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

-------