What it shows:

Gamma ray spectroscopy is used to detect the minute amount of radioactive potassium-40 present in the human body. Using a NaI(Tl) scintillation detector in conjunction with a multichannel pulse-height analyzer (PHA), 1.46 MeV gammas originating from the human body are detected. The source of these gammas is K-40 which has a half-life of 1.26 billion years, and is the main source of radioactivity inside the body. The second most active radionuclide in the body, carbon-14 (5,730 yr half-life), can not be detected with this apparatus because it is a beta emitter.

How it works:

"There are 1.2 radioactive atoms of ⁴⁰K for every 10,000 nonradioactive atoms of potassium. There is of the order of 140 g of potassium in an adult who weighs 70 kg, and 0.0169 g consists of the ⁴⁰K isotope. This amount of ⁴⁰K disintegrates at the rate of 266,000 atoms per minute. Of every 100 disintegrations, 89 result in the release of beta particles with maximum energy of 1.33 MeV, and 11 result in gamma photons with an energy of 1.46 MeV. All of the beta particles and about 50 percent of the gamma rays are absorbed in the body, giving annual doses of 16 mrad from the beta particles and 2 mrad from the gamma rays." ¹ So about 14,600/min of the 1.46 MeV gammas exit the whole body, in all directions. We detect about 34 counts/min. ² Because of the very low count rate, one must make a considerable effort to shield the detector from other natural background radiations (see Going Further at the end of this discussion) as well as count for a sufficiently long time to produce a signal which is significantly above the statistical background so as not to be considered imaginary.

The scintillation detector is shielded by a large 1200 lb. iron cylinder. ³ This cylinder is not only the bulk of the experiment, but also crucial and rather unique: it is actually a slice out of a WWII Navy ship's gun. Since the iron it is made of is quite old, its thick (13 cm) walls provide an excellent radiation shield (low in natural radiation) necessary for low-level counting.

Shown below are two images from the Canberra multichannel pulse height analyzer (PHA). The first is a Cs-137 calibration spectrum. The second is a K-40 spectrum from a person sitting in front of the detector for 10 minutes — the cursor shows 17 counts in the peak.

Setting it up:

Any PHA and regulated high voltage power supply (HVPS) will do the job. Presently we're using a Canberra (Series 20) Multichannel Analyzer which conveniently has a built-in HVPS as well as a video output so that the screen information can be viewed on a large monitor or video projector by the audience. The following settings will give you a quick start in set-up and can be fine tuned:

HVPS: +900 volt bias on photomultiplier tube
Amplifier input: positive
Amplifier gain: 4
ADC gain: 1024
LLD: 2%
Memory size: eighth (512 channels)

An energy calibration with a known radioactive source should be performed once the electronics has warmed up (and won't drift much). Our detector resolution is about 80 KeV FWHM. A Co-60 source is convenient because its two peaks at 1.173 and 1.332 MeV are close to the K-40 line (1.458 MeV). A Cs-137 (0.662 MeV) or Na-22 (0.511 MeV) source, or any combination of sources will also do. Once the energy calibration has been performed, the cursor location will read directly in MeV so that any peak or part of the spectrum can be identified on an energy scale. At this point make sure that the K-40 peak will fall within the energy scale of the chosen memory size (typically = 512 channels); if it doesn't, fine tune the amplifier gain so that the 1.46 MeV line will be on scale. Of course you will have to recalibrate the energy scale.

Have a student sit in front of the iron shield. Since the count rate is low, spectrum accumulation time is long (≈17 min.), so plan to continue lecturing while data is being accumulated. The PHA cursor is used to identify the K-40 peak. 1000 seconds of counting will give a peak about 31 counts high and about 560 total events (integrated counts in peak).

Comments:

An alternative (or additional) experiment is to place a shaker of No Salt® in front of the detector. Being a salt alternative, its main ingredient is potassium chloride and a large (311 g) shaker will produce a convincing K-40 peak within a few minutes. Obviously a jar of potassium (borrowed from your friendly chemistry demonstrator) will yield a considerably higher count rate, but that's getting away from the spirit of the demonstration which is to impress the audience that practically everything is radioactive to some extent, even your own body.

Going Further:

Closely related to this demonstration experiment is the gamma ray spectrum of the "room background" originating from mundane sources such as the cement floor, brick walls, etc., as well as events and interactions happening in the detector itself. This kind of natural background spectrum can also be measured (without the shield!) and shown in the period of a lecture. Most (if not all) of the peaks can be associated with the decay of potassium-40, or one of the elements of the uranium series, or one of the elements of the thorium series. Other peaks present are a result of pair production within the detector by gamma rays having energies in excess of 1.022 MeV and the subsequent escape of one or both of the 511 KeV gammas associated with the annihilation of the positron member of the pair.

A very readable and complete book is by G.F. Knoll, Radiation Detection and Measurement, 2nd ed. (John Wiley & Sons, N.Y., 1989). Other recommended references are Chart of the Nuclides, 13th ed (General Electric Nuclear Energy Operations, San Jose CA, 1984) and an older U.S. Atomic Energy Commission publication by E. Mateosian and M. McKeown, Table of Gamma-Rays Emitted by Radioactive Nuclei, (Brookhaven National Lab, Upton N.Y., 1960) available from the Office of Technical Services, Department of Commerce, Washington 25, D.C.

1 Jacob Shapiro, Radiation Protection, 2nd. ed., (Harvard University Press, Cambridge MA, 1981)
2 Since our detector sees somewhere between 10 and 20 percent of the body being monitored, and subtends approximately 0.2 steradian from it, one would expect something like 35 gammas/min. at the detector and thus a count rate between 15 and 25 counts/min., assuming a detector efficiency of a little more than 50%. This is, of course, a very approximate number and assumes that the potassium is distributed uniformly throughout the body.
3 The cylinder was donated to the Science Center by Jacob Shapiro from Environmental Health & Safety, Harvard University. It measures 44 cm (17") in diameter and is 52 cm long with an 18 cm (7") dia. bore. Those not having access to such a piece of iron will need to secure many lead bricks, low in activity.