Granular Jamming: How a Pile of Discs Turns Solid

A pile of sand is not a phase of matter — until it is

Pour sand from your hand and it flows like a liquid. Stop pouring and let it settle into a pile, and the same grains now support your weight like a solid. Nothing chemical changed — no bonds formed between the grains, no crystal lattice appeared. What changed is purely geometric: the grains ran out of room to rearrange around each other. This is jamming, and it is one of the cleanest examples of a system acquiring rigidity from packing constraints alone rather than from any attractive force.

The simplest laboratory for studying it is not sand but a 2D packing of frictionless, monodisperse hard discs confined in a box. Slowly shrink the box (equivalently, add more discs at fixed box size) and track the packing fraction φ — the fraction of the box area actually covered by discs.

The critical packing fraction φ_J

Below a critical density, discs have just enough slack to slide past their neighbours under any applied shear — the packing flows, however slowly, and carries no static shear stress. Above that density, every disc is geometrically locked against its neighbours: there is no path of rearrangement that doesn't require overlapping discs, so the packing resists shear like a solid. For frictionless, monodisperse discs in two dimensions this threshold sits at the random close packing fraction:

φ_J ≈ 0.84   (2D, monodisperse, frictionless discs — point J)

φ < φ_J   →  packing is UNJAMMED: flows, zero shear modulus, zero pressure
φ > φ_J   →  packing is JAMMED: finite shear modulus, finite pressure

This is not a rounding of some smooth curve — right at φ_J both the shear modulus and the bulk modulus turn on from exactly zero, and just above it they grow as a power law of the excess density, φ − φ_J. The corresponding number for 3D frictionless spheres is φ_J ≈ 0.64, the number every physics student has heard called "random close packing." Both numbers come from the same underlying idea: jamming, not crystallisation, of a disordered assembly of hard particles.

Isostaticity: why the transition is so sharp

The sharpness of the jamming point has a precise mechanical explanation due to James Clerk Maxwell's counting argument for rigidity: a packing of N frictionless discs has 2N degrees of freedom (in 2D) and needs at least 2N contact forces to be mechanically rigid and statically determinate. Right at the jamming point, the mean number of contacts per particle, the contact number z, sits exactly at this isostatic value:

z_iso = 2·d          (d = spatial dimension: z_iso = 4 in 2D, z_iso = 6 in 3D)

z < z_iso  →  under-constrained: "floppy modes" exist, structure can deform for free
z ≥ z_iso  →  over-constrained: no free deformation, packing is rigid

Right at φ_J the packing has exactly z_iso contacts per particle on average, no more and no less — just barely enough to pin every particle in place. This is why jamming looks like a genuine critical point, with excess contact number Δz and excess modulus both scaling as clean power laws of φ − φ_J, and it is the same mathematical language used for rigidity percolation in the spring-network article on this site.

Force chains: rigid but not uniform

Once jammed, the packing does not distribute an applied load the way a continuous elastic solid would. Contact forces concentrate along thin, filament-like force chains that snake through the packing along whichever contacts happen to be favourably oriented, while particles just off these chains carry almost no load. Photoelastic disc experiments — where stressed discs polarise light and light up under crossed filters — make these chains directly visible, and they explain why granular materials can support enormous localised loads (a single boot print on sand) while remaining globally soft and easily disturbed elsewhere.

The jamming phase diagram

Andrea Liu and Sidney Nagel proposed in 1998 that jamming should be viewed as one corner of a much broader unifying phase diagram, with three axes: density (1/φ), temperature, and applied shear stress. Along each axis, increasing density, decreasing temperature, or decreasing shear stress all push a disordered system toward the same rigid, jammed state — which is why foams, emulsions, colloidal glasses, and granular packings, despite being physically very different materials, share the same critical scaling near their respective jamming points. Zero-temperature granular matter is simply the axis where thermal fluctuations are irrelevant and packing geometry alone decides the outcome.

Simulating the compression

A minimal simulation integrates soft-disc overlap forces (a linear or Hertzian repulsion that only switches on when two discs overlap) with damped molecular dynamics, and slowly shrinks the periodic box or grows the disc radii while measuring the total pressure. The pressure stays at machine-zero below φ_J and rises sharply the instant discs are forced into overlap above it — that pressure onset is the cleanest way to locate φ_J numerically without ever counting contacts by hand.

// pairwise soft-disc repulsion, only active on overlap
for each pair (i, j):
  dr = distance(i, j);  overlap = (r_i + r_j) - dr;
  if (overlap > 0) {
    f = k * overlap;              // linear spring, k = stiffness
    applyForce(i, j, f, direction = (pos_i - pos_j) / dr);
  }
// pressure ≈ sum of overlap forces × contact distances / (2 × box area)

Frequently asked questions

Try it live

Everything above runs in your browser — open Granular Jamming and change the parameters while it is running. Nothing is installed, nothing is uploaded, the whole model lives in one tab.