the melting process was insufficient, leaving residual oxides and reaction products in the liquid steel. Second, and more importantly, the foam pattern used in the lost foam castings is made of EPS. When the molten steel enters the mould, the EPS undergoes pyrolysis and vaporization. If the vaporization is incomplete or the generated gases cannot escape completely through the coating and mould, some of the decomposition products, especially carbon-rich residues and tars, remain in the metal. These residues can engulf solid particles and form slag-like inclusions. The observed honeycomb-like macro fracture and the clusters of inclusions in the microstructure strongly support this mechanism.
In addition to slag, the carbon pickup from EPS is a critical issue. The measured carbon content was 0.42%, above the maximum allowable limit of 0.39%. In lost foam castings, the decomposition of EPS produces substantial amounts of free carbon. The following simplified reaction can describe the pyrolysis:
$$ \mathrm{(C_8H_8)_n \rightarrow n\,C + 4n\,H_2} $$
The free carbon can dissolve into the liquid steel, leading to carburization. The extent of carbon pickup depends on the density of the EPS pattern, the pouring temperature, the coating permeability, and the vacuum level. A higher pattern density, as used here (32 kg/m³), results in more carbon residue available for dissolution. The carbon pickup not only raises the carbon content but also creates local concentration gradients that can lead to hard, brittle microstructures. The high carbon equivalent calculated earlier further amplifies the risk of cracking.
The excessive manganese and silicon contents also deteriorate the toughness. Manganese increases the hardenability of the steel, promoting the formation of martensite in cast sections, while silicon can cause grain boundary embrittlement. The low chromium content, on the other hand, reduces the amount of stable carbides and shifts the CCT diagram, making the steel more sensitive to the formation of coarse bainite. All these compositional deviations act together to lower the impact toughness and fracture resistance of the pin rail.
From a mechanical point of view, the presence of inclusions can be characterized by the Griffith criterion for brittle fracture. For an elliptical crack of length \(2a\), the critical stress is given by:
$$ \sigma_f = \sqrt{\frac{2E\gamma_s}{\pi a}} $$
where \(E\) is the elastic modulus, \(\gamma_s\) is the surface energy, and \(a\) is the half-crack length. Inclusions and slag clusters effectively act as pre-existing cracks. Their size and density determine the critical stress. The large slag particles observed in the fracture surface reduce the critical fracture stress to a level below the service stress, leading to catastrophic failure. The lack of plastic deformation indicates that the critical crack size was very small, meaning that even small inclusions triggered failure.
Moreover, the as-cast structure without heat treatment possesses high hardness but low toughness. The tensile specimen exhibited a hardness of 235 HBW, which is on the high side for a structural steel in service. The absence of a tempering treatment leaves the microstructure in a highly stressed state. The martensitic and bainitic constituents are sensitive to internal stresses and stress concentration. When the component is subjected to cyclic loading or impact loads in the scraper conveyor, crack initiation occurs very easily at existing defects and then propagates rapidly through the brittle matrix.
4. Prevention Measures
Based on the failure analysis, I have developed a series of prevention measures to improve the quality of lost foam castings and to avoid similar brittle fractures.
4.1 Strict composition control
In the charge calculation, the carbon content should be targeted to the lower limit of the specification, e.g., 0.32–0.35%, to account for the inevitable carbon pickup during lost foam casting. The carbon pickup is often erratic and can be as high as 0.05% or more. By setting a lower initial carbon content, the final carbon content after casting can be kept within the required range. Similarly, manganese and silicon should be adjusted to the middle or slightly above the nominal composition but not exceeding the upper limit. Chromium should be maintained at the upper half of the specified range to ensure adequate hardenability and strength without sacrificing toughness. A strict chemical analysis should be performed on each heat before pouring, and corrective adjustments should be made if needed.
4.2 Pattern material and structure optimization
The selection of the foam pattern material is of utmost importance in lost foam castings. The standard EPS pattern with a density of 32 kg/m³ is relatively dense and produces a large amount of liquid residue and carbon during pyrolysis. I recommend using low-density copolymers such as EPS-PMMA (polymethyl methacrylate) or pure PMMA. These materials have a lower carbon content and produce a more complete gasification with less residue. The decomposition of PMMA involves unzipping to monomers, which can escape more easily through the coating:
$$ \mathrm{(C_5H_8O_2)_n \rightarrow n\,C_5H_8O_2} $$
This reaction yields fewer solid carbon particles, thereby reducing slag inclusions and carbon pickup. Furthermore, the pattern should be designed with a hollow structure whenever possible. A hollow pattern reduces the total mass of foam material that needs to be gasified, thereby generating less gas and residue. In practice, this can be achieved by assembling the pattern from thin-walled plates or by incorporating internal cavities. The reduced gas evolution allows the gases to escape through the coating and sand more rapidly, minimizing the chance of entrapment.
4.3 Process improvements in coating and vacuum
I also suggest improving the coating permeability. The current coating thickness of 1.0–1.2 mm with 200 mesh brown fused alumina might be too dense for large steel castings. A higher permeability coating or an increased number of vent holes in the pattern may facilitate gas evacuation. The vacuum level should be optimized. In the present process, the vacuum was maintained at -0.053 to -0.067 MPa. I recommend testing higher vacuum levels, but ensuring that the coating does not crack or peel. The pouring temperature should also be carefully balanced. A higher temperature improves fluidity and aids in pattern gasification, but it also increases the risk of metal penetration and sand sticking. The optimum temperature should be determined based on the section thickness and the pattern density. In this case, reducing the pattern density and using a hollow structure would permit a lower pouring temperature, reducing oxidation and slag formation.
4.4 Alternative process: sand casting
If the quality of the lost foam castings cannot be maintained reliably, I recommend switching to conventional sand casting with resin-bonded sand or water glass sand. Sand casting does not involve a foam pattern, and therefore eliminates the problems of carbon pickup and pyrolysis residue. The pin rail geometry is not extremely complex, and sand casting can produce sound castings with appropriate gating and risering. The only disadvantage is the slightly lower dimensional accuracy and surface finish compared to lost foam castings, but for a structural component like a pin rail, this is acceptable. The elimination of slag inclusions and carbon enhancement would significantly improve the ductility and toughness of the material.
4.5 Heat treatment and quality control
Finally, I recommend that all cast pin rails should receive a full heat treatment, such as normalizing plus tempering, or quenching and tempering, before service. The as-cast condition is not suitable for this steel grade because of its high hardenability. A proper austenitizing treatment at 880–900 °C followed by oil quenching and high-temperature tempering at 560–620 °C would produce a tempered martensitic structure with excellent toughness. The heat treatment should be validated by performing tensile and impact tests on representatives from each heat. In addition, non-destructive testing (e.g., ultrasonic or radiographic inspection) should be carried out on critical areas to detect internal inclusions and cracks. A foundry should implement statistical process control for the chemical composition and the pattern quality parameters that affect the final properties.
5. Conclusion
In this paper, I have investigated the fracture failure of ZG35CrMnSi steel pin rails produced by lost foam castings. The failure was characterized by brittle fracture with a honeycomb-like fracture surface and abundant slag inclusions. The chemical composition deviated from the standard specification, with excessive carbon, manganese, and silicon, and slightly low chromium content. The tensile specimen exhibited high strength but very low elongation and no visible necking, confirming the embrittlement of the material. The root causes can be attributed to inadequate deoxidation, poor gasification of the EPS pattern leading to slag inclusions and carbon pickup, and the lack of a proper heat treatment. These factors synergistically reduced the plastic toughness of the pin rail to the point where the service stress exceeded the fracture stress, causing catastrophic failure. To prevent such failures in the future, I have recommended strict control of the initial carbon content, the use of low-density or PMMA-based patterns, hollow pattern designs, improved coating permeability, vacuum optimization, or possibly a change to conventional sand casting. Additionally, a full heat treatment and rigorous non-destructive inspection should be implemented. These measures will enhance the quality and reliability of lost foam castings for structural applications, ensuring that the components meet their service requirements without premature fracture.
As a result of this analysis, I emphasize that successful production of steel castings by the lost foam process requires a comprehensive understanding of the interaction between the foam pattern, coating, sand, vacuum, and metallurgical practices. Only by controlling each stage can the quality of lost foam castings be guaranteed. The lessons learned from this failure are of great significance not only for pin rails but also for other heavy-duty steel castings manufactured by the lost foam method.