A spacetime crystal is a phase of matter that is periodic in both space and time. The particles form a lattice AND they rotate coherently — a clock built from a crystal, or a crystal that happens to tick. Quantum time crystals, predicted by Wilczek in 2012 and observed shortly after, require exotic many-body systems. Liu et al. built one on a table. 3D-printed resin disks, eight millimeters wide, with staggered legs that convert vertical vibration into self-propulsion. Confine a few hundred on a shaking plate, pack them tightly enough, and a triangular lattice forms and begins to rotate as a rigid body. It persists for a day.
The interesting part is how it dies. Reduce the packing fraction and the crystal melts — but not all at once. Three stages, two distinct mechanisms, and the temporal order dies first.
Between packing fraction 0.734 and 0.709, temporal coherence collapses while spatial order holds. The lattice is still there — sixfold symmetry intact, translational order preserved — but the particles have stopped agreeing on which direction to rotate. The directional persistence decays because the many-body interactions that coordinated the rotation weaken at lower density. Clusters of coherent rotation persist in a sea of randomness, then dissolve.
Between 0.709 and 0.687, it is the lattice's turn. Spatial order breaks through a completely different mechanism: topological defect proliferation. Dislocations and disclinations nucleate, multiply, and destroy translational order while preserving local rotational symmetry — the hexatic phase. By 0.687, both orderings are gone. Pure fluid.
The two orderings sit on the same substrate. They emerge together. But they die apart, through mechanisms that have nothing to do with each other. Temporal order dies through the decay of a collective dynamical property — directional persistence. Spatial order dies through the accumulation of geometric flaws — topological defects. One is about agreement. The other is about arrangement. There is no reason these should fail at the same packing fraction, and they don't.
The label “spacetime crystal” hides a composite. What appears to be one phase is two independent symmetries riding the same substrate, each with its own critical threshold and its own mode of failure. The crystal is not a crystal that also rotates; it is a spatial crystal superimposed on a temporal crystal, maintained by different forces, destroyed by different instabilities.
This is general. Any system that looks like one thing but is actually two coupled orderings will come apart in stages, and the stages will reveal which ordering was independent and which was parasitic on the other. Here, spatial order survives the loss of temporal order. The lattice does not need the rotation. The rotation needed the lattice — specifically, it needed the dense-packing many-body interactions that the lattice creates. When those weaken, time melts first.
A crystal in space has a certain durability: you can heat it, stress it, introduce defects, and it resists through well-understood mechanisms. A crystal in time has a different durability: it resists through collective dynamics, and when the collective weakens, the oscillation dies. These are not the same kind of robustness. They are not even the same kind of order. The fact that they coincide at high packing is not a deep unity but a contingent overlap — the conditions that support spatial ordering also happen to support temporal ordering, and when those conditions degrade, the overlap dissolves before the structure does.