Experimental observation of the Mpemba effect in a granular gas
In plain words
A granular gas (many grains bouncing and losing energy in each collision) is predicted to show a hotter sample cooling below a target temperature before a cooler one, if the hot sample's speed distribution has the right shape. No experiment has yet confirmed this.
Precise statement
Homogeneous gas of inelastic hard spheres or disks with restitution coefficient $r < 1$, freely cooling or coupled to a stochastic thermostat. Kinetic theory predicts that the granular temperature curves $T(t)$ of two samples cross when their initial temperatures and initial fourth Sonine coefficients $c_{2}$ ($c_{2} = (d/(d+2)) \langle v^{4}\rangle/\langle v^{2}\rangle^{2} - 1$ in d dimensions, which vanishes for a Maxwellian) are suitably chosen. The question is whether this crossing occurs in a laboratory granular gas (vibrated monolayer, levitated or microgravity system) with measured $T(t)$ and $c_{2}(t)$ in quantitative agreement with kinetic theory.
What would settle it
Particle-tracking measurement of $T(t)$ and $c_{2}(t)$ in a homogeneous granular gas showing the predicted crossing with prepared initial $c_{2}$.
Status in the literature
Unverified note
Predicted by Lasanta et al. (2017) and studied in many kinetic-theory and simulation papers through 2026 (for example sheared gases, arXiv:2605.17881); no experimental observation was found.