Solid-state nucleation is one of the most important processes used to create strong Al alloys. The number density of precipitates nucleated sets a lower limit of the characteristic length scale of the microstructure affecting dislocation motion and controlling the strength. Making stronger alloys by precipitate strengthening invariably means refining the precipitate distribution by catalysing more nucleation.

The starting point for describing solid-state nucleation is usually classical nucleation theory (CNT) [1] and this is the model used in most solid-state precipitation kinetics models (whether they be in commercial programs such as TCPRISMA [2] or MatCalc [3], or in private academic codes such as the KWN-type class models, e.g. [4]). Anyone who has used these codes knows that the nucleation model needs to be tuned to experimental measurements of precipitate number densities and the extrapolative ability is poor. In reality, these CNT models are curve-fits.

Recently, a complementary description of solid-state nucleation was proposed [5]. This new approach is based on the simple idea that nucleation occurs in the matrix at locations where the local composition corresponds to one of the thermodynamically permissible phases. Since even a perfectly random solid solution is non-uniform at the length scale of the nuclei, local regions of solute enrichment (or depletion) exist in all solutions, and these ‘geometrically necessary’ clusters are proposed as sites for nucleation. Instead of assuming the nuclei form stochastically in the matrix in the framework of CNT, this new approach assumes the local sites with the correct chemistry already exist as statistical features of the solid solution. This new approach is especially applicable to solid-state nucleation at low temperatures where atomic mobility (and stochastic cluster formation) is limited.

In this presentation, we consider the number densities of precipitates formed in 2xxx, 6xxx and 7xxx Al alloys and examine how well these number densities can be predicted by this new model for solid-state nucleation. The important effect of Al enrichment (solvent trapping) in the nuclei composition is discussed. Very dilute alloys, such as Al-Sc and Al-Zr are also considered and we attempt to draw conclusions regarding where each of the different types of alloys sit on the nucleation spectrum bounded at one end by CNT and at the other by the geometric cluster model.

References

[1] HI Aaronson, JK Lee, The kinetic equations of solid-solid nucleation theory and comparisons with experimental observations, in Lectures on the Theory of Phase Transformations, TMS, Warrendale, PA, 1999.
[2] https://thermocalc.com/products/add-on-modules/precipitation-module-tc-prisma/
[3] https://www.matcalc.at
[4] M. Perez, M. Dumont, D. Acevedo-Reyes, Implementation of classical nucleation and growth theories for precipitation, Acta Materialia, 2008, 56, 2119–2132.
[5] CR Hutchinson, YJM Brechet, A new approach to solid-state nucleation in kinetically constrained systems, Acta Materialia, 2025, 283, 120521.