Fracture splitting has become an established manufacturing route for connecting rods because it replaces sawing, grinding, and alignment machining of the rod–cap joint with a controlled crack. A laser-cut or wire-cut notch is placed in the big end, a splitting load is applied, and the crack propagates through the intended fracture zone. The resulting interlocking serrations provide a natural mating surface for reassembly. The advantage is a short process chain and a joint that self-aligns, but the material requirement is unusual: the rod body must resist fatigue, bending, and impact, while the fracture zone must fail in a brittle, low-ductility mode without excessive deformation of the big-end bore.
This functional conflict is difficult to satisfy with one steel. Commercial fracture-splitting rods therefore rely on a narrow set of materials such as C70S6, powder metallurgy grades, or malleable irons. A bimetal construction separates the functions: a low-carbon steel carries the body loads, and a high-carbon steel occupies the splitting plane. Investment casting is an attractive way to create such a blank because it can produce a near-net-shape rod and can promote metallurgical bonding between the two steels through a hot ceramic shell. The same slow cooling that favors carbon diffusion also encourages the high-carbon insert to form coarse pearlite and a grain-boundary cementite network, which is undesirable in the body but useful in the fracture zone.
Functional Partitioning of Body and Cracking Zone
The body material in this route is a Chinese 20 steel, equivalent in composition to an AISI 1026-type low-carbon steel. The insert is T10A high-carbon steel with about 1 wt% C. The body provides the toughness and tensile ductility needed for engine service; the insert provides the low ductility needed for cleavage-dominated splitting. The insert is placed across the big-end split plane and extends beyond the body so that it can be located and supported in the ceramic shell.
The manufacturing problem is therefore not simply “cast a connecting rod”. It is a coupled thermal and atmospheric problem: the gating system must avoid shrinkage porosity, gas entrapment, cold shuts, and hot spots; the insert must remain above the body solidus long enough for carbon diffusion without melting or losing its position; and the insert must not oxidize or decarburize during shell calcination. These three requirements conflict, and the process window is defined by where they overlap.
Gating Strategy and Casting Simulation
Two gating layouts were compared with finite-element casting simulation. In the top-gated layout, liquid metal enters from above and a circular riser is placed at the big end. In the side-gated layout, metal enters from the side through three ingates and the sprue acts as the feeding channel; no separate riser is used. Because the simulation package handled one cast alloy at a time, the high-carbon insert was represented as a chill with its own initial temperature. This approximation captures heat extraction from the body metal and the thermal history of the insert, but it does not solve the interfacial concentration field directly.
The top-gated simulation showed an unstable filling pattern. The stream impinged on the big end and the insert, wetting one side of the insert while the other side remained unfilled for part of the pour. Cold shuts and gas entrapment appeared at the small end and near the insert. The riser also solidified before it could feed the big end: at 88.4 s the riser had reached about 80% solid fraction while the big end was only about 46% solid. A persistent hot spot formed near the insert, with local temperature reaching about 1500 °C at 45% solid fraction while the surrounding big end was at about 1456 °C and 80% solid. The local shrinkage-porosity indication near the insert reached 64%, and the maximum solidification time was 533.8 s. The maximum effective stress was 356.6 MPa near the insert and riser. Although this is below the 500 MPa tensile strength of the low-carbon body steel, the combination of stress concentration, hot spot, and feeding deficiency made the top-gated design unacceptable.
The side-gated layout filled from the bottom upward and remained stable, with no significant splashing, cold shuts, or gas entrapment at the small end and insert. The insert region solidified early: at 10.9 s it had reached about 77% solid fraction, which favors a sound bond before the rest of the rod begins to shrink. At 36 s the remaining sections began to solidify from the outside inward. At 231.1 s the sprue was about 40% solid while the casting was about 86.7% solid, confirming that the sprue continued to feed the rod. Shrinkage porosity was confined to the pouring cup and sprue; the rod body itself was sound. The maximum effective stress was 444.5 MPa at the ingate, a location that is removed after casting. The higher numerical stress is therefore not detrimental to the rod, because it lies in the gating rather than in the load-bearing section. The side-gated cluster also produced two rods per pour.
Table 1. Comparison of top-gated and side-gated layouts.
| Feature | Top-gated layout | Side-gated layout |
|---|---|---|
| Filling pattern | Top-down, impingement on big end and insert, unstable | Bottom-up, symmetric, stable |
| Cold shuts and gas entrapment | At small end and insert side | No significant occurrence in the rod |
| Feeding | Riser solidified before big end; hot spot near insert | Sprue fed the casting; outside-in solidification |
| Shrinkage location | Severe near insert; local indication up to 64%; solidification time 533.8 s | Confined to pouring cup and sprue |
| Maximum effective stress | 356.6 MPa near insert and riser | 444.5 MPa at ingate, removable |
| Assessment | Rejected | Preferred |
Thermal Parameter Optimization
The side-gated design was then optimized with a three-factor, three-level orthogonal simulation. The factors were pouring temperature (1580, 1600, and 1620 °C), insert preheat temperature (400, 500, and 600 °C), and shell preheat temperature (1100, 1150, and 1200 °C). The responses were maximum shrinkage-porosity index, maximum effective stress, and maximum deformation. The volume ratio was fixed at 12:1 and the insert was flat.
Table 2. Orthogonal simulation array and responses.
| Run | Pouring temperature (°C) | Insert preheat (°C) | Shell preheat (°C) | Max shrinkage-porosity index | Max effective stress (MPa) | Max deformation (cm) |
|---|---|---|---|---|---|---|
| L1 | 1580 | 400 | 1100 | 0.682 | 441.3 | 0.0375 |
| L2 | 1580 | 500 | 1150 | 0.695 | 422.4 | 0.0362 |
| L3 | 1580 | 600 | 1200 | 0.674 | 424.7 | 0.0365 |
| L4 | 1600 | 400 | 1150 | 0.691 | 444.5 | 0.0375 |
| L5 | 1600 | 500 | 1200 | 0.675 | 373.9 | 0.0341 |
| L6 | 1600 | 600 | 1100 | 0.683 | 398.2 | 0.0357 |
| L7 | 1620 | 400 | 1200 | 0.671 | 392.1 | 0.0349 |
| L8 | 1620 | 500 | 1100 | 0.677 | 439.7 | 0.0373 |
| L9 | 1620 | 600 | 1150 | 0.682 | 416.2 | 0.0367 |
Range analysis showed the same ranking for all three responses: shell preheat had the largest effect, pouring temperature was second, and insert preheat was third. The combined balance selected run L5: 1600 °C pouring temperature, 500 °C insert preheat, and 1200 °C shell preheat. In L5, shrinkage porosity was confined to the pouring cup and sprue. The maximum effective stress was 373.9 MPa at the ingate, and the maximum deformation was 0.0341 cm at the small end, the bottom of the sprue, and the insert region. The coldest node in the insert reached 1437 °C and remained above 1430 °C for more than 60 s, satisfying the thermal condition for interfacial diffusion while avoiding bulk melting of the insert.
Interface Diffusion and Insert Design Window
The body steel has a solidus near 1430 °C, while the T10A insert has a solidus near 1370 °C and a liquidus near 1473 °C. For metallurgical bonding, the practical thermal criterion is that the coldest part of the insert should reach the body solidus and remain above it long enough for carbon diffusion. Temperature distribution in the insert is not uniform. In the width direction, the center reaches about 1436 °C and stays above the solidus for more than 100 s, while the edges cool faster because they are closer to the shell and receive less liquid-metal heat. In the length direction, the center reaches about 1475 °C and remains above the solidus for about 200 s, while edge nodes stay near the solidus for only about 90–120 s. In the vertical direction, upper nodes are slightly hotter than lower nodes because they are closer to the pouring cup and ingate. The overall cold point is an edge node far from both the cup and the ingate. This node is therefore the limiting location for diffusion and is used as the thermal gauge for insert design.
The ratio of liquid body-metal volume at the big end to insert volume controls how long the insert is heated. A ratio that is too low gives insufficient diffusion; a ratio that is too high can melt the insert locally and destroy its position.
Table 3. Effect of body-to-insert volume ratio on the coldest insert node.
| Body-to-insert volume ratio | Peak temperature at coldest insert node (°C) | Time above 1430 °C | Maximum solidification time (s) | Assessment |
|---|---|---|---|---|
| 6:1 | 1385 | 0 | 466 | Insufficient diffusion |
| 12:1 | 1430 | About 20 s | 478 | Marginal but acceptable threshold |
| 24:1 | 1452 | About 180 s | 490.4 | Strong diffusion, but the hottest insert node approaches the 1473 °C liquidus |
The lower bound is therefore about 12:1. Below this ratio, the coldest insert region does not reach the body solidus and the two steels do not bond. The upper bound is about 24:1. At 24:1, the hottest insert node approaches the liquidus, so part of the insert may melt, flow, and lose its intended position. The usable window is 12:1 to 24:1, with the final value set by the required big-end dimensions. The splitting notch is commonly wire-cut with a minimum width around 0.2 mm, so the insert must also be thicker than the notch while still satisfying the volume-ratio window.
Insert geometry provides another degree of freedom. At a fixed volume ratio of 12:1, a serrated insert increases the contact area between the liquid body metal and the solid insert. The serrations raise the coldest-node peak from 1430 °C to 1445 °C and extend the time above the solidus from about 30 s to about 135 s. The maximum solidification time increases only slightly, from 478 s to 482.5 s. The serrated form also creates a tortuous interface that is more resistant to crack opening and retains heat in its recesses.
Table 4. Flat and serrated inserts at a 12:1 volume ratio.
| Insert geometry | Peak temperature at coldest node (°C) | Time above 1430 °C | Maximum solidification time (s) | Assessment |
|---|---|---|---|---|
| Flat | 1430 | About 30 s | 478 | Marginal |
| Serrated | 1445 | About 135 s | 482.5 | More robust diffusion and mechanical interlocking |
The interval between placing the insert and pouring the body metal is also bounded. After the insert is placed in the hot shell, the shell near the insert cools from about 800–924 °C at 17.28 s to 600–700 °C at 42–92 s. By 117–142 s, the shell near the insert has fallen close to the insert preheat temperature and no longer provides useful heating; the probability of oxidation also increases with time. The insert itself, preheated to 400 °C, rises to about 450 °C immediately after placement, reaches about 560 °C within 20 s, peaks near 600 °C at about 88 s, and then falls to about 500 °C at 300 s. The practical pouring interval is therefore 0–20 s. Intervals up to about 92 s still leave the shell warm, but they add oxidation risk without a proportional diffusion benefit.
Suppressing Insert Oxidation During Shell Processing
The first experimental sequence followed conventional investment casting practice: the insert was placed in the pattern or shell and calcined together with the ceramic shell. This failed. The high-carbon T10A insert was severely oxidized and partially decarburized, and some regions melted. Gas generation at the insert–metal interface prevented fusion, and the castings showed misruns and shrinkage. Protective coatings were tested as a possible remedy. A chromium-plated carbon steel nut lost most of its chromium to oxidation during calcination and exposed an oxidized steel surface. A nickel-plated pushpin became embrittled and fractured under slight pressure after the same thermal exposure. Neither coating provided a reliable barrier against the calcination atmosphere.
The workable solution was to reverse the sequence. The wax pattern is built with an insert-shaped protrusion of about 30 mm beyond the body, arranged symmetrically so that the shell cavity can later accept the insert. After shell building and dewaxing, the shell material at both ends of the insert cavity is removed to create openings. The shell is then calcined without the insert. During the final minute of calcination, the insert is preheated separately to 500 °C in a resistance furnace. The calcined shell is taken from the furnace, the preheated insert is quickly inserted, and both ends are sealed with refractory clay to prevent metal leakage. When the shell temperature reaches about 1200 °C, the body metal at 1600 °C is poured. The casting is held under an insulating cover, cooled, and knocked out.
This sequence keeps the high-carbon insert out of the oxidizing calcination atmosphere while still allowing it to be heated by the shell and by the incoming body metal. The trial castings produced with this route showed a continuous bimetal joint, no thermal cracks or shrinkage porosity in the rod body, and no misruns at the insert. The shell and gating still contain the usual shrinkage porosity, but these regions are removed after casting.
Microstructure and Mechanical Performance
Metallography showed a curved, continuous interface between T10A and the 20 steel body. Carbon diffused from the high-carbon side into the low-carbon side. Near the interface, the body steel contained more pearlite and less ferrite than the bulk body, while the T10A side showed a reduced secondary-cementite network. This is direct evidence of metallurgical bonding rather than simple mechanical contact.
Tensile testing of the bimetal sample gave an ultimate tensile strength of 466 MPa and an elongation of 12.60%. The fracture occurred in the 20 steel body, not at the interface. The body steel has a tensile strength range of 355–500 MPa, while T10A is in the 820–920 MPa range. Because the joint survived beyond the body strength, the interface strength exceeded the body strength under this test. Hardness traverses across the joint in a step-like manner. The interface hardness lies between the two steels and is closer to the low-carbon side, with no abrupt jump. This indicates a sound bond but also incomplete carbon homogenization; forging or heat treatment is still needed to broaden the diffusion zone and optimize the interface.
Fracture-Splitting Behavior
The splitting response was evaluated with a notched tensile specimen. A 60° notch with a 0.6 mm edge dimension was cut symmetrically into the insert region. The crack initiated at the notch and propagated through the high-carbon steel. The macroscopic fracture surface was relatively flat and bright, with the appearance of a transgranular brittle fracture. At higher magnification, the surface showed cleavage steps and river patterns formed by intersecting cleavage planes. These features are consistent with the low ductility required for fracture splitting, and the resulting asperities can provide interlocking during rod–cap assembly.
The body region behaved differently. Its fracture surface consisted of fine, equiaxed dimples, which is characteristic of ductile fracture. The contrast between cleavage in the insert and dimple rupture in the body confirms that the two zones performed their intended functions: the insert cracked in a brittle mode, while the body retained ductility and absorbed energy. The fracture path did not follow the bimetal interface, which is further evidence that the joint was stronger than the body in the tested condition.
Forgeability and Preferred Process Window
Investment casting produces a near-net-shape bimetal blank, but it does not remove the need for forging and heat treatment. Both the low-carbon body and the high-carbon insert form austenite at elevated temperature, so they can be forged in a common temperature range. The start-forging temperature should be kept about 150 °C below the solidus to avoid overheating, burning, and grain coarsening. The finish-forging temperature must not be too high, which would leave coarse grains and abnormal structures, or too low, which would cause local cracking. Cooling from the finish-forging temperature to room temperature resembles full annealing; the cooling rate should be controlled so that the high-carbon insert retains a brittle structure while the body retains toughness. The forging ratio must also be a compromise: the high-carbon steel tolerates less deformation than the low-carbon steel. Microalloying the body to form fine carbides offers a possible non-quenched-and-tempered route to higher body strength.
Table 5. Preferred operating window for the bimetal investment-casting route.
| Parameter | Preferred value or condition | Reason |
|---|---|---|
| Gating | Side gating with three ingates | Stable bottom-up filling; sprue feeds the casting; defects confined to removable gating |
| Pouring temperature | 1600 °C | Lowest stress and deformation in the orthogonal array |
| Insert preheat | 500 °C | Balances insert heating against oxidation and melting |
| Shell preheat | 1200 °C | Dominant factor for reducing shrinkage, stress, and deformation |
| Body-to-insert volume ratio | 12:1 to 24:1 | Below 12:1 diffusion is insufficient; above 24:1 the insert may melt locally |
| Insert geometry | Serrated | Longer time above the solidus and a more tortuous interface |
| Insert-to-pour interval | 0–20 s | Shell remains hot; oxidation exposure is short |
| Post-processing | Forge and heat treat | Homogenize carbon diffusion, close porosity, and tune brittle/ductile balance |
The route is technically coherent, but its boundaries are set by coupled thermal and atmospheric constraints rather than by a single casting parameter. The single-alloy simulation approximates the insert as a chill and does not directly solve carbon diffusion; interface quality is inferred from thermal history and then checked by metallography and mechanical testing. The lower volume-ratio bound is controlled by diffusion, the upper bound by insert melting, and the insert-to-pour interval by shell cooling and oxidation. The revised shell sequence solves the oxidation problem without relying on chromium or nickel coatings, but it requires a wax-pattern cavity with well-defined insert protrusions and refractory sealing. Within these limits, investment casting widens the material choice for fracture-splitting connecting rods and produces a bimetal blank in which the body and the cracking zone can be optimized separately.