RADON (THORON) DAUGHTER AUTOMATED,
MEASUREMENTS
PROGRAMMABLE,
W I T H AN
RADIATION MONITOR
J. B I G U
Research Scientist and Radiation Project Leader, Elliot Lake Laboratory, CANMET, Energy, Mines and Resources Canada, Elliot Lake, Ontario, Canada and G. V A N D R I S H
Research and Development Leader, Pylon Electronic Development, Ottawa, Canada
Abstract. A technical evaluation of an automated, programmable, grab-sampler manufactured by Pylon
Electronic Development under the commercial name of WL-1000C has been conducted. Six different methods are implemented in the instrument for analyzing radiation data. Any one method can be used on command and easily selected by means of a keyboard. Available radiation data that can be retrieved on command include radon (thoron) daughter concentrations and radon (tboron) Working Levels. Measurements were carried out under laboratory-controlled conditions in a large (26 m 3) radon/thoron test facility, designed for calibration purposes, and at an underground uranium mine. Data obtained with the WL-1000C have been compared with conventional grab-sampling (e.g., Kusnetz, Thomas-Tsivoglou and Markov methods) and with other automated radiation instrumentation previously tested at our laboratories. Tests were done under constant radiation conditions and also under rapidly fluctuating conditions in order to determine the response of the instruments and methods in these two cases of practial interest. The Working Level used in these experiments was in the approximate range of 0.01 to 10 W L Tests were conducted under a variety of environmental conditions. Good agreement with grab-sampling data was found for radon daughters. Some discrepancies with grab-sampling data were found for radon daughter/thoron daughter mixtures. Disagreement in the latter case is to be expected because of e-energy overlap between RaA and ThC.
I. Introduction Adequate mine air quality control is attained by monitoring radiation levels. Proper radiation monitoring prevents unnecessary risk to occupational workers. Alpha- and ~ monitoring is common in uranium mines and the uranium milling industry. However, unless high grade ores are mined and processed, only a-radiaton is usually considered. The a-radiation variable of interest is in most cases the Working Level (WL), a measure of the a-energy emitted by the radon gas progeny. Radiation monitoring is essentially of two kinds: grab-sampling and continuous monitoring. In the grab-sampling method an air sample is taken and the radiation level is determined. Most of the experimental methods that have been developed for the determination of the WL involve the measurement of a-particles from either RaA or RaC' or both [1-8]. However, techniques based on t-particle count have also been investigated and applied [ 9-12]. Relatively simple techniques involving a single total aor//-count are commonly used when the only radiation variable of interest is the WL Environmental Monitoring and Assessment 6 (1986) 59-70. 9 1986 by D. Reidel Publishing Company.
60
J. BIGU AND G. VANDRISH
[1-5, 10, 11 ]. Other more sophisticated techniques based on several a-counts enable the calculation of the radon daughters concentrations in addition to the WL (7,8). The Kusnetz [2, 3], Markov [1] and Rolle [4, 5] methods, based on a single total a-count, and the fl-particle method, e.g., [10], based on a single total fl-count, permit a simple determination of the WL. A more accurate determination of the WL, in addition to the calculation of the concentration of the radon decay products, is possible if use is made of other more sophiticated techniques, such as the Tsivoglou method [7], the Thomas method [8] or some a-spectrometric methods [13-15]. Furthermore, instrumentation and concepts have been developed that allow the determination of the WL to be made in a matter of a few minutes. The instruments are commonly known as Instant Working Level Meters (IWLM). Evans and Schroeder [16] and later Groer [ 17], Rolle [ 18], Hill [ 19] and James and Strong [20] must be credited with the early development of the concept and/or basic instrumentation that has led to the design of several IWLM prototypes (see also ref. [21 and 22]). Instrumentation and methods operating on continuous monitoring and time-integrating techniques [23, 24] have also been developed. A summary of radiation instrumentation has been given by Bigu elsewhere [25]. The need for a simple, fast and reliable method for the routine determination of WL by a small-sized, automated instrument capable of performing the sampling, counting and raw data processing operations with minimum attendance has been sought for a long time. In addition, it would be highly desirable that the instrument be easily portable, light, electronically reliable, rugged and dust- and moisture-proof as its use is intended for harsh environmental conditions. Pylon Instruments (Ottawa) has recently developed a microprocessor controlled Working Level monitor, commercially available under the trade name WL-1000C, capable of determining the concentrations and WL of the radon progeny by several methods. Furthermore, the instrument permits the determination of the thoron progeny in addition to the radon progeny by using an a-spectroscopy method [ 15, 26], a very desirable feature as thoron is also present in a number of uranium mines and other environments. This paper pertains to a technical evaluation conducted on the instrument under laboratory controlled conditions and in an underground uranium mine, and is a continuation of work conducted on the same instrument [27]. 2. Description of the Instrument The WL-1000C is essentially an automated, programmable, grab-sampler with a-spectroscopy capabilities. The instrument has three energy windows (i.e., channels) of width adjusted to count a-particles from RaA (6.0 MeV) and ThC (6.1 MeV) in the first channel, and RaC' (7.7 MeV) and ThC' (8.8 MeV) in the second and third channels, respectively. The instrument is divided into two main sections. The lower section contains the instrument power supply consisting of several rechargeable Ni-Cd batteries. The upper
RADON (THORON) DAUGHTER MEASUREMENTS
61
section houses a silicon-barrier a-detector as well as a microprocessor and associated electronics. Sampling of air is done by means of a sampling pump/sampling head unit connected to the upper section of the WL-1000C by means of a cable approximately 1 m in length. The instrument uses special filters mounted on plastic cards, thereby eliminating the use of tweezers and simplifying considerably the handling of samples. The WL-1000C only necessitates manual positioning of the sample, i.e., filter, in the sampling head and transfer of the sample from the sampling head to the detector system when the sampling period is over. The rest of the operational procedure is fully automatic, including the sampling period. Several methods for determining the radiation variables of interest are programmed into the microprocessor. A total of six methods are available on command which can easily be selected by means of a keyboard. The methods are: two a-spectroscopy methods, one of the methods permits determination of the radon daughter progeny, while the second method allows the determination of radon daughter/thoron daughter Mixtures [15, 26]; two Kusnetz methods, one of which is a shorter version of the conventional method [2, 3, 26]; the Rolle method [4,5], and the Thomas-Tsivoglou method [6-8]. Also available is a short calibration routine which can be used to determine the instrument's detector a-counting efficiency, e, and to verify the alignment of energy discriminators for the a-spectroscopy methods [27]. The 0~-countingefficiency of the detector, e, and the flow rate of the sampling pump are programmable from the keyboard. Several data are available from the instrument through a read-out unit. Those include total a-count in the energy channels, radon (thoron) daughter concentrations, and radon (thoron) daughter Working Levels. Status report on several instrument variables can also be displayed on command. The WL-1000C is provided with a RS-232 interface output which allows transfer of data from the instrument to an external printer. In addition, the signal from the instrument, i.e., the a-spectrum of the radon and thoron progenies deposited on the filter during and/or after sampling can be studied using an external multichannel analyzer.
3. Experimental Procedure The WL-1000C was tested in the laboratory and in an underground uranium mine. All methods programmed into the instrument were tested and compared with independent grab-sampling by conventional methods, e.g., Thomas-Tsivoglou, Markov and Kusnetz methods, and with two radon daughter continuous monitoring systems. The instrument was tested in radon daughter only atmospheres and in radon daughter/thoron daughter mixtures. Laboratory experiments were mainly conducted in a large radon/thoron room designed for calibration purposes. The radiation levels used ranged from less that 0.02 WL to about 10 WL. Underground WL's were in the 0.2 to 0.5 range. The a-counting efficiency of the detector and the flow rate of the sampling pump were carefully determined prior to the evaluation of the instrument. The WL-1000C was tested under constant and rapidly fluctuating radiation conditions. This was aimed at comparing the time response of grab-sampling instruments,
62
J. B I G U AND G. VANDRISH
e.g., the WL-1000C, versus continuous monitoring systems. A Working Level Monitor operating on time-integrating principles manufactured by EDA Instruments, and known under the commercial name of WLM-300, and a radon daughter-WL semi-continuous system manufactured by Harshaw Instruments, under the commercial name of Radon Daughter Detector Model 101 and Radiation Computer Model 100, were also used in the tests. For simplicity, the Harshaw system will be referred to hereafter as the Harshaw-WLM. Environmental conditions varied from low aerosol concentration (1.0 x 103 to 2.0 • 103 cm -3) and 50-60~/o relative humidity in the radon/thoron room to dusty, humid ( ~ 80%) and high aerosol concentration (> 1.0 x 105 cm- 3) in an underground uranium mine. 4. Experimental Results and Discussion
The results obtained are partly summarized in Fig. 1 to 6 and Table I. Figures l(a), (b), and (c) show ~-spectra from air sampling filters containing radon daughters only, thoron daughters only, and a radon daughter/thoron daughter mixture, respectively. The spectra were obtained by feeding the signal from the detector amplifier output to an external multichannel analyzer, and show the high energy resolution and energy discrimination of the detector and associated circuitry. The symbols W1, W2, and W3 stand for first, second and third energy windows, respectively. Figures 2 and 3 show the radon daughter Working Level, WL(Rn), versus time as determined by the WL-1000C, the Harshaw WLM, the EDA WLM-300 and grabsampling data using the Kusnetz and Markov methods. The tests were conducted in a large radon/thoron test facility under rapidly changing conditions to investigate the response of instruments as compared with grab-sampling data. The results obtained with the WL- 1000C are between those obtained by grab-sampling using the Markov and Kusnetz methods and the other automated instruments. Notice that there is a time lag ( ~ 1 h) between the data by the WLM-300 and other data. This behaviour is predicted theoretically for instruments operating on continuous sampling and time-integrating principles. The time lag for thoron daughters is substantially higer [23, 24]. It is also of interest to notice that the Markov Method gives slightly higher values ( g 10 ~o) than measurements conducted using the Kusnetz and Thomas-Tsivoglou methods. Figures 2 and 3 show that there is fair to good agreement between the WL-1000C and grabsampling data, and data obtained with other automated instrumentation previously tested at our facilities. Figures 4 and 5 show radon daughter concentration, and WL(Rn) and WL(Tn) data obtained in an underground uranium mine with the WL-1000C, and by grab-sampling using the Thomas-Tsivoglou method and several gross a-count methods for WL(Tn) [28-30]. Data with the WL-1000C were gathered using two different a-spectroscopy methods, namely one method to determine radon daughters in radon daughter atmospheres only, and a method specially designed to determine atmospheric radon daughter/thoron daughter mixtures [ 15, 26]. Good agreement was found for RaB and
RADON (THORON) DAUGHTER MEASUREMENTS
63
RaC' 1000 800
9
600 400 200 0
100
50
150 WI
200
W2
Channel number
W3
|
I
ThC' 1000
800 The
600
|
l
400
200 0
100
50
W1
I
!
15~ W2
200 W3
Channel
number
I
ThC'
1400 1200
The I
1000 800
9
6OO 4OO 2OO 0 5O
100 I
W1
I
150 W2
I
200 W3
Channel number
Fig. 1. Alpha-spectrafrom air samplingfilters containingradondaughtersonly (a), thorondaughtersonly (b), and radon daughter/thorondaughtermixtures (c). The symbols W1, W2 and W3 stand for first second and third energy windows, respectively.
RaC whereas substantial disagreement (up to > 50 ~o) was observed for RaA, the values for WL-1000C lying lower than grab-sampling data. This is not surprising as the accuracy of most gross ~t-count grab-sampling methods for RaA is rather limited, particularly at low concentration levels. However, as the contribution of RaA to WL(Rn) is in most cases substantially lower than that corresponding to RaB and RaC, even large differences in RaA concentration do not affect WL(Rn) significantly. The contribution of RaA to WL(Rn) becomes quite significant only for 'young' air, e.g., freshly filtered randon/radon daughter atmosphere. Under equilibrium conditions, the
64
J. BIGU AND G. VANDRISH
1t
L
1.0
0.91-
0.8
0.7
0.4
0.3
0.2
I 8
9
-10
II
12
13
-o-= IZ:~;;0c
14
15
Time, h Fig.2. WL(Rn) versus time for the WL- I000C (Rn a-spectroscopyand Rollemethods) and grab-sampling by the Markov method, and the Kusnetz method (A). Also shown are data obtained with the H a r s h a w - W L M . E x p e r i m e n t s w e r e c a r r i e d o u t in in a r a d o n r o o m .
1.4
'\ 1.2
1.0
,t,~,." /,Z/
~0.8
0.6
\ \ ,/" \? x._.,.,..
o//~'/,/ /~>--
0.4
0.2
0
a
i
i
i
9
10
11
12
i
13 14 Time, h
i
i
I
i
i
15
16
17
18
19
Fig. 3. WL(Rn) versus time for the WL-1000C (Rn ~-spectroscopy, - l - , and Rolle method, - O - ) and the M a r k o v m e t h o d , - I S ] - , a n d the E D A W L M - 3 0 0 ( - . - ) . M e a s u r e m e n t s w e r e c a r r i e d o u t in a r a d o n room.
RADON(THORON)
DAUGHTER
MEASUREMENTS
4
4
~3
3
~176176
2
A'----------A/A
0
o
o z
68
J. B I G U A N D G. V A N D R I S H
to obtain as many samples as possible during a regular working shift and to correct the data accordingly by adequate correction factors experimentally determined. It is a straightforward matter to expand the programmability of the instrument to include other more elaborate methods for determining radon daughter/thoron daughter mixtures [31]. Because the instrument is essentially an automated, quite flexible, 'grab-sampler', which uses open face filter sampling techniques, there is no reason to believe that the accuracy of the instrument should be any different from that of the grab-sampling methods used, utilizing more conventional instrumentation, with which the instrument is compared. Indeed, this is the case, within experimental error, when data from the WL-1000C and by grab-sampling are compared using the same method. Agreement between WL-1000C data and grab-sampling data, obtained by more conventional instrumentation, depends only on: (a) the accuracy with which sampling flow-rate, detector c~-counting efficiency, and timing (e.g., sampling, waiting and counting) are determined in both cases, and (b) purely statistical considerations. (Note: the discussion above is based on the following assumptions (i) simultaneous, parallel, sampling takes place, (ii) uniformity of radiation level over the volume where sampling is done, and (iii) no sampling mutual interference occurs due to proximity of sampling instrumentation, different sampling flow rates and/or other undesirable effects.) Comparison of grab-sampling data obtained with the WL-1000C and data obtained with continuous monitoring, time integrating, instrumentation shows that although the latter is quite 'convenient' as it requires a minimum of manual operation and attendance, it has some inherent limitations because of its response time delay and, hence, difficulty in following 'quantitatively' rapid radiation level fluctuations. Although the WL-1000C requires more manual operation than continuous monitors, its accuracy is inherently better for discrete, spot sampling purposes, and hence for routine calibration purposes. 5. Conclusions
Good agreement for radon daughters was found between the data obtained with the WL-1000C and conventional grab-sampling data using the Kusnetz, Thomas-Tsivogtou and Markov methods. Good agreement was also found between the WL-1000C and other automated instruments previously tested at our laboratories. Some disagreement was found when measuring radon daughter/thoron daughter mixtures, which reflect the difficulty of past and present methodologies for determining radon daughters in the presence of thoron daughters, and vice versa. The main difficulty arises from e-energy overlap from RaA and ThC, and the 'short' a-spectroscopy method used here due to practical constraints. Better agreement is found when use is made of the long e-spectroscopy method, but the latter is rather limited for routine field work. Because the WL-1000C is essentially a grab-sampler with open-face sampling configuration, the accuracy of the experimental data obtained with the instrument is, in principle, as good as the data obtained by grab-sampfing using the same method and more conventional instrumentation. Differences that may be found when comparing WL-1000C and grab-sampling data are mainly related to the accuracy with which the
RADON (THORON) DAUGHTERMEASUREMENTS
69
sampling flow rate and detector a-counting efficiency (plus discriminator alignment in the case of the WL-1000C) are determined in both cases. Once a given method has been chosen as a reference method, the WL-1000C could be used as a reference, 'standard', instrument suitable for calibration purposes. Acknowledgements The authors would like to express their appreciation to M. Grenier for his technical assistance and preliminary calibration work conduted with the instrument. We would also like to thank Ms. G. Szlapetis for her assistance in some laboratory measurements. References [1] Markov, K. P., Ryabov, N. V., and Stas, K. N.: 1962, 'A Rapid Method for Estimating the Hazard Associated with the Presence of Radon Daughter Products in Air', At. Ehnerg. 12, 315. [2] Kusnetz, H. L.: 1956, 'Radon Daughters in Mine Atmospheres', Am. Ind. Hyg. Assoc. J. 17, 1. [3] Kusnetz, H. L.: 1956, 'Radon Daughters in Mine atmospheres - A Field Method for Determining Concentrations', Ind. Hyg. Quarterly. [4] Rolle, R.: 1969, 'Improved Radon Daughter Monitoring Procedure',Am. Ind. Hyg. Assoc. J. 153-160. [5] Rolle, R.: 1972, 'Rapid Working Level Monitoring', Health Phys. 22, 233-238. [6] Tsivoglou, E.C.: 1953, 'Quantitative Description of Successive Transformations in Atmospheric Samples', Ph.D. Dissertation; Ohio State University. [7] Tsivoglou, E. C., Ayer, H. E., and Holaday, D. A.: 1953, 'Occurrence of Non-Equilibrium Atmospheric Mixtures of Radon and Its Daughters'; Nucleonics 11, 9, 40-45. [8] Thomas, J. W.: 1972, 'Measurement of Radon Daughters in Air', Health Phys. 23, 783-789. [9] Lockhart, L. B, Patterson, R. L., and Hosler, C. R.: 1965, U.S.N.R.L. Report 6229. [10] Droullard, R. F. and Holub, R. F.: 1977, 'Continuous Working Level Measurements Using Alpha or Beta Detectors', Report of Investigations RI 8237; U.S. Bureau of Mines. [11] Holmgren, R. M.: 1974, 'Working Levels of Radon Daughters in Air Determined from Measurements of RaB and RaC'; Health Phys. 27, 141-145. [12] Horwood, J. L.: unpublished data. [13] Kerr, G.D.: 1975, 'Measurement of Radon Progeny Concentrations in Air by Alpha-particle Spectrometry'; Oak Ridge; Technical Report ORNL-TM-4924. [14] Martz, D. E., Holleman, D. F., McCurdy, D. E., and Schiager, K. J.: 1969, 'Analysis of Atmospheric Concentrations of RaA, RaB and RaC by Alpha Spectroscopy', Health Phys. 17, 131-138. [15] Tremblay, R. J., Leclerc, A., Mathieu, C., Pepin, R., and Townsend, M. G.: 1979, 'Measurement of Radon Progeny Concentration in Air by a-Particle Spectrometric Counting During and After Air Sampling', Health Phys. 36, 401-411. [16] Schroeder, G.L.: 1969, 'Some Basic Concepts in Uranium Mine Ventilation, Radium and Mesothorium Poisoning and Dosimetry and Instrumentation Techniques in Applied Radioactivity'; Annual Progress Report from Massachusetts Institute of Technology; MIT-952-5 (May 1968); also see MIT-952-6. [17] Groer, P. G., Keefe, D. J., McDowell, W. F., and Selman, R.F.: 1974, 'An Instant Working Level Meter with Automatic Radon-Daughter Readout', Final Report from Argonne National Laboratory (Argonne, Illinois, U.S.A.) to U.S.B.M. [18] Rolle, R.: 1970, 'Determination of Radon Daughters on Filters by a Simple Liquid Scintillation Technique', Am. Ind. Hyg. Assoc. J. 31, 718-721. [19] Hill, A.: 1975, 'Rapid Measurement of Radon, Decay Products, Unattached Fractions and Working Level Values of Mine Atmospheres'; Health Phys. 28, 4, 472-474. [20] James, A. C. and Strong, J. C.: 1973, 'A Radon Daughter Monitor for Use in Mines', Proc 3rdlnt Congr IRPA, Washington; pp. 932-938. [21] Groer, P. G., Keefe, D. J., McDowell, W. F., and Selman, R. F.: 1975, 'Development of a Prototype
70
J. BIGU AND G. VANDRISH
Instant Working Level Meter with Automatic Individual Radon Daughter Readout', Argonne National Laboratory Report PB-249 545 (BuMines OFR 19-76). [22] Bigu, J., Raz, R., Golden, K., and Dominguez, P.: 1983, 'Design and Development of a ComputerBased Continuous Monitor for the Determination of the Short-lived Decay Products of Radon and Thoron'; Division Report MRP/MRL 83-40(J), CANMET, Energy, Mines and Resources Canada. [23] Bigu, J. and Grenier, M.: 1982, Division Reports MRP/MRL 82-36(TR), MRP/MRL 82-65(TR) and MRP/MRL 82-79(TR), CANMET, Energy, Mines and Resources Canada, Reports deal with a technical evaluation of a radon daughter continuous monitor manufactured by EDA Instruments under the commercial name WLM-300. [24] Bigu, J.,Kaldenbach, R.,andGrenier, M.: 1983,'DesignmadPerformanceofaMicroprocessor-Based, Automated, Continuous Monitor with ~-Spectroscopy Capabilities for the Determination of Radon and Thoron and Their Shortlived Decay Products', Division Report MRP/MRL 83-78 (TR). [25] Bigu, J.: 1981, 'Monitoring of Radiation Variables in the Underground Environment', CIMBulletin, 74, 83-92. [26] Cot6, P. and Townsend, M. G.: 1979, 'Effect of Thoron Daughters on Radon Daughter Counting Procedures: Examination of ct- and fl-Counting Methods', Division Report MRP/MSL 79-39(TR); CANMET, Energy, Mines and Resources Canada. [27] Bigu, J. and Grenier, M.: 1982, 'Comparison of Radon (and Thoron) Daughter Measurements by an Automated, Programmable, Grab-Sampler with "Conventional" Grab-Sampling', Division Report MRP/MRL 82-75(TR); CANMET, Energy, Mines and Resources Canada; August 1982. See also Division Report MRP/MRL 82-69(TR). [28] Rock, R. L.: 1975, 'Sampling Mine Atmospheres for Potential Alpha Energy Due to the Presence of Radon-220 (Thoron) Daughters'; MESA Informational Report IR 1015; U.S. Dept. of the Interior. [29] Bigu, J. andLau, W. K.: 1983,'DeterminationoftheThoron Daughter Working Level by a One Gross Alpha-Count', Division Report MRP/MRL 83-13(TR); CANMET, Energy, Mines and Resources Canada. [30] Bigu, J. and Grenier, M.: 1983, 'Thoron Daughter Working Level Measurements by One and Two Gross Alpha-Count Methods', Division Report MRP/MRL 83-57(TR); CANMET, Energy, Mines and Resources Canada. [31] Khan, A., Busigin, A., and Phillips, C. R.: 1982, 'An Optimized Scheme for Measurement of the Concentrations of the Decay Products of Radon and Thoron', Health Phys. 42, 809-826.
MEASUREMENTS
PROGRAMMABLE,
W I T H AN
RADIATION MONITOR
J. B I G U
Research Scientist and Radiation Project Leader, Elliot Lake Laboratory, CANMET, Energy, Mines and Resources Canada, Elliot Lake, Ontario, Canada and G. V A N D R I S H
Research and Development Leader, Pylon Electronic Development, Ottawa, Canada
Abstract. A technical evaluation of an automated, programmable, grab-sampler manufactured by Pylon
Electronic Development under the commercial name of WL-1000C has been conducted. Six different methods are implemented in the instrument for analyzing radiation data. Any one method can be used on command and easily selected by means of a keyboard. Available radiation data that can be retrieved on command include radon (thoron) daughter concentrations and radon (tboron) Working Levels. Measurements were carried out under laboratory-controlled conditions in a large (26 m 3) radon/thoron test facility, designed for calibration purposes, and at an underground uranium mine. Data obtained with the WL-1000C have been compared with conventional grab-sampling (e.g., Kusnetz, Thomas-Tsivoglou and Markov methods) and with other automated radiation instrumentation previously tested at our laboratories. Tests were done under constant radiation conditions and also under rapidly fluctuating conditions in order to determine the response of the instruments and methods in these two cases of practial interest. The Working Level used in these experiments was in the approximate range of 0.01 to 10 W L Tests were conducted under a variety of environmental conditions. Good agreement with grab-sampling data was found for radon daughters. Some discrepancies with grab-sampling data were found for radon daughter/thoron daughter mixtures. Disagreement in the latter case is to be expected because of e-energy overlap between RaA and ThC.
I. Introduction Adequate mine air quality control is attained by monitoring radiation levels. Proper radiation monitoring prevents unnecessary risk to occupational workers. Alpha- and ~ monitoring is common in uranium mines and the uranium milling industry. However, unless high grade ores are mined and processed, only a-radiaton is usually considered. The a-radiation variable of interest is in most cases the Working Level (WL), a measure of the a-energy emitted by the radon gas progeny. Radiation monitoring is essentially of two kinds: grab-sampling and continuous monitoring. In the grab-sampling method an air sample is taken and the radiation level is determined. Most of the experimental methods that have been developed for the determination of the WL involve the measurement of a-particles from either RaA or RaC' or both [1-8]. However, techniques based on t-particle count have also been investigated and applied [ 9-12]. Relatively simple techniques involving a single total aor//-count are commonly used when the only radiation variable of interest is the WL Environmental Monitoring and Assessment 6 (1986) 59-70. 9 1986 by D. Reidel Publishing Company.
60
J. BIGU AND G. VANDRISH
[1-5, 10, 11 ]. Other more sophisticated techniques based on several a-counts enable the calculation of the radon daughters concentrations in addition to the WL (7,8). The Kusnetz [2, 3], Markov [1] and Rolle [4, 5] methods, based on a single total a-count, and the fl-particle method, e.g., [10], based on a single total fl-count, permit a simple determination of the WL. A more accurate determination of the WL, in addition to the calculation of the concentration of the radon decay products, is possible if use is made of other more sophiticated techniques, such as the Tsivoglou method [7], the Thomas method [8] or some a-spectrometric methods [13-15]. Furthermore, instrumentation and concepts have been developed that allow the determination of the WL to be made in a matter of a few minutes. The instruments are commonly known as Instant Working Level Meters (IWLM). Evans and Schroeder [16] and later Groer [ 17], Rolle [ 18], Hill [ 19] and James and Strong [20] must be credited with the early development of the concept and/or basic instrumentation that has led to the design of several IWLM prototypes (see also ref. [21 and 22]). Instrumentation and methods operating on continuous monitoring and time-integrating techniques [23, 24] have also been developed. A summary of radiation instrumentation has been given by Bigu elsewhere [25]. The need for a simple, fast and reliable method for the routine determination of WL by a small-sized, automated instrument capable of performing the sampling, counting and raw data processing operations with minimum attendance has been sought for a long time. In addition, it would be highly desirable that the instrument be easily portable, light, electronically reliable, rugged and dust- and moisture-proof as its use is intended for harsh environmental conditions. Pylon Instruments (Ottawa) has recently developed a microprocessor controlled Working Level monitor, commercially available under the trade name WL-1000C, capable of determining the concentrations and WL of the radon progeny by several methods. Furthermore, the instrument permits the determination of the thoron progeny in addition to the radon progeny by using an a-spectroscopy method [ 15, 26], a very desirable feature as thoron is also present in a number of uranium mines and other environments. This paper pertains to a technical evaluation conducted on the instrument under laboratory controlled conditions and in an underground uranium mine, and is a continuation of work conducted on the same instrument [27]. 2. Description of the Instrument The WL-1000C is essentially an automated, programmable, grab-sampler with a-spectroscopy capabilities. The instrument has three energy windows (i.e., channels) of width adjusted to count a-particles from RaA (6.0 MeV) and ThC (6.1 MeV) in the first channel, and RaC' (7.7 MeV) and ThC' (8.8 MeV) in the second and third channels, respectively. The instrument is divided into two main sections. The lower section contains the instrument power supply consisting of several rechargeable Ni-Cd batteries. The upper
RADON (THORON) DAUGHTER MEASUREMENTS
61
section houses a silicon-barrier a-detector as well as a microprocessor and associated electronics. Sampling of air is done by means of a sampling pump/sampling head unit connected to the upper section of the WL-1000C by means of a cable approximately 1 m in length. The instrument uses special filters mounted on plastic cards, thereby eliminating the use of tweezers and simplifying considerably the handling of samples. The WL-1000C only necessitates manual positioning of the sample, i.e., filter, in the sampling head and transfer of the sample from the sampling head to the detector system when the sampling period is over. The rest of the operational procedure is fully automatic, including the sampling period. Several methods for determining the radiation variables of interest are programmed into the microprocessor. A total of six methods are available on command which can easily be selected by means of a keyboard. The methods are: two a-spectroscopy methods, one of the methods permits determination of the radon daughter progeny, while the second method allows the determination of radon daughter/thoron daughter Mixtures [15, 26]; two Kusnetz methods, one of which is a shorter version of the conventional method [2, 3, 26]; the Rolle method [4,5], and the Thomas-Tsivoglou method [6-8]. Also available is a short calibration routine which can be used to determine the instrument's detector a-counting efficiency, e, and to verify the alignment of energy discriminators for the a-spectroscopy methods [27]. The 0~-countingefficiency of the detector, e, and the flow rate of the sampling pump are programmable from the keyboard. Several data are available from the instrument through a read-out unit. Those include total a-count in the energy channels, radon (thoron) daughter concentrations, and radon (thoron) daughter Working Levels. Status report on several instrument variables can also be displayed on command. The WL-1000C is provided with a RS-232 interface output which allows transfer of data from the instrument to an external printer. In addition, the signal from the instrument, i.e., the a-spectrum of the radon and thoron progenies deposited on the filter during and/or after sampling can be studied using an external multichannel analyzer.
3. Experimental Procedure The WL-1000C was tested in the laboratory and in an underground uranium mine. All methods programmed into the instrument were tested and compared with independent grab-sampling by conventional methods, e.g., Thomas-Tsivoglou, Markov and Kusnetz methods, and with two radon daughter continuous monitoring systems. The instrument was tested in radon daughter only atmospheres and in radon daughter/thoron daughter mixtures. Laboratory experiments were mainly conducted in a large radon/thoron room designed for calibration purposes. The radiation levels used ranged from less that 0.02 WL to about 10 WL. Underground WL's were in the 0.2 to 0.5 range. The a-counting efficiency of the detector and the flow rate of the sampling pump were carefully determined prior to the evaluation of the instrument. The WL-1000C was tested under constant and rapidly fluctuating radiation conditions. This was aimed at comparing the time response of grab-sampling instruments,
62
J. B I G U AND G. VANDRISH
e.g., the WL-1000C, versus continuous monitoring systems. A Working Level Monitor operating on time-integrating principles manufactured by EDA Instruments, and known under the commercial name of WLM-300, and a radon daughter-WL semi-continuous system manufactured by Harshaw Instruments, under the commercial name of Radon Daughter Detector Model 101 and Radiation Computer Model 100, were also used in the tests. For simplicity, the Harshaw system will be referred to hereafter as the Harshaw-WLM. Environmental conditions varied from low aerosol concentration (1.0 x 103 to 2.0 • 103 cm -3) and 50-60~/o relative humidity in the radon/thoron room to dusty, humid ( ~ 80%) and high aerosol concentration (> 1.0 x 105 cm- 3) in an underground uranium mine. 4. Experimental Results and Discussion
The results obtained are partly summarized in Fig. 1 to 6 and Table I. Figures l(a), (b), and (c) show ~-spectra from air sampling filters containing radon daughters only, thoron daughters only, and a radon daughter/thoron daughter mixture, respectively. The spectra were obtained by feeding the signal from the detector amplifier output to an external multichannel analyzer, and show the high energy resolution and energy discrimination of the detector and associated circuitry. The symbols W1, W2, and W3 stand for first, second and third energy windows, respectively. Figures 2 and 3 show the radon daughter Working Level, WL(Rn), versus time as determined by the WL-1000C, the Harshaw WLM, the EDA WLM-300 and grabsampling data using the Kusnetz and Markov methods. The tests were conducted in a large radon/thoron test facility under rapidly changing conditions to investigate the response of instruments as compared with grab-sampling data. The results obtained with the WL- 1000C are between those obtained by grab-sampling using the Markov and Kusnetz methods and the other automated instruments. Notice that there is a time lag ( ~ 1 h) between the data by the WLM-300 and other data. This behaviour is predicted theoretically for instruments operating on continuous sampling and time-integrating principles. The time lag for thoron daughters is substantially higer [23, 24]. It is also of interest to notice that the Markov Method gives slightly higher values ( g 10 ~o) than measurements conducted using the Kusnetz and Thomas-Tsivoglou methods. Figures 2 and 3 show that there is fair to good agreement between the WL-1000C and grabsampling data, and data obtained with other automated instrumentation previously tested at our facilities. Figures 4 and 5 show radon daughter concentration, and WL(Rn) and WL(Tn) data obtained in an underground uranium mine with the WL-1000C, and by grab-sampling using the Thomas-Tsivoglou method and several gross a-count methods for WL(Tn) [28-30]. Data with the WL-1000C were gathered using two different a-spectroscopy methods, namely one method to determine radon daughters in radon daughter atmospheres only, and a method specially designed to determine atmospheric radon daughter/thoron daughter mixtures [ 15, 26]. Good agreement was found for RaB and
RADON (THORON) DAUGHTER MEASUREMENTS
63
RaC' 1000 800
9
600 400 200 0
100
50
150 WI
200
W2
Channel number
W3
|
I
ThC' 1000
800 The
600
|
l
400
200 0
100
50
W1
I
!
15~ W2
200 W3
Channel
number
I
ThC'
1400 1200
The I
1000 800
9
6OO 4OO 2OO 0 5O
100 I
W1
I
150 W2
I
200 W3
Channel number
Fig. 1. Alpha-spectrafrom air samplingfilters containingradondaughtersonly (a), thorondaughtersonly (b), and radon daughter/thorondaughtermixtures (c). The symbols W1, W2 and W3 stand for first second and third energy windows, respectively.
RaC whereas substantial disagreement (up to > 50 ~o) was observed for RaA, the values for WL-1000C lying lower than grab-sampling data. This is not surprising as the accuracy of most gross ~t-count grab-sampling methods for RaA is rather limited, particularly at low concentration levels. However, as the contribution of RaA to WL(Rn) is in most cases substantially lower than that corresponding to RaB and RaC, even large differences in RaA concentration do not affect WL(Rn) significantly. The contribution of RaA to WL(Rn) becomes quite significant only for 'young' air, e.g., freshly filtered randon/radon daughter atmosphere. Under equilibrium conditions, the
64
J. BIGU AND G. VANDRISH
1t
L
1.0
0.91-
0.8
0.7
0.4
0.3
0.2
I 8
9
-10
II
12
13
-o-= IZ:~;;0c
14
15
Time, h Fig.2. WL(Rn) versus time for the WL- I000C (Rn a-spectroscopyand Rollemethods) and grab-sampling by the Markov method, and the Kusnetz method (A). Also shown are data obtained with the H a r s h a w - W L M . E x p e r i m e n t s w e r e c a r r i e d o u t in in a r a d o n r o o m .
1.4
'\ 1.2
1.0
,t,~,." /,Z/
~0.8
0.6
\ \ ,/" \? x._.,.,..
o//~'/,/ /~>--
0.4
0.2
0
a
i
i
i
9
10
11
12
i
13 14 Time, h
i
i
I
i
i
15
16
17
18
19
Fig. 3. WL(Rn) versus time for the WL-1000C (Rn ~-spectroscopy, - l - , and Rolle method, - O - ) and the M a r k o v m e t h o d , - I S ] - , a n d the E D A W L M - 3 0 0 ( - . - ) . M e a s u r e m e n t s w e r e c a r r i e d o u t in a r a d o n room.
RADON(THORON)
DAUGHTER
MEASUREMENTS
4
4
~3
3
~176176
2
A'----------A/A
0
o
o z
68
J. B I G U A N D G. V A N D R I S H
to obtain as many samples as possible during a regular working shift and to correct the data accordingly by adequate correction factors experimentally determined. It is a straightforward matter to expand the programmability of the instrument to include other more elaborate methods for determining radon daughter/thoron daughter mixtures [31]. Because the instrument is essentially an automated, quite flexible, 'grab-sampler', which uses open face filter sampling techniques, there is no reason to believe that the accuracy of the instrument should be any different from that of the grab-sampling methods used, utilizing more conventional instrumentation, with which the instrument is compared. Indeed, this is the case, within experimental error, when data from the WL-1000C and by grab-sampling are compared using the same method. Agreement between WL-1000C data and grab-sampling data, obtained by more conventional instrumentation, depends only on: (a) the accuracy with which sampling flow-rate, detector c~-counting efficiency, and timing (e.g., sampling, waiting and counting) are determined in both cases, and (b) purely statistical considerations. (Note: the discussion above is based on the following assumptions (i) simultaneous, parallel, sampling takes place, (ii) uniformity of radiation level over the volume where sampling is done, and (iii) no sampling mutual interference occurs due to proximity of sampling instrumentation, different sampling flow rates and/or other undesirable effects.) Comparison of grab-sampling data obtained with the WL-1000C and data obtained with continuous monitoring, time integrating, instrumentation shows that although the latter is quite 'convenient' as it requires a minimum of manual operation and attendance, it has some inherent limitations because of its response time delay and, hence, difficulty in following 'quantitatively' rapid radiation level fluctuations. Although the WL-1000C requires more manual operation than continuous monitors, its accuracy is inherently better for discrete, spot sampling purposes, and hence for routine calibration purposes. 5. Conclusions
Good agreement for radon daughters was found between the data obtained with the WL-1000C and conventional grab-sampling data using the Kusnetz, Thomas-Tsivogtou and Markov methods. Good agreement was also found between the WL-1000C and other automated instruments previously tested at our laboratories. Some disagreement was found when measuring radon daughter/thoron daughter mixtures, which reflect the difficulty of past and present methodologies for determining radon daughters in the presence of thoron daughters, and vice versa. The main difficulty arises from e-energy overlap from RaA and ThC, and the 'short' a-spectroscopy method used here due to practical constraints. Better agreement is found when use is made of the long e-spectroscopy method, but the latter is rather limited for routine field work. Because the WL-1000C is essentially a grab-sampler with open-face sampling configuration, the accuracy of the experimental data obtained with the instrument is, in principle, as good as the data obtained by grab-sampfing using the same method and more conventional instrumentation. Differences that may be found when comparing WL-1000C and grab-sampling data are mainly related to the accuracy with which the
RADON (THORON) DAUGHTERMEASUREMENTS
69
sampling flow rate and detector a-counting efficiency (plus discriminator alignment in the case of the WL-1000C) are determined in both cases. Once a given method has been chosen as a reference method, the WL-1000C could be used as a reference, 'standard', instrument suitable for calibration purposes. Acknowledgements The authors would like to express their appreciation to M. Grenier for his technical assistance and preliminary calibration work conduted with the instrument. We would also like to thank Ms. G. Szlapetis for her assistance in some laboratory measurements. References [1] Markov, K. P., Ryabov, N. V., and Stas, K. N.: 1962, 'A Rapid Method for Estimating the Hazard Associated with the Presence of Radon Daughter Products in Air', At. Ehnerg. 12, 315. [2] Kusnetz, H. L.: 1956, 'Radon Daughters in Mine Atmospheres', Am. Ind. Hyg. Assoc. J. 17, 1. [3] Kusnetz, H. L.: 1956, 'Radon Daughters in Mine atmospheres - A Field Method for Determining Concentrations', Ind. Hyg. Quarterly. [4] Rolle, R.: 1969, 'Improved Radon Daughter Monitoring Procedure',Am. Ind. Hyg. Assoc. J. 153-160. [5] Rolle, R.: 1972, 'Rapid Working Level Monitoring', Health Phys. 22, 233-238. [6] Tsivoglou, E.C.: 1953, 'Quantitative Description of Successive Transformations in Atmospheric Samples', Ph.D. Dissertation; Ohio State University. [7] Tsivoglou, E. C., Ayer, H. E., and Holaday, D. A.: 1953, 'Occurrence of Non-Equilibrium Atmospheric Mixtures of Radon and Its Daughters'; Nucleonics 11, 9, 40-45. [8] Thomas, J. W.: 1972, 'Measurement of Radon Daughters in Air', Health Phys. 23, 783-789. [9] Lockhart, L. B, Patterson, R. L., and Hosler, C. R.: 1965, U.S.N.R.L. Report 6229. [10] Droullard, R. F. and Holub, R. F.: 1977, 'Continuous Working Level Measurements Using Alpha or Beta Detectors', Report of Investigations RI 8237; U.S. Bureau of Mines. [11] Holmgren, R. M.: 1974, 'Working Levels of Radon Daughters in Air Determined from Measurements of RaB and RaC'; Health Phys. 27, 141-145. [12] Horwood, J. L.: unpublished data. [13] Kerr, G.D.: 1975, 'Measurement of Radon Progeny Concentrations in Air by Alpha-particle Spectrometry'; Oak Ridge; Technical Report ORNL-TM-4924. [14] Martz, D. E., Holleman, D. F., McCurdy, D. E., and Schiager, K. J.: 1969, 'Analysis of Atmospheric Concentrations of RaA, RaB and RaC by Alpha Spectroscopy', Health Phys. 17, 131-138. [15] Tremblay, R. J., Leclerc, A., Mathieu, C., Pepin, R., and Townsend, M. G.: 1979, 'Measurement of Radon Progeny Concentration in Air by a-Particle Spectrometric Counting During and After Air Sampling', Health Phys. 36, 401-411. [16] Schroeder, G.L.: 1969, 'Some Basic Concepts in Uranium Mine Ventilation, Radium and Mesothorium Poisoning and Dosimetry and Instrumentation Techniques in Applied Radioactivity'; Annual Progress Report from Massachusetts Institute of Technology; MIT-952-5 (May 1968); also see MIT-952-6. [17] Groer, P. G., Keefe, D. J., McDowell, W. F., and Selman, R.F.: 1974, 'An Instant Working Level Meter with Automatic Radon-Daughter Readout', Final Report from Argonne National Laboratory (Argonne, Illinois, U.S.A.) to U.S.B.M. [18] Rolle, R.: 1970, 'Determination of Radon Daughters on Filters by a Simple Liquid Scintillation Technique', Am. Ind. Hyg. Assoc. J. 31, 718-721. [19] Hill, A.: 1975, 'Rapid Measurement of Radon, Decay Products, Unattached Fractions and Working Level Values of Mine Atmospheres'; Health Phys. 28, 4, 472-474. [20] James, A. C. and Strong, J. C.: 1973, 'A Radon Daughter Monitor for Use in Mines', Proc 3rdlnt Congr IRPA, Washington; pp. 932-938. [21] Groer, P. G., Keefe, D. J., McDowell, W. F., and Selman, R. F.: 1975, 'Development of a Prototype
70
J. BIGU AND G. VANDRISH
Instant Working Level Meter with Automatic Individual Radon Daughter Readout', Argonne National Laboratory Report PB-249 545 (BuMines OFR 19-76). [22] Bigu, J., Raz, R., Golden, K., and Dominguez, P.: 1983, 'Design and Development of a ComputerBased Continuous Monitor for the Determination of the Short-lived Decay Products of Radon and Thoron'; Division Report MRP/MRL 83-40(J), CANMET, Energy, Mines and Resources Canada. [23] Bigu, J. and Grenier, M.: 1982, Division Reports MRP/MRL 82-36(TR), MRP/MRL 82-65(TR) and MRP/MRL 82-79(TR), CANMET, Energy, Mines and Resources Canada, Reports deal with a technical evaluation of a radon daughter continuous monitor manufactured by EDA Instruments under the commercial name WLM-300. [24] Bigu, J.,Kaldenbach, R.,andGrenier, M.: 1983,'DesignmadPerformanceofaMicroprocessor-Based, Automated, Continuous Monitor with ~-Spectroscopy Capabilities for the Determination of Radon and Thoron and Their Shortlived Decay Products', Division Report MRP/MRL 83-78 (TR). [25] Bigu, J.: 1981, 'Monitoring of Radiation Variables in the Underground Environment', CIMBulletin, 74, 83-92. [26] Cot6, P. and Townsend, M. G.: 1979, 'Effect of Thoron Daughters on Radon Daughter Counting Procedures: Examination of ct- and fl-Counting Methods', Division Report MRP/MSL 79-39(TR); CANMET, Energy, Mines and Resources Canada. [27] Bigu, J. and Grenier, M.: 1982, 'Comparison of Radon (and Thoron) Daughter Measurements by an Automated, Programmable, Grab-Sampler with "Conventional" Grab-Sampling', Division Report MRP/MRL 82-75(TR); CANMET, Energy, Mines and Resources Canada; August 1982. See also Division Report MRP/MRL 82-69(TR). [28] Rock, R. L.: 1975, 'Sampling Mine Atmospheres for Potential Alpha Energy Due to the Presence of Radon-220 (Thoron) Daughters'; MESA Informational Report IR 1015; U.S. Dept. of the Interior. [29] Bigu, J. andLau, W. K.: 1983,'DeterminationoftheThoron Daughter Working Level by a One Gross Alpha-Count', Division Report MRP/MRL 83-13(TR); CANMET, Energy, Mines and Resources Canada. [30] Bigu, J. and Grenier, M.: 1983, 'Thoron Daughter Working Level Measurements by One and Two Gross Alpha-Count Methods', Division Report MRP/MRL 83-57(TR); CANMET, Energy, Mines and Resources Canada. [31] Khan, A., Busigin, A., and Phillips, C. R.: 1982, 'An Optimized Scheme for Measurement of the Concentrations of the Decay Products of Radon and Thoron', Health Phys. 42, 809-826.
