What you’ll explore
After two half-lives, is it all gone? In this lab you watch 1000 carbon-14 nuclei decay and date a piece of charcoal and test your intuition against a real physics simulation.
Challenge levels
- Two half-lives. A sample starts with 1000 carbon-14 nuclei. After two half-lives (11,460 years), what fraction of the carbon-14 is expected to remain?
- Carbon dating. A piece of charcoal has 12.5% of the carbon-14 a living tree has. How old is it, in years?
- Stop at 10%. Set how long the sample of 1000 nuclei decays so that between 9% and 11% of the carbon-14 is left when the clock stops.
- Is it overdue?. One nucleus has already survived three half-lives. What is the chance it survives one MORE half-life? (Answer as a fraction, 0–1.)
- What if the sample is twice as big?. Start with 2000 carbon-14 nuclei instead of 1000. How many years until half of them are left?
Key ideas
Show the explanations (spoilers: try the levels first)
- Each half-life halves the remaining nuclei: 1 → ½ → ¼. So a quarter is expected to remain (in this run you can count how close the real sample lands: about 250 ± 14).
- 12.5% = 1/8 = (½)³, so three half-lives have passed: 3 × 5730 = 17,190 years. In general t = T½ · log₂(1/fraction).
- 10% = 2^(−t/T½) gives t = T½·log₂10 ≈ 3.32 half-lives ≈ 19,000 years. A real sample of 1000 lands within a few nuclei of 100, so a window of ±1% catches it.
- Decay has no memory: each surviving nucleus faces the same 50% chance over the next half-life, no matter how long it has already lasted. That is why you can never say WHICH nucleus goes next, only how many will.
- The half-life belongs to the isotope, not the sample: 2000 nuclei drop to 1000 in 5730 years, just as 1000 drop to 500. More nuclei only make more decays per year.
About the model
- Each nucleus gets its own decay time from the exponential distribution, tᵢ = T½·log₂(1/uᵢ) with uᵢ uniform in (0, 1). That is exactly “the same chance to decay in every interval, whatever the nucleus’s age” (memoryless). The count still alive at time t is the number of tᵢ greater than t.
- Half-life of carbon-14: 5730 years (the Cambridge value used for dating results; the older Libby value is 5568 years). C-14 decays by beta emission to nitrogen-14; the blue spheres are the nitrogen-14 left behind.
- A real sample holds about 10¹¹ C-14 atoms per gram of fresh carbon; the lab shows 1000 so each decay is visible. The counted sample wanders about the law by roughly ±√(N·p·(1−p)) (±16 at one half-life for 1000 nuclei); bigger samples follow the law more closely.
- Carbon dating as shown assumes the object started with the same C-14 fraction as a living thing and exchanged no carbon since. Real dates are calibrated against tree rings and other records, and the method is limited to about 50,000 years.
- The videos use a sample (seed 286) whose counts happen to land within 2 of the law at 1, 2 and 3 half-lives, so the round numbers in the captions match the screen; the long segment then shows another sample that ends 277 instead of 250, to make the randomness visible. The lab page uses seed 85.
- Every run is time-compressed to 12 s on screen; the HUD shows years. Nuclei are drawn as spheres on a grid; their positions mean nothing.
Every number on screen comes from the simulation, and each lab passes automated checks against the physics formulas. Spotted a mistake? Tell us.