United States
Environmental Protection
ency
Office of Radiation Programs
Radiation
National Air and EPA402-K-92-010
Radiation Environmental Laboratory February 1993
1504 Avenue A
Montgomery. AL 36115-2601
A Summary of EPA Radon
Chamber Tests and Results
for Rounds 3 and 4 of the
National Radon Measurement
Proficiency Program
Ftocyctod/Recyclabte
Primed on paper that contains
at least 50% racydad fiber
-------
A Summary of EPA Radon Chamber Tests and Results
for Rounds 3 and 4 of the National
Radon Measurement Proficiency Program
J. Michael Smith and Edwin L. Sensintaffar
U.S. Environmental Protection Agency
Office of Radiation and Indoor Air
National Air and Radiation Environmental Laboratory
Montgomery, Alabama
February 1993
-------
Contents
Page
Glossary iii
1.0 Introduction 1
2.0 Background Information 2
3.0 EPA Radon/Radon Decay Product Exposure Chambers 3
4.0 Radon and Radon Decay Products Monitoring 9
5.0 Summary of RMP Round 3 and Round 4 Results 15
6.0 EPA Chamber A and Charcoal Canister Testing 18
7.0 Discussion of Possible Reasons for Round 4 High
Bias of Charcoal Adsorbers 33
8 . 0 Summary . . .- 39
References 41
Appendix A - Histograms for RMP4 Performance Round A-l
Appendix B - Notifications of Velocity Sensitivity
of Open-Faced Charcoal Adsorbers B-l
-------
List of Tables
Table Page
1 NAREL Performance in Grab Radon Intercomparisons .... 11
2 NAREL Performance in Grab Working-Level
Intercomparisons 12
3 Methods Completed, Failure Rates and Method
Performance Ratios for RMP Round 3 and
RMP Round 4 16
4 Summary of NAREL Chamber A Radon Tests 19
5 Test 5 Transient Radon Concentrations Measured by
Three Continuous Radon Monitors 30
List of Figures
Figure Page
1 Schematic Diagram of NAREL Radon/RDP Chamber C 6
2 Schematic Diagram of NAREL Radon/RDP Chamber A 8
3 Plan (Top) View of Equipment Racks and Continuous
Monitoring Points in NAREL Radon Chamber A 22
11
-------
Glossary
AT
Bi-214
BOM
CC
cfm
CN
CR or CRM
CW
EML
EP
EPA
ER
GR
GW
L
Lpm
NAREL
ORIA
Pb-214
Po-214
Po-218
PR
Ra-226
radon
alpha track detector
the radioisotope bismuth-214
US Bureau of Mines, Denver Research Center
charcoal adsorber
cubic feet per minute
condensation nuclei
continuous radon monitor
continuous working-level monitor
USDOE Environmental Measurements Laboratory
electret-PERM detector
U.S. Environmental Protection Agency
radon decay products/radon equilibrium ratio
grab radon sampling
grab working—level sampling
liter
liters per minute
National Air and Radiation Environmental
Laboratory, USEPA
EPA's Office of Radiation and Indoor Air
the radioisotope lead-214
the radioisotope polonium—214
the radioisotope polonium—218
performance ratio
the radioisotope radium—226
the radioisotope radon-222
111
-------
Glossary - Continued
Rn-222
RDP
RMP
RMP3
RMP3FU
RMP3P
RMP4
RMP4FU
RMP4P
Rn
RP
RTI
TMC
WL
ZnS
the radio!sotope radon-222
radon decay products
National Radon Measurements Proficiency Program
general reference to Round 3 of RMP Program
the followup test for RMP Program Round 3
the performance test for RMP Program Round 3
general reference to Round 4 of RMP Program
the followup test for RMP Program Round 4
the performance test for RMP Program Round 4
radon
Radon Progeny Integrated Sampling Unit
Research Triangle Institute
USDOE Technical Measurements Center
Working-Level
zinc sulfide
microcurie
IV
-------
1.0 Introduction
The U.S. Environmental Protection Agency's Office of Radiation
and Indoor Air (ORIA) established the National Radon Measurement
Proficiency (RMP) Program in 1986. Through this voluntary program,
participants can demonstrate their ability to measure radon and/or
radon decay products by submitting their detection devices to a
blind test in a designated radon chamber*. The first three test
rounds were conducted using the radon chamber located at the U.S.
Department of Energy Environmental Measurements Laboratory (EML) in
New York, NY. Starting with round 4, EPA chambers have been used
for these tests. The National Air and Radiation Environmental
Laboratory (NAREL)** has had a small radon chamber (chamber C)
since 1979. When the NAREL became involved in testing RMP
participants, it was necessary to construct a new chamber (chamber
A, completed in 1987) which was suitable for handling the large
number of participants involved in the testing.***
In this report, radon chambers A and C will be described as
will the associated chamber monitoring systems used to establish
the official target values for radon and radon decay products
concentrations during RMP Round 4. A summary of the test results
for RMP rounds 3 and 4 will be presented and discussed. The
results of several tests run in chamber A to address the reasons
for the general overresponse of charcoal adsorbers (CC's) in round
4 will be discussed and the most probable reasons for the
overresponse will be identified.
*A11 EPA test chambers have the ability to furnish a test
environment for both radon and radon decay products monitoring
equipment. For simplicity, these chambers will be referred to as
"radon chambers" throughout this report.
**Prior to March of 1990, the National Air and Radiation
Environmental Laboratory was known as the Eastern Environmental
Radiation Facility.
***In 1990, a third chamber (Chamber B) was constructed which was of
the same basic design as Chamber A.
-------
2.0 Background Information
Evaluation and calibration of measurement instruments for
radon (Rn-222) or radon decay products (RDP) require a stable
source of radon and RDP in some form that allows exposure over a
designated measurement period. Most radiation measurement
instruments are tested and calibrated with fixed (physically
contained) sources of radioisotopes of known activities, e.g.,
electroplated alpha sources or sealed gamma sources. Generally,
these sources are constructed to reproduce the geometry of the
actual samples. However, radon is a gaseous radioactive material
that is chemically non-reactive and consequently cannot be produced
as a fixed source. Its short half-life (3.82 days) limits the
usefulness of a given amount of radon as a calibration source.
Radon gas in secular equilibrium with its long-lived parent,
radium-226, is often encapsulated for use as a gamma radiation
source, but this configuration is not useful for radon gas or radon
decay products standardization.
Radon decay products include, among others, polonium—218
(Po-218), lead-214 (Pb-214) , bismuth-214 (Bi-214), and polonium-214
(Po—214), which are short—lived particulates formed when radon
undergoes radioactive decay. In the radioactive decay of each atom
of polonium—218 and polonium—214, an alpha particle is emitted. In
about 0.01 percent of the decays of bismuth—214 atoms, an alpha
particle is emitted. Lead-214 atoms do not emit alpha particles
during radioactive decay. These decay products are electrically
charged when formed and tend to attach to or plateout on other
particles in the air (condensation nuclei) or on nearby surfaces
such as walls, clothes, hair and lung tissue. The half—lives of
all these decay products are much shorter than radon, which
effectively limits their use as standardized calibration sources.
A more practical approach to producing stable sources of radon
and RDP is to construct an exposure chamber where Rn—222 can be
constantly removed from a radium-226 (Ra—226) source of known
-------
activity and mixed with measured quantities of air to create known
concentrations of radon and RDP in an enclosed volume. Although
this is more feasible than the fixed sources, it is still
restricted in practice by the inability to extract completely the
radon from the radium. Several techniques have been used for this
purpose including bubbling air through aqueous radium solutions and
passing air over very thin solid sources, but there is always some
question about the completeness of the radon extraction (de-
emanation) .
A solution to all of these problems has not been found to
date; however, the use of exposure chambers fed by Rn-222 from Ra-
226 sources has been the most practical method for calibration.
The lack of quantitative knowledge about the de-emanation of Rn-222
from the source is generally overcome by making measurements of the
resulting radon- or RDP concentrations. As long as the radium
source produces a constant flow of radon that, when diluted,
provides a range of concentrations of radon similar to those that
are to be measured in the field, then the actual strength of the
source is relatively unimportant. Consequently, the measurements
of the radon and RDP concentrations in an exposure chamber become
the principal determinant in the accuracy of the calibration and
testing of instruments in the chamber.
3.0 EPA Radon Exposure Chambers
The NAREL radon chambers in Montgomery, Alabama utilize
several solid Ra-226 sources to produce Rn-222 that is diluted with
measured quantities of air. The combination of source strength
(based on the manufacturer's measurements) and the dilution with
measured quantities of air allows reasonable estimation of the
radon concentration in these chambers. Selection of source
strength and air flow rates allows maintenance of a stable radon
concentration in these chambers up to about 500 pCi/L. This
maximum could be extended by adding more Ra-226 sources. The
practical minimum is limited to approximately 5 pCi/L by
-------
fluctuations in environmental radon concentrations and measurement
precision.
Radon decay products concentrations in the chambers are a
function of radon concentration, air exchange rate and plateout.
The latter two factors are not as easily controlled as the radon
concentration. The air exchange rate can be varied by adding more
or less dilution air; however, some air exchange due to leakage
occurs. This leakage reduces the maximum concentrations of radon
and RDP that can be obtained in a chamber with a given source
strength.
The radon concentration in air sets an effective upper limit
for the RDP concentration, but that maximum theoretical RDP
concentration is rarely achieved in practice because the RDP's are
constantly removed from the air by both plateout and radioactive
decay. The degree of ingrowth of the RDP with respect to the radon
is given by the equilibrium ratio* (ER). The larger the ER, the
more RDP remain in the air for a given radon concentration. One
way to produce a high ER in an exposure chamber is to use a low
dilution air flow rate and a high concentration of condensation
nuclei (CN). Condensation nuclei are small, airborne particles to
which the RDP become attached and neutralized, thus remaining
*ER = WL X 100
[Rn]
where
WL = RDP concentration in units of working-levels. One
working-level (WL) is any combination of the short-
lived decay products of radon (Po-218, Pb-214, Bi-
214, and Po-214) in one liter of air that will
result in the ultimate emission by the bismuth and
two polonium isotopes of 130,000 MeV of alpha
energy.
[Rn] = Radon concentration in pCi/L.
The ER is a fraction less than unity that may also be
expressed as a percent.
-------
suspended in the air and not plating out. For the same radon
concentration a lower ER can be obtained by increasing the dilution
air flow rate and decreasing the CN concentration.
The three NAREL chambers include one small chamber (Chamber C)
that is approximately 10 years old and two large chambers, Chamber
A which was built in 1987 and Chamber B which was built in 1990.
Chamber C (see Figure 1) has an approximate volume of 3.6 m3 (127.1
ft3) and is designed for radon flow in one end and out the other in
a single pass. Three Ra-226 sources of approximately 50, 7, and 6
microcuries (MCi) are used, either individually or together to
produce the desired radon concentrations (up to ~500 pCi/L). The
radon is purged from the sources with air at a flow rate of ~1 Lpm
(per source) and mixed with environmentally conditioned air at
selectable flow rates over a range of about 10 to 200 Lpm. Air
velocity in Chamber C is very low, essentially less than 0.50 fpm
for all operating conditions. Environmental condition ranges
include temperature (32 to 100° F) , relative humidity (10 to 95
percent), and condensation nuclei (CN) concentration (1,000 to
1, 000, 000/mL). Condensation nuclei are generated by electrically
heating a small diameter nichrome wire located in the chamber air
inlet duct. Changes in CN concentration are induced by varying the
voltage across this wire. The equilibrium ratio (ER) in this
chamber is limited to a maximum of about 35 to 40 percent due to
its small size and high air exchange rates. Therefore, RDP
concentrations in units of working—levels (WL) are obtainable over
a range of approximately 0.01 to 2 WL. Access to this chamber is
through a small door that allows instruments to be placed inside or
through a pass box with glove ports.
-------
•Environmental Enclosure
Small
Radon
Chamber
Access Door
Sampling
Port
(Typ. of 6)
Glove Ports
Passbox
•Baffle
(Typ. of 2)
Condensation Nuclei Generator
Flow Monitor
Filter
Humidifier
Variable
Speed Blower'
(10 to 200 1pm)
Radon Source
(Typ. of 3)
Figure 1: Schematic Diagram of NABEL Radon/RQP Chamber C (Side View)
-------
Chamber A (see Figure 2) has a volume of approximately 44 m3
(1550 ft3) and is designed for recycling of most of the radon to
reduce the source strengths required and to provide uniform mixing
of the radon and decay products throughout the volume. Three
Ra-226 sources of approximately 88, 85, and 28 //Ci can be used,
either individually or jointly, to produce radon that is mixed with
air in the recycling system, which has a flow rate ranging from 50
to 200 cfm. The recycled air can be diluted with outside air at a
flow rate ranging from 0 to 35 cfm to produce the desired
concentrations. The larger size of the chamber and lower air
exchange rates allow higher equilibrium ratios. ER' s in excess of
60 percent are attainable. The Chamber A control systems allow
environmental conditions to be varied over the following ranges:
temperature, 32-100° F; relative humidity 20-90%; and CN
concentration, 3,000-500,000/mL. Access to the chamber includes
walk-in capability through a double-door entry room and a small
door for passing objects in and out. Several 3-inch diameter
access ports are provided on both chambers for collection of
samples (by the operator from outside the chambers) with minimal
effect on the radon or RDP concentrations inside. Condensation
nuclei are generated in chamber A by vaporizing carnauba wax using
a heating system which consists of a Pyrex heating tube located
within a tube furnace (Tu81). The heating tube contains a Pyrex
wax container. The furnace temperature can be varied to alter the
production of condensation nuclei. A low air flow through both the
wax container and the heating tube carries the CN from the
generator into the chamber. Chamber B is very similar in design
and construction to Chamber A.
Only chambers A and C were used in the National Radon
Measurements Proficiency Program round 4 (RMP4). RMP3 performance
tests were conducted in the USDOE Environmental Measurements
Laboratory (EML) chamber in New York, NY. The designs of NAREL
chambers A and B are similar to the EML chamber. All chambers were
constructed using insulated metal—clad panels. All have a double-
-------
00
Outside
Vent
JL
-Personnel-
Doors
Key
CC-Cooling Coil
CN-Condensation Nuclei Generator
EHC-Electric Heating Coil
H-Humidifier
HEF-High Efficiency Filter
PCC-Pre-Cooling Coil
PF-Prefilter
^-Flow Measuring Device
/^-Manual Damper
|/|-Motor Controlled Damper
Insulated Environmental
Chamber
-)
.} Air Disi
Hea<
-)
4
^^ ^K ^** ^*
*
fr-
ier
f-
f
•v i* -r T T
ntal f
1 \'
$ \
>
^ — i\
\ \ \ \i
-CN
H
I
Recirculation Air Leg
l/
[Port
j "
\
•«-Radon Source
r £ (Typ. of
\
" yl '
T c
<• " — ' H ...<•- .- ^ ^ -1 N i P
\ — . n
x
/^
^ — —
/ *
. / 1 ^ P i y — M >
'0-35 cfm
Outside Air Leg
Figure 2; Schematic Diagram of NAREL Radon/BDP Chamber A (Top View)
-------
door, air-lock entry and all recirculate chamber air to conserve
radon and radon decay products. In the EML chamber, recirculated
air enters the chamber through a header located near the floor and
travels upward. In the NAREL chambers A and B, recirculated air
enters the chambers through a header located near the ceiling and
travels downward. All chambers use a condensation nuclei generator
based on vaporizing carnauba wax.
4.0 Radon and Radon Decay Products Monitoring*
Although the radon concentration in the chamber can be
estimated from the radium source strength and the flow rate, the
exact concentration will vary with fluctuations in the flow rate
and with small changes in emanation from the source. Consequently,
it is necessary to measure the concentration of radon and decay
products to have an accurate knowledge of the exposure to
instruments and detectors in the chamber. Continuous monitors,
calibrated by periodic grab samples, are used to document the radon
and RDP concentrations during each test period.
Grab samples (short-term samples) for radon are collected
using scintillation cells. These cells are cylindrical with a
volume of 0.125 liters and are coated inside with zinc sulfide
(ZnS) on all surfaces except the clear window on one end which sits
directly on the photomultiplier tube during counting. The
scintillation cell counting equipment is checked at regular
intervals using a sealed standard containing a known amount of
radium-226 with radon and the decay products in secular
equilibrium. The standard has the same counting geometry as the
scintillation cells. Periodically, the efficiency of each
scintillation cell is determined using a known concentration of
*This description of the radon and radon decay products monitoring
systems and their calibrations represents the equipment and
methodology used in RMP round 4. Since that time, improvements
have been made in both the equipment and calibration procedures.
The current equipment and calibration methodology will be described
in a future EPA report.
-------
radon in air derived from an NIST radium—226 standard reference
solution.
For grab sampling radon decay products, a pump is used to pull
air through a filter which traps the RDP's during a five—minute
sampling period. The alpha activity on the filter is determined by
placing it face down on a ZnS disk placed directly on top of the
photomultiplier tube, which gives "two pi" counting geometry. The
Thomas-Modified Tsivoglou technique (Th72) is used to determine
radon decay products concentrations and working levels. The alpha
scintillation counters are calibrated using a NIST—traceable
electroplated americium-241 (Am-241) alpha source in the same
geometry used for filter counting.
Both of these grab sample measurement techniques are
calibrated independently in a NAREL laboratory, and they are
routinely verified through intercomparisons with other laboratories
(Fi88a, Fi88b, Fi87, Fi85a, Fi85b, Fi83, Fi81, Ge87, Pe87, Pe86,
USBOM82). NAREL results for these intercomparisons through 1988
are shown in Tables 1 and 2. For the radon intercomparisons (Table
1) , NAREL was biased low in the first two tests (April and June
1981) . A change was made in the scintillation cell analysis
procedures before the August 1981 intercomparison. The mean ratio
of NAREL results/reference value for the period 8/28/81 to 8/8/88
is 0.98 +/-3.6% (1 standard deviation, s.d.). For the radon decay
products intercomparisons (Table 2) , the mean ratio of NAREL
results/reference value for the period 5/25/82 to 3/26/87 is 1.00
+/—4.3% (1 s.d.). A review of these data in Tables 1 and 2 shows
that the NAREL has performed well in the intercomparisons over a
considerable period of time. For this reason, grab samples are
used to calibrate the chambers continuous radon and RDP monitors.
10
-------
Table 1: NAREL Performance in Grab Radon Intercomparisons
(a)
Date
Facility
Conducting
Test
Test
Category"3'
Reference
Value (c)
(pCi/L)
Ratio-NAREL/
Reference Value
08/08/88
02/08/88
07/20/87
02/23/87
11/07/86
09/02/86
07/21/86
04/03/86
03/03/86
07/15/85
02/25/85
07/16/84
02/06/84
01/24/83
07/12/82
01/27/82
08/28/81
06/15/81
04/20/81
EML(d)
EML
EML
EML
EML
NAREL
EML
EML
EML
EML
EML
EML
EML
EML
EML '
EML
NAREL
EML
EML
F
F
F
F
RMP3
I
RMP2
RMP1
F
F
F
F
F
F
F
F
I
F
F
6.0 + 0.1
12.5 + 0.2
12.2 + 0.4
67.1 + 0.7
21.8
22.1 + 0.06
61.2
45.0
70.2 + 0.8
70.4 + 1.5
34.9 + 0.8
35.2 + 0.6
82.2 + 1.3
46.5 + 1.2
31.3 + 0.4
36.4 + 0.8
41.0 + 6.0
44.6 + 1.1
50.5 + 0.9
1.02 + 0.13
0.94 + 0.05
0.98 + 0.05
0.93 + 0.03
1.00 + 0.05
0.98 + 0.03
0.97 + 0.02
1.06 + 0.03
0.97 + 0.03
1.00 + 0.07
0.97 + 0.03
1.03 + 0.08
0.96 + 0.04
0.98 + 0.06
0.92 + 0.05
0.99 + 0.03
1.00
0.82 + 0.05
0.88 + 0.02
(a)
(c)
(d)
all errors are 1 sigma.
I=informal intercomparison; F=formal intercomparison;
RMP=National Radon Measurement Proficiency Program; F and RMP
are blind tests where the reference value is not known before
the EPA results are submitted.
Reference value for radon is always the EML value, regardless
of the chamber used for the test.
EML-US DOE Environmental Measurements Laboratory, New York, NY.
NAREL-US EPA National Air and Radiation Environmental
Laboratory, Montgomery, AL.
11
-------
Table 2: NAREL Performance in Grab Working-Level
Intercomparisons(a'b>
Date
Facility
Conducting
Test
Reference
Value (c>
(WL)
Ratio-NAREL/
Reference Value
3/25/87 (Test 1)
3/26/87 (Test 2)
3/25-26/87 (Avg.,
2 Tests)
4/16/86 (Test 1)
4/16/86 (Test 2)
4/16/86 (Test 3)
4/16/86 (Test 4)
4/17/86 (Test 5)
4/17/86 (Test 6)
4/16-17/86 (Avg.,
6 Tests)
9/11/85
9/11/85
9/11/85
9/11/85
9/12/85
9/12/85
(Test 1)
(Test 2)
(Test 3)
(Test 4)
(Test 5)
(Test 6)
9/12/85 (Test 7)
9/11-12/85 (Avg.,
7 Tests)
EML(d)
EML
EML
TMC
TMC
TMC
TMC
TMC
TMC
TMC
TMC
TMC
TMC
TMC
TMC
TMC
TMC
TMC
0.112 +. 0.005
0.061 + 0.004
0.317 ±
0.343
0.067
0.022
0.201 +. 0.013
0.155 0.010
0.093
0.092
0.007
0.007
0.304 i 0.011
0.305 ± 0.008
0.182 i 0.006
0.175 ± 0.006
0.093 ± 0.005
0.092 i 0.006
0.087 + 0.003
(a)
0.98 ± 0.06
0.90 ± 0.06
0.94 + 0.06
1
1
08
00
0.97
1
1
02
04
0.99
1.02 +
0.04
0.94
1.02
0.99
1.02
0.92
1.06
1.01
0.99 + 0.05
5/25-26/82
5/25-26/82
5/25-26/82
5/25-26/82
5/25-26/82
5/25-26/82
5 Tests)
(Test
(Test
(Test
(Test
(Test
(Avg . ,
1)
2)
3)
4)
5)
BOM
BOM
BOM
BOM
BOM
BOM
2
2
2
2
0
.14
.20
.24
.24
.59
+
_+
_+
±
±
0
0
0
0
0
.08
.08
.10
.07
.09
1
1
0
1
1
1
.02
.00
.99
.01
.00
.00 +
0.01
(d)
All errors are 1 sigma.
All data in the table are for formal intercomparisons where the
reference value is not known before the EPA results are
submitted.
Reference value for WL is the mean of all participants.
EML-US DOE Environmental Measurements Laboratory, New York, NY.
TMC-US DOE Technical Measurements Center, Grand Junction, CO.
BOM—US Bureau of Mines, Denver Research Center, Denver, CO.
Each continuous flow radon detector used in the NAREL chambers
12
-------
is comprised of a 0.5 liter flow-through scintillation cell* and
a 5-inch (12.7—cm) photomultiplier tube in a light-tight housing.
Chamber C employs one continuous radon monitor and Chamber A uses
two. To monitor the RDP concentration in the chambers
continuously, a surface barrier semi-conductor detector is mounted
directly adjacent to a filter through which air is drawn. The
detector measures the alpha radioactivity deposited on the filter.
The RDP detector is located inside the chamber and is connected
through a small tube to an air pump with controlled flow set at 1
Lpm. Chamber C has one continuous WL monitor and Chamber A has
two.
The radon and the RDP detectors are connected to computer-
based multichannel analyzers that accumulate data and periodically
print the total number of alpha counts accumulated by each
detector. In addition to the continuous monitors which are part of
each chamber monitoring system, additional self-contained
continuous radon and RDP monitors are operated in each chamber as
backup equipment during an RMP exposure round. For each monitor,
the periodic counts are converted to radon or RDP concentrations
with calibration factors. A calibration factor is the ratio of the
concentration as measured by a grab sample over a short time to the
periodic count generated by the continuous monitor during the
period in which the grab sample was collected. Two calibration
factors are normally computed during each weekday of a test period
and one during each weekend day. An arithmetic mean calibration
factor for each monitor is then computed. The final step is to
multiply each periodic count from a continuous radon or radon decay
products monitor by the arithmetic mean calibration factor for that
*During RMP round 4, the 0.5 L scintillation cells were used.
These 0.5L cells have been replaced with larger cells (some 1.0 L
and some 1.4 L) which has increased the sensitivity for the radon
continuous monitoring systems.
13
-------
monitor to form a time series of periodic concentrations for the
entire test period.
In judging the quality of the data produced by each continuous
monitor during Round 4, EPA computed the standard deviation of the
mean calibration factor for each monitor. The standard deviation
for a monitor was divided by the mean calibration factor and the
result was multiplied by 100 to convert the fraction to percent.
This procedure yielded a percentage standard deviation of the mean
calibration factor. A low percent standard deviation indicated
that the calibration factors used to generate the mean calibration
factor for an instrument were consistently near the mean value.
For RMP round 4, the percent standard deviations for all continuous
monitors used to establish the official target values (against
which participants results were compared) were less than eight
percent.
For the performance and the followup tests in round 4, the
official target values for alpha track detectors (AT), charcoal
canisters (CC), continuous radon monitors (CR), and electret-perms
(EP) were established using data from the continuous radon monitor
record. The target values for continuous working—level monitors
(CW) and radon progeny integrated sampling units (RP) were
established using data from the continuous working-level record.
For the grab radon (GR) detectors, NAREL scintillation cells were
filled as participants sampled. If the sampling time for a
participant's cells (or other devices) was short (generally 20
minutes or less) the mean of the NAREL scintillation cell results
was used to establish the official target value. Generally, if the
participant's sampling time extended more than 20 minutes, the mean
of the NAREL scintillation cell results and the continuous
monitoring record for the sampling period were used to establish
the official target values. For the grab working-level (GW)
participants, NAREL grab—sampled with the participants to obtain
the official target values.
14
-------
5.0 Summary of RMP Round 3 and Round 4 Results
Participant statistics for each detection method tested in
Rounds 3 and 4 are listed in Table 3. EPA categorizes AT, CC, some
GR, EP and RP as passive (or mail-in) methods and CR, some GR, GW
and CW as active (or walk-in) methods. During RMP4P, there were 56
test periods where walk-in participants tested. For test periods
with multiple walk—in participants, all participants failed during
2 periods (3.6 percent of the test periods). There were 5 test
periods when a participant failed and he was the only one testing
that method.
Referring to Table 3, the failure rates for AT, CC, GR, RP and
CW were higher for RMP4P than for RMP3P. An explanation may be
that as the RMP progresses, the new participants entering the
program are generally less experienced in radon and RDP detection
when they enter than the participants who joined in the early
rounds. This may cause a higher failure rate from new entries than
would have been the case for earlier rounds. Failure rates were
lower in RMP4P for CR and GW methods. EP tested for the first time
in RMP4P and all entries passed the performance round.
The method performance ratios (PR) (Table 3) are a measure of
the degree of agreement between the participants for each method
and the NAREL measurements*. In RMP4P, the method PR's for CR,
GR, EP, GW, and CW were all close to 1.0 which indicates generally
good agreement between the participants using these methods and
NAREL. The low bias (15 percent) of RP can be explained by
plugging of the filters on these devices with wax from the chamber
A condensation nuclei (CN) generator, leading to a progressive
'Histograms for all methods tested in RMP4P are included in
Appendix A.
15
-------
Table 3: Methods Completed, Failure Rates and
Method Performance Ratios for RMP Round 3 and RMP Round 4
Method
Methods
Completed'3'
Round
3P 4P 4FU
Failure Rate
(Percent)
Round
3P 4P 4FU
Method Performance
Ratio(b)
Round
3P 4P 4FU
AT
CC
CR
GR
EP
GW
RP
CW
36
120
4
21
0
35
4
17
75
253
15
40
5
46
5
40
9
73
0
9
0
5
1
5
5.
4.
25.
10.
—
34.
5.
6
2
0
0
3
0
9
13
30
6
25
28
20
13
.2
.3(e>
.7
.0
0
.3
.0
.2
22.
20.
—
33.
—
0.
0.
0.
2
5
3
0
0
0
0.
0.
0.
0.
—
0.
0.
0.
92
93
91
96
93
95
98
0.
1.
1.
0.
1.
0.
0.
1.
93
16
01
97
01
98
85(c)
04
1
1
1
1
1
0
.06
.06
—
.02
—
.08
.10
.97
Total 237 479 102
9.7 24.4 19.6(d) -0.93 -1.07 -1.05
(a)
(b)
(c)
(d)
(e)
A method is complete when a participant returns his results
reporting form to RTI and is evaluated against the official
target value.
A detector performance ratio (PR) is the quotient of a
participant's detector reading divided by the official target
value. A participant performance ratio is the average of the
detector performance ratios for all detectors tested by a
participant. The method performance ratio is the mean of the
participants' performance ratios for all participants testing
this method, except for outliers. An outlier is a participant
performance ratio below 0.5 or above 1.5.
The low bias for RP is explained by wax from our CN generator
partially plugging the filter in these units. RP's were run for
shorter exposure times in RMP4FU to alleviate this problem.
4.2 percent of the methods that completed RMP4P failed RMP4.
For RMP4P, there were 74 CC failures. 68 percent of failures
were from primary suppliers and 32 percent of failures were from
participants who supply CC's only for themselves. For the
RMP4FU, there were 15 CC failures. Ten of these failures were
CC's from a single processing laboratory.
16
-------
reduction in air flow rates during the exposure period*. NAREL
staff have experienced similar problems with EPA RP devices in 7 day
field testing in homes occupied by moderate to heavy smokers. This
problem was corrected by reducing the 6-7 day exposure time in the
performance test to 3-4 days in the followup test.
CC detectors were biased 16 percent high and AT detectors were
biased 7 percent low. Experience with the international
intercomparison program (Fi88a, Fi88b, Fi87, Fi85a, Fi85b, Fi83,
Fi81, Ge87, Pe87, Pe86, USBOM82) has shown that the agreement in
radon and RDP measurements between organizations operating radon and
RDP calibration facilities would generally fall within a +_ 10 percent
band. Specifically, EPA staff would normally expect that NAREL and
other U.S. organizations operating radon and RDP calibration chambers
could simultaneously measure the same radon or working-level
environment and produce results that would differ by no more than
+_ 10 percent. In most instances, the agreement would be better.
When the RMP program bias test was established at +_ 25 percent, the
persons who formulated this policy reached a consensus that a bias
check of +_ 15 percent was proper. Ten percent was added to obtain
the +_ 25 percent in recognition that many detectors tested in RMP
would be calibrated in a chamber not used for RMP testing and that,
in general, a difference between measurements of chamber operators of
+.10 percent or less could be expected.
Considering the above discussion, it is not surprising that AT
were biased low by 7 percent since the great majority of the 75 sets
of detectors tested were produced by two manufacturers who had
calibration data from sources other than the NAREL chambers. The 7
percent bias is 3 percent less than was allowed for calibration and
test chamber differences in setting the RMP 25 percent bias criteria.
The 16 percent high bias in the CC participants results is higher
*Wax from the EML CN generator was also present in the EML
chamber during RMP Rounds 1-3.
17
-------
than would normally be expected due to chamber differences. Several
possible contributing factors were identified and a discussion of
each will be given in section 6. To evaluate some of the factors,
radon concentration uniformity testing in chamber A (beyond the tests
already performed) was needed. Also, it appeared that tests of CC
response in chamber A, using EPA CC's, could be enlightening. These
tests and the original chamber A radon concentration uniformity tests
are discussed in section 6.
6.0 EPA Chamber A and Charcoal Canister Testing
Between April and September, 1987, NAREL staff conducted seven
tests in chamber A using EPA CC's. The tests were conducted to
document radon concentration uniformity within chamber A using both
CC's and continuous radon monitors (CRM) and to see if EPA could
cause its own CC's to experience a major bias in chamber A. The
environmental parameter values (temperature, relative humidity, etc.)
and selected results for these tests are listed below and in Table 4.
A discussion of the seven tests and the results shown in Table 4
follow.
The range of values for test parameters which remained nearly
constant are
* chamber temperature, 68-70 degrees F,
* chamber relative humidity, 46-50 percent,
* outside air flow rate, 0 cubic feet per minute (cfm),
* chamber air recirculation rate, 95—110 cfm,
* periodically, a humidifier fan would run to distribute
moisture across the chamber volume to maintain relative
humidity, and
* air velocities across the chamber volume ranged from <10 to
~70 feet per minute (fpm).
18
-------
Table 4: Summary of NAREL Radon Chamber A Tests
Chamber
Average Radon
Concent ration
(pCi/L)
Fcolumnl Test CC
Number Brand
Row Dates [1] [2]
1
2
3
4
5a
5b
5c
5d
5e
6a
6b
7a
7b
1,
2,
3,
4,
5a,
5b,
5c,
5d,
5e,
6a,
6b,
7a,
7b,
4/1-3/87
5/8-10/87
6/19-21/87
6/22-24/87
8/5-7/87
8/5-9/87
8/5-6/87
8/6-7/87
8/5-7/87'
8/25-27/87
8/25-27/87
8/31-9/2/87
8/31-9/2/87
1
1
1
1
1
1
1
1
2
1
1
1
2
Exp.
Time
[3]
-50
-48
-50
-48
-49
-99
-26
-25
-49
-48
-48
-48
-48
CN Cone.
(Particle)
(cm1) [4]
-
yes
yes
yes
-50,000
-50,000
-50,000
-50,000
-50,000
-2,500
-2,500
-250,000
-250,000
Continuous
Monitors
±% Std.
Dev. [5]
No
40
25
24
31
33
29
32
31
25
25
24
24
Reading
.8±1.0%
.2±1.8%
.2±2.3%
.012.5%
.1±3.3%
.4+2.5%
.8±2.6%
.0+2.5%
.0*
.0*
.8*
.8*
CC's
+% Std. No. CC' s
Dev. Exposed
[6] [7]
40.3±4.6%
41.2±4.9%
27.2±4.4%
25.7+7.0%
33.3±7.2%
33.8±5.3%
35.0+5.3%
26.5±5.6%
27.8±5.4%
26.8+3.2%
28.5±3.8%
30
72 -
26
22
48
20
20
5
5
5
5
Perform.
Ratio
(PR) [8]
-
1.01
1.08
1.06
1.07
1.02
1.13
1.06
1.11
1.08
1.15
Continuous
CC Measurements Monitor
Range of Measurements
Deviation Range of
from CC Mean Deviation
(Percent) of CRM' s
for from CRM
Equimpent Shelf Mean Percent
Racks [9] Levels (10] [11]
-5.0 to +3.0 -1.0 to +1.0
-2.9 to +3.9 -1.5 to +1.0
-
- -
-4.2 to +4.2 -3.0 to +3.3 -2.7 to +2.1
-3.8 to +1.9
-2.7 to +2.1
-2.9 to +2.1
- -2.7 to +2.1
- -
_
- -
-
* Only one continuous monitor was run during these tests.
-------
Test 1 was conducted before round 4, using 1 CC at each monitoring
location, and demonstrated good uniformity of radon concentration
across the chamber volume. The other 6 tests were conducted during
and following round 4, using two CC's at each monitoring location.
With the exception of test 1, one or more CRM's were operated so that
the CRM results could be compared with the CC results. For all
tests, the CRM's were calibrated with a series of scintillation
cells (grab samples) filled at different times during testing. The
normal CRM calibration procedure was to establish a calibration point
for each CRM for each cell filled; thus, the sampling points for some
of the CRM's were separated from the point where the scintillation
cell was filled by a few feet. Test 1 results had demonstrated that
the chamber A radon concentration was uniform across the chamber
volume and this calibration method works well for uniform
concentrations. However, in test 5 (the most comprehensive test),
three continuous radon monitors were operated, each calibrated
independently from the others to further document uniformity of radon
concentration within the chamber. These calibration procedures and
test results are discussed with the discussion of test 5.
EPA has used two brands of CC's in the NAREL CC program. Both
brands were initially calibrated in NAREL chamber C which has
negligible air velocity and does not use wax for generating CN.
However, the NAREL scintillation cells were used to calibrate the
CRM's in both chambers A and C. Thus, the results of testing EPA's
CC' s in chamber A should have shown any bias imparted to CC' s by some
condition in chamber A which was different than chamber C but should
not have included any differences due to different organizations
making basic radon measurements to document chamber levels. For
these tests, brand 1 CC's were used exclusively for documenting radon
concentrations across the chamber volume. However, a few brand 2
CC's were used in tests 5-7 with some interesting results (discussed
later).
20
-------
Chamber A has six equipment racks, each with four vertically
spaced shelves. A plan (top) view of the arrangement of the
racks (designated A-F) is shown in Figure 3. The four shelves on each
rack are denoted as level A through level D from top to bottom. For
tests 1, 2 and 5a, the concentration uniformity tests (where CC's
were placed on each shelf of each rack), average concentrations were
calculated for the chamber (using the results for all CC's) and for
each level and each rack. The percent difference between the rack
and level averages and the overall mean chamber concentration were
computed. The range of differences (in percent) between the CC
chamber mean and the means for the 6 racks is a measure of horizontal
radon concentration variation in chamber A. The range of differences
(in percent) between the CC chamber mean and the means for the four
levels is a measure of vertical radon concentration variation.
Test 1.
During test 1 (Table 4, row 1), 30 brand 1 CC's were exposed at
30 chamber A locations for ~50 hours from 4/1-3/1987. No other
monitoring devices were in chamber A during this test. The 30 CC's
were placed on all shelves of each rack (24 shelves) . Three cans
were placed on the floor. The mean radon concentration for the 30
CC's was 40.3 pCi/L +/- 4.6% (1 standard deviation, s.d.). Rack
averages ranged from 5.0% below to 3.0% above the mean CC
concentration. Level averages ranged from 1.0% below to 1.0% above
the mean CC concentration. The average concentration measured by the
CC's on the floor was 0.2 % below the chamber mean.
Test 2.
During test 2 (Table 4, row 2), 72 brand 1 CC's were exposed at
36 chamber A locations (in pairs) for ~48 hours from 5/8-10/1987. No
21
-------
Rack B
Rack A
0
RGM II
Unit 136
Level D
Rack D
Rack C
0
RGM II
Unit 199
Level A
Rack F
0
System 1
Monitor
Level B
Rack E
Notes:
0
Level A
Level B
Level C
Level D
Continuous monitor suction point
Top shelf
Second highest shelf
Third highest shelf
Bottom shelf
Figure 3. Top View of Equipment Racks and Continuous Radon
Monitoring Points in Radon Chamber A.
22
-------
other monitoring devices, except for the chamber continuous monitors,
were in the chamber. As with test 1, the CC's were distributed
evenly across the shelves of the chamber racks. The mean radon
concentration for the 72 CC's was 41.2 pCi/L +/- 4.9% (1 s.d.) . Rack
averages ranged from 2.9% below to 3.9% above the mean CC
concentration. Level averages ranged from 1.5% below to 1.0% above
the mean CC concentration. Two CRN's were operated during this test
and the average radon concentration measured by the CRM's was 40.8
pCi/L + /- 1.0% (1 s.d.). EPA defined the performance ratio (PR), for
these chamber tests, as the average radon concentration measured by
CC's divided by the average radon concentration measured by CRM's.
This ratio gave a measure of the degree of bias of CC' s when compared
to the CRM's, which are a more accurate method of radon measurement.
For this test, the PR was 1.01 which indicated an average high bias
of ~1% for the 72 brand 1 charcoal canisters.
Test 3.
During test 3 (Table 4, row 3), 26 brand 1 CC's were exposed at
13 chamber A locations (in pairs) for ~50 hours from 6/19-21/1987.
This test was run during RMP4 followup testing by placing two EPA
CC's on each tray of participant CC's exposed in chamber A. The mean
radon concentration for the 26 CC's was 27.2 pCi/L +/- 4.4% (1 s.d.) .
The EPA CC's were not distributed uniformly across the chamber
volume; thus, rack and shelf averages were not meaningful and range
of variation data were not computed. However, the average radon
concentration obtained from the three CRM's operating during the test
was 25.2 pCi/L +/- 1.8% (1 s.d.) . The PR calculated for the 26 brand
1 CC's used in this test was 1.08, which indicated an average high
bias of -8% for the 26 CC's.
Test 4.
During test 4 (Table 4, row 4), 22 brand 1 CC's were exposed at
11 chamber A locations (in pairs) for -48 hours from 6/22-24/1987.
As with test 3, this test was run during RMP4 followup testing by
23
-------
placing two EPA CC' s on each tray of participant CC's. The mean
radon concentration for the 22 CC's was 25.7 pCi/L +/- 7.0% (1 s.d.).
As in test 3, the uneven distribution of EPA CC's across the chamber
volume precluded calculation of rack and level range of variation
data. However, the average radon concentration obtained from the
three CRM's operating during the test was 24.2 pCi/L +/- 2.3% (1
s.d.). The PR calculated for the 22 brand 1 CC's used in this test
was 1.06, indicating an average high bias of ~6% for the 22 CC's.
Test 5.
Test 5 was the most comprehensive of the 7 tests discussed in
this report. The test period was 8/5-9/87. Four hundred-eighty
previously used EPA CC's were baked in a furnace to regenerate them
and then distributed uniformly across the chamber volume (and opened)
to serve as a large radon "sink" during testing. These CC's were not
used for radon measurements but simulated a heavy charcoal loading in
chamber A. In addition to these 480 "sink" CC's, six different
groups of EPA CC's were used for radon measurements as listed below:
5a. 48 brand 1 CC's exposed for two days,
5b. 20 brand 1 CC's exposed for four days,
5c. 20 brand 1 CC's exposed for one day (first day of test),
5d. 20 brand 1 CC's exposed for one day (second day of test),
5e. 20 new brand 2 CC's exposed for two days,
5f. 20 regenerated brand 2 CC's exposed for two days.
The total chamber charcoal loading (~42 kg) was almost twice the
maximum load due to participant CC's during RMP4. The test results
for the six groups will be discussed individually.
24
-------
Group 5a.
During test 5a (Table 4, row 5a), 48 brand 1 CC's were exposed
at 24 chamber A locations (in pairs) for ~49 hours from 8/5-7/1987.
The "radon sink" CC's were placed on trays (20 per tray on 24 trays)
and two of the group 5a CC's were placed on each tray. One tray of
CC's was placed on each shelf of each rack. The mean radon
concentration for the 48 CC's was 33.3 pCi/L +/- 7.2% (1 s.d.). Rack
averages ranged from 4.2% below to 4.2% above the mean CC
concentration. Level averages ranged from 3.0% below to 3.3% above
the mean CC concentration.
Three CRM's were operated during this test. The intakes for
these monitors were spatially distributed across the chamber volume
(see Figure 3) such that a chamber air sample was drawn (from left to
right, facing the front of the chamber from the outside) of the
* near the bottom shelf (level D) on the left
side and about equidistant from front and back,
* near the top shelf (level A) in the middle
of the chamber near the front and
* near the second shelf from the top (level B)
on the right side near the rear
chamber. Each CRM was calibrated independently with periodic
scintillation cell grab samples taken within two inches of the CRM
intakes. By calibrating the CRM's independently for this test, EPA
staff were able to use the results to study chamber radon
concentration uniformity and to compare the CRM results to the
results obtained with the CC's. The mean radon concentration for the
three CRM's was 31.0 pCi/L +/- 2.5% (1 s.d.). Individual CRM
readings ranged from 2.7% below to 2.1% above the mean CRM
concentration.
For this test, the PR for the brand 1 CC's was 1.07, which
indicated an average high bias of ~7% for the 48 CC's.
25
-------
Group 5b.
During test 5b (Table 4, row 5b), 20 brand 1 CC's were exposed
at 10 chamber A locations (in pairs) for ~99 hours from 8/5-9/1987.
The mean radon concentration for the 20 CC's was 33.8 pCi/L +/- 5.3%
(1 s.d.). Since only 10 chamber locations were monitored, the brand
1 CC' s were not distributed uniformly across the chamber volume so
that rack and shelf averages would be meaningful. Consequently,
range of variation data was not computed.
The average radon concentration obtained from the three
independently calibrated and spatially separated CRM's was 33.1 pCi/L
+/- 3.3% (1 s.d.) . Individual CRM readings ranged from 3.8% below to
1.9% above the mean CRM concentration.
The PR calculated for these 20 brand 1 CC's was 1.02, indicating
an average high bias of ~2% for the 20 CC's. This was the lowest
bias for any group of CC's exposed during test 5. This was probably
because the strongest calibration data base for the brand 1 CC's is
for four days exposures.
Groups 5c and 5d.
During test 5c, 20 brand 1 CC's were exposed at 10 chamber A
locations (in pairs) for ~26 hours from 8/5-6/1987. During test 5d,
20 brand 1 CC's were exposed at 10 chamber A locations (in pairs) for
~25 hours from 8/6-7/1987. The EPA brand 1 CC's are not well
calibrated for a one day exposure. Due to this lack of calibration
data, the information obtained was not useful and will not be
discussed further. However, the data from the three independently
calibrated CRM's was useful in studying uniformity of radon
concentration across chamber A for the first and the second day of
the test. For test 5c (8/5-6/1987), the mean radon concentration for
the three CRM's was 29.4 pCi/L +/- 2.5% (1 s.d.). Individual CRM
readings ranged from 2.7% below to 2.1% above the mean CRM
concentration (Table 4, row 5c). For test 5d (8/6-7/1987), the mean
26
-------
radon concentration for the three CRM's was 32.8 pCi/L + /- 2.6% (1
s.d.). Individual CRM readings ranged from 2.9% below to 2.1% above
the mean CRM concentration (Table 4, row 5d).
Group 5e.
During test 5e (table 4, row 5e), 20 brand 2 CC's were exposed
at a single chamber A location for ~49 hours from 8/5-7/1987. The
test period was the same as for the brand 1 CC's of group 5a. The
purpose of this group was to compare the response of brand 1 and
brand 2 CC' s in chamber A for the two-day exposure period recommended
for brand 2 CC's. The mean radon concentration for the 20 brand 2
CC's was 35.0 pCi/L +/- 5.3% (1 s.d.) which yielded a PR of 1.13, a
high bias of 13%. The 48 brand 1 CC's (group 5a) exposed during the
same time period yielded a PR=1.07. Thus, during this test, the
brand 2 CC's were biased almost twice as high as the brand 1 CC's in
chamber A.
This test was the first time EPA staff had observed a difference
in response between the brand 1 and brand 2 CC's, which were both
calibrated in chamber C. EPA staff began looking at the differences
between chamber C (where the CC's were calibrated) and chamber A.
Air velocity measurements were made in chambers A and C and
differences in the average air velocities in the chambers were
documented (C has negligible velocity). Also, it seemed apparent
that the air velocity in chamber A caused a minor high bias in brand
1 CC's but a more pronounced high bias in brand 2 CC's. Additional
testing (beyond the seven tests discussed in this report) for
velocity sensitivity of both brands of CC's was performed which
clearly documented the velocity sensitivity of both brands and the
generally greater sensitivity to air velocity of brand 2 when
compared to brand 1 (Gr88) . As an example, one group of tests using
brand 2 CC's was conducted in Chamber C from January 6 through
January 16, 1988, with chamber temperature ~70 degrees F, relative
humidity ~50% and chamber radon concentrations of 111-137 pCi/L. The
CC's were exposed for the recommended two days. The agreement was
27
-------
excellent between the brand 2 CC's and the CRM's where there was no
perceptible air velocity. However, when the brand 2 CC's were
exposed with air velocity (using a fan), they experienced a 16%
high-bias for ~50 feet/minute (fpm), a 23% high-bias for ~100 fpm and
a 27% high-bias for ~200 fpm air velocity.
A possible explanation may be that CC's calibrated under
conditions of negligible velocity probably have a concentration
gradient established between the air near the top charcoal surface
and the bulk chamber air because diffusion from the bulk air would be
the primary mechanism to replace the radon which is adsorbed by the
charcoal. This would lead to a lower radon concentration in the air
next to the charcoal surface than in the bulk chamber air (the air
monitored by the chamber continuous radon monitors used to calibrate
the CC's). Thus, the CC counting rate would be lower in this case
than if the CC charcoal surface were exposed to the bulk air
concentration. When CC's calibrated in this manner are exposed to an
environment where velocity is present, only a small velocity would be
needed for turbulence to dominate, and completely override diffusion,
as the mechanism for radon replacement near the charcoal surface. In
this situation, the air near the charcoal could be kept at the
chamber bulk air concentration. Thus, a CC calibrated in a still
environment would overrespond when exposed in an environment with air
velocity present.
Group 5f.
During test 5f, 20 regenerated brand 2 CC's were exposed at a
single chamber A location (adjacent to the group 5e CC's for ~49
hours from 8/5-7/1987. EPA did not have calibration data for
regenerated brand 2 CC' s and it was obvious that the calibration data
for new CC's would not work with regenerated CC's. Due to the lack
of calibration data, the information from this test was not useful
and will not be discussed further.
28
-------
Transient Response of CRM's
During test 5, the charcoal loading of chamber A was
sufficiently heavy to cause the chamber radon concentration to
decrease to a minimum of ~63% of the initial concentration ~5 hours
after the test began. The concentration did not fully recover during
the four days of testing. Two days after the test began, the radon
concentration had recovered to ~83% and four days after testing began
to ~87% of the initial concentration. However, even with the heavy
charcoal loading and the associated radon concentration transient,
the concentrations across the chamber volume (measured by the three
independently calibrated CRN's) remained essentially constant as is
shown in Table 5.
Test 6 and 7.
Tests 6 and 7 were identical tests with one exception. During
test 6, the condensation nuclei (CN) generator was not running and
the particle count in chamber A was in the range of 5,000
particles/mL. During test 7, the CN generator was running at maximum
output and the particle count (wax particles) was in the range of
500,000 particles/mL. During these tests, 5 each of the EPA brand 1,
new brand 2 and regenerated brand 2 CC' s were exposed at a single
location. The purpose of these tests was to try to define any
difference in response of the EPA CC's exposed with and without wax
particles in the environment. No other monitoring devices, except
for the chamber continuous radon and continuous working-level
monitors used to document chamber concentrations, were in the
chamber. Due to the lack of calibration data for the regenerated
brand 2 CC's, the information obtained was not useful and will not be
discussed further.
29
-------
Table 5: Test 5 Transient Radon Relative Concentrations Measured
by Three Independently Calibrated Continuous Radon Monitors
Relative Radon Concentration*
(% of initial concentration)
Date
8/5/87
8/7/87
8/9/87
Time
0800
0900
1000
1100
1200
1300
1400
1500
1600
1700
1800
0900
0900
RGMII
Unit 136
100
100
94
82
68
64
62
64
64
66
67
83
90
RGMII
Unit 199
100
98
98
86
73
66
63
65
66
67
69
85
86
System
1
100
99
96
82
67
63
63
64
67
63
67
80
85
*The air intakes for these CRM's are spatially separated for
assessment of uniformity of radon concentrations (see Figure 3).
Test 6 (Table 4, rows 6a and 6b) extended over a ~48 hour period
from 8/25-27/1987. During this period, the mean CRM radon
concentration (one monitor) was 25.0 pCi/L. The mean radon
concentration for the 5 brand 1 CC's was 26.5 pCi/L +/- 5.6% (1
s.d.). The PR calculated for the brand 1 CC's was 1.06, indicating
a high bias of ~6%. The mean radon concentration for the 5 new brand
2 CC's was 27.8 pCi/L +/- 5.4%. The PR calculated for the brand 2
CC's was 1.11, indicating a high bias of ~11%.
Test 7 (Table 4, rows 7a and 7b) extended over a ~48 hour period
from 8/31-9/2/1987. During this period, the mean CRM radon
concentration (one monitor) was 24.8 pCi/L. The mean radon
concentration for the 5 brand 1 CC's was 26.8 pCi/L +/- 3.2% (1 s.d.)
which yielded a PR of 1.08, indicating a high bias of ~8%. The mean
radon concentration for the 5 new brand 2 CC's was 28.5 pCi/L +/-
30
-------
3.8%. The PR calculated for the 5 brand 2 CC's was 1.15, indicating
a high bias of ~15%.
Discussion
A great deal of testing was conducted in NAREL chamber A with
EPA CRM's and CC's following RMP4. The purpose was to evaluate
uniformity of radon concentration across the chamber A volume and the
response of two brands of CC's used in the NAREL CC program. The
significant findings which resulted from these tests are stated and
discussed below:
* For air recirculation rates near 100 cfm and operation of the
humidifier fan in NAREL chamber A, air velocity varied across the
volume from <10 to ~70 fpm.
* The response of EPA 4-inch, open faced CC's was shown to be
dependent on the air velocity near the face of the CC (see
explanation of test 5e).
* EPA operates a CC program which has utilized two brands of 4
inch, open-faced CC's. Both brands were initially calibrated in
NAREL chamber C where air velocities are negligible. When these CC's
were exposed in NAREL chamber A under conditions where air velocities
ranged from <10 to ~70 fpm, the weighted-mean high bias for the 198
brand 1 CC's tested was 4.5% and for the 30 brand 2 CC's, 13.0%.
Thus, for these tests, the weighted high bias of the brand 2 CC's was
almost three times the weighted high bias of the brand 1 CC's.
* For air recirculation rates near 100 cfm and cyclic operation of
the humidifier fan, extensive testing using both brand 1 CC's and
three CRM's demonstrated that radon concentrations across the chamber
A volume remained essentially uniform, even with a heavy charcoal
load in the chamber. For the uniformity tests using CC's (tests 1,
2 and 5a), CC's were placed on each level of each chamber equipment
rack. Average radon concentrations were computed for all racks, all
31
-------
levels and the entire chamber. For these three tests (referring to
Table 4) the greatest CC standard deviation of the mean was 7.2% and
rack averages ranged from a low of 5.0% below to a high of 4.2% above
the mean CC concentrations. Level averages ranged from a low of 3.0%
below to a high of 3.3% above the mean CC concentrations. For the
CRM uniformity tests (tests 5a-5d), three CRM' s were spatially
distributed across the chamber volume (see Figure 3 and the
discussion of Group 5a) . For these four tests, the highest CRM
standard deviation of the mean was 3.3% and the individual CRM
readings ranged from a low of 3.8% below to a high of 2.1% above the
mean CRM concentrations (see Table 4) . EPA believes that the
consistently low standard deviation of the mean concentration and the
consistently low rack and shelf ranges of variation from the mean,
using both CC's and CRM's, demonstrated good uniformity of radon
concentration throughout the chamber A volume, even when the chamber
was heavily loaded with charcoal.
Although both measurement methods yielded low ranges of
variations, the CRM data generally displayed less variation than the
CC data. This was probably due to the effects of velocity variation
at different points in chamber A upon the response of brand 1 CC's.
In addition, NAREL staff believe that CRM's are an inherently more
accurate method for radon measurements than CC's.
* When chamber A had a heavy loading of charcoal (~42 kg) with an
air recirculation rate near 100 cfm and operation of the humidifier
fan, the radon concentration in the chamber decreased to about 60-65%
of the initial value after ~5 hours and slowly recovered to ~85-90%
after 4 days. Even during this concentration transient induced by
the heavy charcoal load, chamber A radon concentrations remained
uniform across the volume (see Table 5).
* The brand 1 CC's exposed with wax particles in the chamber air
were biased ~2% higher than those exposed without wax particles (~8%
bias with wax, ~6% bias without wax). The new brand 2 CC's exposed
with wax particles were biased ~4% higher than those exposed without
32
-------
wax particles (~15% bias with wax, ~11% bias without wax).
7. 0 Discussion of Possible Reasons for Round 4 High Bias of Charcoal
Adsorbers
As part of the effort directed toward understanding the reasons
for the high bias of the CC's tested in RMP4P, several possible
reasons were hypothesized. Tests using NAREL CC's in chamber A
(described in section 6) were carried out, where needed, to determine
the plausibility of the various possible reasons. In this section,
the possible reasons considered by EPA for the bias are listed and
discussed.
A. The CC bias could have been due, in part, to incorrect radon
measurements by NAREL during round 4.
EPA staff identified several points which supported the validity
of the NAREL radon target values for round 4. First, the method
performance ratios (PR) for 4 methods for measuring radon (which were
tested simultaneously with CC's) showed good agreement with EPA
target values (see Table 3) . These other methods were continuous
radon (15 participants, PR=1.01), grab' radon (40 participants,
PR=0.97), electret-PERM (5 participants, PR=1.01) and alpha track (75
participants, PR=0.93). Second, NAREL had consistently performed
well in the radon international intercomparison program (see Table
1) . The radon grab sample method that is used in the intercomparison
program is also used to calibrate the NAREL continuous radon
monitors. Data from these continuous monitors was used to establish
the radon target values for methods CC, CR, EP and AT. (The fifth
radon measurement method, GR, was compared directly to the NAREL grab
sample results, as discussed previously). Third, for several
specific exposure periods checked, the average of the NAREL
scintillation cells filled during the periods agreed well with the
official target value established, from the continuous monitors.
Fourth, scintillation cell measurements by Argonne National
Laboratory agreed with NAREL measurements in RMP4P within
33
-------
approximately 4 percent (Lu87). Fifth, in the followup GR testing
for RMP4, NAREL filled its own scintillation cells, four cells from
the U.S. DOE Environmental Measurements Laboratory (EML) in New York,
and the participants' cells at the same time. The EML results agreed
with the NAREL results within approximately 3 percent. Based on the
above, EPA believes that the round 4 radon target values were
correct.
B. There could have been a high bias in a chamber used to calibrate
participant CC' s as compared to the NAREL chamber used to expose CC's
during RMP4P (normally differences should be less than 10 percent, as
discussed previously).
EPA understood that another exposure chamber, used to calibrate
several CC participants entered in RMP4P, experienced a period of
time before RMP4 when the reported radon chamber concentration
appeared to have been overestimated by more than 10 percent. This
could have contributed directly to the high bias in RMP4P.
C. CC participants who underestimated in previous RMP rounds could
have adjusted their calibration factors upward, to yield higher
results, without obtaining additional calibration data.
Two round 4 CC suppliers (representing about 20 of the CC
participants) indicated to EPA staff that, based on RMP3 results,
they made an upward adjustment in their calibration factors without
additional calibration data (one by 15 percent and one by 8 percent) .
There may have been other participants who also made an arbitrary
upward adjustment in their calibration factor between rounds 3 and 4.
D. At times, a large loading of CC's going into chamber A caused
the radon concentration to decrease for 18-24 hours. The continuous
monitors used to establish the official chamber A radon concentration
recorded this initial decrease and recovery. If CC's did not measure
the decrease, due to their time-dependent response, this could cause
an overestimation.
34
-------
NAREL staff examined several exposure periods to ascertain the
magnitude of error if participants' CC's failed to sense the
decrease. An example was the CC exposure period from 0645, 4/24/87
to 0749, 4/27/87, the period where the single largest concentration
decrease occurred due to heavy loading of CC's into the chamber. The
decrease started when CC's were loaded into the chamber, reached a
value of 75 percent of the initial concentration within 5-1/2 hours
and was essentially fully recovered within 24 hours. The radon
concentration remained relatively constant during the rest of the
exposure period. For the 3 day CC exposure period, the continuous
radon monitor record yielded an average radon concentration of
37.7 pCi/L. If one hypothesizes that some 3-day CC's only responded
to chamber concentrations during, say, the last two days of the
exposure period (i.e., they "missed the dip"), the continuous monitor
record indicated an average of 39.0 pCi/L for that period. For this
worst case decrease, the bias created by "missing the dip" would be
+3.4 percent.
E. Temperature in the RMP test chambers could have been different
than in the calibration chamber.
The effect of temperature changes on the response of
participants' CC's should have been accounted for in any rigorous CC
calibration program. The chamber temperature during RMP4 exposures
(~70° F) was measured both electronically and with mercury bulb
thermometers, and was stated on each RMP participant's results
reporting form. One of the purposes of the RMP program is to
identify inadequate calibration data.
F. Inadequate calibration of participants' CC's to account for
relative humidity variation.
Similar to temperature, this effect (if present), should be
accounted for in any rigorous CC calibration program. Chamber
relative humidity during RMP4 was ~50 percent.
35
-------
G. Large quantities of CC's in the RMP chambers might have caused
non—uniform radon concentrations across the chamber volume which
could affect participants' CC's exposed at some locations.
As discussed in Section 6.0, Test 5 was run with the NAREL
Chamber A loaded with approximately 42 kg of charcoal, almost twice
the highest charcoal loading encountered in RMP4. The chamber
temperature, relative humidity and air velocity were virtually the
same as during RMP4. EPA studied chamber A concentration uniformity
under these conditions using 3 spatially distributed and
independently calibrated CRN's and using brand 1 CC's. In four
uniformity tests (using the three CRM's for each test), the greatest
standard deviation of the mean concentration (mean for the three
CRM's) was 3.3% and the individual CRM readings ranged from a low of
3.8% below to a high of 2.1% above the mean. All of the CRM
uniformity tests were conducted while chamber A was heavily loaded
with charcoal. For the three concentration uniformity tests run
using CC's, two were with essentially no charcoal loading and one was
with heavy charcoal loading. Average radon concentrations for each
of six equipment racks (to measure horizontal variation) and for each
of four vertical levels was computed. For the three tests, Table 4
shows that the greatest standard deviation of the mean CC
concentration (for the chamber) was 7.2%. Rack averages ranged from
a low of 5.0% below to a high of 4.2% above the mean CC
concentrations. Level averages ranged from a low of 3.0% below to a
high of 3.3% above the mean CC concentration. EPA believes that the
CRM's provided the highest quality data for these tests since the
brand 1 CC data was probably somewhat sensitive to air velocity
variation across the chamber. However, the percent standard
deviations and ranges of variation were low for both CRM's and CC's.
EPA believes that both sets of data demonstrated good concentration
uniformity throughout chamber A.
36
-------
H. There could have been an aerosol present in chamber A which has
a high solubility for Rn-222 and which is, itself, adsorbed more
efficiently by charcoal than Rn-222. Possible sources are:
a. foam insulation used in manufacture of chamber wall
panels,
b. silicon rubber caulk used to seal interior of
chamber,
c. carnauba wax used in CN generator,
d. coating on wire baskets used to expose CC's, AT's
and EP' s.
As stated earlier in the report, the weighted mean bias for the
NAREL brand 1 CC's used in the seven chamber tests (almost 200 CC's
tested) was a positive 4 percent. If items a—d contributed to a CC
bias in RMP4, the effects should have been present during the seven
chamber tests, since the tests were conducted during and immediately
after RMP4. Also, the effect should have been a negative bias
because the organic compounds which might be released from these
materials would compete with the Rn—222 for charcoal adsorption
sites.
I. It is possible that some CC' s were sensitive to the velocity of
the surrounding air.
The hypothesis was that possible differences in air velocities
between the NAREL chamber used to expose CC's in RMP4 and other
calibration chambers, or between different locations within NAREL
chamber A, had a significant effect on the response of CC's. NAREL
staff ran tests to determine the effect of air velocity on the EPA
4-inch open-faced CC's. NAREL has used two brands of CC's, each with
a different type of charcoal, in the EPA program. The calibration
data for both brands of NAREL CC's was generated in chamber C where
the air velocity is 0.1-0.3 feet per minute (essentially stagnant
conditions). During RMP4, air velocities at some locations in
37
-------
chamber A reached ~70 feet per minute*. This did not appear to be
a significant problem in the initial tests in chamber A with brand 1
CC's. For the seven initial tests, the weighted average bias for
these brand 1 CC's was about 4 percent high as previously stated.
This did not show a significant bias shift in brand 1 CC's between
chambers A and C which have vastly different designs. However, when
the brand 2 CC's were tested they showed a significantly higher bias
in chamber A than did brand 1. Additional testing (beyond these
seven tests) was performed which clearly documented the velocity
sensitivity of both brands of CC's and the generally greater
sensitivity to air velocity of brand 2 CC's when compared to brand 1
(Gr88). As an example (discussed previously) , one group of tests
using brand 2 CC's (exposed for two days) yielded very good agreement
between CC's and CRM's when exposed with no air velocity. However,
when velocity was introduced, the CC's experienced a 16% high-bias
for ~50 fpm, a 23% high-bias for ~100 fpm and a 27% high bias for
~200 fpm air velocity. Using this data, NAREL staff concluded that
air velocity in chamber A during round 4 (maximum was ~70 fpm) could
have been a contributing factor to the overresponse of participants'
CC's. It appears that the RMP4 exposed an inherent weakness in the
performance of open—faced CC's; the sensitivity of CC response to air
velocity. Because of this conclusion, NAREL staff believed that RMP
participants using charcoal adsorbers needed to test their devices to
determine the effect of air velocity on response. A summary of the
EPA test results and a recommendation that participants determine the
velocity sensitivity on their CC detectors was published in two
widely circulated newsletters and in a letter sent to the RMP
participants (see Appendix B).
*The velocity measurements in Chamber A showed velocities
between <10 to ~70 fpm for the RMP4 test conditions. Using the
same instrument used to measure the Chamber A air velocities,
velocity measurements were also made in a local residence with a
forced air heating/cooling system. Velocities at likely CC
monitoring locations in the home ranged from <10 up to ~70 fpm,
similar to RMP4 test conditions.
38
-------
8.0 Summary
In this report, two EPA radon and radon decay products test
chambers (chambers A and C) located at the National Air and Radiation
Environmental Laboratory in Montgomery, Alabama are described. These
chambers were used to expose detectors submitted for testing in Round
4 of the National Radon Measurement Proficiency Program and are used
routinely for calibration purposes. Also described are the
measurement and calibration procedures which were used to establish
the official target values for radon and radon decay products
concentrations during BMP Round 4 testing. The results for RMP
Round 3 (conducted at the U.S. DOE Environmental Measurements
Laboratory radon chamber in New York) and RMP Round 4 (conducted in
the two NAREL chambers) are discussed and compared. Following
Round 4, the NAREL staff analyzed the collective performance for each
measurement method tested in these rounds and found that all methods
agreed with the target values within expected limits except for
RPISU's and charcoal adsorbers. The 15 percent underestimation of
the RPISU's was traced to plugging of the filters by wax. When
retested with shorter exposure time, the problem was resolved. As a
group, CC's were biased 16 percent above the target values in Round 4
whereas they were biased 7 percent below the target values in
Round 3.
After analyzing the RMP4 results, NAREL staff spent several
months evaluating the difference in charcoal adsorber response
between Rounds 3 and 4 by performing radon chamber tests using EPA 4-
inch, open-faced charcoal adsorbers. Several potential causes for
the high-bias of CC's in Round 4 were examined. EPA staff believe
that the round 4 high-bias for CC's was caused by the synergistic
combination of several factors which include:
- Participants who were biased low in previous rounds
adjusting their calibration factors upward to yield higher
results without obtaining additional calibration data;
39
-------
- Overestimation of radon concentrations in a chamber used to
calibrate some RMP4 participants' charcoal adsorbers;
- Inadequate humidity corrections for some charcoal
adsorbers; and
perhaps most importantly,
— Some charcoal adsorbers having a response sensitivity which
varied with air velocity within the velocity range found in
chamber A during RMP4.
These factors, and others which were tested and ruled out, are
discussed in Section 7.0 of the report.
40
-------
Fi88a
Fi88b
Fi87
Fi85a
Fi85b
Fi83
Fi81
Ge87
Gr88
References
Fisenne, I.M., George, A.C., and Keller, H.W. (1988), "The
August 1986 Through February 1988 Radon Intercomparison at
EML", EML-516. U.S. Department of Energy, Environmental
Measurements Laboratory, New York.
Fisenne, I.M., and George, A.C., "Results of the Fifteenth
EML Radon Intercomparison (August 8, 1988)", Environmental
Measurements Laboratory, personal communication to
E.L. Sensintaffar, U.S. EPA, dated September 2, 1988.
Fisenne, I.M. George, A.C., and Keller, H.W., 1987, "The
July 1985 and March 1986 Radon Intercomparisons at EML",
Environmental Measurements Laboratory Report, EML-479, U.S.
Department of Energy.
Fisenne, I.M., George, A.C. and Keller, H.W., 1985, "The
July 1984 and February 1985 Radon Intercomparisons",
Environmental Measurements Laboratory Report, EML-445, U.S.
Department of Energy.
Fisenne, I.M., George, A.C., and Keller, H.W., 1985, "Radon
Intercomparisons at EML, January 1983 and February 1984",
Environmental Measurements Laboratory Report, EML—436, U.S.
Department of Energy.
Fisenne, I..M., George, A.C. and Keller, H.W., 1983, "The
1982 Radon Intercomparison Exercises at EML", Environmental
Measurements Laboratory Report, EML-413, U.S. Department of
Energy.
Fisenne, I.M., George, A.C., and McGahan, M., 1981, "Radon
Measurement Intercomparisons at EML", Environmental
Measurements Laboratory Report, EML-397, U.S. Department of
Energy.
George, A.C. and Tu, K.W., (draft), "Intercomparison of
Radon Daughter Measurement Methods and Equipment in
North America", (March 23-26, 1987), Environmental
Measurements Laboratory Report to be published, U.S.
Department of Energy.
Gray, D.J. and Windham, S.T. (1988), "The Overresponse of
Open—Faced Charcoal Adsorbers Used for Measurements of
Indoor Radon Concentrations", poster paper at "The
41
-------
Lu87
Lu57
Pe87
Pe86
Th72
Tu81
USBOM82
USEPA87
References - Continued
1988 Symposium on Radon and Radon Reduction Technology",
sponsored by U.S. Environmental Protection Agency, Air and
Energy Engineering Research Laboratory and Office of
Radiation Programs, October 17-21, 1988, Denver, CO.
Personal communication to J.M. Smith, USEPA, from
H.F. Lucas, Argonne National Laboratory, October 14, 1987.
Lucas, H.F., 1957, "Improved Low-Level Alpha Scintillation
Counter for Radon", Review of Scientific Instruments,
Vol. 28, p. 680.
Pearson, M.D., 1987, "Radon-Daughter Grab Sampling
Technical Exchange Meeting 14-17 April 1986", Technical
Measurements Center Report DOE/ID/12584-12, U.S. Department
of Energy.
Pearson, M.D., 1986, "Interlaboratory Radon-Daughter
Measurement Comparison Workshop: 9—12 September, 1985".
Technical Measurements Center Report, GJ/TMC-25 US-70A,
U.S. Department of Energy.
Thomas, J.W., 1972, "Measurements of Radon Daughters in
Air", Health Physics, Vol. 23, p. 783.
Tu, K.W., 1981, "A Condensation Aerosol Generator System
for Monodisperse Aerosols of Different Physicochemical
Properties", Journal of Aerosol Science, Vol. 13, No. 5,
pp. 363-371.
US BOM, 1982, "Second Interlaboratory Comparison Test for
Measuring Radon Daughters", U.S. Bureau of Mines, Denver
Research Center, Denver, CO with letter from
Robert F. Holub, US BOM, to E.L. Sensintaffar, U.S. EPA,
dated August 13, 1982, correcting Tables 4 and 8.
U.S. EPA, 1987, "Radon/Radon Progeny Analytical Proficiency
Report (Round 4 Performance and Followup Test)", Research
Triangle Institute, Research Triangle Park, NC.
42
-------
Appendix A
Histograms for RMP4 Performance Round
(Participant Performance Ratio)
Note: A detector performance ratio is the quotient of a
participant's detector reading divided by the official
target value. A participant performance ratio is the
average of the detector performance ratios for all
detectors tested by a participant. The method performance
ratio is the mean of the participants performance ratios
for all participants testing this method, except for
outliers. An outlier is a participant performance ratio
below 0.5 or above 1.5. The histograms were constructed
using participant performance ratios.
A-l
-------
n
-------
7
uo
m
0)
o
I
3
O
V
.a
Charcoal Adsorber Histogram
RMP 4 Performance Round
rfi
ifl
111111111 11111 11111 11 ill 1111111111 11 ill I
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 2.75 3.00 3.25 3.50 3.75 4.00
Performance Ratio
-------
Continuous Radon Histogram
RMP 4 Performance Round
4 H
0)
j ~"
u
o
T
Fail
1 -
IT I II I I I I
0.00 0.25 0.50 0.75 1.00 1.25
11111111
T I I M I I I I M II I II II
1.75 2.00 2.25 2.50 2.75 3.00 3.25 3.50 3.75 4.00
Performance Ratio
-------
Fail
T
ui
o
c
!
o
O
Grab Radon Histogram
4 Performance Round
':' >: i :: I ; I i t, I I I i >: i I i ! •' I I I I I ;, I I i 111 ITT I I I ITTlT I I 1 I
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 2.75 3.00 3.25 3.50 3.75 4.00
Performance Ratio
-------
0)
8
U
JD
2.8 -
2.6 -
2.4 -
2.2 -
2 -
1.8 -
1.6 -
1.4 -
1.2 -
1 —
0.8 -
0.6 -
0.4 -
0.2 -
E—Perm Histogram
RMP 4 Performance Round
Fail
Pas«
111111111111111 M i iff im
Fail
TTH I I I I I rTTTTTTI I I I I M I I I I I I I I I I IT I I I I I I I I I I I I I l"| I I I I I
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 2.75 3.00 3.25 3.50 3.75 4.00
Performance Ratio
-------
W
8
u
3
Grab Working Level Histogram
RMP 4 Performance Round
5 i
4 -i
3 -
Fail
2 H
1 -
Pass
Fail
i i
MM
I I I I I I I I I I I I'lTTTTTTTTTTTT I I I I I I I I ITI I I I I I I I I I I I I I I 1 I I I ! I I I I I I I I I I I I I I I I
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 2.75 3.00 3.25 3.50 3.75 4.00
Performance Ratio
-------
RPISU Histogram
RMP 4 Performance Round
^ •
1.9 -
1.8 -
1.7 -
1.6 -
1.5 -
1.4-
§ 1.3 -
g 1.2 -
o 1.1 -
o 1 -
o o.9 -
jj 0.8 -1
Fail
| 0.7 -
Z 0.6-j
0.5 -j
0.4 -1
0.3-
0.2 -
0.1 -
0 -
0.
0
rm 1 1 1 M F i TTT
0 0.25 0.50 0.
;
:
Pass
1
1
J
75
Ln,
1.00 1.:
Fail
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 M 1 1 1 1 1 1 1 1 1 1 1 1 I 1 1 1 1 1 1 1 1 1 1 1 1 1 I 1 1 1 1 1 1 1 M M
25 1.50 1.75 2.00 2.25 2.50 2.75 3.00 3.25 3.50 3.75 4.00
Performance Ratio
-------
S
a
10
9 -
8 -
7 -
6 -
5 -
4 -
3 -
1 -
Fail
I IT
I II
in r i
Continuous Working Level Histogram
RMP 4 Performance Round
Pass Fail
I 1 II I I I rTTTTTTiTl
IT II I I
TTTTTTTl
ITTT7 H M I i T IT
0.00 a25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 2.75 3.00 3.25 3.50 3.75 4.00
Performance Ratio
-------
Appendix B
Notifications of Velocity Sensitivity
of Openfaced Charcoal Adsorbers
Note: Prior to March of 1990, the National Air and Radiation
Environmental Laboratory was known as the Eastern
Environmental Radiation Facility.
B-l
-------
The discovery, during RMP round 4, of the open—faced charcoal
adsorber response sensitivity to air velocity was communicated to the
radon measurement community in the three documents included in this
appendix:
• Article in Radon Reporter (AARST), Volume 1, Number 2,
Spring 1988.
• Note, Health Physics Society Newsletter, Volume XVI, Number
2, February 1988.
• Letter to RMP Participants from J. Michael Smith, EPA,
February 25, 1988.
B-2
-------
EPA SUGGESTS CHARCOAL CANISTER DESIGN CHANGE
Dear Editor
Round 4 of the Environmental Protection
Agency's (EPA) Radon Measurement Pro-
ficiency (RMP) Program was conducted at
EPA's Eastern Environmental Radiation
Facility (EERF) in Montgomery. AL during
April-June 1987. During Round 4 the EERF
tested four radon measurement tech-
niques (charcoal adsorbers, alpha track
detectors, grab radon monitors, continu-
ous radon monitors) and three radon pro-
geny measurement techniques (RPISU's.
grab radon progeny monitors and contin-
uous radon progeny monitors). The total
number of methods tested was approxi-
mately 500 in the performance test Fol-
lowing Round 4, EERF staff statistically
analyzed the collective performance for
each of these seven measurement tech-
niques and found that all techniques
agreed with our known values within
expected limits except for charcoal
adsorbers. As a group, charcoal
adsorbers were biased 16 percent above
the known values in the performance test
This was puzzling since, in previous
rounds (conducted in a different
chamber), charcoal adsorbers were
biased 6-8 percent below the known
values.
Since the conclusion of Round 4, EERF
staff have spent several months evaluat-
ing this shift in charcoal adsorber
response by performing chamber tests
using our own charcoal adsorbers. We
have examined several potential radon
chamber problems (for example, the pos-
sibility of non-uniform distribution of
radon in our chamber under conditions of
heavy charcoal loading, etc.) and have
demonstrated that none of these potential
problems were present in our chamber
during Round 4 to adversely affect char-
coal adsorber response. We believe that
the 16 percent high bias of charcoal
adsorbers during the performance test in
Round 4 and the shift from a negative to
positive bias between Rounds 3 and 4 is
the synergistic combination of several fac-
tors which may include: participants arbi-
trarily adjusting calibration factors; bias in
a chamber used to calibrate participant's
charcoal adsorbers; inadequate
calibration for differing humidities for cer-
tain charcoal adsorbers; and perhaps
most importantly, certain charcoal
adsorbers having a response sensitivity
which varies with air velocity. These fac-
tors and others which we tested and ruled
out will be discussed in more detail in a
forthcoming EPA report
We believe that the velocity effect is not
well recognized and is of particular impor-
tance to the users of charcoal adsorbers.
During our charcoal adsorber testing in
the past few months, we have obtained
data which indicates that, at constant
temperature, relative humidity and radon
concentration, the air velocities in the vic-
inity of open-faced charcoal adsorbers of
the EPA design significantly affects the
response. For example, in one series of
2-day tests, charcoal adsorbers exposed
under dormant air conditions agreed
within 5 percent with our continuous
radon monitors (most of the time they
agreed within 1 percent); at a velocity of
-50 feet per minute the same type
adsorbers were biased 16 percent high; a
velocity of —100 feet per minute, 23 per-
cent high; and at -200 feet per minute, 27
percent high. Our more recent testing
indicates that adsorbers with diffusion
barriers may exhibit a sensitivity to veloc-
ity, although less pronounced than for
open-faced adsorbers.
During Round 4, air velocities at some
locations within the test chamber were of
the same magnitude as found at some
suitable monitoring locations in residen-
ces with forced air heating/cooling sys-
tems. Since Round 4 and our discovery of
the velocity sensitivity of charcoal
adsorbers, steps have been taken to signif-
icantly reduce the air velocity in EERF's
test chamber used for the RMP Program.
We believe that organizations using
charcoal adsorbers should test their devi-
ces to determine the effect of air velocity
on their charcoal adsorber response. If a
sensitivity to air flow is noted, these organ-
izations may wish to modify their design to
obtain charcoal adsorbers which are less
sensitive to velocity effects. The EPA is
currently testing various detector designs
to find a configuration that minimizes
humidity and air velocity effects while
maintaining adequate sensitivity. Those
organizations who choose not to modify
velocity-sensitive charcoal adsorbers
should take particular precautions to
assure that their adsorbers are exposed in
the field in accordance with EPA protocols
which call for testing in areas with minimal
drafts. However, we believe that the limita-
tions of assuring no air movement in either
field or chamber testing is unrealistic.
Therefore, EPA believes that the proper
solution to this problem is to modify the
design of charcoal adsorbers so they are
less sensitive to air velocity.
We hope that the information in this let-
ter will be helpful to organizations measur-
ing radon levels using charcoal adsorbers.
The EERF staff would be interested in
learning the results of testing conducted
by other organizations regarding the
effects of air velocity on their charcoal
adsorber response.
Charles R. Porter, Director
Eastern Environmental
Radiation Facility
3-3
-------
Problem with Charcoal Adsorbers
Round 4 of the EPA Radon Measurements Pro-
ficiency Program (RMP) was conducted at EPA's
Eastern Environmental Radiation Facility (EERF)
in Montgomery, Alabama, during April-June 1987.
During Round 4 the EERF tested five radon mea-
surement techniques (charcoal adsorbers, alpha
track detectors, electret-perms, grab radon mon-
itors, continuous radon monitors) and three radon
progeny measurement techniques (RPISUs, grab
radon progeny monitors and continuous radon
progeny monitors). The total number of methods
tested was approximately 500 in the performance
test. Following Round 4, EERF staff statistically
analyzed the collective performance for each of
these eight measurement techniques and found
that all techniques agreed with our known values
within expected limits except for charcoal adsor-
bers. As a group, charcoal adsorbers were
biased 16 percent above the known values in the
performance test. This was puzzling since, in
previous rounds (conducted in a different cham-
ber), charcoal adsorbers were biased 6-8 percent
below the known values.
Since the conclusion of Round 4, EERF staff
have spent several months evaluating this shift
in charcoal adsorber response by performing
chamber tests using our own charcoal adsorbers.
We have examined several potential radon chamber
problems (for example, the possibility of non-
uniform distribution of radon in our chamber
under conditions of heavy charcoal loading, etc.)
and have demonstrated that none of these poten-
tial problems was present in our chamber during
Round 4 to adversely affect charcoal adsorber
response. We believe that the 16 percent high
bias of charcoal adsorbers during the performance
test in Round 4 and the shift from a negative to
positive bias between Rounds 3 and 4 is the syn-
ergistic combination of several factors which will
be described in a forthcoming EPA report and/or
article to be submitted to the Health Physics
Journal.
We believe that one of these factors is not well
recognized and is of particular importance to the
users of open-faced (no diffusion barrier) char-
coal adsorbers. During our charcoal adsorber
testing in the past few weeks, we have obtained
data that indicates that, at a constant tempera-
ture, relative humidity and radon concentration,
the air velocities in the vicinity of certain open-
faced charcoal adsorbers significantly affect the
response. For example, in one series of two-day
tests, charcoal adsorbers exposed under dormant
air conditions agreed within five percent with our
continuous radon monitors (most of the time they
agreed within one percent); at a velocity of - 50
feet per minute the same type adsorbers were
biased 16 percent high; at a velocity of - 100
feet per minute, 23 percent high; and at - 200
feet per minute, 27 percent high. A member of
our staff measured air velocities of - 50 feet per
minute or more at points in his own home where
charcoal canisters could be placed.
We believe that organizations performing radon
tests, at least those using open-faced devices
where the charcoal is exposed directly to the air
being monitored, should begin testing their de-
vices to determine the effect of air velocity on
their charcoal adsorber response. If a sensitivity
to air flow is noted, continued home testing using
these devices should be done in strict confor-
mance with EPA protocols which require testing
in areas with minimal or no drafts.
At the time of this writing, we are preparing
a letter, discussing our preliminary findings, for
transmitted to RMP charcoal adsorber participants.
We have already discussed, informally, our find-
ings with officials of the American Association of
Radon Scientists and Technologists, Inc. We will
appreciate your publishing this note in the Health
Physics Newsletter to alert other interested
parties.
Charles R. Porter
Director, EERF
Montgomery, Alabama
B-4
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
\ OFFICE OF RADIATION PROGRAMS
g Eastern Environmental Radiation Facility
* 1890 Federal Drive. Montgomery, AL 36109
(205) 272-3402 • FTS 534-7615
February 25, 1988
Dear RMP Participant:
Round 4 of the EPA Radon Measurements Proficiency Program (RMP) was
conducted at EPA's Eastern Environmental Radiation Facility (EERF) 1n
Montgomery, AL during April-June 1987. During Round 4 the EERF tested 4
radon measurement techniques (charcoal adsorbers, alpha track detectors,
grab radon monitors, continuous radon monitors) and 3 radon progeny
measurement techniques (RPISU's, grab radon progeny monitors and
continuous radon progeny monitors). The total number of methods tested
was approximately 500 in the performance test. Following Round 4, EERF
staff statistically analyzed the collective performance for each of these
seven measurement techniques and found that all techniques agreed with our
known values within expected limits except for charcoal adsorbers. As a
group, charcoal adsorbers were biased 16 percent above the known values in
the performance test. This was puzzling since, in previous rounds
(conducted in a different chamber), charcoal adsorbers were biased 6-8
percent below the known values.
Since the conclusion of Round 4, EERF staff have spent several months
evaluating this shift in charcoal adsorber response by performing chamber
tests using our own charcoal adsorbers. We have examined several potential
radon chamber problems (for example, the possibility of non-uniform
distribution of radon in our chamber under conditions of heavy charcoal
loading, etc.) and have demonstrated that none of these potential problems
were present in our chamber during Round 4 to adversely affect charcoal
adsorber response. We believe that the 16 percent high bias of charcoal
adsorbers during the performance test In Round 4 and the shift from a
negative to positive bias between Rounds 3 and 4 is the synergistic
combination of several factors which may include: participants
arbitrarily adjusting calibration factors; bias In a chamber used to
calibrate participant's charcoal adsorbers; inadequate calibration for
differing humidities for certain charcoal adsorbers; and perhaps most
importantly, certain charcoal adsorbers having a response sensitivity
which varies with air velocity. These factors and others which we tested
and ruled out will be discussed In more detail in a forthcoming EPA report.
We believe that the velocity effect 1s not well recognized and 1s of
particular importance to the users of open-faced (no diffusion barrier)
charcoal adsorbers. During our charcoal adsorber testing In the past few
weeks, we have obtained data which Indicates that, at constant temperature,
relative humidity and radon concentration, the air velocities In the
vicinity of open-faced charcoal adsorbers of the EPA design significantly
B-5
-------
affect the response. For example, In one series of 2-day tests, charcoal
adsorbers exposed under dormant air conditions agreed within 5 percent
with our continuous radon monitors (most of the time they agreed within
1 percent); at a velocity of ~ 50 feet per minute the same type adsorbers
were biased 16 percent high; at a velocity of - 100 feet per minute, 23
percent high; and at ~ 200 feet per minute, 27 percent high. During RMP
Round 4,.air velocities at some locations within EERF's big radon chamber
were on the order of 50 feet per minute. EPA has determined that air
velocities of 50 feet per minute or more occur in homes with forced air
heating and/or cooling systems.
We believe that organizations performing charcoal adsorber tests,
at least those using open-faced devices similar to the EPA design where
the charcoal is exposed directly to the air being monitored, should
begin testing their devices to determine the effect of air velocity on
their charcoal adsorber response. If a sensitivity to air flow is
noted, continued home testing using these devices should be done in
strict conformance with EPA protocols which require testing in areas with
minimal or no ararts.
We hope that the information In this letter will be helpful to RMP
participants enrolled to test charcoal adsorbers. The EERF staff would be
Interested in learning the results of testing conducted by other
organizations regarding the effects of air velocity on their charcoal
adsorber response.
Sincerely,
J. Michael Smith, P.E.
RMP Laboratory Coordinator
Eastern Environmental Radiation Facility
------- |