-------
20
18
ie
14
K,
§ 10
3
jF
^ 8
6'
4'
2
O-
CN

rO
ro
o»
CN
U">
CO
ad
J
^¦919
9I9I?
£i9i9l9
lllllllll
OO OO
P-» CO
lO xri
in
9 0
9 9
9 9 9
III
III
III
9 9 9
III
III
9 9 9
Ml
999
^99
9 9^
g ^ ^
o»
CO
io
o»
i^»
o>
?¦?
919
919
919
919
919
919
919
919
9l9B~
9I9I9
9I9I9
9l9l9i^
9l9l9l9H»
9I9I9I9I9
a>
ro
CM
co *
M
3
e
«
•P
cO
a
p-
w
a
a>
3)
f*
PM
CO
4J
C
•w
o
CM
tn
O
C
o
a>
en
c«
cQ
O
Pu
H
A
3
cd
4J
cO
Q
4J
W
(1)
H
VO

-------
and s, the estimate of the common population standard deviation is defined as
/ (n, - l^2 + (n2 - l)s22
yj	n1 - n2 - 2
Inserting data from figure 1 and figure 2 in equation 1 we calculate t - -2.13.
What does the t-statistic mean? Recall that we are trying to test the two possible hypothese
(described on page 3) that the populations means of the Cm and the Ct data are either the same S
not the same (one data sample is significantly less than or greater than the other).	°r
Suppose we wanted to be 95% sure of which hypothesis is correct. That is to say we want
be 95% certain that the populations of the two curves, Cm and Ct, are either the same
significantly lower or higher. The Student's t distribution provides the so-called crit" °i
values (essentially the area to the left of the critical value--see figure 3 [5]) needed for th
t test, for various degrees of freedom1 --based on the number of samples (in this case n + 8
2) and the level of significance.	1 2
The area (critical value) represents the region of certainty. We want to be 95% certain- -i-v>•
criteria is often referred to as a 5% level of significance. Since our hypothesis is not tS
sided (such a test would entail testing that one population mean is greater than or less than0^'
other) but two-sided (tests that one population mean is either greater than or equal to v.C
other). This means that we would look under a level of significance of 0.05/2 or 0.025. We
to be 95% certain that we are not either too high or low--thus the 95% would be sandwiched bet^"6
two tails which total 5%.	Ween
Now we are comparing the curves in figure 1 and figure 2. The number of degrees of fr
is 5139 + 1052 - 2 or 6189. From the t-statistic tables, (6) the critical value at the
level of significance is 1.96 (taking the number of degrees of freedom as infinity). Since t-v>25
is a two-sided test, we would accept the null hypothesis (HQ) that the sample means of the ma
curve and the test curve are the same if the test statistic is between -1.96 and +1.96. Our Ster
statistic was -2.13.	test
-2.13 < -1.96 (the critical value)
The test statistic is less than the critical value, thus we reject H0 and accept the alte
hypothesis (H )--the two population means are indeed significantly different.	"ate
There is another comparison of two samples--the F-test (one-way analysis of variance)
F-test is used when samples are independent, normally distributed and possess equal sta .The
deviations. The test statistic is defined as:	^dard
Hhe term degrees of freedom "... is the number of independent measurements available f
estimating (the sample standard deviation)..." If one adds the difference X - X (refer °j
to as a residual) for each value of X the result is zero. If we know the sum of n - 1 resltju f
the other residual can be determined. Thus, "...there would be only n - 1 "independent"	iS
use to compute the sample standard deviation (4)
v*lues
7

-------
Area to the left of the critical value
O
3
PC
a>
t-h
m
3
O

T3
05
OQ
(t>
in
CO
Ui
*
OJ
Ul
-J
(W
£
»-(
(D
c
a
(D
3
ri-
ft
cn
rt
•I
K*
V
C
o
p
Student's t distribution
The table gives the critical values of t needed for the / test, for various
degrees of freedom and levels of significance.
Example I. Suppose the test is one-sided and rejects for large val-
ues of t, with level of significance .05. Then the critical value is in the
column headed .95, on the row corresponding to the appropriate
degrees of freedom.
Example 2. Suppose the test is two-sided, and has level of signifi-
cance .01. Then the critical values are in the columns headed .005 and
.995, on the row corresponding to the appropriate degrees of freedom.
If there are 10 degrees of freedom, the test would reject if the observed
value of / is less than —3.17 or greater than 3.17.
Critical
value
Values of t
LJt i^l I t J
»f freedom

.005
.01
.025
.05
1
-
63.66
-31.82
- 12.71
-6.31
2
-
9.92
- 6 96
- 4.30
-2.92
3
-
5.84
- 4.54
- 3.18
-2.35
4
-
4.60
- 3.75
- 2.78
-2.13
5
-
4.03
- 3.36
- 2.57
-2.02
6
-
3.71
- 3.14
- 2.45
— 1.94
7
-
3.50
- 3.00
- 2.36
- 1.90
8
-
3.36
- 2.90
- 2.31
-1.86
9
-
3.25
- 2.82
- 2.26
-1.83
10
-
3.17
- 2.76
- 2.23
-1.81
11
-
3.1 1
- 2.72
- 2.20
-1.80
12
-
3.06
- 2.68
- 2.18
-1.78
13
-
3 01
- 2.65
- 2.16
-1.77
14
-
2.98
- 2.62
- 2.14
1.76
15
-
2.95
- 2.60
-2.13
-1.75
16
-
2.92
- 2.58
- 2.12
-1.75
17
-
2.90
- 2.57
- 2.1 1
-1.74
18
-
2.88
- 2.55
- 2.10
-1.73
19
-
2.86
- 2.54
- 2.09
-1.73
20
-
2.84
- 2.53
- 2.09
-1.72
21
-
2.83
- 2.52
- 2.08
-1.72
22
-
2.82
- 2.51
- 2 07
-1.72
23
-
2.81
- 2.50
- 2.07
-1.71
24
-
2.80
- 2.49
- 2.06
-1.71
25
-
2.79
- 2.48
- 2.06
-1.71
26
-
2.78
- 2.48
- 2.06
-1.71
27
-
2.77
- 2.47
- 2.05
1.70
28
-
2.76
- 2.47
- 2.05
-1.70
29
-
2.76
- 2.46
- 2.04
-1.70
30
-
2.75
- 2.46
- 2.04
-1.70
40
-
2.70
- 2.42
- 2.02
- 1.68
50
-
2.68
- 2.40
- 2.01
-1.68
60
-
2.66
- 2.39
- 2.00
-1.67
80
-
2.64
- 2.37
- 199
-1.66
100
-
2.63
- 2.36
- 1 98
-1.66

-------
where
(n1 - 1)+ (n2 - l)s22
s ^
3 .	n,(X, - *>' - ^
s.
where
t^X, + n2X2
n, + n2
r>f freedom parameters.
there would two degree of free
1. \ "l
2.	-i2 - n1 + n2 - 2
For this example, the F-statistic is 4.52, with ^ « 1 and - 6189. From the f-stan
' - —fi>al value is 5.02.	sttc
For this example, cne 	
tables, (7) the critical value is 5.02.
4.52 < 5.02 (the critical value)
, co < 5 02 (the critical
t ~	the population means are not
U critical value, therefore, the V
tatistic is *-ess than	crltlca
Ilminca"tly di££erent'	. . „ oE the null hypothesis (H„) while the F-
S	.	led to a rejection ot	radon measurements from Firm
,-he T-statistics lea t	hYpothesis (H0) . Thus r
in summary, tW acceptarice of the null nyp
statistic	d a borderline case.	obtained from the confirmatory
A would be deemed	^ ^ ^ ^asure.ehts^btaxne	1
FiRure 4 is the test curve ( t Tbe arithmetic mean i	"	for tMs tw0.sided paired
dJTe (grater than * P^>."tlc is	the t-statlstlc Is between -1.96
^ ^ to 8 p
",4+1'M"	1U6>-1.« ^he critical value,
' h nothesis that the population means are e,ua .
thus we accept the y	2
.	1 25--the critical value
The f-statistic is 1.'
r AO
1.25 < o.	^ Here bQth stattstics confirm
. . the hypothesis that the P^^T^Ld	as providing the DEP with .CCUIK,
thus, confirming the 3T> ^ chls Eltm would
the null hypot o	rad0n test database for Firm C's
re (c ) derived from the volun	y t ls 5 86. Therefore, the
—eent5s	^ t"StatiStlC U betWe6n *
measurements.	that we will accep
t-statistic is -1.33 v
1 96 and +1.96).
.!.55 > .1.96 (the critical value)

-------
1B--
16--
14"»
12- ¦
10- •
8- ¦
6- ¦
4--
2--
m
ro
o o
—I	~—
Ol
,
PpliP
gillllll!
CM
r-»
cs
in
ro
nMflpp
r->
l"">
|
o
• v
CO /J
§
Hi
en
o>
o
ill
2M2M2U9M9.
-18 -14 -10 -6-2 2	6
-16 -12 -8 -4 0	4	8
Difference (pCi/L)
CV|
o>
fO
ro
p
111
lO
to
10
14
o
—i—
18
o
—i—
cfl
¦U
«d
Q
4J
W
0)
H
C
o
X)
«8
OS
ca
3
C
(0
4-1
cd
a
PL.
W
Q

•W
Pt4
12
16
20

-------

IT)
10
CM
cn
ao
a>
o
rri
a>

CM
i
ill
8 -14 -10 -6 -2
-16 -12 -8 -4
6
8
Difference (pCl/L)
<0
•U
cO
a
¦u
w

D)
o
ta
3
C
QO
»0
vo
ro
cs
ro
IP-,
Hill
ro
i fcfl fcg ryi
10
14
CD
—I—
18
12
16
9
20
at
¦u
R)
Q
CL,
M
Q
a>
>-l
3
oo
fa

-------
thus we accept the hypothesis that the population means are equal.
The f-statistic is 2.41--the critical value is 5.02
therefore, confirming the hypothesis that the population means are statistically equal. This
means that the testing firm's readings are acceptable.
Figure 6 is derived from 241 data points each arbitrarily assigned a radon level of 16 pCi/L
(this represents a spiked elevated radon sample). The arithmetic mean is -8.2 and the standard
deviation is 4.27. Therefore, the t-statistic is 19.51
19.51 > 1.96 (the critical value)
thus we reject the hypothesis that the population means are equal (the hypothesis would be
accepted if the t-statistic is between -1.96 and +1.96).
The f-statistic is 380.45--the critical value is 5.02
confirming the hypothesis that the population means are not statistically equal. This means that
this testing firm's readings would not be acceptable.
If both the T-statistic and the F-statistic lead to a rejection of the null hypothesis (HQ)
then the DEP would conduct a canister test at the same time as the testing firm to confirm the
firm's results for several months.
The acceptance criteria for testing firm results can be summarized as follows:
T-Statistic F-Statistic	Decision
1.	Reject	Reject	DEP must conduct side by side-by-side canister tests
with firm for several months
2.	Reject	Accept	Borderline case. If firm
Accept	Reject	is borderline for three (3) consecutive months,
perform side-by-side testing.
3.	Accept	Accept	Testing firms results are deemed accurate
The work described in this paper was not funded by the U.S. Environmental Protection Agency
and therefore the contents do not necessarily reflect the views of the Agency and no official
endorsement should be inferred.
12

-------
30i
20- -
20 ¦ ¦

g w-
lO—
a-.
o>
to
id
o
—t—
10



III
lin
CM
W
ro
0
01
o>
oi
o
—1
6
18 —14 —10 —6 —2 2
-16 -12 -8-4 0 4
Difference (pCi/L)
000000
1	>	1 1	> ¦ '»
10
14
18
12
16
20
CO
3
W)
o
CQ
X.
ao
01
3
C
n)
¦u
<8
Q
a,
w
Q
vo
-------
REFERENCES
1.	Orkin, M. and Drogin, R., Vital Statistics. McGraw-Hill Book Company, c. 1975, p. 65-66.
2.	CDM (Camp Dresser & McKee, Inc.), Summary Report: Scientific Study of Radon. April 1989, p.
204.
3.	Orkin, M. and Drogin, R., Op. Cit.. p. 146.
4.	Ingelfinger, J. A., et. al., Biostatistlcs in Clinical Medicine. MacMillian Publishing
Company, Inc., 1st edition, p. 103.
5.	Orkin, M. and Drogin, R., Op. Clt.. pp. 356-357.
6.	Zar, J. L., Biostatistical Analysis. Prentice-Hall, c. 1974, table D.10.
7.	Zar, J. L., Op. Cit.. table D.ll.
14
-------
Ill—14
TITLE: Grab Sampling as a Method of Discovering Test Interference
AUTHOR: Marvin Goldstein, Building Inspection Service, Inc.
This paper was not received in time to be included in the
preprints so only the abstract has been included. Please check
your registration packet for a complete copy of the paper.
The Abstract is on the following pages.
-------
WHAT HAPPENS WHEN YOU DO FOUR-HUNDRED AND SEVENTY-SEVEN (477)
RADON INSPECTIONS PRECEEDED BY GRAB SAMPLES?
by: Marvin 'D* Goldstein
President-Elect
Pennsylvania AARST
Co-Chairman,
National AARST Legislative Committee
Southampton, PA 18966
ABSTRACT
-------
After compiling the data obtained from conducting 477 radon
inspections, over a four-year period, it can be concluded that performing
a one-hour working level monitor test is useful in determining that
two to seven day E-Perm or charcoal canister tests have not been tampered
with. Duplicate readings increase the confidence that the readings
are accurate.
Of the 477 tests conducted from October, 1987 through June 1990, 107
tests had high readings (above .02 WL or 4pCi/l). Of those 107, 15
one hour Thomson Nielson Monitor Tests revealed levels above .02 WL
when the longer term test had provided readings below 4 pCi/1.
These figures indicate that tampering occured approximately
fifteen percent (15%) of the time when I found high readings.
The working level monitor tests served as an indicator of possible
tampering when compared to the E-Perm or charcoal reading.
Attached to this report is a complete list of all 477 readings.
Six of the fifteen discrepancies were followed up after settlement
and, in every case, the one hour working level monitor result, was
confirmed indicating elevated radon levels.
The following examples illustrate tampering that occured with
multi day tests entrusted to sellers or their agents.
An expensive new house, upon first inspection, had a 1 hour
working level monitor reading of .65. This high reading made the
buyers suspicious and on two occassions they observed all of the
windows were open in the house during the 2 day test period. The
E-Perm reading was 3.0 pCi/1 from this test in the basement.
After the buyers moved in, a follow-up test revealed between
80 and 90 pCi/1 in the home.
At an attorney's house, the realtor tried to talk me out of
conducting a radon test. I insisted that I was hired to do the testing.
The buyer also said that he wouldn't buy without a radon test. The
realtor allowed the testing.
The home measured .08 WL with a one hour working level monitor.
The 2 day test showed between 2 and 2.5 pCi/1. After a $2000.00
escrow account was set up at settlement, a retest showed a reading
of 20 pCi/1.
In another case, I asked the realtor if the home had ever been
inspected before. The realtor said the house had been tested and the
result was a low radon reading.
My first 1 hour working level monitor reading was .05 WL. The
E-Perm test showed 3.5 pCi/1.
-------
When the seller was asked if Che house had been inspected previously,
he responded that the basement and the first floor had been tested.
The readings showed 6.0 pCi/1 in the basement and 2.0 pCi/1 on the
first floor.
My retest showed a basement reading of 6.5 pCi/1 and a first
floor reading of 3.5 pCi/1.
In these examples we can see how grab sampling served effectively
as an early warning system to help detect tampering.
The work described in this paper was not funded by the U.S.
Environmental Protection Agency and therefore the contents do not
necessarily reflect the views of the Agency and no official endorsement
should be inferred.
-------
Ill—15
EXPLORING SOFTWARE DEVICE MANAGEMENT ROUTINES THAT ENSURE:
THE OVERALL QUALITY OF CONTINUOUS WORKING LEVEL AND
CONTINUOUS RADON MONITOR PERFORMANCE IN A FIELD ENVIRONMENT
by:	Richard Tucker
Gemini Research
Cockeysville, MD 21030
Rick Holland
Radonics, Inc.
McLean, VA 22102
ABSTRACT
Various software modules will be discussed that demonstrate how field
measurement problems can be identified for quality assurance purposes prior t
reporting results from actual measurements. Items such as device performance"3
measurement to measurement variances, equipment failure history, and other '
measurement device software tools will be examined.
-------
SOFTWARE MANAGEMENT ROUTINES
It is not difficult to argue that the advent of the micro-processor has
had a profound impact on American business. Computers have changed our lives.
A single micro-computer now allows one person to do what five people did
manually only a few years ago. The efficient and error-free accumulation,
sorting, and processing of massive amounts of data allows American business to
review that data in ways that were once not thought possible. This, of
course, is accomplished using computer software routines.
In the field of radon measurements, automation provides a very powerful
operating advantage. Well-designed software routines are a critical component
of this advantage. In any high volume transaction environment, heavy reliance
on computers is the only practical way to ensure that jobs are handled timely
and efficiently and not confused with other jobs. This is all the more true
in the radon industry. Automation minimizes human error. Automation of any
part of the process associated with performing a radon test (order taking,
scheduling, testing, quality assurance, and reporting) eliminates errors in
that part of the process.
Automating the interface between any two of these independent processes
reduces the possibility of errors even further. Completely automating all of
these processes, as well as the interface between each, reduces possible
errors to their absolute lowest levels: human intervention in the entire
process is virtually non-existent. In this scenario, a completely integrated
system exists. This provides an even greater advantage; a synergistic
advantage. Because the entire process, from order taking to reporting, is
automated (including the smallest component of each process) the entire system
lends itself to detailed computer analysis.
Practically speaking, software routines are a brilliant means for
analyzing data. In a completely automated system, software allows for the
highest levels of quality assurance. Well written programs can easily spot
problems that would not otherwise be obvious. This ensures a test is valid.
Such assurance is nothing less than critical in a real estate-related radon
test where real-time (immediate) quality assurance reviews are absolutely
necessary before results are reported.
Considering that the outcome of a legally binding contract is dependent
on the results of the radon test, errors in the accuracy or validity of that
test which are discovered after results are reported are completely
unsatisfactory.
1
-------
RADON MEASUREMENT EQUIPMENT AND AUTOMATION
Software routines are not possible without an instrument that generates
information in sufficient quantity and detail to allow a computer to analyze
that information. Additionally, software routines are not possible if the
instrument does not have a means for transferring that information into
"machine readable" form (a computer file) for such analysis.
In any continuous based monitor this information may include raw data
representing (1) radon progeny levels, (2) radon gas levels, (3) pump flow
rate, (4) filter pressure, (5) motor voltage, (6) motor current, (7) reference
voltage, (8) monitor status, (9) indoor ambient air temperature, (10)
barometric pressure, etc., as well as the exact time and date the information
was recorded and the length of each measurement period subset (15 minutes 30
minutes, 1 hour, etc.).	'
If an instrument is capable of accurately measuring and recording the
above information, and it is available as a computer file, it may be analyzed
via software programs in two basic ways. First, computer programs can be
written to determine the extent to which the monitor making the measurement
performed its intended functions. In essence, these "internal monitor
performance* routines are a software analysis of how well the machine worked
Second, computer programs can be written to present measurement data and oth*
information in such a way that it allows a human being to more easily	F
determine whether or not a test is valid. In essence, these "quality
assurance" routines are a software analysis of what happened during the
measurement period.
Internal monitor performance routines analyze measurement and internal
monitor data to validate monitor performance. These routines flag abnormal
monitor performance and performance outside nominal ranges as well as
excessive radon variances and other indications of possible error which mav
impact the validity of the test. The quality assurance routines enhance the
ease with which quality assurance personnel review all job data (monitor-
generated data as well as other relevant job data) before validating and
releasing radon test results.
It is important to understand that the software routines used to analy*
a monitor's measurement and performance data or enhance quality assurance
reviews are not designed to form conclusions or take action. They are onlv
designed to provide warnings to the people reviewing the results of these
routines and to prompt action by these people based on the warnings. The
technician in the field and the quality assurance personnel in the office m
decide, based on these software "red flags", whether or not problems have
occurred and whether they require further investigation or action.
Aside from very small corrections in machine performance, software
routines must be designed to tell people of potential problems. Computer
programs cannot decide if a test is valid; they can only flag possible test
problems. Actual decisions must be made by people. As wonderful as compute*,
2
-------
are, they do not have the ability to reason. Reliance should not be placed on
computer programs that attempt to validate a measurement. Artificial
intelligence is still in its infancy and these programs cannot "think." They
should only assist the quality assurance personnel responsible for validating
a test and signing off on the results.
Obviously then, software routines are only as good as the individuals
reviewing the results of those routines. Proper action by these people prior
to releasing results is what is important. That is the reason quality
assurance personnel must be independent of any phase of the test. Scheduling,
test installation, and test pick-up should not be performed by persons having
quality assurance responsibility for reviewing the data or reporting the
results from that test.
SOFTWARE ROUTINES FOR VALIDATING INTERNAL MONITOR PERFORMANCE
Software routines that analyze data for a particular measurement must be
designed to determine the extent to which the monitor performed its intended
functions for that test. Because the health threat associated with radon is
great, and particularly because a legally binding contract rests on the
outcome of the test, it is imperative that the quality assurance process be
real-time (immediate) and that it identify equipment reliability problems
before results are reported. This is done with software in two ways:
preventive routines and detective routines.
Preventive Software Routines
Preventive routines are designed to prevent unanticipated problems before
the measurement begins by preventing or highlighting problems before a monitor
is installed in a house. For example, software routines can be written to
ensure that measurement data from one test is not confused with measurement
data from another test or another house.
This is accomplished in a fully integrated system by programming the
computer to designate a specific job number for each data set; a job number
unique to that data set. When office and field functions are automated and
completely integrated, there is no need for anyone to manually enter or re-
enter job numbers or property addresses in the field after the order is
originally entered into the system. As a result, the computer ensures that
job numbers are not duplicated.
These programs can also ensure that the detailed measurement data, as
well as a separately created job information file (grab samples, technician
notes at installation and pick-up, sketch of property tested, answers to
questions about property tested, etc.) associated with a particular test on a
particular house, are not confused with other tests on that house or with
other tests of other houses.
Other preventive routines can be programmed into the monitor at
instrument power-up in the form of monitor self-checks. For example, a
3
-------
monitor has three primary areas that must be carefully reviewed at power-up to
ensure a 'working* machine is placed in the house: (1) filter integrity, (2)
flow, and (3) detector efficiency. Simply put, the filter cannot leak or the
test is rendered invalid. Even a pin hole leak not visible to the naked eye
can cause large measurement errors and result in reporting false negatives
In addition, air flow and detector efficiency must be exactly known in order
to perform an accurate working level (WL) calculation. WL measurement errors
are directly proportional to flow errors and detector efficiency errorB.
One preventive software routine needed at monitor power-up is to check
filter pressure drop to determine whether or not the filter has a leak. if it
does, the software can warn the install technician to replace the monitor's
filter. The software can then record when such replacement occurred.
Another preventive routine at power-up is a detector check. The software
can be written to request the equipment operator insert a radioactive source
into the filter holder in place of the filter. A two minute count can then
perform to measure the activity of the source. The known activity of this
source and the number of counts measured in the two minute period can be
compared against similar data taken when the monitor was calibrated, which can
be stored in internal monitor calibration tables. The software can ensure
that the activity ratio compares favorably with the reference source at the
time of calibration. If it does not, the software can warn the technician
that the detector is experiencing problems or operating outside its tolerance
etc. and that the monitor must not be installed in the house.	'
Finally, the flow rate must be precisely known. This is typically tested
with either a bubble tube device or with a mass flow meter during calibration
Care must be taken with a mass flow meter to correct for the altitude
barometric pressure and temperature effects on the mass flow meter. if the
equipment includes a mass flow sensor, absolute pressure sensor and
temperature sensor, the absolute pressure sensor can measure the cumulative
effect of the altitude and barometric pressure.
These sensors allow software routines to make accurate volumetric flow
measurements. The results of the flow measurements can be stored with the raw
measurement data for each 15 minute period. if internal flow measurement
capability is not available, then flow measurements should be made both before
and after the test and frequent checks should be performed to demonstrate that
the flow for the device is stable and does not "wander* during tests.
Other preventive software routines can be written to track the
performance history of monitors in the field. This can ensure that monitors
with excessive recent problems are pulled from service before a failure
occurs. A computer program can compress, sort, and analyze all measurements
made by each machine for the last ten weeks. Tests performed the previous
week can be examined especially closely by the software.
For example, impending pump failure caused by slow bearing deterioration
may be easy to spot. A monitor with a malfunctioning pump bearing will show
4
-------
motor current increasing slowly over a period of several weeks. This monitor
can be pulled from service, before the pump actually fails, for pump repair or
replacement. Another good example is monitor detector failure. A monitor
showing increasing detector background noise over a period of recent tests
could also be pulled from service before failure occurs with well written
software analysis routines.
In essence, any machine experiencing difficulty, exhibiting deterioration
of some parameter, or showing sub-par performance can be identified by
computer programs designed to review all monitor measurement and internal
performance data. The software can automatically print this information on a
computer-generated report with a specific request for action by a specific
date for follow-up by quality assurance personnel.
Detective Software Routines
Detective software routines are designed to discover the existence,
presence, or occurrence of monitor malfunction, error, or substandard
performance during a measurement. Detective routines flag problems that
occurred during the test in the quality assurance process prior to reporting
measurement results.
Examples of these software quality assurance routines would be to verify
that instrument internal motor voltages, filter pressures, flow rates,
detector efficiencies, calibration factors, and a variety of other internal
parameters are operating within their prescribed limits. The quality
assurance personnel reviewing this information are able to repeat a test if
the instrument has malfunctioned during the measurement period.
Software routines can also be designed to verify that the date and time
of the test are correct by comparing the actual measurement dates and times
with scheduled dates and times. If for some reason the computer had no record
of the scheduled date and time, a comparison with other known dates would be
valid. For example, the dates the measurement was conducted should occur
after the date it was ordered. Similarly, measurement dates should be less
than the current date, etc.
SOFTWARE ROUTINES FOR VALIDATING QUALITY ASSURANCE
Prior to performing a detailed review of a particular test, quality
assurance personnel should have the ability to review previous tests conducted
on the property. Software routines can query the data base for previous tests
at the property address tested to identify prior tests and to allow for the
detailed review of those tests. Similarly, if multiple machines were used to
measure a particular property, either because of the size of the structure or
characteristics of it requiring more than one machine, these measurements
should be identified and located in the data base and be available for review.
Their results can be analyzed by special software routines that compare their
performance. Well-written software allows for all of this.
5
-------
All of the software routines described in the previous section that
validate internal monitor performance are designed to flag quality assurance
Personnel concerning possible problems with a particular measurement.
Motive software routine. are often designed more for the	^
7,11a to orevent him from installing a monitor that has a problem. These
„u"nes are also useful to quality assurance personnel because of the ongoing
record of each machine's performance history. However, they are some«hat U.s
fn aft-er the fact. Detective routines on the other hand, which have
limited application for field personnel, are enormously valuable for quality
PP	directed more to assist them. Because these
assurance	generally used to review a job after the fact, they are
designed more to help determine whether or not a test is valid and whether or
not that test should be repeated.
in addition to the preventive and detective routines that can be
* h nn * monitor's measurement and internal performance data, there are
^software routines designed for quality assurance. These routines
Enhance the ease with which quality assurance personnel are able to review all
enhance th	^ rticular measurement. Perhaps the easiest way to explain
of the dat	software would be to refer directly to a graph of that
the benefit of this software wo^	^ ^	^ ^	^
data (see Figure	number of raw data (see Figure 2), this picture
and columns of	, is a graphical depiction of the raw data in
Figure 2 Irreality this -picture" represents many hundreds of individual 15
minute data sets.
^v, n„„h in Figure 1 shows clearly radon progeny levels, radon gas
1 ,Mlature, barometric pressure, as well as changes in each over the
levels, tempe	ment period. m addition, this graph also shows
course of the	the presence of individuals in the vicinity of the
information co	measurement period as well as information showing if and
ihenrthenLnitor was moved. In combination this information has real meaning
particularly when graphed.
W rwc*r bv depicting reams of measurement data in an easy to understand
Moreov , yi^ ^	^ ^ qualifcy aBSurance personnel is
graph, the ana y	Changes in radon progeny or radon gas levels,
significantly "^^mperature, b„ometric pressure, and the location of
changes in ind	ence of persons in the vicinity of the monitor are all
the monitor, or	*	addition, abnormalities or tampering with
^"^rcondftSs is often very easy to identify a. well.
* nnmoaes of distinction, it is worth mentioning that the greater an
For purpo	sensitivity, the greater that instrument's ability to
instrument s 1	levels due to tampering versus the diurnal variation
identify chang	^ inBtrument having great sensitivity shows
of radon. A typ J ^ Bharply and dramatically (see Figure 3) whereas
changes in r^°" genaitivity show sharp changes in radon levels only
gradually?Vleaving a question .. to whether or not tampering occurred at all.
6
-------
With the benefit of software routines analyzing data to flag potential
problems, quality assurance personnel are also able to correlate other
relevant data accumulated during the measurement to help determine whether or
not a test is valid. For example, elevated grab samples at test initiation or
test completion may indicate failure of the monitor, or even tampering, if
they do not corroborate data provided by the continuous based monitor.
Also, closed-house conditions, which are explained when a measurement is
scheduled, that are not observed when a technician arrives at the house to
initiate the test, gives good reason for quality assurance personnel to
scrutinize measurement data more carefully. Reluctance or failure of a seller
responsible for closed-house conditions to sign a property owner agreement,
which identifies the conditions necessary for a valid test, may also mean
reason for greater scrutiny of the test data.
CONCLUSION
In summary, well designed software routines offer a tremendous operating
advantage. In any high volume environment, they are critical to ensure
quality.
To the extent that a company's radon measurement equipment is not able to
generate internal information concerning its own performance or transfer that
information into "machine readable" form (a computer file), that company is
severely limited in its ability to know whether or not its instruments
performed properly on a test. And to the extent that a radon measurement
system is not fully integrated (automated) from order taking to reporting,
that system suffers a significantly reduced ability to preserve quality.
Simply put, computers and well written software routines run circles
around human beings trying to perform the same tasks.
The work described in this paper was not funded by the U.S. Environmental
Protection Agency and therefore the contents do not necessarily reflect the
views of the Agency and no official endorsement should be inferred.
7
-------
BAROMETRIC PRESSURE
-TEMPERATURE
.RADON GAS
102100B
2 Lenape Drive
Stanhope, NJ
V RADONICS
The Radon Specialists
T ine
Date
9:00
1/23
13:00
1/23
17:00
1/23
21:00 l:oo
1/23 1/24
TIME OF TEST
5:00
1/24
9:00
1/24
13:00
1/24
I'.:: riSnMEffS'SHEB,1' Mial:;" 'IE
T ine
Date
17:00
1/24
21:00
1/24
1:00 5:oo 9:00 13:00 17:00 21:00
1/25 1/25 1/25 1/25 1/25 1/25
TIME OF TEST	Year of 1991
Figure 1. Graphical depiction of measurement data.
8
-------
BAROMETRIC PRESSURE
-TEMPERATURE
¦RADON GAS
102100B
2 Lenape Drive
Stanhope, NJ
V RADONICS
The Radon Specialists
¦ TAMPER
0.0281
0.0267
0.0253
M0,0239
J 0,0225
ill 0.0211
5 0.0197
11)0,0183
J 0.0169
0,0155
(J 0.0141
Z0.0126
h 0.0112
* 0.0098
1	0.0084
0 0,0070
2	0,0056
0.0042
0.0028
0.0014
T itie
Date
I'M p|	«.
1:00 5:00 9:00 13:00 17:00 21:00 1:00 5:00
1/26 1/26 1/26 1/26 1/26 1/26 1/27 1/27
TIME OF TEST
¦ TAMPER
0.0281
0.0267
0.0253
M 0.0239
J 0.0225
HI 0.0211
20.0197
111 0.0183
J 0.0169
0.0155
(JO. 0141
Z0.0126
m 0.0112
¥ 0.0098
IO.OOB4
0 0.0070
3 0.0056
0,0042
0,0028
0.0014
T ine
Date
9:00
1/27
13:00
1/27
17:00
1/27
2i:oo i:oo
1/27 1/28
TIME OF TEST
5:00
1/28
9:00
1/28
13:00
1/28
Year of 1991
Figure 1 (continued).
-------
T ime
Cnt1
Cnt2
Flo
Bar
FP
MV
MI Ref Status
Flow Temp
I
09:12:03
30
15
161
199
43
86
100 51
0000
0000
SOS
205
255
8
09:27:03
59
20
161
192
42
86
91 51
0000
0000
SDS
205
241
0
09:42:03
90
12
165
188
41
86
86 51
0000
0000
DDS
302
227
0
09:57:03
139
38
163
185
40
84
82 51
0000
0000
DDI
210
217
0
10:12:03
148
28
161
183
40
82
78 51
0000
0000
SSI
025
210
0
10:27:03
166
26
160
181
39
82
76 51
0000
0000
SSI
025
206
0
10:42:03
186
21
160
180
40
82
75 51
0000
0000
SSI
025
203
0
10:57:03
217
27
160
180
39
82
74 51
0000
0000
SSI
025
201
0
11:12:03
240
26
160
179
39
82
73 51
0000
0000
SSI
025
199
0
11:27:03
244
24
160
178
39
82
74 51
0000
0000
SSI
025
198
0
11:42:03
278
28
159
178
40
82
73 51
0000
0000
ISI
032
197
0
11:57:03
268
31
161
178
40
84
74 51
0000
0000
SSI
015
197
0
12:12:03
262
33
161
177
40
84
74 51
0000
0000
SSI
015
196
0
12:27:03
276
26
161
177
40
84
74 51
0000
0000
SSI
015
195
0
12:42:03
277
39
161
177
40
84
74 51
0000
0000
SSI
015
196
0
12:57:03
310
31
161
176
40
84
74 51
0000
0000
SSI
015
195
0
13:12:03
267
26
161
176
40
84
74 51
0000
0000
SSI
015
195
0
13:27:03
294
31
159
176
40
84
73 51
0000
0000
ISI
022
194
0
13:42:03
315
35
159
176
40
84
74 51
0000
0000
ISI
022
194
0
13:57:03
302
26
159
175
40
84
74 51
0000
0000
ISI
022
194
0
14:12:03
319
22
159
175
40
84
73 51
0000
0000
ISI
022
194
0
14:27:03
319
16
159
175
40
84
74 51
0000
0000
ISI
022
193
0
14:42:03
340
32
159
175
40
84
74 51
0000
0000
ISI
022
193
0
14:57:03
320
25
159
175
40
84
73 51
0000
0000
ISI
022
193
0
15:12:03
294
33
159
174
40
84
73 51
0000
0000
ISI
022
193
0
15:27:03
296
31
159
174
40
84
73 51
0000
0000
ISI
022
192
0
15:42:03
307
21
159
174
40
84
73 51
0000
0000
ISI
022
192
0
15:57:03
339
24
159
175
40
84
74 51
0000
0000
ISI
022
192
0
16:12:03
324
20
159
175
40
84
73 51
0000
0000
ISI
022
192
0
16:27:03
322
33
159
174
40
84
73 51
0000
0000
ISI
022
192
0
16:42:03
353
26
159
174
40
84
73 51
0000
0000
ISI
022
191
0
16:57:03
346
34
159
174
40
84
74 51
0000
0000
ISI
022
191
0
17:12:03
326
35
159
174
40
84
74 51
0000
0000
ISI
022
192
0
17:27:03
336
29
159
174
40
84
73 51
0000
0000
ISI
022
191
0
17:42:03
349
20
159
174
40
84
73 51
0000
0000
ISI
022
191
0
17:57:03
348
30
159
174
40
84
73 51
0000
oooo
ISI
022
191
0
18:12:03
342
40
159
174
40
84
74 51
0000
0000
ISI
022
191
0
18:27:03
304
37
159
174
40
84
74 51
0000
0000
ISI
022
191
0
18:42:03
359
24
159
174
40
84
73 51
0000
0000
ISI
022
191
0
18:57:03
314
32
159
174
40
84
73 51
0000
0000
ISI
022
191
0
19:12:03
353
46
159
174
40
84
73 51
0000
0000
ISI
022
191
0
19:27:03
328
23
159
174
40
84
73 51
0000
0000
ISI
022
191
0
19:42:03
360
34
159
174
40
84
73 51
0000
0000
ISI
022
191
0
19:57:03
303
46
159
174
40
84
73 51
0000
0000
ISI
022
190
0
20:12:03
311
24
159
174
40
84
73 51
0000
0000
ISI
022
190
0
20:27:03
351
23
159
174
40
84
73 51
0000
0000
ISI
022
190
0
20:42:03
320
21
159
174
40
84
73 51
0000
0000
ISI
022
190
0
20:57:03
336
35
159
174
40
84
73 51
0000
0000
ISI
022
190
0
21:12:03
317
25
159
174
40
84
73 51
0000
0000
ISI
022
190
0
21:27:03
324
25
159
174
40
84
73 51
0000
0000
ISI
022
190
0
21:42:03
330
30
159
174
40
84
73 51
0000
0000
ISI
022
190
0
21:57:03
336
26
159
174
40
84
73 51
0000
0000
ISI
022
190
0
22:12:03
365
35
159
174
40
84
73 51
0000
0000
ISI
022
190
0
22:27:03
325
30
159
174
40
84
73 51
0000
0000
ISI
022
189
0
22:42:03
347
49
159
174
40
84
73 51
0000
0000
ISI
022
190
0
22:57:03
353
26
159
174
40
84
73 51
0000
0000
ISI
022
189
0
23:12:03
333
28
159
174
40
84
73 51
0000
0000
ISI
022
189
0
23:27:03
338
20
161
174
40
84
73 51
0000
0000
SDI
113
189
0
23:42:03
315
30
161
173
40
84
73 51
0000
0000
SSI
015
189
0
23:57:03
353
32
161
173
40
84
72 51
0000
0000
SSI
015
189
0
Figure 2. Raw data from graph of Figure 1
s
2
0
0
Q
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
10
-------
Time
Cnt1
Cnt2
Flo
Bar
FP
MV
Ml
1 Ref Status
Flow Temp
I
00:12:03
363
30
161
173
40
84
73
51
0000
oooo
SSI
015
189
0
00:27:03
336
27
161
173
40
84
73
51
0000
oooo
SSI
015
189
0
00:42:03
350
37
161
173
40
84
73
51
0000
oooo
SOI
113
189
0
00:57:03
358
30
161
173
40
84
73
51
0000
oooo
SSI
015
189
0
01:12:03
354
24
161
173
41
84
73
51
0000
oooo
SDI
113
189
0
01:27:03
362
19
161
173
40
84
73
51
0000
oooo
SSI
015
189
0
01:42:03
370
35
161
173
40
84
73
51
0000
oooo
SDI
113
188
0
01:57:03
344
41
161
173
40
84
73
51
0000
oooo
SSI
015
189
0
02:12:03
347
40
161
173
41
84
73
51
0000
oooo
SSI
015
189
0
02:27:03
353
37
161
173
40
84
73
51
0000
oooo
SSI
015
188
0
02:42:03
336
41
161
173
41
84
73
51
0000
oooo
SDI
113
188
0
02:57:03
392
27
161
173
40
84
73
51
0000
oooo
SSI
015
189
0
03:12:03
387
34
161
173
40
84
72
51
0000
oooo
SOI
113
188
0
03:27:03
366
42
161
173
40
84
72
51
0000
oooo
SDI
113
188
0
03:42:03
356
33
161
173
41
84
72
51
0000
oooo
SSI
015
188
0
03:57:03
365
26
161
173
41
84
73
51
0000
oooo
SSI
015
188
0
04:12:03
377
31
161
173
41
84
72
51
0000
oooo
SSI
015
188
0
04:27:03
356
40
161
173
41
84
73
51
0000
oooo
SDI
113
188
0
04:42:03
377
35
161
173
41
84
72
51
0000
oooo
SDI
113
188
0
04:57:03
386
37
161
173
41
84
73
51
0000
oooo
SDI
113
188
0
05:12:03
360
28
161
173
41
84
73
51
0000
oooo
SDI
113
188
0
05:27:03
387
39
159
174
42
84
73
51
0000
oooo
IDI
120
188
0
05:42:03
374
31
161
174
42
86
74
51
0000
oooo
SDS
105
188
0
05:57:03
368
24
161
174
42
86
74
51
0000
oooo
SDS
105
188
0
06:12:03
397
27
161
174
42
86
73
51
0000
oooo
SDS
105
188
0
06:27:03
391
29
161
174
42
86
73
51
0000
oooo
SDS
105
188
0
06:42:03
360
34
161
174
42
86
73
51
0000
oooo
SDS
105
188
0
06:57:03
425
37
161
174
42
86
73
51
0000
oooo
SDS
105
188
0
07:12:03
405
24
161
174
42
86
74
51
0000
oooo
SDS
105
188
0
07:27:03
391
31
161
174
42
86
74
51
0000
oooo
SDS
105
188
0
07:42:03
376
28
161
174
42
86
73
51
0000
oooo
SDS
105
188
0
07:57:03
347
24
161
174
42
86
73
51
0000
oooo
SDS
105
188
0
08:12:03
352
24
161
175
42
86
74
51
0000
oooo
SDS
105
188
0
08:27:03
356
29
161
175
42
86
74
51
0000
oooo
SDS
105
188
0
08:42:03
368
23
161
175
42
86
73
51
0000
oooo
SDS
105
188
0
08:57:03
330
31
161
175
42
86
74
51
0000
oooo
SDS
105
188
0
09:12:03
303
34
161
175
42
86
73
51
0000
oooo
SDS
105
188
0
09:27:03
331
20
161
176
42
86
73
51
0000
oooo
SDS
105
188
0
09:42:03
319
33
161
176
42
86
73
51
0000
oooo
SDS
105
188
0
09:57:03
317
20
161
176
42
86
74
51
0000
oooo
SDS
105
188
0
10:12:03
296
29
161
176
42
86
73
51
0000
oooo
SDS
105
188
0
10:27:03
284
30
161
175
42
86
73
51
0000
oooo
SDS
105
188
0
10:42:03
301
33
161
176
42
86
74
51
0000
oooo
SDS
105
188
0
10:57:03
289
21
161
176
42
86
73
51
0000
oooo
SDS
105
188
0
11:12:03
291
32
161
176
42
86
73
51
0000
oooo
SDS
105
188
0
11:27:03
261
31
161
175
42
86
73
51
0000
oooo
SDS
105
188
0
11:42:03
242
36
161
175
42
86
74
51
0000
oooo
SDS
105
188
0
11:57:03
223
29
161
175
42
86
73
51
0000
oooo
SDS
105
188
0
12:12:03
284
34
161
175
42
86
73
51
0000
oooo
SDS
105
188
0
12:27:03
259
39
161
174
42
86
74
51
0000
oooo
SDS
105
188
0
12:42:03
295
29
161
174
42
86
73
51
0000
oooo
SDS
105
188
0
12:57:03
237
39
161
174
42
86
73
51
0000
oooo
SDS
105
188
0
13:12:03
248
27
161
173
42
86
74
51
0000
oooo
SDS
105
188
0
13:27:03
288
36
163
173
42
86
74
51
oooo
oooo
DDS
202
188
0
13:42:03
246
25
163
173
42
86
73
51
0000
oooo
DDS
202
188
0
13:57:03
290
28
163
173
42
86
74
51
0000
oooo
DDS
202
187
0
14:12:03
279
30
163
173
42
86
73
51
0000
oooo
DDS
202
187
0
14:27:03
278
31
163
173
42
86
73
51
oooo
oooo
DDS
202
188
0
14:42:03
284
31
163
173
42
86
73
51
0000
oooo
DDS
202
187
0
Figure 2. (continued).
S
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
11
-------
Time
Cntl
Cnt2
Flo
Bar
FP
MV
MI Ref Status
Flow Temp
I
14:57:03
293
35
163
173
42
86
73 51
0000
0000
DDS
202
187
0
15:12:03
294
33
163
173
42
86
73 51
0000
0000
DOS
202
187
0
15:27:03
274
37
163
173
42
86
73 51
0000
0000
DDS
202
187
0
15:42:03
306
25
163
173
42
86
73 51
0000
0000
DDS
202
187
0
15:57:03
266
25
163
173
42
86
74 51
0000
0000
DDS
202
187
0
16:12:03
296
33
163
173
42
86
74 51
0000
0000
DDS
202
186
0
16:27:03
293
30
163
173
42
86
73 51
0000
0000
DDS
202
187
0
16:42:03
261
29
161
174
42
86
73 51
0000
0000
SDS
105
187
0
16:57:03
295
36
161
174
42
86
73 51
0000
0000
SDS
105
187
0
17:12:03
265
34
161
174
42
86
73 51
0000
0000
SDS
105
187
0
17:27:03
288
42
161
174
42
86
74 51
0000
0000
SDS
105
187
0
17:42:03
323
30
161
175
42
86
74 51
0000
0000
SDS
105
187
0
17:57:03
265
38
161
175
42
86
73 51
0000
0000
SDS
105
187
0
18:12:03
285
30
161
175
42
86
73 51
0000
0000
SDS
105
187
0
18:27:03
277
42
161
175
42
86
73 51
0000
0000
SDS
105
187
0
18:42:03
284
37
161
175
42
86
73 51
0000
0000
SDS
105
187
0
18:57:03
281
51
161
176
42
86
73 51
0000
0000
SDS
105
187
0
19:12:03
268
17
161
176
42
86
74 51
0000
0000
SDS
105
187
0
19:27:03
280
39
161
176
42
86
73 51
0000
0000
SDS
105
187
0
19:42:03
301
38
161
176
42
86
73 51
0000
0000
SDS
105
187
0
19:57:03
241
21
161
176
42
86
73 51
0000
0000
SDS
105
187
0
20:12:03
260
33
161
176
42
86
74 51
0000
0000
SDS
105
187
0
20:27:03
256
24
161
177
42
86
73 51
0000
0000
SDS
105
187
0
20:42:03
233
40
161
177
42
86
73 51
0000
0000
SDS
105
187
0
20:57:03
267
44
161
177
42
86
73 51
0000
0000
SDS
105
187
0
21:12:03
270
52
161
177
42
86
74 51
0000
0000
SDS
105
187
0
21:27:03
268
29
161
177
42
86
74 51
0000
0000
SDS
105
187
0
21:42:03
245
31
161
177
42
86
73 51
0000
0000
SDS
105
187
0
21:57:03
297
46
161
177
42
86
74 51
0000
0000
SDS
105
188
0
22:12:03
276
37
162
177
42
86
74 51
0000
0000
SDS
105
187
0
22:27:03
280
42
161
178
42
86
74 51
0000
0000
SDS
105
187
0
22:42:03
274
39
162
178
42
86
73 51
0000
0000
SDS
105
187
0
22:57:03
274
25
161
178
42
86
74 51
0000
0000
SDS
105
187
0
23:12:03
278
52
162
178
42
86
74 51
0000
oooo
SDS
105
188
0
23:27:03
281
36
162
178
42
86
74 51
0000
0000
SDS
105
187
0
23:42:03
260
38
162
178
42
86
73 51
0000
0000
SDS
105
187
0
23:57:03
285
20
162
178
42
86
73 51
0000
0000
SDS
105
188
0
00:12:03
284
42
162
178
42
86
74 51
0000
0000
SDS
105
188
0
00:27:03
274
14
162
178
42
86
74 51
0000
0000
SOS
105
188
0
00:42:03
275
23
162
179
42
86
73 51
0000
0000
SDS
105
188
0
00:57:03
259
28
162
179
42
86
74 51
0000
0000
SDS
105
188
0
01:12:03
288
46
162
178
42
86
74 51
0000
0000
SDS
105
188
0
01:27:03
261
33
162
179
42
86
73 51
0000
0000
SDS
105
188
0
01:42:03
257
36
162
179
42
86
73 51
0000
0000
SDS
105
188
0
01:57:03
266
23
162
179
42
86
74 51
0000
0000
SDS
105
189
0
02:12:03
264
21
162
179
42
86
73 51
0000
0000
SDS
105
188
0
02:27:03
268
29
162
179
42
86
73 51
0000
0000
SDS
105
189
0
02:42:03
243
33
162
179
42
86
73 51
0000
0000
SDS
105
189
0
02:57:03
265
38
162
179
42
86
73 51
0000
0000
SDS
105
189
0
03:12:03
219
30
162
180
42
86
73 51
0000
0000
SDS
105
189
0
03:27:03
247
29
162
180
42
86
74 51
0000
0000
SDS
105
189
0
03:42:03
212
30
162
180
42
86
74 51
0000
0000
SOS
105
190
0
03:57:03
235
34
162
180
42
86
74 51
0000
0000
SDS
105
189
0
04:12:03
229
19
162
180
42
86
74 51
0000
0000
SDS
105
190
0
04:27:03
217
26
162
180
42
86
74 51
0000
0000
SDS
105
190
0
04:42:03
207
37
162
180
42
86
74 51
0000
0000
SDS
105
191
0
04:57:03
210
31
162
180
42
86
74 51
0000
0000
SDS
105
190
0
05:12:03
198
25
162
180
42
86
73 51
0000
0000
SDS
105
190
0
05:27:03
248
34
162
180
42
86
74 51
0000
0000
SDS
105
191
0
Figure 2. (continued).
S
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
12
-------
T ime
Cnt1
Cnt2
Flo
Bar
FP
MV
Ml
I Ref Status
Flow Temp
I
05:42:03
212
30
162
181
42
86
74
51
0000
0000
SDS
105
191
0
05:57:03
205
30
162
181
42
86
74
51
0000
0000
SDS
105
191
0
06:12:03
237
37
162
181
42
86
74
51
0000
0000
SDS
105
191
0
06:27:03
223
19
162
181
42
86
74
51
0000
0000
SDS
105
191
0
06:42:03
211
30
162
181
42
86
74
51
0000
0000
SDS
105
192
0
06:57:03
228
24
162
181
42
86
74
51
0000
0000
SDS
105
191
0
07:12:03
208
23
162
182
42
86
74
51
0000
0000
SDS
105
192
0
07:27:03
210
41
162
182
42
86
74
51
0000
0000
SDS
105
192
0
07:42:03
234
34
162
182
42
86
74
51
0000
0000
SDS
105
192
0
07:57:03
262
25
162
182
42
86
74
51
0000
0000
SDS
105
192
0
08:12:03
229
40
162
182
42
86
74
51
0000
0000
SDS
105
192
0
08:27:03
244
33
162
182
42
86
74
51
0000
0000
SDS
105
193
0
08:42:03
248
24
162
183
42
86
74
51
0000
0000
SDS
105
192
0
08:57:03
251
20
162
183
42
86
74
51
0000
0000
SDS
105
192
0
09:12:03
203
24
162
183
42
86
74
51
0000
0000
SDS
105
193
0
09:27:03
253
33
162
183
42
86
74
51
0000
0000
SDS
105
193
0
09:42:03
227
28
162
183
42
86
74
51
0000
0000
SDS
105
193
0
09:57:03
218
24
162
183
42
86
74
51
0000
0000
SDS
105
193
0
10:12:03
228
21
162
183
42
86
74
51
0000
0000
SDS
105
193
0
10:27:03
246
25
162
183
42
86
75
51
0000
0000
SDS
105
193
0
10:42:03
222
33
162
183
42
86
75
51
0000
0000
SDS
105
193
0
10:57:03
196
24
162
183
42
86
74
51
0000
0000
SDS
105
193
0
11:12:03
197
36
162
183
42
86
74
51
0000
0000
SDS
105
193
0
11:27:03
193
33
162
183
42
86
74
51
0000
0000
SDS
105
193
0
11:42:03
230
26
162
183
42
86
74
51
0000
0000
SDS
105
192
0
11:57:03
224
35
162
183
42
86
74
51
0000
0000
SDS
105
193
0
12:12:03
207
29
162
183
42
86
74
51
0000
0000
SDS
105
193
0
12:27:03
207
26
162
182
42
86
74
51
0000
0000
SDS
105
193
0
12:42:03
199
27
162
182
42
86
74
51
0000
0000
SDS
105
192
0
12:57:03
190
29
162
182
42
86
74
51
0000
0000
SDS
105
192
0
13:12:03
232
22
162
182
42
86
74
51
0000
0000
SDS
105
192
0
13:27:03
225
28
162
182
42
86
74
51
0000
0000
SDS
105
192
0
13:42:03
209
24
162
182
42
86
74
51
0000
0000
SDS
105
192
0
13:57:03
236
29
162
182
42
86
75
51
0000
0000
SDS
105
192
0
14:12:03
236
23
162
182
42
86
74
51
0000
0000
SDS
105
192
0
14:27:03
228
33
162
182
42
86
74
51
0000
0000
SDS
105
192
0
14:42:03
236
27
162
182
42
B6
74
51
0000
0000
SDS
105
191
0
14:57:03
237
27
162
182
42
86
73
51
0000
0000
SDS
105
191
0
15:12:03
247
35
162
182
42
86
74
51
0000
0000
SDS
105
191
0
15:27:03
296
33
162
182
42
86
74
51
0000
0000
SDS
105
191
0
15:42:03
265
38
162
182
42
86
74
51
0000
0000
SDS
105
191
0
15:57:03
257
30
162
182
42
86
75
51
0000
0000
SDS
105
191
0
16:12:03
251
32
162
182
42
86
74
51
0000
0000
SDS
105
191
0
16:27:03
270
38
162
182
42
86
74
51
0000
0000
SDS
105
191
0
16:42:03
283
33
162
182
42
86
74
51
0000
0000
SDS
105
191
0
16:57:03
272
25
162
182
42
86
74
51
0000
0000
SDS
105
191
0
17:12:03
293
36
162
182
42
86
74
51
0000
0000
SDS
105
191
0
17:27:03
297
31
162
182
42
86
74
51
0000
0000
SDS
105
190
0
17:42:03
308
34
162
182
42
86
74
51
0000
0000
SDS
105
190
0
17:57:03
303
27
162
182
42
86
74
51
0000
0000
SDS
105
191
0
18:12:03
307
37
162
182
42
86
74
51
0000
0000
SDS
105
191
0
18:27:03
296
31
162
183
42
86
74
51
0000
0000
SDS
105
191
0
18:42:03
296
39
162
183
42
86
74
51
0000
0000
SDS
105
191
0
18:57:03
317
39
162
183
42
86
74
51
0000
0000
SDS
105
191
0
19:12:03
339
41
162
183
42
86
74
51
0000
0000
SDS
105
191
0
19:27:03
303
32
162
183
42
86
74
51
0000
0000
SDS
105
191
0
19:42:03
321
22
162
183
42
86
74
51
0000
0000
SDS
105
191
0
19:57:03
276
27
162
183
42
86
74
51
0000
0000
SDS
105
191
0
20:12:03
333
36
162
183
42
86
74
51
0000
0000
SDS
105
191
0
Figure 2. (continued).
S
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
13
-------
Time
Cnt1
Cnt2
Flo
Bar
FP
MV
MI Ref Status
Flow Ten©
I
20:27:03
307
28
162
183
42
86
75
51
0000
0000
SOS
105
191
0
20:42:03
341
16
162
183
42
86
74
51
0000
0000
SOS
105
191
0
20:57:03
316
24
162
183
42
86
75
51
0000
0000
SDS
105
192
0
21:12:03
345
27
162
183
42
86
74
51
0000
0000
SOS
105
192
0
21:27:03
326
50
162
183
42
86
74
51
0000
0000
SDS
105
192
0
21:42:03
336
29
162
183
42
86
75
51
0000
0000
SDS
105
192
0
21:57:03
318
39
162
183
42
86
75
51
0000
0000
SDS
105
192
0
22:12:03
327
33
162
183
42
86
75
51
0000
0000
SDS
105
192
0
22:27:03
336
30
162
183
42
86
75
51
0000
0000
SOS
105
192
0
22:42:03
345
37
162
183
42
86
75
51
0000
0000
SDS
105
192
0
22:57:03
377
34
162
183
42
86
74
51
0000
0000
SOS
105
192
0
23:12:03
315
45
162
183
42
86
75
51
0000
0000
SOS
105
192
0
23:27:03
320
32
162
183
42
86
75
51
0000
0000
SOS
105
193
0
23:42:03
336
42
162
183
42
86
74
51
0000
0000
SOS
105
193
0
23:57:03
342
44
162
183
42
86
74
51
0000
0000
SDS
105
192
0
00:12:03
356
23
162
183
42
86
75
51
0000
0000
SDS
105
193
0
00:27:03
340
27
162
183
42
86
74
51
0000
0000
SDS
105
193
0
00:42:03
354
35
162
183
42
86
75
51
0000
0000
SDS
105
193
0
00:57:03
370
29
162
183
42
86
74
51
0000
0000
SDS
105
193
0
01:12:03
376
25
162
183
42
86
75
51
0000
0000
SDS
105
193
0
01:27:03
364
26
162
183
42
86
74
51
0000
0000
SDS
105
193
0
01:42:03
356
35
162
183
42
86
75
51
0000
0000
SDS
105
193
0
01:57:03
359
27
162
183
42
86
75
51
0000
0000
SDS
105
193
0
02:12:03
371
20
162
183
42
86
74
51
0000
0000
SDS
105
193
0
02:27:03
378
28
162
183
42
86
74
51
0000
0000
SDS
105
194
0
02:42:03
360
22
162
183
42
86
74
51
0000
0000
SDS
105
194
0
02:57:03
364
20
162
183
42
86
75
51
0000
0000
SDS
105
194
0
03:12:03
372
26
164
183
42
86
75
51
0000
0000
DOS
202
194
0
03:27:03
355
33
162
183
42
86
75
51
0000
0000
SDS
105
194
0
03:42:03
386
21
162
183
42
86
75
51
0000
0000
SDS
105
194
0
03:57:03
366
37
162
183
42
86
74
51
0000
0000
SDS
105
194
0
04:12:03
373
30
164
183
42
86
74
51
0000
0000
DDS
202
195
0
04:27:03
336
17
164
183
42
86
75
51
0000
0000
DOS
202
195
0
04:42:03
342
26
164
183
42
86
75
51
0000
0000
DDS
202
195
0
04:57:03
343
33
164
183
42
86
75
51
0000
0000
DDS
202
195
0
05:12:03
352
30
164
183
42
86
75
51
0000
0000
DDS
202
195
0
05:27:03
345
29
164
183
42
86
74
51
0000
0000
DOS
202
195
0
05:42:03
351
21
164
183
42
86
75
51
0000
0000
DDS
202
195
0
05:57:03
336
27
164
183
42
86
75
51
0000
0000
DDS
202
195
0
06:12:03
310
27
164
183
42
86
74
51
0000
0000
DDS
202
195
0
06:27:03
322
22
164
183
42
86
74
51
0000
0000
DOS
202
195
0
06:42:03
318
36
164
183
42
86
74
51
0000
0000
DDS
202
196
0
06:57:03
339
27
164
183
42
86
75
51
0000
0000
DDS
202
195
0
07:12:03
352
20
164
183
42
86
74
51
0000
0000
DDS
202
195
0
07:27:03
349
24
164
183
42
86
74
51
0000
0000
DDS
202
195
0
07:42:03
331
42
164
183
42
86
75
51
0000
0000
DDS
202
195
0
07:57:03
350
37
164
183
42
86
74
51
0000
0000
DOS
202
195
0
08:12:03
333
33
164
183
42
86
75
51
0000
0000
DOS
202
195
0
08:27:03
364
35
164
183
42
86
75
51
0000
0000
DDS
202
195
0
08:42:03
318
31
164
183
42
86
74
51
0000
0000
DDS
202
195
0
08:57:03
336
27
164
183
42
86
75
51
0000
0000
DDS
202
195
0
09:12:03
320
36
164
183
42
86
75
51
0000
0000
DDS
202
195
0
09:27:03
335
35
164
183
42
86
75
51
0000
0000
DOS
202
195
0
09:42:03
290
27
164
183
42
86
75
51
0000
0000
DOS
202
195
0
09:57:03
324
30
164
182
42
86
75
51
0000
0000
DDS
202
195
0
10:12:03
328
31
164
182
42
86
75
51
0000
0000
DDS
202
195
0
10:27:03
361
30
164
182
42
86
75
51
0000
0000
DDS
202
194
0
10:42:03
339
31
162
182
42
86
75
51
0000
0000
SDS
105
194
0
10:57:03
396
28
162
182
42
86
75
51
0000
0000
SDS
105
194
0
Figure 2. (continued).
S
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
14
-------
Time
Cnt1
Cnt2
Flo
Bar
FP
MV
Ml
: Ref Status
Flow Temp
I
11:12:03
389
28
162
182
42
86
74
51
0000
0000
SOS
105
194
0
11:27:03
360
42
164
182
42
86
75
51
0000
0000
DOS
202
194
0
11:42:03
363
38
162
181
42
86
74
51
0000
0000
SDS
105
194
0
11:57:03
354
35
162
181
42
86
74
51
0000
0000
SOS
105
194
0
12:12:03
361
31
162
181
42
86
75
51
0000
0000
SOS
105
194
0
12:27:03
325
28
162
180
42
86
74
51
0000
0000
SDS
105
193
0
12:42:03
354
37
162
180
42
86
74
51
0000
0000
SDS
105
193
0
12:57:03
357
39
162
180
42
86
75
51
0000
0000
SDS
105
193
0
13:12:03
374
39
162
179
42
86
75
51
0000
0000
SDS
105
193
0
13:27:03
354
40
162
179
42
86
74
51
0000
0000
SDS
105
193
0
13:42:03
351
41
162
179
42
86
74
51
0000
0000
SDS
105
193
0
13:57:03
343
45
162
179
42
86
74
51
0000
0000
SDS
105
193
0
14:12:03
387
36
162
178
42
86
74
51
0000
0000
SDS
105
193
0
14:27:03
354
36
162
178
42
86
74
51
0000
0000
SDS
105
192
0
14:42:03
381
29
162
178
42
86
74
51
0000
0000
SDS
105
192
0
14:57:03
389
19
162
178
42
86
74
51
0000
0000
SDS
105
192
0
15:12:03
381
20
162
178
42
86
74
51
0000
0000
SDS
105
192
0
15:27:03
398
29
162
177
42
86
74
51
0000
0000
SDS
105
192
0
15:42:03
424
35
162
177
42
86
74
51
0000
0000
SDS
105
192
0
15:57:03
379
34
162
177
42
86
74
51
0000
0000
SDS
105
191
0
16:12:03
407
31
162
177
42
86
74
51
0000
0000
SDS
105
192
0
16:27:03
433
50
162
177
42
86
74
51
0000
0000
SDS
105
192
0
16:42:03
389
28
161
177
42
86
74
51
oooo
0000
SDS
105
191
0
16:57:03
414
31
161
177
42
86
74
51
0000
0000
SDS
105
191
0
17:12:03
445
43
161
177
42
86
74
51
0000
0000
SDS
105
191
0
17:27:03
438
36
162
177
42
86
74
51
0000
0000
SDS
105
191
0
17:42:03
409
41
161
176
42
86
74
51
0000
0000
SDS
105
191
0
17:57:03
387
34
161
176
42
86
75
51
0000
0000
SDS
105
191
0
18:12:03
374
42
161
176
42
86
74
51
0000
0000
SDS
105
191
0
18:27:03
429
32
161
176
42
86
75
51
0000
0000
SDS
105
191
0
18:42:03
439
34
161
176
42
86
74
51
0000
0000
SDS
105
191
0
18:57:03
412
26
161
176
42
86
74
51
0000
0000
SDS
105
191
0
19:12:03
444
27
161
176
42
86
75
51
0000
0000
SDS
105
191
0
19:27:03
377
32
161
176
42
86
74
51
0000
0000
SDS
105
191
0
19:42:03
419
24
161
176
42
86
74
51
0000
0000
SDS
105
191
0
19:57:03
411
20
161
176
42
86
74
51
0000
0000
SDS
105
190
0
20:12:03
407
32
161
175
42
86
74
51
0000
0000
SDS
105
191
0
20:27:03
420
30
161
175
42
86
74
51
0000
0000
SDS
105
191
0
20:42:03
411
33
161
175
42
86
74
51
0000
0000
SDS
105
191
0
20:57:03
389
29
161
175
42
86
74
51
0000
0000
SDS
105
190
0
21:12:03
389
32
161
175
42
86
74
51
0000
0000
SDS
105
191
0
21:27:03
393
34
161
175
42
86
74
51
0000
0000
SDS
105
190
0
21:42:03
408
26
161
175
42
86
74
51
0000
0000
SDS
105
190
0
21:57:03
398
22
161
175
42
86
74
51
0000
0000
SDS
105
190
0
22:12:03
401
39
161
175
42
86
74
51
0000
0000
SDS
105
190
0
22:27:03
400
29
161
175
42
86
74
51
0000
0000
SDS
105
190
0
22:42:03
388
25
161
175
42
86
74
51
0000
0000
SDS
105
190
0
22:57:03
387
19
161
175
42
86
74
51
0000
0000
SDS
105
190
0
23:12:03
399
25
161
175
42
86
75
51
0000
0000
SDS
105
190
0
23:27:03
382
37
161
174
42
86
74
51
0000
0000
SDS
105
190
0
23:42:03
376
29
161
174
42
86
74
51
0000
0000
SDS
105
190
0
23:57:03
379
28
161
174
42
86
74
51
0000
0000
SDS
105
190
0
00:12:03
439
41
161
174
42
86
74
51
0000
0000
SDS
105
190
0
00:27:03
361
36
161
174
42
86
74
51
0000
0000
SDS
105
190
0
00:42:03
370
26
161
174
42
86
74
51
0000
0000
SDS
105
189
0
00:57:03
351
26
161
174
42
86
74
51
0000
0000
SDS
105
190
0
01:12:03
364
16
161
174
42
86
74
51
0000
0000
SDS
105
190
0
01:27:03
362
25
161
174
42
86
74
51
0000
0000
SDS
105
190
0
01:42:03
411
44
161
174
42
86
74
51
0000
0000
SDS
105
189
0
Figure 2. (continued).
S
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-------
Time
Cnt1
Cnt2
Flo
Bar
FP
MV
HI Ref Status
Flow
Temp
I
01:57:03
372
32
161
174
42
86
74
51
0000
oooo
SDS
105
189
0
02:12:03
394
22
161
174
42
86
74
51
0000
0000
SDS
105
190
0
02:27:03
415
30
161
175
42
86
74
51
0000
oooo
SDS
105
189
0
02:42:03
391
34
161
175
42
86
74
51
0000
0000
SDS
105
189
0
02:57:03
397
38
161
175
42
86
74
51
0000
oooo
SDS
105
189
0
03:12:03
410
31
161
175
42
86
74
51
0000
oooo
SDS
105
190
0
03:27:03
390
36
161
175
42
86
74
51
0000
oooo
SDS
105
189
0
03:42:03
379
27
161
175
42
86
74
51
0000
oooo
SDS
105
189
0
03:57:03
368
34
161
175
42
86
74
51
0000
oooo
SDS
105
189
0
04:12:03
396
34
161
175
42
86
74
51
0000
oooo
SDS
105
190
0
04:27:03
381
37
161
175
42
86
74
51
0000
oooo
SDS
105
189
0
04:42:03
366
38
161
175
42
86
74
51
0000
oooo
SDS
105
189
0
04:57:03
385
40
161
175
42
86
74
51
0000
oooo
SDS
105
189
0
05:12:03
370
26
161
175
42
86
74
51
0000
oooo
SDS
105
189
0
05:27:03
389
31
161
175
42
86
74
51
0000
oooo
SDS
105
189
0
05:42:03
329
30
161
176
42
86
74
51
0000
oooo
SDS
105
189
0
05:57:03
363
36
161
176
42
86
74
51
0000
oooo
SDS
105
190
0
06:12:03
354
34
161
176
42
86
74
51
0000
oooo
SDS
105
190
0
06:27:03
381
31
161
176
42
86
74
51
0000
oooo
SDS
105
189
0
06:42:03
361
21
162
176
42
86
74
51
0000
oooo
SDS
105
189
0
06:57:03
365
38
162
177
42
86
74
51
0000
oooo
SDS
105
190
0
07:12:03
354
32
162
177
42
86
74
51
0000
oooo
SDS
105
190
0
07:27:03
361
34
162
177
42
86
74
51
0000
oooo
SDS
105
190
0
07:42:03
370
28
162
177
42
86
74
51
0000
oooo
SDS
105
190
0
07:57:03
328
38
162
177
42
86
74
51
0000
oooo
SDS
105
190
0
08:12:03
390
29
162
177
42
86
74
51
0000
oooo
SDS
105
190
0
08:27:03
328
37
162
177
42
86
74
51
0000
oooo
SDS
105
190
0
08:42:03
344
36
162
177
42
86
74
51
0000
oooo
SDS
105
190
0
08:57:03
343
27
162
177
42
86
74
51
0000
oooo
SDS
105
190
0
09:12:03
352
36
162
177
42
86
74
51
0000
oooo
SDS
105
190
0
09:27:03
360
30
162
177
42
86
74
51
0000
oooo
SDS
105
190
0
09:42:03
328
20
162
177
42
86
74
51
0000
oooo
SDS
105
190
0
09:57:03
342
18
162
177
42
86
74
51
0000
oooo
SDS
105
190
0
10:12:03
304
36
162
178
42
86
74
51
0000
oooo
SDS
105
190
0
10:27:03
371
29
162
178
42
86
74
51
0000
oooo
SDS
105
190
0
10:42:03
333
22
162
178
42
86
74
51
0000
oooo
SDS
105
190
0
10:57:03
320
36
162
177
42
86
74
51
0000
oooo
SDS
105
190
0
11:12:03
327
36
162
177
42
86
74
51
0000
oooo
SDS
105
191
0
11:27:03
328
36
162
177
42
86
74
51
0000
oooo
SDS
105
190
0
11:42:03
321
23
162
177
42
86
74
51
0000
oooo
SDS
105
190
0
11:57:03
338
37
162
177
42
86
74
51
0000
oooo
SDS
105
190
0
12:12:03
324
29
162
177
42
86
74
51
0000
oooo
SDS
105
190
0
12:27:03
384
31
162
177
42
86
74
51
0000
oooo
SDS
105
190
0
12:42:03
350
24
162
176
42
86
73
51
0000
oooo
SDS
105
190
0
12:57:03
376
34
161
176
42
86
74
51
0000
oooo
SDS
105
189
0
13:12:03
347
36
161
176
42
86
74
51
0000
oooo
SDS
105
189
0
13:27:03
362
37
161
176
42
86
73
51
0000
oooo
SDS
105
190
0
13:42:03
362
39
161
176
42
86
73
51
0000
oooo
SDS
105
190
0
13:57:03
352
23
161
176
42
86
74
51
0000
oooo
SDS
105
189
0
14:12:03
357
32
161
176
42
86
74
51
0000
oooo
SDS
105
189
0
14:27:03
358
35
161
176
42
86
74
51
0000
oooo
SDS
105
189
0
14:42:03
399
34
161
176
42
86
73
51
0000
oooo
SDS
105
189
0
14:57:03
355
39
161
176
42
86
74
51
0000
oooo
SDS
105
189
0
15:12:03
386
45
161
176
42
86
74
51
0000
oooo
SDS
105
189
0
15:27:03
388
34
161
176
42
86
74
51
0000
oooo
SDS
105
189
0
15:42:03
399
30
161
176
42
86
74
51
0000
oooo
SDS
105
189
0
15:57:03
412
26
161
175
42
86
74
51
0000
oooo
SDS
105
189
0
16:12:03
393
47
161
176
42
86
73
51
0000
oooo
SDS
105
189
0
16:27:03
347
32
161
175
42
86
74
51
0000
oooo
SDS
105
189
0
Figure 2. (continued).
S
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
16
-------
Time
Cnt1
Cnt2
Flo
Bar
FP
MV
Ml
I Ref Status
Flow Temp
1
16:42:03
398
33
161
175
42
86
73
51
oooo
0000
SDS
105
189
0
16:57:03
393
43
161
175
42
86
73
51
0000
0000
SDS
105
188
0
17:12:03
438
36
161
175
42
86
73
51
0000
0000
SDS
105
188
0
17:27:03
384
31
161
175
42
86
73
51
0000
0000
SDS
105
188
0
17:42:03
385
29
161
175
42
86
73
51
0000
0000
SDS
105
188
0
17:57:03
434
45
161
175
42
86
74
51
0000
0000
SDS
105
188
0
18:12:03
415
38
161
175
42
86
73
51
0000
0000
SDS
105
188
0
18:27:03
450
29
161
175
42
86
73
51
0000
0000
SDS
105
188
0
18:42:03
424
34
161
175
42
86
74
51
0000
0000
SDS
105
188
0
18:57:03
414
35
161
175
42
86
73
51
0000
0000
SDS
105
188
0
19:12:03
408
32
161
175
42
86
74
51
0000
0000
SDS
105
188
0
19:27:03
428
27
161
175
42
86
73
51
0000
0000
SDS
105
188
0
19:42:03
438
32
161
175
42
86
73
51
0000
0000
SDS
105
188
0
19:57:03
460
39
161
175
42
86
74
51
0000
0000
SDS
105
188
0
20:12:03
381
38
161
175
42
86
73
51
0000
0000
SDS
105
188
0
20:27:03
431
31
161
174
42
86
73
51
0000
0000
SDS
105
189
0
20:42:03
410
32
161
174
42
86
73
51
0000
0000
SDS
105
188
0
20:57:03
422
29
161
174
42
86
74
51
0000
0000
SDS
105
188
0
21:12:03
420
34
161
174
42
86
74
51
0000
0000
SDS
105
188
0
21:27:03
396
33
161
174
42
86
74
51
0000
0000
SDS
105
188
0
21:42:03
453
21
161
174
42
86
74
51
0000
0000
SDS
105
188
0
21:57:03
405
34
161
174
42
86
73
51
0000
0000
SDS
105
188
0
22:12:03
396
30
163
174
42
86
74
51
0000
0000
DDS
202
188
0
22:27:03
433
33
161
173
42
86
74
51
0000
0000
SDS
105
188
0
22:42:03
364
37
163
173
42
86
74
51
0000
0000
DDS
202
188
0
22:57:03
422
36
163
173
42
86
73
51
0000
0000
DDS
202
187
0
23:12:03
442
29
163
173
42
86
73
51
0000
0000
DDS
202
187
0
23:27:03
379
31
163
173
42
86
73
51
0000
0000
DDS
202
188
0
23:42:03
373
22
163
173
42
86
73
51
0000
0000
DDS
202
187
0
23:57:03
388
39
163
173
42
86
74
51
0000
0000
DDS
202
187
0
00:12:03
379
27
163
172
42
86
74
51
0000
0000
DDS
202
187
0
00:27:03
373
28
163
172
42
86
73
51
0000
0000
DDS
202
187
0
00:42:03
426
31
163
172
42
86
73
51
0000
0000
DDS
202
187
0
00:57:03
360
26
163
172
42
86
74
51
0000
0000
DDS
202
187
0
01:12:03
380
41
163
172
42
86
73
51
0000
0000
DDS
202
187
0
01:27:03
365
26
163
172
42
86
73
51
0000
0000
DDS
202
187
0
01:42:03
393
28
163
171
42
86
73
51
0000
0000
DDS
202
187
0
01:57:03
411
32
163
171
42
86
74
51
0000
0000
DDS
202
187
0
02:12:03
450
30
163
171
42
86
73
51
0000
0000
DDS
202
187
0
02:27:03
405
36
163
171
42
86
74
51
0000
0000
DDS
202
186
0
02:42:03
399
29
163
171
42
86
73
51
0000
0000
DDS
202
186
0
02:57:03
385
29
163
171
42
86
74
51
0000
0000
DDS
202
187
0
03:12:03
382
28
163
171
42
86
73
51
0000
0000
DDS
202
186
0
03:27:03
385
39
163
171
42
86
73
51
0000
0000
DDS
202
186
0
03:42:03
400
30
163
171
42
86
73
51
0000
0000
DDS
202
186
0
03:57:03
428
31
163
171
42
86
73
51
0000
0000
ODS
202
186
0
04:12:03
372
42
163
171
42
86
73
51
0000
0000
DDS
202
186
0
04:27:03
371
43
163
170
42
86
73
51
0000
0000
DDS
202
186
0
04:42:03
421
24
163
171
42
86
73
51
0000
0000
DDS
202
186
0
04:57:03
397
21
163
170
42
86
73
51
0000
0000
DDS
202
186
0
05:12:03
406
47
163
170
42
86
73
51
0000
0000
DDS
202
186
0
05:27:03
425
26
163
171
42
86
73
51
0000
0000
DDS
202
186
0
05:42:03
377
43
163
171
42
86
73
51
0000
0000
DDS
202
186
0
05:57:03
361
38
163
171
42
86
74
51
0000
0000
DDS
202
185
0
06:12:03
421
45
163
171
42
86
73
51
0000
0000
DDS
20 2
186
0
06:27:03
461
21
163
171
42
86
73
51
0000
0000
DDS
202
185
0
06:42:03
434
34
163
171
42
86
73
51
0000
0000
DDS
202
185
0
06:57:03
424
27
163
171
42
86
73
51
0000
0000
DDS
202
186
0
07:12:03
436
44
163
172
42
86
73
51
0000
0000
DDS
202
186
0
Figure 2. (continued).
S
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
17
-------
BAROMETRIC PRESSURE
¦TEMPERATURE
¦RADON GAS
1 TAMPER II
0,0217 "
0,0206
0,0195
^ D,0184
JO,0174
tiJ 0,0163
: 0,0152
11)0,0141
j0,0130
0,0119
00,0109
Z n,0098
Hi, 0087
^ 0,0076
(£0,0065
00,0054
1 0,0043
0,0033
0,0022
0,0011
TAMPER
101 Geiiini Road,
Bel Air, MD
VRflDONICS
The Radon Specialists
Unit!,iiiJiiiWMi.li.it.
W	
r
Tine 7'40 11 40 15 40 19:40 23:40 3:40 7:40 11:40
If? 3/12 3/12 3/12 3/12 3/12 3/13 3/13 3/13
Date J/1*	TIME 0F jEST
Figure 3. Graph showing dramatic variances in radon levels.
18
-------
USE OF GRAB SAMPLES AS A QUALITY ASSURANCE TOOL
TO ENHANCE OVERALL RADON MEASUREMENT ACCURACY AND REPRODUCIBILITY
by: Brian Fimian
Radonics, Inc.
McLean, VA 22102
Richard Tucker
Gemini Research
Cockeyaville, MD 21030
ABSTRACT
Discussion regarding the benefit of grab sampling to place measurement
devices, ensure primary measurement device operation, and identify owner
induced variances within the structure (tampering). The paper will
demonstrate how grab sampling as a secondary measurement strategy is becoming
an excellent quality assurance and evaluation tool for radon measurements
during real estate transactions.
-------
Quality assurance is the single most important aspect of any radon gas or
radon progeny measurement. This is true in all cases but it has special
significance when measurements are conducted in the context of a real estate
transaction.
A significant percentage of the residential radon testing that is
currently taking place is being performed during a real estate transaction
Because of the time constraints that are typical in this environment
purchasers of real estate are frequently limited to a single radon measurement
on the property they are purchasing; often a measurement upon which all
parties involved in the transaction place complete reliance. This underscore
the importance of measurement firms producing the highest quality result	*
within the imposed time limitations and demonstrates the importance of
providing the maximum level of quality assurance throughout the measurement
process.
The focus of this paper is to demonstrate how the use of instantaneous
grab samples can provide a meaningful contribution to the quality assurance
process when performing radon/radon progeny measurements in the context of
real estate transaction. It must be noted here that grab sample measurement
are entirely inappropriate as a primary radon measurement method because of 8
their short term nature and inability to correctly represent average
radon/radon progeny concentrations in a given structure.
They are, however, an excellent secondary technique for relative radon
assessment and, if utilized properly, can provide a very useful source of
information to supplement the primary measurement.
There are several advantages to performing grab sampling as a secondar
measurement technique and as a quality assurance tool in support of a prim
measurement strategy. The focus of this paper will be the three most
beneficial uses of instantaneous air samples during radon or radon progeny
measurements. They are as follows:
o Proper deployment location of the primary system.
o Corroboration of data from the primary system.
o Tamper detection.
PROPER DEPLOYMENT LOCATION OF THE PRIMARY SYSTEM
The Environmental Protection Agency (EPA) has stated in the February 190
Interim Protocols For Screening And Follow-up Radon And Radon Decay Product 7
Measurements that 'The screening measurement should be made in the room or
area in which the highest and most stable radon or radon decay product
concentration is expected."
Screening measurements are intentionally conducted in areas of the
property where one expects to find the highest radon concentrations and
additionally are performed under "worst case" conditions (i.e. EPA defined
1
-------
"closed house" conditions in the lowest livable area regardless of whether or
not that area is occupied). This is done to "...estimate the highest
potential concentration to which an occupant may be exposed." This insures
that most properties with the potential for elevated radon are not overlooked
in the screening process (i.e. false negatives).
One of the easiest ways to determine the appropriate placement of a
primary device is to use instantaneous air samples as a guide for locating the
highest and most stable radon/radon progeny concentrations in the structure.
In theory, equal mixing of the radon gas is both predicted and assumed to
take place at distances greater than a few feet from the source, usually the
underlying soil for radon infiltration. What actually occurs, however, is a
phenomenon that may allow concentrations in one area of a basement (say a
bedroom) to be as much as 50% higher than concentrations in another area.
This is due in part to the ventilation patterns within a given structure and
other factors such as the HVAC system design and the building construction
techniques.
To the extent with which the EPA guidance intends to minimize "false
negatives" and to the extent with which a measurement firm is interested in
performing a high quality and "meaningful" service, some form of quality
assurance must take place to insure that the primary measurement system is
properly located within the structure. Instantaneous air samples are one
means of accomplishing this end. In fact, because of the concentration
differences that sometimes occur, even on the same level of a structure, it is
prudent to ascertain the area of the structure which has the highest
concentrations of radon in order to adequately assess health risk. The
primary measurement device should be deployed in that area. Grab sampling is
a very effective tool in achieving that goal.
CORROBORATION OF DATA FROM THE PRIMARY SYSTEM
A second and very important benefit of employing grab samples as a
quality assurance mechanism in the context of a real estate transfer is the
comparison of data from the primary system to the instantaneous air samples.
This is important for a number of reasons, not the least of which is to ensure
that the primary system is operating correctly.
Though grab sampling is limited to providing corroboration with a primary
measurement only at those times when instantaneous samples are possible
(typically during device installation and retrieval), it at least provides
somewhat of a baseline measurement which can be used for comparison. The
normal corroboration that takes place between the two measurement methods
employed often provides a source of credibility to the primary measurement
system and a level of comfort that the reported results are both valid and
meaningful.
This approach for providing quality assurance is not new. The Department
of Energy (DOE) and the EPA have used this technique in their chamber
2
-------
facilities for verifying radon and radon progeny concentrations during the
chamber cross-check and intercomparison program for years.
It is appropriate to note that any corroboration of this secondary
technique with a primary system must be conducted in 'like* units (i.e. both
primary and secondary measurements should be either radon gas or radon
progeny). It is useless to draw any conclusions from the comparison of a
primary and secondary test if, for example, the primary system is a radon gas
measurement and the secondary test is working level or vice versa. This is
primarily because of the unpredictable nature of the equilibrium ratio between
gas and progeny from structure to structure and the difficulty in 'time
calibrating" concentrations between gas and progeny.
Figure 1 illustrates the usefulness of grab sampling as a secondary tool
when measurements are made in "like' units. This example is specific to a
primary continuous data logging device, either working level or gas. As
indicated by the graph, instantaneous air samples are expected to be in the
range (darkened portion of the two bars) of the continuous measurement data
machine time correlated both at test initiation and at test completion when
the results of the primary system are analyzed. If clear corroboration with
the primary measurement system does not exist in the review process, then
perhaps the primary device has either failed or the testing environment has
been compromised. Of course, there is no substitute for a high quality
measurement in the real estate environment and therefore, an effort must be
made to use a high quality primary measurement system.
TAMPER DETECTION
Because of the time-compressed nature of the real estate industry and the
enormous economic incentive of a property owner not to have radon, tampering
with measurement devices and compromising measurement conditions have become
commonplace in residential real estate transfers. Though there are many novel
approaches to tamper detection, not the least of which is taping closed all
the windows and doors in the structure during the test period, the most
obvious means of identifying and deterring tampering is through the use of
continuous data logging measurements. Many continuous monitors have at least
some ability to detect rapid fluctuations in radon or radon progeny
concentrations, especially when they are attributable to an open door or
window. These devices, by their very nature, are therefore much more adept at
tamper detection and even deterrence than are passive devices.
Another technique for detecting impropriety is using instantaneous air
samples in areas of the home which otherwise might be difficult to ventilate
It is not uncommon to detect elevated radon/radon progeny levels in the
interior rooms of a basement or in a closet area when measurements conducted
in open basement areas produce no evidence of elevated radon. ThiB might
indicate tampering especially if the only source of infiltration for the
interior room or closet area is from the underlying soil. Certainly,
consideration must be given to the fact that these rooms are often unoccupied
and have little or no natural ventilation. But what must also be given due
3
-------
consideration the ability of grab sampling to provide measurement firms with a
baseline concentration against which the rest of the structure is measured.
There are some other benefits for performing grab samples that are not
quality assurance related. Grab sampling devices can be simple tools for
qualifying a potential radon hazard. The most common uses are for diagnosis
of relative radon concentrations in a structure known to have elevated radon
levels. These measurements often provide a mitigation contractor with enough
information to allow him to make the most appropriate mitigation design
decisions. Additionally, verifying proper remediation procedures by spotting
leaks on the positive pressure side of a mitigation system or potential re-
entrainment of the exhausted gas is easily accomplished by utilizing
instantaneous air samples.
Still another very important benefit is found in verifying whether or not
measured airborne concentrations of radon are attributable to radon in
household water or simply to the underlying soil. This is most easily
accomplished by running a shower in a bathroom for a few minutes and then
obtaining an instantaneous radon/radon progeny concentration in that area.
Though this is not appropriate for exacting quantitative concentrations of
radon in the water, it does provide a measurement firm with information useful
in isolating the source of the radon problem if levels are elevated.
In concluding, one fact is clear. Great reliance is all too often placed
on a single primary measurement during the course of a real estate transfer.
It is precisely because of this fact that all measurement firms MUST take the
appropriate quality assurance measures specific to their testing techniques
and work to provide the general public with accurate results that are both
meaningful and valid. Only those primary measurement strategies which can
stand the test of time and provide a high level of reproducibility will have a
place in the radon industry. Instantaneous air sampling as a secondary
measurement method is an excellent quality assurance tool for achieving thiB
end.
The work described in this paper was not funded by the U.S. Environmental
Protection Agency and therefore the contents do not necessarily reflect the
views of the Agency and no official endorsement should be inferred.
4
-------

TEST
INITIATION
Figure 1. Grab correlation.
-------
Session III:
Measurement Methods -- PANEL
"Short-term/Long-term Measurement"
-------
III-
PREDICTING LONG-TERM INDOOR RADON CONCENTRATIONS FROM SHORT-
TERM MEASUREMENTS: EVALUATION OF A METHOD INVOLVING
TEMPERATURE CORRECTION
by: T. Agami Reddy, A. Cavallo, K. Gadsby, and R. Socolow
Center for Energy and Environmental Studies
Princeton University
Princeton, NJ 08544
ABSTRACT
Most studies which seek to determine uncertainty bounds in predi t-
long-term indoor radon concentration from short-term measurements d° ^
assuming radon variability to be a random quantity. The objective'of ° So
paper is to evaluate the potential of decreasing these uncertainty bounri^^^®
one assumes indoor radon variations to be in part influenced by certain 1-?
varying known physical driving forces. From daily averaged data from
occupied unmitigated residences, for which continuous measurements were ^ e
for about a year, the stack effect (as also the ambient temperature) has K^n
identified as the predominant physical driving force. We find th t- ®en
uncertainty bounds for predicting long-term radon concentrations, when ex 3
recognition is given to the year-long variation in stack effect, are r rt
drastically in one house, less so in another, and marginally in the
Probable physical causes behind these observations are also discussed
general mathematical equation is derived for predicting these unceri- " ^
bounds in terms of climatic variability, a factor dependent on hou ^
surrounding soil characteristics, and the strength of the physical ^
Though the mathematical equation is correct within the framework	.
assumptions made, more associated studies and analysis involving a lara
base are required before the benefits and scope of this technique co*!
fully appreciated in terms of practical applicability and relevance UJ-C^ fee
This paper has been reviewed in accordance with the U.S. Environ
Protection Agency's peer and administrative review policies and approv™^*1*"^
presentation and publication.1	for
STATEMENT OF PROBLEM
The issue of defining bounds to the uncertainty associated with predict^
xThis work was funded by the U.S. Environmental Protection a
under Cooperative Agreement No. CR-817013.	Agency
-------
the long-term (by which one generally implies, yearly) average2 of indoor radon
concentrations from single short-term (i.e., 1 to 15 day period) observations
has generated great interest in the radon research community. This has arisen
not only because of the practical implications in terms of health hazards to
inhabitants, but also because of mandatory indoor radon testing laws required
for realty transactions. The problem is especially complex given that indoor
radon concentrations vary widely during the day, from day-to-day, and often
show strong seasonal patterns which are house specific since they depend upon
soil type, climate, house construction, house dynamics, and occupant behavior
(1-5). Though the number of studies addressing the issue of predicting yearly
indoor radon levels from short-term screening tests is relatively small, it
would nevertheless be appropriate to start by taking stock of past research in
this area.
There are basically two types of research thrusts: (a) one that involves
analysis of survey data from a large number of houses, for example (6), and (b)
one that involves detailed analysis of a few houses in which continuous
measurements have been performed (3-5). The advantage of approach (a) is that
it enables statistically rational and generalizable uncertainty bounds to be
determined, while the disadvantage is that the uncertainty bounds are rather
large. Ref. (6) finds an uncertainty range of 5 times the short-term screening
value for 95% confidence level when no consideration is given to time of year,
and of the order of 3 times when one explicitly considers the season during
which the single measurement was performed. One way of decreasing these
uncertainty bounds is to perform additional survey tests with stratified
sampling which consists of partitioning the population into groups each of
which is more homogeneous than the population itself. The stratified sampling
could distinguish between season, geographic location, soil type, house
construction and equipment type. Such an approach, which has been used in
previous studies, for example, Ref. (7), could be investigated in the framework
of certain current programs, for example, the Florida Radon Research Program
(FRRP) (8).
The basic disadvantage of approach (b) is that practical and generalizable
uncertainty bounds are difficult to establish given the wide differences from
one house to another. However, what such an approach does provide is insight
into the day-to-day variability of indoor radon concentrations and how and to
what extent these are affected by the various climatic and house-specific
parameters. Such information would also enable sound experimental design and
proper identification of the sample strata in the framework of approach (a).
Indoor radon concentrations vary widely from day-to-day and also show
strong seasonal patterns (4,5). A former study (9) had suggested that an
average of screening measurements taken during two different seasons of the
year would provide a more satisfactory estimate of the yearly average than a
2Current scientific thinking seems to assume that the arithmetic
mean concentration is more representative of the exposure risk than are
other indices, such as median or geometric mean.
2
-------
single measurement. A short-term measurement strategy which involved
performing measurements during each of the four seasons of the year (althouah
impractical in terms of actual implementation) was shown to provide long-term
estimates within 25% accuracy (when the associated instrument error
overlooked).
Parameter sets affecting indoor radon concentrations can be divided into
three groups. The first includes the intrinsic properties of the soil and
the location and concentration of the radon source with respect to the house
The second set is made up of the characteristics of the building sub-structur"
and super-structure and of the equipment in the house. The third set consist
of climatic parameters like ambient temperature, and wind magnitude h
direction. The coupled influence of the second and third sets is general]1
acknowledged to be the primary cause of the day-to-day year-lonq variahi -| >
of indoor radon concentrations, while the mean concentration level is laro—7^
influenced by the first set of parameters. Note that the primary concern -5^
the present study is to capture the variability of radon and not to predict t-K11
magnitude of the mean concentration level as such.	e
Thus, predicting long-term indoor radon concentrations from short-t
measurements is essentially an uncertain process since indoor rajrm
concentrations vary during the year. Most studies, though impliCit?n
acknowledging that this variation is the result of variation in cert •
physical driving forces, have limited themselves to treating (i.e., analyzi^"1
indoor radon concentration data as made up of random observations. n!?
specific objective of this study is to evaluate a technique whereby the ind
radon data are analyzed as being the response of a physical system su' ' °°r
varying input physical forces. Since random effects are bound to be pres
n data are analyzed as being the response of a physical system subject
ing input physical forces. Since random effects are bound to be pres-tQ
in any physical system, the total observed radon variation over a year can
visualized as consisting of two components: a deterministic comr>^~
ent
be
resulting from certain physical forces, and an unexplained random compone6^
The practical relevance of such an approach is that it would have the potent^'
of decreasing the uncertainty bounds, at a prespecified confidence level
predicting the long-term indoor radon concentration when a single short-t- in
measurement is made.	erm
DESCRIPTION OF DATA
This study is based on year-long continuous data collected by Princ***-
University in three occupied residences, designated H2, H21, and H22, in 5on
Princeton area of New Jersey. H2 is a two-story structure with a full basem
made up of hollow cinder-block walls, and an attached garage that was built^1"1*"
1980. It has very little tree cover and sits in the middle of several a
of open land. Heating is provided by a gas fired forced air heating Sv
while cooling is supplied by a central air-conditioner. The house has a err m
bed under the slab, while the soil underneath is relatively impermeable v®-^
H21 is a single-story ranch-style house with a partial basement
remainder of the house being of slab-Qn-grade construction. This house wK*
' w"ich
3
-------
is about 30 years old, is surrounded with trees. The basement walls are of
hollow cinder block. This house also has a gravel bed under the slab. The
heating system is a gas furnace while a central air-conditioner supplies
cooling.
H22 is a 60 year old balloon-construction three-story house with a partial
basement and floor drains. There is tree coverage on two sides while the other
two sides are exposed. Unlike the other two houses, the subslab material is
soil. The house is heated by radiators, while cooling is provided by a central
air-conditioner. Because of this, the air handler is used only for cooling.
Detailed descriptions of the houses and of the continuous data taken during the
period the houses were unmitigated can be found in Ref. (4).
We have screened and reduced the data stored as 1/2 hour averages into daily
averages since this is more appropriate for this study. Periods during which
data were available for all three houses are given in Table 1, while the
parameters selected for analysis are described in Table 2. Variations in
temperature differences are often more appropriate than those of temperature
to explain indoor radon variations (4). For example, differences between TB
and TA have been designated as TBA in this study. Table 3 assembles the mean
and standard deviation of the various parameters over the entire period during
which data were available. One notes that the standard deviations are
generally large compared to the mean values for most parameters.
IDENTIFICATION OF PHYSICAL MODEL
The first step is to describe the system in terms of a model. One
approach is to construct a physical model based on mass balances akin to that
of, say, Ref. (10). This approach is not only involved mathematically but is
also house specific in that the physical geometry of the house dictates the
inclusion, or exclusion, of certain air and radon flow paths, which themselves
may be uncertain. An alternative approach, and the one adopted in this study,
is to formulate a statistical model; for example, a simple linear regression
model. We shall have to identify the model parameters (i.e., the important
driving forces) and the regression coefficients from the data at hand.
Table 4 presents the correlation coefficients [see any statistics book,
say (11), for definition] of the radon quantities (RNB and RNL) with the other
parameters which are described in Table 2. TLA is strongly collinear with TBA
and has not been included in Table 4. We note that the correlation
coefficients of H21 are strongest while those of H2 and H22 are lower [which
is consistent with Ref. (4)]. What is most surprising is that RNL variability
is much better explained (i.e., has stronger correlation coefficients) than
that of RNB. One would have expected the reverse since the conventional
understanding is that soil gas first enters the house via the basement from
where it finds its way to the living area by a combination of several
house-specific pathways. The stack effect, represented by TBA (and TLA), seems
to be the most important [again, consistent with Ref. (4)]. The effect of HAC
on indoor radon values is smaller (correlation coefficients about 0.25).
Moreover, since TLA and HAC are collinear, there does not seem to be much
4
-------
• „ mnriPls for radon with HAC as a second variable
incentive in using «9reS3o.on ~de	^ £or ^ the interpretation of the
(4)\ Though TSB seems c°lll"e1arnarameter may be spurious, given that the
TSB measurement as a phYSxc^parame ^ Jgement to be affected by both
temperature probe ^ =r°tures (4) .
basement and soi	v
fhP adjusted coefficients of determination
Table 5 assembles the values btained by a linear regression of indoor
(11) i e., the adjusted R values , different models. We note that there
radon parameters (RNB and RNL) with four dif^ ^ r2 vaiues) Qf ^
is much greater variation in qua	models # whatever variability the
regression models across hou	, ted by certain house-level factors that are
models fail to account for is domi	meter sets chosen. Radon models f0r
not greatly influenced by the mode p q g _ o>8) degpite the fact that the
H21 are generally satisfac y large portion of the time. On the other
basement window was °?en *u""gextremely poor, an occurrence which could be
hand, models for RNB in H2 a	subsoii is relatively fine-grained and
attributed to the fact that	sistance to radon migration m the soil,
compacted thereby offering large	forcing functions, the resulting
Consequently, for the	«uld be less important than in other houses,
variation in indoor rado	poor. Probable causes are that the house has
Models for RNB in H22	eriments indicate the presence of short-circuit aiE
distinct zones and^prithe attic via the walls,
flow paths from the subsiao
wh
We find that models with TA (Model 1) or TBA (Model 2) are equally good
ile there does not seem to be any advantage in including HAC as an additional
parameter. A previous study (12) had indicated that at half—hour tin»
intervals the physical mechanism affecting radon entry into the basement j6
akin to a one-way valve dictated by temperature differences between soi3_S
basement, and ambient. Consequently, we have also investigated a model
explicitly separating the positive and negative values of TBA. This Pertaint
to Model 3 of Table 5. We note that, though Model 3 has higher R2 values, th
improvement is generally only a few percentage points and does not justif^
the added complexity in the model structure when daily time scales of averagin
are used.
We have also investigated model structures of the form RNB, rnl = f (tb^
c)+ where c is a coefficient to be determined by regression and the + si ~~
indicates that only positive values are retained in the regression analysis**
This model structure, it will be noted, is akin to that used in building ener^"
studies where comfort energy requirements are often regressed against degree'
days (13) . The improvement in R2 values of such a model over those of Model"
30ne of the findings of Ref. (4) was that there was limited, if no
incentive in formulating a model for indoor radon levels over the enti '
year which included HAC as a second variable. However for models on—^
seasonal basis, the inclusion of HAC does improve the models. Th*» &
conclusions are however specific to the scope of Ref. (4) which S
limited to three residences in central New Jersey.	Was
5
-------
1 and 2 was at most a few percentage points, while R2 values were lower than
those of Model 3.
ALTERNATIVE APPROACH TO ANALYZING INDOOR RADON DATA
In this section we shall seek to determine whether, and by how much, the
two following approches of analyzing data narrow down the confidence bounds,
or alternatively, the percentiles (11):
(a)	entire variation of indoor radon concentration over the year is random.
As noted earlier, this is the approach followed by most studies to
date. In this case we shall merely inspect the data series of the
normalized variable (RN/RN) where RN could be either RNB or RNL, and
RN is the long-term (i.e., the annual) average of RN;
(b)	variation of indoor radon concentration is partly the result of
variation in certain known physical forces which drive indoor radon.
Only the residual variation, or the variation in indoor radon not
explained by the model, is random.
The statistical analysis in the previous section suggested [as also did
several studies, say (4)] that the most influential parameter which explains
indoor radon levels is the stack effect, characterized by TBA or by TA. The
following model structure is used to describe the output of the system:
RN± = a' + b' ' TBAt	(1)
where RN could be either RNB or RNL,
a' and b' are the intercept and slope of the linear regression line,
subscript i represents individual observations, and
RN is quantity deduced from the model rather than from measured data.
If TBA and RN are the long-term (i.e., the annual) averages of TBA and RN,
respectively, then
(RNi/RNJ = (a' + b' • TBAj) / (a' + b' ¦ TBA)	(2)
Subsequently, replacing RNt by RNlf we have
/\ 	
RNj. = RN, ' (a'+b* " TBA) / (a'+b' • TBAJ	(3)
A
Note that R^ would be the value of the long-term indoor radon concentrations
predicted from an individual or short-term observation RNt by applying the
temperature correction approach. Finally, this value has been normalized by
dividing it by RN, where RN is the long-term average of RN deduced from data
(and assembled in Table 3). The data available for all three houses have been
processed both as explained above and _also by assuming them to be random; i.e.,
by merely dividing the RNi values by RN.
6
-------
r,f the associated distribution of ^aily vaiues without
^	data series! and with [i.e., of	data series]
[i.e., of (RN,/ >	aiven in Fig. 1. We note that there is a marked
temperature correcJ J; ainty bounds for the indoor radon levels of H21, a
decrease in the u	Y	negligible improvement for H2 . These are
smaller	£ the Associated regres sion model, i.e., higher
consistent with the R values o£ t	interpretatiori o£ the nultoers ih Fig.
the R' value more the	the result^ o£ RN1 for H21 seem to indicate
ih"t Sweracould hope 'to narrow the 90% uncertainty bounds in predicting the
annual radonlevels from a factor of 2.4 with no temperature correction down
to 1.4 When the temperature correction is applie .
STATISTICAL METHODOLOGY TO DETERMINE BOUNDS ON PREDICTION ACCURACY
t +- • /-v.-, -i n the previous section was limited since actual
The scope of the evalua^° the Princeton area were available. Though we
data from only three houses	"advantage in our approach, reappraisal
were unable to f-^^Xs seeras justified. In this section, we shall
in the framework of future	t predict the theoretical uncertainty bounds
derive a mathematical equat.	a^h This wouid permit our approach to be
resulting from our	*7 . t ' different types of houses and soil
generalized to any climate
conditions.
a .imole linear model structure between indoor radon
We shall ass^mJ *	driving force (say, ambient temperature since it is
concentration and a s 9	and has been found to be as good a
= triable easier to obtain uia»
predictor of EN as is TEA) such as:
(4)
RN = a + b ¦ TA
v	in the data and also that the effects of other
„ ,6irfo?«sE6«e overlooked, the model will not be a perfect fit. This can
be"1 represented statistically as (ID:
(5>
RN± = a + b TAi + £i
where e, is the error term in the individual observations.
• « fhat- the observed RN variability could be due to a large
The model implies that f	(i.e., TA) along with a small coupling
variability in the	9 versa_ Thus we have separated the problem of
coefficient (i.e., o)	mtv into a location-dependent climatic effect
long-term indoor rado	location-, and house-characteristics-dependent
and a climate-independent, location ,
effect.
7
-------
If the variables RN and TA are assumed to be normally distributed variables4
with no serial correlation, e will be normally distributed, have zero mean, and
a constant variance of o2 (e) ; i.e., homoscedasticity is assumed in the basic
physical process. Let n be the number of observations and R2 the goodness-
of-fit of the model given by eq. (4) . Then, from the definition of R2 (11) :
R =
I
1-1
.'N — ,
(RNi - RN)2
(6)
I
i-1
(RNj - RN):
where
Also

RNX is the model predicted value [from eq.(4)],
RN, is the observed value, and
RN the long-term average of RN.
n		
Z (RNt - RN)2
i-1
= I
i-1
(TAi - TA):
= b2 • TA ' (n-1)
Introducing this in eq. (6) we have
f TAa
,TA
V
(7)
RN
( RNi ^
RN
b2
TA
O
( TAA
^TA
(8)
From the above and from the definition of R2, we find
1 - R2
(e,) = -
¦ TA
R
( TA^
TA
(9)
The standard deviation of the standardized quantity (e^RN), which is
analogous to the Coefficient of Variation (11), is finally obtained
( eA
VRN,
/ 1 - R2 \
1/2
R2
( TAa
yTA
/ a
+ 1
b ' TA
(10)
Eq. (10) is simply an equation which correlates the variation (quantified
'Several studies [for example Ref. (4)] have found that the variable
TA (and TBA) exhibits a normal distribution over an entire year, while
indoor radon variables have no consistent agreement with either a normal
or a log-normal distribution, though the latter is usually better.
8
-------
by the standard deviation) of RN not explained by the model in terms of three
sets of parameters describing:
(a)	location_ specific variation in the ambient temperature; i.e.
a (TAj/TA) ;
(b)	house and surrounding soil dependent quantity specified by the factor
(a/b), which has units of °C; and
(c)	strength of the regression model between RN and TA designated by th
R2 value. Recall that the physical interpretation of the R2 value 1
that it represents the percentage of the total variation in +-hS
response variable explained by (i.e., directly the result of variati 6
in) the exogenous variable.
Since the variables RNt and TAi are assumed to be normally distributed	th
critical values at different significance levels would correspond to'	t-v6
uncertainty ratios at different probability levels. For example, a	qci
probability level would correspond to (2 ' a) [see Ref. (11)].
The above derivation could be easily extended to linear model structu
with more than one driving force. One could adopt a similar methodology
the more-often-encountered case when the variable RN is not normal
distributed while the variable TA is.	-*-ly
APPLICATION TO ACTUAL DATA
We shall illustrate how eq. (10) could be applied to specific locati
From 1 year's data of daily TA values_provided by NOAA (14), we find f0r°!??'
Princeton area, O (Th) z 8.5°C while TA = 12.8°C. Values of the parameter
and b of eq. (4) are assembled in Table 6 for the three houses. From Tabl 3 3
we find (a/b) factors for RNB to be 78 for H2, 21.8 for H21, and 145 for
Interestingly, H21 is a one-story residence; H2, two-story, and H22, th
story. Thus one notes that (a/b) factors seem to increase with the heiqht-66"
the building. This observation is perhaps premature and needs to be evai,, of
further.	ated
How the theoretical standard deviation of the variable (ex/RN) would
with R2 for a wide range of (a/b) values for the Princeton area is show^^
Fig. 2 generated from eq. (10) . From the limited number of houses stun ^"n
(other than H21 where basement window opening may be an abnormal occurren
values of (a/b) are in the 80-150 range. Even for low values of R2 (=0 2)	'
notes that the temperature correction approach could result in predi'+-?ne
intervals at the 95% confidence level (i.e., 2 standard deviations)0
exceeding 1.4. This is a significant observation since it implies
uncertainty bounds of prediction can be drastically reduced by our nhv
approach even in a house where the indoor radon variability is	Cal
influenced by variation in the stack effect.	We^kly
As a preliminary illustration, Table 7 assembles values of a (rn\	_
and. ta
9
-------
for a few locations in the U.S for different averaging times. To within one
decimal accuracy, the arithmetic mean is essentially independent of averaging
time interval while the standard deviation decreases with length of averaging
interval. If the standard deviation values for each location are normalized
with respect to the 1-day value, we note that the decrease with averaging time
is fairly linear and location independent (Fig. 3) . Thus, we note that an
averaging interval of 15 days will lead to a 35% decrease in the standard
deviation of the ambient temperature variability over the year as compared to
a 1-day time scale of averaging, while an averaging interval of 1 week would
result in a 20% decrease. Though an exponential fit to these data points would
be more accurate, we find that a linear fit to the normalized standard
deviation versus averaging time (in days) yields an R2 of 0.93 with a slope of
-0.023 (SEM = 0 .002) .
The possible range of variation of the factor (a/b), representative of the
soil conditions and house construction practices prevalent in widely different
geographic locations in the U.S., is unknown at present. Either analyzing
existing radon survey data or gathering data explicitly for this purpose may
be tasks worth evaluating in the framework of future radon projects. An
alternate, and perhaps more promising, approach is to infer the parameters a
and b from the house response when certain simple "stressed" experiments on the
house are performed. Such experimental protocols have yet to be satisfactorily
formulated and validated, but initial attempts are underway in the Research
House Study of the FRRP (8) . Efforts such as the above would, hopefully,
permit numerical values of a and b to be specified dependent on generic
building construction type and soil conditions.
CONCLUSIONS
The physical approach advocated in this study, whereby one visualizes indoor
radon variations as the response of a physical system acted upon by certain
varying and known forces, has been shown to have the potential of decreasing
the uncertainty bounds associated with the problem of having to predict long-
term indoor radon levels from short-term screening tests. The physical system
can be described by a regression model with the stack effect as the single most
influential driving force. How such a model approach fares with respect to the
conventional procedure, of assuming indoor radon variability to be random, has
been evaluated with daily averaged data for over a year in three occupied
houses. It has been found to be distinctly advantageous in one house,
moderately advantageous in another, and marginally so in the third.
The theoretical uncertainty bounds of prediction resulting from the physical
approach can be predicted from a mathematically derived equation expressing the
normalized standard deviation of the variation of indoor radon not explained
by the model (i.e. the random component), in terms of three sets of parameters:
location-dependent statistics of ambient temperature, a factor describing the
coupling between the soil and the house, and the strength of the regression
model. How the equation could be applied to individual geographic locations
has been illustrated by generating a figure of the theoretical uncertainty
bounds for the Princeton area. An important observation is that the strength
10
-------
of the regression model is not a significant parameter provided the
corresponding Rz values of the regression model are greater than about 0 2
thereby suggesting that the approach could be potentially useful over a variet'
of housing stock and soil conditions. However, more associated studies and
analysis involving a larger data base are required before the benefits and
scope of the present technique could be fully appreciated in terms of practical
applicability and relevance.
ACKNOWLEDGEMENT S
We would like to acknowledge critical comments and discussions with W. r0
of PU/CEES, R. Moseley, and D. Sanchez of US EPA/AEERL, R. Leadbet'ter
University of North Carolina, and R.Sextro of LBL.
REFERENCES
1.	Nazaroff. W.W. and Nero, A.V. (Eds.), Radon and Its Decay Prortn^t-
Indoor Air, John Wiley and Sons', NY, 1988.	"	——
2.	Hopke, P.K., (Ed.), Radon and Its Decay Products. American Chem'
Society, 1987.	"U-cal
3.	Yuill, G.K. and Associates Ltd., A study of the statistics of r
measurements in houses, Report to Energy, Mines & Resources, Divi •°n
of Environmental Affairs, Ottawa, Canada, File 9290/92900604 MaSl°n
1989.	' rch'
4.	Reddy, T.A., Molineaux, F.B., Gadsby, K.J., and Socolow, r
Statistical Analyses of Radon Levels in Residences Using Weekly and n" ^ '
Averaged Data, PU/CEES Report No. 249, March 1990.	D^ily
5.	Hull, D.A., and Reddy, T.A., Study on the reliability of short—t-
measurements to predict long-term basement radon levels in a resid""
paper presented at the 1990 International Symposium on Radon and pnCe'
Reduction on Technology, Feb. 19-23, Atlanta, Georgia, 1990.	adon
6.	Roessler, C.E., Revell, J.W., and Wen, M.J., Temporal patterns of i
radon in North Central Florida and comparison of short-term monit^^0*
to long-term average, paper presented at the 1990 International Symt>°ring
on Radon and Radon Reduction Technology, Feb. 19-23, Atlanta r ^°sium
1990.	' 0lr9ia,
7.	Cohen, B.L. Surveys of radon levels in homes by University of Pitt h
Radon Project, paper presented at the 1990 International Symposi h
Radon and Radon Reduction Technology, Feb. 19-23, Atlanta, Georgia 11*9 90°
8.	Sanchez, D.C., Dixon, R., and Williamson, A.D., The Florida
research program: Systematic development of a basis for Stat-*"3*?00
standards, paper presented at the 1990 International Symposium on pW'*"c*e
and Radon Reduction Technology, Feb. 19-23, Atlanta, Georgia, 199Q °n
11
-------
9. Harley, N. and Terilli, T., Predicting annual average 222 Rn exposure,
Health Physics, 1988.
10.	Hubbard, L.M., Bolker, B., Socolow, R.H., Dickerhoff, D. and Mosley,
R.B., Radon dynamics in a house heated alternately by forced air and by
electric resistance, in Proceeding: the 1988 Symposium on Radon and Radon
Reduction Technology, Volume 1, EPA-600/9-89-006a (NTIS PB89-167480) ,
March 1989.
11.	Sachs, L. Applied Statistics, 2nd Ed. Springer-Verlag, New York, 1984.
12.	Hull, D., Time series linear regression of half-hourly radon levels in
a residence, paper presented at the 1990 International Symposium on Radon
and Radon Reduction Technology, Feb. 19-23, Atlanta, Georgia, 1990.
13.	ASHRAE, Fundamentals, American Society for Heating, Refrigerating and
Air-Conditioning Engineers, Atlanta, Georgia, 1985.
14.	National Oceanic and Atmospheric Administration (NOAA), "Local
Climatological Data - Monthly Summary," National Climatic Data Center,
Asheville, NC, 1978.
12
-------
TABLE 1. PERIODS DURING WHICH DATA WERE AVAILABLE.
House	Period	No. of
Months
H2	10/15/1986 - 6/24/1987	8
H21	1/15/1988 - 10/31/1988	10
H22	3/11/1988 - 9/28/1989	18
TABLE 2. DESCRIPTION OF VARIOUS PARAMETERS CHOSEN FOR THIS STUDY.
TA	-	Ambient dry-bulb temperature, (°C)
TB	-	Basement temperature, (°C)
TL	-	Living area temperature, (°C)
TBA	-	Difference between basement and ambient air temperatures, (°C)
TLA	-	Difference between living area and ambient air temperatures, (°C)
HAC - Fraction of the time during which the heating and air-conditio '
equipment was on,	lr*g
RNB	- Radon level in the basement, (pCi/L)
RNL	- Radon level in the living area, (pCi/L)
13
-------
TABLE 3. MEAN AND STANDARD DEVIATION OF CERTAIN IMPORTANT PARAMETERS FOR
ALL THREE HOUSES OVER THE ENTIRE PERIOD OF DATA AVAILABILITY.



H2


H21


H22



Arith-

Geo-
Arith-

Geo—
Arith-

Geo-


metic
St.
metric
metic
St.
metric
metic
St.
metric


Mean
Dev.
Mean
Mean
Dev.
Mean
Mean
Dev.
Mean
TA
(°C)
7.4
8.52
-
12.5
8.64
-
13.8
8.80
-
TB
<°C)
16.1
2.16
16.0
19. 4
2. 06
19.3
23.2
3.75
-
TL
(°C)
19.8
2.73
19.3
21.1
2.81
20.9
21.4
2.44
21.3
TBA
(°C)
8.6
6.72
-
6.8
7.92
-
9.4
9.48
-
TLA
(°C)
12.1
7.99
-
8.2
6.61
-
7.2
6.82
-
HAC
(" )
0.20
0.151
-
0.18
0.214
-
0.16
0.273
-
RNB
(pCi/L)
22.8
10.21
21.4
93.0
106.25
42.4
63.6
46.44
49.9
RNL
(pCi/L)
15.3
5.44
13.6
36.8
39.82
19.8
13.6
11.03
8.2
TABLE 4. CORRELATION COEFFICIENTS OF RADON WITH PHYSICAL PARAMETERS USING THE ENTIRE
DATA SET. THE VARIABLE TLA HAS NOT BEEN INCLUDED SINCE IT IS STRONGLY
COLLINEAR WITH TBA, AND THE STRENGTH OF THE CORRELATION COEFFICIENTS OF
THIS VARIABLE WITH RADON LEVELS IS ESSENTIALLY SIMILAR TO THAT OF TBA.


H2



H21




H22


TA
TBA
HAC
TSB
TA
TBA
HAC
TSB
TA
TBA
HAC
TSB
RNB
0.12
-0.09
-0.24
-
o
OD
H4
0.78
-0.20
0.04
0.05
-0.12
00
CM
O
1
0. 04
RNL
-0.52
0.49
0.29
-
-0.84
0.87
-0.02
0.51
-0.56
0.54
-0.29
1
o
o
14
-------
TABLE 5. ADJUSTED R2 VALUES OF DIFFERENT INDOOR RADON MODELS USING DAILY
AVERAGE DATA.

Model 1

Model 2


Model 3
Model 4
RNB







H2
0.02

0.01


0. 02
0.04
H21
0. 66

0. 62


0.64
0. 67
H22
0.02

0.01


0.04
-
RNL







H2
0.26

0.23


0.26
0.26
H21
0.71

0.76


0. 81
0. 81
H22
0.31

0.29


0.30
~
Model 1:
RNB, RNL = f
(TA)
Model
3:
RNB, RNL
= f [ (TBA)*,
(TBA) "]
Model 2 :
RNB, RNL = f
(TBA)
Model
4 :
RNB, RNL
= f (TA, HAC)

TABLE 6. VALUES OF THE REGRESSION COEFFICIENTS OF DAILY AVERAGE
INDOOR RADON USING THE LINEAR MODEL IN TA (Eq. 4).

Intercept
(pCi/L)
RNB
SEM Slope
(pCi/L/°C)
SEM
Intercept
(pCi/L)
RNL
SEM
Slope
(pCi/L/
SEM
°C)
H2
H21
H22
21.72
218.30
50. 62
0.94
6. 70
4.70
0.28
-9.99
0.35
0.16
0.44
0.29
17 .71
85.41
23.35
0.43
2.33
0.93
-0 . 33
-3.87
-0 . 71
0.04
° - 15
0.06
TABLE 7. YEARLY MEAN AND STANDARD DEVIATIONS OF DAILY AVERAGE AMBIENT
TEMPERATURE FOR A FEW LOCATIONS [FROM DATA SUPPLIED BY REF.(14)
FOR 1978]. THE ARITHMETIC MEAN VALUE FOR ALL LOCATIONS IS
ESSENTIALLY NOT AFFECTED BY THE TIME SCALE OF AVERAGING.
City
State
Mean

Standard
Deviation
(°C)

(°C)
1-day
3-day
7-day
15-day
1.Atlantic City
NJ
12 . 8
9.8
9.0
8.2
6.2
2.Houston
TX
19.3
7.9
7.2
6.5
5.2
3.Miami
FL
24.3
4.4
3.8
3.5
2.7
4.Newark
NJ
12.6
9.6
8.7
7.6
5.1
5.Portland
OR
12.9
6.0
5.5
5.0
4 . 0
6.Tallahassee
FL
19.1
7.4
6.8
6.3
4.9
15
-------
3 •
2-S -1
2j6 ¦
2.4 ¦
2-2 H
2 ¦
"1
1-6 -I
1.4 •
u H
i •
0 8 ¦
06 -
0.4 -
0.2 -
0 ¦
H2
~ Without
+ With
R2=0 . 01
so
65	70	73	80
PERCENTILES
i-q
a
&
3
24
26
2 <
22
2
U
14
1.4
12
\
OS
0.6
0.4
0 2
0
H2
~ Without
+ With
R2=0 .23
GO5 70 ~T> ST
PERCENTILES
2J
2
H21

D Without

+ With
RJ=0 . 62
6: io is to
PERCENTILES
k)
os
Without
With
R =0 . 7 6
<5	70 75	t0
PERCENTILES

H22

Without

With
0.1 ^
0.6 -I
0.4 J
24 4
H22
RJ=0 .01
" H

Without
With
OS
06 -I
0 .4 H
PERCENTILES	PERCENTILES
Figure 1. Percentiles of 1-day ratios of normalized basement and living
area radon concentrations with and without temperature
correction for all three houses. The associated R2 values of
the regression model given by eq. (1) are also shown.
-------
Model RN = a + b * TA
MODELR SQUARE
9 variation of the theoretical uncertainty bounds versus model
Figure 2. ^ar	.	(l0) for different values of the factor (a/b)
d for a day-to-day ambient temperature variation
corresponding to the Princeton, NJ;area.
-------
z
o
H
<
>
w
Q
Q
H
o
Cs
m
N
"<
s
DC
O
z
1
0.9
0.8
0.7 H
0.6
0.5
0.4 H
0.3
0.2
0.1 -\
0
1
-8,
n	Atlantic City
+	Houston
o	Miami
a	Newark
x	Portland
v	Tallahassee
"T"
4
"T~
6
i
10
~i I V
TIME SCALE OF AVERAGING (DAYS)
11
12
13
i
14
15
Figure 3. Normalized standard deviation values of ambient temperature
versus time scale of averaging for six locations. The
variation is close to being linear and is fairly location
independent.
-------
HI-18
CORRELATION BETWEEN SHORT- AND LONG-TERM
INDOOR RADON CONCENTRATIONS IN FLORIDA HOUSES
By: Susan McDonough and Ashley Williamson
Southern Research Institute
Birmingham, AL 35255-5305
and
David C. Sanchez
Air and Energy Engineering Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
ABSTRACT
th,0 oerformance standard for radon-resistant
In support of a	F1orida, a protocol is needed to provide
ruction for the State of Flori , ^ order to relate the result® of
cost-construction indoor ra^on_™®	inferred annual average concentrations, a
short term compliance ^easureme"egions of Florida known to have potential for
study is in progress in four r g	# in polk^ Alachua/ Dade, and Leon
elevated indoor radon. Eighty stu y	sing long-term (quarterly and annual
Counties are being	chambers) and short-term monitors
alpha track and	Q£e®^l.ter and short term electret-ion chambers) .
. p -	and barrier charcoal caniB«	and 2_ weelc intervals. a
£?oetrets are deployed contil?"ou®iYUB£na pylon AB-5 continuous radon monitors.
E1®c"etJ the houses are monitored usingi Pylon	Florida housing
were selected to be *ePrn^	the 2-20 pCi/L range. Data hav2
Houses	with indoor radon concentr variations and to derive confidence
construction, ^t late By8tematic	or multiple short-
^mits ^or predicted long-term (annual) ^^r%roto^olB. For relevant
£em measurements according^ th	thre8hoids have been determined below
combinations of device and samplingpe ' ide Bpecified confidence that the
S^.W^^iSrir,^P«V/L?'	result. h«v. b..n
.Mf4M, in accordance with the U.S. Environmental
Thi. P»per	*raSr»STinU«.tlve review pollcle. and .pprov.a for
Protection Agency's peer ana
presentation and publication.
INTRODUCTION AND BACKGROUND
Many studies have been conducted nationwide to determine the extent *
elevated indoor radon concentrations in the U.S. The majority of these studi
have employed short-term screening techniques, ranging from 1 to 90 days, usi *
either open-faced or diffusion barrier charcoal canisters or alpha' trau
detectors according to EPA protocols. Several factors prevent the developme°£
of a direct relationship between short-term measurements and long-term indo
radon concentrations. Primarily, radon concentrations have been shown to va°r
considerably with time; diurnal and seasonal variations are prominent in
houses and suggestions of weekly or other periods have been made. Some of the
variations clearly correlate with house construction or occupant behavi**
patterns, such as heating, ventilation, and air-conditioning (HVAC) equipment
usage patterns, and the use of natural or mechanical ventilation during mild
periods. However, no general means of computing the effect of these factors
resulting levels of indoor radon has been demonstrated. Added to thi*1
uncertainty due to fluctuations in actual radon concentrations is a small
measurement uncertainty due to the radon measurement devices themselves. Ea'h
-------
possible sampling periods. This paper reports preliminary results of a study of
short-term and long-term variations in radon concentration in approximately 80
houses in the state of Florida. The study involves comparative sampling using
the most common radon measurement technologies, and extends over a year to date.
It is probably the most extensive study of its kind.
This project was commissioned by the state of Florida, in cooperation with
the U.S. Environmental Protection Agency, as one portion of the Florida Radon
Research Program (FRRP) (1). The purpose of the FRRP is to provide technical
support for a statewide building standard for radon-resistant construction
currently in the rulemaking process. The FRRP includes several projects targeted
for technical support of specific standard elements. In this case the
information provides technical background for a post-construction radon test
specified as a performance element of the standard. Other projects address
prescriptive elements of the code such as specifications on soil and fill
characteristics, barrier or sealing techniques, HVAC systems, and active subslab
depressurization systems.
The philosophy of the proposed performance standard can be briefly stated
as a compromise between conflicting needs in the light of measurement
uncertainty. First, as described below, estimates of long-term radon exposure
from single short-term radon measurements are subject to measurement uncertainty.
Second, the State needs to have confidence that a building actually will conform
to the long-term radon concentration standard set [currently 4 pCi/L, considered
as equivalent to 0.02 Working Level (WL)] by the State's Department of Health and
Rehabilitative Services (DHRS); therefore, the needs of the State are best
served either by a longer testing period or multiple measurements (either of
which decreases measurement uncertainty) or by a conservative performance
threshold (i.e., lower than the DHRS standard). Third, builders and developers
need to minimize delays between construction and occupancy; therefore, the
construction industry is best served by as short a test period as is feasible.
The proposed standard was written to offer options in measurement device and
sampling period to address both needs.
Thus the objectives of this study were conceived to provide the specific
information required for the threshold levels incorporated in the codes. A goal
of the project is to provide short-term (less than 2 weeks) measurement options
which would provide adequate confidence that the long-term average indoor radon
concentration does not exceed a specified level (in this case, 4 pCi/L). To
achieve this goal, supporting objectives include documentation of the variability
of indoor radon in typical houses in the state, characterization of this
variability as measured by the most probable candidate radon measurement devices,
separation of seasonal trends in radon concentrations in the state, and
evaluation of regional climatic or construction factors which affect radon
variability.
While most studies of this type have been performed outside the state of
Florida, and many reflect sampling situations inappropriate for Florida housing
(e.g. basement screening measurements), the major features of other research
studies are corroborated by several studies which have been conducted within the
state to identify factors which contribute to the variability of radon
concentrations in Florida homes (2-4). These studies suggest that both short-
term and seasonal variability can cause uncertainties of a factor of 2 or more
in predicting long-term averages from single short-term measurements. These
studies were limited, however, in devices used, region of the state, and number
of houses studied. The current project was designed to supplement these earlier
findings with a more definitive database.
EXPERIMENTAL METHOD
In order to provide an adequate statistical basis for the development of
code recommendations for Florida, the current project includes the monitoring of
approximately 80 houses for over a year using parallel measurements with
different sampling devices. The selected study homes represent a sampling from
-------
tfeiBbbewi		
housing stock such as:
Single family, single level, slab on grade housing with forced >u
heating and cooling	,-
-	Low to moderate radon level - 2 to 20 pCi/L
-	Unmitigated
-	Air handler characteristics! split between houses with air handl
inside building shell (closet) and outside shell (garage, attic* *'
-	Natural ventilation; attempt to select about half of the'houaes wKi
never use natural ventilation for cooling.
Five radon measurement devices were employed in the study for the CUrn«.
of identifying acceptable methodologies for estimating the annual avarnna t
radon concentration as well as developing appropriate predictive relation**.?*®®
between short-term measurements and long-term (annual) average concentrafcI
The devices selected and their deployment periods were:
-	Alpha Track Detectors (ATD; quarterly and annual deployment)
-	Short-Term (EPS) and Long-Term (EPL) Electret Passive Environment-.*
Radon Monitors (deployed continuously; EPS read on a 1-week i,
2-week cycle; EPL read biweekly or monthly)
-	seven day passive diffusion barrier (CC7) and two day open face tr<
charcoal canisters (deployed once per month in each house)
-	Pylon AB-5 Continuous Radon Monitor with a Passive Radon Detector
(deployed for month-long periods in subset of houses)
Each county researcher devised a sampling schedule based on the
guidelines and homeowner schedules. The homeowners were asked to keeD **
homes closed during the charcoal canister deployment period, but were ,n ,
ventilate their houses according to their normal habits otherwise. Tr! ***
county the data were gathered, checked for consistency, and entered I
regional database. The regional databases were combined at least ouart«pi° *
a quality control (QC) survey was performed on the entire database. Oua£? *****
data analyses were performed on the combined data set.	'
In order to assess seasonal trends, the quarterly boundaries were
to isolate the peak heating and cooling seasons as defined by historic*?
outdoor temperatures in the state. The study began the first week of n*
1989, with 40 houses per region. After completion of the first winter
at the end of February 1990, the study was increased to 80 houses
Although some houses were lost during the study, at least three quarter•%
was available for 71 houses by the end of November 1990, the last fall m
incorporated in this paper. The study was scheduled to continue another
quarter until the beginning of March 1991.	er
RESULTS AND DISCUSSION
in order to assess the radon variability displayed in the studv homo-
quarterly and annual arithmetic average radon concentration, standard
(STD), and coefficient of variation (COV — defined as the ratio of th®
the mean, expressed as a percentage) were calculated for each house ana *1 , to>
from all observations made during the period. For the electret raeaau*
time-weighted averages were used due to the variable sampling interval^n«f«
k8 f ? , ° radon concentrations among the houses is illustrated
short-term electret quarterly average results presented in Figure 1.
diat-v< k!?«- .®enera*/ the sample population can be approximated by a loo	'
ThI ^ 5 " typical of studies in larger, randomly selected pODula*? **
than 20 ^ /
-------
of 2-20 pCi/L. The median radon concentration was 3.41 pCi/L, and the geometric
mean and standard deviation were 3.61 pCi/L and 2.1, respectively. The
arithmetic mean and standard deviation among the study houses were 4.71 and 3.95
pCi/L, respectively.
The error structure of the quarterly average radon measurements is depicted
in Figure 2. In Figure 2, the standard deviations of the 7 or 14 day electret
measurements in a quarter are plotted against the quarterly time weighted mean
for the house. Within a significant degree of scatter, the standard deviation
tends to vary linearly with the mean. This suggests that a variance stabilizing
transformation (either performing a log transformation on the data or normalizing
all concentrations to the long-term mean radon) is justified prior to any
regression analysis of the time variability of the data.
Figure 3 shows the pairwise comparison of the quarterly average short-term
electret radon concentrations to those measured by each of the other devices.
In general, the devices agree quite well with each other. Regressions for the
alpha track and long-term electret, which were continuously deployed with the
short-term electret, show slopes near unity and R2 of about 0.95. The 7 and 2
day charcoal canisters, which were deployed 1 week or 2 days each month, showed
somewhat greater scatter (R2 of 0.93 and 0.91, respectively). A more detailed
description of the results of this investigation is beyond the scope of this
paper.
SEASONAL VARIATION
One key issue in the variability of radon measurements is the seasonal
component of this variability. In order to compare pooled seasonal trends across
the study houses, the quarterly average radon concentration data were normalized
by dividing each quarterly average by a longer-term average radon concentration
measured by the same device in the same house. In order to include the houses
which were added in the Spring of 1990, all data were normalized to the average
of the last three quarters of the study (March - November 1990). For the 40
houses which were in the study an entire year, this three-quarter average was
typically less than the annual mean (by an average ratio of 95%). To simplify
data presentation, this investigation will focus on the outcome of the short-term
electret data, although similar plots for the other devices have been developed.
Figure 4 shows the frequency distribution of these normalized quarterly
average concentrations in the study house pool. An examination of the seasonal
plots reveals several clear qualitative differences. Winter, as a rule, is found
to be the season with highest relative radon, as in other parts of the country.
Spring, as a rule, had the lowest radon, then summer and fall.
More striking is the range of normalized quarterly averages. The fall
quarter data correlates best with the long-term average, with 50% of the
normalized concentrations falling within ±0.07 of the mean value (1.094). By
contrast, the winter quarter distribution has a "tail" of houses with higher
relative concentrations, and the inner 50% of the normalized concentrations fall
over the range from 0.98 to 1.61. Thus, given nothing but quarterly average
radon, the moBt precise estimate of the annual average in a given house appears
to be 96% of the fall quarter mean. The winter quarter mean had the largest
range of variation relative to the long-term average radon. While most of the
houses fell within ±0.30 of the long-term average, 25% of the houses had winter
concentrations over 1.5 times the mean for the rest of the year, resulting in a
broad distribution ranging from 0.65 to 2.1. Since the study was continued in
the full set of over 70 houses last winter, it will be of special interest to see
if this behavior is repeated.
VARIABILITY OF RADON MEASUREMENTS
As noted in Figure 2, the standard deviation of short-term E-Perm
measurements during a calendar year was, on the average, proportional to the
quarterly mean, with a constant of proportionality of 0.26. Thus the
-------
distribution of the coefficient of variation should cluster around 26% The rt
for long- term E-Perms are similar. Figure 5 illustrates the distr"ibutl«
quarterly COV value in the study houses by quarter. One might expect that
measurements taken during the summer cooling season would vary less than th
for the other three seasons, in which occupants are more prone to ventilate tH°f®
houses. There is indeed a slight tendency toward higher mean COVs for the ¦ ?
and fall as compared to the summer, but the variability among houses in
season is greater than this seasonal effect. Therefore, the short
variability in relative radon concentrations can be assumed to be of	^
magnitude in all seasons.
If the distribution of normalized radon concentration is assumed unl*
for the houses in the pool, the upper or lower confidence limits ca w1
calculated for certain distributions. Using a lognormal model simila *
Roessler, et al. (3), one-sided upper confidence levels were calculated *to
different combinations of device and sampling period. These thresholds K r
in Table 1, were incorporated into the proposed building standard current? °Vn
the rulemaking process. The values in Table 1 represent threshold levels
device/time combinations listed at the left of each row and the confidence 1
shown in the column headings. In order to predict within the specified j Ve*-
that the long-term average radon concentration in a house will be less th
pCi/L, the results of a single measurement must be lower than the corresno *
threshold level in Table 1. The model from which Table 1 was generated doe ®
include seasonal effects, but was based on the three quarters of data avail k°^
at the time of the calculation. Nonetheless, the table gives a good indi	*
of the way our observed level of uncertainty can be incorporated	n
conservative building standard.	nco a
CONCLUSIONS
This study has provided the most detailed database of which we are awa
of the time variation of a significant number of occupied houses with moderate?*
elevated radon concentrations. We Bee clear evidence of seasonal trends in rad
concentrations from four regions of the state of Florida. Winter concentration"*
are typically higher than for the rest of the year, although the degree o*
elevation varies strongly over the pool of study houses. Fall quarterly averaa®
concentrations correlate best with the annual mean concentration. The patte
of variability suggests that models with logarithmic scaling can be used t*1
estimate expected uncertainties in long-term average radon from short-term
measurements.	™
REFERENCES
1.	Sanchez, D.C., Dixon, R. and Williamson, A.D. The Florida Radon Reeearov,
Program: Systematic Development of a Basis for Statewide Standard
Presented at The 1990 International Symposium on Radon and Radon Reduce"*
Technology; Atlanta, GA, February 19-23, 1990.	l°n
2.	Nagda N.L., Koontz M.D.,	Fortman R.C., Schoenborn, W. A., and Mehen»«
M.A., 1987. Florida	Statewide Radiation Study.	Publicat?
No.05-029-057, Florida	Institute of Phosphate Research
PB88-213970) .	* 1 us
3.	Roessler, C.E., Revell/ J.W., and Wen, M.J. Temporal Patterns of Indor»
Radon in North Central Florida and Comparison of Short-Term Monitorina *
Long Term Averages. Presented at The 1990 International Symposium «
Radon and Radon Reduction Technology; Atlanta, GA, February 19-23, 1990
4.	Roessler, C.E., Roessler, G.S., and Bolch, E. Indoor Radon Proem*
Exposure In The Florida Phosphate Mining Region: A Review.
Physics, 45:389, 1983.
-------
TABLE 1. THRESHOLD RADON CONCENTRATIONS FOR SINGLE RADON MEASUREMENT
CORRESPONDING TO SEVERAL CONFIDENCE LEVELS OF FINDING LONG-TERM AVERAGE
CONCENTRATIONS UNDER 4 pCi/L.
CONFIDENCE LEVEL
Device/Days*
0.5
0.6
0.7
0.75
0.8
0.85
0.9
0.95
CRM-1
4.18
3.87
3.56
3.39
3.22
3.03
2.81
2.51
CRM-7
4.02
3.81
3.61
3.50
3.39
3.25
3.10
2.88
CRM-14
4.00
3.83
3.65
3.56
3.46
3.35
3.21
3.02
EPS-7
4.22
3.85
3.49
3.31
3.11
2.90
2.66
2.33
EPS-14
4.23
3.88
3.54
3.37
3.18
2.98
2.74
2.43
EPL-14
4.39
3.88
3.39
3.15
2.90
2.63
2.33
1.95
EPL-28
4.32
3.91
3.51
3.31
3.10
2.87
2.60
2.26
CC2"
4.78
4.30
3.84
3.61
3.37
3.11
2.81
2.42
CC7
4.20
3.81
3.43
3.23
3.03
2.81
2.55
2.22
~Where CRM = Continuous Radon Monitor
EPS = Short-Term (High Senaitivity) Electret-Ion Chamber
EPL = Long-Term (Low Sensitivity) Electret-Ion Chamber
CC2 = Open Face ("2 day") Charcoal Canister
CC7 = Diffusion Barrier ("7 day") Charcoal Canister
CC2 values may be overestimated due to observed bias in study sample.
-------
60
w
E->
w
§
CO
><
i-3
W
a
o
c*
w
«
501
40-
4 6 8 10 12 14 16
RADON CONCENTRATION (pCi/L)
-F^r
18 20
Figure 1. Distribution of quarterly average radon concentration in study
houses.
-------
10
9
6
7
6
5
4
3
2
1
0
i
! i
i
Y=0.029+0.261X
fT2=0.620
5	10	15	20	25
QUARTERLY AVERAGE RADON (pCi/L)
30
Figure 2. Standard deviation of individual short-term electret
concentrations compared to quarterly mean concentration
in study houses.
-------
EPSJVTD ALLQ
30
0 25i
a.
£20
0
J 15
1	i° h
5
j#

ATO=1.04764+0.0374t(EPS)
(t~2=0.945
10 IS 20 23
EPS RADON-CM pCi/L
30
35
EPSfPL ALLQ
EFLBl.02235'f0.31l53(EPS)
IT2=0.i35
10	15	20
EPS RADCN-C M pCi/L
• AID
' EPL
EPS.-CC7 ALLQ
_ 25-
\
§.20-
Z
o 15
^ 10
5-

• 1'
* f '	CC7=O.I44#6+0.34422(EP3)
IT2=0.»2«
10	15	20
EPS RADOdjC M pCi/L
25
30
EPS:CC2 ALLQ
CC2=«.893B5-H>.0275»(EPS)
R~2=0.9I1
10	15	20
EPS RADON-CM pCi/L
25
30
« CC7
¦ CC2
Figure 3. Linear regression between the quarterly average radon concentrations measured by
short-term electret and each other device.
-------
EPS WNTER QUARTER
0.3 0.5 0.7 O.t 1.1 1.3 1.5 1.7 1.1 2.1
RAOGN-C RAID
EPS SPRMG QUARTER
0.1 1.1 1.3 1.3
RADONJC RATO
EPS FALL QUARTER
EPS SLAAER QUARTER
Figure 4. Relative frequency distribution of the quarterly average radon
concentrations normalized to a 9-month average.
-------
WNTER QUARTER

mm
20 40 to 10
RMZNTCaV.
100
SPRWG QUARTER
20 40 CO BO
RxcBtr c.av.
100
SlAiO QUARTER
FAIL QUARTER
Figure 5. Distribution of quarterly coefficient of variance values froa short-term electret data.
-------
Ill-19
RELATIONSHIP BETWEEN 2-DAY SCREENING MEASUREMENTS OF 222Rn
AND ANNUAL LIVING AREA AVERAGES IN BASEMENT AND NONBASEMENT HOUSES
by: S. B. White, N. F. Rodman and B. V. Alexander
Research Triangle Institute
Research Triangle Park, N.C. 27709
J. Phillips and F. Marcinowski
U.S. Environmental Protection Agency
Office of Radiation Programs
Washington, D.C. 20460
ABSTRACT
As part of an EPA/State cooperative program, a random sample of 41.648
houses from 30 of the 48 conterminous states have been screened for 222Rn
over the past four years. Charcoal canisters were placed in the lowest
livable level and exposed for two days. In addition, 1-year alpha track
detectors were used in a random subsample of houses with at least one
detector placed on each livable level.
This paper describes the relationship between annual living area
averages (ALAA) and wintertime, closed-house, 2-day screening measurements.
Both 2-day and 1-year measurements of 222Rn were made on 995 houses located
in 13 states. A broad range of climates, geologic conditions, and housing
types are represented in the sample. Equations for predicting ALAA are
derived for screening measurements taken in the basement and on the first
floor of nonbasement houses. These relationships are used to obtain
predicted values of ALAA for the 41,648 houses for which screenings
measurements are available. The distribution of predicted values of ALAA
by house type are then characterized. To the extent that the 30 states
represent the 48 conterminous states, these distributions apply to the
nation as a whole.
This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved
for presentation and publication.
-------
INTRODUCTION
Short-term screening tests for 222Rn are used to determine if
additional testing (usually one year duration) is needed to more accurately
characterize health risks to Rn exposure. The extent to which screening
tests can properly identify houses needing further testing is governed by
the degree to which short- and long-term measurements are related. This
relationship has not been studied extensively. In a review of the
published literature, Ronca-Battista (1) found only nine studies in which
this issue was addressed and most of these had sample sizes less than 100
houses. This study attempts to provide a better understanding of this
relationship.
A component of the EPA/State Indoor Radon Surveys involves two types of
222Rn measurement devices in a subsample of houses. Each participating
house is tested with a 2-day charcoal canister placed in the lowest livable
level and 1-year alpha track detectors (ATDs) placed on each livable level
The 2-day test is carried out in the winter season under closed-house
conditions. A total of 995 houses provided data for establishing the
relationship between 2-day measurements and 1-year measurements.
OBJECTIVES
The purposes of this study were (1) to examine the overall relationshi
between 2-day screening measurements and annual living area averages
(ALAA), (2) to determine if a screening measurement can be effectively u<**rt
to predict the ALAA for an individual house, and (3) to examine the
distribution of predicted values of ALAA for some 40,000 randomly selects
houses for which screening measurements are available.
METHODOLOGY
Two indoor radon measurements (X, ALAA) were obtained from houses
covering a 13-state area.1 X is the 2-day charcoal canister measurement
observed in a given house and ALAA 1s the annual living area average
obtained by averaging all ATD readings taken on that house. In multiple
level houses, a single ATD was placed on each livable level with a maximum
of four ATDs per house. Two ATDs were used 1n one-story nonbasement
houses. Averaging measurements from each level 1s one of several ways of
characterizing the annual concentration in a house. Other ways include
lStates providing both short-term and long-term measurements include:
Alaska, Arizona, Indiana, Iowa, Maine, Massachusetts, Minnesota, Missouri,
North Dakota, Ohio, Tennessee, Vermont, and West Virginia.
-------
using only the first floor ATD measurement or using a weighted average of
the ATD measurements from each level, where the weights reflect the
proportion of time spent on each level.
This report examines the relationship between X and ALAA in basement
and nonbasement houses. Values of X and ALAA for a given house are
considered usable in the analysis if 1) the canister floor code matched the
lowest floor code of the ATDS, 2) the ATDS used in calculating the ALAA had
been exposed between 305 days and 425 days, 3) the canister was exposed
within 30 days of the beginning of the ATD exposure period, and 4) a valid
ATD reading was reported for each ATD originally placed in the house. A
total of 997 houses provided data that met these requirements. After
examining the data, two houses were excluded as outliers (one in
Massachusetts and one in Tennessee). The relationships reported herein are
consequently based on 995 houses--609 basement houses and 386 nonbasement
houses.
A scatter plot of the data shows that ALAA is linearly related to X and
that the variation in ALAA tends to increase as X Increases. A
relationship between X and ALAA is derived using a model which reflects
these visual observations in the data. A specification of the model is
given below.
The results in this paper employ a mathematical model that assumes that
long-term measurements of 222r0 are nnearly related to short-term
measurements and have variances that are proportional to their expected
values. That is,
ALAAi = (a + £X.) + aZ. (a + ^X^172	(1)
where
ALAAi = annual living area average calculated for the 1*^ house,
Xi = canister measurement on the house,
a,p,a = parameters to be estimated, and
Zi = random error for 1*h house, assumed to be normally distributed
with mean 0 and variance 1.
In order to convert (1) to a model having a homogeneous error structure, we
divide by (a +	and substitute \ALAAi f°r (a + /JXi)1/2 on the left
hand side of (lj:
JALAA1 = (a + flXj)1'2 + ffZr
(2)
-------
The parameters in (2) were estimated using nonlinear least squares. The
/N	A	A
prediction equation J ALAA = (a + /JX)1'2 was squared to obtain predictions
of long-term concentrations for given short-ternnrieasurements. Similarly
endpolnts of the 95% confidence Interval for J ALAA were squared to obtain a
corresponding Interval estimate for the long-term concentration.
RESULTS
SHORT- VS LONG-TERM RELATIONSHIP
Results of fitting equation (2) to data from basement houses and from
nonbasement houses are given in Table 1. For each type of house, Table 1
TABLE 1. EQUATIONS FOR PREDICTING ANNUAL LIVING AREA AVERAGES
FOR BASEMENT AND NONBASEMENT HOUSES
Sample	Correlation Residual
Type of House Size Prediction Equation (X, ALAA) Error (5)
Basement 609 ALAA = 0.69 +	0.54X	0.82 n m
(0.08)*	(0.02)	D1
Nonbasement 386 ALAA = 0.53 +	0.61X	0.90 n ia
(0.04)	(0.02)	4
* (Standard error of parameter estimate.)
gives the sample size, the prediction equation, the correlation between X
and ALAA, and the standard deviation, a, from the fitted model. The
prediction equations are
Basement House: ALAA = 0.69 + 0.54X	(3)
/\
Nonbasement House: ALAA = 0.53 + 0.61X	(4)
where ALAA 1s the expected (or predicted) value of the annual living area
average 1n a house that has a screening measurement of X on the lowest
livable level. The prediction equation for nonbasement houses reflect the
exclusion of two (X, ALAA) data points considered to be suspect - (24.0
2.2) and (39.6, 3.6). If these data points are Included the prediction'
equation becomes A^AA = 0.61 + 0.52X.
-------
Scatter plots of the data for basement and nonbasement houses are shown
in Figures 1 and 2, respectively. Note that as the canister measurements
get larger the ALAA measurements show greater dispersion. As noted
previously, this increase in variability 1n long-term measurements is taken
into account by the model used in the data analysis. Superimposed on each
scatter plot are three lines. The center line is the prediction equation.
The other two lines (designated as UCL and LCL) represent the estimated
upper and lower 95 percent confidence limits on the predicted value for an
individual house. The interpretation of the confidence limits in Figures 1
and 2 is as follows: if a 2-day canister reading is X for a given house,
there is a 95 percent chance that the true ALAA for that house would be
covered by the interval falling between the upper and lower lines
corresponding to X. For instance, for a basement canister measurement of
X = 10 pCi/L, we can be 95 percent confident that the interval (2.2-12.0)
will cover the true ALAA for that house. The vertical spread in the data
for a given value of X (as reflected by the distance between the upper and
lower confidence limits) indicates that the ALAA varies widely among houses
having the same canister measurement.
FALSE POSITIVE/NEGATIVE ERRORS
EPA currently recommends additional testing if the screening
measurement exceeds 4 pCi/L. Furthermore, EPA recommends mitigation if a
1-year test exceeds 4 pC1/L. In this case, a perfect screening test would
correctly classify a house as to whether its annual concentration would
exceed 4 pCi/L. Although there 1s no perfect test, one can, however,
assess the performance of a screening test by characterizing the
probability of an Incorrect decision. One of two Incorrect decisions can
be made on the basis of a screening measurement—If a screening measurement
1s £4 pC1/L, one may Incorrectly conclude that the house annual
concentration is *4 (false negative); 1f a screening measurement exceeds
4 pCi/L, one may Incorrectly conclude that the house annual concentration
is also greater than 4 pC1/L (false positive).
The probability that the ALAA will exceed 4 pC1/L, given a specified
screening measurement, X, is given by
'(Z<
1§_
bX)1/2- 2
(5)
where Z is a standard normal deviate, a and b are the estimated model
parameters, and a is the standard deviation from the fitted model. This
probability was calculated for screening measurements, X, ranging from 1 to
16 pC1/L for basement and nonbasement houses by substituting the
appropriate parameter estimates from Table 1 Into equation (5); the results
are shown 1n Figure 3. The regions of false positive and false negative
errors are noted and the probability of an error associated with a given
screening measurement can be determined directly from the plotted curves.
-------
80 "
UCL
70-
ALAA = 0.69 + 0.54X
60
A
L
A
A
,LCL
40
8
i
/
L
30 "

20 -
* »»
10 -
110
100
90
70
80
50
60
30
40
20
0
10
2-Day Canister Measurement. pCi/L
Figure 1, Relationship Between 2-Day Charcoal Canister Measurement
And Annual Living Area Average - Basement Houses
-------
30
27
24
21
18
15
12
9
6
3
0
ALAA
UCL
0.53 + 0.61X
LCL
—I
24
l
26
I
28
-r-
30
-r
36
—T~
2
"T"
4
10
-r~
12
16
-T-
18
14 16 18 20 22
2-Day Canister Measurement, pCi/L
32
34
Figure 2. Relationship Between 2-Day Charcoal Canister Measurement
And Annual Living Area Average - Nonbasement Houses
-------
.0
9
8
.7
.6
.5
.4
.3
.2
.1
0
False Positives
(above the curve)
Nonbasement
Houses—#
Basement Houses
False Negatives
(below the curve)
Screening Measurement, X
Figure 3. Probability that annual living area average exceeds 4 pCI/L
as a function of screening measurement
-------
For instance, the probability of a false negative error is approximately
0.17 for a screening measurement of 3 pCi/L in a basement house and
approximately 0.09 for a screening measurement of 3 pCi/L in a nonbasement
house. On the other hand, the probability of a false positive error is
approximately 0.41 (= 1.00 - 0.59) for a screening measurements of 7 pCi/L
in a basement house and approximately 0.29 (= 1.00 - 0.71) for a screening
measurement of 7 pCi/L in a nonbasement house.
DISTRIBUTION OF ALAA
The relationships between short-term and long-term measurements of
indoor 222r0 as given by equations (3) and (4) provide an opportunity to
use an existing data base of screening measurements to characterize the
distribution of ALAA.
Under the EPA/State Indoor Radon surveys initiated in 1986, 30 of the
48 conterminous states have conducted statistically designed surveys. A
probability-based sample of owner-occupied main residences having a listed
telephone number, a permanent foundation, and at least one floor at or
below grade level was selected in each state. Sample houses were tested
with a 2-day charcoal canister placed 1n the lowest livable level. Tests
were conducted during the heating season under close-house conditions.
Although other surveys have used probability sampling (2,3), and other data
sets include more test houses (4,5), the state surveys collectively provide
the largest existing data base formed from studies that 1) use
probabilities in making house selections, 2) have common objectives,
3) utilize the same measurement method, 4) employ the same protocol, and
5) sample the same target population. The 30 state surveys have produced
20,768 basement measurements, and
20,880 first floor measurements 1n nonbasement houses.
The basement screening measurement, X, for a given house was
/ % /\
substituted Into equation (3) to obtain a value of ALAA for that house. In
making the translation from X to ALAA, the sampling weight for the house
was retained for use in future analyses. This process was repeated for all
basement screening measurements and produced 20,768 ALAA values and
associated sampling weights from a random sample of basement houses
covering a 30-state area. Similarly, each first floor screening
measurement was substituted into equation (4). This generated 20,880 ALAA
values and associated sampling weights from a random sample of nonbasement
houses covering a 30-state area.
Table 2 gives, in tabular form, the weighted cumulative distribution of
AU\A for basement houses, for nonbasement houses, and for all houses in the
30-state area. In addition, the distribution for basement and for
nonbasement houses are presented graphically 1n Figure 4. Summary
statistics (weighted) relating to these distributions of ALAA are given in
-------

TABLE 2. CUMULATIVE
DISTRIBUTIONS OF ALAA FOR 30-STATE
AREA
ALAA
Basement Houses
Nonbasement Houses
All Houses
.4
.6
.8
1.0
1.2
0.0
0.1
1.2
5.4
15.7
3.4
12.9
28.6
45.2
58.3
1.6
5.9
13.7
23.6
35.1
1.4
1.6
1.8
2.0
2.2
26.2
33.0
41.3
48.1
52.4
70.5
76.2
80.4
84.8
87.4
46.4
52.7
59.1
64.8
68.4
2.4
2.6
2.8
3.0
3.5
57.2
61.5
65.1
67.6
74.6
89.3
90.6
92.1
93.0
94.7
71.9
74.8
77.4
79.2
83.8
4.0
4.5
5.0
6.0
8.0
79.1
82.4
85.1
89.4
93.6
96.1
97.0
97.6
98.4
99.2
86.8
89.1
90.8
93.5
96.1
10.0
15.0
20.0
25.0
95.8
98.0
98.8
99.2
99.5
99.9
99.9
99.9+
97.5
98.8
99.3
99.5
-------
40
30
20 -I
10
_ 8
SJ 6 -J
4 -
C	2 ~
§
o
O	i.o
5	B
^5	.6
.4 -
.2 -
A A
AA
AA
• Basement Houses
a Nonbasement Houses
1—i	1—i—i—i—i—i—i	1	1	1—r
5 10 20 30 40 50 60 70 80 90 95 98 99 99.5
Cumulative Percent
Figure 4. Normal probability plot of ALAAfor houses in 30-state area
-------
Table 3. It is clearly evident that ALAA 1s higher 1n basement houses
For example: the median (50th percentile) 1s 2.05 pC1/L for basement *
houses as compared to 1.02 pC1/L for nonbasement houses and the arithmetic
mean for basement houses is more then double that for nonbasement hous*c
(3.4 pCi/L as contrasted to 1.4 pCi/L).
Figure 4 indicates that the distribution of ALAA cannot be
satisfactorily approximated by a lognormal distribution since the data
points for each house type depart substantially from a straight line plot
This was anticipated and can be explained by examining the prediction
equations (3) and (4). Prediction equation (3), for example, converts the
variable X, basement screening measurement, Into the variable ALAA through
the relationship ALAA = 0.69 + 0.54X. X 1s assumed to be lognormally
distributed and there is strong evidence to support this assumption. Under
this assumption, ALAA is lognormally distributed only if the intercept
term in the relationship Is zero. If the Intercept is zero, then In (ALAA^
is normally distributed since it 1s the sum of a normally distributed
variable (In X) and a constant (In 0.54). The data show, however the
intercept (estimatedat 0.69) to be statistically greater than zero.
The empirical distribution of AGU shows more houses 1n the tall of th
distribution than the number obtained by using a lognormal distribution
For instance, the empirical distribution (Table 2 or Figure 4) shows 1 2%
of basement houses have ALAA exceeding 20.0 pC1/L as contrasted to an
estimate of 0.2% based on a lognormal distribution with a geometric mean nf
2.4 pC1/L and a geometric standard deviation of 2.1. Applying these
percentages to a base of several millions of houses produces an enormous
difference in the two estimates of the number of houses with ALAA exceedi
20.0 pCi/L. The empirical distribution is based on tests from more than
40,000 houses and should be used 1n estimating proportions rather than
using a geometric mean and geometric1standard deviation.
The percentage of houses with ALAA exceeding 4, 10 and 20 pC1/L are
respectively, 13.2%, 2.5% and 0.7% (Table 3, 3rd column). In contrast,' [he
distribution of annual average radon concentration 1n U.S. houses reported
by Nero (6) shows: 7.4% of the houses above 4 pCi/L; 1.0% above 10 pCi/L-
and, 0.13% above 20 pC1/L (these percentages were calculated using a
geometric mean of 0.9 pCi/L and a geometric standard deviation of 2.8).
The differences may be attributable, 1n part, to differences in the
dependent variables, to the way basement houses are defined, and to the
sampled populations. For this study, a basement house 1s defined as any
house where the lowest livable level has at least one wall built against
earth.
A scientific study 1s now under way by EPA to characterize the
nationwide distribution of annual concentration of Indoor 222Rn in
residential houses (7). This study should resolve many Issues/questions
-------
TABLE 3. ALAA SUMMARY STATISTICS FOR 30-STATE AREA
Parameter
Basement
Houses
Nonbasement
Houses
All
Houses
Arithmetic Mean*
3.4
1.4
2.5
Geometric Mean*
2.4
1.1
1.7
Geometric Standard Deviation
2.1
2.0
2.2
Median*
2.05
1.02
1.50
% > 4 pCi/L
20.9
3.9
13.2
% > 10 pCi/L
4.2
0.5
2.5
% > 20 pCi/L
1.2
0.1
0.7
* Units of measurement - pCi/L.
-------
relating to levels of 222Rn to which occupants are exposed. Until such
times that results from this national assessment study become available,
the information provided by the distribution of ALAA serves to add to the
existing body of data on nationwide annual concentration of 222r0 jn ^he
living area.
CONCLUSIONS
Two-day charcoal canisters and 1-year alpha track detectors were used
to measure 222Rn in 609 basement houses and 386 nonbasement houses.
Results from this two-year study show there is a strong positive
relationship between 2-day screening measurements and annual living area
averages (ALAA). The equations for predicting ALAA from a screening
measurement for basement and nonbasement houses are, respectively,
ALAA = 0.69 + 0.54X and ALAA = 0.53+ 0.61X. The results also show that
ALAA varies widely among houses having the same screening measurement. The
derived relationships were used to obtain predicted values of ALAA for a
probability-based sample of 41,648 houses covering a 30-state area. A
characterization of the distribution of predicted values for basement and
for nonbasement houses is given. For example, an estimated 7.0% of the
nonbasement houses have predicted values exceeding 3.0 pCi/L as compared to
32.4% for basement houses. To the extent that the 30 states represent the
48 conterminous states, the distributions of ALAA shown herein apply to the
nation as a whole.
REFERENCES
1.	Ronca-Battista, M. Radon 1n U.S. homes: A summary of the available
literature on annual and screening radon concentrations, season
variations, and differences between floors. Unpublished report
prepared for U.S. Environmental Protection Agency, Washington, D.C.
September 1989.	'
2.	Perritt, R.L., Hartwell, T.D., Sheldon, L.S., Cox, B.G., Clayton, C A
Jones, S.M., Smith, M.L., and Rizzuto, J.E. Radon-222 levels in New *'
York state homes. Health Phys. 58: 147-155, 1990.
3.	Rahman, M., Hoyle, W., and Tuccillo, K. Radon measurements in New
Jersey. Paper presented at the 1988 Symposium on Radon and Radon
Reduction Technology, Denver, Colorado, October, 1988.
4.	Cohen, B.L. and Gromicko, N. Variation of radon levels 1n U.S. homes
with various factors. JAPCA 38: 129-134, 1988.
5.	Alter, H.W. and Oswald, R.A. Nationwide distribution of Indoor radon
measurements: A preliminary data base. JAPCA 37: 227-231, 1987.
6.	Nero, A.V., Schwehr, M.B., Nazaroff, W.W., and Revzan, K.L.
Distribution of airborne radon-222 concentrations in U.S. homes.
Science 234: 992-997, 1986.
-------
7. Bergsten, J.W., Holt, N.A., Lucas, R.M., and Smith, M.L. The national
residential radon survey design report. Unpublished report prepared
for the Environmental Protection Agency, RTI/4240/02-04F, 1989.
-------
in-:
TITLE: The Use of Multiple Short-Term Measurements to Predict Annual
Average Radon Concentrations
AUTHOR: Frank Marcinowski, EPA
This paper was not received in time to be included in the
preprints so only the abstract has been included. Please check
your registration packet for a complete copy of the paper.
EPA has been working to establish a policy for addressing
radon in the context of a real estate transaction. Much of the
work to date has focused on policy alternatives such as escrow
accounts and long-term contract clauses. The principle
difficulty with addressing radon in a housing transaction is the
short testing period necessary to get information in time for
house closing. The use of multiple short term measurements has
been suggested as one method for reducing the uncertainty
inherent in the use of short-term tests to characterize long-term
radon levels. A statistical model was established to compare the
expected performance of short-term procedures which apply
simultaneous tests to the expected performance of a single short
term test. The analysis examined multiple measurements that were
separated temporally (taken at different times) and spatially
(taken simultaneously on one or more house levels).
-------
Session III:
Measurement Methods -- POSTERS
-------
III?
ntianarTRRT ZftTTON OF STRUCTURES USING SIMULTANEOUS SINGLE SOURCE CONTINUOUS
WORKING LEVEL AND CONTINUOUS RADON GAS
MEASUREMENTS
By: Brian Fimian
Radonics, Inc.
McLean, VA 22102
John E. McGreevy
Radonics, Inc.
McLean, VA 22102
ABSTRACT
Hundreds of simultaneous single-point working level/radon gas
measurements have been used for the following study in order to demon
the behavior of the time synchronized equilibrium relationships between r^6
gaa and progeny. Various normalizing equilibriums schemes will be explored0"
for different structures, homestyles, and HVAC systems. The "equal-mix'
theory will be discussed in great detail as it applies to radon gas and^"9
progeny and to real world measurement data.
-------
INTRODUCTION AND SCOPE
A commonly held belief by the scientific and radon measurement community
is that radon variances from season to season are far greater than any
measurement device's accuracy in a given measurement situation. When
considering the large number of variables in the radon assessment equation,
one can easily become uncomfortable with a single measurement approach for the
purposes of a mitigation decision. Practical reality, however, has dictated
that the EPA move towards a one measurement scenario for the purposes of a
mitigation decision. Therefore, much attention must now be paid towards
providing the best measurement possible in a single-measurement scenario.
Homeowners are very reluctant to test initially, and on those rare
occasions in which they do, it is almost impossible to motivate them to
perform a long term follow-up test as the EPA recommends. Real estate
transactions prevent follow-up measurements due to the severe time constraints
of the transaction itself. Escrow of funds is commonly objected to by
realtors and sellers because of the 'transaction tail* it creates and because
of possible future legal liability problems.
So then, in light of the practical considerations in both homeowner and
real estate testing, one is left with but one option and that is to make a
single measurement and base the decision of whether or not to mitigate on that
measurement.
Assessing true health risk now becomes a completely impractical
objective. A screening measurement can only assess "radon potential*. Radon
potential is determined by placing a structure under what would be considered
to be the worst-case conditions for the purposes of maximizing the probability
of finding elevated elevated radon. The measurement made under these
conditions would determine if any radon potential actually exists which could
pose a health threat.
IDENTIFYING RADON POTENTIAL- MAJOR ISSUES TO ADDRESS:
o Validation of EPA closed-house conditions.
o Accuracy and precision of the measurement device itself.
o Correctly identifying the highest and most stable radon
concentration in the structure (i.e. lowest liveable area, least
ventilated area, closest to the source).
o Assurance that short-term weather conditions will not unfairly bias
the test in one direction or the other.
o Provide as much assurance as possible with current measurement
technology that internal house variables do not unfairly bias the
test high or low.
1
-------
o	Ensure that standard house operating conditions do not produce
misleading results (i.e. furnace on or off, house occupied or
unoccupied)
o Ensure that post-mitigation measurements provide as much assurance
as possible that the mitigation system will be effective for 'the
long-term.
Simultaneous measurements of both radon and its progeny can eliminate
many of the variables present and improve the chances of obtaining a truly
meaningful measurement for the purposes of assessing radon potential. The
definition of a 'good" radon test is simply a test that will withstand all
future radon measurements in the structure. Radonics is focused on this
definition of a "good" radon test which has led to a preference whenever
possible for making simultaneous radon gas and progeny measurements.
Unfortunately, the machines that provide these kinds of
measurements are 8ti.li
fairly expensive. As test volumes increase and requirements for a single
measurement for the basis of a decision become commonplace, the cost of both
the equipment and services will decrease significantly.
VALIDATION OF CLOSED HOUSE CONDITIONS AND DETECTION OF INTENTIONAL TAMPERING
Equilibrium relationships are a good indication of the maintenance off
closed house conditions due to the fact that under normal situations
equilibrium variances remain within a fairly narrowly defined range,
approximately 15%. Some of this normal variance is due to limitations within
the measurement device itself in that it is very hard to calculate
instantaneous equilibrium relationships (1). a lag exists between the
infusion of radon gas into the structure and the ensuing creation of the
progeny for that gas (See Fig.l). To correctly calculate an instantaneous
equilibrium relationship, such things as time delay for creation of progeny,
calculation from continuous to instantaneous measurement for both gas and
progeny, and instrumentation lag times must be all be assessed before a
meaningful number can be derived.
Violation of closed house conditions produces dramatic changes in the
equilibrium relationships over very short periods of time. This volatility
can sometimes be correlated with large temperature swings that occur in the
structure close to the times of the large equilibrium changes, if temperatu
swings do occur simultaneously with these large equilibrium changes, the	*
reason would most likely be the opening of doors and/or windows in the same
level of the structure where the measurement device is located (See Fig. 2)
The more difficult to detect violations of closed-house conditions come from
more knowledgeable occupant. Opening of upstairs windows coupled with the *
operation of the HVAC system blower will dilute the basement air gradually
over a few hours with the fresh air that exists in the upper levels of the
structure. This of course causes a gradual transition towards lower
radon/radon progeny levels. Equilibrium relationships during this transition
move slightly downward at a very gradual pace. This gradual movement is
sometimes very difficult to detect.
2
-------
Data from a radon gas and progeny measurement that does not follow or
track one another also leads the measurement firm to hold the final results
suspect. Such things as operation of electronic air filters and ion
generators can cause progeny levels to deviate from their normal
correspondence with the radon gas levels (See Fig. 3). This is simply another
indication that the closed-house conditions of the structure are being
violated in order to defeat the radon measurement.
Finally, the use of a HVAC blower running constantly reduces both the
radon gas and progeny levels, irrespective of whether or not the house is open
upstairs. A "mixing bowl" action occurs within the air handler, providing for
the entire volume of the house to be mixed with the basement air. This of
course causes a significant dilution of the basement radon levels, typically
down to about one third (depending on upstairs volume) and can significantly
alter the test results.
EPA Protocols have yet to address the HVAC blower situation. Caution
should be applied to ensure that both finished (open return duct) and
unfinished (no return duct) basements are well characterized. Our measurement
data has clearly indicated that there is a dramatic reduction in radon and
radon progeny levels when an HVAC system is run constantly in either a
finished or unfinished basement. Presumably this is due to the leakage
factor on both the positive and negative side of the HVAC system in unfinished
basements and the open return air duct present in the lowest liveable area in
finished basement.
VARIATIONS IN HOUSE TYPES THAT INFLUENCE RESULTS
Newer homes in the United States typically have somewhat lower air
exchange rates than older homes. Different areas of the country have
different housing make-ups that account for the variables in the equation.
For instance, New England's housing stock is typically older than housing
stock in Atlanta or Washington, DC. When measurements are constantly made in
older homes of any type there is a general trend towards slightly lower
equilibrium relationships because of the increased air exchanges per hour in
these structures. On the other hand, newer homes produce equilibrium
relationships that are somewhat higher because of lower exchange rates.
Testing in the lowest liveable area, which is typically an unfinished
basement, produces equilibrium relationships (in houses that are less than 25
years old) of 60% to 70%. Finished basements in the same type of structures
with forced hot air systems tend to produce equilibriums closer to 50% because
of the mixing bowl action that is occurring within the air handler from the
on/off action of the HVAC system. One point of concern in this area comes up
when measuring older homes that are likely candidates for future energy
improvements. In such homes marginal radon levels could become elevated radon
levels once the home is insulated and such things as new windows are
installed.
3
-------
HVAC TYPES
The following are four groups of HVAC systems commonly found in the
United States:
1.)	Forced air HVAC with either electric, heat pump, oil, or
natural gas with blower unit or air handler in the lowest
liveable area.
2.)	Hot water baseboard or radiant heat with electric, oil,
or natural gas with boiler in the lowest liveable area.
3.)	Electric baseboard heat.
4.)	Forced air heat with the air handler not in the lowest
liveable area.
5.) A combination of forced-air (typically air conditioning)
and baseboard type heat.
Forced hot air systems with the air handler in the lowest liveable area
tend to have equilibrium relationships in the 50% range. Upstairs radon
levels in the winter season tend to be similar to the basement radon levels
due to the operation of the furnace. This is predominantly because of the
mixing action that occurs in the air handler and the distribution of the
basement radon level throughout the structure. Summer radon levels in this
type of house also tend towards being somewhat evenly-distributed when AC
units are operating.
Forced air systems with the air handlers in upper floors tend to act v©
similarly to hot water baseboard and electric baseboard systems in that there
does not seem to be nearly the mixing action that occurs with all floors. The
exception to this rule occurs when there is not a door at the top of the
stairs that leads down to the basement which will allow air mixing between the
first and second levels. Also, hot water baseboard or radiant systems in
general tend to have much greater concentrations in the lowest liveable area
because of the inability of the radon to effectively mix itself with the
floors above. Electric baseboard heat also acts in a very similar manner
The two largest sources of error occur when either an air-handler is
operating continuously or in situations where houses are tested with the
heating system not in operation and the structure is unoccupied. These homes
that are tested when they are unoccupied tend to act like baseboard heat
houses, raising the measured radon levels.
POST-MITIGATION
Post-mitigation performance can only be truly assessed using simultaneous
radon/ radon progeny measurements. Most pre-mitigation levels in the United
States lie between four and eight picoCuries. It is currently impossible to
4
-------
remove all of the radon gas with a sub-slab mitigation system. Typically, 20%
to 30% of houses that are mitigated have radon gas levels between two and four
picoCuries. These radon levels can create progeny levels that are in excess
of the EPA action guideline. Newer homes that are energy efficient and have
unfinished basements may produce a low radon gas level while still having
quite elevated post-mitigation progeny levels.
Thoron has also been discovered to be a problem in a number of U.S.
homes. Radon gas measurement devices do not typically measure thoron and
hence miss detecting a potential health risk (2).
COST
The cost of simultaneous radon/radon progeny measurement hardware and the
associated quality assurance requirements will continue to decrease. However,
in the overall measurement equation it is certainly not the measurement device
alone that determines the cost. The key elements are manpower, transportation
costs, scheduling, quality assurance, and the consultation with the customer
after the results are reported. The equipment costs typically only run 10 to
14% of the overall transaction cost, whether it be active devices capable of
performing simultaneous working level/gas measurement or passive devices
without this capability.
CONCLUSION
With the entire radon measurement process becoming a test/fix
environment, tremendous focus and energy will be put into the requirement of
making a 'good" measurement, i.e. one that stands the test of time. As
customer demand for 'good' radon measurements increases, firms will encourage
the use of simultaneous radon/radon progeny measurements to assess radon
potential in the worst case home use scenario. Should a radon gas measurement
be made without consideration for the progeny and a portion of the house is
changed ( i.e. older homes being tightened up), elevated progeny could be
present and the homeowner would be completely unsuspecting. Conversely,
should a progeny measurement alone be made with the right set of conditions
present for elevated gas and low progeny, and these conditions changed, the
homeowner could then also be harmed. The logical conclusion is to measure
both radon gas and progeny.
The work described in this paper was not funded by the U.S.
Environmental Protection Agency and therefore the contents do not necessarily
reflect the views of the Agency and no official endorsement should be
inferred.
5
-------
» TAMPER
0.0281
0.0267
0.0253
H 0.0239
J 0.0225
W 0.0211
D0.0197
W 0.0183
J 0.0169
0.0155
0 0.0141
2 0.0126
h 0.0112
Y 0.0098
££ 0.0084
0 0.0070
2 0.0056
0.0042
0.0028
0.0014
T ine
Date
BAROMETRIC PRESSURE
¦TEMPERATURE
•RADON GAS
V RADONICS
The Radon Specialists
PI P
9:00
1/23
13:00
1/23
17:00
1/23
21:00 l:oo 5:00 9:oo i3:oo
1/23 1/24 1/24 1/24 1/24
TIME OF TEST
• TAMPER
0.0281
0.0267
0.0253
to 0.0239
J 0.0225
W0.0211
D 0.0197
LJ 0.0183
J 0.0169
0.0155
(3 0.0141
Z 0.0126
h0.0112
*0.0098
ff0.0084
0 0.0070
2 0.0056
0.0042
0.0028
0.0014
mam

T ine
Date
17:00
1/24
21:00
1/24
Tmdo 5:°° 9:00 i3:oo i?:oo
1/25 1/25 1/25 1/25 1/25
TIME OF TEST
21:00
1/25
Year of 1991
Figure 1. Radon gas leads progeny.
-------
BAROMETRIC PRESSURE
¦TEMPERATURE
¦RADON GAS
TAMPER ¦«
0,0217
Bel Air, «D
VRflDONICS
The Radon Specialists
Tine 7:40 11:40 15:40 19:40 23:40 3:40
Date 3/12 3/12 3/12 3/12 3/12 3/13
TIME OF TEST
7:40 11:40
3/13 3/13
Figure 2. Tampering.
-------
BAROMETRIC PRESSURE
• TEMPERATURE
-RADON GAS
1764 Old Meadow Lane
McLean, VA
1RADONICS
The Radon Specialists
.¦'li'.iMliVW
iilllilkiilili
W3 lit!) tfwv
if.' 'lil'';!'
lli'lt!I'lJl
Tine 16:10 20:10 0:10 4:10 8:10 12:10 16:10 20:10
Date 10/30 10/30 10/31 10/31 10/31 10/31 10/31 10/31
TIME OF TEST
Figure 3. Progeny Deviation.
-------
REFERENCES
1.	Busigin, Van Der Vooren, Phillips. "Interpretation of the Response of
Continuous Radon Monitors to Transient Radon Concentrations'". Health Physics,
Vol. 37, pp. 659-667, November 1979.
2.	Martz, Falco, Langner. 'Time-Averaged Exposures to Radon and Radon Progeny
in Colorado Homes". Health Physics, Vol. 58, No. 6, pp. 705-713, June 1990.
-------
IIIP-.
TITLE: Pennsylvania Department of Environmental Resources Radon in Water
Measurement Intercomparison
AUTHOR: Douglas Heim and Carl Granlund, Pennsylvania Department of
Environmental Resources
This paper was not received in time to be included in the
preprints so only the abstract has been included. Please check
your registration packet for a complete copy of the paper.
Abstract
An intercomparison of laboratories offering radon in water
measurement services to homeowners was performed. Nine laboratories
including the Pennsylvania Department of Environmental Resources and
the Environmental Protection Agency participated. Six different
wells located near Harrisburg were sampled in triplicate. Sampling
locations were arranged to cover a wide range of concentrations,
e.g., 60 pCi/1 to over 100,000 pCi/1. Laboratory information was
also gathered on sampling procedures, sample analysis, and data
reduction. Information gathered showed a wide range of sampling and
analysis techniques. Reported results were within +/- 25%, with the
exception of one laboratory. Reported lower limit of detection
(LLD) values were as high as 300 pCi/1. Recommendation for future
work includes developing a standardized sampling technique and
standardized analysis methods, the liquid scintillation counting and
the electret methods.
-------
I IIP — 3
TITLE: a Field Comparison of Several Types of Radon Measurement Devices
AUTHOR: Elhannan L. Keller, Trenton State College
This paper was not received in time to be included in the
preprints so only the abstract has been included. Please check
your registration packet for a complete copy of the paper.
A six month comparison study of activated carbon canisters, alpha
tracks, electret passive environmental radon monitors (E-PEKM),
cellulose nitrate foil track etch detectors, and at various times a
continuous radon monitor, was conducted on the campus of Trenton State
College throughout the winter heating season. Monitors were placed at ten
sites in various buildings (dormitories and offices) throughout the
campus. The monitors were subjected to actual environmental conditions
for this time period. Temperature, humidity and barometric pressure
were measured biweekly, with canisters being exchanged at the same rate.
E-PERMs were read and calculated biweekly as well. Both types of track
etch detectors were exchanged on a one, three and six month basis. Ail
data were evaluated and compared relative to the effects of temperature,
humidity and pressure. A preliminary analysis of the data has shown that
temperature and humidity have a significant effect on certain of these
measurement devices.
-------
IIIP
RADON AND WATER VAPOR CO-ADSORPTION ON SOLID ADSORBENTS
by
Neguib M. Hassan, Tushar K. Ghosh, Sudarshan K. Loyalka
and Anthony L. Hines*
College of Engineering
University of Missouri-Columbia
Columbia, MO 65211
and
Davor Novosel
Gas Research Institute
Chicago, IL 60631
* To whom correspondence should be directed
-------
ABSTRACT
The potential health hazard posed by radon and its daughters has led to increased efforts to
develop effective methods for reducing radon concentrations in indoor air. One promising method
for reducting radon in homes is to adsorb it on solid adsorbents such as activated charcoal, silica
gel, and molecular sieves. Activated charcoals are currently used for the measurement of radon
concentrations in indoor environments, whereas silica gel and molecular sieves are finding
increasing application in desiccant based air conditioning systems. Although radon measurements
are carried out in humid air, literature data describing radon adsorption in the presence of water
vapor are limited.
An experimental system has been designed to measure radon concentrations in the solid and
gas phases simultaneously at equilibrium. The uptake of radon by commercially available BPL
activated charcoal, silica gel (grade 40), and molecular sieve-13X were measured at room
temperature (298 K) from dry nitrogen and moist nitrogen under both dynamic and static
conditions. Radon adsorption isotherms were of Type III for all the three adsorbents with
adsorption capacities being highest for activated charcoal and the lowest for silica gel. The presence
of water vapor reduces the adsorption capacities for radon on BPL activated charcoal considerably.
The applicability of these data to improve the radon measurement by charcoal canisters and to
design a removal system are also discussed.
1
-------
introduction
, „ ™ls to adsorb and retain radon has been employed m .he
The ability of activated <*"">	^ concemratio„s indoors. The amount of
design of charcoal canisters that are use	^ # Kmperature is a function of the gas
radon that is adsorbed per unit weight o	^ ^ by , curve known as "isotherm".
phase concentration. This relation^,in a canister can reach equilibrium corresponding
Therefore, radon adsorbed from air by	^ ^ From (he knowledge of ^
,o its gas phase concentration, pro^e	^ ^ ^ ^ _ the ncentmion of
phase radon concentration, the	of lhe charcoal such as surface area, pore
radon in the gas phase. ^	play an imporlant role in the uptake capacity.
size distribution, and c^emlC||	influence the adsorption process.	.
The presence of water vapor a *c	^ ^ parameIers on ^ radon adso^uon
Several researchers have	^ ^	the tempera,„re or relat.ve
capacities of charcoals. Most	However, Cohen (1) stated that the correction due
humidity or both decreases t e	^ 6% However, George (2) Ren and Lin (3),
l0 ,he change in relative humidity was	^ &	from thelr expenmentai
scaripitta and Harley (4) and Ronc^ ^ ^ ^ ^ ^ ^ as 50%
results that relative hum,dims^cand	^ ^ ^ ^ ^ (2) a)so reported that (he
humidities changed from 20 0 ¦	whe„ temperature was changed
amount of radon adsorbed by	otoved .d change of 1.5% in adsorption capacity
from 17 to 27°C. on the other han ,	^ of temperature and humidity,
for each 1"F change in temperature. ,	mance of a dlffusi0n barrier charcoal canister,
both theoretically and	b m when the temperature was increased from 13
The adsorption capacity for ra on	tjs .ncrcas(,d from 15 to 90% at 3S°C.
to 35°C, and by a factor of 3 when	,urc on the adsorption capacity of radon by
The effects of relative humid y	a "calibration factor" which has a dimension
charcoal have been taken into account y e	^ ^ ^ ^ EnvironmeMal Protection
of Lmin-1. This procedure has »ls0	.n hom£S Recently, ,he General Accounting Office,
Agency for measuring radon con^	recomme„ded by the EPA can give an average
GAO (7), found that the radon	chatcoal canister type detector can give an error
concentration error ranging rom	^
as high as 133* with an average err	^	^ determine radon concentrations, there
Although adsorption isotherm	Uterature. Most studies were conducted from
does not appear to be any recent data in	^ ^ ^ adsorbenls that were specially
1950 to 1960 (8-12). Also, the equilibrium
2
-------
prepared or are not commercially available at the present time. In those previous studies, the
equilibrium data were obtained by using air, argon, hydrogen, nitrogen, and carbon dioxide as
carrier gases at atmospheric pressure and near room temperature. The experimental data from these
studies also indicated that the radon adsorption capacities were less for silica gel and molecular
sieves than for activated charcoal.
Although activated carbons are used to measure radon concentration, they have not been
employed in the design a radon removal unit. In the past, however, activated charcoal has been
used to control radon concentrations in uranium mines (13-16). Such a radon removal method
could have several advantages over present radon removal techniques (increased ventilation,
sealing of cracks and and joints, and source control). For example, an adsorption process may be
more energy efficient than one that employs increased ventilation, since radon can be allowed to
decay in the bed and the bed would not require a great deal of energy for regeneration. A few
studies have been reported that relate to the development of a commercial radon removal unit (16-
19). In those studies, the effects of temperature, relative humidity, radon concentration, air flow
rate, and the concentration of carbon dioxide on radon adsorption were investigated. Recently,
Bocanegra and Hopke (20) investigated the adsorption of radon on several types of activated
carbons in the presence of several pollutants including iso-octane, ethylene chloride, and
formaldehyde. The dynamic adsorption coefficients and the number of theoretical stages also were
obtained at 20 °C. No dynamic studies have been reported in the literature where either silica gel or
molecular sieves were employed as adsorbents.
In the present study, the equilibrium uptake of radon by commercially available BPL
activated charcoal, silica gel (grade 40), and molecular sieve-13X were measured at 298 K from
dry and moist nitrogen under both dynamic and static conditions. The equilibrium data were
correlated with a modified Freundlich equation. The applicability of these data to improve the radon
measurement by charcoal canisters and to design a removal system are discussed.
EXPERIMENTAL SECTION
Materials: The adsorbents used in this study were BPL activated charcoal provided by Calgon
Carbon Corporation, and molecular sieve 13X and silica gel (grade 40) provided by Davidson
Chemical Division of the W.R. Grace & Co., Baltimore, Maryland. The properties of these
adsorbents are presented in Table I. Radon gas was generated from a model Rn-1025 Pylon flow-
through source by flowing dry nitrogen gas from a cylinder. The source is a sealed container,
containing dry powder of 22^Ra with a stated activity of 22.6 kBq. The source is capable of
producing a constant radon gas at 2.847 Bq/min (76.87 pCi/min). A certified gamma calibration
source, model Ra-226-Sc, obtained from The Nucleus, Inc., Oak Ridge, Tennessee, was used as a
3
-------
reference to calibrate the NaI(Tl) spectroscopy system before each run. It had a stated activity of
548 Bq (14800 pCi).
ppnrF.niJRES/.
adsorption studies were carried out using an all glass
E0*—'heTy" em is shown in Figure 1. The sample holding tube was
apparatus. A schematic . g ^ ^ ^ ^ ^ ^ ^ boitom (o ma|ch (he geomeny of the
a 0.5 m * 48 mm d.ame 8	^ (he saraple holding tube was
gamma calibration sou •	whhin + 0. l K of the desired value by circulating water
USed to control	mbe was positioned to provide direct contact with a lead-
through the coil. The	^ which was coupied to a multichannel pulse height
shielded 50 mm x 50 mm	a ^ ^ continuously flushed into a 2.8-liter glass
analyzer with associate e ec	• Hnder through the source at a constant flow rate. The
chamber by flowing	mtrogen was vented through a fume
flow rate was ^	of caUbration prior to each run and radioactivity calculation
hood equipped with	Adsorbent samples were spread uniformly on the
was described in °Ur	has^; same geometry as the gamma calibration source,
bottom of the samp e	g ^ ^ sufficient t0 achieve statistically significant
By trial and em*. .1^	^ ^^ sample was generally stacked in a 2-3 layers a, the
counts from the Na < )	^ ^ ^ ^ amounK o( ^ samp,e were used> ^
bottom of the samp e » .	^ ^ ^ ^ gamma	decreased signiflcamly
was adsorbed on th p	y ^ ^ ^ waj required befae any signiflcam
before being detecte	^	than 5 g of (he amp]c were used
='—slm the detector were insignificant relative to the background level, even
after adsorption was continued f0^° he	under vacuum at a temperature of 573 ± 0.1 K
The sample was	was coo)ed t0 the desired adsorption temperature, and
for 12 hours. After regen" ^ ^ Lucas cell was recorded. The sample holding tube was then
a background count obtaine rom	^ radon-laden nitrogen was introduced into the
placed direC'^ ""Isrincrements of approximately 50 mmHg. After each increment. 3.5 hours
system in small p	^ adsorbent t0 reach radioactive equilibrium; It
were allowed for radon and its aug	adsorption equilibrium, which was
shou,d be noted that this e<,u tbnum	^ jn	„ minute, Thjs was
—~ rol —ive equilibrium was reached, the gas phase and soiid phase
4
-------
counts were obtained simultaneously at ten minute intervals. Subsequent data points were obtained
by admitting more radon-laden nitrogen into the system and following the same procedure.
The isotherm data of radon and water vapor co-adsorption were obtained by initially
equilibrating the adsorbent with water vapor up to the desired relative humidities i.e., 40%, 60%,
80%. The radon-laden dry nitrogen was then introduced into the system in small pressure
increments of approximately 50 mmHg, and radon and its daughters were allowed to equilibrate in
the usual manner. The dry radon-laden nitrogen stream from the radon source was bubbled
through water in two saturators in series. The saturators were immersed in a constant temperature
bath whose temperature was controlled within ±0.1 K to maintain the desired relative humidity of
the gas stream. The gas stream could be admitted into either a 2.8-liter glass chamber, so that
humid radon in nitrogen can be introduced into the system in small pressure increments, or directly
passed through the packed bed during the dynamic study.
Dynamic Adsorption: The adsorption column was packed with approximately 50 grams of
adsorbent sample. A bed diameter to particle ratio of 17.5 was maintained in the bed to avoid
channeling and wall effects. A superficial gas velocity of 0.198 Lmin"! was maintained in the bed
during experimental runs. The concentration of radon in the inlet nitrogen stream was maintained at
220 ± 20 pCi/L throughout the experiments. The bed was regenerated by flowing dry nitrogen at
473+ 0.1 K for 10 hours. After regeneration, the adsorbent bed was cooled to the desired
adsorption temperature, and a background measurement was obtained using the Lucas flow-
through cell. The nitrogen was directed to the source, which was flushed for one hour at a constant
flow rate. The dry or moist radon-laden nitrogen was then passed from the top of the column, and
the radon activity at the column exit was measured by the Lucas cell along with the AB-5 radiation
monitor at five minute intervals. The radon was allowed to flow into the column until the
breakthrough was complete and the counts of the effluent reached a maximum constant value,
equal to that of the inlet gas stream.
BF.SIfl.TS AND DISCUSSION
F.quilibrium Isotherm Data: The equilibrium data of pure water vapor were first obtained on BPL
activated charcoal, silica gel, and molecular sieve 13X at 298 K. As can be seen from Figure 2, the
shapes of the isotherms are different for different adsorbents. The isotherms are Type I on silica
gel, Type II on molecular sieve-13X, and Type V on activated charcoal, which suggest different
adsorption mechanisms. Small amounts of water vapor were adsorbed on charcoal for relative
humidities below 40%, followed by a sharp rise in the uptake at relative humidities from 40 % to
60%. This is mainly due to the pore filling of the capillaries.
5
-------
The experimental data for radon adsorption on BPL activated charcoal, silica gel (grade
40), and molecular sieve 13X at 298 K are shown in Figure 3. The error in the count rate of the
gas phase was determined at the one sigma significance level and it ranged from 4.79 % to 11 92
%. The minimum detectable activity due to background count was found to be 21.2 counts
minute at the three sigma significance level. Adsorption measurements were repeated for each
to check the reproducibility of the data. Although radon adsorption is a random process the
experimental data were reproducible, with an average error of less than 5%, as indicated in Figure
3. The uptake of radon by activated charcoal was considerably higher than that of silica gel and
molecular sieve 13X. However, molecular sieve 13X exhibited a higher affinity for radon than was
found for Type 5A (10). This may be due to the larger pore diameter of Type 13X. The silica gel
exhibited a relatively low radon adsorption capacity. The equilibrium data exhibited a Type m
isotherm. Once an atom or a molecule is adsorbed, adsorbate-adsorbate interactions promote the
adsorption of further atoms or molecules so that the isotherm becomes convex to the pressure *
The Type III isotherms were also obtained by Przytycka (11), Brutt and Kurbatov (12) and
Coleman et al. (10).
The equilibrium data of radon on BPL activated charcoal, preadsorbed with water vapor at
40%, 60%, and 80% relative humidities are compared in Figure 4 with that obtained for d
nitrogen. The isotherms were also of Type III in the presence of water vapor. However the
amount of radon adsorbed on a pre-equilibrated activated charcoal was lower than that adsorbed
a dry charcoal. It may be noted that when radon-laden nitrogen was introduced into the system
small amount of water vapor was desorbed from the charcoal. In a separate run, radon free
nitrogen was introduced into the system under similar conditions and the same results
observed, suggesting that probably some other impurities in nitrogen is displacing water
from the charcoal surface. The interaction between radon and charcoal is not strong enough^
displace the preadsorbed water vapor. Thomas (17) and Strong and Levins (18) made the
conclusion from their dynamic adsorption studies; however, the results contradicted oth^
investigators' conclusions (6,2).
Dynamic Adsorption Data : Breakthrough curves for radon on BPL activated charcoal silica
(grade 40), and molecular sieve 13X are shown in Figure 5. The bed lengths for'act" * ^
charcoal, silica gel, and molecular sieve were 10.16, 11.43, and 13.97 cm, respectively^^
superficial velocity for nitrogen gas through the adsorbent bed was 0.198 cm/min, and the Dr *
drop across the bed varied from 10 to 15 mmHg. The flow rate of nitrogen through the ^
source was maintained approximately constant at 0.35 Lmin"1. This flow rate produces 9^°°
radon concentration of about 220 pCi/L (radon partial pressure of 1.2 x 10" 12 mmHg) with
fluctuations of ± 20 pCi/L for all experimental runs. It was noted that the various solid adsorb^
-------
exhibit different steady state outlet concentrations for the same inlet radon concentration. According
to Madey (22), the steady state concentration of radioactive gas at the outlet of a column of length L
is related to the inlet concentration through the expression
C ouuci=C exp( - yL)	(1)
where y is the decay constant for the radioactive gas. The breakthrough curves for radon from the
radon-water vapor mixture on BPL activated charcoal were measured at different relative
humidities. The results are shown in Figure 6. The steady state radon concentration at the column
outlet was found experimentally to be more than the inlet concentration. This may be due to
increased radon plate out on the inside of the Lucas cell. A similar observation was reported earlier
*
by Boncanegra and Hopke (20), who found that the presence of water vapor causes neutralization
of 218Po ions and results in their deposition in the detector. The adsorption of radon on BPL
activated charcoal from humid nitrogen is considerably less than that from a dry nitrogen stream.
The presence of water vapor in the nitrogen stream reduced the uptake capacities for radon due to
the strong competition by water vapor molecules for the available sites. However, it is interesting
to note that radon broke through the bed approximately at the same time irrespective of the relative
humidities of the nitrogen stream. Water vapor molecules, having a stronger affinity for activated
charcoal than radon, might be adsorbed near the inlet section of the bed at a faster rate. As
adsorbed water front progresses through the bed it is displacing radon, resulting in a faster radon
breakthrough from the bed. As shown in Figure 7, repeated experiments for radon adsorption from
dry and moist nitrogen on the activated charcoal verified the reproducibility of the results.
Equilibrium Data Correlation: The Freundlich equation has been modified to correlate the
equilibrium adsorption data of radon from dry nitrogen. The modified equation can be written as
q=ki(P~)n	^
O
or
P
lnq=lnk^+nln(—)	(3)
Po
where q is the amount of gas adsorbed, P is the system pressure, P0 is a reference pressure and is
arbitrarily set to 10"^ mmHg. Here, kj is a measure of the volume of gas adsorbed per unit mass
of adsorbent and n is the intensity of adsorption. Figure 8 shows good agreement between the
7
-------
experimental data and the predicted values. The value of n was set to 1.75 and kj has the following
temperature dependence:
k.|=2.10x10-11-6.58x10-14 T	(4)
Thus, the amount of radon adsorbed on BPL activated charcoal corresponding to the gas phase
concentration can be expressed as
a J	1 A (	\1. 75
q=(2.10x10" -6.58x10" T)
V1x10-14 J
(5)
Knowing the solid phase concentration, the gas phase concentration can be obtained in pCi/L from
the equation
	5.48x103 q0 571 T~1
C (2.10x10_11-6.58x10-14 T)0571 ^
The equilibrium isotherms of radon, such as the ones shown above, can be used for
measuring radon concentrations in homes provided accurate isotherm data are available for the
activated charcoal. From the knowledge of the solid phase radon concentration, the equilibrium
isotherms corresponding to the particular field conditions, temperature and humidity, can be used
to determine the concentration of radon in the gas phase.
Another application for the dynamic and static adsorption data is in the design of an
adsorber unit for removing radon from indoor air. The design of such a unit may range from a
single packed column to a complex system of multiple columns.
Acknowledgements: This work was supported by the Gas Research Institute, Chicago, IL. and
ASHRAE, Atlanta, GA. under contract No. 5089-246-1821, and a grant from the Weldon Spring
Funds. The authors gratefully acknowledge Calgon Carbon Corporation for providing the activated
carbon samples, Davidson Chemical Division of W. R. Grace & Co., Baltimore, Maryland for
providing the molecular sieve and silica gel samples, and Porous Materials, Inc., Ithaca New
York for carrying out the pore size analyses of the adsorbents.	'
8
-------
REFERENCES
1.	Cohen, B. L., and Nason, R., "A Diffusion Barrier Charcoal Adsorption Collector for
Measuring Rn Concentrations in Indoor Air,", Health Physics; 50(4), pp. 457-463
(1986a).
2.	George, A. C., "Passive, Integrated Measurement of Indoor Radon Using Activated
Carbon", Health Physics; 46(4), p. 867 (1984).
3.	Ren, T. and Lin, L., "A Passive Integrating Indoor Radon Detector with Activated
Carbon," Radiat. Prod. Dosim. 19, 121 (1987).
4.	Scarpitta, S. C., and Harley, N. H., (New York University Medical Center), "Adsorption
and Desorption of Noble Gases on Activated Charcoal II: Radon Studies in a Monolayer
and Packed Bed.", (Prepared for: U.S. Department of Energy) 1989.
5.	Ronca-Battista, M.; Gray, D., "The Influence of Changing Exposure Conditions on
Measurements of Radon Concentrations with the Charcoal Adsorption Technique",
Technical Exchange Meeting on Passive Radon Monitoring; Grand Junction, Colorado
(Sept. 21-22, 1987).
6.	Pojer, P. M., Peggie, J. R., O'Brien, R.S., Solomon, S. B. and Wise, K. N.,
"Performance of a Diffusion Barrier Charcoal Adsorption ^^Rn Monitor Under
Conditions of Varying Humidity and Temperature," Health Phys. 58(1), 13 (1990).
7.	GAO, United States General Accounting Office, Report GAO/RCED-90-25 Radon Testing
(1989).
8.	Gubeli, V. O., and Stammbach, K., "Zur Adsorption von Radon an Aktivkohle und
Silicagel", Helvetica Chimica Acta; 34(154), pp. 1257-1263 (1951).
9.	Gubeli, V. O., and Stori, M., "Zur Mischadsorption von Radon an Aktivkohle mit
verschiedenen Tragergasen", Helvetica Chimica Acta.; 37(260), pp. 2224-2231 (1954).
10.	Coleman, R. D.; Kuznetz, H. L.; Woolrich, P. F., and Holaday, A. D., "Radon and
Radon Daughter Hazards in Maine Atmosphere", American Industrial Hygiene Association
Quarterly; 17 (1957).
11.	Przytycka, R., "Sorption of Radon on Activated Polish Charcoals", Nukleonika; 6(1), pp.
23-32 (1961).
12.	Burtt, B. P., and Kurbatov, J. D., "Mixed Adsorption of Radon and Argon on Silica Gel",
Journal of American Chemical Society; 70, pp. 2278-2282 (1948).
13.	Ackley, R. D., (Oak Ridge National Laboratory), "Removal of Radon-220 from HTGR
Fuel Reprocessing and Refabrication Off-Gas Streams by Adsorption (Based on A
9
-------
14.
15.
16.
17.
18.
19.
v Reoort No. ORNL-TM-4883, (Prepared for: Chemical Technology
Literature Survey) > V
Division (ERDA)) (WTO- r ^ Qno s „0n the Swdy of Rado„ Removal with Active
Fusamwa, .	^ (Jap). 79> pp 590.597 (1963).
Carbon , PP	Nati0„al Laboratory), "The Low-Level Gamma Counting Room:
Lucas, H.	rontrol" Report No. ANL-7360, (Prepared for: DOE) (1963).
Radon Remova an	R p (BatteUe> Pacific Northwest Laboratory), "Cleanup
"llaiion Study Effluent Report No. BNWL-SA-5541, (Prepared for: U.S.
U onH DeveloDment Administration) (1975).
Energy Researc	Ac,ivated Carbon Canislers for Radon Protection i
Th°maS' «• Report No. HASL-280 (1974).
Uranium i ns D M "Dynamic Adsorption of Radon on Activated Catbon
DOEN^ AidingConference; p.Conf-780819 (1978).
G • Porstendorfer, J., and Wicke, A., "Radon Adsorption m a Gas F
Scheibe, ¦	, Heaith Physics; 38, pp. 83-85 (1980).
Activated Charcoa^ ^ R _ „Radon Adsorption on Activated Carbon and the
Eftets"^ Some Airborne Contaminants", The Science of the Total Environment; 76, pp.
193-202 nT Ghosh, T. K„ Hines, A. L„ Loyalka, S. K„ and Ketring, A., » New
ExTrimental Apparatus for Measuring Radon Adsorption on Solid Adsorbent", Submitted
for PuW^atl°^ |"j^Bgebe^M. R.. and Stephenson, T. E„ "Adsorption of Krypton-
^ — carbon", Trans. Am. Nucl. Sc., 3(2), pp. 373 (1960).
15th
Flow by
21.
22.
nomenclature
first empirical constant in the modified Freundlich equation
n constant in the modified Freundlich equation
p equilibrium partial pressure of radon (mmHg)
p„	reference pressure (mmHg)
q volume of radon adsorbed per unit weight (cm' Rn/g adsorbent)
r gas constant
T temperature (K)
10
-------
LIST OF	TABLES
Table 1	Properties of the Microporous Solid Adsorbents.
LIST OF FIGURES
Figure 1.	Schematic flow diagram of radon adsorption apparatus.
Figure 2.	Adsorption isotherms of water vapor on various solid adsorbents at 298 K.
Figure 3.	Adsorption isotherm data of radon on various solid adsorbents at 298 K
Figure 4.	Radon and water vapor co-adsorption data on BPL activated charcoal at 298 K.
Figure 5.	Experimental breakthrough curves for radon on solid adsorbents at 298 K.
Figure 6. Breakthrough curves for radon adsorption from moist nitrogen stream on BPL
activated charcoal at 298 K.
Figure 7. Reproducibility of radon breakthrough data on BPL activated charcoal.
Figure 8. Correlation of radon adsorption data with the modified Freundlich equation.
11
-------
Table I. Properties of Microporous Solid Adsorbents.
Property BPL Activated Silica Gel	Molecular Siev?"
	Charcoal	(Grade 40) 13X
Particle size, (A)b 6x16 mesh	6x12 mesh	8x12 mesh
Surface area, S(m2/g)a
micropores	294
meso and macropores	50 9	101
total	874 672	395
Pore volume, V (cnvVg)a
micropores	^.47 0.38	0.14
meso and macropores	0.10 0.02	0.27
total	0.57 0.40	0.41
Average pore diameter,
4V/S(A)a	26 24	41.7c
Bulk density (g/cm^	0.60	0.72
0.72
Equilibrium water
capacity (% wt.)b	"	-	5
Moisture content
as shipped (% wt.)b	<1	-	<15
a Analysis made by Porous Materials, Inc. Ithaca, New York,
b Analysis provided by the manufacturer.
c based on total surface area and pore volume.
12
-------
Cahn D-lOO
Electrobalance
Vacuum Bottle
Pressure
Gauge
Pressure
Gauge
Control Recorder
Unit
Pressure
Gauge
Balance
Thermocouple
Hangdown
Tubes
Sample
Lucas Cell
Nal(TI) Detector
Vacuum
Trap
Dry Ice
Pylon AB-5
Radiation Monitor
Photomultiplier
Tube
Diffusion
Water Container Arrangement for Multicomponent
Study
From Radon
Gas Supply
System
Vacuum
Pumps
Reader
Amplifier MCA System
Photomultiplier Base
and
Preamplifier
HV Power Supply
Figure 1. Schematic Flow Diagram of Radon Adsorption Apparatus
-------
0.35
0.3
0.25
CD
r,
£ 0.2
u
o>
¦go.15
a.
0.1
0.05
0
0	5	10	15	20	25
P (mmHg)
Figure 2. Adsorption isotherms of water vapor on various solid adsorbents
at 298 K.
Molecular Sieve-13X
Silica Gel (Grade 40)

&
-A-
•0

A

-

.i
¦i
-i


^ ^ & Activated Carbon BPL $
D'
a

¦-^^oqow ?°yQ4


-------
10
o 8
x
c
js 0
s_
O
1/5
¦o
<
J5JD 4
-------
4
s
i>
jd
s_
O
¦c 2
<
fc£
"c
o
"D
a
X i
0
I t I I I I I I I I I
"I	I I I	1	1	1 I I	1 I I I I
I I
/o
/
/
/
A

/
A
A '
/
/
A
/
So /

A
*
~ A
~

/ A/
/ A ° ^
J	I	L
J	L.
I
j	¦ ¦
_L
_i	1	i
—A- - Radon only
— Jt— 40% RH
60% RH
80 % RH
_l	I	I	I	''I'	L
0	4	8	12	16	20	24	28	32
P (mmHg) x 1014
Figure 4. Radon and water vapor co-adsorption data on BPL activated carbon at 298 K.
-------
240
200 -
U
o.
160 -
5
c
o
CD,
CQ
h 120
c
a>
tj
c
o
w 80
c
©
¦o
CQ
* 40
o
5
%
o
*
I
1
/


a/
/

£ — k— BPL activated carbon -
—	~— Silica gel
—	O— Molecular sieve-13X
A 1 A 1 A i-rf- A1 A		1	L
4	8	12
0	4	8	12	16	20	24
Time (hr)
Figure 5. Experimental breakthrough curves for radon on solid adsorbents at 298 K.
-------
250
j 200 -
O
a.
I 150
03
= 100
&
0
£ f t
Inlet radon concentration: 220 pCi/L_
Gas stream flow rate: 0.35 L/min
~ Radon only
~ 20% RH
O 60% RH
12	18
Time (hr)
24
30
36
Figure 6. Breakthrough curves for radon from moist nitrogen stream on BPL
activated carbon at 298 K.
-------
320
s240
a.
160
e
o
ea
s.
+*
c
a>
v
C
O
(J
c
¦§ 80
CS
X
0
0
A	Radon only (R# 1)
~	Radon only (R# 2)
O	40% RH (R# 1)
¦	40% RH (R# 2)
I I



*
'o
.A

v
/
/D
4/
Q.
S-'
4	6




J	i	i	i	L
* ¦ * ¦
8
10
12
14
Time (hr)
Figure 7. Reproducibility of radon breakthrough curves on BPL activated
carbon 298 K.
-------
V- 1 0
Equation
P (mmHg)
Figure 8. Correlation of radon adsorprion data with the modified Freundlich equation.
-------
IIIP-5
CALIBRATION OF MODIFIED ELECTRET ION CHAMBER
FOR PASSIVE MEASUREMENT OF RADQN-220 (THORON\ IN AIR
by: P. Kotrappa, Ph.D., and J. C. Dempsey
Rad Elec, Inc.
5310 H Spectrum Drive
Frederick, MD 21701 USA
ABSTRACT
A commercially available electret ion chamber known as
the E-PERMr* (Electret-Passive Environmental Radon Monitor) is
now widely used for both indoor and outdoor radon-222 (radon)
measurements. This unit was designed to minimize the response
to radon-220 (thoron) by restricting the diffusion entry time.
The chamber of such a unit was modified by increasing the
filtered diffusion area from 0.3 cm2 to 30 cm2 to allow thoron
to enter the chamber in and out with very little delay time.
Such an E-PERM is termed a Radon-Thoron E-PERM or RT E-PERM
since it responds to both radon and thoron. A steady state
thoron concentration was generated in a room using a Pylon
thoron source, a pump and a large fan. Several R E-PERMs
(radon) and RT E-PERMs were positioned at a location from
where a 24-hour sample was also drawn by a large double filter
unit (18 liters) for measuring the average thoron
concentration at that location. This paper describes the
calibration procedures and the use of an E-PERM and RT E-PERM
side by side to obtain both radon and thoron concentrations in
a location. The units have sufficient sensitivity to measure
1 pCi/L of thoron in five days. Such units are useful in
thorium rich areas or thoron refineries or in some special
cases where thoron can either be a problem or an interference.
*E-PERMR is the registered trademark of the product
patented and manufactured by Rad Elec Inc., Frederick, MD
21701 USA.
-------
INTRODUCTION
A commercially available electretion chamberknown as an
t? pprm (Electret-Passive Environmental Radon Monitor) is now
SldeW 'used for both indoor and outdoor radon-222 (radon)
measurements. This unit was designed to minimize the response
radon-220 (thoron) by restricting the diffusion entry time.
are fully described in literature (lf 2). To
IncSase the responte o£ these devices to thoron, it is
possible to modify the unit by increasing the area of the
filtered inlets so that thoron can get in and out more
readily. Such a modified unit is a Radon-Thoron E-PERM or RT
v pfrm since it responds both to radon and thoron. The
purpose of the present work was to calibrate the units and
describe a procedure for making a measurement of thoron in
air The principle and the analysis procedure is somewhat
similar to those made by Pearson (3) for alpha track
detectors.
DESCRIPTION OF RADON AND RADON-THORON E-PERMS
Figure 1 gives a schematic view of the two units. The
left half of Figure 1 shows the standard E-PERM. This has a
small filtered inlet area ( 0.3 cm2) to restrict the entry of
side of Figure 1 shows the modified unit
thoron. The right s:
FILTERED
INLET
ELECTRET
8.1 cm

"S" Chamber Schematic
(210 mL)
RADON E-PERM®
FILTERED
INLET
ELECTRET
"S" Chamber Schematic
(210 mL)
RADON-THORON E-PERM®
Figure 1
-------
A series of holes were drilled in the body of the
chamber, and the holes were subsequently closed using an
electrically conducting filter paper. The area of the opening
was about 30 cm2. In both the units, the electret can be
covered or uncovered bringing the unit to the "off" or "on"
position, respectively (1, 2). Standard short-term electrets
are used in these devices (1, 2).
CALIBRATION
To calibrate these devices for thoron, it was necessary
to produce a steady state thoron concentration in the room
air. An upper floor closed room with a dimension of about 12
x 12 x 8 feet was chosen for generating a steady state of
thoron. This room was known to have a radon concentration of
less than 0.5 pCi/L. A PylonR** Model TH-1025 flow-through
thoron gas source was used as a source of thoron. An air flow
rate of 5 liters per minute was established through the source
using an appropriate pump and a flowmeter. The outlet of the
source was taken through a polyethylene tube and was
positioned at the center of a large fan to disperse the thoron
into the room air. The system was run for at least one day to
achieve a steady state thoron concentration.
A small table was located at a distance of about 7 feet
from the fan in the direction of the air flow. This location
was chosen as a reference location for measurement of thoron.
An air sample was drawn from the center of the table
through a large (18 liter volume) double filter unit (4, 5)
for 24 hours. The second filter paper of the double filter
unit was counted for alpha activity in a standard alpha
scintillation counting unit after a delay of 12 hours.
Standard procedure was used for analyzing the results. Table
1. gives the thoron concentration in air over a period of
eight days from the 24-hour samples collected each day.
A set of three R E-PERMs and set of three RT E-PERMs were
also located on the same test table for the entire period of
eight days. The electrets were read at the end of each day
and reloaded back into the units for the next run. Table 1.
gives successive electret readings taken at the end of each
day.
The calibration factor is obtained (Table 1.) by dividing
the net average electret voltage drop per day by the average
thoron concentration over the experimental period.
**PylonR is a registered trademark of Pylon Electronic
Development Company, Canada and USA.
-------
TABLE 1. CALIBRATION OF R AND RT E-PERMS FOR THORON. SECOND
ROW GIVES THE AVERAGE THORON CONCENTRATION ON THAT DAY. R-l
to R-3 GIVE SUCCESSIVE ELECTRET VOLTAGE READINGS OF R-EPERMS.
RT-1 TO RT-3 GIVE SUCCESSIVE ELECTRET VOLTAGE READINGS OF RT
E-PERMS^
Time
(Davs)
0
1
2
3
4
5
6
7
8
Thoron
(pCi/L)
_
26.6
24.4
23.0
24.8
24.5
23.5
25.2
26.0
R-l
(volts)
515
507
501
491
479
471
463
453
447
R-2
(volts)
518
512
504
492
482
472
464
456
448
R-3
(volts)
550
542
534
524
514
502
494
488
480
RT-1
(volts)
702
660
611
560
507
452
398
355
308
RT-2
(volts)
701
659
606
552
494
435
376
330
280
RT-3
(volts)
438
384
331
283
237
196
155
113
-
AVERAGE DAILY THORON CONCENTRATION
AVERAGE VOLT DROP PER DAY FOR R E-PERM
NET*
AVERAGE VOLT DROP PER DAY FOR RT E-PERM
NET*
AVERAGE CALIBRATION FACTOR FOR R E-PERM
AVERAGE CALIBRATION FACTOR FOR RT E-PERM
24.8	pCi/L
8.7 volts
7.0 volts
49.6 volts
47.9	volts
0.2823 **
1.9314 **
* Subtracting contribution from gamma background of 10 uR/h
~~Calibration factors are in units of volts per pCi/L-day
-------
RESULTS
Table 1. gives the results. It can be seen that thoron
concentration did not vary from day to day in any significant
way justifying taking the average thoron concentration over
the entire period as nearly constant. The voltage drop per
day also remained nearly the same. The E-PERMs are recommended
to be used with electrets in the voltage range of 200 to 700
volts (1/ 2). The data indicated that the calibration factor
did not change significantly with the operating voltage over
the voltage ranges studied. More sophisticated and accurate
work may indicate a marginal dependence of the calibration
factor with the operating voltage as is found with E-PERMs for
radon (1/ 2). It is not expected to be more than 6% to 8%.
DERIVATION OF AN EQUATION
FOR CALCULATING THORON CONCENTRATION
FROM THE SIMULTANEOUS DATA OBTAINED WITH R AND RT E-PERMS
Let the radon concentration be R pCi/L.
Let the thoron concentration be T pCi/L.
Let the gamma radiation background be 10 uR/h.
Let both detectors be exposed to D days at the same location,
and let the voltage drops be VI and V2 volts, respectively.
Detector #1 (known as the Radon E-PERM) and #2 (known as
the Radon-Thoron E-PERM) have the same responses for radon and
gamma radiation but different responses for thoron because of
modifications done to detector #2.
Let the calibration factor for radon be C(R) volts per pCi/L- .
day.
Let the gamma response be G volts per 10 uR/h -day.
Let the calibration factor for thoron for detector #1 be C(T1)
volts per pCi/L-day.
Let the calibration factor for thoron for detector #2 be C(T2)
volts per pCi/L-day.
Then VI and V2 are given by equation (1) and (2).
VI = C(R) x D x R + C(T1) xDxT + GxD 	 (1)
V2=C(R) xDxR + C(T2)xDxT + GxD 	 (2)
Solving equation (1) and (2) for T leads to:
T - ( VI - V2 ) / ( D x ( C(T1) - C(T2) ) ) 	 (3)
Example: Data on R-l and RT-1; VI = 394 volts; V2 = 68 volts;
D = 8 days;
T = (394 - 68) / ( 8 x ( 1.9314 - 0.2823 ) ) = 24.5 pCi/L
-------
recommended procedure
Colocate a R and a RT E-PERM at the location where the
thoron concentration has to be measured. Leave them for 2 to
7 days. Use the data for analyzing the thoron as illustrated
in the earlier section. Data on the R E-PERM can be used for
calculating the radon concentration using the standard
procedure (1, 2). If one wishes to measure cumulative thoron
concentrations which are relatively higher (more than 250
pCi/l-days), a long-term electret (1, 2) can be used. Long-
term electrets have calibration factors which are
approximately lower by a factor of 10.5 .
DISCUSSIONS
Work described by Pearson (3) indicated that a
simultaneous use of two alpha track detectors (one sensitive
to thoron and the other not sensitive to thoron) as a method
of determining the thoron concentration in air. The
calibration was done (3) at very high concentrations (1,000 to
20,000 pCi/L-days) compared to the concentrations used in the
current work (25 to 200 pCi/L-days) mainly because of the
sensitivity limitations of alpha track technology. The E-PERM
technique appears to give a practical method of measuring
thoron concentrations usually found in homes with an
acceptable accuracy and in relatively short time intervals.
GENERAL
The work described in this paper was not funded by the
U.S. Environmental Protection Agency, and therefore the
contents do not necessarily reflect the views of the Agency
and no official endorsement should be inferred.
-------
ACKNOWLEDGEMENTS
The authors are grateful to Mr. Carl Granlund of the
Department of Environmental Resources in Harrisburg, PA, for
loaning the thoron source and the alpha counting unit. The
authors are also grateful to Leena Kotrappa for her help in
experiments and in data analysis and to Lu markland for
editorial assistance.
REFERENCES
1.	Kotrappa, P., Dempsey, J. C., Hickey, J. R., and Stieff,
L. R. An electret passive environmental Rn-222 monitor
based on ionization measurement. Health Physics 54:47-
56, 1988.
2.	Kotrappa, P., Dempsey, J. C., Ramsey, R. W., and Stieff,
L. R. A practical E-PERMR environmental radon monitor
system for indoor radon measurement. Health Physics
58:461-467, 1990.
3.	Pearson, M. D. Calibration of alpha-track monitors for
measurement of thoron (Rn-220). In: The 1990
International Symposium on Radon and Radon Reduction
Technology. United States Environmental Protection
Agency. Atlanta, GA, 1990. Paper III-P-6.
4.	Kotrappa, P., Soman, S. D., and Mayya, Y. S. Modified
double filter apparatus for measuring radon/thoron in the
environment and in exhaled breath. In: IAEA
International Symposium on Advances in Radiation
Protection Monitoring. Stockholm, 1978. Paper IAEA-SM-
229-31, page-413.
5. Thomas, J. W. Thoron determination by two-filter method.
USAEC Health and Safety Report. HASL-TM-71-1. 1971.
-------
TITLE: Unit Ventilator Operation and Radon Concentrations in a
Pennsylvania School
AUTHOR: William P. Brodhead, WPB Enterprises
This paper was not received in time to be included in
preprints and the abstract was not available. Please check
registration packet for a complete copy of the paper.
-------
Session IV:
Radon Reduction Methods
-------
IV-
CAUSES OF ELEVATED POST-MITIGATION RADON CONCENTRATIONS
IN BASEMENT HOUSES HAVING EXTREMELY HIGH PRE-MITIGATION I FVFi c
by: D. Bruce Henschel
Air and Energy Engineering Research Laboratory
U. S. Environmental Protection Agency
Research Triangle Park, NC 2771 1
Arthur G. Scott
AMERICAN ATCON, INC.
Wilmington, DE 19899
ABSTRACT
Forty basement houses in Pennsylvania which had received EPA-sponsored indo
radon mitigation systems in 1985-87 as part of an earlier project, were re-visited in 1989-gn
to permit further testing. These houses had generally had very high pre-mitigation rad
concentrations (commonly 50 to 600 pCi/L, or 2 to 22 kBq/m3); a significant fraction still ha°n
residual (post-mitigation) levels greater than EPA's original guideline of 4 pCi/L (148
based upon alpha-track detector measurements. The objective of the follow-up testing w
to assess why levels were still elevated, and what additional steps would be required in ord^
for these houses to achieve both the original guideline of 4 pCi/L, and a more challenaino n 6|
of 2 pCi/L (74 Bq/m3).	y 8 9°al
In houses having sub-slab and drain-tile depressurization systems, the primary sin I
cause of elevated residual levels was re-entrainment of the high-radon fan exhaust; airbo °
radon resulting from radon in well water was an important secondary contributor in so™6
houses. Care in design of the system exhaust, and treatment of the water, would be requi^
to reduce these houses below 2 pCi/L. In only one house with a sub-slab system did'th
elevated residual levels clearly appear to be due to inadequate depressurization beneath th6
slab. However, in houses having block-wall depressurization systems, inadequate sub-sl K
depressurization appeared to be the major cause of the residual levels; exhaust re-entrainm
and well-water radon also played a role in some houses with block-wall systems.
Elevated outdoor radon concentrations, and emanation of radon from poured concr
slabs and foundation walls, were not major contributors to the residual indoor concentratio
with each of these factors contributing on the order of 0.2 pCi/L (7 Bq/m3).	°S'
This paper has been reviewed in accordance with the U. S. Environmental Protect'
Agency's peer and administrative review policies, and approved for presentation
publication.	anc"
-------
INTRODUCTION
During the period June 1985 through June 1987, developmental indoor radon
reduction systems were installed and tested in a total of 40 houses in the Reading Prong
region of eastern Pennsylvania (Reference 1). Most of these installations involved some form
of active soil depressurization (ASD), including sub-slab depressurization (SSD), drain-tile
depressurization (DTD), and block-wall depressurization (BWD). Other mitigation approaches
tested in a few of the houses included active soil pressurization, heat recovery ventilators
(HRVs), and radon removal from well water. All of the houses had basements, sometimes
with an adjoining slab-on-grade or crawl-space wing. These houses were generally difficult
to mitigate, for two primary reasons:
1)	The source term was often extremely high, with soil gas concentrations as high as
50,000 pCi/L (1.8 MBq/m3) measured in one case. As a result, pre-mitigation indoor
concentrations were very high, commonly in the range of 50 to 600 pCi/L (about 2 to
22 kBq/m3). The high source term requires careful treatment of all entry routes and
care in avoiding re-entrainment of ASD exhaust, among other considerations.
2)	Communication beneath the basement slabs was sometimes poor or uneven
complicating the application of ASD systems.
The radon concentrations in the basements and living areas of these houses have been
measured using alpha-track detectors (ATDs) with 3- to 4-month exposure periods, during
each of the winter quarters since the mitigation systems were installed (References 1, 2, and
3). In addition, an annual ATD measurement in the living area was completed during the
period December 1988-December 1989 (Reference 4). The average winter-quarter concentra-
tions for each house, and the annual average living-area concentration, are presented in Table
1. As shown in the table, of the 38 houses still participating in the program, the average
basement concentration over the past two or three winters has been above 4 pCi/L (148
Bq/m3) in 18 of them, and above 2 pCi/L (74 Bq/m3) in 28 of them. The average winter-time
living area concentration has been above 4 pCi/L in 11 of the houses (about 30%), and above
2 pCi/L in 22 (about 60%). The annual average readings in the living area are somewhat more
favorable than the winter-quarter results, with about one-quarter of the houses above 4 pCi/L
and half above 2 pCi/L according to the annual measurement.
Thus, even though the percentage radon reductions were substantial in essentially all
of these high-level houses, a significant number have residual (post-mitigation) radon levels
greater than EPA's original guideline of 4 pCi/L. An even greater number have residual levels
above 2 pCi/L, suggesting that there could be difficulty in achieving the goal of near-ambient
indoor concentrations, specified in the Indoor Radon Abatement Act of 1988
Accordingly, during the winter of 1989-90, additional testing was carried out in all of
these difficult houses in order to better understand why residual radon levels were still
elevated, and what additional steps would be necessary to reduce the indoor levels to near-
ambient. Five possible explanations for the elevated residual levels were investigated-
-------
4)
5)
'.toH hv ASD systems to adequately extend beneath
11	leavin9 some s°"9as en,'v 'outes inadequa,e,y
treated;
2)	re-entrainment of	exhaust ,0. the ASD systems baclc into .he house;
3)	release into .he air o. radon contained in well water,
contribution of ambient (outdoor) radon to indoor levels; and
emanation of radon from concrete slabs and foundation walls.
annroaches not involving ASD, another consideration is possible inherent
/"I'KSSSS o> L mitigation approach,
limitations in the eneci
RESULTS
»of suctianfiElds-gffisffltcdtlY A?r> ?YSterna
= that the suction fields being generated by the ASD systems
The first concern was that the sue	^ ^ ^ adequately be panting
might not be adequately ext^d'"9	the extremely elevated soil gas concentrations at
J gas entry into block walls. In£»« ^ ^ ^ a signlficant impact 0n indoor
many of these houses, any unite
leV0'S'	.en system between 4 and 22 test holes were drilled
In each house having en ASD sy ¦	w|ns t0 permi, measurements o«
through the basement slab and the slab	m Usual|y a test hote was drilled in each
sub-slab depressurization being created y	Mes d(llled in that quadrant where the
corner of the slab, with	„ed to be poorest based upon the results
depressurization being """"^" measurements were made with a micromanometec
from the corner hole. Sub P	>|( Mst h0|es plugged except thei one et vvhich
sensitive to ±0.001 in. WGl±«	of ,humb it ,s estimated that the sub-slab
?he measurement was being made. As ^ ^ 0.015 in. WG (about 4 Pa) in order to
depressurization at a given poin	openings at that point. This value of 0.015 in.
reliably prevent soil gas flow up thro 8	, stack depressurization created in the
WG approximately equals the	^ lt is believed that a sub-slab depressuri-
ha<»pment of a two-story house duringI	smal) percentage of the time by weather
zation of 0.015 in. WG will bean added safetv margin, a depressurization of 0.04
(10 pllshould almost never be overwhelmed.
, u clah rommunication, the sub-slab depressurizations
As a separate measurement of su	mitigation system off, with suction being
at these test holes were also m0a^red Usi a simp|e mathematical model, the results
generated by an industrial vac.^ c,ea~r; ^\0 calculate a "Standard Suction Distance"
from these vacuum cleaner diagnostics jwe	e over which suction drawn through a 4-
(SD) for each slab. T^e is nomman.;, ^ ^ i % Qf tha( being maintained under the
in no-cm) diameter SSD suction nu.«
-------
slab immediately under the SSD pipe. One percent of the suction under the SSD pipe would
typically be about 0.005 to 0.010 in. WG (about 1 to 2 Pa), of the magnitude of the 0.015
in. WG rule of thumb considered above. In general, a SD greater than 1,000 ft (about 300
m) is interpreted as very good communication, suggesting that one SSD suction pipe should
easily treat the entire slab. A SD less than 10 ft (3 m) is interpreted as poor communication,
indicating the need for multiple SSD pipes.
The results of these measurements are summarized in Table 2 for those houses having
ASD systems. As shown, almost all houses having SSD systems have sub-slab depressuriza-
tions at all test holes greater than 0.015 in. WG, sometimes by an order of magnitude. In
many of the SSD houses, most or all of the sub-slab readings are above the more conservative
value of 0.04 in. WG. Of the houses with SSD systems having residual radon levels greater
than 2 pCi/L, in only one case -- House 39 -- does the elevated level appear to be due to
inadequate distribution of a suction field under the slab by the system. It is noted that
effective sub-slab depressurizations are generally being maintained even in houses where the
SD is less than 10 ft. This is due to the fact that most of the SSD systems were
conservatively designed with multiple suction pipes (usually between three and seven).
However, even this number of SSD pipes should be insufficient in the poorest-communication
houses, if the SD were in fact an accurate predictor of the distance over which a single pipe
can provide treatment. The SD consistently over-predicts the number of SSD pipes actually
required.
ASD systems other than SSD are less effective at depressurizing the sub-slab. Of the
five houses (Houses 10, 12, 15, 26, and 27) having exterior DTD systems (i.e., drain tiles
outside the footings), three houses have at least one sub-slab reading below 0.015 in. WG.
Understandably, the suction being developed around the exterior of the footings is impeded
in extending into the sub-slab region. However, all three of the houses with at least one
marginal depressurization measurement are below 4 pCi/L, and two are below 2 pCi/L. Thus,
it would not appear that inadequate suction field extension is responsible for elevated residual
levels in the houses with DTD systems. Testing to be described later tends to confirm that
the residual radon in these houses is indeed due to factors other than inadequate sub-slab
depressurization. Exterior DTD systems probably function primarily by diverting soil gas away
from the footings (preventing entry into the block walls), and perhaps by intercepting the gas
before it reaches the immediate sub-slab region; thus, maintenance of high depressurizations
immediately under the entire slab might not be necessary for successful performance.
Sub-slab measurements were permitted in five of the houses (Houses 3,8,14,16, and
20) having BWD systems, or systems with a significant BWD component. All five of these
houses have multiple readings below 0.015 in. WG (although it is noteworthy that the BWD
systems dfi produce some depressurization of the sub-slab). It is likely that the marginal sub-
slab depressurizations in the BWD houses are partly responsible for the elevated residual radon
levels in many of these houses. However, inadequate depressurization of the sub-slab is not
the only problem. Other testing in some of the BWD houses demonstrated that good
depressurization of the sub-slab by an SSD system in those houses was not sufficient, by
itself, to provide the desired radon reductions. Thus, part of the problem with the BWD
systems (and with the SSD systems that were also tested in some of these houses) is that
they were not adequately treating the block walls.
-------
•	¦	Honrocciiri73tion of the sub-slab appears to be largely or partly
In summary, ,nad®^^e , ,s in SSD House 39, and at least partially responsible
"the BWD hou'ses However, i. is no, generally responsible ,or the significant number of sti..-
elevated houses having SSD and DTD systems.
Rfi-gntrainmRnt of ASP Fan Exhaust
~	th0 a<;d exhaust piping indicated radon concentrations ranging from
Measurements in the ASDexl^st^pMgny of th0 SSD systems had
10 to 27,000 pCi/L (0.37 to I.	^ l (3? tQ ?4 kBq/m3} At these levels,
exhaust concentrations excee . g	^	jnto the hQuse cou,d create
re-entrainment of even a Traci,on Ul
indoor concentrations exceeding 4 pCi/L.
„ ^ « thP flow rate and radon concentration of the exhaust, and upon the volume
Based upon the flow rat	^	g ca|cu|ation was made 0f the indoor
and estimated natural vent	one-thousandth) of the exhaust
radon concentration.hat would	re-entrainment would cause an
was re-entrained. The ca	Bo/m3) in nine of the houses, and of more than
incremental increase of more-than1 pC./L (37 Bq/m> ^	^ of ^ 14,
0.5 pCi/L (18 Bq/m 1of TD Measurements exceeding 4 pCi/L, suggesting a possible
cSion beween're-entrainmen, and elevated residual radon levels.
• m thft^ ASD installations have the exhaust fan mounted outside the
house aTgradeTevel, exhausting straight upward immediately beside the house. Thisexhaus,
configuration is conducive to re-entrainment.
-r ~ of testina were conducted to quantify the effects of re-entrainment on
Two types of testing	approach, 9 houses from among the top 14
residual indoor levels,n ^.®se configurations modified, with Pylon measurements in the
were selected to,hav i their e ^ ^	modifications on indoor radon. In the second
approacheVf3ive£ ofthe houses were selected for perfluorocarbon tracer (PFT) gas
measurements.
it« of the exhaust modification testing are summarized in Table 3 For each
The results of the_ exna	^ tested gre ,isted> a,ong W)th resulting
house, the alternative ex ®	measured in the basement and/or living area. Each radon
radon concentrations that	^ measurements with a Pylon continuous
result is the average ot /. to	y . the exhaust modifications: reduced three of
radon monitor. As si°	22 25 and 34); reduced another two below 4 but not
the houses be,ow	f^ed't0 reduce the other four houses below 4 pCW.
rnIOa"leasP, one story (Houses ,0. 13. 20. and 24,.
. .	„raHo oyhausts, directed 90° away from the house, were
From Table 3, h°nz°nt®n "®h eave exhausts in two houses (Houses 20 and 24). In
modified to become vertica -	sianjficant reduction in re-entrainment by converting to
both houses, there	1 h ther house originally having a horizontal exhaust
the above-eave	levels were fairly low to begin with (2.4
W concentrations in the exhaust (8,000 PCi,L.
-------
or 296 kBq/m3). Extension of the exhaust piping 15 ft (about 5 m) away from the house was
required to achieve a significant additional reduction in indoor levels. Thus, horizontal exhaust
at grade might be as acceptable as the above-the-eave method of exhausting ASD systems,
especially when radon concentrations in the exhaust are not very high, as long as the
horizontal exhaust is directed 90° away from the house. However, from the other results in
Table 3, it would never appear appropriate to exhaust horizontally at grade parallel to the
house (or at an angle significantly less than 90°), nor would it ever appear appropriate to
exhaust vertically at grade immediately beside the house.
The actual reductions in indoor radon concentrations achieved by these exhaust
modifications, shown in Table 3, were compared against the calculated increase that 0.1%
re-entrainment should contribute to indoor levels, discussed earlier. This comparison should
suggest the degree of re-entrainment that was eliminated by re-directing the exhaust. In all
cases except House 22, the measured reductions in indoor levels suggested that re-
entrainment was reduced on the order of 0.1 %. In House 22, the reduction was about 2%,
consistent with the high re-entrainment that might have been expected based upon the
original exhaust configuration in this house (horizontal at grade parallel to the house,
underneath an overhung bay window).
In view of the residual radon levels following the modifications to the system exhausts,
it is doubtful that the modifications eliminated all re-entrainment in any of the houses. Rather,
re-entrainment was simply reduced to some lesser value.
In an effort to obtain a more quantitative measure of the actual re-entrainment with the
different exhaust configurations, PFT tracer gas measurements were made in five of these
houses. In each case, one specific PFT gas ("lime") was released into the ASD exhaust
piping. To quantify house ventilation rates, "red" PFT was released into the house upstairs,
and "gold" PFT was released into the basement. PFT detectors were deployed on both levels.
From these results, it should have been possible to quantify the amount of re-entrainment on
both stories of the house.
The results from the PFT testing are summarized in Table 4. Unfortunately, some of
the detectors were lost during shipment to the analytical laboratory, so that results for some
of the exhaust configurations in some of the houses are missing. Table 4 compares basement
radon concentration that would be predicted based upon the PFT results, with the actual
measured concentration for the particular exhaust configuration, from Table 3. As shown,
the PFT-predicted basement levels are always significantly greater that the levels actually
measured, suggesting some problem with the technique by which the tracers were used in
this study, and preventing any meaningful interpretation of the results.
Contribution of Well Water to Airborne Radon
All but five of the study houses in this project are served by private wells. The radon
concentrations in the well water ranges between 530 and 266,000 pCi/L (20 and 9,800
kBq/m3) from house to house. Much of this waterborne radon is released into the indoor air
when water is used in the house.
-------
The widely used rule of thumb — based upon typical water usage rates, house volumes,
and house ventilation rates — is that 10,000 pCi/L (370 kBq/m3) of radon in well water will
contribute approximately 1 pCi/L (37 Bq/m3) to the airborne concentration, on the average
over time. Using this rule of thumb, the well water in these houses could be contributing
between <0.1 and 7.5 pCi/L (<4 and 278 Bq/m3) to the airborne concentrations (excluding
the one house originally having 266,000 pCi/L, which has since been provided with a water
treatment unit). Eleven of these houses could have a water contribution to the air levels
greater than 1 pCi/L.
To confirm the practical accuracy of this rule of thumb, "temporary" granular activated
charcoal (GAC) units were installed to remove the radon from the water in four houses where
the water could be contributing more than 1 pCi/L to the air concentrations. To determine the
effect of water treatment, radon measurements were made in the basement and upstairs
using Pylon monitors, over 2-week periods both immediately before, and immediately after,
the GAC units began treating the water.
The "temporary" GAC units consisted of a standard fiberglass water-softener cylinder
filled with 0.2 ft3 (6 L) of charcoal. These units were being marketed locally for organics
removal; they were not specifically designed for radon removal, and thus could be subject to
a deterioration in radon removal performance over time. However, water radon measurements
indicated that these units were providing high radon removals (94 to 99.6%) for the relatively
short duration of the current study.
The effects of the GAC units on airborne radon concentrations are summarized in Table
5. The table includes not only the current results for the four houses tested here, but also the
results from two permanent GAC units installed and tested in two other houses in 1986,
during the original project.
In four of the six houses in Table 5 (Houses 10, 23, 30, and 34) the ratio of the water
radon to its apparent airborne contribution ranges between 7,900:1 and 12,800:1; i.e., within
about j_ 25% of the 10,000:1 rule of thumb. Thus, this rule of thumb generally appears to
be a rough but reasonable predictor of water effects. The expected role of waterborne radon
in contributing to the residual airborne levels in these houses is thus confirmed. Except
perhaps for House 23, none of these houses could be reduced below 2 pCi/L (74 Bq/m3)
without permanent water treatment.
House 20 is the one house with reliable data where the observed ratio differs from the
10,000:1 rule of thumb by greater than ±_ 25%. In this house, the apparent actual contribu-
tion of waterborne radon (3-1 PCi/L, or 115 Bq/m3) is only about half of the 7 pCi/L (259
Bq/m3) that would have been predicted. It is not clear why this should have been the case.
The owners have small children, and operate the washing machine frequently; thus, lower-
than-usual water usage is not the explanation. The house is somewhat larger than average
(about 2,600 ft2, or 240 m2)/ but n°t sufficiently to explain the significant deviation from the
rule of thumb. A higher-than-average natural ventilation rate of the house would also help
explain the elevated ratio; it's n°t known what the ventilation rate of this house is. A
reduced fraction of radon released from the water upon use in the house would also help
explain this ratio, but there is no reason to expect the release rate from the water to be
unusually low.
-------
The apparent ratio in House 2 would also appear to be dramatically different from the
10,000:1 rule of thumb. However, the results from House 2 are so uncertain, for the reasons
indicated in the table, that these results are not felt to be meaningful.
Contribution of Outdoor Levels to Indoor Radon
In view of the highly elevated soil gas radon concentrations in some locations, it was
considered that higher-than-average ambient (outdoor) radon concentrations could possibly
be contributing to the elevated residual indoor levels.
To assess the extent of this contribution, measurements of outdoor concentrations
were made near seven of the study houses distributed around the study area. Three alpha-
track detectors, shielded by weather-protection cups, were hung from trees near the houses
(but well away from the ASD exhausts). The detectors were deployed in December 1989 and
returned to the laboratory for analysis in February 1990, after 3 months' exposure. The
measured concentrations over this exposure period at the seven sites ranged from 0.0 to 0.8
pCi/L (0 to 30 Bq/m3). Excluding the one site (near Oley, PA) giving the 0.8 pCi/L, the other
six sites averaged 0.2 pCi/L (7 Bq/m3), definitely no higher than the national average.
Accordingly, it would appear that the ambient levels are not contributing unduly to the
indoor concentrations.
Radon Emanation from Building Materials
It was not anticipated that building materials were generally a major contributor to
indoor radon. Gamma measurements in all of the houses had shown indoor readings (5 to 13
/yR/hr, or 13 to 34 x 1010 C/kg air/hr) somewhat lower than the outdoor readings (averaging
between 5 and 20ywR/hr, or between 13 and 52 x 10'10 C/kg/hr). On this basis, it would be
expected that the concrete slabs and foundation walls did not contain unusually elevated
radium concentrations, and should not be contributing an amount of indoor radon significantly
greater than might be expected in other parts of the country.
Typical concretes contain roughly 1 pCi of radium per gram of concrete. This radium
content will commonly result in an emanation of 10 to 40 pCi of radon/hr/ft2 (4 to 16
Bq/hr/m2). Depending upon the house ventilation rate, and whether the basement has poured
concrete foundation walls, this typical emanation could contribute approximately 0.25 pCi/L
(approximately 10 Bq/m3) to indoor levels.
As a more quantitative estimate of the emanation from the concretes of these houses,
a flux test was conducted on the slab and concrete foundation wall of Houses 33 and 34
under the current project. Inverted stainless steel bowls having a volume of 0.2 ft3 (6 L) were
sealed over the slab and wall, and the increase in radon concentration was measured inside
the bowls after 1 hour. For the dimensions of these bowls, an increase of 1 pCi/L/hr (37
Bq/m3/hr) inside the bowl would correspond to a radon emanation rate of 8 pCi/hr/ft2 (3.2
Bq/hr/m2). The changes in radon concentration in the bowl over 1 hour during this testing
were small, in the range of 1 pCi/L, indicating approximate emanation rates of 2.3 pCi/hr/ft2
(1 Bq/hr/m2) from the slab, and 12 pCi/hr/ft2 (5 Bq/hr/m2) from the walls in House 33. In
House 34, emanation from the slab was comparable to House 33, and emanation from the
-------
.¦ U*. u;nhor (Oft r>Pi/hr/ft^ or 12 Bq/hr/m2). Because of the short duration of the
walls was sightly higher (28pO/h /n , <*	uncertainties in these
,est and the small	VTo SX (A4 Bq/hr/m'). However, it is clear that the
emanation rates are g ¦ ^ compared t0 rates from slabs in other parts of the country.
Tboth houses, the emanation rates would suggest that the concrete is contributing less than
0.2 pCi/L (7 Bq/m3) to the indoor concentrations.
In conclusion, it would appear that building materials are not a significant contributor
to the residual indoor radon concentrations in these houses.
Inherent Limi»tir.ns of Certain Mitigation Approaches
in several of the houses not having ASD systems, the failure of the house to have been
reduced below 2 pCi/L (74 Bq/m') is felt to be the result of inherent lim.tat.ons .n the
effectiveness of the selected mitigation approaches.
z u u 11M« HW.nn hlock-wall oressurization systems (Houses 2, 5, and 9)
All three of the h^s	wan ^ ^ ^ ^
have bas^®™ anent'Vproblem of wall pressurization systems in establishing an effective
pressure/flowTfield to prevent soil gas entry into the block cores, or through slab cracks.
T « m thp three houses having HRVs have residual concentrations of greater than 4
Two of the three houses na ^g ^ ^ ^ ^ hrv	{House 2Q) |#	^
9rn °These^esults reflect the fact that ventilation techniques such as HRVs are inherently
limited To achiev^g no greater than moderate ,50 to 75%, radon reduct,ons.
The one house being treated solely with a GAC wen water removal unit (House 30) is
The one	^ s|mp|y rof|ects that, while water treatment can be very
effective at reducing the waterborne source of radon, it cannot address soil-gas-related entry
mechanisms.
CONCLUSIONS
Based upon the testing and assessment
Err,esidua',eve,s orea,er than 2 PCi,L
(74 Bq/m3). These reasons are summarized in Tab e
Pnr qSD and DTD systems, the primary single cause of residual elevated levels is re-
For SSD ana ui u sy	followed in some houses by airborne radon resulting
entrapment of ^,g^dh design 0'f the exhaust, and treatment of the water, would be
from well water. Care in	a	^ one hQuse wjth ^ sgD system did thQ
Zateed'eswtaMeUsecleaUrly appear to be due to inadequate depressurization beneath the
slab.
-------
For BWD systems, inadequate depressurization beneath the slab by the BWD system
is probably the major contributor. Re-entrainment and well-water contributions are probably
also playing some role in some of the houses.
For other than ASD systems, inherent limitations in the systems are commonly the
primary single cause of the elevated residual levels.
Elevated outdoor radon concentrations, and radon emanation from the poured concrete
slabs and foundation walls (where present), do not appear to be significant contributors to the
elevated residual indoor levels. These factors apparently contribute on the order of 0.2 pCi/L
(7 Bq/m3) each to the indoor concentrations.
ACKNOWLEDGEMENTS
The authors wish to express their appreciation to the Pennsylvania Department of
Environmental Resources, for their support throughout the project, and especially to the
homeowners who made their homes available for this testing.
REFERENCES
1.	Scott, A. G., A. Robertson, and W. 0. Findlay, "Installation and Testing of Indoor
Radon Reduction Techniques in 40 Eastern Pennsylvania Houses," EPA-600/8-88-002,
January 1988 (NTIS PB88-156617).
2.	Scott, A. G., and A. Robertson, "Follow-Up Alpha-Track Monitoring in 40 Eastern
Pennsylvania Houses with Indoor Radon Reduction Systems (Winter 1987-88)," EPA-
600/8-88-098, September 1988 (NTIS PB89-110035).
3.	Scott, A. G., and A. Robertson, "Follow-Up Alpha-Track Monitoring in 40 Eastern
Pennsylvania Houses with Indoor Radon Reduction Systems (Winter 1988-89)," EPA-
600/8-89-083, October 1989 (NTIS PB90-134172).
4.	Scott, A. G., and A. Robertson, "Follow-Up Annual Alpha-Track Monitoring in 40
Eastern Pennsylvania Houses with Indoor Radon Reduction Systems (December 1988-
December 1989)," EPA-600/8-90-081, November 1990 (NTIS PB91-127779).
-------
TABLE 1. SUMMARY OF POST-MITIGATION ALPHA-TRACK DETECTOR RESULTS
FROM PENNSYLVANIA STUDY HOUSES
Post-Mitiaation Radon (oCi/L)
House Mitigation
Pre-Mitigation
Winter-Quarter Averaoes4
Annual Average
NO-
System1
Radon (DCi/L)2 3
Basement
Livino Area
(Livino Area)
2
Wall press.
413
4.3
6.9
s
3
BWD + SSD
350
3.3
2.1
1.8
4
SSD
25
1.0
0.9
0.5
5
Wall press.
110
4.8
4.4
4.0
6
SSD
60
3.5
3.6
2.3
7
SSD
402
4.5
3.3
8
8
BWD
183
3.4
1.4
1.1
9
Wall press.
533
11.5
14.8
S
10
DTD
626
11.5
8.4
12.1
12
DTD
11
2.5
2.3
1.3
13
SSD + DTD
64
2.5
2.9
6
14
BWD
36
0.8
1.0
S
15
DTD
18
1.2
1.2
0.9
16
BWD
395
5.3
1.8
1.5
17
HRV
9
8.1
5.1
2.7
18
HRV
12
11.7
3.5
3.6
19
BWD
32
31.3
0.7
s
20
SSD + BWD





+ DTD
210
6.9
9.7
10.0
21
SSD
172
2.3
2.7
3.7
22
SSD
24
9.0
3.8
6
23
SSD
98
2.5
1.6
1.6
24
SSD
66
4.1
4.0
3.2
25
SSD
122
6.8
4.8
6.4
26
DTD
89
1.3
1.4
1.0
27
DTD
21
4.5
2.2
3.9
28
HRV
21
3.6
4.9
3.6
29
DTD + SLD
61
1.9
1.9
3.0
30
Water
17
3.6
1.7
1.9
31
SSD
485
2.3
7.0
s
32
SSD
6
0.9
3.6
4.0
33
SSD
82
5.6
1.0
0.6
34
SSD
470
5.3
4.9
5.8
35
SSD
144
1.4
0.9
0.7
36
SSD
300
1.2
0.8
0.7
37
SSD
87
0.9
1.0
0.9
38
SSD
309
7.8
7.2
6.6
39
SSD
111
7.5
1.8
4.1
40
SSD
148
1.9
1.2
s
Fnntnntes for Table 1
' SSD = sub-slab depressurization; DTD = drain-tile depressurization; BWD = block-
wall depressurization; SLD = sub-liner depressurization (crawl spaces); HRV = heat
recovery ventilator; wall press. = block-wall pressurization.
1 1 pCi/L = 37 Bq/m3
J Pre-mitigation measurements were usually made in the basement by the Pennsylvania
Department of Environmental Resources using ATDs, prior to the mitigation project.
4 Each reported radon value is the average of winter-quarter ATD measurements, usually
for two or three winters.
8 Annual average ATD measurement was not successfully completed in this house,
usually because system was turned off, or was not fully operational, during part of the
measurement period.
-------
TABLE 2. SUB-SLAB DEPRESSURIZATIONS CREATED BY MITIGATION SYSTEMS
(HOUSES WITH ASD SYSTEMS ONLY)
Range of Sub-Slab
Depressurizations
House
Mitigation
No. of
Created by System
Range of SD
(ft)13
No.
Svstem
SSD PiDes
(in. WG)',J
3
BWD + SSD
14
0.004-0.012
1,600 to >30,000
4
SSD
6
0.008-0.234
0.3 to 6
6
SSD
3
0.129-0.194
2 to 45
7
SSD
7
0.093-0.375
90 to > 30,000
8
BWD
04
0.004-0.007
3,900 to >30,000
10
DTD
0s
0.056-0.085
> 30,000
12
DTD
oB
0.014-0.018
8,800 to >30,000
13
SSD + DTD
4
0.109-0.605
3 to >30,000
14
BWD
04
0.006-0.012
110 to >30,000
15
DTD
0B
0.014-0.072
1 to 580
16
BWD
04
0.001-0.006
3,300 to >30,000
19
BWD
04
Owner did not permit measurements.
20
SSD+BWD




+ DTD
54
0.008-0.202
1 to 25
21
SSD
1
0.117-0.169
> 30,000
22
SSD
4
0.322-0.399
170 to 2,200
23
SSD
4
0.669-0.706
45 to >30,000
24
SSD
3
0.847-1.109
75 to 190
25
SSD
4
0.020-0.274
6 to 270
26
DTD
0s
Pos.-0.008
2 to 990
27
DTD
O5
0.056-0.081
>10,000
29
DTD + SLD
0s
0.625-0.685
>30,000
31
SSD
6
0.113-0.738
5 to 380
32
SSD
7
0.282-0.706
2 to 4
33
SSD
1
0.322-0.637
6,100 to >30,000
34
SSD
6
0.685-1.391
1 to 40
35
SSD
4
0.014-0.171
1 to 30
36
SSD
5
0.056-0.181
80 to >30,000
37
SSD
6
0.968-1.012
>30,000
38
SSD
2
0.044-0.258
45 to > 30,000
39
SSD
3
0.001-0.102
0.7 to 2
40
SSD
20
0.001-0.256
1 to 3
Footnotes for Table 2
1	The range of depressurizations and 1 % suction distances (SDs) reflect the range of
results from the different test holes.
2	1 in. WG = 248 Pa
3	1 ft = 0.30 m
4	House has a block-wall depressurization system only, or a SSD system with a major
BWD component; thus, depressurization beneath the slab will be low in comparison
with typical SSD systems.
6 House has a drain-tile depressurization system. In all cases except House 29, the drain
tiles are outside the footings; thus, sub-slab depressurizations will be low in
comparison with typical SSD systems.
-------
TABLE 3. PYLON RESULTS FROM MODIFICATION OF ASD EXHAUST CONFIGURATIONS
Radon in
House Exhaust
No. fpCi/L)
7 3,500
Exhaust Configuration
1.	Vertical at grade, immediately beside
house (original configuration).
2.	Stack extended up to eaves; elbow directs
exhaust horizontally, 90° away from house,
at eave level.
3.	As in 2 above, except stack ends vertically
above eaves.
Average Pylon Result
(pCi/L)
Basement
5.2
4.9
2.1
Living
10 2,300 1. Vertical at grade, immediately beside	9.4	5g
house (original config.) Incl. water treatment.
2. Elbow on fan outlet directs exhaust horizon- 2.1	10 8
tally at grade level, at a 20° angle away from
house (i.e., almost parallel). Water treatment.
13 580 1. DTD fan exhausting vertically at grade (original 7.3
configuration). SSD system off.
2. Elbow on DTD fan outlet directs exhaust	15.6
horizontally at grade level, at 60° angle away
from house, toward corner of house. SSD off.
20 2,200 1. Horizontal at grade, directed 90° away from	4.6	~5-10
house (original config.). Incl. water treatment.
2. Stack extended up outside house, vertical	--	5 2
discharge above eaves. Incl. water treatment.
22 1,550 1. Vertical at grade, immediately beside house	14.5
(original configuration).
2. Elbow on fan outlet directs exhaust horizon-	1.6
tally at grade level, 90° away from house;
hose on horizontal outlet of elbow leads
exhaust 10 ft away from house.
(continued)
-------
TABLE 3 (continued)
Radon in
House Exhaust
No. (pCi/L)
24 2,000
Exhaust Configuration
1.	Horizontal at grade, directed 90° away from
house (original configuration). (Fan reduced.)
2.	Stack extended up outside house, vertical
discharge above eaves. (Fan reduced.)
Average Pylon Result
(pCi/L)
Basement Living
5.4
4.9
25 1,200 1. Horizontal at grade, parallel to house,	4.6
under deck (original configuration).
2. Horizontal at grade, directed 90° away from	0.5
house, with exhaust pipe extending 10 ft
away from house (to end of deck).
27 650 1. Vertical at grade, immediately beside side	6.9
of house (original configuration).
2.	Horizontal at grade, directed 90° away from	2.7
rear of house, with exhaust pipe extending 4 ft
away from rear of house (under deck stairs).
3.	Stack extended up outside of house, vertical 2.4
discharge above eaves.
34 8,000 1. Horizontal at grade, directed 90° away from	2.4	3.4
rear of house by sliding glass door (original
configuration). (Temporary well water treat-
ment system also operating.)
2.	Horizontal at grade; 90° elbow on fan outlet	3.5
directs exhaust parallel to rear of house, with
a 14-ft length of pipe directing the exhaust to
the corner of the house, where it is discharged
parallel to the rear but 90° away from the side
of the house. (Temporary water treatment
system operating.)
3.	As in 2 above, except horizontal exhaust piping 1.4
extended an additional 15 ft, diagonally away
from the corner of the house. (Water treated.)
-------
TABLE 4 PREDICTED INDOOR RADON CONCENTRATIONS BASED UPON PFT RESULTS,
COMPARED WITH MEASURED RADON LEVELS
House
No. Exhaust Configuration1
10 2. Horizontal at grade
22	2. Horizontal at grade
23	Vertical above eaves
24	1. Horizontal at grade
25	1. Horizontal at grade,
parallel to house
Bsmt Radon Expected Basement Radon	Radon
Tracer Release3 Cone, from Re-Entrainment	Measured
- "	(Based Upon PFT Results)4	in Bsmt6
(oCi/L)		(pC'/L)
Ratio2 (pCi/hr)
(x 107) (x 10 7)
0.4 45
1.1
0.9
6.5
1.5
20
32
12
27
18
22
29
78
40
2.1
1.6
0.9
5.4
4.6
34 1. Horizontal at grade, 1.3 39
directed 90° away
2.	Horizontal at grade, 1.9 39
extended to corner
3.	As in 2 above,	1.0 39
extended 1 5 ft
51
74
39
2.4
3.5
1.4
38 Horizontal at grade
1.4
24
34
5.1
Configuration numbers shown here are identified in Table 3.
2	The ratio of (Lime PFT concentration in basement, in PFT units/L):(Lime releasp rato
ASD exhaust, in PFT units/hr).
3	The rate of radon release from the ASD exhaust, in pCi/hr, determined from the exhaust
flow rates and radon concentrations.
4	The predicted basement radon concentration, based upon PFT measurements 1
calculated by multiplying the radon release rate times the PFT tracer ratio, (basement PFT
concentration)/(PFT exhaust rate from ASD system).
5 The measured basement radon concentration listed here is generally the average of the
4-day Pylon measurement made during, or just before, the PFT measurements.
-------
TABLE 5. EFFECT OF WATER TREATMENT UNITS ON AIRBORNE RADON LEVELS
House
No. Story
Water 	Airborne Radon (pCi/L)	
Radon1 Without Water With Water
(pCi/L) Treatment Treatment Reduction
Current Testing
10 Upstairs	26,200
10 Basement3	26,200
20 Basement3	69,900
23 Basement3	11,500
34 Upstairs3	26,800
Prior Testing (Reference 1)4
2 Basement3	53,200
30 Basement3	206,000
7.4
10.1
8.2
1.7
5.4
2.85
29.1
4.1
7.1
5.1
0.8
2.8
2.2
5.2
3.3
3.0
3.1
0.9
2.6
0.6
23.9
Water Radon:
Airborne Reduction2
7,900:1
8,700:1
22,500:1
12,800:1
10,300:1
Questionable5
8,600:1
1 For houses tested under current project, the water concentrations shown here are the
averages of two pre-treatment measurements, made in December 1989 and January
1990. For the houses tested under the original project (Houses 2 and 30), the values
shown are the average of the original 1985-86 analyses and of several analyses made
during the period August 1986 through March 1987, since these were made closer to the
time that the airborne radon measurements were made with the GAC on and off.
2 The ratio of the water radon concentration to the reduction in airborne levels achieved by
operating the GAC system, which should approximately equal the contribution of
waterborne radon to the airborne levels. For comparison against the 10,000:1 rule of
thumb.
3	Washing machine is on this story.
4	The measured effects of the GAC units on airborne radon are thought to be much less
accurate in the prior testing, since the GAC on/off measurements were not made back-to-
back in the earlier testing, and the measurements under "GAC on" and "GAC off"
conditions were shorter than the 7 days used in the current project.
5	Results from House 2 very uncertain because: Pylon measurement with GAC off far too
short (only 20 hours in duration); possible basement ventilation by owner during
measurement period makes results uncertain.
-------
TABLE 6. APPARENT REASONS WHY STUDY HOUSES ARE STILL ABOVE 2 pCi/L
House Mitigation Pre-Mitigation
No. System Radon (pCi/D1
Houses greater than 4 pCi/L
Post-Mitigation
Radon (PCi/L)2
2
Wall press.
413
4.3
5
Wall press.
110
4.8
7
SSD
402
4.5
9
Wall press.
533
11.5
10
DTD
626
11.5
16
BWD
395
5.3
17
HRV
9
8.1
18
HRV
12
11.7
19
BWD
32
31.3
20
SSD + BWD
210
6.9

+ DTD

9.0
22
SSD
24
24
SSD
66
4.1
25
SSD
122
6.8
27
DTD
21
4.5
33
SSD
82
5.6
34
SSD
470
5.3
38
SSD
309
7.8
39
SSD
111
7.5
Houses between 2 and
4 DCi/L

3
BWD + SSD
350
3.3
6
SSD
60
3.5
8
BWD
183
3.4
12
DTD
11
2.5
13
SSD + DTD
64
2.5
21
SSD
172
2.3
23
SSD
98
2.5
28
HRV
21
3.4
30
Water
17
3.6
31
SSD
485
2.3
Reasons for Elevated
Residual Radon
System limitations; water.
System limitations.
Re-entrainment.
System limitations; water.
Re-entrainment; water.
Inadequate sub-slab depressurization.
System limitations.
System limitations.
Inadequate sub-slab depressurization.
Water; perhaps re-entrainment;
marginal sub-slab depress.
Re-entrainment.
Re-entrainment.
Re-entrainment.
Re-entrainment.
Unsealed entry route.
Re-entrainment; water.
Probably re-entrainment; water.
Inadequate sub-slab depressurization.
Inadequate sub-slab depressurization.
Probably re-entrainment; water.
Inadequate sub-slab depressurization.
Marginal sub-slab depressurization;
probably re-entrainment; water.
Re-entrainment.
Probably re-entrainment.
Water; perhaps re-entrainment.
System limitations.
System limitations.
Probably re-entrainment; water.
1	1 pCi/L = 37 Bq/m3
2	Post-mitigation radon level is average of two or three winter-quarter ATD
measurements in the basement.
-------
IV-2
A Measurement and Visual Inspection Critique
to Evaluate the Quality of Sub-Slab Ventilation Systems
by: Richard W. Tucker
Gemini Research, Inc.
Cockeysville, Maryland 21030
Keith S. Fimian
Radonics, Inc.
McLean, Virginia 22102
ABSTRACT
The reliability of radon testing and the effectiveness of radon mitigation
systems are critical areas of concern because of the detrimental health effects
that can result when a home owner may believe that his radon exposure is less
than he is actually experiencing. This paper provides measurement and inspection
criteria that are oriented towards ensuring that an installed radon sub-slab
depressurization system is actually performing properly and is likely to continue
to do so for several years. Particular attention is paid to the typical house
that is experiencing mitigation where the pre-mitigation levels were between four
and eight picocuries. Continuous-based data logging measurements are used to
show the reaction of certain dwellings to particular mitigation work. A visual
inspection list is provided to identify installation deficiencies which would
lead to the possibility of long-term or short-term operational problems which
could result from improper mitigation system installation.
1
-------
OVERVIEW
~	i natailed in dwellings to reduce the levels of
Mitigation system	dwellinqs. Mitigation efforts are undertaken when
harmful radon, Pr°^fnJ 1 exceed those felt to have acceptable health risk. This
levels are detected which exceea tno^ either the Qwner of the dwelling, the
determinatlon of h®a ^ dwelling or in some cases by regulation or legal
potential owner of the	y fiaure of 4 pCi/1 or .02 WL is used as the
determinations. In most mitiaation efforts. The distribution of radon in
level at which to initia	such that, greater than 60% of dwellings
residential dwellings in	or 02 WL contain levels between 4 pCi/1 and
having levels in excess of 4 pci/i or
8 pCi/1 or .02 WL and .04 WL.
nf radon reduction available. Caulking and
There are ™™e*°^ilation are used in the majority of cases. The typical
sealing and sub-s ab ven	fco rform his own caulking and sealing work. In
home owner wl11 ^J^ork has been completed, there is no additional testing
many cases, after thisw	efforts were effective since the levels were less
as the assumption is th	cases where additional testing is performed,
than 8 pci/l to begin with. In the case	^	^ ^ reduction and
the home owner will usu' Y	lB# If the owner decides to proceed with the
possibly an increase in m{4-jaati0n system, sub-slab ventilation is usually
installation of an active9 ^ improperly installed active systems, these
2-S-beBfnst^ by a professional mitigation contractor.
This paper addresseeb^en^ta^leS JnTLn^r
an operational sub_s a	nd long-term protection and does not cause other
to provide both	and	operational and mechanical evaluations.
collateral problems. The roc	fci tests should be performed during the
Other types of operaUonal ar^d dia^ evaluations but are not addressed in
initial dwelling and system installation e	e across fche glab should fee
testing
The EPa does not ^
The EPA does' .how®d®^'ad^ Reduction Techniques for Detached Houses" (Techniques)
documents, entitled R	lication of Radon Reduction Methods" (Methods) dated
dated January 1988 and pp	regulate the installation of radon
August 1988. Because the EPA^doe ^ ^ ^ ,g ^ ^ ^ ^ referring
mitigation systems, the_s g^ ^ technical guidance is "preferred" and
to sPecific,f®atUTrhese preferred and recommended practices are given for the
"recommended. inebe y.	dwellinq.
protection of the occupant of the dwelling
THE RATIONALE FOR POST MITIGATION PERFORMANCE TESTING AND INSPECTION
improper installation of ad£el1ing*from* many Causes, depending upon the
danger to the occupants of the	^ ^ CQmmon enough and serious enough,
nature of the installation, i ingtallation of active radon mitigation systems
that our firm recommends that the ^ezaj. mitigation firms, certified by the EPA
only be performed by professsi	current EPA "recommended" and "preferred"
and in accordance with all or	the homeowner installed his own sub-slab
procedures. Our sample of home	bJ_	In general, it would not be going
depressurization syste™' "	run the homeowner is at more risk after the
4-^,-n far to say that in tnts iuwy
system installation than before.
2
-------
The dangers that can result from the improper installation of a sub-slab
depreBSurization system are several. First is the danger associated with the
radon itself. This particular danger comes in two forms. The first danger comes
from a system failing to perform its primary mission. In this situation, the
radon level is allowed to exceed the intended maximum level due to some system
malfunction or due to the inability of the system to deal with certain dwelling
operating conditions or changes in outside environmental conditions such as rain,
low pressure systems or high wind conditions.
The second danger from radon is even more dangerous than the first in most
cases. Most houses that are mitigated are less than 20 pCi/1 before mitigation.
If the system simply fails to work, the radon levels in the dwelling will
probably only rise to their former level. If caught within a few weeks or
months, this does not represent a serious increase in health threat. If however,
the system fails in such a manner that the potentially huge levels of radon that
typically exist below the slab are introduced into the living areas of the
structure, even short term failures can lead to significant increases in health
risks to the occupants of the dwelling.
In addition to dangers from radon, there is the potential for danger to the
occupants from several other factors. Many of these other potential dangers are
addressed under the local and national code guidelines and regulations. These
are areas such as fire, electrical, and structural installation considerations.
The final area of danger from an incorrectly installed sub-slab
depressurization system, arises from possible alterations in the pressure field
in the houses vis-a-vis the outside pressure and the effect on devices and
systems in the house that are concerned with the handling of combustion input
materials or by-products. In particular there are many potential dangers that can
result when a sub-slab depressurization system also results in an inordinate
reduction of the pressure field within the house, interfering with the ability
of combustion systems to efficiently remove toxic by-products from the dwelling.
VISUAL INSPECTION
EPA technical guidance for radon mitigation contained in Techniques and
Methods lists many different ways to install a sub-slab depressurization system.
However, EPA technical guidance "recommends" a very precise system design using
a very limited number of system features. These EPA "preferred" and
"recommended" system features are less failure-prone and more efficient than EPA
techniques merely described in EPA technical guidance that are not "preferred"
or "recommended." These "preferred" and "recommended" features may not be
required to get the levels in a structure below the desired level, but they do
provide long-term operational benefits. Therefore, EPA "preferred" and
"recommended" techniques should be followed at a minimum to insure the best
possible system based on current technology. The visual inspection of a system
is designed to ensure that a system contains these EPA "preferred" and
"recommended" features.
A set of questions in Appendix I provide assistance in the evaluation of
a sub-slab depressurization system. Appendix I questions answered with a
negative response are intended to identify deficiencies that may exist in a
system visually inspected in light of EPA recommendations. Each of the
categories of questions in Appendix I are discussed in some detail here.
Specific references to EPA documentation are also given.
3
-------
*7 i f Methods S3.VS/ "it is advissbl© to xnstall an dl&rin
Alarm. Section 7.1 of M	Y house occupants "if the fan becomes
on a radon mitigation	blocked or if the system fails in any other
ineffective", if the pipe becoj"®® J1
°or otherwise detected without the use of
way. Radon cannot be seen,	t' A Bystem that does not audibly or visually
sophisticated measurement equ p	nce of a partial or total system failure
aLef normee^Th^^eVTrTments for long term system operation and may pose a
substantial increase in health risks.
The system alarm ^°ul^d	^o^can occur*as \ result of several
dpiS^!ifncSUrebiockage due to
^uXVe^%^re!SorivL«BPevenqwhile the .ppe.,3 to b.
running at its normal rate.
•v ui r-i no should be provided to the alarm to ensure that a
Separate circuit wirg	system does not impair the functionality of
current disruption to the mi g	circuit breaker to the fan circuit could go
SLSSi	.l»o on the _ circuit	r.
Fan. section 7.1 Method, .ays ££IT'oTtl
to weather conditions, c*P^leh°al8) at a fLow rate of 150 to 200 CFM (.071 to
1.0 inches WC (124 to 228 Pj1	» numbers in this EPA guidance have been
.094 CMS)." The nununum f;lj« r.te^ ^ wifchout of the EPA. The current
discussed to a great extent o	^ higher than required under normal
prevalent conBenju^iB^tM generally believed to be the minimum flow required
for°good system performance.
this much power should be specifically designed
Fans capable of generati g	Bathroom or kitchen fans not designed for
for the purpose of radon r®™ed u£je"d These fans are not designed to run
radon remediation must not	experience a significant reduction in
continuously at high speed. They le
capability when operated in this way
^ J	"in all cases, care should be taken to insure
section 7.3 of Metho .s ar(d fans installed." Vibration caused by these
adequate support for all p:^ ificant. If the fan is not properly and securely
powerful ®Y^e™/baration will accelerate the incidence of system leakage.
. 7 3 of Methods says, "all fans should be mounted
Fan Mounting. Bectron . collecting and all horizontal runs of pipes
vertically^^to^prevent^water^^^^^^	point aQ fchat condensed water can drain
back to the soil.
.	an average radon remediation system handles
The EPA estimates	d in an average house. This volume of
approximately two quarts of	P location of the system fan if the fan is
condensation will aiccurrnul	mounted in a low point along a horizontal
mounted horizontally or it tne
run of pipe.
4
-------
EPA also recommends that the fan be outside the negative pressure field of
the house so that radon leakage will not contaminate the house. The negative
pressure field of the house constitutes all interior portions of the house,
including basements, crawl spaces, and garages beneath or adjacent to living
areas of the house. This means, at a minimum, the fan and the pipe on the
positive pressure side of the fan (the portion of pipe between the fan and the
exhaust) should be located in the attic. The safest operation occurs when the
fan is located completely outside of the house.
Section 7.2 of Methods says, "where the pipe penetrates the roof, the fan
should be mounted either in the attic or on the roof." Mounting the fan on the
roof has the advantage of reducing noise and the risk of re-intrainment.
Mounting the fan in the attic has the advantage of protecting the fan from the
effects of weather. The EPA recommendation is based on the fact that the
constant vibration applied by the fan to nearby system elements can result in
structural fatigue and system leaks.
If a fan is located in a garage and develops a substantial leak, not only
could very high levels of radon be pumped into the garage, but the fan could
pressurize the garage to the point where gasoline fumes that accumulate on the
floor of the garage would be forced into the living quarters of the house. This
poses both an explosive risk and a toxic fumes risk.
sump. Section 7.2 of Methods says, "For the sump ventilation to be
effective, the cover must be sealed airtight. This cover can be made of sheet
metal, plywood, or another suitable material. It will usually be convenient to
fabricate the cover in two pieces so it can be fitted around the pipes which
penetrate the sump. The possibility of needing to service the sump pump should
be taken into consideration when designing the sump cover. Caulk and sealants
can be used to insure an airtight fit. The cover should be secured to the floor
with masonry bolts. If water sometimes enters the sump from the top of the Blab
then an airtight seal that allows water to drain must be installed."
Section 7.2 also states, "When the sump is covered, it is recommended that
the existing sump pump be replaced by a submersible pump if such a pump is not
already present. The submersible pump is recommended to avoid problems of
corrosion with the pump motor and/or for ease of sealing the sump."
Section 7.2 continues with, "The ventilation pipe that penetrates the sump
cover must extend up through the house shell to exhaust the soil gas extracted
through the sump. Figure 9 shows two alternative exits for the exhaust pipe.
In one, the pipe penetrates the house shell through the band joist and extends
up outside the house. It is recommended that the exhaust be above the eaves of
the house and away from windows in such an instance. In the other case, the pipe
extends up through the house to the roof and exhausts soil gas above the roof
line."
Pipe. Section 7.3 of Methods says, "piping used to construct ventilation
systems should be made of plastic, such as PVC sewer pipe for durability as well
as for corrosion and leak resistance. Flexible hose such as clothes dryer vent
hose is not recommended because it is easily damaged and not conducive to
draining water that condenses in the line. It will tend to sag under condensed
water creating traps which could result in reduced effectiveness of the
ventilation system." For these reasons, flexible hose is not acceptable.
5
-------
Section 7.3 further states, "In EPA1 a experience, the ventilation ava^A»
usually consists of 4 inch PVC pipes." Also, "The size of the pipe can alS
influence system performance. If the diameter of the pipe is too small the £»
cannot depressurize the soil because of increased pressure drop in the oi
Long runs of pipe or turns and elbows have a similar effect. Since
diameter pipe takes up less space and is more easily hidden, it may be desiraSt
*-n „br small pipe in some instances."
diameuei. — -
to use small pipe in some instances.
EPA provides no further guidance or a specific "recommendation" concerni
the size of the PVC pipe. The interior diameter of the pipe is not critical 9
long as a sufficient pressure drop across the slab is maintained. Small**
diameter PVC pipe may sufficiently reduce radon levels depending* upon Jh*
characteristics of an individual property.
Section 7.2 of Methods advises, "the pipe must be supported with mount*
brackets either on the basement wall or at the floor penetrations. Horizoj.**®
piping runs should be supported by clamps or brackets attached to floor joists i
Vibration of the pipe and normal wear and tear caused bv weatn
conditions, system fans, and general operation will accelerate the incideneK
system leakage if the pipe is not adequately and securely mounted.
Pipe slope. Section 7.2 of Methods says, "horizontal rune of pipe 8ho
be sloped slightly so that condensed water can drain to the ground or te»
outside drain. It is imperative that no low points exist in the line T**n
natural trap exists in the exhaust line condensed water can collect and block ^ *
air flow." Section 7.3 further states, "all horizontal runs of pipe shoulrt w*
sloped toward the sub-slab vent point so that condensed water can drain	^
the soil."	c* *0
«vhaust. Section 7.1 of	it be Extended above t£2
systemffe^^ff^df'^Hs recorded that the exhaust be above the eave*
eaves." Secti°Vawav from windows."
fhp house ana away
of the noua	avhil„(,Hna the soil gas above the eaves of
section 7 3 ^/^trating through the roof from inside the house or
the house include jitlH P outside the house.
extending	exhaust side of the fan is indoors, it
-If any part of the line oni the	release radon in the house if it
should be carefully leak tested be,u the attio, on th. roof, or
i^oVa For this reason
leak . her:ever possible.
outside w	powerful and they operate continuously.
... fanB in these systems are p £ caused by these fans will likely
, 1% exposure to the continuous vibration^ ^ p*nhole aized leak in the
prolonge P nearby joints to even	* f pipe after the fan) will pump
positive pressure 8^®.°*^nto the ^Li^ibfT quarters if the fan is ^.t.1
high concentrations
inside the house	.. exhaust is mounted on the roof or
section 7.1 ^^^eJa^on"should be given to the ^possibility that it
.uay from the hou'e^ con.i^	or by B„ow and ice.
could become covered,
6
-------
Section 7.3 adds that, "vents through the roof should be capped with a rain
guard that does not impede air flow. The possibility that the outlet could be
covered by snow accumulation or drifts should also be considered." Therefore,
the exhaust port should extend high enough above the roof surface to ensure that
snow accumulations that could be expected for the area in which the system is
being installed would not prevent proper system performance.
System insulation. Section 7.3 of Methods says, "in cold climates
insulation might be needed on the exhaust pipe to prevent ice from blocking it."
If the system is equipped with an adequate alarm capable of detecting when air
flow is impeded due to system blockage caused by ice, snow or other conditions,
the alarm would alert the occupants of this fact. Preference should be given to
extending the ventilation pipe up through the interior of the house shell in cold
climates.
If schedule 40 or greater PVC (or equivalent) is used, 5000 degree days is
considered to be a cold climate. Should lass than schedule 40 PVC or equivalent
be used, then 4200 degree days is considered to be a cold climate.
Electrical, Mechanical. Building Code Compliance. Local building codes
must be followed in the installation of any mitigation system. Local electrical
code must be followed to insure that electrical current provided to a system has
been wired in a manner that would prevent electrical shock to persons working or
playing around the system and that no fire hazard is created. Depending on the
location of the of the fan, some localities may require ground fault interruption
circuits be installed. To insure wiring has been installed in accordance with
local electrical code, evidence of inspection by a qualified electrician must be
provided by the radon remediation company. Other mechanical considerations
include insuring that fire wall penetrations are protected with fire dampers.
These types of requirements depend heavily on the local code requirements. The
inspection process should ensure that the necessary inspections have been
performed.
OPERATIONAL TEST AND EVALUATION
Once a system has been determined to meet the aforementioned visual
inspection requirements, an actual measurement of the radon levels in the
dwelling should follow. These measurements are currently being made in several
ways. Two preferred methods for this measurement are given here however.
The first preferred method is performed with a combination of a short term
passive test and a long term passive test. A short term test is conducted
shortly after the completion of the mitigation work, with enough time allowed for
the house to stabilize with respect to the new conditions. A waiting period of
about 24 hours is recommended. The short term test should be conducted in
accordance with the requirements of the device being used. It should be
remembered that in a post-mitigation environment where sub-slab depressurization
was performed, the levels should normally be in the range of 0.5 to 2.0 pCi/1.
The length of test should be sufficient for the device being used to have
reasonable accuracy at those levels. In any case a minimum two day test should
be performed. Three days is recommended. If the short term test indicates that
the radon levels have been sufficiently reduced, then a one year test should be
performed. This approach does not guarantee that radon levels may not at some
points be very high, but it does indicate the long term exposure.
7
-------
The second preferred method is performed by making a short term test with
a continuous logging active monitor. A device with good resolution over th
range of 0.2 pCi/L to 10 pCi/L must be used. A measurement period of two or
three days should be used. The data provided by this method will yield not onlv
an average level for the test, but can show the performance of the system as
living conditions and barometric pressure vary.
Our data for post mitigation tests shows that radon levels in dwellina
with adequate sub-slab depressurization, do not vary significantly with chanoes
in barometric pressure, rainfall or living patterns. When the pressure
differential between the area above the slab and the area below the slab i
maintained so that the pressure below the slab is sufficiently less than thS
pressure above the slab, radon levels are consistently abated.	6
The use of a single short term passive test in a post mitigation
environment is not recommended. It provides no information about what kind of
variations are occurring and may also provide a poor indication of the lono term
performance.
Figure 1 illustrates a sequence of three tests made with continuous logoincr
equipment. The top plot (Figure 1A) shows the initial screening test. The radon
level averaged over the entire test period was 0.0501 WL. The maximum variance
in the radon levels was about two to one. The homeowner next attempted t
mitigate the house himself by the use of caulking and sealing. As is tyoica?
after homeowner caulking and sealing, the new average radon level was within
few percent of the original reading. The middle plot (Figure IB) shows the teat
made after the caulking and sealing which yielded an average level of ,0467 WL
Again the radon variations are about two to one. After homeowner caulkinq and
sealing, the levels are higher than before the mitigation as often as they a
lower. When caulking and sealing is done by professional mitigators the results
may be a little better, but usually not markedly so.
The bottom plot (Figure 1C) shows the results after a sub-Blah
depressurization system was installed by a professional mitigator. The averaa
radon level was .0050 WL. At no time did the level exceed .01 WL. This ayste
exhibited fairly good performance, although the variance of almost three to on
would be a concern if the maximum levels were higher.
The next example shows a dwelling with a great amount of radon variance
On the initial test (Figure 2A), the average radon level was .0214 WL. A aub-
slab mitigation system was installed and an additional test performed The
second test (Figure 2B) showed great variance in the radon levels and yielded a
average of .0300 WL. The system was tuned by the contractor and again wa2
tested. The levels now rose to .0840 WL with peaks to .1760 WL. Additional work
was performed. The average radon level got back to the pre-mitigation level of
.0214 WL. The maximum level of .0417 WL, however indicates that the system i
far from performing adequately. At this point, the frustrated contractor put i
an air to air heat exchanger. The final test yielded an average radon level of
.0061 WL. Again there was an excessive amount of radon variation, but the )<»vai
were consistently below .01 WL.	a
The work described in this paper was not funded by the U.S. Environmental
Protection Agency and therefore the contents do not necessarily reflect the views
of the Agency and no official endorsement should be inferred.
8
-------
01B302C
13 Old Woods RoAd
Brookf i«td, CT 06AM
0.077QI
0.07411
0.0704 I
0.06A7
W 0.0630,
0	0.0593!
R 0.0556
< 0.0519j
: 0.048?
N 0.0445,
c 0.0408
0.0370]
t 0.0333
E 0.0296
V 0.0259
1	0.0222
I 0.0165'
s o.ouaj r
0.0111;
o.oo74!rvxv-.{Ar
0.0037
10:40
7/03
22:40 2:40 6:40
7/03 7/04 7/04
TIME Of TCSI
10:40
7/04
U:40 15:40 22:40
7/04 7/04 7/04
Tear of 1990
2:40
7/05
6:40
7/05
10:40
7/05
U:40
7/05
' NOTE: The darker area on the graph it above 0.02 Wl (4.0 pCi/l).
021101F
13 Old Woods ftoad
Brookfield, CT 06804
NOTE: The darker area on ihe graph ts above 0.02 VI (4.0 pCi/l)
021905A
13 Old Wood* load
Brookfield, CT 06804
11:00 15:00 19:00 23:00
8/12 8/12 8/12 8/12
Tear of 1990
Figure 1. Mitigation Case 1
9
-------
9110G1C
25 Easiwood Cn.e
Ha.JUoc, CI 0W.43
14:40 18:40 22:40
5/03 5/03 5/03
Tear of 1989
0.02 Wl (4.0 pCi/c)
MOTE: 1h« darker oreo
VliJOlf
25 futwood 0riv«
Hadiion, CT 06*41
0.0737
0.0702
0.0667
0.0632
W 0.0597
0	0.0562
II O.OS26
K 0.0491
1	0.0456
it 0.0421
C 0.0586
0.0551
l 0.0316
£ 0.0281
V 0.C246
£ 0.0211
I 0.0175
S 0.0140
0.0105	I
O.00JS ^,.y ™
(4.0 pCi/L>.

9154011
25 Eaitwood Circle
Nad(»on, cr 06443
0.1760
0.1676
0.1592
0.1509
W 0.H25
0	0.1341
B 0.1257
t 0.1173
1	0,1090
U 0.1006
C 0.C92?
o.caia
L 0.C754
E 0.C670
V O.C5A7
I 0.0503
L O.Oi )9
S O.ayj
0:30 4:30 8:30 12;30 16;30
6/19 6/19 6/19 6/19 6/19
graph it above 0.02 Wl (4.0 pCI/L).
Figure 2. Mitigation Case 2
10
-------
9173001
25 Eastwood Drive
Kadi son, CT 06443
0.0417
0.0397
0.0377
0.0357
u 0.0336
0 0.0318
ft 0.0298
K 0.0278
I 0.0258
N 0.0238
5 0.0218
0.0199
0.0179

0.0155
V 0.0139
j*;' ¦•••	*h '.T«,
l 0.0119
L 0.0099
S 0.0079
0.0060
0.0040
i*
0.0020
7:50 11:50 15:50
6/25 6/25 6/25
TIME OF TEST
23;50 3:50 7:50
6/25 6/26 6/26
Year of 1989
The darker area on the iraph U above 0.02 Wl (4.0 pCi/l).
0.0099
0.0094
0.0090
0.008S
U 0.0080
0	0.0075
R 0.0071
X 0.0066
1	0.0061
N 0.0057
C 0.0052
0.0047
l 0.0042
E 0.0038
V 0.0033
E 0.0028
L 0.0024
S 0.0019
0.0014
0.0009
0.0005
9)930)1
25 Eastwood Drive
Haditon, CT 06443

Time 18:40 22:40
Oate 7/17 7/17
2:40
7/18
6:40 10:40 14:40
7/18 7/18 7/18
TIME Of TEST
18:40
7/18
22:40 2:40 6:40
7/18 7/19 7/19
Year of 1989
10:40
7/19
14:40
7/19
18:40
7/19
22:40
7/19
7/20
Figure 2 (continued) Mitigation Case 2
11
-------
APPENDIX I
Alarm.
Fan.
(1)	Does the sub-slab depressurization system have an alarm?
(2)	Is the alarm triggered by reduced air flow and/or differential
pressure?
(3)	Is the system alarm wired to a separate electrical circuit or
backed up by a battery, should the mitigation system's electrical
circuit fail?
(4) Is the system fan capable of sustaining a pressure differential of
at least .5 inches WC at a flow rate of greater than or equal to
60 CFM (standard 4 inch mitigation fan or larger?)
Fan mounting.
(5)	Is the system fan mounted vertically?
(6)	Is the fan properly mounted and adequately supported?
(7)	Is the system fan mounted outside the negative pressure field of
the house?
Note: Fans located inside the house, garage, or crawlspace are inside
the negative pressure field of the house.
If the fan is located in the attic answer questions 8, 9, and 10.
(8)	Does attic have external air vents?
(9)	Is the attic free from a permanent stairwell (not including a
pull-down stairwell) to living areas below?
(10)	Is the attic free from a chase that enters the attic from the
living areas below?
Sumg.
If the house contains a sump, answer questions 11-14.
(11)	Is the sump capped?
(12)	Is the sump capped with a plastic, metal or wood cover?
(13)	Is the sump cover caulked and sealed to the floor?
(14)	Is the sump cover secured to floor with masonry bolts?
If the sump contains a pump or if a pump was present prior to mitigation
answer questions 15 and 16.
(15)	Does the sump contain a submersible pump?
(16)	Does the sump discharge line contain a reverse flow valve?
If the sump was used as a floor drain prior to mitigation, answer question 17
(17)	Does the sump cover contain an air tight water drain that allows
water accumulating on the basement floor to drain into the sump?
If the floor drain drains to sump, answer question 18.
(18) Is the floor drain trapped at the drain or at the point where the
drain line enters the sump?
12
-------
Sump/Drain Tile Suction System.
If any floor drains, window well drains, gutter down-spouts, etc. connect to
the drain tile, answer question 19.
(19)	Have these connections to the drain tile system been properly
trapped to prevent exterior air from entering the system?
If an exterior drain tile suction system is used, answer question 20.
(20)	Is there a check valve or trap in the piping between the fan and
the drain tile?
If a trap exists, answer question 21.
(21)	Does the trap design permit the owner to check water level in the
trap and add water?
Does the system use PVC pipe or a durable equivalent?
Is the system free of dryer vent hose or flexible pipe?
Are horizontal runs in pipe sloped toward the sub-slab vent point
so that condensed water drains to the ground?
Is the pipe free of low points in horizontal pipe runs that can
collect condensed water and block air flow?
S^stem_exhaust.
(26)	Does the sub-slab depressurization exhaust extend above the eaves
of the house?
(27)	Is the exhaust port at least six feet from the structure if vented
through garage roof or other lower level roof?
(28)	Is the exhaust port at least 8" above the roof line so that it can
not be blocked by snow?
(29)	Is the exhaust pipe capped with a rain guard and or covered by a
protective screen?
If the exhaust port exits near dormers or skylights, answer question 30.
(30)	Is the system exhaust port at least 10 feet from windows and
skylights.
Pipe.
(22)
(23)
Pioe_slo]2e-
(24)
(25)
System insulation.
If schedule 40 or greater PVC (or equivalent) is used in a climate with
greater than 5000 degree heating days, answer questions 31 and 32.
(31)	Are all interior exhaust pipe runs (i.e. pipe runs in unheated
crawlspaces or attics) that are located in untempered space
insulated?
(32)	Are all exterior exhaust pipe runs insulated?
If less than schedule 40 PVC (or equivalent) is used in a climate with greater
th«in 4200 degree heating days, answer questions 33 and 34.
(33)	Are all interior exhaust pipe runs (i.e. pipe runs in unheated
crawlspaces or attics) that are located in untempered space
insulated?
(34)	Are all exterior exhaust pipe runs insulated?
13
-------
Electrical. Mechanical. Building Code Compliance.
(35) Has the installation of the system passed local code requirements
by a county/city inspector and proof thereof been produced?
14
-------
IV-3
TITLE: Correlation of Diagnostic Data to Mitigation System Design and
Performance as Related to Soil Pressure Manipulation
AUTHOR: Ronald f. Simon, R.F. Simon Company
This paper was not received in time to be included in the
preprints so only the abstract has been included. Please check
your registration packet for a complete copy of the paper.
The U. S. Environmental Protection Agency has identified the Montclair, Glen Ridge, West
Orange, New Jersey site as being contaminated with radium tailings. In order to control indoor
radon concentrations as a product of these tailings, the Environmental Protection Agency has
initiated a state-of-the-art diagnostic and mitigation program. Through this program extensive
diagnostics have been performed, directed toward mitigation strategies that incorporate soil
pressure manipulation.
Typically, soil depressurization in permeable soils can readily be accomplished. However, in a
tight soil configuration, soil pressure manipulation becomes more difficult due to the inherent
properties of the soil.
The information collected during diagnostics is compared to installed system performance to aid
in effective system design and installation . The information presented in this paper will allow
the diagnostician and installer to interpret and apply site specific diagnostic data directly to the
design and installation of soil pressure manipulation systems. This data represents a missing link
that will allow systems to be sized appropriately and reduce the number of system failures.
-------
IV-4
TITLE: Pressure Field Extension Using A Pressure Washer
author: William P. Brodhead, WPB Enterprises
This paper was not received in time to be included in the
preprints so only the abstract has been included. Please check
your registration packet for a complete copy of the paper.
abstract
Radon remediation is typically done with sub-slab ventilation
systems. Sub-slab ventilation installation failures are often
due to an incomplete pressure field extension that allows radon to
continue to enter the building. Over half the homes we mitigate do
not have a good gravel base under the slab. This project
investigated a technique for extending the pressure field in tight
soils from a single suction point by the creation of sub-floor
tunnels using commonly available high pressure washers. Houses with
the appropriate tight non-rocky soil were tested for radon and
pressure field extension before and after tunneling with the high
pressure washer. Different equipment and techniques were used and
described.
The tunneling under the slab was an effective method for
extending the pressure field. This technique holds good promise for
mitigators dealing with tight soils and restrictions with where thev
can locate suction holes.	^
-------
IV- 5
A VARIABLE AND DISCONTINUOUS BUBSLAB VENTILATION SYSTEM
AND ITS IMPACT ON Rn MITIGATION
WILLY V. ABEELE.
NEW MEXICO ENVIRONMENTAL IMPROVEMENT DIVISION
1190 St FRANCIS DRIVE
SANTA FE, NM 87 503
Abstract- A house, with a high specific area in contact with earth
materials, was chosen as the site for a long-term Rn mitigation
study. Close to 30 000 Rn readings were collected and intensive use
of statistics was made to determine locations/ time periods and
•xternal parameters promoting high Rn activity. Several Rn
mitigation methods were studied such as passive subslab
ventilation, active subslab pressurization and active continuous
and discontinuous subslab depressurization. Varying degrees of
subslab depressurization were also combined with discontinuous fan
activation to determine the most cost-effective method of Rn
mitigation. Recommendations are for a -50 Pa subslab
depressurization either on a full-time or a part-time basis. The
most cost-effective method used for Rn mitigation was sealing of a
slab opening. The minimum Rn concentrations were obtained, whether
the opening was sealed or not, upon activation of the subslab
depressurization system. The influence of cold weather and
^subsequent increased stack effect is clearly reflected in higher Rn
concentration readings.
INTRODUCTION
Lack of adequate ventilation in a house may allow Rn and its
decay products to reach levels well above the average outdoor
levels. The potential primary sources of Rn in the house under
study are the adjacent earth materials and existing building
materials. The significant sources of Rn, as well as its primary
pathways, will be examined, as will be the influence of subslab
ventilation on indoor Rn concentration. Passive subslab
ventilation, active subslab pressurization and depressurization
will be examined as potential remediation. A simple analysis will
be used to determine the potential role of each source. This will
be derived from the efficiency with which a particular remedial
system is controlling the Rn level in the house. Whereas in a
typical house (including garage), an average volume of 500 m3 has
a 200 m2 surface in contact with earth materials, with a subsequent
specific contact surface of 0.4 m'1, the house under study
(including garage), has a volume of 500 m3, with 300 m2 in contact
with earth materials and a subsequent specific contact surface of
0.6 m"1. This high proportion of surface in contact with earth
materials (50% higher than average) is due to the house under study
being built into the side of a hill. The house design is one of
-------
n K n„or fnotina which eliminates the vertical floor/wall
slab over *°°ti 9	noteworthy that the building has all its
transition 30int.	south with the exception facing east. The
Sj!?SKg>S to be considered'very tight. wi?h little or no cross-
ventilation.
The purpose of this study is to examine the close to 30 000 Rn
The purpos	,, ted over a time period of two years ending
1990 Readings taken during the summer period when windows
in June 1990. Kea y r-|Qck were recorded but not incorporated
i£rthestudy" She study ran from September 1988 to June 1989 and
from September 1989 to June 1990.
RADON emanation and exhalation
To study radon emanation from soils and its exhalation fro*n
To stuay	correct assessment of parent material and
building	a correct	materials is necessary. The
long-lived prog Y P. built Qf concrete. Polystyrene forms were
h°Ur rn shape tSe walls These forms were then filled with
USnrrete The polystyrene remained subsequently in place and served
as an internal and external insulation layer. Approximately 120f
11 concrete™ Ts^or^
t^i^hln but does not include patios, walkways detached walls
the kitcnen d	g was taken every estimated 10 m3. The twelve
etc. . A cone	,p , during construction, in Marinelli beakers
^ aesfeesds'mednt. with n-12. ^Ra averaged 37.3 Bq kg-
J?th a standard deviation of 8.1 Bq kg ', whlie	averted 41.5
nrr lfti'1 with a standard deviation of 5.5 Bq kg	, ™ias
at 80 Bq kg"1 with a standard deviation of 24.9 Bq kg1.
Seated to *|n lut radiologically significant, »u averaged 1.7
Bq kg"1 While 212Pb had a mean of 28 Bq Kg .
The highest calculated transmission fraction was for 2"pb, a
T nigi	nhoton enerqy of 351.9 keV. Fifty mm of concrete
R" ^Tavye' a ^rans^ssron factor of 0.532 for «*Pb, while ««,
would have one o£ 0.4 34 at 186 keV through 50 mm of concrete.
^ -t c=,Yn^-i<*c= taken around the house foundation revealed a 226Ra
Soil samples taken aroun	23fl ^ averaged 29 Bq kg-i and
concentration	^ 234u mean was 26 Bq kg-i and 230Th
averagtH Concentration'of 34 Bq kg^. The unrelated radioisotope
S ignificance, 2i2Th, measured 36 Bq kg .
The slightly higher 226Ra activity found in concrete does not
aVa for the much smaller emanation coefficient or escape to
on ratio of 222Rn found in concrete. It is unlikely that
concrete will be £°und to be a major source of
The assumption was made, subject to a revision based on
observation, that the most important Rn entry process i. the
2
-------
pressure driven flow of Rn through the substructural system
(soil+slab). This is normally orders of magnitude higher than Rn
entry rates from building materials, water and outdoor air. Entry
rates by diffusion directly through the masonry substructure is
even less (1). The Rn entry rate due to pressure-driven flow is
primarily a function of a pressure differential driving this flow
(2), soil Rn activity and substructural (soil+slab) permeability.
Pressure differentials that activate Rn entry are, among others
the easily identifiable ones triggered by temperature differentials
and combustion devices that draw indoor air needed for the
combustion process.
INSTRUMENTS AND DESIGN
The 222Rn activity was measured using charcoal canisters and
the Working Level Reader (WLR) in conjunction with several Workinq
Level Meters (WLM) from Eberline1 for continuous sampling. The WLMs
are really measuring the equilibrium equivalent concentration of Rn
(EER), which is that activity concentration of Rn in radioactive
equilibrium with its short—lived daughters which has the same
potsntidl alpha—energy concentration as the actual non—equilibrium
mixture (3) . This will be reported in this paper more simply as Rn
activity.
The WLM provides the function of sample collection and data
storage. These data points are stored in memory until retrieved by
the WLR. The WLM microcomputer turns the pump on at the preset
starting time, and the activity on the filter paper is counted for
the total time period specified. Calibration of the WLM at the
Technical Measurement Center, Grand Junction, Colorado , showed the
instruments to be highly precise. Occuring inaccuracies were
corrected through calibration. All the WLM readings were on the low
side and had to be corrected by factors varying from 1.437 to
2.031. On the other hand, the repeatability of the measurements no
matter how originally inaccurate, yielded an average coefficient of
variation of 3.19%, which is a measure of the precision of the
instrument.
Subslab ventilation consisted of a network of perforated pipe
installed horizontally underneath the existing slab. Such a
comprehensive system is likely to provide a better performance than
when a vertical pipe perforates the slab and a relatively strona
pressure gradient with limited pressure field is induced.
In any case, good subslab communication is required. The
subslab material consists of a 0.1 m layer of gravel with assumed
high permeability. Ventilation could be passive or active. Active
ventilation could result in subslab pressurization or
1 Eberline Instrument Corp, Airport Rd, Santa Fe, NM
3
-------
The values in parentheses represent actual measurements,
calculated averages, a derivation from measurements (such as the
soil gas concentration derived from soil 226Ra analyses) or a best
guess (such as the ventilation rate) . Accordingly, a soil would
have to have a permeability of 4.25*10"10 m2 to sustain a Rn entry
rate of 0.0142 Bq m"3 s"1, which in turn would lead to an indoor Rn
concentration of 142.8 Bq m3, which is the average Rn
concentration measured inside the bedroom in 1988—89 A
permeability value of 4.25*lO"10 m2, although high, is indeed
acceptable. Average soil permeabilities range anywhere from 10*16
to 10 m but may be impermeable to the point of reaching values
of 10 m ¦ which are ideal for waste containment and are indeed
the permeabilities evaluated to exist at the Waste Isolation Pilot
Plant in the salado formation in Carlsbad, New Mexico (4) .
RESULTS (1988-1989)
When averaged, the Rn activity peak was found to be located at
around 23.2 hrs (11.2 P.M.), while the minimum activity seemed to
be centered around 10.8 hours (with standard deviations of 4.10
hours and 3.87 hours respectively). This seemed to correspond well
with the computed timings of maximum and minimum depressurization.
Maximum depressurization and consequent peak Rn activity seem to
occur earlier than in the average home (5) . This could be occuring
because the heat is not controlled by thermostat and the house is
mostly responding to solar heating patterns. The fact that the
temperature is solely controlled by a solar heat sink could be at
the origin of a maximum indoor-outdoor temperature difference
occuring earlier than in a thermostat controlled home because of an
early drop in indoor temperature and, consequently, earlier maximum
temperature differential and the maximum depressurization that
inevitably follows.
A computed depressurization of 1.87 Pa corresponded with a Rn
activity of 142.7 Bq m' . The regression analysis of Rn activity on
depressurization was run on the computed corresponding dailv means
The correlation coefficient between depressurization and Rn
activity is 0.32, which with 196 degrees of freedom (d.f ) is still
highly significant (at the 1% level).
The correlation is significant at the 1% level because of the
high degrees of freedom. The remarkable aspect of this regression
analysis is that high depressurization was always associated with
high Rn activity, although the reverse was not necessarily true
High Rn activities were also noticed at low depressurizations#
This would lead to the obvious conclusion that other factors
besides thermally induced depressurization play a role in causing
high Rn emanation rates into the house. Two of the factors wind
velocity and direction (6,7), and soil moisture (8), known to
influence indoor Rn emanation, were not studied because of lack of
equipment, although soil temperature was monitored. Because of the
5
-------
depressurization and is produced by an on-line centrifugal fan well
suited to conditions of moderate static pressures. The fan in use
is a 90 Watt T-2 centrifugal fan from Kanalflakt1 with a flow rate
of 0.1275 m3 s"1.
STACK EFFECT
Temperature differentials produce pressure differentials
across vertical walls. This pressure differential is directly
proportional to the height of the walls. Making a few assumptions
about temperature uniformity, the expression:
dp = (r*g*z*dT)/(T,-+273)
reflects the pressure difference at any distance z from the neutral
pressure plane, with dT the temperature difference and T- the
indoor temperature. The soil gas density (in kg m"3) is expressed
by r while g is the acceleration due to gravity (in m s"2) . This
expression can be simplified, after filling average values in for
r and g, to reveal an average depressurization of 0.04 Pa °C"1 nf1
In the house under study, this amounted to an averaae
depressurization of 0.1 Pa °C~ . The effect of soil temperature was
considered separately.
The Rn entry rate in the bedroom, whose floor averages a decth
of 2 m below the soil surface, is expressed by the eguation (1) *
E =( (C*L*dp)/ (V*P) ) * (G/ (12W3) +AC0SH ( (2Z) /W) /(PI*K) ) "1
(in Bq m"3 s*1)
where	V = volume of house (500 m3)
C = soil gas concentration (40000 Bq m'3)
L = crack length (10 m)
G = slab thickness (0.15 m)
dp = pressure differential (1.873 Pa)
P = soil gas viscosity (1.7*10"5 Pa s)
W = floor crack width (0.002 m)
K = soil permeability (4.25*10"10 m2)
Z = floor depth below soil surface (2 m)
E = Rn entry rate (Bq m"3 s"1)
The steady state mass balance equation for the correspondin
indoor Rn concentration can consequently be calculated.	^
Rn = (E+ (N-E/C) *Rn0) / (N+d)	(in Bq m-3)
where N = ventilation rate (10"4 s'1)
Rn0 = outdoor Rn concentration (4 Bq m"3)
d = decay constant of Rn (2.1*10"6 s"1)
1 Kanalflakt, 1121 Lewis Ave., Sarasota, Fl.
4
-------
particular microclimate of a hillside topography, wind velocities
and directions could not be assumed to be related to the ones
measured at the airport, located on a plain on the other side of
town. Acquisition of an anemometer and wind vane was not considered
because of cost and dubious results originating from the warped
topography. Cost was also a factor in not measuring soil moisture
although studies show that the emanation coefficient is stronqlv
influenced by it. (9) show that the emanation coefficient increases
nearly four times as the moisture content by mass increases from
0.2 to 5.7% to drop drastically as the soil becomes saturated.
Analysis of the Rn activity in the bedroom shows that in 64%
of the cases, the nighttime average is significantly higher than
the daytime readings, while in 28% of the cases daytime averages
are significantly higher. In 8% of the instances there is no
significant difference between daytime and nighttime averages, it
is also noteworthy that in 46% of the instances, nighttime averaqes
exceeded 150 Bq m"3, while the 200 Bq m"3 level was exceeded 22% of
the time, the 300 Bq m'3 level 6% of the time and the 400 Bq m"3
level was exceeded only once. A t-test of daytime vs nighttime
means show a p-value of 0.002 6 which demonstrates a verv
significant difference between those two averages. The maximum
hourly average ever recorded was 1.33*103 Bq m"3.
A woodstove was ignited on 16 nights during the study period
Measurements show that Rn activity was 235 Bq m'3 or 164.4% of
average during that period, which seems to indicate* that
woodstoves, or low outside temperatures, or cloudy days accompanied
by snow on the ground (thereby additionally capping the soil and
decreasing Rn exhalation) may be linked to an increased pressure
differential.
Simultaneous depressurizations and Rn activity levels were
measured or calculated simultaneously for the bedroom and the
garage. Despite the fact that the garage floor was crisscrossed bv
shrinkage cracks, the Rn activity measured consistently lower in
the garage. This could be due to a lower depressurization in the
garage. If simultaneous depressurization and Rn activity readinas
were taken in the bedroom and the garage, time related uncertainty
elements would be eliminated. In this case, the coefficient of
correlation rose to 0.87 with 24 degrees of freedom (instead of
0.3 2 with 196 degrees of freedom where time and fluctuations
thereof were a factor).
When the indoor Rn levels in the bedroom were compared to the
indoor levels in the bathroom, a remarkable similarity emerged
Although the bathroom levels were consistently higher, the periods
of maximum Rn activity in the bathroom and the bedroom show a hiah
degree of concurrence with the maximum centered around 23.2 hours
and r = 0.98, while the minimum centered around 10.8 hours and
= 0.95. The readings in both rooms are in almost nerfpnf
synchronization.
6
-------
The Rn daughter activity, as measured with the WLM (W) , is
related to the Rn activity, measured with the charcoal canister
(C), by the equation W = -65.8622 + 0.8439*C, with r = 0.57 and 54
degrees of freedom.
It is important to remember that even if readings were
gathered every hour, all the above statistical analyses are based
on computed daily averages. All the above experiments were
performed with the venting system blocked off and inoperative. The
subslab venting system was put in operation shortly before the
annual deadline dictated by the arrival of summer (which meant a
radical increase in room ventilation and subsequent Rn removal
other than by quantitatively controlled means such as subslab
ventilation controlled by a regulated fan).
The Rn activities, now measured by the hour because of the
short study period remaining in 1988-1989, show a drastic drop when
either convectional venting or active subslab pressurization was
applied.
Table 1 shows the Rn activity in the bedroom before the system
was in operation (I) , when the system was convectionally venting or
passive (P) , and when the subslab was actively and continuously
pressurized (A).
It is important to notice, that the data for P and A are
statistically much less significant than the data for I because
they cover a much shorter period of time (hours instead of days for
I) . It is also important to note the drastic drop in the standard
deviation or the coefficient of variation (c.v.) when the system is
activated.
When the subslab is pressurized, the trend of maximum and
minimum activity seems to be curbed. This is reflected in the
smaller standard deviation of the readings. House depressurization
does not seem to influence Rn entry noticeably because of the
overwhelming effect of subslab pressurization.
Four rooms were regularly checked and their Rn activity could
be ranked as follows by decreasing order of activity: bathroom,
bedroom, living room and kitchen. The fact that the remedial system
equalizes the indoor Rn activity points the finger at the soil as
the main source of Rn since the subslab pressurization only
inhibits the soil gases entry but does nothing to prevent the Rn
emanation from tap water and could only activate the emanation of
Rn trapped in the slab. The subslab pressurization system affects
Rn inhibition equally strongly in both bathroom and bedroom
pointing again at the soil as the main source (water, available in
the bathroom but not in the bedroom, does not seem to be a main
source of Rn) .
It is important to note that active subslab ventilation seems
7
-------
more effective in reducing Rn activity in the house than room
ventilation.
First, the effectiveness of the
Two re^a	onstrated by the appearance of an ice plume
passive system w	can be explained by the fact that even a
at the vent outlet. This	Yto 100%. As the soil
dry soil has a	their saturation point is reached as the
gases escape, ini w	^ temperature is low enough, the condensate
temperature drops.	f f the escapeI Second, it is believed
freezes to	effective, the system must
that for subsiab pres umau ^ ^ ^ subslab gas> The same
create	£ d when SUbslab depressurization takes place.
£ltl Z "S1 authors believe subslab pressurization to be less
effective than depressurization (10).
RESULTS (1989-1990)
NO SUBSLAB VENTILATION
x.v. •	the dav-to-day correlation between Rn
D^ri^on andP hous^ depresLrization due to temperature
concentration an	than during the previous season and was
differential "^/^^^"si^ifiLnt. Only on a long-term basis
consequently foun	d> The Rad0n concentration increased
could a trend d	3 through January 12, as the average
steadily from Septem	decrease. Figure 1 seems to indicate a
temperature continue	average ambient temperature, and
strong	ebssurilation, Ind indoor Rn activity. More
consequent room d p ivit seems equally closely related to the
importantly, the Rn	„lays a pivotal role in influencing the
soil temperature whic*plays a^p ^ ^ buiit infco the side Qf
depressurization prtiai pressure is consequently for a good
a hill and	t-pmoerature (The soil temperature underwent
part governed by	P^ period, which meant increased stack
a steady drop dur g	d house depressurization followed by
effect and |°"Yntake) During tte earlier part of the testing
£e?ioS! occasional opening of doors and windows took place as
comfort requirements mandated.
m moa=ure the impact of subslab depressurization on Rn
TO ™eas r , f , nEriods of high and low Rn concentration in
infiltration, cy	be established for that season. Daily
the building ha 1	showed a significant difference between
t_tests ^ere/^i^^heoncen?ration It was therefore determined
nighttime anc^ daytime Rn^concentra ^ ^ ^ ^ readings day) into
to divide the z	resoectively centered around a maximum and
two uninterrupted halves respectively^	^ unlnte	ted use out
^"tTaloffour:rfor continuous rotation purpose. One WLH was
— to checK how well last
-------
seasons' results could be replicated.
Based on the various t-tests, it was decided to compare, in
both the bathroom and the bedroom,the results obtained from 19:00
hrs to 6:00 hrs (night) against those obtained from 7:00 hrs to
18:00 hrs (day). The maximum readings (fig 2) occured around the
same time period as the previous year (23.2 hrs). To check the
effectiveness of subslab depressurization, it was determined to run
a t-test of day vs. night on the Rn concentration obtained over a
period of two and a half months in both the bathroom and the
bedroom, without any ventilation taking place.
The Rn readings were again always higher in the bathroom. This
was confirmed by readings obtained using Rn canisters located at
regular interval in connecting rooms. The canisters were situated
in the bathroom, the bedroom, the living room and the kitchen, with
the bedroom, living room and kitchen canister located along an
airway respectively 10 m, 20 m and 30 m from the canister in the
bathroom. These readings were repeated ten times and, without any
exception, the decreasing order of the activities remained
unchanged: bathroom, bedroom, living room and kitchen. This seemed
to indicate that the bathroom is the main entry route for Rn into
the building. Although there is no ideal statistical method to
express the existing relationships, some type of quantification of
the strong path evidence can be demonstrated by applying a
regression analysis which yielded:
Y = 263.9 - 5.06X with r = 0.999
where Y = Rn concentration in Bq m'3
X = distance from the alleged source in m
Concurrent readings obtained from the WLMs showed that,
without any single exception, and despite rotation of the WLMs,
readings in the bathroom, which were taken during the day as well
as during the night, were always higher than in the bedroom.
In both cases, day or night, the bathroom readings were more than
50% higher than in the bedroom (fig 3) . Although the parallelism in
the readings is as good as during the previous year, there is a
greater discrepancy in the activity levels during the 1989-1990
season. This is mainly due to a strong drop in Rn levels in the
bedroom.
Parallelism in the readings and, by extension, precision, can
j^e concluded from a multiple regression analysis where one of the
I^XjMs was chosen at random as the dependent variable whereas the
three others were designated as independent variables. The adjusted
coefficient of multiple determination, R2, was found to be, after
106 consecutive measurements, equal to 0.985 (which is highly
significant) . The same test demonstrated further evidence of the
precision through the low coefficients of variation (4.71%)
existing between the various instruments in use.
9
-------
a •»--test oerformed on the Rn measurements taken in the
bathroom^ d^nsS't'ed ttat for
and a daytime ®v®*®g®h ce that the two sets of samples (day and
night) ^ight belong to the same population (or not be different),
is very slim indeed.
After applying the Behrens-Fisher correction where necessary,
« ~»P4-tlU «d	bathroom readings were significantly
it was found that 80% of the Datnr	^ ^
hig^ra?tdid9not show any significant difference and 6% of the
JSas^Sents showed significantly higher daytime values (at the 5%
significance level).
Thp t-test performed on the Rn measurements taken in the
v, ^	.,, nest-rate that for a nighttime average of 85.9 Bq m and
bedroom lllustr	. _ -3 tfte p-value was 4.2*10"*, which
^f^rmonsSated a very significant difference between nighttime
still demonstrat	nighttime readings in the bedroom were
^ ^YcanTlv1T&erinof the cases, not significantly
significantly g the cases and 7% of the readings showed a
significantly higher daytime reading. Table 2 compares the 1988-
1989 with the 1989-1990 measurement period.
c-aon these Rn activities are quite a bit lower than
As can s ,	before. This is also confirmed by the Rn
the ones	about the effect of the
can^ter readings. 0^	that occured in the 1989_199o fall and
sunny (and oft	^ depressurization and consequent lower Rn
Wi Orations (fig 5). Much more frequent use of the woodstove
concentration (g^ winter seemed to correspond to higher Rn
during the p	h	The drop was also found to be much more
drastic ^.n the bedroom (which happens to be much closer to the
stove).
The averaged daily coefficients of variation 
-------
Depressurization time: 6 hours/dav at -175 Pa.
The fan was activated from 0:00 hrs until 6:00 hrs. The
decrease of Rn activity in the house was measured to be within one
hour of start of activation, so that the time of maximum Rn
activity remained at 23:00 hrs. Radon activity in the house
bottomed out about 5 hours after fan activation to 35 Bq m'3 or
less and remained near that level for about 10 hours, so that the
period of lowest Rn activity did not correspond to the period of
subslab depressurization.
nepressurization time: 12 hours/dav at -175 Pa.
The daily subslab depressurization period lasted from 18:30
hrs until 6:30 hrs. The time of maximum activity in the house was
now measured at 19:00 hrs, so that one could conclude that Rn
abatement was measurable within one and a half hour of subslab
depressurization. The half-day periods measuring the highest Rn
activity were from 14:00 hrs until 1:00 hrs. Again, Rn activity
bottomed out about 5 hours after fan activation. Although of
questionable value, since the data sets are not independent, a t-
test of "high" activity vs "low" activity showed that the
difference was still significant.
n<*r>ressurization time: 2 4 hours/dav at -17 5 Pa (from 6:00 hrs
until 6:00 hrs).
Depressurization occurs from 6:00 hrs until 6:00 hrs the next
morning, only to be deactivated for the next 24 hours and
reactivated again the following day at 6:00 hrs. As was the case
previously, fan activation caused an immediate lowering of the Rn
activity with the readings again bottoming out after 5 hours and
resulting further in a curve sharply reduced in amplitude. After
the fan was deactivated the next morning at 6:00 hrs, a rather
rapid rise in Rn activity occured at 16:00 hrs or about 10 hours
after the fan was deactivated. The maximum readings obtained during
the deactivation period were around 23:00 hrs. During
depressurization of the subslab no trend at all was apparent.
nepressurization time: 24 hours/dav at -175 Pa (from 18:00 hrs
nnf.il 18:00 hrs).
Depressurization occurs now from 18:00 hrs until 18:00 hrs,
ending consequently around the time that Rn activity normally
starts to climb. Within a few hours after fan deactivation (at
18:00 hrs), a rapid rise in Rn activity is now witnessed (fig 6).
While the subslab was depressurized, on the other hand, high Rn
activity was inhibited, so that during this period, a flat curve
appeared, contrasting sharply with the curve obtained after the fan
was deactivated (before the diurnal peaks of Rn activity).
Since Rn levels remained low up to 10 hours after fan
deactivation, it was concluded that activating the fan 12 hours/day
11
-------
during the period corresponding to that of highest indoor Rn
activity was the most cost-effective way to use the discontinuous
subslab depressurization system (at -175 Pa).
SUBSLAB DEPRESSURIZATION AT VARYING FAN SPEEDS.
It was obvious at this stage that, regardless of any remedial
action taken, the bathroom measurements remained significantly
higher than measurements taken in any other room. On investigation
as to the probable cause,and removal of a trapdoor accessing the
bathtub, a large slab opening was found. After sealing that opening
with expanding polyurethane foam, only a sporadic and intermittent
difference remained between the Rn concentration found in the
bathroom and the rest of the house. The Rn levels now average 68 Bq
m"3 throughout the house without any subslab ventilation taking
place (average of the last 5 weeks in both the bathroom and the
bedroom; table 3). Table 4 indicates the hourly maxima and minima
obtained under varying circumstances. It is noteworthy that subslab
depressurization results are not significantly different if
measurements are taken before or after sealing of the slab opening.
Investigation of the influence of varying subslab
depressurization on Rn concentration indicated that after sealing
the slab opening, no drastic decrease in indoor Rn activity took
place beyond -50 Pa depressurization, which is the smallest
depressurization attainable through fan activation (Fig 7) . The
influence of warmer weather and subsequent decreased stack effect
can be seen once more as time progresses. Weekly measurement cycles
featuring daily increases in depressurization (from 0 to -175 Pa)
show a trend of decreasing Rn concentrations as weeks (wk) progress
towards springtime (table 3). Subslab depressurization appears to
be effective if the fan is activated during the peak Rn activity
hours (18:30 hrs until 6:30 hrs the next morning). Practically no
activity occured until the fan was left deactivated during the peak
Rn activity period (fig 8).
CONCLUSIONS
, ,	mriQt- cost-effective method used for Rn mitigation
Probably themost cost etre	bathtub. For research
was the sealing of th^slab^opening u ^ ^ piace ^ ^ ^
purposes, it was	indoor Rn activities were strongly
study.	dealing the slab opening and some relationships even
disappeared tot^Ay thereafter (Such as the distance fro "source-
and Rn activity relationship).
The intermittent activation of the fan shows that the Rn
The in™f" effective in most cases, long after fan
mitigation is	certain degree of "exhaustion" of Rn as a
(^obably replaced by atmospheric gases) . This rule does
not seem tHpply if fan deactivation occurs around the tin. that
12
-------
indoor Rn activity normally starts to climb.
Equally low Rn concentrations could be obtained with the
depressurization system in operation, regardless of whether the
slab opening was sealed or not.
Before sealing the slab opening, decreases in Rn activity of
95% were obtained through subslab depressurization (at -175 Pa)
because of the high initial Rn concentration. A noticeable drop in
temperature (-2°C) was also experienced when the system was fully
depressurized. After sealing the slab opening, it appears that the
increased benefits obtained from running the fan at full speed are
marginal and that an overall decrease in Rn activity of 85.3 % of
maximum (obtained at a subslab depressurization of -175 Pa) can be
obtained by running the fan at -50 Pa. Due to the much lower
initial Rn concentration, this only amounts to a decrease of 51.6
% of the incipient Rn activity. A t-test shows no significant
improvement to be obtained by depressurizing the subslab at -175 Pa
instead of -50 Pa (p-value :0.033). A satisfactory reduction in Rn
activity was obtained by depressurizing the subslab at -50 Pa for
only 12 hours/day during the peak Rn activity hours (18:30 hrs
until 6:30 hrs).
The work described in this paper was not funded
by the U.S. Environmental Protection Agency and
therefore the contents do not necessarily reflect
the views of the Agency and no official endorsement
should be inferred.
Acknowledgments-The support of Kirkland Jones, Neil Weber and
Eloy Montoya is acknowledged for making this report possible.
Thanks are also going to Benito Garcia and Bill Floyd for liberal
use of their equipment, to Loren Berge of the State Laboratories
for the radiological analyses and to Ed Essington of the Los Alamos
National Laboratory for the graphics.
REFERENCES
Journal 1. Mowris, R. J. and Fisk, W. J. Modeling the effects
of exhaust ventilation on 222Rn entry rates and
indoor 222Rn concentrations. Health Physics Vol. 54
No. 5 pp. 491-501, 1988.
2.	Nazaroff, W. W. Entry by pressure-driven flow or
molecular diffusion? A reassessment of Rn
concentrations measured in an energy efficient
house. Health Physics Vol 55: 1005-1014? 1988.
3.	Swedjemark, G. A. and Makitalo A. Recent Swedish
Experiences in 222Rn Control. Health Physics Vol.
13
-------
Report	4.
5.
Journal 6.
7.
8.
9.
Symposium 10.
System
Inactive
Passive
Active
58 No 4 pp 453-460, 1990
Lappin, A. R. and Hunter, R. L. Systems Analysis,
Long-Term Radionuclide Transport and Dose
Assessments; Waste Isolation Pilot Plant (WIPP),
Southeastern New Mexico; 1989; SANDIA REPORT
SAND89-0462*UC-70; DOE contract DE-AC04-76DP00789
Hall, A. Unpublished Rn data from New Mexico
Environmental Improvement Division taken from over
250 houses in Northern New Mexico, 1986-1987.
Jonassen, N. On the effect of atmospheric pressure
variations on the Rn concentrations of unventilated
rooms. Health Physics Vol 29:216-220? 1975.
Stranden, E. and Berteig, L. Radon in dwellings and
influencing factors. Health Physics Vol 39: 275-284?
1980.
Schery, S. D.; Gaeddert, D. H. and Wilkening, M. H.
Factors affecting exhalation of Rn from a gravelly
sandy loam. J. Geophys. Res. Vol 89: 7299-7309;
1984 .
Strong, K. P. and Levins, D. M. Effect of moisture
content on radon emanation from uranium ore and
tailings. Health Physics Vol. 42 No 1 pp. 27-
32,1982.
Henschel, D. B. ; Scott A. G. ; Findlay, W. O. and
Robertson, A. Testing of indoor radon reduction
methods in 16 houses around Dayton, Ohio. 1988
symposium on radon reduction technology. Denver,
Colorado.
Table 1: Rn activity (Bq m"3) in 1988-1989
Mean	Std. Dev.
142.7	66.5
82.9	38.6
30.6	6.66
d. f.	c.v. (%)
196	46.6
166	46.6
200	21.7
14
-------
Table 2: Comparison of Rn concentrations
in bedroom (E) during 1988-89 and during 1989-90
and in bathroom (A) during 1989-90
88-89(E) 89-90(E) 89-90(A)
Xn (Bq m"3)	155	85.9	147.8
Xd (Bq m"3)	129	56.4	85.6
p-value	2.6*10"3	4. 2*10~5	7.6*10'6
Xn > Xd (%)	64	70	80
Xd > Xn (%)	28	23	6
no difference	(%) 8	7	14
Table 3: 0 to -175 Pa consecutive (8) weekly depressurization
cycles and their influence on Rn (in Bq m"3) .
(A refers to bathroom and E to bedroom)
Depress.(Pa)
0
-50
-75
-100
-125
-175
Rn (A)
138




25.0
.. (E)
132




19.6
« (A)
122




36.5
" (E)
128




24.2
" (A)
190
43.5




H (E)
167
35.0




" (A)
92.9
35.0
40. 1
38.2
35.6
36.4
« (E)
87.9
39.1
45.0
39.0
34.8
36.3
» (A)
51.0
36.0
32.7
40.5
38.7
31.0
.. (E)
44.9
28.6
25.6
33 . 9
33.9
25.4
" (A)
65. 2
43.0
35.9
43.7
40.7
26.9
" (E)
52.7
35.6
28.3
34 . 6
33.3
21.5
" (A)
83.3
31.6
34.7
37.7
29.2
25.6
" (E)
73.9
27.0
32.0
34 . 5
26.8
22.5
" (A)
73.1
28.4
30.6
29. 5
38.9
31.0
« (E)
50.7
24.7
23.6
22.6
29.1
24.1
Table 4: Hourly extremes (in Bq m"3) :
Max	Min
Before sealing	1330	37
After sealing	299	12.3
-175 Pa depressurization 50.8	2.73
15
-------
10	15
Time of Day (hr)
80
Time (d)
Fig 1. Rn Levels as a Function of Time	Fi9- 2. Hourly Rn Levels
(Day 1 = 3 Sept 1989)	(Winter 1989-90)
400-
400-1
300
200
cr
CD
100-

t, '¦
I l~
50	100	150
E (Bq m"3)
Fig. 3. Rn Levels in Bathroom (A)
vs. Bedroom (E)
200
300
200
100
I Q-
in
i P"
- -P
a-n
0 t-
50 100 150 200
d (Bq m"3)
Fig. 4. Rn Levels, Day (d)vs. Night (n)
250
-------
300
D
r^ifc
[

i r)i1 r7171

1
[£¦
u
TO
D
& ana li
^~f^p
0.5	1	1.5	2
Depressurization (-Pa)
Fig. 5. Room Depressurization
vs. Rn Activity
2.5
200
CO
E
cr
m
1001
u
a
n
n
~
Lb
c.i
n
Li
a
CT<
?•11;,.,• nr« i.n, ^
10
20
Time (hr)
30
40
Fig. 6. Depressurization Ending at 18:00 Hrs
150i
y
100 I
cr
CD
c
cr
50
V

04"	
-200
-150	-100	-50
Depressurization (Pa)
Fig. 7. Indoor Rn as a Function
of Subslab Depressurization
-30 -20 -10 0 10
Time (hr)
20
30
Fig. 8. Indoor Rn as a Function of Time
(-50 Pa depressurization ending at 0 hrs)
-------
IV -6
Natural Basement Ventilation as a Radon Mitigation Technique
A. Cavallo, K. Gadsby, T.A. Reddy
Center for Energy and Environmental Studies
Princeton University
Princeton, NJ 08544 USA
Abstract
basement ventilation has al»ays been recommended as a
Natural basement	, els in houses. However, its efficacy
means of reducing ra	generally been assumed to be a
has never been docume ' strategy since it was believed that
SXttSU'wsiSS mechanism by which radon levels were reduced.
t,l3f(0n has been studied in two research houses
NatTrVt£eXsu™er cooling season and the winter heating
during both the su	environmental and house operating
season. Vent ^^"/^vels have been monitored; it can be
parameters, and ra	reduce radon levels two
concluded that f	fs by simple dilution. The second,
ways. The firs^'	providing a pressure break which reduces
less obvious, way is y v	amount of radon
»d»r?nto the structure .
Thus, ^a^^grthaiawasnpreviously believed. It might be
mitigation strategy	ith x w radon concentrations (of the
especially useful m houses	^	^ cannot be
order of 10	tively with conventional technology.
mitigated cost-etrec^ivc y
heen reviewed in accordance with the U.S,
This paper has	Agency's peer and administrative review
pol icie^and approved for presentation and publication.
Introduction
Radon emanation from naturally occurring soils, as
distinguished from building materials and mine tailings used as
construction fill, has been suspected of being a significant
source of indoor air pollution in single family houses since the
1
-------
early 1980s [1,2,3,4]. This concern grew out of studies
undertaken after the first energy crisis in 1973 to understand
energy consumption patterns in homes and to reduce energy
consumption, among other ways, by sealing up structures and
reducing building air exchange rates [5]. It was immediately
realized that reducing ventilation rates had the undesirable side
effect of causing an increase in trace gases such as volatile
organic compounds, oxides of carbon and nitrogen, and moisture,
decreasing both comfort and safety.
It was initially believed that the effect of ventilation on
indoor radon concentration was the same as for all other indoor
air pollutants, that is that ventilation reduced indoor radon
levels by dilution. This is based on a very simple model [6,7]:
if the radon entry rate SRn is assumed to be constant and set
equal to the removal rate, we have: SRn = X,vCRn, where Xv is the
air exhange rate and is the radon concentration.
Results from initial experiments [8,9] in which it was found
that basement radon concentrations were inversely proportional to
the ventilation rate, as predicted by the above equation, seemed
to confirm this hypothesis. Thus, to reduce radon levels by a
factor of 10 would require an increase in the air exchange rate
by that same factor, which in most cases is neither practical nor
desirable. The experiments were done using an air to air heat
exchanger to control the basement ventilation rate. An air to
air heat exchanger operates in a balanced mode with inflow and
outflow equal and would neither pressurize nor depressurize the
basement.This is actually very different from natural ventilation
in which a basement window is opened, providing a pressure break;
nevertheless it resulted in ventilation's being thoroughly
discredited as a means to control indoor radon.
However, the mechanisms which bring radon into a structure
are completely different from those causing high levels of many
other indoor air pollutants. Most often, the source of
undesirable indoor chemicals is found within the structure
itself, such as poorly sealed paint cans and cleanser containers,
or rug pads and foam stuffing in furniture. Radon entry into a
building is dominated by pressure-driven flow of soil gas rather
than by emissions from building materials. The subsoil pressure
field of the building is caused by the following factors: wind
generated depressurization of the structure, basement
depressurization caused by air handler operation, and most
importantly, by basement depressurization induced by the
temperature difference between the outdoor environment and the
building interior (the stack effect).
2
-------
rii emission that the radon entry
It is clear from	be a function cf the
e rannot be a consta	_ f ntial Thus, basement
basement to subsoil Pr®s^fL deduce indoor radon levels both by
nrnation can theoretl ^ LessUre break which reduces the
dUution and by Prf ^^/differential which reduces the radon
basement to subsoil f
entry rate [10] •
Experiments
basement ventilation, that is opening
The effect of naturaJn^ Jadon levels has been examined in two
basement windows, on -f^^ouses (PU31 and PU21) during the
Pri?ertheating6season and the summer cooling season.
have been instrumented as follows:
The houses have Deen
¦ ic across the building shell and between
!. Pressure dif	^stairs (PU21 only) are measured with
differential p^essure^transducers^ ^ ^	temperatures
areBmonitored using thermistors ^ in-the-block radon levels
3 Basement, living area^ su^ # ^ (La#jre[,ce Berkeley
(PU21 only) '^Monitor) or a PRD (Pylon passive radon
rnntinuous Raaot*
detector) .	ive humidity is monitored with a CS 201 relative
4.	Basement reiat
humidity Prob®' ,r conditioning system usage is monitored using a
5.	Heating and air
oaii switch.	t-varpr) system is used to measure
6 A PFT (PerflU°r°^rra?e and interzonal flows. Up to four
building air exc*a,^e *iS system, but for these experiments only
aases may be used in this y	per 2Dne) are placed in
So were	.^f^efs in the basement and living area.
temperature regu
station at Princeton University
in addition, a weat^Jall, relative humidity, barometric
Hi- a as well as house dynamics data are
The weather sta^10nn!faveraged over 30 minutes, while the air
read every 6 se"^Je^onal flow measurements are averaged over a
infiltration and inter
minimum of 2 day
-------
EXPERIMENTS IN RESEARCH HOUSE PU21
Natural ventilation experiments have been carried out in
research house PU21 during the winter heating season; the results
of these experiments are summarized here.
The research house has the following characteristics:
SIZE: 1970 ft2* living area, 525 ft2 basement.
TYPE: Modified ranch. The living room/dining room has a
cathedral ceiling with a large window area facing
almost due south. A cinderblock basement underlays
about one third of the house, with the remainder being
built on a slab. There is a cinderblock chimney stack
in the center of the house.
FIREPLACE: Large fireplace in the living room.
HEATING SYSTEM: Central gas forced air heat, furnace in
basement.
COOLING SYSTEM: Central air conditioning.
HOT WATER:	Gas hot water heater located in basement.
RADON LEVEL:	-120 pCi/L in basement.
The house had been mitigated with a subslab mitigation system
which was turned off during the ventilation experiment. The
perimeter floor/wall shrinkage crack had also been sealed and
Dranger© basement drain seals installed as part of the
mitigation.
The effect of opening a basement window on indoor radon
levels and the basement/outdoor pressure differential in PU21 is
illustrated using continuous radon and pressure data in Figs, la
and lb. Data points are 30 minute averages of the parameters;
the experiment was carried out be'tween Julian Date (JD) 47, 1990
(90047) and JD90050.5. Shown in Fig. la are basement radon
levels as measured with a pumped CRM, which has a response time
of less than 30 minutes, and upstairs radon levels as measured
with a Pylon PRD, which has a response time of about 3 hours.
Plotted in Fig. lb is the pressure differential across the south
wall of the basement (positive values indicate that the basement
is depressurized relative to the outdoors). A normally closed
basement window was opened at times JD90048.4 and JD90049.45,
and closed at times JD90048.83 and 90049.8.
"Readers more familiar with metric units may use the factors
at the end of this paper to convert to that system.
4
-------
The b..ewnt/o»tdoor.^r«.^	of the window with a -1.5
immediately to the	. (50te that, even with the window
Pa change in thisfp"*™f Remains depressurized relative to the
open, the basement still re .ndication that the racion entry rate
outdoors.) This is a st" 9	his is ln fact the case, as
into the basement must	^ experiments of building air
verified by ^measurement s^i^	levels<	and radon entry
rates.
« i rtfi^PT" D6riod of tinis to s
Radon levels resP°^	This is to be expected since the
window's owning or closxng. (defined as the flow of outdoor
total basement air	living area into the basement) is
air plus the "ow,£r®"	per hour (ACH) , and the building air
approximately 1 ax^ q 1-0.6 ACH. Thus, the time necessary to
exchange rate is about 0.3	^ ^ the order of 2 or 3 hours,
achieve a new steady stat	^ ^ upstairs radon detector is
in addition, the respoi? hours, which is why there is such a
itself of the order of ^	e of the upstairs and basement
difference in tne
radon levels.
¦i mnnrtance to note that there are natural
It is also of some i P behavior which are of the same order
variations in the buiiainy	opening a basement window. An
of magnitude as those'c	time JD 90048f The deCrease in
example of this occurs ar	surizati0n in this time period
indoor radon and basemen J\nter temperature spike m which the
was caused by an unus^ d fell by 8 'C in a 12 hour period,
outdoor temperature rose	temperature differential and the
changing the ^do	effect. It is essential that an
magnitude of the stac* e duration t0 be able to average over
experiment be ot sum
such excursions.
<- • i at- ion experiment in PU21 was conducted over
The natural ventJ:®^arY. tw0 periods of 2 and 3 days each
a 17 day period in Fee * line building conditions (windows
Sere used to defermine the base^ ^ ^ ^ determine thfi
closed), and three 4 aay v ^	& single basement window (-2.2
building operating Par* ._ 2 thr0ugh 4, described below,
ft2window area) open, x ;?ods when the basement window was
experiments 1 ana o	*	. 4 are periods when the basement
closed, and experiments
window was open.
uont" i lation on basement and upstairs
r»doneievelsti»fsho"'"in Fig. 2. with the windows closed,
-------
basement radon levels were about 120 pCi/L, while upstairs levels
were about a factor of 2 or less lower (80 pCi/L). This is a
fairly typical result and is a consequence of the basement's
being isolated from the living area. With one basement window
open, the upstairs levels were about a factor of 2 higher than
the basement levels. This is quite unusual and indicates a radon
entry route into the living area which bypasses the basement.
This result was checked by making two simultaneous continuous
measurements of the upstairs radon levels. A similar result was
noted in the measurements made in the summer of 1989 on PU31 and
will be discussed further; this indicates one way that basement
ventilation, while certainly reducing indoor radon levels, might
not be as effective in reducing living area radon levels as in
reducing basement levels.
Another consequence of a reduction in basement radon entry
rate is an increase in subslab and basement radon levels. This
is indeed observed, as shown in Fig. 3, in which basement and
subslab radon levels are plotted for the different experiment
periods. The strong decrease in basement radon levels with the
window open and the simultaneous increase in subslab radon levels
are clearly present. The expected magnitude of the increase in
subslab radon levels is not obvious, since it would depend on the
details of the amplitude and spatial distribution of subslab soil
permeability, moisture, and radium content. Qualitatively, the
effect is certainly present.
A critical factor in this experiment is to quantify the
effect that basement ventilation has on the building air change
rate, since the observed reduction in radon levels could be
caused by a large increase in the ventilation rate. This has been
done using the perfluorocarbon tracer (PFT) system, and results
are illustrated in Fig. 4, in which building air exchange rate
and basement radon levels are plotted. The building air exchange
rate increases by a factor of 2, from 0.3 to 0.6 ACH, when the
basement window is opened. Note that the basement radon levels
decrease by a much larger factor (~6-8), again indicating that
dilution cannot account for the entire decrease in radon levels.
The doubling of the air exchange rate corresponds to a
ventilation rate of 115 cfm, very roughly comparable to that
achieved by a subslab depressurization system, which for this
house reduces radon to much lower levels than basement
ventilation. However, the main application of natural ventilation
is expected to be in lower level homes where installation of a
subslab system might not be justified.
Using the interzonal flows and tracer gas concentrations
6
-------
Art bv the PFT system, the basement and living area radon
measured by tn	calculated. The two zone system of flows and
entry rates	illustrated in Fig. 5. Radon entry rates
tracer concentrat	uJated tw0 „ayE. The first method is to
S""	ates deduced from tracer gas measurements but
use the flow	^ radon concentrations in zones 1
Tas-enU an'd 2 diving area), respectively.
Sir„ = V S*
equal in this case, about 5 jiCi/h. with the basement windoUt
closed the basement radon entry rate, approximately 20 UCi/h
predominates. This does add an extra complication to the
7
-------
natural ventilation as a mitigation strategy. It remains to be
seen how widely this effect is observed.
Therefore, measurements in PU21 clearly demonstrate the
mechanisms by which natural ventilation acts to lower radon
levels. Both dilution and reduction of the basement/outdoor
pressure differential and the concomitant reduction in radon
entry rate are factors, with the second effect being the more
important.
EXPERIMENTS IN RESEARCH HOUSE PU31.
Natural ventilation experiments have been conducted in
research house PU31 over a complete seasonal cycle; that is,
during the summer cooling season and the winter heating season.
The results of these experiments are summarized for both.
Research house PU31 has the following characteristics:
SIZE: 1600 ft2 living area, 1300 ft2 basement.
TYPE: Ranch with full attic and full basement, half of
an attached slab-on-grade, two-car garage
converted to TV room, cinderblock wall basement
with a sump, and cinderblock chimney stack in the
center of the house.
ATTIC: Two 1100 cfm attic fans, thermostatically controlled;
insulated with 8 in. blown-in insulation.
FIREPLACES:	Two: one in l-iving room, one in kitchen.
HEATING SYSTEM: Central gas forced air heat, furnace
located in basement.
COOLING SYSTEM: Central air conditioning.
RADON LEVEL:	~80 pCi/L in the basement.
Research house PU31 has been instrumented in a similar
fashion to PU21, except that subslab and cinderblock wall radon
are not measured, and the pressure field of the basement is
measured at three heights on each basement wall and at three
subslab locations.
cooling Season Measurements
The summer season natural ventilation experiment was
conducted in the following manner. A 17 day period was used to
establish an operating baseline for the house. During this time
8
-------
the house functioned normally/ e.g., thermostatically control l oh
attic fans operated automatically. Basement and upstairs wi £
were kept closed, as is normally the case since the' house is °WS
centrally air conditioned. (Upstairs windows were of excelle
quality and could be closed tightly. The basement windows wer«
low quality steel frame casements which could not be shut v
tightly.)	er^
After the baseline operating conditions of the building
established, two basement windows (one on the west wall and"616
other on the east wall, each 2.2 ft2) were opened and the
relevant parameters compared to those obtained in the basel"
conditions.	lne
The effect of opening two basement windows on basement
levels and the soil to basement pressure differential is h n.
Figs. 8 and 9. Basement radon levels are shown in Fig 8^ th" in
is clearly a significant drop in this parameter, from an av« ®
of about 90 pCi/L to about 10 pCi/L when the windows are on e
on JD89220.6. The magnitude of this drop was completely Pened
unexpected. The large diurnal variation in basement radon 1
is due to the operation of the attic fans which depressuriz V6ls
entire house, increasing the'ventilation rate as well as th th&
radon levels. Measurements of a typical differential pr
transducer are illustrated in Fig. 9 (positive pressure ind^^
that soil pressure is above that of the basement). The 1	tes
peaks (-3 Pa) in soil/basement pressure differential are d*^6
the operation of the attic fans. There is an abrupt pressi^ t0
drop when the windows are opened, indicating that the Dreccf6
field of the building has been modified, it is clear that
this house only, a very small pressure differential (~o 5 p
needed to drive the radon level to 10 pCi/L. This result -iS
strongly suggests that a modification of the basement/soila9ain
pressure differential is important in reducing the basemenl-
level; however, the measurement of the building air excha rad°n
and interzonal flows and calculation of the radon entry	r&te
essential for a definitive evaluation of this problem.	are
The behavior of the basement air exchange rates and ba
radon level is shown in Fig. 10/ these two parameters are?60*
for seven experiments, each of 3-4 days duration. This p otte<3
time was needed to obtain reasonable levels of the PFT QPer^OCi
capillary adsorption tubes. Baseline conditions for theh
(with the attic fans thermostatically controlled) were ah
ACH for the entire building with an average basement	0,3
of about 80 pCi/L.	radon level
9
-------
With the basement windows opened, the building air exchange
rate increases by about a factor of 2, to 0.6 ACH. Basement
radon levels decrease to about 12 pCi/L, a factor of about 7
below the levels with the windows closed. This decrease is far
larger than the increase in the building air exchange rate (about
a factor of 2), and indicates that the change in the pressure
field of the building is much more important in decreasing radon
levels than the increase in the building air exchange rate.
To investigate the impact of the attic fans on building air
exchange rates, the two basement windows were left open and the
attic fans switched off. The building air exchange rate dropped
by about a factor of 2, while the basement radon level dropped by
about 20%. Such a large decrease in the air exchange rate
without any increase in radon level is yet another indication
that the modification of the pressure field of the basement and
thus the entry rate S1Rn (which is a function of the soil to
basement pressure differential) is of prime importance in
determining the radon level of this basement.
As for house PU21, the basement radon entry rate of house
PU31 can be computed using the air infiltration and interzonal
flow measurements. Results from this calculation using Eq. 3 are
shown in Fig. 11. If the baseline house operation (Experiments
1-5 of Fig. 11) is compared to house operation with the attic
fans off and the basement windows open (Experiments 7-8 of Fig.
11), radon entry rate decreased by about a factor of 7. For
house operation with attic fans off and basement windows closed
(Experiment 6) compared to that with the fans off and windows
open (Experiments 7-8), the basement radon entry rate decreased
by about a factor of 3. This demonstrates clearly that the radon
entry rate decreases significantly with natural basement
ventilation.
Although basement radon levels have been emphasized in the
above analysis, radon levels in the living area are of most
concern. These have also been measured during the natural
ventilation experiments. With all windows closed, the upstairs
radon level (~62 pCi/L) was lower than the basement radon level
(-80 pCi/L), as would be expected. However, with basement
windows opened, the upstairs radon level (~25 pCi/L) was about
2.5 times higher than the basement radon level (~10 pCi/L) (see
Fig. 12). Instrumental error has been carefully ruled out in
this case. It is clear that radon can enter the upstairs zone of
this house two ways. The first is the usual one in which soil
gas is drawn into the basement and then flows into the upstairs
zone. The second entry route must bypass the basement but could
10
-------
not be localized. It may be associated with the central
cinderblock chimney stack or the slab-on-grade garage which h
been converted into a TV room. This second route is unaffe i>aS
by the pressure break provided by the open basement windows? ^
Heating Season Measurements.
A series of measurements on natural basement ventilation
conducted in PU31 during the winter heating season; a tempora^6
mitigation system was installed in the house in January 1990 ^
This system was turned off and the vent pipe capped during th
ventilation experiment.	y e
Measurements to determine the house baseline operating
conditions were begun in December 1990. Radon levels in th
living area of 70 pCi/L were routinely found, and it was deem h
advisable to install a temporary mitigation system immediate?
This was done on January 5, 1990, and reduced upstairs rado
levels to about 4 pCi/L. The mitigation system was turned Sff
JD90030 and an attempt made to measure another baseline poi r °n
Radon levels were about a factor of 2 less than those found • "
other baseline measurements (Compare Fig. 13, Experiment l t*1
Experiment 5, 6, or 7.) It appears either that it takes se ^
days for the house to return to its unmitigated operating
from the time when the mitigation system is turned off L
this was an exceptional case, perhaps because of some other *
change in the house operating point. Since the building ai
exchange fate was 40% lower for this experiment than for oth
experiments with the windows closed in this series (see Fi
Experiment 1 compared to Experiments 5,6,7), this change in"**3''
operating point certainly could explain much of the discrena
Experiment 1 is included for completeness, but the baselin
experiments (windows closed) to which others will be cornea6
(windows open) will be Experiments 5, 6, and 7.	C
Basement radon and building air exchange rate for PU31
shown in Fig. 13 for the winter ventilation experiments Th*
baseline air exchange rate is about a factor of 2 larger th
that found in the summer measurements (0.3 ACH, summer* o fi?1*
winter). This is due to the larger indoor/outdoor tempera* *CH'
differentials which occur in the winter. The air exchan
doubles, from 0.65 to 1.2 ACH, when either one or two wind rat%
(2.2 ft2 per window) are opened. Basement radon levels al **
higher than the summer values, decreased by more than a f
10, from -130 to -12 pCi/L with the east and west windows^0* of
with the west window open. The west window is just above	n °r
pump and -10 ft away from installed instrumentation it 7
At is not
11
-------
clear why the west window should be more effective in reducing
basement (and upstairs) radon levels than the east window, but it
may be that providing a pressure break immediately above the sump
pump, which may be a strong source, is more efficient than
locating the pressure break at a distance of 44 ft.
Basement and upstairs radon levels are shown in Fig. 14.
Both are strongly reduced by natural basement ventilation, but
the reduction in upstairs radon is about a factor of 2 less than
that by which basement radon is reduced. This is to be expected
when the radon source is located in the basement, and can be
understood from the interzonal flow and infiltration and
exfiltration measurements.
In contrast to the measurements made during the cooling
season, there is no indication that upstairs radon levels are
higher than basement radon levels with the basement windows open,
and no indication of an entry path which bypasses the basement.
It is not clear why this change has occurred.
The radon entry rate and basement radon levels are shown in
Fig. 15 for the winter natural ventilation experiments. The
first data point shows an anomalously low entry rate and radon
level as discussed above. With either the east and west windows
open or only the west window open, the radon entry rate is
reduced by about a factor of 5, compared to with the windows
closed. Note that, vizh only the east window open, the entry
rate is approximately the same as when the windows are closed,
although the radon levels are about a factor of 2 lower. This
may be the result of an ineffective pressure break with only
dilution reducing basement radon levels.
Thus, heating season natural ventilation experiments in PU31
indicate that radon in houses is reduced both by dilution and by
the introduction of a pressure break when basement windows are
opened. The factor by which radon levels are reduced is even
larger in the winter than in the summer: basement radon levels
are reduced from much higher winter levels to about the same
value as in the summer measurements.
CONCLUSIONS
Natural ventilation experiments conducted during the summer
cooling season and the winter heating season in research house
PU31 and during the winter heating season in research house PU21
have demonstrated that basement ventilation can reduce indoor
12
-------
radon both by reducing the radon entry rate and by dilution
Calculations based on measurements using the PFT gas system*
the effects of dilution and entry rate reduction to be delin "
and quantified: a decrease in the basement radon entry rate eJte
-------
system.
REFERENCES
1.	van Assendelft, A.C.E., and Sachs, H.M., Soil and Regional
Uranium as Controlling Factors of Indoor Radon in Eastern
Pennsylvania, Princeton University Report PU/CEES-145, 1982.
2.	Gross,S., and Sachs, H.M., Regional (Location) and Building
Factors as Determinants of Indoor Radon Concentration in
Eastern Pennsylvania, Princeton University Report
PU/CEES-146, 1982.
3.	Sachs, H.M., Hernandez,T.L., and Ring,J.W., Regional Geology
and Radon Variability in Buildings, Environ. Int. 8., 97
(1982) .
4.	Scott, A.G., Ch 10 in Radon and Its Decay Products in Indoor
Air, W.W. Nazaroff and A.V.Nero, editors, John Wiley and
Sons, New York, NY, 1988.
5.	Socolow, R.H. ed., Saving Energy in the Home, Ballinger
Publishing Co., Cambridge, MA, 1978.
6.	See also the discussion in Chapters 1, 2, and 5 of Radon and
Its Decay Products in Indoor Air, W.W. Nazaroff and A.V.
Nero, editors, John Wiley and Sons, New York, NY, 1988.
7.	Radon Reduction Methods: A Homeowner's Guide, U.S.
Environmental Protection Agency, OPA-86-005, p 4, August
1986.
8.	Nazaroff, W.W., et al., Radon Transport into a Detached One
Story House with a Basement, Atmos. Environ.19, 31 (1985).
9.	Nazaroff, W.W., et al., The Use of Mechanical Ventilation
with Heat Recovery for Controlling Radon and Radon Daughter
Concentrations in Houses, Atmos. Environ. JL£, 263 (1981).
10.	Cavallo, A., Berkner,C., and Gadsby,K., Use of Ventilation to
Control Radon in Single Family Dwellings, Proceedings, Fifth
International Conference on Indoor Air Quality and Climate,
Vol. 3, p 489, Toronto, Canada, August 1990.
14
-------
Hubbard,L., Gadsby,K., Bohac,D., Lovell,A., Harrje,D.,
Socolow,R., Matthews,T., Dudney,C., and Sanchez,D., Radon
Entry into Detached Dwellings: House Dynamics and Mitigation
Techniques, Rad. Prot. Dos. 24, 491 (1988).
D'Ottavio,T.w., Senum,G.I., and Dietz,R.N., Error Analysis
Techniques for Perfluorocarbon Tracer Derived Multizone
Ventilation Rates, BNL 39867, Brookhaven National Laboratory,
Upton, NY, June 1987.
15
-------
Conversion Factors
Readers more familiar with metric units may use the following
factors to convert to that system.
Non-metric
Times
Yields
cfm
0 .00047
m3/s
ft
0.30
m
ft2
0.093
m2
gal.
0 .0038
m5
in.
2.54
cm
pCi/L
37
Bq/m3
16
-------
200
o
Q.
S
CC
100 H
Up Rn
Bsmt Rn
90047.0 90048.0 90049.0
Julian Date
90050.0
Figure 1a. Basement, Upstairs Radon Level vs Julian Date
Sequence of Window Open and Window Closed, PU21
0=0pen; C=Closed; T= Temperature Spike
5
4
3
2
1
c/>
-O 0 "•
3 90047.0
o
90048.0
90049.0
Julian Date
90050.0
Figure 1b. Outdoor/Basement Pressure Differential vs Julian rw
Sequence of Window Open and Window Closed PU21
0=0pen; C=Closed; T=Temperature Spike
-------
0 H	1	1	1	1 i	1- 0
1 2 3 4 5
Experiment
Figure 2. Basement, Upstairs Radon,PU21
Experiments 1,5 Window Closed;
Experiments 2,3,4 Window Open
200
1200
800 B
Experiment
Figure 3. Basement, Subslab Radon PU21
Experiments 1,5 Window Closed;
Experiments 2,3,4 Window Open
-------
200
o
a.
c
o
tr 100
§
E
0)
c8
CO
T


-
£ ,
\ j
% /


\ I •
* /
\ /
_ 4 /


\ 'T -
\ /
\j


\ ' A
V
A
/ \
/ \
J *
7 t

/ \

/ \
t \r
1 1 1	1	1	1-
0.8
0.7
0.6
0.5
0.4
0.3
0.2
X
o
<
03
rr>
c
03
JCZ
o
X
LU
< —«-•
C 	*-
3
CO
1 2 3 4 5
Experiment
Figure 4. Building ACH, Basement Radon,PU21
Experiments 1,5 Window Open;
Experiments 2,3,4 Window Closed
Ro2
Cij Zone 2
R20
R01 '
R10
1R12
BsmtRn
Bid ACH
Zone 0
R21'
Cu Zone 1
Cu - Concentration of Tracer i in Zone J
Rij- Flow from Zone i to Zone j
Figure 5. Flows and Tracer Concentrations for Two Zones
-------
300
200-
"O
-10
s> 100-
	B-	B
LU
CD
T
5
T
1
T
3
T
4
2
Experiment	—•— Entry Rte
Figure 6. Basement Radon Entry Rate, Basement Radon,PU21
Experiments 1,5 Windows Closed;
Experiments 2,3,4 Windows Open
•10 JS
	¦— Eq.3
	Eq. 1
1 2 3 4 5
Experiment
Figure 7. Entry Rate Calculations Compared, PU21
Experiments 1,5 Window Closed;
Experiments 2,3,4 Window Open
-------
200
o
Q.
C
o
"O
to
cc
100 -
/

89216
89218
89220
89222
Julian Date
Figure 8. Basement Radon vs Julian Date, PU31. In This Experiment
Two Basement Windows Were Opened (O) at JD 89220.6.
cs
a.
<
C
-------

I
o
<
-------
Q_ 60 -
fa--_
G
2
3 4 5 6 7
Experiment
--a-- BsmtRn
	4— Up Rn
Figure 12. Basement, Upstairs* Radon Level,PU31 Summer
Experiments 1-6, Baseline (normal house operation)-
Experiments 7-8, Windows Open
200
o
a.
c
o
cc
«—>
c
0)
E
CD
CC
m
100
--a-.-
Bsmt Rn
Bid ACH
Experiment
Figure 13. Basement Radon, Building Air Exchange, PU31 Winter
Experiments 1,5,6,7: Windows Closed;
Experiment 3: East and West Open;
Experiment 2: East Open;
Experiment 4: West Open
-------
200
K"1
C
o
0
Bsmt Radon
Up Radon
1 2 3 4 5 6 7
Experiment
Figure 14. Basement, Upstairs Radon; PU31 Winter
Experiments 1,5,6,7: Windows Closed;
Experiment 3: East and West Open;
Experiment 2 :East Open;
Experiment 4: West Open
Experiment
Figure 15. Basement Radon, Entry Rate PU31 Winter
Experiments 1,5,6,7: Windows Closed;
Experiment 3, East and West Open;
Experiment 2, East Open;
Experiment 4, West Open
_i
¦-" Bsmt Rn
Entry Rte
2 3 4 5 6 7
-------
IV—7
TITLES Attic Pressurization - A Radon Mitigation Technique for
Residential Structures
AUTHOR: Myron R. Edelkind, Southern Mechanical
This	paper	was	not	received	^n	^me	^to^be^includeti^in^the
you^ registration	complete copy of the paper.
In residences, normally occuring attic depressur1..£ion can
be a primary source of reduction of basement air pressures and,
therefore, a considerable factor as a driving force for R222 (Radon).
The literature does not well document the effects of upward draft
in residences created by solar insolation on the roofs of structures.
Venturi forces generated as a result of wind, and roof design, and
the draft induction resultant of the operation of attic ventilation
equipment.
In order to guarantee and achieve mitigated radon levels below
2 pCi/L the radon mitigation contractor may wish to apply more than
. •	u„ chou that the alleviation of attic de-
one reduction technique. We snow
pressurization is highly effective in reducing radon levels in
residential structures. Techniques and equipment mod!f,cat>ons are
de scr ibed .
-------
Session IV:
Radon Reduction Methods -- POSTERS
-------
RADON MITIGATION FAILURE MODES
by: William M. Yeager
Research Triangle Institute
Research Triangle Park, NC 27709
D. Bruce Harris
U.S. Environmental Protection Agency
Air and Energy Engineering Research Laboratory
Research Triangle Park, NC 27711
Terry Brennan and Mike Clarkin
Camroden Associates
Oriskany, NY 13424
ABSTRACT
An EPA study solicited anecdotal information on failure modes of rad
mitigation systems from practicing mitigators, state government agencies °k-
monitor radon mitigation, and EPA radon mitigation project officers and	^
contractors. This study identified three categories of failures: des'
flaws, component problems, and occupant activities which compromised °
mitigation systems. This paper reviews several examples of failure mode
each of these categories.	ln
Radon mitigation systems, like other mechanical systems, are subject
failure and should be designed accordingly. Mitigators should design 3° t
to minimize the probability of failure and to readily detect failures t^S 1113
occur. The system design should include a monitor which occupants can &t d°
determine whether or not the system is operating properly. Occupants USS t0
realize that even well-designed and properly installed systems have so^St
chance of failure; they should check the system monitor periodically an
-------
INTRODUCTION
BACKGROUND
For several years, the U.S. Environmental Protection Agency (EPA) has
been funding radon mitigation demonstration projects in various states. These
projects have developed diagnostic measurements and procedures to select the
most appropriate mitigation technique for a particular house. A variety of
mitigation techniques have been tested in over 170 houses (1). In most
houses, post-mitigation measurements have shown that radon concentrations in
the living areas were reduced below the EPA's guideline of 4 picocuries per
liter (pCi/L).
The EPA has monitored the long-term effectiveness of these mitigation
systems with radon measurements during successive heating seasons. Most
houses have shown little degradation in the effectiveness of the systems, but
in a few, the systems have stopped working altogether. In others, the systems
are much less effective than they were initially.
PURPOSE
This project was undertaken to study the failure modes of radon mitiga-
tion systems. The study focused on systems which once worked satisfactorily,
but stopped working either completely or nearly completely. The study was not
intended to deal with "problem houses," where the installed mitigation system
never performed satisfactorily, or with systems whose performance has degraded
somewhat, but is still generally satisfactory.
Research Triangle Institute (RTI) solicited information on mitigation
system failures from practicing mitigators, state government agencies which
monitor radon mitigation, and EPA radon mitigation project officers and con-
tractors. During the EPA radon conference in February 1990, RTI convened an
impromptu discussion group of approximately 50 attendees to discuss failure
modes of radon mitigation systems. Some of them later provided additional
details about problems that they had experience or observed. They asked
about design flaws, component problems, and resident activities which
compromised mitigation systems. This paper discusses the wide variety of
radon mitigation system failures noted.
Although the study did not involve any measurements, people who worked
for government agencies were asked if they had a data base from long-term
follow-up radon measurements or if they knew of anyone who might have one.
Unfortunately, the response to this inquiry was uniformly negative. Some
data were received on immediate post-mitigation radon measurements from two
sources: the New Jersey Department of Environmental Protection (NJDEP) and EPA
Region 3 (Philadelphia).
-------
ORGANIZATION
The rest of this paper summarizes the anecdotal information collected
during this study. Most of the information refers to subslab depressurization
systems, as this is the most common mitigation technique used by commercial
mitigators. Sections 2.0 through 4.0 discuss failure modes in the three
categories which were established: design flaws, component problems, and
resident activities. Section 5.0 draws conclusions and suggests some areas
for future work on residential mitigation failure modes.
DESIGN FLAWS
were concerned that conscientious and competent mitigators
Several Pe°PX* unscrUpulouS or incompetent ones. If mitigation
could not compete	measurements immediately after installation,
systems are }udgea y j	quality components may not be distinguishable
poorly designed	^,tiL ™y f.vor the poorer systems. The
from better ones	V. whQ have pasSed EPA's Radon Contractor
recent listing o proq?am (2) should help homeowners to identify competent
Proficiency (^CP)dJ^on( several states distribute similar lists of mitiga-
tiri9»horhav. L^isfi.d it.te requirements.
faf,.nr in the radon mitigation business is real estate transac-
A manor facto	n radon leveis below 4 pCi/L. Under these
tions which are co	^	incentive for a quick and inexpensive
circumstances, the	which is seen as a radon measurement > 4 pCi/L,
solution to the pro e '	k unless the health risk is recognized,
rather than a J^-ter^health^is^ ^ ^	^ ^	which
mustrbe°surmounteiyas quickly and inexpensively as possible.
CONDENSATION OF SOIL GAS MOISTURE
contacted knew that soil gas is very moist and that ducts which
Everyone	^ned with a positive slope so that the inevitable
exhaust it should be	duct. Everyone had also seen mitigation
condensation will drain^^ ^ ^ water trap. Sometimes the trap was part of
systems which fail	^	provided. Such drain lines tend to clog
the design and ad	freeze in cold climates. The trap then fills with
flow in the duct, several „itig.tors reported
rerouting ducts to eliminate such water traps.
ronnrted water accumulating in long horizontal ducts in
Some mitigator P	developed or was not originally noticed. All
si^.
FROZEN PRECIPITATION OR CONDENSATION
• - * oositive slope, they may be subject to con-
Even when ducts.^ long rung in unheated or exterior space,
densation problems it	, ir flows. Condensation may freeze to the in-
particularly if the^.haV® n draining down, gradually choking the air flow,
side of the duct rather than or
-------
If the duct is exposed to alternate heating and cooling, ice may form and then
break loose, dropping down the duct into the fan. One mitigator who works for
a national company mentioned that they have a guideline which requires that
exterior ducts be insulated if the winter season has more than 5,000 heating
degree days.
FAN MOUNTING
Improper fan mounting can lead to a variety of problems with mitigation
systems. The EPA recommends that fans be vertically oriented so that con-
densation will drain through without accumulating in the fan housing. The
Agency also recommends that fans be located outside the building envelope so
that all ducts inside the building are under negative pressure (3). Thus, if
any leaks develop in the duct, indoor air will be pulled in rather than radon-
laden soil gas being pushed out. The fans used in radon mitigation systems
have powerful motors which tend to vibrate and must be securely mounted to a
sturdy support. Two mitigators cautioned about securing fan supports to a
frame wall because the wall may act as a sounding board, amplifying the fan
noise. One mitigator reported a failure where the fan housing was supporting
the weight of a vertical duct and warped enough to bind the fan blade.
Mitigators should also consider the environment in which the fan must
operate. Florida attics are hot in the summer; Minnesota attics are cold in
the winter. It may be difficult to imagine temperatures of -20 or 120 °F (-29
or 49 °C) when working on a roof in April, but a fan which is mounted there
will experience a wide range of environments. Even if the fan is rated for
the entire range of environmental conditions which it will encounter, extreme
temperatures may contribute to premature failure. Insulating the fan housing
or shielding it from direct exposure to wind, rain, and sunlight may moderate
effects of extreme conditions.
FOREIGN DEBRIS
Several mitigators mentioned unpleasant experiences with small animals
which had entered a duct through an unscreened exterior opening. One noted
that children put toys and trash into such openings. Systems which use out-
door air to ventilate or pressurize inside space should have a filter as well
as a screen. These filters should be cleaned or changed frequently during
times of the year when plant debris (seeds, flower parts, leaves, etc.) may be
airborne.
HIGH WATER TABLE
During their pre-mitigation inspection, some mitigators look for a de-
watering system or for water stains on basement walls as an indicator of a
"problem house." A subslab depressurization system which is blocked by water
will not be effective. Even when there is no standing water, some soils will
expand when wet and will close off subslab communication. If subslab suction
is the selected mitigation technique and there is any indication of an
occasional high water table, the pit excavated under the duct penetration
through the slab should be enlarged and the duct should extend a minimal
distance below the slab. This should provide sufficient volume to accommodate
some water accumulation without restricting radial air flow.
-------
Homes in areas with a high water table may have an existing sump whi h
can be used as a suction point for a radon mitigation system. A very effe
tive way to extend a pressure field under the slab is by depressurizing a
which is connected to footing drains. The sump should be sealed with an ?UKip
tight cover, which must be removable to allow servicing or replacement
pump. If the existing pump is not submersible, it should be replaced with
that is, since rusting of the pump will accelerate when the sump is sealed °n6
The cover should contain a drain to allow the sump to collect water from
above, as well as below, the slab. This drain should have a seal which
water to pass while maintaining suction in the sump. If this seal fails °WS
suction will be reduced. This could seriously reduce the effectiveness of
mitigation system, particularly if there is a low flow rate of soil gas the
RE-entrainment
RE-ENTKAilNi'ic,!" x
lines some people mount fans inside buildings
in spite of the EPA 9U1	positive pressure. A few mitigators had
so that some of the duct is un Peither from leaks in ducts which were under
seen problems with re"®n	^hich terminated immediately outside a build-
positive Pressure, or from	importance of following the EPA guidelines for
ing wall. This J1^*t^8Juilding envelope and terminating ducts where re-
mounting fans outside the du	y	^ ^ exhaust is at or near grade, it
entrainment will not be a p	^ preVent re-entrainment and in an area
should be far enough f:romt.*	away frora patios or gardens) . Pre-
of the yard not utilized by P P ^	&hove fche roof to prevent
ferably, the exhaust sh°nl re_entrainment through windows or chimneys. Some
blockage by snow, as we	plumbing vents terminate at least 2 ft (0.6 ro)
building codes	} horiz0ntally from any openings,
vertically and 10 ft u mi
* fh» ootential for leaks in the vent from an aeration
One person mentioned	P	^	water. The air vented from such
system installed to	don concentrations than soil gas. If the fan
systems may have much hign	inside the house near the aeration unit,
which exhausts the vent is	amounts of radon into the house,
any leak in the duct could introa
COMPONENT PROBLEMS
FANS
A long-term follow-up study of 40 houses in Pennsylvania mitigated by an
EPA contractor found that 5 of 36 houses with active soil ventilation systems
had experienced fan failures (4). Four were due to capacitor failures in the
fans' split-phase motors. When the capacitor fails, the motor continues to
run at reduced efficiency, but cannot be restarted after a power interruption.
Although the fan's performance is greatly reduced, the failure may not be
detected unless there is a monitor of air flow or pressure drop across the
a radon monitor.

-------
This failure mode was discussed at the EPA Radon Symposium in February
1990: mitigators were specifically asked about their experience with fan
failures. Most mitigators have experienced some failures, but this EPA pro-
ject had a failure rate far higher than that experience by these mitigators.
A distributor who sells over 700 fans per month for radon mitigation reported
that less than 1% fail within the 3-year warranty period. Failures may be due
to either bearings or capacitors, but bearing failures are more noticeable
because the fan begins to produce more noise. Several mitigators reported
that fan failures seem to occur within a few months rather than after a year
or more.
SYSTEM MONITORS
As mentioned above, drain lines from water traps may freeze in unheated
spaces. A similar failure mode exists when condensation accumulates and
freezes in the tubes which connect a pressure monitor or switch to the duct.
If either tube is blocked, the switch or monitor will not function properly.
System monitors which are electronic or which trigger an electrically
powered alarm should be wired to a different circuit than the system itself.
SEALANTS
Most mitigation systems involve some sealing of floor/wall joints as well
as of cracks in a slab or wall. Unless the surface is properly prepared, the
sealant will not adhere to it. Even with proper preparation, an appropriate
sealant must be used. For example, silicone caulk will not stick to concrete,
but urethane will. Any sealant used for radon mitigation should last as long
as the house. While not technically a sealant failure, it is not uncommon for
new cracks to develop in a slab or wall after mitigation. It may be that the
drying of soil by a mitigation system stimulates cracks.
Ducts are usually constructed from sections of polyvinyl chloride (PVC)
or acrylonitrile-butadiene-styrene (ABS) pipe. PVC pipe can be glued, but ABS
pipe must be caulked. It is important that joints fit snugly and be
thoroughly cleaned, and that an appropriate adhesive be used to ensure a
permanent seal. Metal ducts are a special problem. The joints which are near
a fan may be subjected to considerable vibration. The fan should be connected
to the duct with rubber couplings to reduce vibration and provide a better
seal between the fan and the duct.
PIPES
Since plastic pipe is readily available and easy to work with, it is
probably the most common duct material. Some plastic, however, is affecte<^ ^y
sunlight; it becomes brittle and more susceptible to impact damage. Only
plastic pipe stamped "DWV" (drain, waste, vent) should be used outdoors unless
it will be insulated or otherwise protected from sunlight.
-------
RESIDENT activities
intentional actions
• ¦ i , after paying hundreds of dollars for mitigation systems.
Surprisingly, atter P * *	the most common reasons are to save
some people turn them otr.	resident thinks that radon is only a
energy or to eliminat-e noi	^ ^ ^ may fcurn off the mitigation
problem during the heati 9	especially if windows are left open (5).
system during the warmer	^ typical mitigation system fan uses less
Often people do not realiz	lb. 0ne mitigator felt that renters had a
electricity than a 100 w ixg ^ th&n homeowners and were more likely
I itigaticn .yt--. <»«•"¦* —
d svstems which were turned off by new owners
Several mitigators repor	^ 0ne new owner had been told that the
who did not understand their p	'of gewer gas . Another had been advised
system was intended t^cont	^ unnecessary.
by the realtor that cn y
unintentional actions
ted that residents had temporarily turned off
Several mitigators rep	^ Acoustic or electrical noise
systems and forgotten to tu	0ne mitigator reported that a system was
seemed to be the most c°mm	e the fan noise interfered with conversation,
turned off during a party	inter£erence with radios and television. Some
There were several reports	or elecCriCal components of the mitigation
of these were due to fau y	realize that the system could be fixed or
system. Often residents d	^ noise. Rather than call the mitigator.
adjusted to reduce or elim	^ ^ particulariy offensive (6) .
they turned the systems off when
1iance mitigation systems which are plugged into an
Like any other appnar ^enfcally unpiugged. If the system does not make
electrical outlet can be acci	fcake some time for a resident to realize
much noise and has no alar . ^ probably a design failure, stimulated by the
that it is not running.	electrician and possibly an inspection. Radon
desire to avoid the cost o	^ go that they cannot be accidentally un-
mitigation systems	mitigators as to whether it is better to use
plugged. Opinions	pxisting circuit. Some felt that a separate circuit
a dedicated circuit or an	erence with a radio or television. Others
would minimize electrical *	circuit used for lights or appliances would
felt that tapping into an	^ ^ mitigation system were
it mor. notic.abl. if the P
interrupted.
home renovation or remodeling
• • t-ors contacted warned homeowners that a mitigation
Many of the mitigator	by some typicai home renovation or remodeling
system may be adverSf* DlaCing the heating/cooling system, making an addi-
projects. These inclu*®	the basement. One EPA contractor reported
tion to the house, ^urization system in a crawlspace had been severely
that a submembrane depressur
-------
damaged by workmen replacing a furnace. Although the contractor had provided
a walkway to the furnace, apparently the workmen had dragged the old unit out
across the membrane, damaging it.
CONCLUSIONS AND RECOMMENDATIONS
The experiences related in this report show that residential radon miti-
gation systems do fail for a variety of reasons and that such failures may not
be immediately recognized. Mitigators should design systems to minimize the
probability of failures. The system design should include a monitor which
residents can use to determine whether the system is operating properly.
Homeowners must realize that even systems with good design and components have
some chance of failure; they should check the system monitor periodically and
measure radon levels annually as long as the house is occupied.
SYSTEM MONITOR FOR THE HOMEOWNER
Only a few mitigators reported using system monitors with which they were
satisfied; one had personally designed and built the monitor. Some research
and development of a suitable monitor for residential radon mitigation systems
is needed. The monitor need not have high resolution as it will not be used
to monitor minor variations in system performance. It need only be capable of
detecting change by a factor of 2 or more. An ideal monitor would have the
following characteristics:
•	The monitor should be inexpensive so that there is little in-
centive for mitigators to omit it to cut costs. It could monitor
the system operating parameters (e.g., pressure drop) rather than
radon concentrations. Such monitors are 2 orders of magnitude
less expensive than the least expensive continuous radon
monitors.
•	The monitor should be adjustable so that the mitigator can set it
for the system installed in that house. Mitigators may want to
check the settings after a break-in period; two mitigators
mentioned that flow rates tend to increase and pressure drops
decrease over the first few weeks after system start-up.
•	The monitor should be simple enough to be useful to the vast
majority of residents. Several mitigators reported that most
people do not check monitors when they are provided. Some of
those who do check their monitors call the mitigator about minor
fluctuations.
•	The monitor should be durable. It should not require any adjust-
ment by the resident, who should be able to test whether it is
functioning properly. Several mitigators said that many of the
reports of mitigation system failure to which they responded were
actually failures of the system monitor.
-------
SYSTEM DOCUMENTATION FOR THE HOMEOWNER
It is essential that residents understand the basic principles of the
mitigation system and how to interpret the system monitor. If residents are
to avoid activities which could compromise mitigation systems and to recognize
problems when they occur, they should receive verbal explanation and instruc-
tion when the system is installed, as well as written documentation which they
may refer to in future years or pass on to a new owner if the house is sold
Such documentation should include:
•	Radon concentrations before and after mitigation. The measure-
ment method, duration, and time of year should be documented.
•	A description of the principles and specifications of the mitiga-
tion system. The basic principle of operation could be taken
from EPA's homeowner's guide to radon reduction methods (7). The
location of ducts, wires, fans, switches, and the system monitor
should be sketched or described. System operating parameters
(e.g., pressure drop and air flow) after a break-in period of at
least 24 hours should be available.
•	An explanation of the system monitor. This would include whether
the monitor indicated air flow or pressure drop, and the nominal
range for the indicated parameter. If there is an audible or
visual alarm, conditions that trigger it and what to do if the
alarm goes off.
|
•	A scheidule and procedure for periodic inspections. This might
simply be to check the monitor monthly.
•	A description of any preventive maintenance and of the warranty
on any components (e.g., the fan) or on the system as a whole.
Homeowner or resident activities that might void the warranty
should be listed. Who should be called if there is a problem
should be identified.
•	The appropriate state or local health department to contact in
case of a problem that cannot be resolved by the original miti-
gator.
•	A discussion of the sensitivity of the system to typical home
remodeling or renovation projects.
•	The importance of measuring radon concentrations annually as long
as the house is occupied, even when the mitigation system appears
to be operating normally.
•	A short, simple summary of all of the above.
This may seem like a tremendous burden for a commercial mitigator but
most of them are already providing such documentation. An EPA survey of com-
mercial mitigators (8) found that over 80% prepare a written mitigation plan
and give a copy to their clients; over 60% provide clients with written in-
structions on how to maintain the systems.
-------
The EPA might develop model documentation which could be copied or
modified by commercial mitigators. Most of this documentation could be
"boilerplate" which should be easy to assemble for each mitigation technique,
with blanks to fill in specifics like radon concentrations and operating para-
meters. It is essential that the documentation be written so that most
residents can understand it; otherwise the mitigation system will remain a
"black box." The homeowner or resident will not feel competent to monitor its
operation and may not appreciate the need for long-term follow-up radon
measurements.
In addition to the documentation described above, the mitigation system
should be clearly and permanently labeled with a warning that it is a radon
mitigation system, that it protects the residents' health, and that residents
should measure radon annually. The label should also identify whom to contact
if a problem is identified or suspected.
LONG-TERM FOLLOW-UP
Based on the experiences of the mitigators contacted, few homeowners or
residents recognize the potential for failure of their radon mitigation
system. When a system monitor is provided, they do not check it regularly.
When radon detectors are provided during subsequent heating seasons, they do
not expose them. Like any mechanical system, radon mitigation systems are
subject to failure. Some way to communicate this fact to current and future
residents must be found.
A study involving long-term follow-up radon measurements in a national
sample of mitigated houses could show the rate of mitigation system failures.
Publicity about such a study might inspire many people to check the
performance of their mitigation systems.
REFERENCES
1.	Henschel, D.B., U.S. Environmental Protection Agency, personal
communication, January 29, 1991.
2.	U.S. Environmental Protection Agency. The national radon contractor
proficiency (RCP) program: proficiency report. EPA-520/1-90-004 (NTIS
PB90-204744), Office of Radiation Programs, Washington, DC, April 1990,
108 pp.
3.	Henschel, D.B. Radon reduction techniques for detached houses: technical
guidance (second edition). EPA-625/5-87-019 (NTIS PB88-184908), U.S.
Environmental Protection Agency, Research Triangle Park, NC, January 1988.
4.	Scott, A.G., and Robertson, A. Long-term performance and durability of
active radon mitigation systems in eastern Pennsylvania houses. Presented
at the 1990 International Symposium on Radon and Radon Reduction
Technology, Atlanta, GA, February 19-23, 1990.
-------
Harrje, D.T., et al. Long term durability and performance of radon
mitigation subslab depressurization systems. Presented at the 1990
International Symposium on Radon and Radon Reduction Technology, Atlanta,
GA, February 19-23, 1990.
Prill, R., Fisk, W., and Turk, B. Evaluation of radon mitigation systems
in 14 houses over a two-year period. J. Air Waste Manage. Assoc. 40:
740-746, 1990.
U.S. Environmental Protection Agency. Radon reduction methods: a
homeowner's guide. OPA-87-010, office of Research and Development,
Washington, DC, September 1987.
Hoornbeek, J., and Lago, J. Private sector radon mitigation survey.
Presented at the 1990 International Symposium on Radon and Radon Reduction
Technology, Atlanta, GA, February 19-23, 1990.
-------
IVP-2
MITIGATION BY SUB-SLAB DEPRESStTRIZATIQN UNDER STRUCTURES
FOUNDED ON RELATIVELY IMPERMEABLE SAND
by: Donald A. Crawshaw, M.S. (Civ. Eng.), MICE
Geoffrey K. Crawshaw
Pelican Environmental Corporation
Wellesley, Massachusetts 02181
ABSTRACT
Effective sub-slab depressurization requires the pressure
field to cover the entire area under the slab. This is readily
achieved by means of a low pressure, high flow fan system when the
sub-slab material is permeable crushed stone or gravel. However,
the occurrence of relatively impermeable sub-slab sand presents
the mitigator with a number of problems to overcome. Traditional
solutions have included using multiple suction points, digging
pits and filling them with permeable material and using more pow-
erful in-line fans. Such solutions can not always be used, and may
be ruled out by aesthetic considerations, particularly when the
mitigation work has to be located in a part of the structure that
is fully furnished.
The paper documents results from using a high pressure, low
flow (hplf) fan system that has been developed to address these
problems, and successfully used to mitigate radon levels, in vari-
ous structures founded on relatively impermeable sand.
-------
INTRODUCTION
Active sub-slab depressurization (SSD) has proved to be an
effective means of collecting radon in soil gas from beneath slabs
in contact with soil. An active SSD system consists essentially of
a fan connected to a piping system that collects soil gas from
beneath the slab for venting above the roof line. The slab acts as
a membrane to form the upper boundary of the required sub-slab
pressure field. Ideally the pressure field should cover the total
area under the slab and should also extend under the exterior
wall/floor joint, this being a usually significant radon entry
route.
The type of material immediately under the slab is an important
factor governing the design of every SSD system. Crushed stone
aggregate under a slab is relatively permeable and typically
requires a centrifugal type blower that can move soil gas in some
volume (125 to 270 cfm in free air) and generate a maximum static
pressure of less than 2 inches WC. On the other hand, sand or dirt
under a slab is relatively impermeable in comparison to crushed
stone aggregate and requires a fan that can generate considerably
greater suction pressure than 2 inches WC to move a lesser volume
of soil gas (1).
Traditionally, effective sub-slab depressurization in sand or
dirt has required breaking into the slab, excavating a large
amount of sand, replacing it with crushed stone and even with per-
forated PVC piping, and then recasting the slab. Only then can a
SSD system with a centrifugal type blower be used to extract the
soil gas. This extensive construction work may disturb the occu-
pants, particularly if the work is to be done in the furnished
part of the building. The difficulties and costs associated with
this method led to the development of the patent-pending Pelican
hplf soil gas reduction system for SSD under structures founded on
relatively impermeable material (2).
-------
CLASSIFICATION OF SUB-SLAB MATERIAL
Evaluation of the communication of suction pressures through
the sub-slab material between various test holes is a well known
diagnostic technique used for designing SSD systems. After con-
ducting diagnostic evaluations for many structures, it became ap-
parent that additional data to help in classifying the sub-slab
material can be collected using the same vacuum equipment, hosing
and pressure gauges that are used for the communication tests.
This entails taking two readings of suction pressure at each test
hole with the vacuum equipment operating at full suction:-
(i)	with the end of the hose in air
(ii)	with the end of the hose tightly inserted in the test hole.
The net difference between these two pressure gauge readings gives
the Pelican Permeability Number (PPN). Permeability of a soil is a
property that determines the rate of flow through the soil and the
PPN is a simple measure in inches W.C. of the resistance to air
flow of the sub-slab material subjected to suction pressure ap-
plied at the test hole. Figure 1. shows the results obtained from
numerous tests of sub-slab material encountered in Massachusetts
with a standard 2.25 HP Sears Wet Dry vac having been used to gen-
erate the suction at the test hole.
18"l
16-
14-
12-
PPN 10-
8-
6-
4-
2-
Stone Gravel Coarse Sand
Fine Sand Very Fine Sand
Material beneath Slab
Figure 1. PPN soil classification of material beneath slabs (1)
-------
Visual inspection down the test holes may provide additional
confirmation of the class of material, but the PPN value is an in-
situ test result that takes into account such variables as par-
ticle size, grading and lamination of the soil that are not appar-
ent to the'naked eye. The test can be repeated at a number of test
holes during the diagnostic evaluation. With this method, PPN val-
ues can be readily obtained by a diagnostician without the need of
special permeameter equipment and the recorded values are meaning-
ful for the designer of the SSD system in selecting the required
type of fan to be used. It is recommended that diagnosticians
should construct their own soil classification charts for the sub-
slab material which they encounter in their locality, using their
vacuum equipment.
The paper covers results obtained with 56 HPLF systems that
were installed to reduce radon levels where the sub-slab material
was relatively impermeable in comparison to crushed stone or
gravel. The paper does not deal with SSD where the sub-slab mate-
rial is clay .
DESIGN OF HPLF SYSTEMS FOR RELATIVELY IMPERMEABLE
SUB-SLAB MATERIAL
PIPE SELECTION
EPA's Reducing Radon in Structures Manual (3) includes a de-
sion guide for soil depressurization in various types of sub-slab
material. A minimum pipe diameter of 1 1/2" is suggested in the
manual where the material under the slab is sand. In practice,
this 1 1/2" diameter piping has proved to be very suitable for
typical residential applications, particularly in finished living
areas as the piping can be run inconspicuously along beams, in
suspended ceilings, behind dry walls, and in closets. Installation
of the piping is further facilitated by using thick-walled
flexible piping to negotiate awkward bends. Two inch diameter pip-
ing is used where the pipe runs are lengthy or in offices or
schools where the piping is potentially vulnerable to damage as a
-------
result of the large number of people using the building. All pip-
ing is Schedule 40 PVC.
BLOWER SELECTION
The PPN value is useful for determining the type of blower to
be used. The Pelican HPLF system was selected for PPN values be-
tween 3 and 16. Figure 2 shows the fan curve of a S-3 blower that
was used in 42 of the 56 projects described in this paper.
Powtr Curvt
FlowCuiva
Flow (CFM)
Figure 2. Air flow vs. vacuum pressure and power for S-3 blower
The EPA Manual (3) cites certain criteria that are important in
fan selection and which were addressed under the following head-
ings : -
a. Mr flow
in the normal operating zone, the air flow moved by the S-3
blower in sand is 19 to 26 cfm, which is low in comparison to that
moved by SSD systems in crushed stone or gravel with a centrifugal
fan. (It is also low in comparison to a typical natural infiltra-
tion rate of more than 100 cfm for basements.) This meets the de-
sign requirement to minimize the amount of air that the SSD system
can potentially remove from the house so as to minimize energy
-------
bills during the heating and cooling seasons and to avoid the risk
of downdrafting and spillage from combustion devices.
b.	Maximum static pressure
The typical air flow from HPLF systems using this
blower,where the sub-slab material falls between dirt to coarse
sand ( as shown on Figure 1.), has been found to be in the range
of 19 to 26 cfm. These operating conditions correspond to a vacuum
pressure range of 14 to 4 inches WC. The maximum static pressure
of 26 inches WC at 0 cfm air flow has proved to be sufficient for
most residential applications in sand or dirt.
c.	Electric power consumption
In the typical operating zone, the power curve in Figure 2
indicates power consumption of approximately 165 watts, when oper-
ating at 7.5 inches WC vacuum pressure. For an electric power cost
of 10 cents/KWH this amounts to a monthly cost of $12.05. This
cost can be offset against the reduced energy costs during the
heating and cooling seasons as compared to a higher energy costs
for a standard centrifugal blower used in SSD systems in crushed
stone that may remove considerably more air from the house.
d.	Noise
The blower housings are lined with industrial soundproofing.
The blowers are often installed in attics and the soundproofing
enables them to be located even directly over bedrooms. The 4 inch
diameter Schedule 40 PVC pipe that discharges effluent from the
blower to atmosphere above the roof line has been acoustically
designed as a muffler. When the blower is suspended from the
structure, a vibration isolator is used to eliminate any low fre-
quency vibrations from entering living areas (4).
e.	Long service life
To meet the design retirement of a long service life, Peli-
can hplf systems incorporate a special housing so that the blowers
run in a temperature stable environment. The S-3 blowers have
-------
double sealed bearings that minimize maintenance.
f.	Ease of installation
Installation of the blower is facilitated by customized in-
door and outdoor hanging kits and other accessories. Clic hangers
are used to install the 1 1/2" diameter Schedule 40 PVC piping.
Heavy duty 1 1/2" flexible PVC piping can replace multiple bends
and reduce air flow friction losses at the bends; it is glued into
standard PVC fittings (4).
g.	No leaks from blower housing
The blower housing is under negative pressure to ensure a
"safe-leak" design; this ensures that a leak in the housing will
suck air into the fan. The blower is mounted in attics out of liv-
ing areas or outside the structure.
h.	Moisture resistance
The blowers are weatherproofed and can be installed outdoors,
being totally encased in the cylindrical housing shell.
NUMBER OF SUCTION POINTS
The footprint area of each structure is useful information
for the designer in estimating the number of suction points to be
used. Under favorable conditions, the pressure field generated by
the S-3 blower can cover up to 1000 square feet in fine sand but
it is prudent to assume coverage of 500 square feet per suction
point in such material. One suction point should be near the cen-
ter of the footprint that is to be covered by the pressure field.
The final choice of number and location of suction points should
be left to the installation crew as they may gather additional
soil data after core drilling through the slab on the day of in-
stallation. The extension of the pressure field must be checked
with the system in operation as additional suction points may be
required and can be readily added by means of extending the 1 1/2
inch diameter piping system at that time.
-------
CONDENSATE CONTROL
Pipe runs must be sloped so that condensate will always
gravitate back to drain under the slab. The higher vacuum that is
required for HPLF systems in relatively impermeable material works
against the condensate, which is draining under gravity, so more
slope is needed on the drainage pipes. When designing the piping
layout, it is necessary to designate a drain point and then divide
the piping network into drain and non-drain zones (4) . This is
CONDENSATE BYPASS
The condensate bypass arrangement around the blower is shown
in Figure 4. It is designed to move condensate from the 4 inch
diameter effluent stack to the intake piping where it can safely
drain back to beneath the slab. This prevents condensate from run-
ning back into the blower or from forming a slug which would block
the effluent exhaust (4).
To Slower
To dlower
i
Drain zone piping [ft 41
\ r&g
Non-drain zone piping
Drain zone piping
Non-drain zone piping
Drain Point
illustrated in Figure 3
Figure 3
Condensate zones in piping network
-------
Seal with silicone
Figure 4. Condensate bypass arrangement in attics
INSTALLATION
Closure of any openings in the slab and some sealing is done to
improve the integrity of the slab and to enhance the sub-slab
pressure fields of each active HPLF system, but not to act as a
primary mitigation system. In the installations described in the
paper, major cracks discovered in the slabs during the diagnostic
evaluation were sealed with polyurethane. Only unusually wide
wall/floor joints were similarly sealed.
Effective slab penetrations are important in order to extend
the sub-slab pressure fields and thereby achieve maximum radon
reductions. Suction pressures should radiate horizontally through
the sub-slab material so five inch nominal diameter holes were
core drilled through the floor slabs to allow easy hand excavation
of a plenum under the slab at each suction point. Two to five g»1-
-------
Ions of sub-slab material were excavated to form each cavity, with
the larger amount being removed when the material was dirt and the
lesser amount when the material was coarse sand. The end of each
suction pipe was securely covered with aluminum insect screen to
prevent sand from rattling in the lowest part of the pipe (4) .
Figure 5. shows a typical installation where the blower was lo-
cated in the attic.
Exhaust
To drain point
To drain point
Figure 5. Typical Pelican HPLF attic installation.
In cases where there was more than one sub-slab suction point,
the pipes were manifolded into a single pipe which was typically
routed through the side of the basement foundation wall, up the
outside of the house and into the attic through the gable. (The
piping was run up through closets if they lined up from the base-
ment to the attic.) The blower housing was connected to the 4 inch
diameter exhaust muffler which vented the effluent through the
roof to atmosphere.
All electrical connections of the blower to the power supply
-------
were made in accordance with the Massachusetts electrical code by
a qualified electrician .
Dwyer U-tube manometers were fitted on the HPLF piping sys-
tems in locations convenient for the homeowner to inspect.
Make-up air was ducted to the proximity of the furnace to
supply air for combustion and to guard against the possibility of
backdrafting flue gases into the basement, wherever this was a
concern.
With the mitigation systems operating, sub-slab pressure test-
ing was performed to determine the extent and strength of the
negative pressure fields beneath the slab.
RADON RETESTING
On completion of the work, and after the mitigation systems
had been operating for at least two days, radon concentrations
were retested with two charcoal vials exposed for two days. Retest
locations were typically in the basement and on the first floor
levels. The homeowners mailed the vials to Niton Corporation for
testing and analysis.
In those cases where the work was done for clients such as
relocation companies, in addition to retesting with canisters,
retesting with a continuous monitor was carried out by Radonics,
inc.
RESULTS
The paper deals with 56 of the Pelican HPLF systems that have
been installed. In 42 of them the S-3 blower was used and, for
various reasons, HPLF blowers with different fan curves were used
in the other 14 homes. Most of the HPLF systems required two suc-
tion holes, one of the holes preferably being near the center of
the slab and the other being located near the footing of the foun-
-------
dation wall for the purpose of draining condensate. The 56 homes
with HPLF systems had an average footprint area of 1115 square
feet with an average footprint area per suction point of 500
square feet.
The PPN value was recorded for 27 of the installations and
generally ranged from 6 (coarse sand) to 15 (dirt) with an average
of 10.9 (fine sand). One installation was in gravel with a ppn of
3.
Short term retest results showed that the radon concentra-
tions in the basements of the 56 homes were reduced by an average
of at least 96.4%. The words "at least" are used because the low-
est retest values were taken to be 0.4 pCi/L, having been reported
by the laboratory as <0.4 pCi/L. The average pre-mitigation base-
ment radon concentration was 19.8 pCi/L and the average post-
mitigation basement radon concentration was 0.72 pCi/L. The
highest retest result in a basement was found to be 1.8 pCi/L. and
76% of the basement retest results were below 1 pCi/L.
Manometer readings recorded for 43 of the installations had
an average value of 6.2 inches W.C. with a maximum value of 16.5
inches and a minimum value of 1.0 inch. Manometer readings for the
S-3 blower averaged 5.6 inches WC with a maximum value of 14.3
inches and a minimum value of 1.4 inches.
Table 1. lists data obtained in 20 HPLF installations during the
diagnostic evaluation as well as the associated manometer reading
with the system operating.
TABLE 1.	DIAGNOSTIC DATA AND MANOMETER READINGS.
Applied vacuum, inches W.C.
Pelican Permeability Number
Sub-slab pressure at 10ft
Smoke test 0 (none) - 3 (greatest)
System Vacuum, inches W.C.
Applied vacuum, inches W.C.
Pelican Permeability Number
Sub-slab pressure at 10ft
Smoke test 0 (none) - 3 (greatest)
System Vacuum, inches W.C.
1
2
3
4
5
5
7-
8
9
10
47.5
13.5
0.250
2.0
15.0
47.0
11.5
0.005
?
10.5
47.0
10.5
0.003
0.0
4.5
47.0
12.0
0.100
2.0
11.0
45.5
10.0
0.250
3.0
3.0
45.0
12.0
0.250
3.0
4 1
44.0
14.0
0.010
0.0
7.8
43.0
11.0
0.130
2.0
3.7
43.0"
14.0
0.062
2.0
1.8
42.0
12.0
0.004
0.0
5.7







11
12
13
14
15
16
17
18
15"
?(r
42.0
10.0
0.120
3.0
10.5
42.0
10.0
0.007
2.0
2.1
41.0
10.0
0.155
3.0
3.8
40.0
14.0
0.095
1.0
5.3
40.0
10.0
0.020
1.0
6.5
37.0
3.0
0.003
1.0
2.5
36.0
9.5
0.002
1.0
1.2
35.0
7.0
0.025
3.0
14.3
33.0_
7.5
0.007
1.0
1.5
33TT
6.5
0.007
?
1.3
-------
Figure 6. charts the diagnostic data and the manometer readings
for the 20 HPLF installations in Table 1.
Vacuum
Inches
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Figure 6. Diagnostic data and manometer readings for 20 HPLF
installations.
Figure 7. charts the diagnostic data in Table 1 and the sub-
slab pressure at 10 ft from the test hole at which the vacuum was
applied, during the diagnostic evaluation.
100
Vacuum
PPN
10V
Inches 1
6
~
«
.~
~ .
0.01 - -
Sub-slab pressure at 10 feet from vacuum
0.001 -L
Figure 7. Diagnostic data and sub-slab pressure at 10 ft from
applied vacuum for 20 HPLF installations.
-------
DISCUSSION
Mitigation of radon concentrations in homes founded on rela-
tively impermeable material can be achieved in a number of ways.
Verified reduction of the radon concentration is of prime impor-
tance but coupled with this requirement, the owner of the home or
building has other important needs that must be addressed by the
mitigator for successful completion of the mitigation contract.
For mitigation by any sub-slab depressurization system, these con-
Riderations include noise reduction , visual impact, condensation
control, acceptable running costs, reliability and longevity of
the blower. The Pelican HPLF System was designed to meet all these
requirements and has proved to be effective for sub-slab mitiga-
tion of radon concentrations in structures founded on sand or dirt
in Massachusetts and New Hampshire.
Basement radon concentrations were reduced to below 2 pCi/L
in all of the 56 homes, and 76% of them were reduced to less than
1pCi/L, despite the fact that the slabs rested on such relatively
impermeable material.
For a single-storey Federal building, which had an addition
founded on sand, the S-3 blowers were located on the flat roof.
Aloha-track retest results showed that the radon retest results
were all less than 1.0 pCl/L in the office area. This project is
not included in the results discussed in this paper.
For structures founded on sand or dirt, the designer has the
ootion of selecting a HPLF blower coupled with 1 1/2 inch piping,
this is particularly useful where the system has to be installed
in a furnished part of the dwelling, such as a fully finished
basement. The sound proofed housing enables the S-3 blower to be
installed near living areas, even directly over bedrooms in at-
tics, without disturbing the occupants.
Placing suction points away from the center of the slab near
the wall/floor joint can result in "bypassing ¦, where basement
air rather than soil gas is collected by the pressure field.The
initial Pelican HPLF installation was carried out with four
-------
suction points located next to the middle of the four basement
foundation walls in a house founded on dirt. The radon retest
results were acceptably low but a considerable amount of piping
was used. On subsequent HPLF installations, it became apparent
that it was preferable to locate one suction point nearer the
center of the footprint.
The PPN value has been found to be useful for quickly and
simply identifying which blower system should be selected for sub-
slab depressurization. It can be a useful number for broadly clas-
sifying the sub-slab material, particularly for discussing the
project and blower selection with people who were not present at
the diagnostic evaluation. No apparent relationship was found be-
tween the PPN value and the manometer reading except that the PPN
usually exceeds the installed manometer reading.
Smoke tests were not found to be a satisfactory indicator of
sub-slab communication when the sub-slab material was fine sand or
dirt. This is because the relatively impermeable nature of the
sub-slab material obstructs the flow of smoke, in a number of
cases, although the smoke test was inconclusive, the PPN value
indicated that HPLF sub-slab depressurization was a suitable miti-
gation method.
in the HPLF installations reported in this paper, no apparent
relationship was found between the PPN value and sub-slab communi-
cation pressure test results at 10 feet distance from the point of
suction with vacuum applied to a 3/4" diameter inspection hole. It
appears that it is more reliable to base the choice of the blower
on the PPN value than on the sub-slab communication test result at
10 feet when the sub-slab material is sand or dirt.
When using blowers with higher suction pressure, it is very
important to slope piping correctly to enable condensate to be
effectively drained to beneath the slab.
The work described in this paper was not funded by the
U.S. Environmental Protection Agency and therefore the
contents do not necessarily reflect the views of the
Agency and no official endorsement should be inferred.
-------
REFERENCES
1.	Crawshaw, d.a. and Crawshaw, G.K. Mitigation of elevated VOC
concentrations in buildings arising from contaminated groundwater,
in: Proceedings of HMC- Great Lakes '90. Hazardous Materials Con-
trol Research Institute, Cleveland, Ohio. September 26-28, 1990.
2.	Crawshaw, G.K. Pelican HPLF System, radon solution for slabs
poured on sand or dirt. Pelican Environmental Corporation,
Framingham, MA. 1990.
3.	O.S. Environmental Protection Agency. In: Reducing radon in
structures manual. 2nd. Edition, 1989. Considerations in mitiga-
tion system design and installation, p. 5-23,24
4.	Crawshaw. ..K. In: installation and Design Guide, Pelican HPLF
System. Pelican Environmental Corporation, Framingham, MA. 1990
-------
IVP-3
TITLE: A Laboratory Test of the Effects of Various Rain Caps on
Sub-Slab Depressuirzation Systems
AUTHOR: Mike Clarkin, Camroden Associates, Inc.
This paper was not received in time to be included in the
preprints so only the abstract has been included. Please check
your registration packet for a complete copy of the paper.
ABSTRACT
Many sub-slab depressurizatlon systems are Installed wtth some type of rain
cap intended to Keep rain water from entering the exhaust pipe. In order to determine
the effect these rain caps have on the pressures developed in the sub-slab
depressurization system, a series of tests were performed to determine: 1. the static
pressure losses associated with the use of the rain caps, and, 2. the effect of wind on
the system with the various rain caps installed. The results of these tests are
presented in this paper.
-------
I VP
ANALYSIS OF THE PERFORMANCE OF A RADON MITIGATION SYSTEM BASED DN
CHARCOAL BEDS
P. Wasiolek, N. Montassier, P.K. Hopke
Clarkson University
Department of Chemistry
Potsdam, NY 13699-5810
R. Abrams
RAd Systems, Inc.
Westboro, MA 01581
ABSTRACT
The performance of a radon mitigation system based on adsorption of
radon onto charcoal beds (RAdsorb system) combined with an electronic
air cleaner (EAC) installed in a single family house in Shrewsbury, MA
was studied in a series of tests. Semi-continuous measurements were
made of the radon gas concentration, potential alpha energy
concentration (PAEC), particle concentration with size distribution and
radon decay product activity-weighted size distribution with and without
additional aerosol sources. The instruments used were a radon gas
monitor (EBERLINE, RGM-3), WL-meter (Thomson & Nielsen), and a
differential mobility particle sizer (DMPS) by TSI. For measurements of
the activity size distribution, an Automated, Semi-continuous Graded
Screen Array (ASC-GSA) developed at Clarkson University was utilized.
During the time of tests, the conditions in the basement of the house,
without the mitigation system in operation, were as follow: radon
concentration up to 800 Bq nf3, PAEC up to 650 nJ m'3 (30 mWL) , particle
concentration below 1000 cm"3, and the fraction of PAEC and ziePo in the
smallest size range 0.5- 1.58 nm was up to 0.6 and 0.9, respectively.
The tests were designed to study the influence of the combined system as
well as the separate components of the mitigation system: fan, charcoal
bed and EAC on the all of the measured parameters. When all the
components of the mitigation system were working, the achieved
reductions were radon concentration below 150 Bq m 3 (4 pCi L""1) and PAEC
below 100 nJ m"3 (5 mWL) with the smallest sized fraction of PAEC (0.5-
1.58 nm) of about 0.4. The tests proved that under certain conditions,
the charcoal bed/EAC mitigation systems can be a potentially valuable
technique for reducing a health risk due to indoor radon.
-------
INTRODUCTION
Inhalation of the short lived decay products of radon (222Rn):
218Po, 21APb, 21ABi and 21APo, is thought to be the second largest cause of
lung cancer after cigarettes smoking. To reduce this potential risk, it
is presently recommended that the remedial measures should be taken when
the level of radon gas in a home is found to exceed 150 Bq nf3 (4 pCi L~
x) (1). Several mitigation methods have been tried in houses with
elevated radon levels. These techniques might be divided into two main
categories:
a)	Ones based on the reduction of the radon entry rate into the
house, that sometimes required changes in a house construction or
house modification e.g." subslab ventilation", "crawl space
ventilation",
b)	Others based on the removal of radon from indoor air
(ventilation, filtration, radon adsorption).
The RAdsorb system built by the RAd Systems Inc. is a carbon
adsorption system. The system has been installed in a single family
house in Shrewsbury, MA. The RAdsorb radon removal system is based on
activated carbon adsorption of radon. A radon gas removal efficiency
evaluation was performed by the producer yielding values up to 97%
reduction in radon gas concentrations (2). In addition, for this study,
an electronic air cleaner (EAC) (Honeywell Model F50E) has been added to
the RAdsorb system. The influence of the operation of the RAdsorb
system on the indoor radon and its decay products concentrations (PAEC)
and activity weighted size distributions are important from the health
risk point of view and were the objective of measurements made in this
house during September 1990.
HEALTH RISK DUE TO INDOOR RADON
The health risk associated with radon in indoor air is not from
radon itself but rather from radon's short lived decay products. Radon
as an inert gas with a half-life of more then 3 days may be inhaled and
subsequently exhaled with little decay while in the human lung. The
decay products of radon, however, are reactive and when inhaled, may
deposit within the lung. Since they have short half-lives, further
radioactive decay will occur prior to particle clearance from the
respiratory tract. The alpha energy emitted during decay is therefore
fully deposited in the lung tissue, possibly causing damage to the DNA
within the target cells. If the DNA is damaged, the abnormal cell may
reproduce and may result in a cancer. The deposition of the radon decay
products within the lungs depends to a great extent on the attributes of
the particles to which it is attached. The efficiency of deposition of
particles in the lung varies with the particle size and hence, knowledge
-------
of the particle size distribution and the activity size distributions
are important in evaluating the risk attributed to radon progeny. The
fraction of radon progeny atoms or ions possibly clustered with other
molecules such as H20 is traditionally defined as the "unattached"
fraction. The most recent studies strongly suggest that so-called
"unattached" fraction is actually an ultrafine particle mode in the
0.5 - 3.0 to 5.0 nm size range (3). In the absence of active particle
sources, the radon decay product activity size distribution may be
thought of as bimodal, with a fairly sharp small-diameter mode near the
molecular size corresponding to the "unattached" fraction and a broader
large-diameter mode corresponding to the "attached" fraction (4). Two
physical parameters used in all lung dosimetry models estimating
radiation doses from inhaled radon decay products, are the activity
median diameter of the "attached" radioactive aerosols and the
"unattached" fraction of 218Po. The 218Po is of particular interest
because it is the first short-lived decay product in radon chain with a
half-life of only 3.1 minutes.
The dosimetric calculations for evaluation of the absorb dose in
lung tissue per unit exposure suggest that the dose per unit exposure
from the "unattached" fraction could be up to 25 times higher then that
for the "attached" fraction (5).
In the most recent dose estimates (6), particle size has been
taken into consideration. The basal cell and the secretory cells in the
bronchial epithelium were considered as target cells. The resulting
dose conversion factors per unit exposure from monodisperse activity Di,
are presented in Figure 1 as a function of breathing rate.
5^
O
H
V
o
o
a
o
%
o
o
o
Q
500
400
V 300
200
100
"" 1 	 		
r
3 -1
, v 	 0.45
m hr
~ ' V — 1.6
m hr
/ ^
3 -t
^ - - 3.0
\
m hr
A\ v

/ \ \

\ \

7 \ v

\ V

\ V

\ \


-
y \v




y
		.				 TT-rrrr^Mw
mmmm—r
1	10	100	1000
Particle Diameter (nm)
Figure 1. Dose to bronchial secretory cells as a function of the size of
radon decay products for an adult male (6)
-------
The graph shows the dose to secretory cells for three different
breathing rates equivalent to resting, light activity and heavy work and
that for all cases the conversion factor is strongly dependent on the
activity median diameter especially for particles smaller then 10 run.
Therefore, to calculate the dose per unit exposure to secretory cells,
the following formula applied:
-------
CLEAN AIR RETURNED TO ROOM
RADON GAS
<$=J IN
ROOM AIR
RADON GAS
TO
-------
HOUSE CHARACTERISTIC
The house consists of a living room and kitchen on the first floor
and three bedrooms on a second floor. The initial concentration of
radon in house basement before mitigation ranged up to 1100 Bq m 3 (30
pCi I/1) . The RAdsorb system was chosen by house owners as the easiest
way of reducing radon levels without significant construction work and
changes in a house operation. The dimensions of the basement were 8 m x
7.5 m x 2.3 m, with a volume of about 138 m3. Standard doors connected
the basement with the kitchen and with the outdoors. The sampling
location was in the basement close to the RAdsorb/EAC system outlet and
near to the outside door. This location was necessary because of the
use of the basement as a workshop and storage room by the house owner.
The radon concentration on the day of arrival to the house was about 660
Bq m~3 (18 pCi L"1) with particle concentration of 10000 cm"3. The
average temperature in the basement during the measurements was up 30° C
with very high humidity.
INSTRUMENTATION
The physical parameters measured during testing the RAdsorb/EAC
system were: radon concentration, particle concentration, potential
alpha energy concentration, and activity-weighted size distribution of
the radon decay products.
Radon gas
For radon gas concentration measurements, an EBERLINE RGM-3 radon
monitor was used. The RGM-3 is a portable, microcomputer-based radon
gas measuring instrument which utilizes a 3.3 liter, scintillation cell
detector and microcomputer controlled 8 1pm pump to sample radon gas.
The instrument allows the operation in the grab sampling mode and a
continuous mode. That provides the radon gas concentration at one hour
intervals. The microcomputer predicts decay products plateout as a
function of time during the first hours of operation and compensates for
it. The sensitivity of the device was 0.12 cps/pCi L"1.
Particle Concentration
To measure the airborne particle concentration and size
distribution, a TSI Model 3932 Differential Mobility Particle Sizer
(DMPS) was used. The DMPS measures the size distribution of
submicrometer aerosols by the electrical mobility detection technique.
The aerosols are classified with Model 3071 Electrostatic Classifier and
their concentration measured with Model 3086 Electrometer.
-------
A microcomputer controls the system, collects the raw data, performs the
data inversion to obtain the particle concentration as a function of
particle diameter. The diameter range measured in these experiments is
0.01 pra to 0.4 ^m with a concentration in the range of 103 to 105
particles per cm3.
Activity-Weighted Size Distributions
The activity weighted size distribution was measured with the
automated, serai-continuous graded screen array (ASC-GSA) described by
Ramamurthi (9) and Ramamurthi and Hopke (10). The ASC-GSA measurement
system involves the use of combination of six sampler-detector units
(see Figure 3) operated in parallel.
SCREEN
ALPHA DETECTOR
AA FILTER
FILTER HOLDER
CRITICAL ORIFICE
TO PUMP
Figure 3. The cross-section of the sampler unit
Each sampler-detector unit couple wire screen penetration, filter
collection and activity detection with a solid state detector in a way
as to minimize depositional losses. The system samples air
simultaneously in all of the units, with a flow of about 15 1pm through
the sampler slit between the detector and filter holder section in each
unit. The sampled air is drawn through a Millipore filter (0.8 fjm, Type
AA). The combination of wire screens wrapped around the samplers are
presented in Table 1.
-------
TABLE 1. THE PARAMETERS OF THE SIX SAMPLERS OF THE ASC-GSA
SYSTEM
Unit Sampler Slit Width Sampler Diameter Screen Mesh Dp50 (0.5-350 run)
[cm]	[cm]	[run]
1	0.5	5.3
2	0.5	5.3	145	1.0
3	0.5	5.3	145x3	3.5
4	0.5	5.3	400x12	13.5
5	1.0	12.5	635x7	40.0
6	1.0	12.5	635x20	98.0
One of the sampler-detector units is operated with an uncovered
sampler slit, thus providing information on the total radon decay
product concentrations. To detect alpha particles emitted by 218Po and
2uPo atoms collected or formed on the filters, ORTEC Model DIAD II, 450
mm2 surface barrier alpha detectors are used. The signals from the
detectors are amplified and routed through a multiplexer to PC-based
multichannel analyzer (ORTEC-MAESTRO) installed in an IBM-compatibile
laptop computer. The collected spectra are saved on the hard disk of
the PC for further analysis. The block diagram of the ASC-GSA system is
presented in Figure 4.
MULTIPLEXER
PUMP
LAP-TOP
COMPUTER
1.2.3.4.5.6 SAMPLERS
a,b,C,cU,f- AMPLIFIERS
Figure 4. The block diagram of the ASC-GSA system
The computer control of sampling, counting and analysis permits
automated, semi-continuous operation of the system with a sampling
-------
frequency between 1.5 to 3 hours. The activities of each radon progeny
are estimated from alpha spectra collected during two counting
intervals: the first one during sampling and the second 20 minutes after
end of sampling. The observed concentrations of 218Po, 21**Pb and 2uBi
are used to reconstruct the corresponding activity-weighted size
distributions using the Expectation-Maximization algorithms (11).
The ASC-GSA system allows the determination of the activity
weighted size distributions in six inferred size intervals in geometric
progression within the 0.5 - 500 nm size range. The performance of the
ASC-GSA system was tested during laboratory (9) and field (12)
intercomparison measurements showing very good agreement with systems
from other leading laboratories.
RESULTS
To study the performance of the RAdsorb/EAC radon mitigation
system on radon and radon decay products the experiments were designed
to :
a)	Test the effectiveness of RAdsorb/EAC in removal of Rn gas and
progeny.
b)	Determine the changes in the size distributions of Rn-d caused
by the RAdsorb/EAC system.
The design approach was to run each component of the RAdsorb/EAC
system: Fan, RAdsorb, EAC independently and in combination, establishing
the baseline before and after each run. As a control parameter to test
the potential health effects of the action during the tests, the dose to
secretory cell for a resting adult male was calculated by the method
described earlier. The reference levels (the "background" values) of
222Rn concentration, PAEC and activity fractions to which the
comparisons were made, were taken as:
1)	The mean values of measurements after assuming that the steady-
state conditions were established,
2)	The mean values of the "background" measurements performed on
the day of arrival and on the last day of tests (see Table 4).
The second approach was considered to present the changes in
measured quantities in relation to the conditions when no devices were
operated and which could be treated as a true "background".
The exposure to PAEC was calculated as follow:
= PAEC^-n	(2)
where,
-------
Ep - exposure to PAEC [WLM],
PAEC - potential alpha energy concentration [WL],
8760 - numbers of hours per year,
170 - number of hours per working month,
n - occupancy factor (n=0.8 was assumed).
"Background" Conditions
To establish the "background" conditions and the operational
parameters of the instruments, the first measurement was performed on
the day of arrival with the RAdsorb/EAC system turned off 40 hours
earlier.
The measured "background" conditions are presented in Table 2.
TABLE 2. THE "BACKGROUND" CONDITIONS IN THE SHREWSBURY HOUSE
ON THE DAY OF ARRIVAL
Particle concentration [cm"3]
10000
222Rn concentration [Bq m"3]
659
218Po concentration [Bq m"3]
307
21APb concentration [Bq m~3]
122
21*Bi concentration [Bq m"3]
78
PAEC [mWL]
33.1
Equilibrium factor
0.19
"Unattached" fraction of 218Po
0.65
"Unattached" fraction of PAEC
0.35
The "background" conditions were tested again, after the
RAdsorb/EAC system had been turned off for 15 hours during the last day
of measurements. The measured variables are presented in Table 3.
TABLE 3. THE "BACKGROUND" CONDITIONS IN THE LAST DAY OF
MEASUREMENTS
Particle concentration [cm-3]
4000
222Rn concentration [Bq m"3]
599
218Po concentration [Bq m"3]
377
21<,Pb concentration [Bq nf3j
93
214Bi concentration [Bq m"3]
52
PAEC [mWL]
28.4
Equilibrium factor
0.18
"Unattached" fraction of 218Po
0.87
"Unattached" fraction of PAEC
0.61
-------
The size distributions of radon decay products and PAEC without
RAdsorb/EAC system working are presented in Figure 5. The low particle
concentration in the basement for the two background samples resulted in
65% and 87% of the 218Po activity in the smallest inferred size interval
with a mid-point diameter of 0.9 nm. The corresponding 214Pb and 21*Bi
distributions showed activity in the 0.5 to 1.6 nm range below 20% and
50%, respectively. The resulting PAEC distribution followed a standard
bimodal distribution with maximums in the range 0.5 to 1.6 nm and 160 to
500 nm. The estimated doses to secretory cells and mean values of PAEC
and Z22Rn concentrations in "background" conditions are presented in
Table 4.
TABLE 4. AVERAGE VALUES OF SOME PARAMETERS IN "BACKGROUND"
CONDITIONS
222Rn [Bq nf3]
PAEC [mWL]
0.5-1.58 nm
Secretory Cell Dose


PAEC fraction
[mGy y"1]
630
30.8
0.428
55.8
Fan
To investigate the influence of the operation of the RAdsorb
system's fan, the charcoal canister was blocked allowing free
circulation of the air through the device. According to some studies, a
fan itself can act as a removal unit by increasing the plateout rate of
radon decay products (13). This effect was observed as well during
operation of the RAdsorb's fan operating. The results of the
experimental runs with fan ON and OFF are presented in Figure 6. As
expected, radon gas concentration (Figure 6 a) was not effected by
turning on the fan. The fan caused a decrease both in the PAEC and
2i8p0 concentrations (Figure 6 b) and d). This result is due to better
mixing of indoor air and an increase m the deposition rate of the
progeny on room surfaces. The activity size distributions of PAEC and
218Po were not affected by the fan in any significant way.
Fan /E AC
To study the effect of the combined operation of the RAdsorb unit
fan together with its attached EAC, the EAC was turned on while the fan
was operating. The results are also presented in Figure 6. The
concentrations of 2Z2Rn and 218Po did not show any drastic changes that
could be attributed to operating the fan/EAC. PAEC has shown a
reduction of a factor of 2 from about 40 mWL to 22 mWL (mean values from
four measurements under steady-state conditions before and after turning
the device on). For the reference values from the "background"
measurements (see Table 4), the reduction of PAEC was about 29%. A much
-------
SAMPLE No. 1
1.00
222—Rn: 659 Bq/m3
CN conc.: 10000 /cm3
CZI — PAEC: 688 nJ/m3 (33.1 mWL)
~ - 218-Po: 307 Bq/m3
A - 214-Pb: 122 Bq/m3
O - 214-0 i: 7B Bq/m3
0.80-
y 0.60-
P 0.40-
).5 1.0	10.0	100.0
PARTICLE DIAMETER [nm]
SAMPLE No. 53
o
f—
CJ
<
cr
£
>
I—
o
<
0.80-
0.60-
0.40-
0.20
0.00
222—Rn: 599 Bq/m3
CN conc.: 4000 /cm3
~	-PAEC: 591 nJ/m3 (28.4 mWL)
~	- 218-Po: 377 Bq/m3
A - 214-Pb: 93 Bq/m3
O - 214-B i: 52 Bq/m3
0.5 1.0	10.0	100.0
PARTICLE DIAMETER [nm]
Figure 5. Typical activity size distributions in "background" conditions
(no mitigation devices in operation)in the Shrewsbury house.
-------
E
f a#4N«*
y om m
TT
a)
4	8	12 16
T I M E [h]
100.0
^ ao.o
E
' 60-0
£ 40.0
20.0
0.0
b)
Ctg«r«U Tv«n FAN
~tT Tf
«M/ec
8 12 16 20
T I M E [h]
700
eoo
500
400
300
200
100-
0-
ClMrilU Tereh
y* vr
d)
¥ .
8 12
T I M E [h]
20
24
Z 0.8
O
f=
P
O
¦< 0.2
SIZE RANGE 0.5-1.58 nm: PAEC
1«nsh PAN
C^FF* y
*1.
C)
6	12 16
T I M E [h]
2 0.8
O
^ 0.6
P
< 0.2
0.0
SIZE RANGE 0.5-1.56 nrru Po-218
O'VonUa 7«rch PAN
mj* cyfvo* err
e)

• •
8	12 16
T I M E [h]
Figure 6. The influence of a fan and electronic air cleaner (EAC) on
indoor radon and its decay products:
a)	222Rn concentration,
b)	potential alpha energy concentration (PAEC),
c)	activity fraction of PAEC in the size range 0.5-1.58 nm,
d)	218Po concentration,
e)	activity fraction of218Po in the size range 0.5-1.58 nm
-------
1
£ BOO
RAd **a/UC i
m mtrr
n

12 16 TO 24 28 32 34
T I M £ [h]
a)
BO.O-
70.0
60.0
r—i
5t 50.0
^ 40.0
o
§ 30.0
Q>
20.0
(0.0
0.0
b)
I Ad RM/tAC dOMRt *M/VC
9Jt 1* OWPf	<*f
I
•I'"
8 12 16 20 24 28 32 36
T I M E [h]
1.0-
Z 0.8
i-
fe 0.4
^ 0.2
0.0
SIZE range; D.5-1 .sa nm: PA£C
c)
ww nu/ttc ciam**
k
" I • • ;

Jh
6 12 16 20 24 20 32 36
T I M E [h]
500
K>
.E 400
&
^ 300
^ 200
to
V 100
o
0.
0
RAd
r

d)
agar*!* Mdj
T

4 8 12 16 20 24 28 32 36
TIME [h]
SIZE RANGE 0.5-1.58 nm: Po-218
2 O.B
O
P
< 0.6
0.4.
%
^ 0.2
0.0
e)

. •
¦ *
"«¦
B 12 16 20 24 2B 32 36
T I M E [h]
Figure 8. The performance of the RAdsorb system (RAd) alone and combined
with a electronic air cleaner (EAC):
a)	22ZRn concentration,
b)	potential alpha energy concentration (PAEC),
c)	activity fraction of PAEC in the size range 0.5-1.58 nm,
d)	218Po concentration,
e)	activity fraction of218Po in the size range 0.5-1,58 nm
-------
TABLE 6. THE CHANGES OF 222Rn CONCENTRATION, PAEC, SIZE
DISTRIBUTION AND RESULTING DOSE DUE TO OPERATION OF of
RAdsorb/EAC
RAdsorb/EAC
222Rn [Bq m"3]
PAEC [mWL]
0.5-1.58 nm
PAEC fraction
Secretory
Cell Dose
[mGy y"1]
OFF
666
55.6
0.061
37.5
ON
163
8.0
0. 339
10.5
The operation of the combined RAdsorb unit with the EAC yielded a
substantial reduction in the radon gas concentration of about 76% and
PAEC of about 86%. This improved removal efficiency was enough to
compensate for the potential increase in the health effect due to
changes in the radon decay products size distribution (5 times increase
in the 0.5-1.6 nm fraction of the PAEC). The estimated dose to
secretory cells of 10.5 mGy y"1 was 72% lower then the initial value.
The estimation of the changes because of the operation of the combined
RAdsorb/EAC system was performed using the measured "background" values
(see Table 4). In relation to those values, the radon gas was reduced
by 76%, the PAEC by 74% and the dose to the secretory cells by 81%
The results suggest that the combined use of the RAdsorb and electronic
air cleaner (EAC) provided greater dose reduction than either operating
alone. The data suggests that the EAC is more effective in reducing the
dose from radon decay products when radon concentrations are lower (e g
less than 200 Bq nf3) . It was only when the RAdsorb lowered the
concentrations that the EAC provided some dose reduction. Since
the EAC are often installed to provide removal of pollen and other
irritants, the possible ancillary benefit of a reduction in radon
progeny dose at low radon concentrations warrants further investigation
Figure 8 a) presents the hourly measurements of radon gas. The data
shows a first sharp decrease in the radon concentration reaching the
lowest point of about 111 Bq m~3 in about 6 hours. Later, the radon
level increased and fluctuated around 150-200 Bq m~3. This pattern was
observed during all of the experiments with the RAdsorb unit.
SUMMARY
The influence of the RAdsorb/EAC radon mitigation system installed
in a single family house in Shrewsbury MA, was studied in a series of
tests. The radon gas concentration, PAEC and radon decay products
activity-weighted size distributions were measured on semi - continuous
bases.
The results obtained confirmed the theoretical predictions:
-------
larger effect was observed in the size distributions both of 218Po and
PAEC. The combined operation of the fan/EAC caused an increase in the
fraction 0.5-1.6 nm of 218Po from 0.445 to 0.754 (1.7 times increase)
and for PAEC from 0.158 to 0.626 (4 times increase).
Using the values obtained in the investigated house (decrease in
PAEC of about 50% and the changes in size distributions), the estimated
dose to secretory cells was 53 mGy y"1 before and 51 mGy y"1 after
turning the EAC/fan on. For the measured "background" parameters, the
estimated dose was 56 mGy y"1 (see Table 4). Therefore, no benefit in
reducing the health risk was observed.
The increase in "unattached" fraction without substantial
reduction in PAEC could lead to an actual increase in the radiation
dose, especially considering the relationship between dose per unit
exposure and size of particles described earlier (Figure 1). These
observations agree with the EPA recommendation not to used air cleaners
alone as a device for controlling the risk due to indoor radon.
RAdsorb
The results of operation of the RAdsorb system without the EAC
attached to the room air inlet are summarized in Figure 7 and Table 5.
The data included in table are mean values of measurements performed
after establishing the steady-state conditions.
TABLE 5. THE CHANGES OF 222Rn CONCENTRATION, PAEC, SIZE
DISTRIBUTION AND RESULTING DOSE DUE TO OPERATION OF RAdsorb
RAdsorb
222Rn [Bq m"3]
PAEC [mWL]
0.5-1.58 nm
Secretory Cells Dose



PAEC fraction
[mGy y"1]
OFF
670
55.6
0.061
37.5
ON
289
22.7
0.074
15.5
The operation of RAdsorb system caused a decrease in radon gas and
PAEC of about 60%, and an increase in 0.5-1.6 nm fraction of PAEC of
about 21%. The resulting decrease in dose to secretory cells was also
about 60%. For the measured "background" conditions (see Table 4), the
reductions in radon gas, PAEC, and dose were 54%, 26% and 72%,
respectively.
RAdsorb/EAC
The fully assembled RAdsorb system with the EAC unit attached to
the room air intake was operated continuously for 12 hours. After about
three to four hours, a new steady-state was established. The influence
of the device on Rn, PAEC, size distribution and dose are presented in
Figure 8 and Table 6.
-------
•ir
li
ll
II
II
li
'9
a)
T I M E [h]

70.0

60.0
T
50.0
*


*U.U
i__i

CJ
.VI .0
9

Q-
20.0

IO.O

a.o
Ogtratti	RAd Cmar«4k«
metr	wopwi
ii m
b)
•w

6	12 16
T I M E [h]
, iOO
5
N 100
©
ctamft ~
°%r
d)
8	12 16
T I M E [h]
SIZE RANGE 0.5-1.56 nm: PAEC
SIZE RANGE 0.5-1,58 nm; Pq-216
T 1
Z 0.8
< 0.2
C)
4 8 12 Ifl 20
T I M E [h]
z Q.e
o
; o.6
: 0,4-
0.0
e)
Oaofwtt# RAd Clac
o^rr	omwdW

8	1Z 16 20 24
T I M E [h]
Figure 7. The influence of the RAdsorb system on indoor radon and its
decay products:
a)	222Rn concentration,
b)	potential alpha energy concentration (PAEC),
c)	activity fraction of PAEC in the size range 0.5-1.58 nm,
d)	218Po concentration,
e)	activity fraction of218Po in the size range 0.5-1.58 nm
-------
1)	No substantial changes in measured parameters were observed
during only the operation of the fan,
2)	The EAC caused a shift of the size distribution towards smaller
particles,
3)	The RAdsorb system decreased the radon gas concentrations
without substantial changes in the progeny size distributions,
4)	The combined RAdsorb/EAC reduced the radon concentration by
about 76%, with the shift in the size distribution towards smaller
particles.
To study the effect of the increase in the "unattached" fraction
(0.5 - 3 nm), the doses to bronchial secretory cells of adult male
resting were evaluated. The estimation of doses before and during the
operation of the EAC gave similar results. By comparison, the combined
operation of the RAdsorb/EAC system not only substantially decreased the
radon gas concentration to a value around the EPA recommended limit of
150 Bq nf3 (4 pCi L"1) , but also yielded an 86% reduction in the PAEC.
The resulting dose reduction was 76% with assumption that the new
steady-state conditions were established. If the levels of 222Rn, PAEC
and activity fraction measured in the "background" conditions were taken
as the point of reference, the dose reduction was about 81%.
The dose estimates presented in the study, are based on the most
recent dosimetric calculations. However, it is possible that the
conversion factors applied in this study may change in the future due to
new development in dosimetric calculations.
In conclusion, the overall performance of the combined operation
of the RAdsorb/EAC system was very good in reducing both exposures to
and dose from indoor radon and its decay products.
The work described in this paper was not funded by the U.S.
Environmental Protection Agency and therefore the contents do not
necessarily reflect the views of the Agency and no official endorsement
should be inferred.
REFERENCES
1.	Environmental Protection Agency (EPA). A Citizen's Guide to Radon.
ODA-86-004, EPA, 1986.
2.	Abrams, R.F. Field Demonstrations of Radon Adsorption Units.
Paper presented at the 82nd Annual Meeting of the Air and Waste
Management Association, Anaheim, CA, June 25-30, 1989.
3.	Hopke, P.K. A Critical Review of Measurements of the "Unattached"
Fraction of Radon Decay Products. Technical Report Series DOE/ER-
0451P, U.S. DOE, 1990.
-------
k. Li, C-S. Field Evaluation and Health Assessment of Air Cleaners
in Removing Radon Decay Products in Domestic Environments. Ph.D.
Thesis, Department of Civil Engineering, University of Illinois at
Urbana-Champaign, Urbana, IL, 1990.
5.	James, A.C., Strong, J.C., Cliff, K.D. and Stranden, E. The
Significance of Equilibrium and Attachement in Radon Daughter
Dosimetry. Radiat. Prot. Dosim. 24: 451, 1988.
6.	James A.C. Dosimetry of Radon Exposure: What Does It Imply for
Risk and Measurement of Indoor Exposure? Paper presented at the
Twenty-Ninth Hanford Symposium on Health and the Environment,
Richland, WA, October 15-19, 1990.
7.	Abrams, R.F. and Rudnick, S.N. Radon Removal Unit Development.
Paper presented at the EPA 1988 Symposium on Radon and Radon
Reduction Technology, Denver, CO, October 17-21, 1988.
8.	Bocanegra, R. and Hopke, P.K. Theoretical Evaluation of Indoor
Radon Control Using a Carbon Adsorption System. JAPCA 39: 305,
1989.
9.	Ramamurthi, M. The Detection and Measurement of the Activity Size
Distributions (dp> 0.5 nm) Associated with Radon Decay Products in
Indoor Air. Ph.D. Thesis, Department of Civil Engineering,
University of Illinois at Urbana-Champaign, Urbana, IL, 1989.
10.	Ramamurthi M. and Hopke P.K. An Automated, Semi - Continuous System
for Measuring Indoor Radon Progeny Activity-Weighted Size
Distributions, dp:0.5-500 nm. Aerosol Sci. Technol., 1990 (in
press)
11.	Maher, E.F. and Laird, N.M. EM Algorithm Reconstruction of
Particle Size Distribution from Diffusion Battery Data. J. Aerosol
Sci. 7: 127, 1986.
12.	Hopke, P.K., Wasiolek, P., Knutson, E.O., Tu, K.W., Gogolak, C.,
Cavallo, A., Gadsby, K. and Van Cleef, D. Interc.omparison of
Activity Size Distribution Measurements with Manual and Automated
Diffusion Batteries - Field Test. Submitted for presentation at
the 1991 International Symposium on Radon and Radon Reduction
Technology, Philadelphia, PA, April 2-5, 1991.
13.	Abu-Jarad, F. and Sextro, R.G. Reduction of Radon Progeny
Concentration in Ordinary Room Due to a Mixing Fan. Radiat. Prot.
Dosim. 24: 507, 1988.
-------
IVP-5
CONTROL OF RADON RELEASES IN INDOOR COMMERCIAL WATER TREATMENT
by
D. Bruce Harris and A. B. Craig
U. S. EPA
Air and Energy Engineering Research Laboratory
Research Triangle Park, NC 27711
ABSTRACT	Water used in some commercial operations is
subject to conditioning processes inside buildings which
could cause radon to be released into the building's air.
The U. S. Fish and Wildlife Service recently found elevated
radon levels (100-300 picocuries per liter(pCi/L)) in some
of their National Fish Hatcheries (NFHs) even with
relatively low (400-600 pCi/L) levels in the incoming water.
The EPA's Air and Energy Engineering Research
Laboratory/Radon Mitigation Branch investigated possible
control techniques at the Neosho, MO, NFH. Data collected by
the NFH indicated that the nitrogen stripping packed tower
was removing up to 60% of the waterborne radon from 500,000
gal./day* and discharging it into the air above the fish
tanks. Two methods were tried to remove the radon: one used
countercurrent stripping and the other relied on hooding the
area immediately around the column discharge point. The 4 ft
height of the column prevented the low pressure fan normally
used in radon mitigation from establishing sufficient
countercurrent air flow to remove the radon. The pilot test
of the local hooding technique proved to be sufficient to
control the emissions. Final control was obtained by vacuum
stripping the incoming water rather than treating each tank
feed separately. Some city and industrial water treatment
facilities have reported elevated radon levels in treatment
rooms and adjoining offices that may have a similar origin
and may be amenable to similar control techniques.
This paper has been reviewed in accordance with the U.
S. EPA's peer and administrative review policies and
approved for presentation and publication.
BACKGROUND
Ground water is used as the source for many municipal and
industrial water systems. Some of the process treatment or use
takes place indoors. If radon is present in the water, the
(*) Readers more familiar with the metric system may use the
factors listed at the end of this paper to convert to that
system.
-------
possibility exists for radon to be released from the water and
exhausted into the interior of the process building. Fish
hatcheries are one such industrial facility.
The U. S. Fish and Wildlife Service has been testing
National Fish Hatcheries (NFHs) for radon as part of the general
testing program of federal buildings. Elevated levels were
measured in the air of buildings at the Neosho, MO, NFH. Initial
levels above 100 picocuries per liter (pCi/L) were found in the
tank room and adjoining offices (Table 1). Discussions with EPA
Region 7 staff led to a request for assistance from EPA's Office
of Research and Development.
PROCESS DESCRIPTION
The Neosho NFH uses water from several springs fed by
gravity to eight fish tanks inside the main building and several
outdoor tanks. Water flows at 50 gpm through a 4 ft high packed
nitrogen stripping/aeration tower and into each tank. This system
is similar to that shown in Figure 1 except the pipe extension
through the ceiling and the fan are not included and the top of
the tee is covered with a plate to prevent splashing. The plate
is not sealed, allowing some air into the water, but most of the
aeration takes place at the discharge of the column. The 400
pCi/L of radon found in the water wouldn't normally be considered
a major source of airborne radon. However, the tower is
approximately 60% efficient in stripping the radon as well as the
nitrogen. Given the water throughput, calculations show that up
to 500 pCi/L could be reached in the hatchery room air assuming l
air change per hour (ACH).
-------
The radon could be prevented from reaching the tank room air
by removing the radon at one of three points in the process: (1)
treating the water prior to entry into the hatchery, (2)
reversing the flow of air through the stripping tower and
exhausting it out the roof, and (3) collecting the tower effluent
gases with a hood and exhausting it.
MITIGATION SYSTEMS DESIGN AND TESTING
Neosho personnel modified the water inlet and aeration
column to fish tank No. 6 as shown in Figure 1 except for of the
fan which was installed by AEERL to test option (2). A
Kanalflakt T-2 fan was installed at the top of the column for
preliminary tests. This fan pulls air at 270 cfm at no head and
110 cfm at 1 in. WC head, the highest level listed on the
performance curve furnished by the manufacturer.
The fan was turned on and the column inlet water was
adjusted to 50 gpm. Under these conditions, the air flow rate at
Test Point l (Figure 1) was only 20 cfm and the pressure at Point
2 was -1.8 in. WC. The pumping action of the water passing
through the column was much greater than had been expected and a
larger fan (T-3B) would be needed to operate within its design
range.
The front half of the tank was then covered with plastic
film and the tank filled with water to determine the radon (Rn)
levels in the air exiting both ends of the column. When the
water was turned on, the film ballooned indicating that air was
exiting the bottom of the column as expected. The exhaust fan
was then turned on and, surprisingly, the film over the tank
-------
continued to balloon, although not quite as much. This indicates
that greater than 20 cfra of air is being released by the spring
water as it passes through the column.
Rn levels in the air exiting both ends of the column were
measured using a Pylon AB-5 continuous monitor. When the fan was
on, the air exiting the top of the column contained 15-20 pCi/L
and the air exiting the bottom of the column (measured next to
the column when the tank was covered by plastic) about 40 pCi/L.
When the fan was turned off, the radon in the air in the plastic-
covered tank rose to 60-80 pCi/L.
Based on these results and further theoretical
considerations, this type of fan installation would not be
expected to completely eliminate the flow of air containing radon
out the bottom of the column. Consequently, option (3), to
enclose the head end of the fish tank and keep that area under a
negative pressure with the use of a fan system similar to that
used in subslab depressurization systems, is a more viable
solution. This approach was tested using plastic sheeting to
make a temporary hood over the tank end around the water inlet.
Smoke studies showed that this captured the air above the water
easily with bleed air entering countercurrently just above the
tank water surface.
Figures 2 and 3 show how this option could be implemented to
enclose the free space over the column end of the fish tank.
This plan evolved during conversations with Neosho NFH personnel
as a simple but practical way of enclosing the column end of the
tank for evacuation with a minimum effect on day-to-day operation
of the fish tank. The tank top would be made of a heavy gauge
-------
aluminum (or perhaps plastic) cut as wide as the outside of the
tank (about 4 ft) and as long as the desired enclosure plus
enough to bend down a lip at a 90° angle to extend into the water
about 2 in. when the tank is in normal operation. Two corners of
the sheet would be notched so that the lip would just clear the
inside of the tank. The cover would be bolted to the top of fish
tank for ready removal when access is needed. It could not be
removed with the tank in operation. The top would be made
airtight with a bead of caulking applied under the lid before
bolting down. (A thick soft rubber gasket would be a viable
alternative.) The lip would need to be sealed to the side of the
fish tank, probably with caulking. Depending upon the fan
selected and the amount of air being pumped into the hood by the
tower, provisions for bleed air in the end of the cover may be
needed.
Two holes in the cover would be necessary for the 8 in.
aeration column and a 4 in. suction pipe. These pipes should
extend through the cover and be sealed to the cover to prevent
air leakage. This can be done very easily as shown in Figure 2
by cutting the pipe and placing a coupling on it at a point to
allow the coupling to ride on the cover and, if a short piece of
pipe is extended from the coupling through a hole in the cover
cut to the OD of the pipe, allow an easy caulk seal. The water
column would also have to be supported at the top to carry its
weight when operating. Water entry should be through a tee in
the column as was done in the experimental setup.
The top of the aeration column should be sealed from the
tank room and be supplied with outdoor air to prevent
-------
Table 1. NEOSHO NFH RADON LEVELS
Location
Device
Readina. pCi/L
Office
At Ease
108 average
Office
At Ease
98.2 last 12 hrs
Office
At Ease
108 current
Office
Sniffer
150
Office
E-Perm
116
Secretary's office
At Ease
99.4 average
Secretary's office
At Ease
106 last 12 hrs
Secretary's office
At Ease
106 current
Tank room
E-Perm
241
Tank room
Sniffer
150
Tank room
At Ease
222 average
Tank room
At Ease
2 63 current
Covered empty tank
E-Perm
270
Covered empty tank
E-Perm
260
Covered empty tank
Sniffer
300
Covered tank with water
E-Perm
>456*
Covered tank with water
E-Perm
>467*
Covered tank with water
Sniffer
475
Spring house
E-Perm
128
Visitor's rest room
E-Perm
20.3
* The E-Perm electrets read zero when checked, so reported reading
is an estimate.
-------
Fernco
F i tti ngs
Air to Outdoors
Test Point 1
Packing
Retai ner
Screen
^onfrol—
Water-

"I CT)
.
lv
rO
/-Test Point 2
¦Packing Retainer Screen
¦Attic Floor
L
Cei1ing
r
Splash Cover
¦ /*	Packing Retainer Screen
Water
I	f s
Air
Fish Tank
FIGURE 1. MODIFIED STRIPPING/AERATION COLUMN
-------
Exhaust to
1 r. Outdoors
.Outdoor Air
Fan
Damper
Attic Floor
Joi st
	^ 	
Cei1ing
Fernco Fitting
Water 	>¦
CD
9 / Cover Detailed in Figure 3
) Conpling
¦=F
Air & Radon
Water Surface
Fish Tank
Screen (Optional)
Water
FIGURE 2. TANK RADON MITIGATION SYSTEM
-------
U-- Outside Tank Width -~*j
Bolt Holes
o
To Water
+ 2 in.
o
4.5 in. Hole for
Suction Pipe
8.5 in. Hole for
Deaeration
Column
\
O
Break 90°
Inside Tank
Width
Figure 3. HOOD/TANK COVER
-------
JR.
\
p-"
1800
5/3/90
~\	1	1	1	1	1	1	r
2400
1	r
5/4/90

k
'0
0800
TIME, hrs
Figure 4. NEOSHO NFH RADON CONCENTRATIONS
-------