Operational Topics

A THERMOLUMINESCENT DOSIMETRY INTERCOMPARISON IN OPERATIONAL POWER STATION FIELDS Robert A. Facey,* David A. Agnew,t C. Ross Hirningt and Murray L. Walsht parison is given here; complete details are provided in reports of the Atomic Energy Control Board (AECB) of Canada (AECB 1989, 1991).

Abstracr-A dosimetry intercomparison was held among the five agencies in Canada that are recognized by the Atomic Energy Control Board as competent to perform external dosimetry. Exposures of thermoluminescent dosimeter badges were made under operational conditions to radiation fields in Candu nuclear generating stations. Details of the method are described including the large, block-type phantoms (with a rotating front face so that all badges were equally exposed) and a small device to measure the depth-dose distribution. Thirty-six exposures (or “runs”) were made, exposing 522 badges for periods of 1 h-2 d. Normalization between the runs was based on the absorbed dose at 1,000 mg cm-* for each run, as measured by the depth-dose device. Using this method, the average relative readings for the five participants ranged from 1.01-1.40 (dimensionless). Health Phys. 63(6):702-709; 1992 Key words: thermoluminescence dosimetry; phantom; power plants, nuclear; radiation, beta

History and aims

The initiative for the intercomparison came from an AECB Working Group on external dosimetry. The Working Group originally came together to review the various dosimetry models and practices in use among dosimetry processors in Canada. It recommended that a laboratory intercomparison be conducted, to be followed by a field (i.e., operational) intercomparison. The two intercomparisons were intended to show the degree of accord among the processors by comparing reported doses: (a) under ideal irradiation conditions, and (b) under the conditions found in a working reactor environment. In the first intercomparison, each processor sent dosimeters to the Chalk River Nuclear Laboratories (CRNL) where they were irradiated to known doses in gamma-ray, x-ray, and/or 90Sr-90Ybeta-ray beams. Dosimeters were then returned to the processor for reading. A detailed description of the first intercomparison has been given in an AECB report (AECB 1989). In the second intercomparison, the irradiations were done in pressurized heavy water (Candu) nuclear generating stations operated by Ontario Hydro (OH). The results of the second intercomparison are reported here.

INTRODUCTION PERSONAL radiation dosimeters worn by atomic radiation workers should perform within their design specifications at all times and under all operational circumstances. Calibrations under controlled laboratory conditions, along with quality control and performance testing, are normally considered sufficient to ensure this. In addition, intercomparisons between different dosimetry systems may help to increase confidence in performance. As with calibrations, intercomparisons are commonly carried out in laboratories using controlled irradiations in specified fields. We repoit here a dosimetry intercomparison using a variety of mixed radiation fields found in nuclear generating stations. For all the participants, this represented a “worst case” trial since the radiation fields were varied and complex. Only a summary of the intercom-

Participants

There are five agencies in Canada recognized by the AECB as competent to perform external dosimetry, and all five use thermoluminescent dosimeters (TLDs). The five agencies are listed in Table 1. Of the participants, CRNL is a branch of Atomic Energy of Canada Ltd., while New Brunswick Electric Power Commission (NBEPC), Hydro-Quebec (HQ), and OH are public electric utilities. All four of these agencies provide dosimetry only for their own employees and visitors to their sites. The fifth agency, the Department of National Health and Welfare Canada (DNHWC), operates the National Dosimetry Service

* Consultant, 665 Chipmunk Street, Pickering, Ontario, Canada L1 W 2W2; ‘Health and Safety Division, Ontario Hydro, 757 McKay Road, Pickering, Ontario, Canada L1 W 3C8. (Manuscript received 21 August 199I; revised manuscript received6 May 1992, accepted 18 May 1992) 00 17-9078/92/$3.00/0 Copyright 0 1992 Health Physics Society 702

Thermoluminescent dosimetry and power station fields 0 R. A. FACEY ET AL.

703

Table 1. Agencies recognized by the AECB as competent to perform external dosimetry in Canada. Name Chalk River Nuclear Laboratories” Bureau of Radiation and Medical Devices New Brunswick Electric Power Commission

H ydro-Quebec Ontario Hydro a

Affiliation Atomic Energy of Canada Ltd? Department of National Health and Welfare Canada Government of New Brunswick Government of Quebec Government of Ontario

Abbreviation used in text CRNL DNHWC NBEPC HQ OH

Now called “Chalk River Laboratories.” Now called “AECL Research Ltd.”

which provides external dosimetry to all other radiation workers in Canada (Grogan et al. 1984, 1990; Ashmore and Davies 1989; Sont and Ashmore 1984, 1988).

EXPERIMENTAL METHODS There were 36 runs in this intercomparison. A “run” consisted of one exposure at one location, using badges mounted on a large phantom. The duration of the run was long enough to give a tissue dose of 1- 10 mGy (0.1-1 rad) to the dosimeters. Exposure times ranged from 1-48 h. Fifteen TLD badges (three from each participant) were normally exposed per run. Design of the large phantom An overriding constraint of phantom design was our need to ensure that all the badges in one run would “see” the same incident field. Some fields would be highly divergent, due to restricted size, from a physically small source. In such cases, large changes of field could occur on a distance scale of centimeters. Thus, the only practical solution was to rotate the badges sequentially through the same orbit in space. This led to a square phantom design with a circular, rotating front surface to which the badges were attached by Velcro strips. Drawings, with dimensions, are shown in Fig. 1. The size was dictated by the need to accommodate 15 badges, reasonably spaced, and not too close to the edge. The actual distance from the TLD chips to the rotor edge was about 70 mm. An edge effect, if present, would be about the same for all badges and for the petty phantom detectors. The rotating disk, ball-bearing-mounted on a horizontal axle, had an air gap clearance of 2-4 mm from the stationary block. This was considered small enough, and deep enough into the phantom (33 mm), to be unimportant. The rotor was driven at 6 rpm by a motor mounted at the rear. The phantom material was “pressed wood,” which has a long history of use in radiology as an inexpensive soft-tissue simulator. The block was built up from 14 sheets laminated together, and the disk from three sheets. The density was 0.9 g cmT3,and the total mass was 45 kg, each. A photograph is shown in Fig. 2. The X-shaped

structure at the front is made of semirigid plastic tubing. This was necessary since all equipment required bagging in plastic before being carried into contaminated environments. The plastic tubes, bent into the shape of a bow, prevented the bagging from snarling on the rotating badges. Design of the petty phantom The radiation fields were characterized by measuring their depth-dose distributions in the front of the phantom. For this purpose, we designed and built a petty phantom (where “petty” is used in the sense of minor or subordinate). The petty phantom consisted of a rectangular stack of nine slabs of polystyrene, each one 50 X 50 mm2in area and 3 mm thick. Four shaliow cylindrical recesses, just sufficiently large to hold one TLD chip of dimensions 3.2 x 3.2 x 0.9 mm3, were machined into the front surface of each polystyrene slab. The recesses formed a spiral pattern, to ensure that none of them lay directly behind any other. An exploded drawing of a petty phantom is shown in Fig. 3. Each of the 36 recesses was loaded with a chip of 7LiF.SThe front surface of the first slab was covered with 14 mg cm-2 of aluminized mylar tape. The stack was assembled and held together with four nylon bolts at the corners. A photograph of a partially-assembled petty phantom is shown in Fig. 4. After assembly, the petty phantom was press-fitted into a square hole in the rotating disk of the large phantom, with the thin mylar cover flush with the front surface of the disk. In Fig, 2, the petty phantom is just to the left of the 12 o’clock position on the disk, at the same radius as the dosimeter badges. Nine petty phantoms were made. This allowed preloading of the 324 TLD chips necessary for 1 wk of nine runs. The stacks remained bolted up until readout time, around 1 wk after exposure. This preserved the chip identity (in sets of four for each depth) in all petty phantoms. After readout and annealing, the exposed chips rejoined a pool of several hundred chips. This pool had been selected from a larger population of chips TLD-700, Solon Technologies Inc. (formerly Harshaw/Filtrol Partnership), 6801 Cochran St., Solon, OH 44139.

704

Health Physics

December 1992, Volume 63, Number 6

PHANTOM

ELEVATlON

-

50.5 cm

c

Square Hole (5x 5 cm) in Rotor for

Petty Phantom

50.5 cm

I I

!b

0.3cm

Clearance J

I

’

between Stator & Rotor 02.0.4 cm

Fig. 1. Drawings of the large phantom. The “petty phantom” was press-fitted into the square hole in the rotating disk, flush with the surface. Badges were attached in a circle, at the same radius as the square hole. Exposures were equalized by rotation.

and was homogeneous to better than d 5 % [for 1 standard deviation (S.D.)]. The petty phantoms provided 36 readings for each run, in sets of four at each of the nine depths. The mean depths of each level (taken to be the middle thickness of each chip) were in the range 134-2,830 mg cmV2in steps of 337 mg cm-2. Description of dosimeters Four of the five dosimetry processors use similar TL dosimeters and readers; these are CRNL, DNHWC, HQ, and OH. Their dosimetry systems are variants of the automatic-reader system originally designed and developed at CRNL (Jones 1971). The dosimeter consists of an aluminum plaque that cames two LiF TLD chips of different thicknesses, mounted on thin Kaptong tape. The reader heats the TLDs using the hot anvil

Fig. 2. Photograph of large phantom with attached badges. The two attached quartz-fiber pocket dosimeters (near the center) allowed checking of progress during the run. X-shaped plastic tubes are explained in text.

technique and identifies the plaque by reading either a hole code or a bar code. A plastic badge case holds the plaque so that the whole-body chip, 0.9 mm thick (240 mg cm-2), is sandwiched between two aluminum filters that are 2 mm thick (540 mg cm-2). Thus, the detector is at an average depth of 660 mg cm-* beneath the surface. The skin chip is 0.4 mm thick (100 mg cm-2) and its front surface lies under 7 mg cm-’ of aluminized mylar tape. The main difference among the four processors using the CRNL-type system is in the choice of phosphor. DNHWC uses lithium fluoride with the natural isotopic mix of lithium, TLD-100, for both chips. OH and HQ use lithium fluoride made from lithium enriched in 7Li, TLD-700, for both chips, to reduce their sensitivity to thermal neutrons. CRNL uses TLD- 100 for the thick chip and TLD-700 for the thin chip, and both chips are sensitized by the Mayhugh, Fullerton, Jones method (Mayhugh and Fullerton 1975; Jones 1980; Jones and Richter 1982). The fifth processor, NBEPC, uses a commercial dosimetry system by Panasonic.” In this system, each badge contains a plastic plaque carrying four “elements” of thin, granular lithium borate powder (copperactivated) on a plastic substrate. The type of plaque used” carries three elements of lithium borate made with the odd-numbered isotopes ’Li and “B (and therefore insensitive to thermal neutrons). Two of these elements measure deep dose by being filtered behind 1 g cm-2 of plastic. The third is a shallow element with a thin window that results in a mean depth of about 22 ‘I

Trademark of E. I. du Pont de Nemours Inc., Wilmington, DE 19898.

Panasonic Inc., Two Panasonic Way (7E-4), Secaucus, NJ

07094.

’ Model No. UD-8 14AS9.

Thermoluminescent dosimetry and power station fields 0 R. A. FACEYET AL.

705

PE?Ty PHANTOM

Aluminized

DISASSEMBLED VIEW

Fig. 4. View of partially disassembled “petty phantom.” ASSEMBLED VIEW

stripe

&

Nylon bolt for compression

Fig. 3. Exploded view to give details of “petty phantom” design. Nine petty phantoms, with 36 chips in each, allowed preloading of chips for one week’s runs.

mg cm-2 to the phosphor center. The fourth element has similar TLD properties, though it differs isotopically. The lithium borate granules are made from the even-numbered isotopes 6Li and log, making the element very sensitive to thermal neutrons. In the badge, this element lies behind 300 mg cm-* filtration of tissueequivalent plastic. It is used qualitatively to indicate neutron exposures and to serve as an element for emergency evaluation, if an overexposure is suspected. The badges are read using an automatic reader, the Panasonic Model UD-7 IOA. A photograph of badges and plaques, representative of the CRNL designs, and the NBEPC badge and plaque are shown in Fig. 5.

Care and management of badges Thirty-six runs took place during four “active weeks” at the power stations, at the rate of nine runs wk-I. Between active weeks, there was a spare week to

Fig. 5. Close-up of badge cases and plaques for four of the processors shown in the open position. The CRNL badge was similar to those of OH and HQ, on the left. The third badge from left is the DNHWC, shorter than OH and HQ because it lacks criticality detectors (gold and sulfur). The badge on the right is the NBEPC Panasonic design, with the four detectors displayed.

allow participants to read out and report their badge readings, as well as allowing OH to read out, manually, the 324 loose TLD chips from the petty phantoms. Also during this week, freshly loaded badges (27 badges for nine runs, plus a few extra badges to measure dose acquired during transportation) were received by express mail from the four outside participants for use during the next active week. The exposed badges were mailed back to the participants at the end of each active

706

Health Physics

week, for readout. A summary of the organization of the intercomparison is given in Table 2. This timetable was adhered to quite successfully. The only failure occurred in the first active week when the CRNL badges were delayed inadvertently, missing the first six runs. All badges, except those being currently exposed on the three large phantoms, were kept at all times at a secure site, in a low-radiation-background area of the nuclear stations. Of the 540 badges received, only one badge showed any contamination, and this was readily removed. Data handling Each processor read its own badges, using routine procedures. The net badge readings, after subtraction of the transportation dose readings, were reported back to OH fortnightly. For the processors doing the readouts, this testing was essentially "blind." There was no knowledge at readout time of the results of any other processors, or even of the numbering codes that related each participant's badge and plaque numbers to the phantom number and number of the run on which they were exposed. At OH, the readings were manually entered into a personal computer for analysis. Stations visited The first station visited was Bruce Nuclear Generating Station-A (BNGS-A). This is located on the eastern shore of Lake Huron at Tiverton, Ontario. It is a four-unit station, with four Candu reactors of 79 1 MW, (gross) each. Two active weeks were spent there. The last two active weeks were spent at Pickering Nuclear Generating Station (PNGS). This is located just east of Toronto at Pickering, Ontario, on the north shore of Lake Ontario. There are eight Candu reactors of 540 MW, (gross) each, the whole station being housed within a single structure. Most of the reactors visited were under full power at the time, except for two in planned shutdown. Types of fields encountered The radiation sources for phantom exposures ranged in size from a short length of iron pipe, 5 cm in diameter, to reactor end shields, 8.5 m in diameter. About 60% of the fields were essentially pure gamma rays, with dose rates of 100-30,000 pSv h-l (10-3,000 mrem h-I). The rest were mixed fields, with beta dose rates up to 100 mGy h-' (10 rad h-I). Previous neutron surveys had shown that there were no significant neutron fields in most of the areas where the measurements were made. In one case where neutrons were expected to be present, a survey was done with an Anderson-Braun-type remmeter and a BF3 thermal neutron counter. The results are given in Table 3 and show that, even in this case, the contribution of neutrons to the dose rate was insignificant.

December 1992, Volume 63, Number 6

Table 2. Organization of the Canadian Dosimetry Intercom-

parison (CDI).

a

Unit

Runs

Badges per uarticivant

Petty vhantoms

Total badges

1 run 1 week TotalCDI

1 9 36

3 27 108

1 9 36

15 135 522"

Actual badges irradiated. Badge irradiations planned 540.

Table 3. Dose equivalent(DE)rates for two classes of neutrons at a typical location, with detectable neutron fields. The neutron DE rate is of the order of 5% of the total DE rate. Component of dose eauivalent

+

Neutron (fast intermediate) Neutron (thermal to 0.55 eV) Gamma

DE rate mSv h-'

mrem h-'

Remmeter

0.24

24

BF3counter

0.012

OH badees

4.0

How measured

1.2 403

Petty phantom reference doses The readings of the petty phantoms, after averaging at each level, were used to calculate a reference dose by linear interpolation. The reference dose for each run is the best estimate of the dose at 1,000 mg cm-2. Conversion to absorbed dose The lithium-fluoride chips were calibrated to measure exposure in free air, by irradiating them with I3'Cs gamma rays under conditions of charged-particle equilibrium. In order to calculate absorbed dose in soft tissue, the lithium-fluoride chips in the petty phantoms were treated as cavities of lithium fluoride inside an extended medium of polystyrene. An exposure-to-dose conversion factor of 37.40 tissue Gy (C kg-')-' [0.965 rad (roentgen)-'] was calculated from cavity theory, over the energy range from 0.1-6 MeV, using the method of Burlin (1966, 1968) for dose partitioning in cavities of intermediate size. Details of the calculation are given by Facey and Hirning (1989). Although there were air gaps around the edges of each chip, the mass of air was very small relative to that of the chip, and the gaps would therefore not significantly perturb the radiation field. A more exact calculation of the conversion factor could be done using Monte Carlo techniques, but the resources to do such a calculation were not available for this study. RESULTS AND DISCUSSION Petty phantom results The petty phantoms were used to characterize the various fields by penetrability into tissue equivalent material. Some typical results are shown in Table 4. The measured depth doses are shown relative to the

Thermoluminescent dosimetry and power station fields 0 R. A. FACEY ET AL.

Table 4. Some typical petty phantom results. Penetrability in tissue was inversely quantified by the surface/depth ratio (surface dose/deepest dose). Surface/depth ratios ranged from 25.1 down to 1.06. Beta fields are prominent in runs 28 and 10. Run number: Source:

28

10

35

Ram seal filters

Fueling machine snout

Reactor feeder cabinet

Normalized depth doses:

Relative dose (dimensionless)

TLD mean deoths 47 1 808 1,145 1,482 1,819 2,156 2,493 2,830

Deepest dose equivalent: 2,830

Reference dose:

25.05 6.58 1.94 1.15 1.10 1.04 1.02 1.01 1.oo

8.45 2.41 1.28 1.15 1.08 1.04 1.03 1.oo 1.oo

Measured dose (mSv) 1.02

1.36

12.6

Interpolated dose (mSv) ~~

1.OOO me cm-*

1.21 1.19 1.16 1.10 1.11 1.08 1.07 1.04 1 .oo

1.52

~

1.64

14.2

dose at the deepest level. The reference dose at 1,000 mg cm-2 depth is also given. The second column of Table 5 gives R, the ratio of the shallowest dose reading to the deepest dose reading (numerically the same as shallowest relative dose). This ratio provides a single number for each run which is a measure of penetrability for shallow depths. In mixed beta-gamma fields, this number is a good direct indicator of the importance of the beta component. Reporting of badge readings Each participant in the intercomparison reported the readings of their three badges for each run. For the two-chip dosimeters, each reading was for a single 0.9mm TLD chip. For the four-element Panasonic dosimeters, the deep-dose reading reported was the average of the two deep-dose elements that were insensitive to thermal neutrons. The mean of the three reported readings was taken as the “official reported reading” for each agency. The relative S.D. for this averaging was usually about 5% or less, but occasionally up to 12%. The reading of the thermal-neutron-sensitive element in the Panasonic badges provided a sensitive indication of the presence of thermal neutrons: A ratio of the reading of this element to the average reading of the two deep-dose elements >1.2 was taken as an indication that thermal neutrons were present, and a flag was set for that dosimeter. All three Panasonic dosimeters were flagged in six runs, two out of three

707

Table 5. Shallow-to-deep-doseratio, R, and whole-body readings normalized to reference dose at 1,000 mg cm-’. There are no data in this table for runs 29 and 30. Run number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 31 32 33 34 35 36

Mean: S.D.:

R

CRNL

DNHWC

NBEPC

HQ

OH

1.21 1.20 1.20 1.20 6.06 1.13 1.15 1.18 1.06 8.45 7.62 10.23 1.13 1.15 1.19 1.19 1.26 1.17 1.19 1.19 1.28 1.21 1.34 1.29 1.20 1.16 1.11 25.05 1.30 1.32 1.35 1.19 1.21 1.11

-

0.93 0.96 0.97 1.53 1.53 1.62 0.94 0.95 0.96 0.93 0.95 0.92 0.89 0.87 0.96 0.86 0.95 0.97 0.93 0.93 0.88 2.90 1.03 1.02 0.96 0.89 0.98 0.88

1.42 1.34 1.31 1.29 1.82 1.39 1.26 1.32 1.31 2.06 2.18 2.26 1.49 1.52 1.47 1.50 1.54 1.52 1.09 1.10 1.24 1.67 1.17 0.73 1.15 1.17 1.05 2.95 1.12 1.13 1.09 0.98 1.09 0.99

0.85 0.88 0.83 0.87 1.00 0.95 0.85 0.85 0.87 1.58 1.53 1.80 0.99 1.06 1.11 1.01 1.07 0.94 0.82 0.82 0.90 1.25 0.98 0.49 0.91 0.91 0.83 1.36 1.07 1.35 1.17 1.04 1.05 1.15

1.04 1.11 0.97 1.05 1.28 1.05 1.05 1.03 1.02 1.53 1.58 1.74 1.09 1.12 1.12 1.11 1.13 1.1 1 1.02 1.01 1.07 1.03 1.06 1.07 1.04 1.10 1.01 2.34 1.10 1.13 1.08 1.03 1.12 1.05

0.93 0.90 0.87 0.94 1.16 0.98 0.95 1.00 0.98 1.25 1.32 1.40 0.92 0.96 0.97 0.94 0.98 0.94 0.86 0.85 0.94 0.90 0.96 0.94 0.91 0.94 0.88 2.21 0.95 0.98 0.96 0.91 0.98 0.92

1.07 0.41

1.40 0.43

1.03 0.26

1.16 0.27

1.01 0.24

-

-

were flagged in three runs, and one out of three in seven runs. Normalization of badge readings The reference dose, determined from the reading of the petty phantom, was used to normalize all runs. The results are shown in Table 5 . This table gives the mean results for each participant for 34 of the 36 runs. In the table, overall means and S.D.s are given for each processor at the bottom. A loss of identity of the petty phantom readout data occurred on one occasion. The readings for runs 29 and 30 became mixed and could no longer be identified with certainty. Therefore we are unable to report in Table 5 on normalization of these two runs by the petty phantom method. The same data, averaged across all participants, are displayed as a bar chart in Fig. 6. Most runs show average relative doses close to 1.0. However, there are five runs that stand out above the others; all of these runs also had high values of the ratio R. High average doses are therefore related to a large beta component

708

Health Physics

in the field. This is because the four processors with dosimeters at a mean depth of 660 mg cm-2 are measuring dose at a depth that is substantially shallower than the depth at which the reference dose is measured. From the sample petty-phantom depth-dose curves shown in Fig. 7, it is clear that, for runs where R is high, there is a high-energy beta component that contributes significantly to the dose at 660 mg cm-2. Thus,

c

2

i

2 ...............................................................................

.....................

......................

1

3 5 7 9 11 13 15 17 19 21 23 25 : 2 4 6 8 10 12 14-16 18 -20 2 2 2 4 26

31 33 35 32 34 36

Run Number

Fig. 6 . Bar chart display of overall average of all processor's results, normalized by the petty phantom reference dose. Five runs with exceptionally high results are due to large beta fields from the fueling machine snout at BNGS-A (#5, 10, 1 I , and 12). The fifth (run 28) is a ram seal filter from a Pickering fueling machine. All other runs show an average very close to 1.o.

December 1992, Volume 63, Number 6

these processors will over-report the deep dose in fields with a significant fluence of high-energy beta particles.

Overall performance The overall performance of the participants is shown in Table 5. The mean relative readings (averaged over all runs except runs 29 and 30) ranged from 1.O 1 to 1.40 (dimensionless). These results are reassuring in that no agency had a mean relative reading of