Quartz Thermal Shock: Heat, Quench and Crack a Crystal

Heat a quartz crystal in a furnace, dip it in a bath and watch exact transient conduction build surface tension, the 573 C alpha-beta step and the chance of cracking.

Switch to 2D

Presets

Playback

Parameters

Camera

Layers

Quality

Model check

The science

What you see. A quartz crystal is soaked in a furnace until it is uniformly hot, lifted out by a hoist, carried over and dipped into a water bath. The crystal in the picture is a hexagonal prism; the calculation treats it as a sphere of the chosen radius (an "equivalent sphere"), which is what makes an exact solution possible. The picture is not to scale: the crystal is drawn larger than the radius slider says.

Heat flow. The temperature obeys the heat equation with convective cooling at the surface, characterised by the Biot number Bi = hR/k. For a sphere the exact solution is a Fourier–Bessel series, (T−Tw)/(T0−Tw) = Σ Cn sin(λnx)/(λnx) exp(−λn²Fo) with 1 − λn cot λn = Bi and Fo = αt/R². The diffusion time R²/α (about 4.5 minutes for a 2 cm radius) sets the pace of everything.

Stress. A cooler skin that wants to shrink is held by a hot core, so the surface is in tension and the core in compression. For a free elastic sphere with thermal strain e(T) the tangential stress at the surface is σ = E/(1−ν) (⟨e⟩ − e(Ts)), where ⟨e⟩ is the volume average; at the centre it is 2E/(3(1−ν)) (⟨e⟩ − e(Tc)). The strain contains the ordinary linear expansion and a small smooth step near 573 °C for the α–β transition of quartz. The largest possible surface stress is the fully constrained value EαΔT/(1−ν), reached only for a very fast quench (large Bi). A thermal-shock figure follows directly: ΔTc = σf(1−ν)/(Eα), about 34 K for the values used, which is why quartz is easily cracked by sudden temperature steps.

Fracture. Cracks start from surface flaws, so the chance of failure is a Weibull law, P = 1 − exp(−(A/A0)(σ/σ0)m) with the sphere's surface area A. Repeating the quench n times is treated as n independent trials. The model says nothing about how many fragments or how much dust a failure makes.

Honest simplifications. E = 95 GPa, ν = 0.17, α = 13×10−6/K, k = 4 W/(m K), ρ = 2650 kg/m³, c = 1000 J/(kg K), a flawed-crystal strength of 50 MPa and a Weibull modulus of 8 are round, textbook-scale values; real quartz is anisotropic (its expansion and conductivity depend on the crystal axis), its properties change with temperature, and every natural crystal has its own flaws. The transition step size is only approximate. The heat transfer coefficient h is a free parameter: stirred water gives a few thousand W/(m²K) but a very hot surface first sits in a vapour film that lowers it, which this model ignores. Cracks are not allowed to change the stress, there is no sub-critical crack growth, and the time spent moving between furnace and bath is neglected (the crystal starts the quench uniformly hot). The glow colour and the boiling bubbles are visual aids, not simulated.

Checks. (1) the series against an independent finite-difference solution of the radial heat equation; (2) the heat leaving the surface against the heat lost by the whole crystal; (3) the stress field satisfies radial equilibrium and a stress-free surface, and the surface value agrees with the exact series; (4) the large-Biot thin-layer limit and the closed-form critical quench step.

Loading model...
Loading Blender model...
Surface, centre and mean temperature in the bath
Surface and centre stress against the crystal strength
Fracture probability against furnace temperature
Made with 🟠 Blender · Three.js