Measuring the radon concentration always means counting the decays that take place in a measuring chamber within a period of time. A measuring device can only register a part of these decays. The parameter sensitivity (counting efficiency) indicates how many decays are actually detected at a given radon concentration. The sensitivity depends on the measuring principle used and the size of the measuring chamber. A high level of sensitivity is good, but it is expensive due to the technical outlay or is associated with restrictions in the area of application. Therefore, look at all parameters of the measuring device when you are looking for the optimal device for your application!

Radioactivity is a statistical process, ie even with a constant radon concentration, a different number of radon atoms decay in every time interval. This number fluctuates around a mean (better expected value). The larger the expected value itself, the smaller the range of fluctuation around this expected value. A measuring device with high sensitivity registers a larger number of decays within a time interval, which results in a smaller deviation of an individual measurement from the expected value. This deviation is also known as statistical uncertainty or statistical errors. For devices with low sensitivity, the time interval must therefore be selected to be very long in order to register a sufficiently large number of decays for an acceptable statistical error.

The following applies to devices with higher sensitivity:

• With identical (e.g. maximum permissible) statistical uncertainty, lower radon concentrations or smaller changes in concentration can be measured.
• Identical radon concentrations can be measured with a lower statistical error for the same measurement duration.
• Identical radon concentrations can be measured in a shorter time with the same fluctuation range of the measured value.

However, the statistical error does not increase linearly with the number of detected pulses, but with their square root. If you compare the statistical errors of a device with a high sensitivity with those of a device with a lower sensitivity, you can see that their fluctuation ranges differ greatly at low concentrations and / or short measuring times. With increasing concentrations, however, they are getting closer and closer.

Example: A less sensitive device 1 detects 10 decays at a radon concentration of 100 Bq/m³, device 2 with ten times sensitivity corresponds to 100 decays.

However, this simple estimate only applies if all detected decays can be assigned to the current radon concentration. In field use, however, nuclides also decay in the measuring chamber, to which this does not apply and which a device cannot distinguish from one another without spectroscopy. These decays are called the background signal, which must be estimated and subtracted from the total number of decays detected. Otherwise the measured value would be too high.

An underground signal is generated by:

• Residual activity from previous measurements in quick successive measurements
• inseparable Thoron and its relatively long-life progeny
• Long-term Polonium-210 contamination caused by the radon measurement itself
• Ambient radiation and spontaneous ionization in ionization chambers

Since both the number of impulses detected as a whole and the estimated background signal are associated with statistical uncertainty, both add up during subtraction in accordance with Gaussian error propagation. The uncertainty of the corrected measured value increases more or less depending on the size of the background.

Devices with real α-spectroscopy differentiate the different nuclides energetically so that they can be counted separately. A background subtraction is therefore not necessary.

The sensitivity required depends on the application and the measuring principle used. The following table contains reference values for some example scenarios.

cpm = counts per minute (number of decays detected within one minute)