Skip to content
Radiation Lab
Radioactive decay examples: how much is left

How much is left? Radioactive decay worked examples

By Radiation Lab · Published · Reviewed

Most decay questions are one question: you have a known amount of one nuclide, some time passes, and you want to know how much is left. This note works it four ways. First a textbook problem with the half-life printed in it, then the same problem with an evaluated half-life, then a check source brought from its certificate date to today, and last the reverse step from a present reading back to a certificate. Each answer can be reproduced in the radioactive decay calculator; where the calculator cannot do a step, the note says so.

The rule: count half-lives

Divide the elapsed time by the half-life to get n, the number of half-lives that have gone by. The fraction left is 2−n, or written out, fraction remaining = 2−t/T½.

Time and half-life must be in the same unit, and the result is a pure ratio. It applies equally to the activity, the number of atoms and the mass of that nuclide, since all three are proportional to one another. Nothing requires n to be a whole number.

Half-lives elapsed, nFractionPercent left
0.5—70.7
11/250.0
1.5—35.4
21/425.0
31/812.5
41/166.25
51/323.13
71/1280.781
101/10240.0977

To ask how long instead of how much, turn the rule around: t = T½·log₂(A₀/A). Falling to 1% takes log₂(100), about 6.64 half-lives whatever the nuclide, which for Ra-228 is 38.2 years.

Ra-228 after 17.4 years: 1/8 or 12.3%?

A typical homework version reads: “The radioisotope Ra-228 has a half-life of 5.8 years. What fraction remains after 17.4 years?” Here n = 17.4 / 5.8 = 3, so the fraction is 2−3 = 1/8, or 12.5%. The numbers were chosen to come out clean.

Now run the same 17.4 years through the calculator with Ra-228 selected. It answers 12.3%. Nothing has gone wrong. The calculator takes its half-life from evaluated nuclear data, the IAEA Livechart value of 5.75 years, so n = 17.4 / 5.75 = 3.026 and 2−3.026 = 0.1228. The two answers differ by 0.22 of a percentage point, which comes from the rounded half-life in the question. Keep that in proportion: the evaluated value is itself 5.75 ± 0.03 years, enough to spread the answer from 12.1% to 12.4%.

Which one to give depends on who is asking. On a problem set, use the half-life the problem gives you: the expected answer is 1/8. For a real sample, use the evaluated value and say which one you used. The calculator has no field for typing your own half-life, so a textbook figure such as 5.8 years has to go through the rule above by hand.

The fraction also describes Ra-228 alone. Its daughter Ac-228 and then Th-228, half-life 1.91 years, grow in behind it, so the total activity of an old radium sample does not follow this curve. The calculator follows one nuclide and does not build decay chains; the note on natural decay series shows the ingrowth.

A Cs-137 check source, from certificate to today

A Cs-137 check source has a certificate reading 370 kBq on 1 June 2016. What does it hold on 1 November 2026?

  1. Elapsed time: 3,805 days, which is 10.418 years.
  2. Half-lives: n = 10.418 / 30.08 = 0.3463.
  3. Fraction: 2−0.3463 = 0.7866.
  4. Activity: 370 × 0.7866 = 291 kBq.

The calculator with these inputs gives the same figure. The reason to do this is what you then compare it with. A meter with a net reading of 12,000 cpm on this source on the certificate date should give about 9440 cpm net now, in the same geometry and at a rate where dead time does not matter. That 21.3% fall is decay and says nothing about the detector. Only a difference from the corrected figure, beyond your tolerance, raises a question about the instrument or the source. A gamma detector is counting the 662 keV photon from Ba-137m, half-life 2.55 minutes, which stays in equilibrium inside a sealed source, so its rate follows the Cs-137 and the same factor applies.

How exact does this need to be? Cs-137 loses 2.28% a year, so a date wrong by a month moves the answer by 0.19%. Replacing the 30.08-year half-life used here with the 30.018-year value from the Decay Data Evaluation Project moves it by 0.05%. Set both against the uncertainty printed on your certificate before spending time on them.

Going backwards: from a reading to the certificate date

Suppose a Co-60 source is assayed at 52.4 kBq on 1 November 2026, and its certificate, dated 15 March 2019, says 140 kBq. Do the two agree? The past activity was larger, so here you divide: A₀ = A / 2−t/T½.

  1. Elapsed time: 2,788 days, which is 7.633 years against a half-life of 5.271 years.
  2. Fraction: 2−1.4481 = 0.3665.
  3. Activity on the certificate date: 52.4 / 0.3665 = 143.0 kBq.

That is 2.1% above the certificate. Whether it counts as agreement depends on the two uncertainties combined; against a certificate stated at ±5% and a comparable assay, it does.

The elapsed-time field in the calculator does not accept negative values, so it will not run decay backwards directly. There are two ways round that. Run the certificate forward from 140 kBq for 2,788 days and compare its 51.3 kBq with your reading, or enter a starting activity of 1, read the fraction, and divide your reading by it.

The usual slip is to multiply: 52.4 × 0.3665 gives 19.2 kBq, which is less than the source holds today. A past activity below the present one cannot come from decay, so that check catches the error at once. For planning ahead, count from today's value: the “When does it reach” mode says this source falls from 52.4 to 10 kBq in 12.6 years.

What these answers leave out

Check it yourself

Sources