Skip to main content

Simulating radioactive decay and carbon dating in JavaScript

Updated 2 min readMarkdown

Radioactive decay is random for one atom but predictable for many, and a 20-line Monte Carlo simulation shows it. I simulate 100,000 carbon-14 nuclei, compare the count with N = N₀e^(−λt), and then turn a measured fraction into an age. (I build PractiSci, where this runs as a free virtual lab.)

What is the half-life, in code?

Carbon-14 has a half-life of 5730 years, so λ = ln 2 / 5730 per year. Each nucleus has the same chance of decaying in a time step, p = 1 − e^(−λΔt), no matter how old it is.

const lambda = Math.LN2 / 5730;
const N0 = 100000;
let seed = 12345;
const rand = () => (seed = (seed * 1664525 + 1013904223) % 4294967296) / 4294967296;

const step = 500;                          // years per step
const p = 1 - Math.exp(-lambda * step);
let n = N0;
for (let t = 0; t <= 11460; t += step) {
  let d = 0;
  for (let i = 0; i < n; i++) if (rand() < p) d++;
  n -= d;
}

Does the simulation match the formula?

Years Simulated nuclei N₀e^(−λt)
0 100000 100000
2500 73955 73903
5000 54570 54616
7500 40408 40363
10000 29911 29829

The two agree within about 0.3%. The small difference is the statistical noise you expect from random decays, and it shrinks as N₀ grows.

How do I get an age from a measured fraction?

Invert the law: t = −ln(N/N₀)/λ.

const age = (frac) => -Math.log(frac) / lambda;
console.log(age(0.5).toFixed(0), age(0.25).toFixed(0));  // 5730, 11460

A sample with half its carbon-14 left is 5730 years old, and with a quarter left it is 11,460 years old (two half-lives).

FAQ

Can I predict when one atom will decay?

No. Only the probability per unit time is fixed, which is why we speak of half-life for large samples.

Is the decay constant the same as the half-life?

No. λ = ln 2 / T½, so they are inversely related.

Where can I try it?

The free radioactive decay lab lets you choose the isotope and watch the count fall, and the carbon dating lab turns a measured ratio into an age.

No comments yet. Be the first to comment.

More from this blog