Removal of doubt: between certainty and uncertainty
by Col. Denis Giordan, Deputy Director of Savoie Fire and Rescue Services, France
The presence of radioelements emitting radiation imperceptible to the human senses makes necessary to use specially adapted detectors. These detectors, generally used for radiation protection, are called survey meters. They make it possible either to detect the presence of radioactive radiation or to assess the quantity of energy transferred to matter - one of the factors directly leading to biological effects. The count rate meter is used to remove doubt in the presence of radioactive materials, contamination of people or objects and to ensure the location of these (sources sealed or not). This device translates the number of radioactive radiations perceived – a direct function of the activity – into shocks or pulses per second. The minimum detectable radioactivity activity depends mainly on three parameters: - Detector sensitivity - Background noise - Duration of the measurement cycle It is imperative for all operators to be familiar with their equipment, the desired radiations, and the measurement conditions (distance to the source, reproducibility of the measurement, etc.). Two parameters are important to define precisely if there is added radioactivity to natural radioactivity:
Search of radiological smoke detectors after a fire
I/ DECISION THRESHOLD AND DETECTION LIMITS Regularly, the emergency specialized teams are called upon for intervention in the field of radiological risk. The first step is to validate or deny the presence of abnormal radioactivity; this is the "removal of doubt": Is there or not radioactivity added to natural radioactivity? From what value read on a detector can the Incident commander or Command and control center be informed of the absence of a proven radioactive risk? It is understood that this is a delicate operation, the result of which, if misinterpreted, could have serious consequences. This is the reason why well-trained personnel must carry out the removal of doubt under the best possible measurement conditions. In the field of ionizing radiation, detectors always measure "something". This quantity has three simultaneous origins and of unequal importance: 1 - The own movement generated by the detector itself, particularly its electronics 2 - The background noise induced by the natural ambient radioactivity of the place 3 - The count rate induced by the added radioactivity (the one we are trying to find out) We must therefore be able to quantify this "unnatural something" and compare it to natural values. A measurement of added radioactivity can sometimes be difficult to identify amid an ambient "background noise" which varies from moment to moment. The measurement of radioactive phenomena, made up of numerous and random events, is a privileged area of the application of statistical laws. Thus, by unanimously accepted simplification, there is a notion of distribution of events according to the so-called “normal” statistical law. It is from the study of this law that we can define a decision threshold DT and a detection limit DL. While the ISO 11929-7 standard defines the rules to be applied for integrated counting over time, this standard is not intended for the types of measurements carried out by specialized responders during rescue operations. Indeed, the controls carried out in the field do not pretend to achieve measurement precision that can be obtained in the laboratory, there are factors liable to induce evaluation errors.
These main factors are: 1. Self-absorption of radiation by the source matrix. 2. The presence of screens between the active matrix and the detector. 3. The variations in background noise at the control site.
1) Definitions "Relative certainty" of the absence of added radioactivity (at XX%) The decision threshold (DT) is a “border” making it possible to define whether a measurement indicates added radioactivity. This statement is associated with a certain probability of error. The decision threshold is expressed in c/s. Another way to put it is that - This threshold makes it possible to take a decision on the presence of radioactivity added to the background noise, with a probability of error of XX%). - Above the DT, a measuring device gives a measurement result which, statistically, is not natural background noise. If the measured value is less than the Decision Threshold (DT), it can be said, with plausible and conservative assumptions, that added radioactivity is not detected if the value of the latter is lower than that of the decision threshold of the device used. The decision threshold is set, taking an “alpha” risk of deciding that there is a radioactive sample present when it is a background count.
The factor k(1-a) will be used as calculated in the table at the left.
"Relative certainty" of the presence of added radioactivity (at XX%) The detection limit (DL) is the minimum value for which we are XX% sure that the measurement represents added radioactivity. In other words, the detection limit is the smallest value of the true signal having a probability of at least equal to XX% of being detected: when I exceed the detection limit, the value corresponds to the signal (measurement + Bkgd), for example, the presence of surface contamination or the presence of a source nearby, and not Bkgd. The detection limit is equal to twice the decision threshold: DL = 2 * DT. Another way to put it is that the detection limit is an almost certain limit to measure added radioactivity. From this low limit, the probability of detection is very high and, conversely, the probability of non-detection is very low. This concept is particularly used when the measured values of the background noise and the counting rate are close. The limit of detection is set, taking a “beta” risk of deciding that it is background noise when there is a radioactive sample present.
The factor k(1-b) will be used as calculated in the table at the left.
Between "relative certainties" of the absence and presence of added radioactivity It is considered that in this area, the probability is higher to have added radioactivity than to have only background noise. Another way of saying it is that this zone, between threshold and limit, shows added radioactivity without the certainty being absolutely controlled.
2) Synthesis diagram A diagram illustrates this definition and answers the question "is the sample I am measuring radioactive?"
II/ IMPLEMENTATION 1) Basic formulas This formula makes it possible to take into account a different counting time between Bkgd and measurements. The general formula for calculating the net decision threshold is:
With: k (associated with a level of certainty) given in the table above Bkgd: background noise given in c/s tbcgd: background noise measurement time, given in seconds tm: measurement time of the sample or the surface, given second As a reminder, the gross decision threshold will be:
The general formula for calculating the net detection limit is:
With: k1-a et k1-b (associated with a level of certainty) given in the table above Bkgd: background noise given in c/s tbcgd: background noise measurement time, given in seconds tm: measurement time of the sample or the surface, given second As a reminder, the gross detection limit will be:
In the particular case where the background noise counting and measurement times are identical and equal to 1 second The formulas used to calculate decision threshold and decision limit are mathematically:
2) Examples The first examples are related to instantaneous measurements Detection limit and "two to three times the background noise": 95%, time = 1 s A measurement by means of an X probe on a count rate meter gives a count rate of 8 c/s. This is the expression of background noise. This is the minimum value of the count rate indicated by the probe under the current measurement conditions. We will retain the certainty of 95 %.
Then we can say that: - from 16 c/s, it is likely that the count is more related to the added radioactivity than to the background noise - from 24 c/s, it is likely that the count is more related to the added radioactivity than to the background noise. Hence, the famous rule of "2 to 3 times the background noise". If the count rate remains below 16 c/s, the information to be communicated to the Authority: «The count is not significant of added radioactivity with 95% certainty». If the count rate is greater than 24 c/s, the information to be communicated to the Authority will be: "We are in the presence of a significant count of added radioactivity with 95% certainty" Decision threshold and the end of "two to three times the background noise": 99.7%, time = 1 s A measurement by means of an X probe on a count rate meter gives a count rate of 8 c/s. This is the expression of background noise. This is the minimum value of the count rate indicated by the probe under the current measurement conditions. We will retain the certainty of 97,7 %.
We can say that: - from 20 c/s, it is likely that the count is more related to the added radioactivity than to the background noise - from 32 c/s, there is a 99.7% chance that what I am measuring is added radioactivity. The famous rule of "2 to 3 times the background noise" is no longer suitable. If the count rate is smaller than 20 c/s, the information to be communicated to the Authority will be: "The count is not significant of added radioactivity with 97,7 % certainty". If the count rate is greater than 32 c/s, the information to be communicated to the Authority will be: "We are in the presence of a significant count of added radioactivity with 99,7 % certainty" Rising the level of certainty induced, statistically, the increase of counts per second (you have less doubt thanks to expansion of values).
The following examples use the integration mode: Numerous detectors may count during a define time. It is often named "integration mode" or "sum mode” or “scaler”. Thus, it is possible to increase the acquisition time of the background noise or of the “sample/source/contaminated surface” count-rates.
MIP 10D - Sum mode
Radiagem - Sum (preset) mode “keystrokes: Log+Light”
Increase of the measurement time of Bkgd to 5 s.: 95% A measurement using a “soft beta” probe (GM pancake) gives a count rate of 1 c/s, the background measurement time is 5s. This is the expression of background noise. This is the minimum value of the count rate indicated by the probe under the current measurement conditions.
Increasing the background time allows for decreasing detection threshold and detection limit, inducing a more precise removal of doubt. The same results are mathematically obtained if we increase the acquisition time on the “sample/source/contaminated surface”.
Increase of the background noise acquisition and sample measurement times to 30 s.: 99.7% A measurement by means of an X probe on a count rate meter gives a count rate of 8 c/s on average over 30 s. acquisition of the background noise, the measurement time of the sample or the surface is also 30 s. This is the minimum value of the count rate indicated by the probe under the current measurement conditions. We will retain the certainty of 99,7%.
The decision threshold and detection limit are particularly low, even with a very certainty of 99.7%. However, to put these measurement conditions in place, time is needed. This will remove any doubt about the presence of radioactive material on a surface or in an object.
III/ OPERATIONAL IMPLEMENTATION From these examples, it may be possible to extrapolate examples of control procedures and the formula to use
Through the use of these mathematical formulas, specialized responders may be easily in capability to remove the doubt on the presence of added radioactivity. Of course, it is necessary to define a level of certainty. Moreover, it is important to choose the most favourable operational measurement conditions, depending on the situation: - measurement time - level of natural background - distance of detection
The removal of doubt on a supposed contaminated glove of a specialized responder
The removal of doubt on a piece supposed to be contaminated
A U T H O R
Denis Giordan is a Colonel and Deputy Director of Savoie Fire and Rescue Services in France. He is also a member of the technological risks and CBRN Commission of French Firefighter’s National Federation, and of the HazMat commission of the International Association of Fire and Rescue Association (CTIF).
He performs Nuclear and Radiological technical advisor tasks as head of the "Radioprotection and radio radionuclear risks and threats” section. He is a former zonal R.N. technical advisor from the Paris civil defence zone and from East civil defence zone. Col. Giordan wrote numerous papers on technological risks and CBRNe threats, being co-writer of the French reference book “CBRNe terrorism”. Additionally, he is a teacher in the Upper Alsace University, participating in the “CBRN risks and threats” Master's Degree and university diploma.