“Pile up earth into a mountain, and wind and rain will rise from it; gather water into a chasm, and dragons will be born within. Without accumulating half-steps, one cannot reach a thousand miles; without gathering tiny streams, no river or sea can form.”Xunzi · Classical Philosopher of the Late Warring States Period
Once temperature drops below the equilibrium melting point \(T_{\mathrm{m}}\), solid already has lower bulk free energy than liquid—why can ultrapure water be cooled to \(-40\,^\circ\mathrm{C}\) and clean liquid-iron microdroplets undercool by \(\sim 300\text{ K}\) without freezing?
Adding just \(0.01\%\) \(\mathrm{TiB_2}\) inoculant turns coarse columnar aluminum ingots into fine equiaxed grains—why do fewer than \(1\%\) of particles actually nucleate grains, and why does prolonged holding before pouring cause “inoculation fade”?
Crystallizing alike from disordered fluids, why does ice always grow into sixfold dendritic snowflakes while \(\mathrm{NaCl}\) forms sharp cubes—and why are metallic solid–liquid interfaces atomically rough and rounded whereas semiconductor silicon grows flat facets?
Nose T*
vδ/D_L = 1.20
v = 4 μm/s
Without foreign interfaces, pure water relies on homogeneous nucleation (\(-40\,^\circ\mathrm{C}\) still needs \(\approx 290\) molecules crossing \(70\,k_{\mathrm{B}}T\)). Small-\(\theta\) \(\mathrm{TiB_2}\) slashes the barrier by \(f(\theta)\), yet the coarsest particles activate first and release latent heat (recalescence) that suppresses the remaining \(99\%\); settling fade is solved by late-stream inoculation.
Disordered vapor breaks rotational symmetry into hexagonal \(6/mmm\) dendritic snowflakes, while \(\mathrm{NaCl}\) breaks into cubic \(m\bar{3}m\) with \(\alpha\gt 2\) cube facets; metals (\(\alpha\lt 2\)) grow continuously on atomically rough fronts, whereas \(\mathrm{Si}\{111\}\) (\(\alpha\approx 2.7\gt 2\)) grows smooth facets.
Next Lecture (From Mechanisms to Real Castings & 3D Printing) — Solidification & Crystallization II: three ingot zones & Hunt's CET, single-crystal grain selection, Scheil microsegregation, Chvorinov riser design, and laser additive manufacturing.
1. A metal has \(\gamma=0.1\text{ J/m}^2\), \(\Delta H_{\mathrm{m}}=10^9\text{ J/m}^3\), and \(T_{\mathrm{m}}=1500\text{ K}\). At \(\Delta T=300\text{ K}\), what is homogeneous \(r^*\)? At \(\theta=45^\circ\), how do heterogeneous \(r^*\) and \(f(\theta)\) behave?
\(|\Delta G_v|=10^9\times 300/1500=2\times 10^8\text{ J/m}^3\) gives \(r^*=2\gamma/|\Delta G_v|=1.0\text{ nm}\); wetting leaves \(r^*\) unchanged while \(f(45^\circ)\approx 5.8\%\).
2. An alloy (\(C_0=2\%\), \(k_0=0.3\), \(m_L=-5\text{ K/}\%\), \(D_L=10^{-9}\text{ m}^2/\text{s}\)) solidifies at \(v=10^{-4}\text{ m/s}\). What minimum \(G_L\) avoids supercooling, and how does eutectic \(\lambda\) scale if \(v\) rises \(100\times\)?
\(\Delta T_0=|m_L|C_0(1-k_0)/k_0=23.3\text{ K}\), so \(G_L\ge v\Delta T_0/D_L=2.33\times 10^6\text{ K/m}\); from \(\lambda^2 v=K_{\mathrm{JH}}\), \(100\times\) faster \(v\) refines \(\lambda\) to \(1/10\).
3. Which statement regarding Jackson's roughness factor \(\alpha\) and thin-film heteroepitaxial growth modes is completely accurate?
Metals have \(\alpha\lt 2\) (rough continuous growth), whereas \(\mathrm{Si}\{111\}\) has \(\alpha\approx 2.7\gt 2\) (faceted); \(7\%\) misfit strain triggers SK quantum-dot assembly.