
model radioactive decay, explain half-life, and calculate the age of a rock or fossil from its parent-daughter ratio.
Instruction
Isotopes and decay. Atoms of an element all have the same number of protons, but can have different numbers of neutrons: these are isotopes. Some isotopes are unstable. An unstable parent nucleus decays into a daughter nucleus, releasing radiation. Uranium-238 decays through a chain to lead-206; potassium-40 decays partly to argon-40; carbon-14 decays to nitrogen-14.
Random for one atom, predictable for many. No one can say when a single nucleus will decay. But in a large sample, a fixed fraction decays in a fixed time. The half-life is the time for half of the parent atoms to decay. After one half-life, 1/2 remains; after two, 1/4; after n, (1/2)n. Half-lives are not changed by heat, pressure or chemistry, which is what makes them clocks.
| Parent → daughter | Half-life | Useful for |
|---|---|---|
| Uranium-238 → lead-206 | 4.47 billion years | Zircons, meteorites, the oldest rocks |
| Potassium-40 → argon-40 | 1.25 billion years | Volcanic rocks, from about 100,000 years old upward |
| Carbon-14 → nitrogen-14 | 5,730 years | Once-living material up to about 50,000 years |
The age equation. If a mineral started with no daughter atoms, every daughter atom now present was once a parent. So parent + daughter = original parent. The number of half-lives elapsed is n = log2(1 + D/P), and the age is t = n × half-life.
Worked example 1. A volcanic ash layer has 3 atoms of argon-40 (from decay) for every 1 atom of potassium-40 left. D/P = 3, so 1 + 3 = 4 = 22: two half-lives, 2 × 1.25 = 2.5 billion years. (A detail for the careful: only about 11% of potassium-40 decays to argon; geologists correct for this, and the problem assumes the correction is done.)
Worked example 2. A zircon crystal has 0.25 lead-206 atoms per uranium-238 atom. n = log2(1.25) ≈ 0.322, and t ≈ 0.322 × 4,468 million ≈ 1,440 million years. That is close to the ages U.S. Geological Survey geologists measured in zircon from Missouri’s St. Francois Mountains, 1.48 to 1.45 billion years; a closed zircon from the 1.47-billion-year Granite of Castor River Shut-Ins should hold about 0.256 lead-206 atoms per uranium-238 atom.
Worked example 3. A piece of charcoal from an ancient campsite has 25% of the carbon-14 of living wood: two half-lives, 11,460 years. Carbon-14 cannot date a dinosaur: after 66 million years, over 11,000 half-lives, none would remain. After 50,000 years only about 0.24% remains, which is why the method stops there. Missouri’s ice-age record sits inside that range: Missouri State Parks reports that people had reached present-day Missouri by at least 12,000 years ago, and at the Kimmswick Bone Bed a Clovis spear point was found with mastodon bones, which is about two half-lives of carbon-14 (2 × 5,730 = 11,460 years).
Choosing the right clock. A good clock has a half-life comparable to the age you are measuring, and the rock must have stayed a closed system, neither gaining nor losing parent or daughter atoms. Zircon is ideal for uranium-lead dating because it takes in uranium but rejects lead when it crystallises, and it survives erosion and metamorphism. In the lab, coins or candies stand for parent atoms: each shake, about half ‘decay’. Your data will be noisy, like real decay counts, and you will see why larger samples give smoother curves.
📖 Read for this lesson
The reading opens over the lesson; close it to come back to exactly where you were. The checks below draw on it.
Lab: Modeling Radioactive Decay with Coins or Candies
Question. Does a large set of randomly ‘decaying’ objects follow a predictable half-life curve, and how does sample size affect how closely it follows?
Materials
- 100 pennies or other coins (or 100 paper squares marked on one side, or 100 small candies printed on one side, such as M&Ms or Skittles)
- A shoebox, or a cereal bowl with a plate on top
- A tray, baking sheet or paper plate
- Graph paper or a spreadsheet
- Calculator
- Do not eat candies used in the lab: they have been handled and poured onto surfaces. Wash hands before and after.
- If anyone at home has a food allergy, use coins or marked paper squares instead of candies.
Procedure
- Count 100 ‘parent atoms’ into the box, all heads up (or printed side up). Record 100 at shake 0.
- Close the box, shake it well for three seconds, and pour onto the tray.
- Remove every tails (or unprinted) piece: these have ‘decayed’ into daughter atoms. Count and record the parents that remain and the daughters removed.
- Return only the parents to the box and repeat until no parents remain or 8 shakes have passed.
- Reset all 100 pieces and repeat the whole run until you have eight runs (at least four if time is short). Add the runs together for each shake: this is your pooled total.
- Graph run 1’s parents remaining against shake number, and on the same axes graph the pooled total scaled to 100 and the ideal curve 100 × (1/2)ⁿ.
Data
| Shake (half-life) | Parents remaining (run 1) | Daughters (total) | Pooled total parents (all runs) | Ideal 100 × (1/2)ⁿ |
|---|---|---|---|---|
| 0 | 100 | 0 | 100 | |
| 1 | 50 | |||
| 2 | 25 | |||
| 3 | 12.5 | |||
| 4 | 6.25 | |||
| 5 | 3.1 | |||
| 6 | 1.6 |
Illustrative decay data (Option B). One run of 100 coins and a pooled total from eight runs (800 coins). Constructed teaching numbers, not real measurements.
| Shake | Parents remaining, run 1 (of 100) | Pooled total parents (of 800) | Pooled total scaled to 100 | Ideal 100 × (1/2)n |
|---|---|---|---|---|
| 0 | 100 | 800 | 100.0 | 100 |
| 1 | 47 | 403 | 50.4 | 50 |
| 2 | 21 | 198 | 24.8 | 25 |
| 3 | 12 | 102 | 12.8 | 12.5 |
| 4 | 7 | 49 | 6.1 | 6.25 |
| 5 | 3 | 23 | 2.9 | 3.1 |
| 6 | 1 | 12 | 1.5 | 1.6 |
| 7 | 0 | 5 | 0.6 | 0.8 |
Analysis
- Claim: does your decay follow a half-life pattern? Use your data and the ideal curve as evidence, and explain why each shake removes about half.
- Compare your single-run curve with the pooled curve. Which is closer to the ideal, and why does a larger sample reduce scatter?
- Suppose a run ended with 13 parents and 87 daughters but the shake count was lost. Use the ratio D/P to estimate how many shakes (half-lives) had passed, and check it against the table.
- Sources of error: name two ways the model differs from real atoms (for example, the probability of decay per shake, or pieces landing on edge) and how each affects the result.
If you cannot do this lab in person: Choose one. All three earn full credit and end in the same graphs against the ideal curve and the same four analysis answers. Option A — shake real objects. The procedure above. Pennies, any coins, printed candies, or 100 squares of paper with a mark on one side all work, and a cereal bowl with a plate on top is a good shaker. Option B — the illustrative decay data in the data section above. No materials and no internet. Graph the single-run curve, the scaled pooled curve and the ideal curve on one set of axes, then answer every analysis question, including which curve is closer to the ideal and why. Option C — generate your own randomness. Fill 100 cells of a free spreadsheet with =RAND() and count how many are below 0.5: those atoms decayed. Keep only the survivors and repeat for eight shakes, then repeat the whole thing eight times and pool the runs to build your own pooled total. A free online coin-flip tool set to 100 flips does the same job. Because you generate the numbers yourself, add one extra sentence to your answer about scatter: say whether your pooled 800 came closer to the ideal than your single 100, and by how much.
Formative check
Work through these before moving on. They are not graded — they tell you whether the standard below has landed yet.
Simulation & tools
Stable and unstable isotopes of carbon
On the Isotopes screen, build carbon: 6 protons. Add neutrons one at a time from 6 to 8. For each, record the mass number, the symbol shown (12C, 13C, 14C) and whether the sim marks it stable or unstable. Then build nitrogen-14 (7 protons, 7 neutrons). Explain in one sentence how carbon-14 can become nitrogen-14 when one of its neutrons changes into a proton.
No sim? Fill in the same three rows from here instead. Carbon always has 6 protons. With 6 neutrons it is carbon-12 (mass number 12, stable, and about 99% of all carbon); with 7 it is carbon-13 (stable); with 8 it is carbon-14 (unstable, half-life 5,730 years). Nitrogen-14 has 7 protons and 7 neutrons and is stable. Write the same sentence explaining how carbon-14 becomes nitrogen-14 when one neutron turns into a proton, and say what the mass number does when that happens.
Open in a new tab ↗ · PhET Interactive Simulations, University of Colorado Boulder · CC BY 4.0
A zircon contains 7 lead-206 atoms for every 1 uranium-238 atom (assume no lead at the start). How many half-lives have passed?
Carbon-14 dating can find the age of a 66-million-year-old dinosaur bone.
After 3 half-lives, % of the parent isotope remains; after 5 half-lives, about %.
How confident are you that you can explain half-life and calculate a radiometric age from a parent-daughter ratio?
Practice
Work these on paper or in your notebook, then open Check your answer. Aim for all of Fluency and Application; try at least one Challenge.
Printable version: this unit’s practice workbook (PDF)
Fluency
Build speed and accuracy with the core skill.
- An isotope has a half-life of 4 days. How long until only 1/64 of the original sample remains?
Check your answer
Answer: 24.0 days (6 half-lives)1/64 = (1/2)6. - A sample starts with 640 g of an isotope whose half-life is 8 years. How much remains after 32.0 years?
Check your answer
Answer: 40.00 g32.0 years is 4 half-lives; halve 4 times. - A sample starts with 640 g of an isotope whose half-life is 30 years. How much remains after 150.0 years?
Check your answer
Answer: 20.00 g150.0 years is 5 half-lives; halve 5 times. - A sample starts with 80 g of an isotope whose half-life is 30 years. How much remains after 120.0 years?
Check your answer
Answer: 5.00 g120.0 years is 4 half-lives; halve 4 times.
Application
Use the skill in context. Show your reasoning.
- A volcanic ash layer has 7 atoms of argon-40 (from decay) for every atom of potassium-40 left (assume the branching correction is done). How old is it? Half-life 1.25 billion years.
Check your answer
Answer: 3.75 billion years1 + 7 = 8 = 23: three half-lives. - Charcoal from an ancient fire pit has 12.5% of the carbon-14 of living wood. How old is it?
Check your answer
Answer: about 17,190 years12.5% = (1/2)3: three half-lives of 5,730 years. - A zircon has 0.50 lead-206 atoms per uranium-238 atom. Find its age (half-life 4.468 billion years).
Check your answer
Answer: about 2.61 billion yearsn = log2(1.5) = 0.585; t = n × 4.468. - A zircon from a Missouri granite in the St. Francois Mountains is 1.47 billion years old. What lead-206 to uranium-238 ratio should it hold (half-life 4.468 billion years, no starting lead)?
Check your answer
Answer: about 0.256D/P = 21.47/4.468 − 1 = 20.329 − 1.
Challenge
Stretch problems. Expect to think before you write.
- Only one-third of the original potassium-40 remains in a rock. How old is it (half-life 1.25 billion years)? Show why the answer is between one and two half-lives.
Check your answer
Answer: about 1.98 billion yearsRemaining 1/3 means D/P = 2; n = log23 = 1.585; 1/3 is between 1/2 and 1/4.
Review
Keep earlier skills sharp.
- Review (Unit 1): a galaxy recedes at 7,000 km/s. With H0 = 70 km/s/Mpc, how far away is it?
Check your answer
Answer: 100 Mpcd = v / H0.
Recommended Half-Lives, an Eight-Run Coin Record, and Every Age in the Lesson Recomputed
DDEP/LNHB Table de Radionucleides for U-238, U-235, K-40 and C-14 (lnhb.fr), retrieved November 6, 2026; du Bray et al., USGS Professional Paper 1866 (2021), retrieved November 6, 2026; SLS original illustrative coin-shake record of eight runs, built to the pooled totals printed in the lesson · SLS original (illustrative lab record and worked computations), citing DDEP half-lives and USGS ages as published data
Why this reading. The lab prints the counts from one run of 100 coins and a pooled total for eight runs, but not the eight runs themselves, so the question “which curve is closer to the ideal, and why” has only two curves to compare. Here are all eight runs, built so that they add up exactly to the printed totals. The half-lives the lesson uses are checked against the evaluated values that laboratories actually use, and one of them does not match. Every age in the lesson is recomputed below so you can see where each digit comes from.
Table 1. Half-lives: the lesson against the evaluated values
The Decay Data Evaluation Project (DDEP), published by the French national laboratory LNHB, is the set of recommended values that metrology institutes, including NIST, use. Values as printed in the DDEP tables, retrieved November 6, 2026.
| Parent | Lesson’s half-life | DDEP recommended value | Agreement |
|---|---|---|---|
| Uranium-238 | 4.47 billion years (4,468 million in the worked examples) | 4.468 ± 0.005 × 109 years | Yes |
| Uranium-235 | not used in this lesson; see Lesson 2.5 | 704 ± 1 × 106 years | — |
| Potassium-40 | 1.25 billion years | 1.2522 ± 0.0027 × 109 years; 10.34% of decays go to argon-40, 89.56% to calcium-40 | Yes (the lesson’s “about 11% to argon” is this 10.34%) |
| Carbon-14 | 5,730 years | 5,700 ± 30 years | No. 5,730 is the older “Cambridge” value still used by convention in radiocarbon laboratories to report raw ages. The evaluated physical half-life is 5,700 ± 30 years. |
What the carbon-14 disagreement does to the lesson’s numbers: the charcoal at 25% of living carbon-14 has passed two half-lives, which is 11,460 years at 5,730 or 11,400 years at 5,700; the amount left after 50,000 years is 0.24% at 5,730 or 0.23% at 5,700. The conclusions do not change. When you report a radiocarbon age, say which half-life you used.
Table 2. Eight runs, 100 coins each: parents remaining after each shake
Illustrative constructed data. Run 1 is the run printed in the lesson; runs 2 to 8 were constructed so that each row sums to the lesson’s pooled total. Procedure as in the lab: 100 pennies heads up, three-second shake, tails removed, the same 100 coins reset for every run. n = 8 runs × 100 coins = 800 “atoms”.
| Shake | R1 | R2 | R3 | R4 | R5 | R6 | R7 | R8 | Pooled total | Total scaled to 100 | Ideal 100 × (1/2)n |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 800 | 100.0 | 100 |
| 1 | 47 | 53 | 49 | 48 | 45 | 49 | 59 | 53 | 403 | 50.4 | 50 |
| 2 | 21 | 26 | 24 | 23 | 17 | 26 | 34 | 27 | 198 | 24.8 | 25 |
| 3 | 12 | 14 | 11 | 13 | 8 | 13 | 17 | 14 | 102 | 12.8 | 12.5 |
| 4 | 7 | 6 | 4 | 7 | 3 | 7 | 7 | 8 | 49 | 6.1 | 6.25 |
| 5 | 3 | 2 | 1 | 5 | 1 | 4 | 4 | 3 | 23 | 2.9 | 3.1 |
| 6 | 1 | 1 | 0 | 3 | 1 | 2 | 3 | 1 | 12 | 1.5 | 1.6 |
| 7 | 0 | 1 | 0 | 1 | 0 | 0 | 2 | 1 | 5 | 0.6 | 0.8 |
Read across shake 1: the eight runs range from 45 to 59 survivors, a spread of 14 around an ideal of 50. Their total, 403, is off the ideal 400 by less than 1%. Read down run 7: it ran high on the early shakes, 59 then 34 then 17, and still had only 2 parents left by shake 7. Read down run 5: it ran low, 45 then 17 then 8. Neither run is wrong; a coin does not know what the other coins did. The pooled curve is smoother because the high and low runs cancel, which is the whole answer to analysis question 2, and the reason real laboratories count many atoms.
The lab’s lost-count problem
A run ends with 13 parents and 87 daughters, but the shake count was lost. Original parents = 13 + 87 = 100. Number of half-lives n = log2(100/13) = log2(7.69) = 2.94. So about three shakes. Check against Table 2: after three shakes the runs hold 8 to 17 parents, and 13 is squarely inside that range; after two shakes they hold 17 to 34, and after four, 3 to 8. Three is the only shake count consistent with the record.
Every age in the lesson, recomputed
| Problem | D/P | n = log2(1 + D/P) | Half-life used | Age | Lesson says |
|---|---|---|---|---|---|
| Ash layer, argon-40 to potassium-40 | 3 | log2 4 = 2 | 1.25 billion years | 2.50 billion years | 2.5 billion |
| Zircon, lead-206 to uranium-238 | 0.25 | log2 1.25 = 0.3219 | 4,468 million years | 1,438 million years | about 1,440 million |
| Charcoal, 25% of carbon-14 left | 3 (75 daughters per 25 parents) | 2 | 5,730 years (or 5,700) | 11,460 years (11,400) | 11,460 |
| Quiz zircon, 7 lead per uranium | 7 | log2 8 = 3 | 4.47 billion years | 13.4 billion years: impossible, older than the solar system, so lead was added | 3 half-lives, reject the grain |
| Missouri check: a zircon from the 1.47-billion-year Granite of Castor River Shut-Ins | 21.47/4.468 − 1 = 0.256 | 1.47 / 4.468 = 0.3290 | 4,468 million years | 1,470 million years (working forward from the published age) | not in the lesson; compare the worked zircon |
| Fill-in: 3 half-lives | — | 3 | — | (1/2)3 = 12.5% remains | 12.5 |
| Fill-in: 5 half-lives | — | 5 | — | (1/2)5 = 3.125% | 3.1 |
To compute log2 of a number on any calculator, take its ordinary logarithm and divide by log 2 = 0.30103: log2(1.25) = 0.09691 / 0.30103 = 0.3219. The potassium-argon example assumes the 10.34% branch to argon has already been corrected for, as the lesson says; without the correction, 3 argon atoms per remaining potassium would imply far more decays than actually occurred.
The worked zircon and Missouri’s oldest rocks
The lesson’s worked zircon, 0.25 lead-206 atoms per uranium-238 atom, gives 1,438 million years. U.S. Geological Survey geologists dated zircon from the St. Francois Mountains of southeast Missouri at 1.48 to 1.45 billion years, and the Granite of Castor River Shut-Ins at 1.47 billion years (USGS Professional Paper 1866, 2021). Working forward from 1.47 billion years, a closed zircon from that granite should hold 20.3290 − 1 = 0.256 lead-206 atoms per uranium-238 atom, only slightly more than the worked example’s 0.25. Small differences in the ratio make differences of tens of millions of years in the age, which is why the laboratories measure ratios to several decimal places.
Two things the model gets wrong, with numbers
First, a coin decays with probability exactly 1/2 per shake; a real nucleus has a fixed probability per second, and the half-life is whatever time makes that probability add up to 1/2. The curve shape is the same. Second, the model cannot show that the daughter can escape. Argon is a gas; if the volcanic rock is reheated, argon leaks out, the ratio D/P drops, and the computed age is too young. Zircon holds its lead, which is why the lesson calls it the ideal clock. When you name sources of error, name one from the coins (pieces landing on edge, miscounts, a shake too gentle to randomize) and one from the atoms (an open system), and say which way each pushes the age.
Answer key for the Lesson 2.1 practice section: P, Q, R, S deposited (superposition); intrusion I cuts them (cross-cutting); the fault offsets P to S and, since I is not offset on the drawing where it meets the fault, the order of fault and intrusion cannot be settled from the sketch alone; the lava flow L is unbroken across both, so it is youngest.
Sources: DDEP/LNHB Table de Radionucleides, U-238 (lnhb.fr); DDEP/LNHB Table de Radionucleides, U-235 (lnhb.fr); DDEP/LNHB Table de Radionucleides, K-40 (lnhb.fr); DDEP/LNHB Table de Radionucleides, C-14 (lnhb.fr); LNHB nuclear data table (index) (lnhb.fr); USGS Professional Paper 1866, Petrology and geochronology of 1.48 to 1.45 Ga igneous rocks in the St. Francois Mountains terrane, southeast Missouri (2021) (pubs.usgs.gov)