TiB2 grain refiners prevent solidification cracking in continuous billet casting by changing the solidification structure before the crack can form. They act as heterogeneous nucleation sites for α-Al, promoting a columnar-to-equiaxed transition that shortens interdendritic feeding paths, disperses thermal contraction strain across many grain boundaries, and reduces segregation. The effect is measurable: controlled studies report roughly a 29% reduction in hot-cracking sensitivity with TiB2 addition, and industrial practice treats grain refinement as a baseline requirement for avoiding billet center cracks. The limiting condition is dispersion — TiB2 must be present as fine, individual particles, because agglomerates become crack-initiating inclusions rather than nucleation sites.
Why Solidification Cracking Occurs in the Last Fraction of Liquid
Continuous casting cracks are not simply a strength problem. They form in the mushy zone at high solid fraction, where dendrites have already interlocked into a coherent network but a thin liquid film remains between them. The billet is simultaneously subjected to thermal contraction, mold friction, and withdrawal tension. If the liquid film cannot be replenished as the grain boundaries separate, a crack opens.
The two governing conditions are therefore tensile strain and insufficient interdendritic feeding. Preventing the crack means improving feeding access and distributing strain more uniformly — not making the alloy “stronger” in the conventional sense. This is exactly where grain refinement intervenes.
The TiB2 Mechanism: Heterogeneous Nucleation → Equiaxed Structure → Better Feeding
TiB2 particles provide low-energy substrates for α-Al nucleation. The effect depends on synergy between TiB2 nucleation sites and solute Ti, which generates constitutional undercooling ahead of the solidification front. Together, they promote the columnar-to-equiaxed transition.
That structural change suppresses cracking through three linked effects:
1. Shorter, more distributed feeding paths. A long columnar dendritic network creates narrow, tortuous interdendritic channels. Residual liquid struggles to reach the center as solid fraction rises. An equiaxed network shortens those paths and makes feeding more uniform.
2. Lower strain concentration. Thermal contraction strain is shared across many small equiaxed grain boundaries rather than concentrated in a few long columnar boundaries.
3. Reduced segregation. Grain refinement also limits macro-segregation of solute and second phases, which directly affects hot-cracking sensitivity.
TiB2 does not prevent cracking by strengthening the billet. It prevents cracking by restructuring the solidifying network so that feeding and strain are spatially redistributed.
Quantified Effect: About 29% Lower Hot-Cracking Sensitivity
The effect is not merely qualitative. In controlled Al-Cu experiments, TiB2 addition at 3 wt.% reduced the hot-cracking sensitivity coefficient by 29%, while increasing fluidity length by 20%. Better fluidity and lower cracking are two sides of the same mechanism: improved liquid availability in the interdendritic region.
Constrained-rod casting studies on 5083, 6061, and 7075 aluminum alloys found that TiB2 grain refiner minimized hot-cracking sensitivity across all three alloy systems. This matters because the mechanism is not limited to one composition window — it applies across common wrought alloy families.
Industrial practice reflects the same conclusion: grain refinement is widely treated as necessary to avoid billet cracking, and grain refiners are added specifically to prevent center cracks.
The Boundary Condition: Dispersion Determines Whether TiB2 Helps or Harms
TiB2 only works when it is dispersed as fine particles that actually participate in nucleation. Two limits are critical.
Agglomeration. TiB2 is denser than liquid aluminum and tends to settle and cluster. Typical TiB2-related inclusions in aluminum billets are agglomerates ranging from 10 μm to 700 μm. These large clusters no longer provide useful nucleation interfaces; they become rigid internal defects. Excess Al-Ti-B addition has been associated with internal cracks in extruded billets — the opposite of the intended effect.
Low effective nucleation ratio. A commonly cited estimate is that only about 1 in 500 TiB2 particles actually nucleates α-Al. Most particles remain unused in the solidified structure. Adding more refiner does not linearly increase nucleation efficiency; it increases the population of unused particles that can collide, agglomerate, and enter the billet.
Poisoning and fading. Zr, Cr, and V can poison Al-Ti-B refiners, and refinement effectiveness fades during long melt holding. In Zr-containing alloys, TiB2 nucleation efficiency is notably reduced. TiB2 performance is therefore alloy-specific and time-sensitive, not a fixed additive effect.
Engineering Control: The Real Question Is “How Much and How Dispersed”
| Control Variable | Target Condition | Consequence of Deviation |
|---|---|---|
| Effective nucleation particle count | Sufficient to convert columnar to equiaxed structure | Too low: columnar network persists; feeding paths remain narrow |
| TiB2 dispersion | Individual particles or fine clusters, uniformly distributed | Agglomerates of 10–700 μm become internal crack sources |
| Addition level | Minimum effective amount to reach target grain size | Excess: more unused particles, higher agglomeration risk |
| Alloy compatibility | Account for Zr, Cr, V poisoning | Refinement fades; addition strategy or refiner type must change |
| Melt holding time | Cast soon after addition | Long holding causes fading and reduced nucleation effectiveness |
Published addition levels range widely — from 60 ppm B (about 300 ppm Ti) in Al-5Ti-1B practice to 1–2 kg/MT of master alloy, with target Ti concentrations around 100–200 ppm. The wide range confirms that there is no universal optimum. The process should be controlled by target grain size and dispersion quality, not by a fixed ppm value. The goal is the minimum TiB2 addition that produces a fully equiaxed structure with minimal residual agglomerates.
Key Takeaways
-
Solidification cracking occurs in the high-solid-fraction mushy zone, driven by tensile strain and insufficient interdendritic feeding. TiB2 addresses the feeding condition, not the strength condition.
-
TiB2 promotes heterogeneous nucleation and a columnar-to-equiaxed transition, shortening feeding paths and dispersing thermal strain across many grain boundaries.
-
The quantified effect is significant: about a 29% reduction in hot-cracking sensitivity in controlled studies, with minimized hot-cracking sensitivity observed in 5083, 6061, and 7075.
-
Industrial practice treats grain refinement as a baseline requirement for avoiding billet center cracks, but effectiveness depends on dispersion and active nucleation.
-
Agglomeration reverses the benefit. TiB2 clusters of 10–700 μm are crack-initiating inclusions; excess refiner has been linked to internal cracks.
-
Control should target minimum effective addition with good dispersion, accounting for Zr/Cr/V poisoning and melt holding time — not simply “more refiner.”