Index > 3 Characterisation of radioactively contaminated sites >

3.3.7 Criteria for the selection of direct measurement, scanning and sample collection methods

Contents
3.3.7.1 Introduction
3.3.7.2 Detection sensitivity
3.3.7.3 Direct measurement sensitivity
3.3.7.4 Scanning sensitivity
3.3.7.5 Scanning sensitivity for beta and gamma emitting nuclides
3.3.7.5.1 Signal detection theory
3.3.7.5.2 The two stages of scanning
3.3.7.5.3 Determination of the minimum detectable count rate (MDCR) and use of surveyor efficiency
3.3.7.5.5 Determination of scan MDCs for building/structures
3.3.7.6 Determination of scan MDCs for land areas and structure surfaces
3.3.7.6.1 Determination of scan MDCs for land areas
3.3.7.6.2 Determination of scan MDCs for building/structure surfaces
3.3.7.7 Determining a scan MDC for alpha emitters
3.3.7.8 Sensitivity of mobile systems with integrated positioning systems

3.3.7.1 Introduction

The presence of radioactive and hazardous chemical wastes (mixed wastes) at a site can influence the survey design [EA P2-178]. The external exposure rates or radioactivity concentration of a specific sample may limit the time that workers will be permitted to remain in intimate contact with the samples, or may dictate that smaller samples be taken and special holding areas be provided for collected samples prior to shipment. These special handling considerations may conflict with the size specifications for the analytical method, normal sampling procedures, or equipment. There is a potential for biasing sampling programs by selecting samples that can be safely handled or legally shipped to support laboratories. Because final status surveys are performed to demonstrate that a site can be safely released, issues associated with high levels of radioactivity are not expected to be a concern.

There are three methods for collecting radiation data while performing a survey:

  • Direct measurement. A direct measurement is obtained by placing the detector near or against the surface or in the media being surveyed and reading the radioactivity level directly.
  • Scanning. An evaluation technique performed by moving a portable radiation detection instrument at a constant speed and distance above the surface to semi-quantitatively detect elevated areas of radiation.
  • Sampling. The process of collecting a portion of an environmental medium as representative of the locally remaining medium. The collected portion of the medium is then analyzed to determine the radionuclide concentration.

In practice, there has to be obtained a proper balance among the use of various measurement techniques. In general, there is an inverse correlation between the cost of a specific measurement technique and the detection levels being sought. Depending on the survey objectives, important considerations include survey costs and choosing the optimum instrumentation and measurement mix.

A certain minimum number of direct measurements or samples will be needed to demonstrate compliance with the release criterion based on the non-parametric statistical tests (see Section 3.3.10). In addition, the potential for areas of elevated contamination will have to be considered for designing scanning surveys. Areas of elevated activity may also affect the number of measurements; however, scanning with survey instruments should generally be sufficient to ensure that no areas with unusually high levels of radioactivity are left in place. Some measurements may also provide information of a qualitative nature to supplement other measurements. An example of such an application is in-situ gamma spectrometry to demonstrate the absence (or presence) of specific contaminants.

Table 3.12 presents a list of common contaminants along with recommended survey methods that have proven to be effective based on past survey experience in the decommissioning industry. This table provides a general indication of the detection capability of commercially-available instruments. In the next section more detailed information can be found on detection sensitivity.

Table 3.12 may be used to provide an initial evaluation of instrument capabilities for some common radio-nuclides at the example DCGLs listed in the table. For example, consider the contamination of a surface with 241Am.

Table 3.12 indicates that 241Am is detectable at the example DCGLs, and that viable direct measurement instruments include gas-flow proportional (α-mode) and alpha scintillation detectors. Table 3.12 should not be interpreted as providing specific values for an instrument’s detection sensitivity, which is discussed in Section 3.3.7. In addition, NRC draft report [USNRC-2006], [USNRC-1995] provides further information on factors that may affect survey instrumentation selection.

Structure surfaces Land areas Direct measurement instruments2

Nuclide Example DCGL1 (Bq/m2) Detectable Example DCGL1 (Bq/m2)

Detectable Surface activity Soil activity Exposure rate
3H 1.6×106 No 1.5×104 No ND6 ND ND
14C 4.7×105 Yes 1.4×103 No GPβ ND ND
54Mn 1.3×104 Yes 450 Yes GPβ7, GM γS, ISγ PIC, γS, ISγ
55Fe 1.8×106 No 4.1×105 No5 ND ND (ISγ) ND (ISγ)
60Co 3.1×103 Yes 110 Yes GPβ, GM γS, ISγ PIC, γS, ISγ
63Ni 1.5×106 Yes 2.8×105 No GPβ ND ND
90Sr 6.0×103 Yes 420 No5 GPβ, GM ND (GM, GPβ) ND
99Tc 6.5×105 Yes 1.9×103 No GPβ, GM ND ND
137Cs 8.2×103 Yes 400 Yes GPβ, GM γS, ISγ PIC, γS, ISγ
152Eu 6.6×103 Yes 240 Yes GPβ, GM γS, ISγ PIC, γS, ISγ
226Ra( C )3 970 Yes 210 Yes GPα, αS γS, ISγ PIC, γS, ISγ
232Th( C )3 340 Yes 320 Yes GPα, αS, GPβ γS, ISγ PIC, γS, ISγ
U4 560 Yes 710 Yes GPα, αS, GPβ, ISγ γS, ISγ,GPβ PIC, γS, ISγ
239Pu, 240Pu, 241Pu 120 Yes 70 No5 GPα, αS ND (ISγ) ND
241Am 110 Yes 70 Yes GPα, αS γS, ISγ PIC, γS, ISγ

Table 3.12 Selection of direct measurement techniques based on experience

1 Example DCGLs based on values given in NRC draft report (USNRC-1994)
2 GPα = Gas-flow proportional counter (α-mode).
GM = Geiger-Mueller survey meter.
GPβ = Gas-flow proportional counter (β-mode).
PIC = Pressurized ionization chamber.
αS = Alpha scintillation survey meter.
γS = Gamma scintillation (gross).
ISγ = In-situ gamma spectrometry.
3 For decay chains having two or more radio-nuclides of significant half-life that reach secular equilibrium.
The notation “( c )” indicates the direct measurement techniques assume the presence of progeny in the chain.
4 Depleted, natural and enriched.
5 Possibly detectable at limits for areas of elevated activity.
6 Not detectable.
7 Bold indicates the preferred method where alternative methods are available.

Sample characteristics such as sample depth, volume, area, moisture level, and composition, as well as sample preparation techniques which may alter the sample, are important planning considerations for Data Quality Objectives. Sample preparation may include, but is not limited to, removing extraneous material, homogenizing, splitting, drying, compositing, and final preparation of samples. As is the case for determining survey unit characteristics, the physical sample characteristics and sampling method should be consistent with the dose or risk pathway modelling that is used to determine radio-nuclide DCGL’s. If a direct measurement method is used, it should also be consistent with the pathway modelling.

For example, a sample depth of 15 cm (6 in.) for soil samples might be specified during the DQO process for a final status survey because this corresponds to the soil mixing or plow depth in several environmental pathway models. If contamination exists at a depth less than this, a number of models uniformly mix it throughout this depth to simulate the soil mixing associated with plowing. Similarly, models may be based on dry weight, which may necessitate either drying samples or data transformation to account for dry weight.

The DQOs and subsequent direction to the laboratory for analysis might include removal of material not relevant for characterizing the sample, such as pieces of glass, twigs, or leaves. Table 3.13 provides examples of how a particular field soil composition of fine-, medium-, and coarse-grained materials might determine laboratory analysis DQOs for particular radio-nuclides. Fine materials consist of clay (less than 0.002 mm) and silt (0.002 to 0.062 mm). Medium materials consist of sand, which can be further divided into very fine, fine, medium, coarse, and very coarse sand. Coarse materials consist of gravel, which is composed of pebbles (2 to 64 mm), cobbles (64 to 256 mm), and boulders (greater than 256 mm).

Separate out and evaluate fine-grain material because re-suspension is associated with the fine grain fraction for the air pathway.
If contamination resides on sand, pebbles, and cobbles, analyze these materials for direct exposure pathway and analyze the fine-grain fraction for the air pathway.
Separation and homogenization are not necessary for analyses because direct exposure pathway depends upon the average concentration and presence of cobbles will usually not impact laboratory analysis.
Determine if pathway modelling considered the presence of cobbles.
Separate, homogenize, and evaluate fine-grain material because plant root uptake is associated with the fine-grain fraction for the plant ingestion pathway.
Separate, homogenize, and evaluate fine-grain materials because of their relevance for the contaminant source term for contaminant migration to the sub-surface for the water pathway.

Table 3.13 Example of DQO planning considerations

3.3.7.2 Detection sensitivity

The detection sensitivity of a measurement system refers to a radiation level or quantity of radioactive material that can be measured or detected with some known or estimated level of confidence. This quantity is a factor of both the instrumentation and the technique or procedure being used.

The primary parameters that affect the detection capability of a radiation detector are the background count rate, the detection efficiency of the detector and the counting time interval. It is important to use actual background count rate values and detection efficiencies when determining counting and scanning parameters, particularly during final status and verification surveys.

When making field measurements, the detection sensitivity will usually be less than what can be achieved in a laboratory due to increased background and, often times, significantly lower detection efficiency. It is often impossible to guarantee that pure alpha emitters can be detected in-situ since the weathering of aged surfaces will often completely absorb the alpha emissions. The report [USNRC-1997] contains data on many of the parameters that affect detection efficiencies in-situ, such as absorption, surface smoothness, and particulate radiation energy.

3.3.7.3 Direct measurement sensitivity

Prior to performing field measurements, an investigator must evaluate the detection sensitivity of the equipment proposed for use to ensure that levels below the DCGL can be detected. After a direct measurement has been made, it is then necessary to determine whether or not the result can be distinguished from the instrument background response of the measurement system. The terms that are used in this manual to define detection sensitivity for fixed point counts and sample analyses are:

  • Critical level (LC);
  • Detection limit (LD);
  • Minimum detectable concentration (MDC).

The critical level (LC) is the level, in counts, at which there is a statistical probability (with a pre-determined confidence) of incorrectly identifying a measurement system background value as “greater than background.” Any response above this level is considered to be greater than background. The detection limit (LD) is an a priori estimate of the detection capability of a measurement system, and is also reported in units of counts. The minimum detectable concentration (MDC) is the detection limit (counts) multiplied by an appropriate conversion factor to give units consistent with a site guideline, such as Bq/kg.

The following discussion provides an overview of the derivation contained in the well known publication by Currie [CURRIE] followed by a description of how the resulting formulae should be used. Publications by Currie [CURRIE], NRC [USNRC] and Altshuler and Pasternak [Altshuler] provide details of the derivations involved.

The two parameters of interest for a detector system with a background response greater than zero are:

  • LC the net response level, in counts, at which the detector output can be considered “above background”;
  • LD the net response level, in counts that can be expected to be seen with a detector with a fixed level of certainty.

Assuming that a system has a background response and that random uncertainties and systematic uncertainties are accounted for separately, these parameters can be calculated using Poisson statistics. For these calculations, two types of decision errors should be considered. A Type I error (or “false positive”) occurs when a detector response is considered to be above background when, in fact, only background radiation is present. A Type II error (or “false negative”) occurs when a detector response is considered to be background when in fact radiation is present at levels above background. The probability of a Type I error is referred to as α (alpha) and is associated with LC; the probability of a Type II error is referred to as β (beta) and is associated with LD. Figure 3.3 graphically illustrates the relationship of these terms with respect to each other and to a normal background distribution.

Figure 3.3 Graphically represented probabilities for Type I and Type II errors in detection sensitivity for instrumentation with a background response
Figure 3.3 Graphically represented probabilities for Type I and Type II errors in detection sensitivity for instrumentation with a background response

If α and β are assumed to be equal, the variance (σ2) of all measurement values is assumed to be equal to the values themselves. If the background of the detection system is not well known, then the critical detection level and the detection limit can be calculated by using the following formulae:

LC = k√(2B)
LD = k2 + 2k√(2B) …………………………………………………………………………………. (3-6)

where:
LC = critical level (counts);
LD = detection limit (counts);
K = Poisson probability sum for α and β (assuming α and β are equal);
B = number of background counts that are expected to occur while performing an actual measurement.

The curve to the left in the diagram is the background distribution minus the mean of the background distribution. The result is a Poisson distribution with a mean equal to zero and a variance, σ2, equal to B. Note that the distribution accounts only for the expected statistical variation due to the stochastic nature of radioactive decay. Currie assumed “paired blanks” when deriving the above stated relationships [CURRIE] which are interpreted to mean that the sample and background count times are the same.
If values of 0.05 for both α and β are selected as acceptable, then k = 1.645 (from Appendix E, Table E.12 or Table 3.33) and Equation 3-6 can be written as:

LC = 2.33√(2B)
LD = 3 + 4.65√(2B) …………………………………………………………………………………. (3-7)

Note: In Currie’s derivation, the constant factor of 3 in the LD formula was stated as being 2.71, but since that time it has been shown [BRODSKY] and generally accepted that a constant factor of 3 is more appropriate. If the sample count times and background count times are different, a slightly different formulation is used.

For an integrated measurement over a preset time, the MDC can be obtained from Equation 3-7 by multiplying by the factor, C. This factor is used to convert from counts to concentration as shown in Equation 3-8:

MDC= C ×(3 + 4.65√(B)) ………………………………………………………………………… (3-8)

The total detection efficiency and other constants or factors represented by the variable C are usually not truly constants as shown in Equation 3-8. It is likely that at least one of these factors will have a certain amount of variability associated with it which may or may not be significant. These varying factors are gathered together into the single constant, C, by which the net count result will be multiplied when converting the final data. If C varies significantly between measurements, then it might be best to select a value, C’, from the observed distribution of C values that represents a conservative estimate. For example, a value of C might be selected to ensure that at least 95% of the possible values of C are less than the chosen value, C’. The MDC calculated in this way helps assure that the survey results will meet the Data Quality Objectives. This approach for including uncertainties into the MDC calculation is recommended in both [USNRC-1984], [ANSI-1996]. Underestimating an MDC can have adverse consequences, especially if activity is later detected at a level above the stated MDC.

Summary of direct measurement sensitivity terms:

  • The MDC is the a priori net activity level above the critical level that an instrument can be expected to detect 95% of the time. This value should be used when stating the detection capability of an instrument. The MDC is the detection limit, LD, multiplied by an appropriate conversion factor to give units of activity. Again, this value is used before any measurements are made and is used to estimate the level of activity that can be detected using a given protocol.
  • The critical level, LP~C~], is the lower bound on the 95% detection interval defined for LD and is the level at which there is a 5% chance of calling a background value “greater than background.” This value should be used when actually counting samples or making direct radiation measurements. Any response above this level should be considered as above background (i.e., a net positive result). This will ensure 95% detection capability for LD.
    From a conservative point of view, it is better to overestimate the MDC for a measurement method. Therefore, when calculating MDC and LC values, a measurement system background value should be selected that represents the high end of what is expected for a particular measurement method. For direct measurements, probes will be moved from point to point and, as a result, it is expected that the background will most likely vary significantly due to variations in background, source materials, and changes in geometry and shielding. Ideally, the MDC values should be calculated for each type of area, but it may be more economical to simply select a background value from the highest distribution expected and use this for all calculations. For the same reasons, realistic values of detection efficiencies and other process parameters should be used when possible and should be reflective of the actual conditions. To a great degree, the selection of these parameters will be based on judgment and will require evaluation of site-specific conditions.
  • MDC values for other counting conditions may be derived from Equation 3-8 depending on the detector and contaminants of concern. For example, it may be required to determine what level of contamination, distributed over 100 cm2, can be detected with a 500 cm2 probe or what contamination level can be detected with any probe when the contamination area is smaller than the probe active area. Table 3.14 lists several common field survey detectors with estimates of MDC values for 238U on a smooth, flat plane. As such, these represent minimum MDC values and may not be applicable at all sites. Appropriate site-specific MDC values should be determined using the DQO Process.
Approximate sensitivity

Detector Probe area (cm2) Background (cpm) Efficiency (cpm/dpm) LC LD MDC (Bq/m2)a

Alpha proportional 50 1 0.15 2 7 150
Alpha proportional 100 1 0.15 2 7 83
Alpha proportional 600 5 0.15 5 13 25
Alpha proportional 50 1 0.15 2 7 150
Beta proportional 100 300 0.20 40 83 700
Beta proportional 600 1500 0.20 90 183 250
Beta proportional 15 40 0.20 15 32 1800

Table 3.14 Examples of estimated detection sensitivities for alpha and beta survey instrumentation. (Static one minute counts for 238U calculated using Equations 3-7 and 3-8)
aAssumes that the size of the contamination area is at least as large as the probe area.

Example 3.9: Calculation of the MDC in Bq/m2 of an instrument with a 15 cm2 probe are

Sample Calculation 1:
The following example illustrates the calculation of an MDC in Bq/m2 for an instrument with a 15 cm^2^ probe area when the measurement and background counting times are each one minute:

B = 40 counts
C = (5 dpm/count)(Bq/60 dpm)(1/15 cm2 probe area)(10,000 cm2/m2)
= 55.6 Bq/m2-counts

The MDC is calculated using Equation 3-8:
MDC = 55.6 × (3 + 4.65√(40)) = 1,800 Bq/m2 (1,100 dpm/100 cm2)

The critical level, LC, for this example is calculated from Equation 3.7:
LC = 2.33√(B) = 15 counts

Given the above scenario, if a person asked what level of contamination could be detected 95% of the time using this method, the answer would be 1,800 Bq/m2 (1,100 dpm/100 cm2). When actually performing measurements using this method, any count yielding greater than 55 total counts, or greater than 15 net counts (55-40=15) during a period of one minute, would be regarded as greater than background.

3.3.7.4 Scanning sensitivity

The ability to identify a small area of elevated radioactivity during surface scanning is dependent upon the surveyor’s skill in recognizing an increase in the audible or display output of an instrument. For notation purposes, the term “scanning sensitivity” is used throughout this section to describe the ability of a surveyor to detect a pre-determined level of contamination with a detector. The greater the sensitivity, the lower the level of contamination that can be detected.

Many of the radiological instruments and monitoring techniques typically used for occupational health physics activities may not provide the detection sensitivities necessary to demonstrate compliance with the DCGLs. The detection sensitivity for a given application can be improved (i.e., lower the MDC) by:

  • Selecting an instrument with a higher detection efficiency;
  • Selecting an instrument with a lower background;
  • Decreasing the scanning speed,
  • Increasing the size of the effective probe area without significantly increasing the background response.

Scanning is usually performed during radiological surveys in support of decommissioning to identify the presence of any areas of elevated activity. The probability of detecting residual contamination in the field depends not only on the sensitivity of the survey instrumentation when used in the scanning mode of operation, but is also affected by the surveyor’s ability – i.e., human factors. The surveyor must make a decision whether the signals represent only the background activity, or residual contamination in excess of background. The greater the sensitivity, the lower the level of contamination that may be detected by scanning. Accounting for these human factors represents a significant change from the traditionally accepted methods of estimating scanning sensitivities.

An empirical method for evaluating the detection sensitivity for contamination surveys is by actual experimentation or, since it is certainly feasible, by simulating an experimental set-up using computer software. The following steps provide a simple example of how one can perform this empirical evaluation:

  • A desired nuclide contamination level is selected.
  • The response of the detector to be used is determined for the selected nuclide contamination level.
  • A test source is constructed which will give a detector count rate equivalent to what was determined in step 2. The count rate is equivalent to what would be expected from the detector when placed on an actual contamination area equal in value to that selected in step 1.
  • The detector of choice is then moved over the source at different scan rates until an acceptable speed is determined.

The most useful aspect of this approach is that the source can then be used to show surveyors what level of contamination is expected to be targeted with the scan. They, in turn, can gain experience with what the expected response of the detector will be and how fast they can survey and still feel comfortable about detecting the target contamination level. The person responsible for the survey can then use this information when developing a fixed point measurement and sampling plan.

The remainder of this section is dedicated to providing the reader with information pertaining to the underlying processes involved when performing scanning surveys for alpha, beta, and gamma emitting radio-nuclides. The purpose is to provide relevant information that can be used for estimating realistic scanning sensitivities for survey activities.

3.3.7.5 Scanning sensitivity for beta and gamma emitting nuclides

The minimum detectable concentration of a scan survey (scan MDC) depends on the intrinsic characteristics of the detector (efficiency, physical probe area, etc.), the nature (type and energy of emissions) and relative distribution of the potential contamination (point versus distributed source and depth of contamination), scan rate, and other characteristics of the surveyor. Some factors that may affect the surveyor’s performance include the costs associated with various outcomes – e.g., fatigue, noise, level of training, experience – and the survey’s a priori expectation of the likelihood of contamination present. For example, if the surveyor believes that the potential for contamination is very low, as in a Class 3 area, a relatively large signal may be required for the surveyor to conclude that contamination is present. NRC draft report [USNRC-1997a] provides a complete discussion of the human factors as they relate to the performance of scan surveys.

3.3.7.5.1 Signal detection theory

Personnel conducting radiological surveys for residual contamination at decommissioning sites must interpret the audible output of a portable survey instrument to determine when the signal (“clicks”) exceeds the background level by a margin sufficient to conclude that contamination is present. It is difficult to detect low levels of contamination because both the signal and the background vary widely. Signal detection theory provides a framework for the task of deciding whether the audible output of the survey meter during scanning is due to background or signal plus background levels. An index of sensitivity (d’) that represents the distance between the means of the background and background plus signal (see Figure 3.3 for determining LD), in units of their common standard deviation, can be calculated for various decision errors (correct detection and false positive rate).

As an example for a correct detection rate of 95% (complement of a false negative rate of 5%) and a false positive rate of 5%, d’ is 3.29 (similar to the static MDC for the same decision error rates).

The index of sensitivity is independent of human factors, and therefore, the ability of an ideal observer (theoretical construct), may be used to determine the minimum d’ that can be achieved for particular decision errors. The ideal observer makes optimal use of the available information to maximize the percent correct responses, providing an effective upper bound against which to compare actual surveyors. Table 3.15 lists selected values of d’.

3.3.7.5.2 The two stages of scanning

The framework for determining the scan MDC is based on the premise that there are two stages of scanning. That is, surveyors do not make decisions on the basis of a single indication. Rather, upon noting an increased number of counts, they pause briefly and then decide whether to move on or take further measurements. Thus, scanning consists of two components:

  • Continuous monitoring;
  • Stationary sampling.

In the first component, characterized by continuous movement of the probe, the surveyor has only a brief “look” at potential sources, determined by the scan speed. The surveyor’s willingness to decide that a signal is present at this stage is likely to be liberal, in that the surveyor should respond positively on scant evidence, since the only “cost” of a false positive is a little time.
The second component occurs only after a positive response was made at the first stage. This response is marked by the surveyor interrupting his scanning and holding the probe stationary for a period of time, while comparing the instrument output signal during that time to the background counting rate. Owing to the longer observation interval, sensitivity is relatively high. For this decision, the criterion should be more strict, since the cost of a “yes” decision is to spend considerably more time taking a static measurement or a sample.

False positive proportion True positive proportion

0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95
0.05 1.90 2.02 2.16 2.32 2.48 2.68 2.92 3.28
0.10 1.54 1.66 1.80 1.96 2.12 2.32 2.56 2.92
0.15 1.30 1.42 1.56 1.72 1.88 2.08 2.32 2.68
0.20 1.10 1.22 1.36 1.52 1.68 1.88 2.12 2.48
0.25 0.93 1.06 1.20 1.35 1.52 1.72 1.96 2.32
0.30 0.78 0.91 1.05 1.20 1.36 1.56 1.80 2.16
0.35 0.64 0.77 0.91 1.06 1.22 1.42 1.66 2.02
0.40 0.51 0.64 0.78 0.93 1.10 1.30 1.54 1.90
0.45 0.38 0.52 0.66 0.80 0.97 1.17 1.41 1.77
0.50 0.26 0.38 0.52 0.68 0.84 1.04 1.28 1.64
0.55 0.12 0.26 0.40 0.54 0.71 0.91 1.15 1.51
0.60 0.00 0.13 0.27 0.42 0.58 0.82 1.02 1.38

Table 3.15 Values of d’ for selected true positive and false positive proportions

Since scanning can be divided into two stages, it is necessary to consider the survey’s scan sensitivity for each of the stages. Typically, the minimum detectable count rate (MDCR) associated with the first scanning stage will be greater due to the brief observation intervals of continuous monitoring – provided that the length of the pause during the second stage is significantly longer. Typically, observation intervals during the first stage are on the order of 1 or 2 seconds, while the second stage pause may be several seconds long. The greater value of MDCR from each of the scan stages is used to determine the scan sensitivity for the surveyor.

3.3.7.5.3 Determination of the minimum detectable count rate (MDCR) and use of surveyor efficiency

The minimum detectable number of net source counts in the interval is given by si. Therefore, for an ideal observer, the number of source counts required for a specified level of performance can be arrived at by multiplying the square root of the number of background counts by the detectability value associated with the desired performance (as reflected in d’) as shown in Equation 3-9:

si = d’√(bi) ………………………………………………………………………………………. (3-9)

where the value of d’ is selected from Table 3.14 based on the required true positive and false positive rates and bi is the number of background counts in the interval.

Example 3.10: Determination of the minimum detectable count rate (MDCR) for scanning surveys

Suppose that one wished to estimate the minimum count rate that is detectable by scanning in an area with a background of 1,500 cpm. Note that the minimum detectable count rate must be considered for both scan stages – and the more conservative value is selected as the minimum count rate that is detectable. It will be assumed that a typical source remains under the probe for 1 second during the first stage, therefore, the average number of background counts in the observation interval is 25 (bi = 1500 × (1/60)). Furthermore, as explained earlier, it can be assumed that at the first scanning stage a high rate (e.g., 95%) of correct detections is required, and that a correspondingly high rate of false positives (e.g., 60%) will be tolerated. From Table 3.15, the value of d’, representing this performance goal, is 1.38. The net source counts needed to support the specified level of performance (assuming an ideal observer) will be estimated by multiplying 5 (the square root of 25) by 1.38. Thus, the net source counts per interval, si, needed to yield better than 95% detections with about 60% false positives is 6.9. The minimum detectable source count rate, in cpm, may be calculated by:

MDCR = si × (60/i) ……………………………………………………………………. (3-10)

For this example, MDCR is equivalent to 414 cpm (1,914 cpm gross). Table 3.16 provides the scan sensitivity for the ideal observer (MDCR) at the first scanning stage for various background levels, based on an index of sensitivity (d’) of 1.38 and a 2-second observation interval.

Background [cpm] MDCR [net cpm] Scan sensitivity [gross cpm]

45 50 95
60 60 120
260 120 380
300 130 430
350 140 490
400 150 550
1000 240 1240
3000 410 3410
4000 480 4480

Table 3.16 Scanning sensitivity (MDCR) of the ideal observer for various background levelsa
a The sensitivity of the ideal observer during the first scanning stage is based on an index of sensitivity (d’) of 1.38 and a 2-second observation interval.

The minimum number of source counts required to support a given level of performance for the final detection decision (second scan stage) can be estimated using the same method. As explained earlier, the performance goal at this stage will be more demanding. The required rate of true positives remains high (e.g., 95%), but fewer false positives (e.g., 20%) can be tolerated, such that d’ (from Table 3.15) is now 2.48. One will assume that the surveyor typically stops the probe over a suspect location for about 4 seconds before making a decision, so that the average number of background counts in an observation interval is 100 (bi = 1,500 × (4/60)). Therefore, the minimum detectable number of net source counts, si, needed will be estimated by multiplying 10 (the square root of 100) by 2.48 (the d’ value); so si equals 24.8. The MDCR is calculated by 2.48 × (60/4) and equals 372 cpm. The value associated with the first scanning stage (this example, 414 cpm) will typically be greater, owing to the relatively brief intervals assumed.

Laboratory studies using simulated sources and backgrounds were performed to assess the abilities of surveyors under controlled conditions. The methodology and analysis of results for these studies are described in [USNRC-1997 and 1997a]. The surveyor’s actual performance as compared with that which is ideally possible (using the ideal observer construct) provided an indication of the efficiency of the surveyors. Based on the results of the confidence rating experiment, this surveyor efficiency (p) was estimated to be between 0.5 and 0.75.
EURSSEM recommends assuming an efficiency value at the lower end of the observed range (i.e., 0.5) when making MDC estimates. Thus, the required number of net source counts for the surveyor, MDCR~surveyor~, is determined by dividing the MDCR by the square root of p. Continuing with this example, the surveyor MDCR is calculated by 414 cpm/0.707, or 585 cpm (2,085 cpm gross).

3.3.7.6. Determination of scan MDC’s for land and structure surfaces

The survey design for determining the number of data points for areas of elevated activity (see Section 3.5.1) depends on the scan MDC for the selected instrumentation. In general, alpha or beta scans are performed on structure surfaces to satisfy the elevated activity measurements survey design, while gamma scans are performed for land areas. Because of low background levels for alpha emitters, the approach described here is not generally applied to determining scan MDCs for alpha contaminants – rather, the reader is referred to Section 3.3.7.7 for an appropriate method for determining alpha scan MDCs for building surfaces. In any case, the data requirements for assessing potential elevated areas of direct radiation depend on the scan MDC of the survey instrument (e.g., floor monitor, GM detector, NaI(Tl) scintillation detector).

3.3.7.6.1 Determination of scan MDCs for land areas

In addition to the MDCR and detector characteristics, the scan MDC (in pCi/g) for land areas is based on the area of elevated activity, depth of contamination, and the radionuclide (i.e., energy and yield of gamma emissions). If one assumes constant parameters for each of the above variables, with the exception of the specific radionuclide in question, the scan MDC may be reduced to a function of the radionuclide alone. NaI(Tl) scintillation detectors are generally used for scanning land areas.

An overview of the approach used to determine scan MDCs for land areas follows. The NaI(Tl) scintillation detector background level and scan rate (observation interval) are postulated, and the MDCR for the ideal observer, for a given level of performance, is obtained. After a surveyor efficiency is selected, the relationship between the surveyor MDCR (MDCRsurveyor) and the radio-nuclide concentration in soil (in Bq/kg or pCi/g) is determined. This correlation requires two steps – first, the relationship between the detector’s net count rate to net exposure rate (cpm per μR/h) is established, and second, the relationship between the radio-nuclide contamination and exposure rate is determined.

For a particular gamma energy, the relationship of NaI(Tl) scintillation detector count rate and exposure rate may be determined analytically (in cpm per μR/h). The approach used to determine the gamma fluence rate necessary to yield a fixed exposure rate (1 μR/h) – as a function of gamma energy – is provided in [USNRC-1997]. The NaI(Tl) scintillation detector response (cpm) is related to the fluence rate at specific energies, considering the detector’s efficiency (probability of interaction) at each energy. From this, the NaI(Tl) scintillation detector versus exposure rates for varying gamma energies are determined. Once the relationship between the NaI(Tl) scintillation detector response (cpm) and the exposure rate is established, the MDCR~surveyor~ (in cpm) of the NaI(Tl) scintillation detector can be related to the minimum detectable net exposure rate. The minimum detectable exposure rate is used to determine the minimum detectable radionuclide concentration (i.e., the scan MDC) by modelling a specified small area of elevated activity.

Modelling (using MicroshieldTM) of the small area of elevated activity (soil concentration) is used to determine the net exposure rate produced by a radionuclide concentration at a distance 10 cm above the source. This position is selected because it relates to the average height of the NaI(Tl) scintillation detector above the ground during scanning.

The factors considered in the modelling include:

  • Radio-nuclide of interest (considering all gamma emitters for decay chains);
  • Expected concentration of the radio-nuclide of interest;
  • Areal dimensions of the area of elevated activity;
  • Depth of the area of elevated activity;
  • Location of dose point (NaI(Tl) scintillation detector height above the surface);
  • Density of soil.

Modelling analyses are conducted by selecting a radionuclide (or radioactive material decay series) and then varying the concentration of the contamination. The other factors are held constant – the areal dimension of a cylindrical area of elevated activity is 0.25 m2 (radius of 28 cm), the depth of the area of elevated activity is 15 cm, the dose point is 10 cm above the surface, and the density of soil is 1.6 g/cm3. The objective is to determine the radio-nuclide concentration that is correlated to the minimum detectable net exposure rate.

Example 3.11: Calculation of scan MDC for a NaI(Tl) scintillation detector for a land area

The scan MDC for ^137^Cs using a 1.5 in. by 1.25 in. NaI(Tl) scintillation detector is considered in detail. Assume that the background level is 4,000 cpm and that the desired level of performance, 95% correct detections and 60% false positive rate, results in a d’ of 1.38. The scan rate of 0.5 m/s provides an observation interval of 1 second (based on a diameter of about 56 cm for the area of elevated activity). The MDCRsurveyor may be calculated assuming a surveyor efficiency (p) of 0.5 as follows:

bi = (4,000 cpm) × (1 sec) × (1 min/60 sec) = 66.7 counts

MDCR = (1.38) × (√(66.7)) × (60 sec/1 min) ) = 680 cpm

MDCRsurveyor = 680/√(0.5) = 960 cpm

The corresponding minimum detectable exposure rate is determined for this detector and radionuclide. The manufacturer of this particular 1.5 in. by 1.25 in. NaI(Tl) scintillation detector quotes a count rate to exposure rate ratio for 137Cs of 350 cpm per µR/h. The minimum detectable exposure rate is calculated by dividing the count rate (960 cpm) by the count rate to exposure rate ratio for the radio-nuclide of interest (350 cpm per µR/h). The minimum detectable exposure rate for this example is 2.73 µR/h.

Both 137Cs and its short-lived progeny, 137mBa, were chosen from the MicroshieldTM library. The source activity and other modelling parameters were entered into the modelling code. The source activity was selected based on an arbitrary concentration of 5 pCi/g. The modelling code performed the appropriate calculations and determined an exposure rate of 1.307 µR/h (which accounts for build-up). Finally, the radio-nuclide concentrations of 137Cs and 137mBa (scan MDC) necessary to yield the minimum detectable exposure rate (2.73 µR/h) may be calculated using the following formula:

scan MDC = (5 pCi/g) x (2.73 µR/h) / 1.307 µR/h = 10.4 pCi/g ……………………. (3-11)

It must be emphasized that while a single scan MDC value can be calculated for a given radionuclide – other scan MDC values may be equally justifiable depending on the values chosen for the various factors, including the MDCR (background level, acceptable performance criteria, and observation interval), surveyor efficiency, detector parameters and the modelling conditions of the contamination. It should also be noted that determination of the scan MDC for radioactive materials – like uranium and thorium – must consider the gamma radiation emitted from the entire decay series. The document [USNRC-1997] provides a detailed example of how the scan MDC can be determined for enriched uranium.

Table 3.17 provides a number of scan MDCs for common radio-nuclides and radioactive materials in soil. It is important to note that the variables used in the above examples to determine the scan MDCs for the 1.25 in. by 1.5 in. NaI(Tl) scintillation detector – i.e., the MDCRsurveyor detector parameters (e.g., cpm per µR/h), and the characteristics of the area of elevated activity – have all been held constant to facilitate the calculation of scan MDCs provided in Table 3.17. The benefit of this approach is that generally applicable scan MDCs are provided for different radioactive contaminants. Additionally, the relative detectability of different contaminants is evident because the only variable in Table 3.17 is the nature of the contaminant.

Radionuclide / Radioactive material 1.25 inch by 1.5 inch NaI detector 2 inch by 2 inch NaI detector
Scan MDC (Bq/kg) Weighted cpm/µR/h Scan MDC (Bq/kg) Weighted cpm/µR/h

241Am 1,650 5,830 1,170 13,000
60Co 215 160 126 430
137Cs 385 350 237 900
232Th 111,000 4,300 78,400 9,580
226Ra (in equilibrium with progeny) 167 300 104 760
232Th decay series (sum of all-radionuclide in the thorium decay series) 1,050 340 677 830
232Th (in equilibrium with progeny in decay series) 104 340 66.6 830
Depleted uraniumb (0.34% 235U 2,980 1,680 2,070 3,790
Natural uraniumb 4,260 1,770 2,960 3,990
3% Enriched uraniumb 5,070 2,010 3,540 4,520
20% Enriched uraniumb 5,620 2,210 3,960 4,940
50% Enriched uraniumb 6,220 2,240 4,370 5,010
75% Enriched uraniumb 6,960 2,250 4,880 5,030

Table 3.17 NaI(Tl) scintillation detector scan MDCs for common radiological contaminantsa

aRefer to text for complete explanation of factors used to calculate scan MDCs. For example, the background level for the 1.25 inch by 1.5 inch NaI detector was assumed to be 4,000 cpm, and 10,000 cpm for the 2 inch by 2 inch NaI detector. The observation interval was 1 second and the level of performance was selected to yield d’ of 1.38.
bScan MDC for uranium includes sum of 238U, 235U, and 234U.

As noted above, the scan MDCs calculated using the approach in this section are dependent on several factors. One way to validate the appropriateness of the scan MDC is by tracking the residual radioactivity (both surface activity and soil concentrations) levels identified during investigations performed as a result of scanning surveys. The measurements performed during these investigations may provide an a posteriori estimate of the scan MDC that can be used to validate the a priori scan MDC used to design the survey.

3.3.7.6.2 Determination of scan MDCs for building/structures

The scan MDC is determined from the minimum detectable count rate (MDCR) by applying conversion factors that account for detector and surface characteristics and surveyor efficiency. As discussed above, the MDCR accounts for the background level, performance criteria (d’¬), and observation interval. The observation interval during scanning is the actual time that the detector can respond to the contamination source – this interval depends on the scan speed, detector size in the direction of the scan, and area of elevated activity. Because the actual dimensions of potential areas of elevated activity in the field cannot be known a priori, EURSSEM recommends postulating a certain area (e.g., perhaps 50 to 200 cm^2^), and then selecting a scan rate that provides a reasonable observation interval.

Scan MDC = MDCR / [ √(p) εi εs (probe area/100 cm2) ] ……………………………………… (3.12)

where:
MDCR = minimum detectable count rate;
εi = instrument efficiency;
εs = surface efficiency;
p = surveyor efficiency.

Example 3.12: Determination of a scan MDC for building/structure surfaces

The scan MDC (in dpm/100 cm2) for 99Tc on a concrete surface may be determined for a background level of 300 cpm and a 2-second observation interval using a hand-held gas proportional detector (126 cm2 probe area). For a specified level of performance at the first scanning stage of 95% true positive rate and 60% false positive rate (and assuming the second stage pause is sufficiently long to ensure that the first stage is more limiting), d’ equals 1.38 (see Table 3.15) and the MDCR is 130 cpm (see Table 3.16). Using a surveyor efficiency of 0.5, and assuming instrument and surface efficiencies of 0.36 and 0.54, respectively, the scan MDC is calculated using Equation 3-12:

Scan MDC = 130 / {√(0.5) (0.36) (0.54) (1.26)} = 750 dpm/100 cm2

Additional examples for calculating the scan MDC may be found in [USNRC-1997].

3.3.7.7 Determining a scan MDC for alpha emitters

Scanning for alpha emitters differs significantly from scanning for beta and gamma emitters in that the expected background response of most alpha detectors is very close to zero. The following discussion covers scanning for alpha emitters and assumes that the surface being surveyed is similar in nature to the material on which the detector was calibrated. In this respect, the approach is purely theoretical. Surveying surfaces that are dirty, non-planar, or weathered can significantly affect the detection efficiency and therefore bias the expected MDC for the scan. The use of reasonable detection efficiency values instead of optimistic values is highly recommended. Appendix D contains a complete derivation of the alpha scanning equations used in this section.

Since the time a contaminated area is under the probe varies and the background count rate of some alpha instruments is less than 1 cpm, it is not practical to determine a fixed MDC for scanning. Instead, it is more useful to determine the probability of detecting an area of contamination at a predetermined DCGL for given scan rates.

For alpha survey instrumentation with backgrounds ranging from < 1 to 3 cpm, a single count provides a surveyor sufficient cause to stop and investigate further. Assuming this to be true, the probability of detecting given levels of alpha surface contamination can be calculated by use of Poisson summation statistics.
Given a known scan rate and a surface contamination DCGL, the probability of detecting a single count while passing over the contaminated area is:

P(n > 1) = 1 – e-G E d / 60 v ………………………………………………………………………(3-13)

where:
P(n > 1) = probability of observing a single count;
G = contamination activity (dpm);
E = detector efficiency (4π);
d = width of detector in direction of scan (m);
v = scan speed (m/s);

See Appendix D for a complete derivation of these formulas.
Once a count is recorded and the guideline level of contamination is present the surveyor should stop and wait until the probability of getting another count is at least 90%. This time interval can be calculated by:

t = 13,800 / (C A E) ………………………………………………………………………………….(3-14)

where:
t = time period for static count (s);
C = contamination guideline (dpm/100 cm2);
A = physical probe area (cm2);
E = detector efficiency (4 π).

Many portable proportional counters have background count rates on the order of 5 to 10 cpm, and a single count should not cause a surveyor to investigate further. A counting period long enough to establish that a single count indicates an elevated contamination level would be prohibitively inefficient. For these types of instruments, the surveyor usually will need to get at least 2 counts while passing over the source area before stopping for further investigation.
Assuming this to be a valid assumption, the probability of getting two or more counts can be calculated by:

P(n > 2) = 1 – P(n = 0) – P(n = 1)
= 1 – (1 + (G x E + B) x t / 60) (e-(G x E + B) x /60) …………………………………………………(3-15)

where:
P(n > 2) = probability of getting 2 or more counts during the time interval t;
P(n = 0) = probability of not getting any counts during the time interval t;
P(n = 1) = probability of getting 1 count during the time interval t;
B = background count rate (cpm).

All other variables are the same as for Equation 3-13.

Appendix D provides a complete derivation of Equations 3-13 through 3-15 and a detailed discussion of the probability of detecting alpha surface contamination for several different variables. Several probability charts are included at the end of Appendix C for common detector sizes. Table 3.18 provides estimates of the probability of detecting 300 dpm/100 cm2 for some commonly used alpha detectors.

Detector type Detector efficiency (cpm/dpm) Probe dimension in direction of scan (cm)

Scan rate (cm/s) Probability of detecting 300 dpm/100 cm2
Proportional

0.20 5 3 80%
Proportional

0.15 15 5 90%
Scintillation

0.15 5 3 70%
Scintillation

0.15 10 3 90%

Table 3.18 Probability of detecting 300 dpm/100 cm2 of alpha activity while scanning with alpha detectors using an audible output (calculated using Equation 3-13)

3.3.7.8 Sensitivity of mobile systems with integrated positioning systems

In recent years, the advent of new technologies has introduced mobile sensor systems for acquiring data that include fully-integrated positioning systems. Portable and vehicle-based versions of these systems record survey data while moving over surfaces to be surveyed and simultaneously recording the location data from either a roving DGPS receiver or local microwave/sonar receiver. All measurement data are automatically stored and processed with the measurement location for later posting (see Section 3.10.8.5 for a discussion of posting plots) or for mapping the results. These systems are designed with a variety of detectors for different applications. For example, alpha or beta detectors have been mounted on a robot a fixed distance over a smooth surface. The robot moves at a predetermined speed over the surface to provide scanning results, and also records individual direct measurements at predetermined intervals. This type of system not only provides the necessary measurement data, but also reduces the uncertainty associated with human factors. Other systems are equipped with several types of radiation detectors, magnetometers, electromagnetic sensors, or various combinations of multiple sensors. The limitations of each system should be evaluated on a site-specific basis to determine if the positioning system, the detector, the transport system, or some combination based on site-specific characteristics will represent the limits of the system.

Use current sensitivity standard: ISO 11929-2010
– by Rafael Garcia-Bermejo Fernandez almost 3 years ago