Skip to content

Phase-Diagram Mapping

Coexistence-mapping model — instantiates Controlled Demixing and Domain Formation

Charts where a mixture stays mixed, where it turns metastable, and where it spontaneously splits — the coexistence landscape every other demixing move steers by.

Before you can tune a demixing process you need a map of the terrain: which combinations of temperature, composition, and other controllable conditions leave the substrate happily mixed, which leave it metastable (mixed but poised to separate if disturbed), which make it unstable (separating on its own), and what the coexisting phases look like once it splits. Phase-Diagram Mapping builds that map — empirically, computationally, or both — and draws the boundaries on it: the binodal that divides one-phase from two-phase, and the spinodal inside it that divides metastable from unstable. It is the one mechanism here that produces knowledge of the landscape rather than an action on the substrate; every trigger, quench, and arrest downstream is a trajectory drawn on this chart. Its defining move is delineating the boundaries and coexisting states — not crossing them.

Example

A bioprocess team wants to purify a protein by partitioning it into one layer of a two-phase aqueous mixture of PEG and phosphate salt — an aqueous two-phase system. Which formulations actually split into two layers? They titrate: fix a PEG concentration, add salt in steps, and record the composition at which the clear solution first resolves into two coexisting layers. Repeating across concentrations traces the binodal curve — below it, one phase; above it, two. They add tie-lines connecting the compositions of the two coexisting phases, so that for any overall mix the lever rule gives the volume and composition of each layer. The finished diagram tells them, before a single purification run, which formulations land safely in the two-phase region with the phase ratio they want — and which sit so close to the binodal that a small temperature drift would collapse the split.

How it works

  • Locate the binodal. Sweep the controllable conditions and detect the onset of separation (cloud point, turbidity, a calorimetric signature) to find where one phase becomes two.
  • Find the spinodal. Distinguish the metastable interior from the unstable interior by probing how the mixture responds to a perturbation — the boundary between "separates only if nucleated" and "separates on its own."
  • Connect the coexisting phases. Draw tie-lines and mark critical points, so the diagram predicts not just whether it splits but into what.
  • Keep it an equilibrium object. It says where the system wants to go, deliberately leaving how fast and by which morphology to the kinetic mechanisms.

Tuning parameters

  • Resolution vs. effort — how finely the condition grid is sampled. Denser sampling resolves narrow two-phase windows and critical points but multiplies experiments or compute.
  • Empirical vs. computational — measured cloud points versus a thermodynamic model (activity coefficients, interaction parameters). Models extrapolate cheaply; measurements catch what a model omits.
  • Choice of axes — temperature, composition, pH, pressure, salt: the diagram is only as useful as the conditions you chose to vary; a hidden variable left off-axis distorts everything on it.
  • Equilibration time — how long each point is allowed to settle before it's read. Too short and you map a kinetic artifact, not the true boundary.

When it helps, and when it misleads

Its strength is that it converts "will this mixture separate, and into what?" from trial-and-error into a read off a chart, and it becomes the shared reference every other mechanism steers by — a quench aims for a region on it, an extraction targets a tie-line on it.

Its central failure mode is that it is an equilibrium map, and the classic misuse is reading it as if it predicted the pathway — assuming that landing in the two-phase region delivers the equilibrium phases promptly and cleanly, when kinetics can trap the system in a metastable morphology for a very long time.[1] Mapped boundaries also drift with impurities and with any condition left off the axes. The discipline is to treat the diagram as the map of destinations, not the route, and to pair it with the kinetic mechanisms that govern how the system actually travels.

How it implements the components

  • phase_coexistence_map — its primary output: the chart of mixed, metastable, unstable, and coexisting regions across controllable conditions.
  • stability_margin_and_separation_criterion — the binodal and spinodal it draws are the criterion for whether (and how forcefully) separation occurs, and the margin any operating point sits from them.

It does not drive the system across those boundaries — that is Temperature or Composition Quench and the other triggers — nor measure how constituents actually partitioned once split (that is Composition-Partition Assay). A map names destinations; it does not travel to them.

Notes

Phase-Diagram Mapping is a static artifact, built once and revised when the formulation changes; it is distinct from the Phase-Boundary Monitor, which watches where the current batch sits relative to these boundaries during a run. The map defines the boundaries; the monitor reports your live position on it. Keeping them separate lets one hard-won diagram be reused across many runs.

References

[1] Equilibrium phase diagrams omit time entirely, which is exactly why metallurgy also keeps time–temperature–transformation (TTT) diagrams — they add the kinetic axis the equilibrium diagram leaves out. Reading a coexistence map as a kinetic prediction is the error TTT diagrams exist to prevent. ↩