Title: CCMPARISON CONTINUOUS A - Digital Library/67531/metadc698846/m2/1/high... · Los Alamos NATIONAL LABORATORY CCMPARISON OF CONTINUOUS AIR ... Tom Voss, ESH-1 ... An upgrade

Title:

Author(s):

Submitted to:

Los Alamos N A T I O N A L L A B O R A T O R Y

CCMPARISON OF CONTINUOUS A I R MONITOR UTILIZATION: A CASE STUDY

J.C. Rodgers, ESH-4

J.T. VOSS, ESH-1 J.J. Whicker, ESH-4

DOE Contractors Air Monitoring User Group Annual Meeting - Miamisburg, Ohio April 28 - May 1, 1997

Los Alantos National Laboratocy. an afflrmative actionlequal oppochmity empk&er, k operated by the U M t y of California for the U.S. Department of Energy undec contrad W-74WEW-36. By acceptance of this atikle, the puMkher recognizes that lhe U.S. Oovernment rstalns a mxdusive. royalty-frw Gcense 10

repmduce the pub&hed form of this contdbutbn. or lo a l k ~ others to do bo, for U.S. &ovumunent purposes. The Los Alarms National Laboratory that lhe PubaSherideMify tMs article as work performed under the auspices ofthe U.S. Department of Energy. Fonn No. 836 RS

ST 2629 10/91

COMPARISON OF CONTINUOUS AIR MONITOR UTILIZATION: A CASE STUDY

John Rodgers, ESH-4 Jeff Whicker, ESH-4

Tom Voss, ESH-1

ABSTRACT:

The Chemical Metallurgy Research (CMR) building has been upgrading to different continuous air monitors (CAMs) over the past several years. During the transition, both the newer and older CAMs were positioned in the rooms for field testing and comparison. On December 19, 1996, an accidental release of plutonium aerosol occurred into a laboratory in the CMR building. The event occurred while the room was unoccupied, and no personnel were exposed fi-om this incident. There were two fixed air samplers (FASs) and three CAMs operating in the room at the time the release occurred, including two of the recently installed Canberra Alpha Sentry CAMs and one older Eberline CAM. The apparent cause of the release was a procedure carried out in the basement involving the replacement of the HEPA filter in the ventilation exhaust of a slot-box in the laboratory. For a short period, the ventilation was disconnected from the slot-box in this room, but not fi-om the chemical hood exhaust on the opposite side of the laboratory. Therefore, a condition was created where backflow could occur out of the slot-box and into the room. Eventually all three CAMs in the room alarmed, and the situation was successklly monitored and brought under control by health physics personnel. Data on CAM performance were logged, and Pu activity collected on CAM and FAS filters were measured. A comparison of the new and old continuous air monitoring programs was performed and many interesting lessons on CAM performance and CAM utilization were learned. Overall, this comparison showed the advantages of remote monitoring, timely spectral information, and concentration measurements resolved in time and space.

INTRODUCTION

Continuous air monitors (CAMs) are positioned in plutonium laboratories to provide real-time monitoring of radioactive aerosol levels. When the airborne radioactivity levels exceed a preset threshold, the CAM alarms warning workers of the potential for a release. Continuous air monitors have been used at the Chemical Metallurgy Research (CMR) building for many years. Personnel at the CMR building primarily perform chemistry on various actinides, including 23?Pu.

Historical perspective

An upgrade of the continuous air monitor program has been ongoing at the CMR building over the past several years. This upgrade included the change from older Eberline Continuous Air Monitors (CAMs) to newer Canberrra CAMs. Field testing of the newer CAMs was done during the transition. In many cases, the newer CAMs were placed next to the older CAMs. This arrangement allowed side-by-side operational comparisons between the two CAMs under field conditions.

- The upgrade to the newer CAMs and the associated data processing equipment occurred for many reasons including manufacture claims of 1) sensitive detection of alpha-emitting aerosols, 2) time-resolved information, 3) remote monitoring of radioactive aerosol from a safe location, and 4) alpha energy spectral information.

The placement of FAS and CAM monitoring equipment in wings of the CMR building was recently evaluated using NUREG-1400 (Hickey et al. 1993). In general, CAMs were placed in rooms near one or more of the open-front hoods. This reflects the presently used principle that the best location for CAMs are at exhaust points within the room [i.e., open-fkont hoods (slot- boxes) ]. Such a principle has commonly been applied in other laboratory rooms at Los Alamos (whlcker et al. 1997), laboratories in the United Kingdom (Brunskill and Hermiston 1966), and elsewhere in the DOE Complex (Alvarez et al. 1994, Mishima et al. 1988). In addition, CAMs were placed above the slot-boxes because the majority of releases in the CMR building occur at the slot-boxes. Finally, room operations were evaluated to place CAMs near boxes where more hazardous operations are conducted or larger amounts of radioactive material are used.

The equipment, CAM, and fixed air sampler (FAS) layout for Room 3 1 1 1, the room where the release occurred, is shown in Figure 1. The ventilation air duct entry is at the ceiling at the far south end of the room. There were three exhaust points in the room: one was through the open- fkont chemical hood at the far end of the room near the entrance, a second was the floor register under the chemical hood, and the third was through the open-front hood near the middle of the room. A smoke test study in the room indicated that under normal ventilation conditions the tendency was for smoke streamers to move upward, indicated by the letter U, near the glove boxes and open-fiont hood on the west wall, and down, indicated by the letter D, near the chemical hood and the counter top along the east wall. Somewhat paradoxically, there was an upward tendency found in the middle of the room near the inlet end that suggests a vortex motion in the incoming air stream. Upward flow patterns along the glovebox line and open-face hood lends credence to the placement of CAMs on the top of these boxes. Horizontal movement was recorded with arrows showing the general direction of the smoke.

Each of the two types of CAMs in the room have different alarm strategies and background subtraction logic. The Canberra CAM uses both an acute release alarm algorithm and a chronic release alarm algorithm. The acute release algorithm is a calculation based on the net number of counts in the region of interest and is compared to the background every 30 s. The chronic release algorithm is a calculation based on the subtraction of an exponential fit to the tail of the background spectrum extending into the region of interest, using user defined count time intervals (options range from 5 minutes to 99 minutes). The Eberline CAM used in the CMR building subtracts a preset fraction of the background counts from the number of counts in the region of interest. Ifthe number of net counts in this region exceeds 40 counts per minute, an alarm is sounded.

3pen-front 2hemical 4ood

a

Counter

I I

I I I I I I I I I I

----,,-J

I

Entrance

2 D

FAS 31 11-A

D

D

D

Eberline CAM U

D Canberra CAM 3A1

FAS 31 11-8

U

U U

GBI

GB2

I 3pen-front

iood D

GB3

GB4

Canberra CAM 381

VENT. SUJPLY

t

Figure 1. Layout of room 31 11 at the time of the release including the gloveboxes (GBI-GBS), ventilation supply, open-faced hood, chemical hood, fixed air samplers (FASs), and CAMS. This room is approximately 12 feet by 24 feet by 10 high. Also shown are the results of a air flow study using smoke tubes. The letter “U” signifies upward movement of air, the letter “D” signifies downward movement of air, and the arrows indicate direction of air in the horizontal plane.

- Ventilation and air sampling/monitoring systems

The equipment and air sampling/monitoring layout at the time of the release is also shown in Figure 1. The measured face velocity in the open-front hood ports was 1 15 feet per minute (fpm) implying a volumetric rate of 75 cubic feet per minute (cfin). The flow into the slot created by a partially closed sash in the chemical hood (fixed and locked into place) was 880 fpm. The total flow into that hood and its associated bypass floor register was about 1225 cfm. The open-front hood and the two gloveboxes to the north of the open-fiont hood on the west wall are all connected to a common exhaust duct in the basement that is HEPA filtered before discharge into the facility stacks in that wing where the exhaust air again passes through another HEPA filter. The chemical hood has a separate exhaust duct and discharge path. Clearly then most exhaust flow was through the chemical hood. The volumetric air exchange rate for the room was about 30 changes per hour, which is a relatively high exchange rate compared with many at Los Alamos plutonium facilities. Two FASs were in use at the time of the release: one positioned over the sash of the chemical hood, and a second over the open-front hood on the west wall. There were three CAMs: a new Canberra CAM (CAM 3B 1) over the south glovebox maintained at reduced pressure and inerted with argon, and the two others side-by-side over the open-front hood in the middle of the west wall. One of the latter was a Canberra CAM (CAM 3A1) and the other is one of the older Eberline models. Their positioning relative to openings in this hood is roughly illustrated in Figure 2 ( the photo is from another room). Note in Figure 2 that flow into the open-front hood is restricted to two ports cut in a plexiglass plate covering the slot (shown by dotted line). The purpose of the plate was for better contamination control as restriction increases the entry velocity into the hood.

Release event

The events leading up to the release began on December 19 at 9: 15 A.M. with the closing of a damper on the HEPA filter servicing the glove box line in preparation for replacement. This filter apparently had been in service for more than 20 years. A containment tent had been built around the HEPA location to contain any spill that might occur when the filter was removed. A flexible duct section had to be removed on each end of the HEPA to allow replacement. At about 9:45, plates were to be inserted into the ductwork to cap off the duct once the old filter had been removed. Up to this point in time, the CAM filter activity log showed no release. The HEPA filter was then replaced. Between 9:45 and 10:00, during the replacement process, possibly while the exhaust duct was open to the basement air before cap-off could occur, a back-flow of air up the ductwork into the gloveboxes and open-front hood occurred. Air was apparently being pulled out of the open-fi-ont hood by the continued action of the very large exhaust into the chemical hood.

RESULTS

The resultant discharge may have been very brief, but the episode of elevated air concentration possibly lasted 50 minutes, including the ramp-up and clearance phases, based on the count and alarm logs of the Canberra CAMs. The chronology of events is described in Figure 3. Both the Canberra and Eberline CAMs alarmed during the release. The Canberra CAM placed over the

. . .

Figure 2. Older CAM (left) and newer CAM (right) above an open-front box. Daskied line represents plexiglass plate with two ports for access.

-Figure 3. Chronology of events around the Pu aerosol release in room 31 11 and the integrated DAC-hr concentration over time.

1,000

300

100

30

DAC 10

hrs 3

1

FAS31118Filter 1o:wAM 10:35AM chan@at13:DSPM A

CAM3A1 /

9:57AM

9 4

open-front hood (3Al) responded with an acute alarm at 9:49 at an exposure of only 5.5 DAC-hr, and then again with a chronic alarm 15 minutes later (10:04) at a peak exposure of 453 DAC-hr. A Radiological Control Technician (RCT) responded to the CAM alarms at 1O:OO. Following standard procedures, he donned Personal Protective Equipment (PPE) and entered the room where he investigated the situation and went to change the filters on both the Eberline and Canberra CAMS for fbrther analysis. A second ingrowth phase was initiated on these filters during the clearance process. Meanwhile, the second Canberra CAM (3B1) also began detecting a slow rise of plutonium activity after a brief lag. By 957 the exposure at the south end of the room had reached 6.45 DAC-hrs and by 10: 13 it was 19.55 DAC-hrs. Since the filter was not changed by the RCT during the entire episode, it was possible to ascertain the point in time when activity ceased to accumulate on the filter, signaling the end of the clearance phase and the end of the episode. This was approximately at 10: 15 with a cumulative exposure of 20 DAC-hrs. At about 1:03 P.M. the RCT again changed the CAM filters and the FAS filters. Filters were sent to the count room for gross-alpha and beta counting. Using the fact that the count room data for the CAM 3B 1 filter was 276 DPM while the instrument response was 81 cpm, an implied efficiency of 0.293 can be estimated. This in turn suggests that the CAM 3A1 filter must have had more than 6700 DPM of activity on it when it was removed (based on a count rate of 1974 cpm), which is about three times the activity on the Eberline CAM filter removed at about the same time (2 100 DPM). CAM. During the recovery cleanup later, it was found that there was floor contamination of approximately 2000 to 5000 DPW100 cm2 at the base of the slot-box; there was about 600 DPW100 cm2 on the top of the glove boxes; and about 300 dpd100 em2 on the counter top along the east wall.

This is consistent with a higher sampling rate and more efficient inlet in the Canberra

The Canberra alpha CAM filter count rate data are plotted in Figure 4 against time fiom the onset of the release. There are two plots for CAM A shown because the RCT changed the filters during the incident (CAM-A( 1st) before and CAM-A(2nd) after the filter change). The plot of CAM- A(1st) was extrapolated beyond the filter exchange by adding the successive filter counts from CAM-A(2nd) to the last count from CAM-A(1st). The plateau of count rates indicated that this was a short-term release event. To fbrther explore the character of the release, these count data were cast in a form that indicated both the incremental increasing and decreasing responses, and then a %Jellmixed' box model was used to generate an analytical model of the event as seen at CAM A and CAM B. The model was of the form:

for time (t) less than time at C, or,

Figure 4. Alpha CAM Responses Room 31 11

1000

U CAM-A(lst)

-0- CAM-A(2nd)

4 3 CAM-B

0 20 40 60 80 100120140160180200

Time (min)

. for t greater than time at C,, . Where C,, is the peak concentration; m, and m2 are the d n g rates in the ingrowth and decline phases, respectively; Q is the air exchange rate; V is the effective miXing volume; and tb and t, are the times before and after the maximum concentration was reached, respectively. A nonlinear least square fit of this model was made to the data with the fit shown in Figure 5, with the prediction that mixing was near unity (well-mixed), the effective air exchange rate was 105 cfin, and the dilution volume was 1030 cubic feet. This suggests that the CAM responses were consistent with exposure to well-mixed contaminated air, and a state of rapid air exchange.

DISCUSSION

Many lessons were learned by detailed analysis of the release, the health physics monitoring measurements, and the subsequent response of the CAMs. Advantages of the newer continuous air monitoring system include time-resolved information, alpha energy spectral information, remote supervision of the air concentrations in the affected room.

Advantages of time-resolved information

Identifjhg the cause of a release is a first step to preventing fbture releases, and the time resolved output fiom the Canberra CAMs allowed quick determination of the time of the release. This information was helpful in correlating the release to the work on the HEPA filters. We further analyzed the hypothesis that these particles could have been swept off the glovebox surfaces during the back-flow. We estimated that a threshold friction velocity of about 30 c d s (60 $m) would be required to initiate particle movement off the surfaces of the box @ombrowski et al. 1993, Bagnold 1941). Although the conditions in the open-front hood were not measured at the time, it is possible that the friction velocities in portions of the box were great enough to resuspend particles off the surfaces. Given the evidence available, a working hypothesis is that there were some relatively recently deposited plutonium particles in the open-front hood, possibly a dusting of interior surfaces or deposits near the exhaust port of the hood near the back. A fraction of these were resuspended when reverse flow began, entrained in high speed turbulent eddies that were created by air being drawn out of the hood by the main room exhaust. Mechanical disturbances such as vibrations in ductwork (i.e., from pulses in ventilation air) could also have resuspend particles from the surfaces.

Having information on CAM responses resolved in time and space was also useful for retrospectively evaluating the CAM placement strategy. Knowing the approximate time of releases, the time of the CAM alarms, and their sensitivity of the CAM as compared with measured concentrations by the fixed air samplers at other locations in the room a continuing evaluation of CAM placement is possible. Here, the strategy of placing the CAMs on top of the slot-box based on potential for a release was reasonable. Although the information for this release was qualitative because only an approximate time of the release was known, the alarm occurred within minutes after the aerosol was released into the room, and the CAMs on top of the hood responded with enough sensitivity to alarm. Alternatively, CAM 3B 1 did not respond quickly and

3 a

Figure 5. Results from mixin model ana lys i s . Room 31 11 dixing Model

2200 2000 1800 1600 1400 1200 1000 800 600 400 200 0

100

m 80 a E! 9 0 60 a L

% 40 a

20

0

0 20 40 60

r o Puff response CAM A I

0 20 40 Time (min)

- effectively to the release and therefore was not positioned well for a release out of the slot-box. In fact, C A M 3B 1 was placed in this location because of the amount of material in that glovebox and not for protection for employees working in the slot-box. The results from this release support current operational procedures that state that work in the open-front hood cannot be performed if CAM 3A1 is not operational.

Although important information was recorded by the Canberra CAMs, only qualitative evaluation of the adequacy of the placement of the CAMs for this release was permitted because of the limited number of concentration measurements recorded over time (every 15 minutes) and space (essentially 3 sampling points in the room including FASs). Quantitative testing of aerosol dispersion using tracer aerosol releases and multiple optical counters could provide valuable information on the effectiveness of the current placement of the CAMs and FASs. This incident illustrates the need for more timely measurement recording because the time scale for the aerosol dispersion is much shorter than 15 minutes.

It is instructive to try to understand the response of CAM 3B1 to the release because this CAM superficially seemed to be upwind from the release location. One possible explanation for the alarming of CAM 3B 1 later is that turbulent eddies carried materials upward from the openings in the slot-box and then thoroughly mixed the contamination into the ventilation air. Such turbulence not only could have brought material upward toward the top of the glovebox line, it also transported plutonium particulate in a counter-current fashion back toward the ventilation inlet at the south end of the room where CAM 3B 1 was situated. While the airborne activity was an order-of-magnitude smaller than at the slot-box, the fact that such a flow existed indicates the existence of a large scale vortex structure connecting the far ends of the room. The diluted aerosol was not concentrated enough for CAM 3B 1 to activate an acute alarm, and the first alarm was a chronic alarm (based on a 15 minute count cycle) at 10: 13.

Finally, time-resolved information allowed for modeling concentrations over time. Although the resolution of the time-dependent concentrations is coarse (1 5 minutes), this information provides qualitative estimates on the degree of mixing that occurs in a room. This type of information can be valuable in evaluating placement strategies. For example, placement of CAMs in rooms with high miXing values may not have to be as precise as in room where the mixing values are low.

Advantages of having energy spectra information

Spectral information was he1pfi.d in identifling the source of the radioactive material. The Canberra CAM system stores the alpha spectrum of the last alarm cycle. The spectrum measured in the release by CAM 3B 1 is shown in Figure 6. The Pu-239 is prominently seen in the tail region of the 7.68 MeV peak of PO-214. The background spectrum is remarkably similar to an outdoor radon progeny spectrum in that there is very little evidence of the 6.0 MeV spectrum of Po-2 18, the first short-lived progeny. This is consistent with there being no filtration of incoming air. Another striking feature is the virtual absence of a 5.5 MeV peak from Am-241. This suggests that the plutonium involved in this release was not well-aged, and therefore would not likely have come from old materials dislodged from the HEPA filter during the exchange. This is

Figure 6 .

C

c 0 EI 60 % 40 Q to -E 3 20 s

CMR Room 3111CAM Spectra (After filter change)

. . . . .. . . . .. . . . __. . . ___. _ _ _ ___. . . . ___. . __. . . _. .. .... . . . ... .. .. . . . . . . ... . .. .. . ..

2 4 I O Energy (MeV)

consistent with the reported absence of any contamination in the tent enclosure in the basement when the old HEPA filter was removed, implying that perhaps there was little or no loose materials dislodged during the HEPA removal process. The breadth of the Pu-239 peak indicated that the particles collected may have been fairly large with diameters 3 - 10 micrometers. This hypothesis was supported by the presence of detectable radioactive deposits on horizontal surfaces of the room, especially at the base of the slot-box. These deposits are likely caused by gravitational settling of large particles (Sehmel 1971). For comparison, a 10 ,um Pu particle with a density of 10 g/cm3 will have a terminal settling velocity of about 3 c d s ; whereas, a lpm settles at a rate of only 0.03 c d s @rids 1982).

Also, spectral information often provides a quick answer to one of the first questions asked after a CAM alarm: was the alarm caused by Pu aerosol (a true alarm) or from Rn progeny (a false alarm)? Historically, this question was answered by having a RCT enter the room to retrieve the CAM filters for nuclear counting and, if necessary, do successive counts over time to quanti@ the decay of the short-lived Rn progeny. In this release, the Canberra CAMS provided information required to answer the question if the release was true and what the current aerosol concentrations were, and this information was accessed remotely. Figure 6 clearly shows a well- defined Pu peak, and so the release could have quickly been classified as a “true” release without entry into the room.

Advantages of remote monitoring information

Based on the potential for increased risk and clean up costs, often waiting outside the room is prudent to observe the air concentrations until they decrease. Although protected with respirators and personal protective clothing, responding RCTs entered rooms with elevated air concentrations at some risk and expense. The level of risk posed to the responding RCTs is difficult to quanti@, but a range of values can be estimated. On the low end of this range, the incremental CEDE to responding RCTs can be zero. At LANL, this risk is most often very small based on nose swipe results. The upper bound of the range of potential CEDE can also be calculated. For this calculation, this CEDE can be estimated as the product of the measured average air concentration, a respiratory protection factor (i.e., NRC 1991 allows a respiratory protection factor of 50 for a full-face respirator), the exposure time, and a dilution factor which accounts for the observation that breathing zone concentrations are often higher than area air samplers (Brunskill and Holt 1967, Scripsick et al. 1979). In this particular release, the results of the RCT’s nose swipe showed no detectable contamination.

The cost of recovery using the remotely obtained data can also be less. Without remote monitoring, retrieval of the CAM filters was necessary to determine if the alarm was real. Responding to CAM alarms under the older protocol while the air concentrations are elevated can increase the financial cost to recover the room from the person-hrs for the retrieval and analysis of the filter and from the cost of discarding personnel protective clothing if they become contaminated from aerosol attaching to the clothing from gravitational settling or turbulent impaction. Also, the technician may inadvertently spread the contamination to previously uncontaminated locations as they move about in the room to retrieve filters.

- CONCLUSIONS

The CMR building has been transitioning to the Canberra continuous air monitoring system over the past several years. During the transition, the preexisting continuous air monitoring system continued to operate allowing for field comparisons between the old and new systems. The release of Pu aerosol fi-om a slot-box and the relative responses of the two systems to the release are described. Overall, this comparison showed that the transition to the newer CAMs has many advantages such as remote access to concentration and spectral information from the CAMs, which may reduce cost and worker risk. Although there is a need for shorter intervals between recorded concentration measurements, the time-resolved information obtained was helpfil in determining characteristics of aerosol dispersion in the work space and for planning recovery of the room. Finally, the analysis suggest that the position of the CAMs and the current radiological procedures for room operations are good.

I REFERENCES

Alverez, J.L.; Bennett, W.S.; Davidson, T.L. Design of an airborne plutonium survey program for personnel protection. Health Phys. 66:634-642; 1994.

Bagnold, R.A. The physics of wind-blown sands and desert dunes. Methuen, London, 1941.

Brunskill, R.T.; Hermiston, S.T. The detection and measurement of plutonium airborne contamination in major plutonium facilities. In: Proceedings of the first international congress of radiation protection. Rome, Italy; 1966.

Brunskill, R.T.; Holt, F.B. Aerosol studies in plutonium and uranium plants at the Windscale and Springfield Works of the United Kingdom Atomic Energy Authority. In: Proceedings of the symposium assessment of airborne radioactivity; Vienna: IAEA; Report No. SM-95/3 0:463 -476; 1967.

Department of Energy. Radiation protection for occupational workers. DOE Order 5480.11; 1989.

Dombrowski, E.A.; Foumeny, D.B.; Ingham, D.B.; Qi, Y.D. Air entrainment of particles from a flat plate. Atmos. Env., Vol. 27A(15):2449-245 1, 1993.

Hickey, E.E.; Stoetzel, G.A.; Strom, D.J.; Cicotte, G.R.; Wiblin, C.M.; McGuire, S.A. Air sampling in the workplace. U. S. Nuclear Regulatory Commission document NUREG-1400.

Hinds, W.C. Aerosol technology: Properties, behavior, and measurement of airborne particles. New York: John Wiley and Sons; 1982.

Mishima, J.; Hunt, J.; Kittinger, W.D.; Langer, G.; Ratchford, D.; fitter, P.D.; Rowan, D; Stafford, R.G. Health physics manual for the prompt detection of airborne plutonium in the workplace. PNL-6612, UC-607; 1988.

Nuclear Regulatory Commission. Standards for protection against radiation. Federal Regulation 10cFR20; 1991.

Sehmel, G.A. Particle difhsivities and deposition velocities over a horizontal smooth surface. Jour. of Colloid and Interface Sci. 37(4):891-906.

Scripsick, R.C.; Stafford, R.G.; Beckman, R.J.; Tillery, M.I.; Romero, P.O. Evaluation of a radioactive aerosol surveillance system. In: Advances in radiation protection monitoring, proceedings of an international symposium on advances in radiation protection monitoring conference held by the International Atomic Energy Agency. Stockholm: IAEA-SM-229/62; 1979.

I Whicker, J.J.; Rodgers, J.C.; Fairchild, C.I.; Scripsick, R.C.; Lopez, R.C. Evaluation of continuous air monitor placement in a plutonium facility. Health Phys. 75(5):734-744; 1997.