Rubber Elasticity: Entropy, Cross-Links, and the Gent Upturn

Entropy, not stretched bonds, does the pulling

Stretch a steel spring and you are bending atomic bonds away from their lowest-energy separation; let go and the bonds pull back toward that minimum. Stretch a rubber band and almost none of that happens — the polymer chains between cross-links are already coiled at random, and pulling them straighter does not raise their bond energy so much as it lowers their entropy: there are vastly fewer ways for a chain to stay nearly straight than to wander in a random coil. Rubber elasticity is overwhelmingly an entropic effect, which is also why warming a stretched rubber band makes it pull harder, the opposite of a metal spring.

The Neo-Hookean model: the simplest network theory

The statistical theory of rubber elasticity treats the material as a network of freely-jointed chains between fixed cross-links and derives a stress–stretch law directly from the entropy of a Gaussian chain. The simplest strain-energy density that comes out of this — Neo-Hookean — depends on a single material constant set by the cross-link (network chain) density:

W = (G/2) * (I1 - 3)                  strain energy density
sigma = G * (lambda - 1/lambda^2)     uniaxial true stress

  G      = shear modulus  ~  n * k_B * T
  n      = number density of network chains between cross-links
  lambda = stretch ratio (length / original length)
  I1     = lambda^2 + 2/lambda   (first strain invariant, uniaxial)

That G ~ n·kB·T relationship is the single most useful fact in rubber elasticity: stiffness comes directly from cross-link density and temperature, not from any intrinsic bond stiffness, so a more densely cross-linked rubber (more sulfur in a vulcanised tyre, for instance) is stiffer purely because it has shorter, more constrained chains between junctions.

Where Neo-Hookean breaks down: Mooney–Rivlin and Gent

Neo-Hookean theory fits real rubber well at small-to-moderate stretch but systematically undershoots the measured stress at larger stretch. Mooney–Rivlin adds a second invariant term with an empirical constant C2, which mostly corrects the moderate-stretch region and is still widely used for engineering seals and tyres because it fits with only two parameters. Neither model, though, captures the sharp upturn seen just before rubber tears, because an ideal Gaussian chain has no maximum length — real chains do. The Gent model fixes exactly that by building in a finite chain extensibility limit Jm and letting the stress diverge as the stretch approaches it:

W_Gent = -(G * Jm / 2) * ln(1 - (I1 - 3)/Jm)

  Jm -> limiting value of (I1 - 3) as chains approach full extension
  as (I1 - 3) -> Jm,  W -> infinity   (captures the stiffening upturn)

Why the upturn matters for real parts

That late stiffening is not a curve-fitting nuisance — it is a safety feature. As a rubber seal or a tyre sidewall approaches the strain where its network chains are nearly fully extended, the sharply rising stress resists further stretch well before the material tears, giving a soft, forgiving response over most of its working range and a stiff warning right before failure. Neo-Hookean and Mooney–Rivlin fits, calibrated only on the working range, will silently under-predict stress near failure and are a common source of over-optimistic seal and gasket designs.

Cross-link density in practice

Vulcanisation — heating raw rubber with sulfur — is literally the act of setting n, the cross-link density, by forming sulfur bridges between polymer chains. Lightly cross-linked rubber (soft, low G, large extensibility Jm) behaves like a rubber band; heavily cross-linked rubber (high G, low Jm) behaves like a hard, low-stretch gasket material, and the same statistical framework — just with different n and Jm — describes both.

Frequently asked questions

Try it live

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