How Crystalline works

Every specimen grows from a fluid that fills a cavity in rock. The fluid's temperature, pH, salinity and dissolved load change through time, following a script for one type of mineral deposit, or your own changes to it.

Chemistry

For each mineral the saturation index SI = log(IAP/K) is computed from the pocket fluid. Equilibrium constants are temperature-dependent fits: fluorite (Nordstrom & Jenne 1977), calcite and the carbonic-acid system (Plummer & Busenberg 1982), quartz (Fournier 1977), baryte (Blount 1977), and rhodochrosite (extrapolated from its 25 °C value), as tabulated in the WATEQ4F database and older versions of PHREEQC. Activity coefficients use the Davies equation. Carbonate ion is speciated from total dissolved carbon and pH, so raising pH or heating (calcite has retrograde solubility) precipitates calcite.

The pocket exchanges fluid with an inflow. During each step the amount precipitated is solved so that the crystals' growth rate and the depleted fluid are consistent. When crystals could grow faster than the fluid supplies material, growth is limited by transport, as is usual in nature, and the fluid stays close to saturation.

Growth and shape

Crystals are bounded by their real crystallographic forms, with face normals computed from lattice parameters and point-group symmetry (for example fluorite m3̄m: cube {100}, octahedron {111}, dodecahedron {110}; quartz 32: prism m and rhombohedra r and z). Each form advances at its own rate, and the slowest forms end up as the largest faces (kinetic Wulff construction). No real crystal is fed evenly, so each face of each crystal also gets a small fixed difference in growth rate, and cubes come out slightly lopsided. The rates depend on temperature and supersaturation, following well-documented tendencies: fluorite is commonly cubic in low-temperature deposits and octahedral in hotter ones, calcite forms equant rhombs when it grows slowly and c-elongated crystals when it grows fast, and pyrite changes from cubes to octahedra to pyritohedra as supersaturation and temperature rise (Murowchick & Barnes 1987). Quartz grows by the Rimstidt & Barnes (1980) rate law. Other rate constants are effective values calibrated so that centimetre crystals grow in 1,000 to 100,000 years.

Nucleation follows classical theory, J ∝ exp(−B / ln²Ω). It is easier on host-rock grains of the same mineral (epitaxy). New crystals may settle on earlier crystals. Fluorite can form penetration twins on [111].

The record inside

Each crystal stores its growth history: face positions and the fluid's trace-element content at each moment. A point inside belongs to the growth zone of the moment the growing crystal first enclosed it, and to the sector of the form that enclosed it. This reproduces phantoms, colour banding and sector zoning (such as amethyst colour concentrated under the r faces). Dissolution events are recorded too.

Colour

Most colours in fluorite and quartz are defect centres created by natural radiation from K, U and Th in the host rock. They need a trace element, a dose and low enough temperatures that they do not anneal. Each growth zone accumulates its own dose and anneals with Arrhenius kinetics, so colours deepen after the fluid has gone, and fade if the rock is heated. The doses at which colours saturate are chosen so that they develop within the model's shortened time, not measured. Fluorite (Bill & Calas 1978): purple (calcium colloids), blue (yttrium-associated F centres), rose (YO₂ centres), yellow (O₃⁻), green (Sm²⁺). Quartz: smoky (Al), amethyst (Fe, in quartz grown at low temperature), citrine (amethyst heated above about 450 °C). Rhodochrosite is intrinsically red from Mn²⁺. The UV lamp gives both shortwave and longwave light, as fluorescent-mineral displays do. Under it Eu²⁺ makes fluorite glow blue-violet and Mn²⁺ makes calcite glow orange-red (Yan et al. 2026).

Light

Each crystal is ray-traced in a shader. Light refracts at the surface (with the measured indices and dispersion of each mineral) and is absorbed and scattered by the zone it passes through. At every face it partly reflects (Fresnel equations, including total internal reflection) and partly leaves, with each colour leaving at its own angle. In strongly birefringent minerals (calcite, rhodochrosite) the ordinary and extraordinary rays are traced separately, with the extraordinary ray's walk-off, so calcite shows its famous double image. Faces carry their characteristic micro-relief: horizontal striations on quartz prisms, striated pyrite cubes, hillocks and terraces, and frosting after etching. Light from the case's main lamp is traced through the clear crystals too, each colour bent by its own index, and lands on the rock as bright, colour-fringed patches inside the crystals' shadows.

Grow your own specimen

Your five choices become a fluid recipe that runs on the same chemistry and growth models as the deposits. The recipe card gives the real value behind each ingredient, and "Why it works" gives the reasons, with sources. So that your choices come true on any rock, the model gives your mineral an easy start and limits how many crystals start.

Simplifications

• The Davies equation is used beyond its range (ionic strength is capped at 0.5); real brines need Pitzer models. Complexing (CaF⁺, PbCl₄²⁻) is ignored.
• Sulfide saturation (pyrite, galena) is schematic: no redox or chloride complexing.
• Carbonate constants are held at their 250 °C values above that temperature.
• Quartz solubility (Fournier 1977) is calibrated up to about 250 °C at low pressure; hotter fluids, as in Alpine clefts, extrapolate it without a pressure term.
• Dissolution is taken to be twice as fast as growth at the same distance from saturation.
• The time since the crystals grew is shortened: the clock gives each deposit's real age, but runs fast once growth has ended.
• Habit rules are empirical tendencies, not molecular models. Faces grow flat: no hopper growth or curved faces. Crystals that meet stop growing into each other, but a crystal cannot grow around another one.
• Colour-centre spectra are approximated by three-channel absorption coefficients.

Scenarios are simplified from published fluid-inclusion and paragenetic studies of each deposit type; they are illustrations, not reconstructions of particular mines. Tap "The science behind this" on any label for the sources behind a mineral or a deposit.