metastability: one construct, many systems

In materials science and physical chemistry, metastability is the condition of a phase that sits in a local free-energy minimum and so persists for long times, even though a thermodynamically more stable phase is available; reaching the stable phase requires surmounting an energy barrier. Wilhelm Ostwald's 1897 step rule first framed the idea that a system often forms the nearest metastable phase before the true equilibrium one, as set out in R. A. van Santen's analysis The Ostwald step rule in The Journal of Physical Chemistry. The referent is a long-lived non-equilibrium phase of matter, not a transient voltage hovering near a logic threshold.[1]

Unit

One metastable phase, characterized by its excess free energy above equilibrium

Recovery means

Identify a phase that persists yet transforms once an energy barrier is overcome (retained austenite, diamond, glass)

Measured by

Detected by calorimetry and phase diagrams, with the barrier inferred from lifetime to transformation

In digital electronics, metastability is the condition in which a bistable element such as a flip-flop, after a setup or hold violation, settles at neither a valid logic 0 nor a valid 1 but hovers at an intermediate level for an unbounded time before resolving. Thomas Chaney and Charles Molnar documented the effect in Anomalous Behavior of Synchronizer and Arbiter Circuits in IEEE Transactions on Computers. The referent is a circuit's indeterminate electrical state and its resolution-time statistics, not a long-lived phase of matter.[2]

Unit

One metastable event, characterized by its resolution time

Recovery means

Observe a synchronizer output that resolves late or glitches after an asynchronous input violates setup/hold

Measured by

Quantified by the resolution time constant and mean-time-between-failures (MTBF)