Proceedings of the American Association of Radon Scientists and Technologists 2008 International
Symposium Las Vegas NV, September 14-17, 2008. AARST © 2008
CONCLUSION OF A MULTI-YEAR STUDY ON THE ELEVATION
EFFECTS ON SCINTILLATION CELL COUNTING EFFICIENCY
FOCUSING ON THE PYLON™ MODEL 300
James F. Burkhart and Benjamin Abrams
Physics Department
University of Colorado-Colorado Springs 80918
Phillip H. Jenkins
Bowser-Morner, Inc.
Dayton, Ohio 45424
Abstract
This is a continuation of a study begun in 2005 when the authors first reported on the
theoretical possibility of a non-trivial error in calibration of radon chambers caused by
the current use of scintillation cells. The ramifications of this error are just now being felt
by the radon industry as experts around the world attempt to better define radon
concentration standards and insist upon measurement results more accurate than the
traditional 25 %. In this report, the authors conclude their study by comparing the
theoretical error and the actual experimental error of using a certain popular scintillation
cell made by Pylon™(1) with a volume of 271 ml. A graph is presented which predicts the
cell calibration error, as a function of the difference in elevation of any secondary
chamber and the primary calibration facility, which would be introduced using this
Pylon™ cell. The authors summarize their study by presenting a second graph that
predicts the cell calibration error introduced by using any right cylinder scintillation cell
of any dimensions that can be modeled as a scaled-up or scaled-down version of the
popular Rocky Mountain Glassworks cell.
Background
Both theoretical and experimental investigations indicate that the performance of
scintillation cells varies when filled at different elevations, as a result of the dependency
of alpha particle range on air density (George, 1983;Eberline 1987; Burkhart, 2005). In
fact, it has previously been shown that this discrepancy in cell counting efficiency for the
popular Rocky Mountain scintillation cell(2) (a right cylinder cell, 7 cm in diameter and
9.7 cm long, 360 ml in volume) can cause calibration errors between different elevations
in the U.S. as large as 9.1% (Burkhart, 2006). As the specific cell geometry influences
the magnitude of this error, correction factors must be determined for other cells, cells
that are not simply scaled-up or scaled-down versions of the Rocky Mountain cell. This
study investigates, therefore, the error in cell counting efficiency at different elevations
for the Pylon™ Model 300 scintillation cell, using the same apparatus as was used in the
earlier Rocky Mountain cell study.
(1) Pylon Electronic Development Company, Ltd., 147 Colonnade Road, Ottawa, Ontario, Canada.
(2) Rocky Mountain Scientific Glass Blowing Co., 4900 Asbury Ave., Denver, CO 80222
1
Proceedings of the American Association of Radon Scientists and Technologists 2008 International
Symposium Las Vegas NV, September 14-17, 2008. AARST © 2008
Theoretical Considerations
Alpha particles are the type of radiation emitted from 222Rn and two of its decay
products: 218Po and 214Po. It is well understood that the range (R) of an alpha particle
traveling in air is inversely proportional to the air density (ρ):
ρlRl = ρhRh ,
(Equation 1)
which means that alpha particles travel further at higher elevations (denoted by an “h”
subscript) where the air is “thinner” (Lapp, 1963). For example, the alpha particles that
are emitted by 222Rn, 218Po and 214Po travel more than 20% further in Colorado Springs,
Colorado, which is at 6000 feet elevation, than they do at sea level. Alpha particles that
are traveling within a scintillation cell may, therefore, strike the interior of a cell which is
filled at a high elevation when they would be quenched within the cell air volume, not
striking the scintillator, when the same cell is filled at a low elevation.
However, despite having such a significant impact on alpha particle path length, the
influence of differing air densities due to elevation has on alpha particle path is not
incorporated into current scintillation cell calibration techniques. An important example
of why this cell calibration error is commonly ignored can be best understood by
following the process by which secondary laboratories calibrate their chambers to the
U.S. EPA’s primary facility in Las Vegas, Nevada (elevation of approximately 2,200
feet). Briefly, the EPA sends a sample of radon from their chamber (using a standard
grab sampling method with a scintillation cell) to one of the secondary locations for
analysis. After analyzing the sample, the secondary location can then adjust its
calibration factor for the cell/counting system such that their equipment will report the
same radon concentration as the EPA. Indeed, this technique has proven very effective,
as it allows the secondary locations to read subsequent cell samples from Las Vegas to
within a few percent of the target value. In addition, as long as cells that are filled by a
secondary chamber (subsequent to calibration) are sent only to other secondary chambers
that have also calibrated with the EPA with a cell with identical geometry, all such
secondary chambers will agree on the radon concentration within the cell. Thus, any
intercomparison between secondary chambers (which are all calibrated with the EPA
chamber using this cell geometry) will not uncover any underlying problems attributable
to the air density within the cell. The authors are convinced that this is the reason that the
problems caused by using grab cells as an intercomparison have not received much
attention in the past. Nonetheless, problems do exist and will show up under some
circumstances, causing as much as a 9 % to 10 % error.
Specifically, the resultant calibration factors (counts per minute divided by decays per
minute; cpm/dpm) that the secondary facilities force onto their systems are dependent
upon the radon/air mixture and the difference in air pressure between the Las Vegas
chamber and the location of the secondary chamber. Errors will become evident when
the cell is refilled at the secondary location (if it is at a different elevation than Las
Vegas), and is subsequently used to calibrate a radon instrument or a tertiary chamber
because the number of decays per minute necessary to achieve a desired counts per
2
Proceedings of the American Association of Radon Scientists and Technologists 2008 International
Symposium Las Vegas NV, September 14-17, 2008. AARST © 2008
minute will change from the earlier calibration, depending upon the difference in
elevation between Las Vegas and the secondary chamber. In other words, the cell
calibration, cpm/dpm, will be incorrect for the subsequent fill. Clearly, secondary
chambers at a high elevation will require less radon (smaller dpm) to achieve the same
cpm compared to a chamber at lower elevation.
Experimental Procedure
In order to fill scintillation cells in a way which duplicated the elevations of Las Vegas
and other locations (from sea level to 6000 feet), a 25-liter tedlar(3) bag that is typically
used for transporting radon was filled over a 12-minute interval from a commercial radon
source manufactured by PylonTM. The bag was placed in a large metal cylinder (40 cm
diameter and 150 cm length), which was then sealed for a pressure tight fit. The chamber,
with the radon bag inside of it, was then brought to one of four pressures by pumping in
outside air. The pressure was read by a digital gauge and maintained to within 0.01
pounds per square inch (psi) of the desired value through the use of a pressure regulator.
While held at the desired pressure, radon was extracted from the bag at a rate of 5 SCFH
(2.4 L/min), and passed through the scintillation cell for four minutes. Before entering
the cell, the decay products that had accumulated in the radon bag were filtered out. The
valve located at the exhaust port of the cell was then closed, and the cell was allowed to
equilibrate with the pressure of the chamber/bag for one minute. See figure1:
Air Back
into Room
Air Pressure Gauge
Pressure
Regulator
Fill Pump
Valves
Filter
Radon
bag
Cell
Air
compresso
Air Compressor
r
Figure (1): An air compressor brings the chamber to a desired pressure. The pressure is maintained to
within 0.01 pounds per square inch with a gauge and a regulator. Using a second, independent pump, the
radon is extracted from the bag (under pressure), filtered, and passed through the Pylon™ cell.
(3) Tedlar bags purchased from Environmental Measurements, Inc. 215 Leidesdorff St., San Francisco, CA
94111 and SKC Inc., 863 Valley View Road, Eighty Four, PA 15330-9613, www.skcinc.com.
3
Proceedings of the American Association of Radon Scientists and Technologists 2008 International
Symposium Las Vegas NV, September 14-17, 2008. AARST © 2008
The same cell was filled sequentially at ambient pressure (820 hPa) in Colorado Springs
(6,000 feet above sea level), an over-pressure of 1 psi (888 hPa, equivalent to 4,000 feet
above sea level), an over-pressure of 2 psi (957 hPa, equivalent to 2,000 feet above sea
level), and an over-pressure of 3 psi (1026 hPa, roughly equivalent to sea level). The cell
was flushed between samples, and the background was re-established for each run. Two
correction factors were incorporated into the calculations in order to eliminate
confounding errors. Specifically, the loss of radon due to radioactive decay during the
time between filling the bag and taking samples was accounted for. Also, when sampling
at higher pressures, the increased density of air/radon that entered the cell was corrected
for by multiplying by the ratio of the ambient pressure divided by the chamber pressure
(according to the ideal gas law).
Results
Each “run” is defined as a minimum of two fills of a cell: the first fill was done under one
of the three over-pressures and the second fill, using the same cell and the same radon in
the bag, was done at ambient pressure. A minimum of five runs was completed at each
“elevation” below 6000 feet and at ambient pressure (6000 feet elevation), so that a
percent error in counting efficiency could be calculated. Considerable difficulty was
encountered because of frequent small holes occurring in the tedlar bags causing leaks,
especially when the bag was under pressure; many runs had to be repeated. In addition,
the bag only held enough radon to measure the error at two or three “elevations” per run.
As a consequence, missing values are represented by “X’s” in table 1, below.
Run Number Error at Sea Level
1
X
2
X
3
0.08
4
X
5
X
6
X
7
0.04
8
0.09
9
0.04
10
0.11
11
X
12
X
13
X
14
X
15
X
16
X
Average +/- σ
0.072 +/- .031
Renormalized
0.00 +/- 0.00
Error at 2000’
0.05
0.03
0.07
X
0.00
0.08
X
0.07
X
X
X
X
X
X
X
X
0.050 +/- .030
0.025 +/- .030
Error at 4000’
X
-.02
X
.04
-.05
.08
X
X
0.07
0.03
0.04
0.03
0.04
0.02
0.02
0.03
0.027 +/- .037
0.050 +/- .037
Error at 6000’
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.078 +/- .031
Table (1): Chart of errors in counting efficiency at different elevations found experimentally for the
Pylon™ Model 300 scintillation cell. The errors were calculated using Equation 2, found below. The final
row shows the renormalized errors calculated using an error of zero for sea level. See comments section.
4
Proceedings of the American Association of Radon Scientists and Technologists 2008 International
Symposium Las Vegas NV, September 14-17, 2008. AARST © 2008
The errors in table 1 were initially calculated by defining the error on the counting
efficiency at 6000 feet elevation as zero error where cell counting efficiency is here
defined as cpm/dpm and the number of decays per minute was held constant for any one
run (shown as one row in table 1). Radon was held constant by using radon from the
same bag, hence, once corrected for half-life decay, the radon was identical for each run.
Since the dpm/cell volume was, thereby, held constant (after correcting for the air volume
changes at different pressures) the dpm term in the cell calibration cancels out for each
run and only the cpm needed to be compared. Therefore, after counting for 60 minutes,
the number of counts (corrected for decay and pressure) at 6000 feet (Nh) was compared
to the number of counts (corrected for decay and pressure) at the respective lower
N ! Nl
elevations (Nl). Thus:
Error = h
.
(Equation 2)
Nh
These are the values shown in the second to the last row in table 1. In the previously
published theoretical model (Burkhart, 2005), however, the error was calculated by
dividing by Nl instead of Nh as is done in equation 2 above. However, to minimize
incidents of tedlar bags leaking, it was experimentally adventitious to always perform a
run at ambient pressure (6000 feet) and finding Nh instead of an over-pressure of 3 psi
(sea level) which would have allowed us to find Nl. Therefore, we were able to ascertain
a value of Nh for each and every run, making division by Nh more practical than division
by Nl, which was frequently not available. Then, in order to more easily compare to
previous theoretical and experimental results, all of the error values were renormalized by
forcing the error at sea level to be 0 % and correcting the other errors accordingly. These
are the values shown in the last row of table 1. (See the simple equation used for
renormalization in the comments section at the end of this paper).
Discussion of Results
There is an error, linear with elevation, when comparing cells filled at sea level with cells
filled at other elevations, with the error maximizing at .078 (7.8 %) for cells filled at 6000
feet elevation. This latter result agrees reasonably well with the theoretical prediction(4)
shown for the Pylon™ cell in table 2, below.
Cell Manufacturer
Cell diameter
Cell length
EDA
PylonTM
Rocky Mountain
5.3 cm
5.3 cm
7.0 cm
7.5 cm
13 cm
9.7 cm
Theoretical Error
8.0 %
8.8 %
9.8 %
Table (2): Chart of theoretical differences in counting efficiency between cells used at sea level and at 6000
feet of elevation. The theoretical errors are defined as the Error = 1 – (Nh-Nl)/Nl.
In addition, the results confirm the qualitative expected dependence of this error on cell
geometry: the error found here for the Pylon™ Model 300 cell is less than that of the
(4) The numerical analysis leading to this theoretical prediction, done by Robert E. Camley at the
University of Colorado-Colorado Springs, can be found in our earlier work (Burkhart, 2005).
5
Proceedings of the American Association of Radon Scientists and Technologists 2008 International
Symposium Las Vegas NV, September 14-17, 2008. AARST © 2008
Rocky Mountain cell. We believe this smaller error arises from the relative dimensions
between the two types. Specifically, the Pylon™ Model 300 has a relatively smaller
diameter compared to its length than the Rocky Mountain. (The ratio of the diameter to
the lengths for the PylonTM Model 300 is 0.41, while that of the Rocky Mountain is 0.72.)
As previously discussed (Burkhart, 2005), the narrow profile of the Pylon™ Model 300
lessens the extent to which the difference in alpha particle range at different air densities
affects counting efficiency.
0.00
0.01
E
r
r
o
r
Pylon Cell
0.02
0.03
0.04
0.05
0.06
0.07
0.08
Rocky Mountain Cell
0.09
0.10
0 Ft.
2000 Ft. 4000 Ft. 6000 Ft.
Elevation above sea level
Figure (2) This graph shows the error introduced by using the PylonTM cell for calibration of secondary
chambers. The maximum error between two chambers occurs when both chambers calibrate to the Las
Vegas U.S. EPA radon chamber while one chamber is at sea level and the second is at 6000 feet above sea
level. That error is 0.078, or 7.8 %. With the Rocky Mountain cell, that same maximum error was found to
be around 9.1 %, once renormalized to sea level (Burkhart 2006).
Secondary chambers, which used either the Pylon™ cell or the Rocky Mountain cell for
their calibration with the U.S. EPA, can use figure 2 in order to correct their reported
radon to represent the “true” radon in their chamber when exposing a radon instrument or
when calibrating a tertiary chamber. By “true” radon, we mean the actual radon in the
Las Vegas chamber.
6
Proceedings of the American Association of Radon Scientists and Technologists 2008 International
Symposium Las Vegas NV, September 14-17, 2008. AARST © 2008
In order to use figure 2 to calculate the “true” radon in the secondary chamber, find the
elevation of the secondary chamber and determine the error of its cell calibration by using
figure 2 and the correct curve (either a Pylon™ cell or a Rocky Mountain cell). Then
subtract the error for the elevation of the primary calibration chamber. This difference is
the net error that should be applied to the secondary chamber’s reported radon value.
As the first example, let us take a chamber in Colorado Springs, elevation 6000 feet using
a Pylon™ cell. Reading from figure 2, we see the error is 0.078. Assuming that the
chamber was calibrated by the U.S. EPA chamber in Las Vegas, we use figure 2 to find
an error of about .03 for Las Vegas, which is at 2000 feet. Subtracting the Las Vegas
error from the secondary chamber error, (0.078-0.030), we end up with a net error of
0.048. Thus, in order to find the “true” radon from the reported radon, using the original
calibration factors, it is necessary to reduce the reported radon of the secondary chamber
by .048, or 4.8 %. In other words, because Colorado Springs is at such a high elevation, it
took about 5 % less radon to produce the same number of counts per second as the U.S.
EPA used during its intercomparison with the Colorado Springs chamber. Also, any
device which the Colorado Springs chamber exposes should use the chamber radon value
reduced by about 5 %, assuming that the owner of the device wants to be calibrated or
compared to the “true” radon value.
As a second example, let us take a chamber at sea level. Reading from figure 2, we see
that error is 0 for the Pylon™ cell. Using the Las Vegas chamber for calibration gives us
the .03 error again. Subtracting the Las Vegas error from the former, (0-.03), we get - .03
which means that the sea level chamber needs to raise its reported radon by about 3 % in
order to reflect the “true” Las Vegas value. Failure to make these corrections among
secondary chambers could result in an accumulated error as much as 7.8 % if, for
example, a radon device is calibrated in a secondary chamber at sea level and is
subsequently sent to a chamber at 6000 feet for an intercomparison. If, on the other hand,
the Rocky Mountain cell was used in the initial calibration with the EPA chamber and the
secondary chambers, the maximum error between chambers could, as seen in figure 2, be
as high as 9.1 %.
Finally, for cells that have not been specifically studied, one can use a general theoretical
predictive model introduced in an earlier paper (Burkhart, 2005). The graph, table 3,
shows the maximum error between a chamber at sea level and a second chamber at 6000
Percent Error
12
10
8
6
4
2
0
0
2
4
6
Scale Factor
8
10
Table (3): This graph shows the predicted maximum error caused by using a cell that can be modeled as a
Rocky Mountain cell scaled up to 9 times larger or down to 1/3 of its manufactured size. The arrows on the
graph are used in a simple example that follows below which shows how to apply the graph to other cells.
7
Proceedings of the American Association of Radon Scientists and Technologists 2008 International
Symposium Las Vegas NV, September 14-17, 2008. AARST © 2008
feet. The Rocky Mountain cell is 9.7 cm in length and 7.0 cm in diameter. As an
example, then, a cell which is two times as large, say, about 20 cm in length and 14 cm
in diameter, would be expected to have a maximum error, from table 3, of around 9.5 %,
distributed linearly from sea level to 6000 feet.
Comments
As was discussed in the above paper, it was necessary, for practical reasons, to
experimentally calculate the error by counting each cell at one of the over-pressures and
at ambient (6000 feet) and calculating the error by dividing by the ambient counts, i.e.,
N ! Nl
Error = h
.
(Equation 2)
Nh
However, in order to compare with previous theoretical work, it was necessary to convert
this equation to one in which the error was determined by dividing by Nl, i.e.:
Nl ! N h
.
(Equation 3)
Nl
This was done by taking the error in each row of table 1, which is a fractional error
compared to 6000 feet (ambient) and changing it to a fractional error compared to sea
level:
N ! Nl
Start with
Error = h
.
Nh
Subtract both sides from 1 and divide both sides by the error
Error =
N h ! Nl
)/Error .
Nh
Substituting the value for the error from equation 2 into the right hand side, we get:
(1 - Error)/Error = (1 -
N h ! Nl
N ! Nl
)/( h
).
Nh
Nh
Taking the reciprocal of both sides and dividing through by the denominator on the right,
we get:
N ! Nl
Error/(1-Error) =1/ (1 /( h
- 1 )), which, after multiplying
Nh
and dividing the right hand side by Nh-Nl gives:
(1 - Error)/Error = (1 -
Error/(1-Error) = (Nh – Nl)/(Nh-Nh-Nl)
And canceling the Nh’s in the denominator and multiplying and dividing the right hand
side by -1, gives:
Error/(1 – Error) = (Nl – Nh)/Nl,
8
Proceedings of the American Association of Radon Scientists and Technologists 2008 International
Symposium Las Vegas NV, September 14-17, 2008. AARST © 2008
which is equation 3, the definition of the error as measured from the lower elevation.
References
Burkhart, J.F., Jenkins, P.H. and Camley, R.E., “Elevation Effects on Radon Cell
Counting Efficiency”, American Association of Radon Scientists and Technologists,
Proceedings of the 2005 International Radon Symposium, San Diego, CA, September,
2005. Volume 9-27-1:15, Pages 1-9.
Burkhart, J.F., Jenkins, P.H. and Moreland, E.L., Using a Radon Pressure Chamber to
Determine Cell Counting Efficiencies as a Function of Elevation”, American Association
of Radon Scientists and Technologists, Proceedings of the 2006 International Radon
Symposium, Kansas City, MO, September, 2006. Volume 1, Pages 67-75.
George, J.L., 1983. “Procedures Manual for the Estimation of Average Indoor Radon
Daughter Concentrations by the Radon Grab Sampling Method”, Bendix Field
Engineering Corp., Grand Junction, Colorado, GJ/TMC-11 (83) UC 70A, as referenced in
the “Indoor Radon and Radon Decay Product Measurement Device Protocols”, U.S.
Environmental Protection Agency, Office of Air and Radiation (6604J), EPA 402-R-92004, July 1992 (revised). Page 2-38.
Eberline-A Subsidiary of Thermo Instrument Systems, Inc., “RGM-3 Radon Gas Monitor
Technical Manual”, Santa Fe, NM 87504, March 1989. Pages 25 and 26.
Lapp, Ralph E. and Andrews, Howard, L., Nuclear Radiation Physics, Third Edition,
Prentice-Hall Inc., Englewood Cliffs, NJ, 1963, pages 117-119.
9