United States Air and Energy Environmental EPA/600/9-90/005d
Environmental Protection Research Laboratory January 1990
Research Triangle Park NC 2771 1
Research and Development
v>EPA The 1990 International
Symposium on Radon
and Radon Reduction
Technology:
Volume IV. Preprints
Session VII: Radon
Reduction Methods
Session D-VII: Radon
Reduction Methods—
POSTERS
Session D-IX: Radon in
Schools and Large
Buildings—POSTERS
February 19-23,1990
Stouffer Waverly Hotel
Atlanta, Georgia
-------
Session VII:
Radon Reduction Methods
-------
VII-1
EVALUATION OF SUB-SLAB VENTILATION FOR INDOOR RADON REDUCTION
IN SLAB-ON-GRADE HOUSES
by: D. Bruce Henschel
Air and Energy Engineering Research Laboratory
U. S. Environmental Protection Agency
Research Triangle Park, NC 27711
Arthur G. Scott
Andrew Robertson
AMERICAN ATCON, INC.
Wilmington, DE 19899
William 0. Findlay
Acres International Corp.
Amberst, NY 14228
ABSTRACT
Sub-slab ventilation (SSV) radon reduction systems were tested in nine
slab-on-grade houses around Dayton, Ohio. The testing addressed the following
house design/construction variables: slab size; foundation wall material of
construction; presence or absence of sub-slab forced-air supply ducts; and
presence of sunken living rooms as an obstruction and soil gas entry route in
the slab interior. SSV design/operating variables addressed include: placement
of the ventilation pipes inside or outside the house; location and number of vent
pipes; SSV fan capacity; and operation of the SSV system in suction vs. in
pressure. The results suggest that large slabs, block foundation walls, and sub-
slab ducts can sometimes necessitate additional care in SSV design (number,
location of vent points), but that, in general, radon levels can be reduced to
2 pCi/L or less with one or two points if there is a good aggregate layer under
the slab. SSV from inside and outside the house give generally comparable
performance, although in some cases interior might be preferable for large
houses. Increasing the number of suction pipes from one to two, and increasing
fan capacity from half to full speed, generally (but not always) appear to
improve SSV performance. Only in the case of one large house with sub-slab ducts
were two pipes required to reduce levels below 4 pCi/L. Low fan speed was
generally sufficient to reduce levels below 4 pCi/L. Operation of SSV systems
in pressure never gave better reductions than did operation in suction. Results
from testing of slab-on-grade houses in Florida indicate that SSV systems
generally require more suction pipes than were found necessary in Ohio, when the
sub-slab material has poor permeability.
This paper has been reviewed in accordance with the U. S. Environmental
Protection Agency's peer and administrative review policies, and has been
approved for presentation and publication.
-------
INTRODUCTION
Much of the previous testing of residential radon reduction techniques has
focused on houses having basements, including basements with adjoining slab-on-
grade or crawl-space wings. Testing has been more limited in houses having
solely a slab-on-grade substructure. Slab-on-grade houses offer various design
and construction features which could impact the installation and performance
of traditional sub-slab depressurization systems, and which, in some cases, might
offer potential for application of alternative radon reduction techniques beyond
sub-slab systems.
Previous reports (1,2) presented results from initial (Phase 1) testing
of indoor radon reduction methods in four slab-on-grade houses in Dayton, Ohio.
This paper expands those previous results to include further testing and data
analysis for those four Phase 1 houses, plus testing on five additional slab-
on-grade houses as part of a Phase 2 effort in the Dayton area. To provide
additional perspective on this substructure type, recent results are also
summarized from testing in 1 slab-on-grade house in Maryland, and in 14 slab-
on-grade houses in Florida (3).
A number of house design and construction variables can impact the design
and performance of radon reduction systems in slab-on-grade houses. The house
design/construction variables addressed in the Ohio study include: slab size;
nature of the foundation stem wall (hollow-block vs. poured concrete); presence
or absence of forced air heating supply ducts underneath the slab; the presence
or absence of interior slab openings/soil gas entry routes, such as plumbing
openings under bathtubs; and sub-slab obstructions, such as sunken living rooms.
Among other house variables that could be significant but that are not
specifically addressed by the study in Ohio are: absence of aggregate underneath
the slab (all of the Ohio houses had a good aggregate layer); nature of the
interface between the floor slab and the foundation; sub-slab obstructions other
than sunken living rooms (e.g., interior grade beams); and the impact if the
forced-air ducts under the slab are return ducts (at negative pressure) rather
than supply ducts (positive pressure). Some of these other variables are being
addressed in testing underway or planned in slab-on-grade houses in Florida and
New Mexico.
The characteristics of the native soil underlying the house can also be
important, especially if there is no aggregate underneath the slab. All nine
of the slab-on-grade houses in Ohio (and the one house in Maryland) were
underlain by clay soil, with a layer of aggregate between the clay and the bottom
of the slab. The Florida houses were on top of fine, moist sand with no
aggregate, providing relatively poor sub-slab communication; as discussed later,
these characteristics greatly influence the design of sub-slab ventilation (SSV)
systems for the Florida houses. Future testing in New Mexico will address houses
with no aggregate over expansive clays and coarse, dry sands.
SSV was the radon reduction method emphasized in the Ohio testing, and is
the technique discussed in this paper. Mitigation design variables studied
during this testing included: placement of the ventilation pipes vertically down
through the slab from inside the house ("interior" SSV) vs. horizontally through
-------
the foundation wall from outdoors ("exterior" SSV); location of the vent pipes;
number of vent pipes; SSV fan capacity; and operation of the SSV system to draw
suction on the sub-slab region vs. to pressurize the sub-slab. Other radon
reduction approaches which were tested in addition to SSV -- but which were less
effective and which are not covered in this paper -- were: continuous operation
of the central furnace fan in houses having sub-slab forced-air supply ducts,
in an effort to pressurize the sub-slab region; closure of the slab opening where
the bathtub plumbing comes up through the slab; and "site ventilation." Site
ventilation involves suction on a pipe embedded in the ground outside the house,
in an effort to draw soil gas from the entire lot.
Table 1 lists the nine slab-on-grade study houses in Ohio, including the
results of pre-mitigation alpha-track measurements in each house over a 3-month
period during the winter. Various pertinent house variables are also summarized.
Soil gas is the predominant source of the radon in all of the houses; well water
and building materials were determined not to be significant radon sources.
MEASUREMENT METHODS
For short-term measurements of mitigation system performance, radon gas
was measured using a Pylon Model AB-5 continuous radon monitor equipped with a
285 mL (17.4 in.3) Lucas scintillation cell. The Pylons were programmed to
measure radon hourly. Because of the significant day-to-day variability in
indoor radon concentrations observed in the Ohio houses, the short-term radon
measurements consisted of at least 48 hours (and commonly up to 96 hours) of
Pylon readings both before and after the mitigation system was activated. System
on/off measurements were made back-to-back, to the extent possible, to reduce
temporal variations. Measurements were made in the main living area under
closed-house conditions. Essentially all of the monitoring was completed during
the heating season (between December and March), except for a few of the Phase
1 measurements.
Diagnostic testing was generally limited primarily to measurements of
suctions, flow rates, and radon concentrations in the piping of operating SSV
systems. Sub-slab communication measurements were performed in some of the
houses, but were generally not conducted in houses with sub-slab ducts. The
ducts commonly radiated out from a centrally located furnace, effectively
subdividing the sub-slab region into multiple zones. Initial sub-slab
communication measurements in houses with ducts confirmed that suction did not
extend between the sub-slab zones created by the ducts, presumably because of
air leakage through the ducts; however, the suction field did extend well through
the aggregate within a single zone.
Three-month alpha-track detector measurements were made in the houses prior
to mitigation, as reported in Table 1. To measure long-term system performance,
alpha-track measurements will be repeated for at least 1 year with the systems
operating. For quality assurance, these detectors are deployed in clusters of
two or three; in addition, unexposed detectors, and detectors exposed to known
radon levels in a chamber, are sent blind to the analytical laboratory.
-------
SUB-SLAB SYSTEM DESIGN AND INSTALLATION
In houses having poured concrete foundation walls, an exterior suction
point was created by coring a 4-in. (10-cm) diameter hole horizontally through
the wall below slab level, to expose the sub-slab aggregate. A limited amount
of aggregate in the immediate vicinity of the hole was removed; otherwise, no
effort was made to excavate an opening under the slab. A fan was mounted on
the foundation wall over the hole, exhausting horizontally away from the house
at grade level. In the final installations, this fan was mounted on a piece of
marine-grade plywood that was sealed onto the foundation wall over the hole.
No piping extended from the fan through the hole into the sub-slab region. In
houses having hollow-block foundation walls, exterior SSV points involved a 5-
in. (13-cm) hole drilled through the foundation wall; a short segment of 4-in.
FVC pipe was inserted through the hole, and expanding foam was injected into the
annular gap between the pipe and the hole in the wall. This step was taken to
reduce the amount of air drawn into the fan through the block cores (i.e., to
reduce the block wall ventilation component of this system), since significant
leakage of air through the block cores might be expected. The fan was then
mounted on the outdoor end of the pipe segment, exhausting away from the house
at grade.
For interior SSV systems, vertical PVC pipes penetrated the slab, generally
in a furnace room, utility room, or bedroom closet. Again, aggregate immediately
under the hole was removed, but no effort was made to excavate a pit under the
slab. Each pipe extended up through the ceiling into the attic. If there were
more than one suction pipe, the multiple pipes would be manifolded together in
the attic, and connected to a single fan. The fan was mounted in the attic, and
exhausted through the roof at a convenient location. In House 58, to determine
whether interior SSV could be simulated from outside the house, a PVC pipe was
inserted horizontally through the foundation wall from outdoors, for a distance
of 6 to 7 ft (about 2 m) under the slab, simulating an interior pipe through
the slab at that location.
Both the "exterior" and "interior" approaches were tested in four of the
houses, to permit direct comparison of the two approaches in the same house.
The interior systems in three of the houses included two suction pipes, located
in different parts of the house, permitting testing of either pipe by itself,
or both simultaneously. This multiple-pipe installation was to enable an
assessment of the effect of the number and location of the vent pipes. The
exterior system in the largest house (House 77) had two vent points, which could
be operated one at a time, or both simultaneously (each with its own fan). In
several houses where only exterior SSV was tested, the single fan was tested at
two locations. In most houses, different fan speeds were tested.
In all cases, Kanalflakt T2 plastic-bodied fans were used, capable of
drawing 270 cfm (127 L/s) at zero static pressure, and capable of 1.4 in. H,0
(350 Pa) suction at zero flow.
-------
RESULTS
SSV SYSTEMS IN SUCTION
The best performance of sub-slab suction systems in each of the slab-on-
grade houses is shown in Table 2. Where best performance was obtained during
operation of two suction points simultaneously, these cases are marked in the
table; in all other cases, only one suction pipe was operating. In houses where
both exterior and interior SSV were tested, the results of the best installation
of each are shown in the table for comparison. In houses where single exterior
SSV suction points were tested at two locations, the performance is reported for
each location (designated Site 1 and Site 2). In all cases, best performance
was observed with the fan operating at high speed, although good performance was
also attained at reduced fan speed. In general, these best results were obtained
with the slab opening under the bathtub having been foamed closed. The pre- and
post-mitigation radon levels presented here are short-term (48- to 96-hour)
Pylon measurements, generally made immediately before and immediately after
activation of the SSV fan, for back-to-back system on/off comparison. The first
5 hours of data taken immediately after the fan was turned on or off are
excluded, in order to reduce the effect of the transient on the reported results.
Several conclusions are apparent from Table 2. Substantial reductions
-- generally to a residual radon level of less than 1 to 2 pCi/L (37 to 74
Bq/m3)* -- were achieved in all houses, with no more than two suction points (and
generally only one). The generally good performance was probably aided by the
very low air-flow resistance of the good layer of aggregate under the slab. As
indicated in the table, the suctions in the SSV system piping were generally
above 0.5 in. H20 (125 Pa), and the flows were generally below 140 cfm (65 L/s) ,
which is half fan capacity. These suctions/flows generally represent good SSV
operation, and confirm that large amounts of air were not short-circuiting into
the systems via sub-slab ducts or block foundation walls.
The apparent effects of the house variables and SSV system design/operating
variables are discussed below.
Effect of Sub-Slab Forced-Air Supply Ducts--
The initial concern had been that houses with sub-slab heating ducts would
not achieve high radon reductions with a reasonable number of SSV pipes, because
the ducts would prevent effective distribution of the suction field underneath
the slab. However, the results in Table 2 indicate that -- with the aggregate
present under these houses -- any effect of the ducts is limited. The two houses
without sub-slab heating ducts (Houses 70 and 72) did not consistently perform
better than did comparable houses with ducts. On one hand, House 70 (large
house, block foundation, no ducts) appeared to give better results with two-pipe
interior SSV (0.2 pCi/L, 99% reduction) than did House 77, which was comparable
in characteristics except that it had ducts (1.7 pCi/L, 89% reduction). But on
* Radon in Bq/m3 - 37 x value in pCi/L.
-------
the other hand, House 72 (moderate size, block foundation, no ducts) did not
seem to do as well with one-pipe interior SSV (1.6 pCi/L, 93% reduction) as did
House 58, which was comparable but had ducts (0.8 pCi/L, 97% reduction). These
radon levels are sufficiently low that -- given temporal and house-to-house
variability, and measurement uncertainties -- it is not possible to quantify,
from these data, any limited effect that the sub-slab ducts were having on SSV
performance. The results from House 70 vs. House 77 are further compared later.
Effect of Interior vs. Exterior SSV--
From the standpoint of aesthetics, the choice between interior and exterior
SSV for a given house will depend on the layout of the house and the preferences
of the homeowner. It was desired to determine whether technical considerations
might suggest one over the other under certain circumstances. For example,
exterior SSV might be aiding in good performance in houses with radial sub-slab
ducts, because the suction is drawn at the perimeter and thus might extend
through the loose fill around the perimeter slab/foundation wall junction. In
houses with block foundation walls below the slab, exterior SSV might be helpful
or be a disadvantage, depending on the effects of the wall ventilation component
that is likely to result. Large slabs with potential interior entry routes (in
addition to the perimeter wall/floor joint) might be less effectively treated
with exterior SSV.
In the four houses in Table 2 where both interior and exterior SSV were
tested, the two SSV approaches appeared to perform about equally well in three
of the houses (Houses 31, 58, and 77). These three houses cover the range from
a small house with a poured foundation (31) to a large house with a block
foundation (77); all had sub-slab ducts. In the fourth house, House 70, the
interior SSV system appears distinctly superior; however, it must be noted that
the interior SSV system utilized two suction pipes, while the exterior system
drew on only one point. The large size of this house, and its block foundation,
might also have contributed to this result. It should also be noted that in one
of the houses where interior and exterior SSV appeared comparable (House 77),
a two-point interior system connected to a single fan is being compared to a two-
point exterior system where each exterior point has its own fan; it is not known
whether the two-point exterior system would have performed as well as the
interior system if both exterior points had been connected to a single fan.
Effect of Slab Size--
Re gar ding the effect of house size, the largest house (House 77, with a
slab of 2,600 ft2, or 240 m2) required two suction pipes to reduce radon levels
below 4 pCi/L; one pipe was insufficient. The other large house (House 70, with
2,350 ft2, or 220 m2) is also shown in Table 2 as having achieved its best
performance with two interior suction pipes; however, operation of any one of
the pipes was sufficient to reduce levels below 2 to 3 pCi/L. The likely reason
for the better performance with one point in House 70 is that it did not have
sub-slab ducts; in other respects, Houses 70 and 77 are similar (large houses,
block foundations), although House 77 did have the added complexity of a sunken
living room. For all of the other houses (ranging in size from 1,100 to 1,700
ft2, or 90 to 160 m2), one suction point was sufficient to reduce levels below
-------
2 pCi/L. The effects of multiple suction points are discussed further later.
As indicated previously, it appears that interior SSV might sometimes be
preferred in houses with large slabs; but exterior and interior SSV appeared to
perform equally well in houses with smaller slabs.
Effect of Block vs. Poured Concrete Foundation--
Regarding the effect of foundation material of construction, the houses
having poured concrete foundations would appear to have most consistently
achieved the best reductions. All four of the houses with poured foundations
(Houses 15, 31, 47, and 58) were reduced to below 1 pCi/L (97 to 99% reduction)
with each SSV approach (interior or exterior) tested in that house -- even though
all of these houses had sub-slab ducts. By comparison, only one of the five
houses with block foundations was reduced below 1 pCi/L, with a two-point
interior SSV system (House 70, having no sub-slab ducts). The other block-
foundation houses were reduced to levels of 1.6 to 3.7 pCi/L (81 to 93%
reduction). In part, the poorer reductions in the block houses might be
explained by the fact that the block houses were generally larger; except for
House 85, all of the block houses were 1,700 ft2 or larger, whereas all of the
poured-foundation houses were 1,400 ft2 or smaller. But size might not be the
only explanation: House 85, with a block foundation, was only 1,100 ft2, but
it was not reduced below 1.6 - 2.5 pCi/L. Perhaps air leakage through the block
foundation walls is reducing the effectiveness of suction distribution under the
slab.
Effect of Number and Location of Suction Points--
In three of the houses, two-point SSV systems were tested. Table 3
compares the results obtained in these three houses when the system is operated
with each of the suction points by themselves, and when the system is operated
with both points. For Houses 70 and 77, the results reported previously
in Table 2 reflected operation with both points. For all cases reported in
Table 3, the mitigation fan is being operated at high speed.
Several observations can be made from Table 3. In House 31 -- a small
house with a poured foundation and sub-slab ducts -- the best performance is
achieved when only the one centrally located pipe in the furnace room is
operated; connecting the second pipe, from the bath, causes a deterioration in
performance compared to the furnace room pipe alone. And, not surprisingly,
operation of the bath pipe by itself gives the poorest performance of all
(although still reducing levels to below 3 pCi/L). The reason for this effect
is not clear from the flow/suction measurements in the piping. The suction pipe
in the bath maintains reasonable suction (0.4 to 0.8 in. H20, or 100 to 200 Pa)
and reasonable flows (29 to 49 cfm, or 14 to 23 L/s), with the lower end of each
range reflecting operation of this pipe in combination with the furnace room
pipe, and the upper end reflecting operation of this pipe by itself. From these
measurements, the bath pipe does not appear to be degrading performance by
providing a short-circuit for air leakage into the system. The furnace room pipe
appears to be best intercepting the radon source, and operation of this pipe by
itself -- resulting in maximum suction (0.6 in. HjO, or 150 Pa) and maximum flow
(69 cfm, or 32 L/s) in this pipe -- appears clearly desirable. The apparent
-------
conclusion is that more suction pipes are not always better.
In Houses 70 and 77, operation of two suction points is always better than
operating only one. However, in House 70, operation of either point by itself
is sufficient to reduce the house below 4 pCi/L. By comparison, in House 77,
operation of any one point is only sufficient to reduce levels to between 6 and
8 pCi/L; operation of both points is required to achieve sufficiently low levels.
Both of these houses are large with block foundations; the probable explanation
for the better performance of single points in House 70 is that it does not have
sub-slab heating ducts, and House 77 does have sub-slab ducts. The presence of
a sunken living room in House 77 could also be contributing to the poorer
performance of the single points in that house. It is noted that slightly better
single-point performance with the interior SSV in House 70 was achieved with
Point #2, which was located in the right front quadrant of the house, rather than
with Point #1, which was centrally located. Thus, central location of a single
suction pipe is not necessarily required, even in a large house, if there is good
communication without obstructions such as sub-slab ducts. It is also noted
that, in House 77, the individual exterior SSV suction points seemed to perform
slightly better than did the individual interior pipes. This result could be
suggesting that -- where there are insufficient suction points in houses having
sub-slab ducts -- suction at the perimeter of the slab might be helpful because
it "taps into" the permeable region where excavation for the footings had
occurred during construction, a region which extends around the entire perimeter
and which is less interrupted by the sub-slab ducts. With the block foundation
in House 77, this result could also be suggesting that the wall ventilation
component that might exist with exterior SSV systems could be important,
especially when there is an insufficient number of points. (As noted previously,
steps were taken in most of the block-foundation houses to reduce the wall
ventilation component of exterior SSV systems.)
In the two houses where exterior SSV was tested at different locations,
there appeared to be no effect of location in one (House 15), and some possible
effect in the second (House 65). Both of these houses are small, and have sub-
slab ducts; House 15 has a poured foundation, and House 85 a block foundation.
Effect of the Wall Ventilation Component in Block-Foundation Houses--
An effort was made In House 1 to investigate the significance of the wall
ventilation component of exterior SSV in block-foundation houses. The original
Phase 1 exterior SSV system in this house had consisted of a T2 fan mounted
directly over the hole through the foundation wall; there was no pipe through
the wall, nor were the block cores surrounding the hole mortared or foamed
closed. Under these circumstances, it would be expected that there would be a
significant wall ventilation component to this sub-slab-plus-block-wall
ventilation system. As discussed in Reference 1, this Phase 1 installation
resulted in indoor levels of 1.3 pCi/L (during mild weather), the highest
residual radon level among the four Phase 1 slab-on-grade houses. The level
increased to 10.6 pCi/L when the central furnace fan was operated continuously
(to pressurize the sub-slab, counteracting the efforts of the SSV system to
depressurize the sub-slab); continuous fan operation had not caused such a
problem in the other three Phase 1 houses.
-------
The first step taken at House 1, as part of the Phase 2 effort, was to
remove the original fan, and to insert a 4-in. (10-cm) diameter pipe horizontally
through the foundation wall into the sub-slab aggregate. The T2 fan was then
re-mounted on this pipe. Expanding foam was injected to close the block voids
around where the pipe penetrated the foundation. These steps should have greatly
reduced the wall ventilation component of the system. Following these system
modifications, radon levels in the house averaged 7.8 pCi/L over a 96-hour period
in December, higher than had been observed with the original system. The pipe
and foam were then removed, and the fan mounted on the foundation wall over the
hole, as with the original installation. With this original configuration, the
indoor levels averaged 2.8 pCi/L over 94 hours in January (somewhat higher than
the 1.3 pCi/L reported earlier, perhaps because the earlier measurements had been
made in April). The conclusion would appear to be that the wall ventilation
component can be important in some cases. Taken together with the results from
House 77, discussed above, it would appear that the wall ventilation component
might be most important where the SSV system is marginal (e.g., in large houses
having only one suction point). Since interior SSV performed well in the block-
foundation houses (70, 72, and 77) when sufficient suction points were provided,
it would appear that a major direct wall ventilation component is not required
when the system is adequately designed.
Effect of Fan Capacity--
In many of the houses, the SSV system was operated with the Kanalflakt T2
fan both at maximum suction/flow and at about half suction, to investigate the
effect of fan capacity. The performances at the two different fan capacities
are compared in Table 4. The ability to achieve good performance at reduced fan
capacity is important. Some percentage (probably between 30 and 70%) of the gas
exhausted by the fan is air withdrawn from the house; thus, if fan capacity can
be reduced without greatly reducing the radon reduction performance, the heating
and cooling penalty of the SSV system will also be reduced.
As apparent from Table 4, operation of the fan at high capacity almost
always resulted in lower indoor radon concentrations than did operation at
reduced capacity. The exceptions are Houses 15, 31 (with interior pipe #1), and
85 (exterior site 1), where performance was the same at both capacities. In the
majority of cases, operation at reduced capacity was sufficient to reduce the
house below 4 pCi/L, although operation at high capacity reduced levels even
further (generally below 2 pCi/L). The most dramatic improvements from operating
at high capacity occur in those cases where high-capacity operation is not
sufficient to reduce levels below about 2 pCi/L -- Houses 1, 31 (interior pipe
#2), 70 (exterior, and interior pipe #1), and 77 (all single-suction-point
cases). For these cases, the average radon level with the fan at high capacity
was 4.0 pCi/L, while the average at reduced capacity was 9.2 pCi/L. For the
other cases, where high-capacity operation achieved 2 pCi/L or less, reducing
the fan to about half capacity increased radon levels from an average of 1.1
pCi/L (at high capacity) to an average of 2.1 pCi/L (at reduced capacity). Only
in the case of House 77 (both interior points) did high-capacity operation
achieve less than 2 pCi/L while reduced-capacity operation did not reduce levels
below 4 pCi/L. In summary, it would appear that houses which are "easy" to treat
-------
with SSV can be reduced below 4 pCi/L with fans of this size operating at reduced
capacity; however, even with "easy" houses, there is usually (but not always)
some marginal additional radon reduction that can be achieved by operating at
high capacity.
The long-term objective defined by the Indoor Radon Abatement Act of 1988
is to reduce indoor radon concentrations to the levels which exist outdoors.
To approach this objective, operation of the fans at high capacity in all cases
would be preferred, since radon is generally reduced to lower levels at high
capacity. Comparing the system flows indicated in Table 4, flows average about
35 cfm (16 L/s) less when the fan is at low capacity. Electrical consumption
at low capacity is roughly one-third that at high capacity. Depending upon
specific assumptions regarding the fraction of the 35 cfm which is withdrawn from
the house, the local climate, and energy costs, the increased heating penalty
and electricity cost resulting from operating the fan at high capacity rather
than low capacity would be roughly $55 per year.
SSV SYSTEMS IN PRESSURE
In four of the Phase 2 slab-on-grade houses, the SSV fans were reversed
so that the system was pressurizing the sub-slab region rather than drawing
suction. In almost all cases, these sub-slab pressurization tests were conducted
with the exterior SSV systems, since there is easy access to the external fans;
only in House 70 was pressurization tested with an interior SSV system having
the fan mounted in the attic. The results are shown in Table 5. Table 5 also
presents the SSV pressurization results reported previously for the houses tested
during Phase 1. In all cases, the suction vs. pressure comparisons shown in the
table are with the suction and pressurization systems being operated at
comparable conditions: the same ventilation points, and the same fan capacity.
These results, now with seven houses, confirm the earlier conclusions drawn
from the three houses tested in the Phase 1 effort. In all cases, operation of
the SSV system in suction gives greater radon reductions (88-99%) than does
operation in pressure (43-90%). In several cases (Houses 31, 47, and 70), the
difference between suction and pressure is dramatic.
For sub-slab pressurization to be effective, the system probably must
create sufficient flows of fresh air under the slab to dilute the radon in the
sub-slab gas. Because the sub-slab is under pressure, any radon in the sub-slab
gas will be forced up into the house. Unlike sub-slab suction, where the system
works by reversing the flows of sub-slab gas, establishing a pressure field under
the slab may not be sufficient by itself with pressurization. With the geology
present in Dayton, flows in the sub-slab systems were apparently not sufficiently
great to provide the sub-slab dilution necessary to make pressurization work
effectively.
COMPARISON WITH RESULTS FROM OTHER STUDIES
The one slab-on-grade house tested by Infiltec in an EPA-sponsored project
in Maryland was a relatively large house (2,700 ft2, or 250 ma) with a block
foundation stem wall and with forced-air supply ducts under the slab, similar
-------
to House 77 in Ohio. The geology in Maryland was also similar to Ohio -- clay
soil, with a good layer of aggregate underneath the slab. However, the results
with interior SSV in the Maryland house were distinctly better than those
observed in House 77. A single interior SSV pipe, located approximately in the
middle of the slab and operating with a Kanalflakt T2 fan in suction, reduced
radon concentrations in the Maryland house from a pre-mitigation level of 6 to
21 pCi/L, down to 0.7 pCi/L, a reduction 90% or greater. In House 77, a single
interior pipe had been sufficient only to reduce levels from 16 to about 7 pCi/L
(see Table 3). It is not clear why the performance is so different between these
two houses.
Fourteen slab-on-grade houses tested in Florida (eight of which are
reported in Reference 3) had no aggregate underneath the slabs. The slabs rested
directly on fill material consisting of fine, compacted sand and clay which often
had limited permeability to soil gas flow, especially when moist. The
permeability of this fill may have been poorer than that of the native soil
underlying the aggregate under the Ohio and Maryland houses, since the fill is
specifically selected and compacted to support the elevated slabs characteristic
of Florida slab-on-grade houses. The Florida houses varied in size between 1,500
and 2,570 ft2 (140 and 240 ma), had either block or poured concrete foundation
walls, and did not have forced-air ducts underneath the slab. Several had sunken
living rooms. Performance of interior and exterior SSV in these houses has been
far less effective than it was in the Ohio and Maryland houses. In the Florida
houses, two or three suction pipes have been necessary to reduce radon
concentrations to post-mitigation levels of 3 to 9 pCi/L (generally 70 to 90+%
reduction), compared to the 81 to 99% reduction achieved with only one or two
pipes at the other sites. The two or three suction pipes were needed in Florida
despite efforts to improve performance by excavating pits underneath the slabs
where the pipes penetrated, and by using high-suction fans (capable of suctions
up to 33 in. HaO, or 8,100 Pa). In general, exterior SSV gave comparable
reductions to interior SSV in Florida, although, in one house, exterior SSV was
significantly less effective. These interior vs. exterior results appear
generally comparable to those from Ohio (where the two approaches also generally
compared well except in one house). However, in Florida, performance of exterior
systems was reduced by high air leakage rates into the system, a problem far less
pronounced in Ohio. In the Florida houses, the exterior suction pipes were often
inserted horizontally some distance toward the interior of the slab in order to
reduce leakage, making these exterior systems more similar to interior SSV.
It would appear that major obstructions such as sub-slab supply ducts (in
the Ohio and Maryland houses) can fairly readily be handled by SSV systems, as
long as there is a good aggregate layer. But where there is no aggregate and
where the soil (or fill) under the slab is not reasonably permeable, design of
the SSV system can be made more difficult, even when there are no apparent sub-
slab obstructions, as indicated by the experience in Florida.
-------
ACKNOWLEDGEMENTS
The authors wish to express their appreciation to: the homeowners who made
their homes available for this testing; the Dayton Regional Air Pollution Control
Agency, which provided substantial support during the selection of the houses
for testing; the Ohio Department of Health, which provided support throughout
the project; and Air Chek, Inc., which contacted candidate homeowners prior to
the house selection process.
REFERENCES
1. Henschel, D. B., A. G. Scott, W. 0. Findlay, and A. Robertson, Testing of
indoor radon reduction methods in 16 houses around Dayton, Ohio. In:
Proceedings: The 1988 Symposium on Radon and Radon Reduction Technology.
Vol. 1. Symposium Oral Papers. EPA-600/9-89-006a, U. S. Environmental
Protection Agency, Research Triangle Park, NC, and Washington, D. C., March
1989. 746 pp. (NTIS No. PB89-167480)
2. Findlay, W. 0., A. Robertson, and A. G. Scott, Testing of indoor radon
reduction techniques in central Ohio houses: Phase 1 (Winter 1987-1988).
EPA-600/8-89-071, U. S. Environmental Protection Agency, Research Triangle
Park, NC. July 1989. 315 pp. (NTIS No. PB89-219984)
3. Fowler, C. S., A. D. Williamson, B. E. Pyle, F. E. Belzer III, D. C.
Sanchez, and T. Brennan, Sub-slab depressurization demonstration in Polk
County, Florida, slab-on-grade houses. In: Proceedings: The 1988
Symposium on Radon and Radon Reduction Technology. Vol. 1. Symposium Oral
Papers. EPA-600/9-89-006a, U. S. Environmental Protection Agency, Research
Triangle Park, NC, and Washington, D. C., March 1989. 746 pp. (NTIS No.
PB89-167480)
-------
TABLE 1. SUMMARY OF SLAB-ON-GRADE HOUSES TESTED DURING PHASES 1 AND 2
OF THE RADON REDUCTION FIELD PROJECT IN DAYTON, OHIO
House
ID Ho.
15°
31"
47"
58
70
72
77
85
Pre-mitigation
Radon (pCi/Ll*
Slab size
17.5
20.9
16.3
12.6
14.0
22.9
24.1
21.6
15.4
17.9
16.0
10.3
(kitchen)
(bedroom)
(bedroom)
(kitchen)
(bedroom)
(kitchen)
(bedroom)
(liv. area)
(liv. area)
(liv. area)
(liv. area)
(liv. area)
1,700
950
950
1,100
1,400
2,350
1,700
2,600
1,100
Foundation
Material
Block
Poured
Poured
Poured
Poured
Block
Block
Block
Block
Sub-slab
Ducts?
Yes
Yes
Yes
Yes
Yes
No
No
Yes
Yes
Duct
Config.
Radial
Radial
Radial
Radial
Radial
Radial
Radial
Sunken
LR?
No
No
No
No
No
No
No
Yes
No
Pre-mitigation results are from alpha-track detectors exposed during the period December 19B7 -
March 1988 (Phase 1 houses) or November 1988 - January 1989 (Phase 2 houses), prior to completion
of mitigation systems. Detectors exposed in clusters of two (Phase 2) or three (Phase 1); figures
shown here are averages.
Houses initially tested during Phase 1.
Conversion factors: Bq/m1 = 37 x value in pCi/L. Area in ma = 0.093 x value in ft2.
-------
TABLE 2. BEST PERFORMANCE OF SUB-SLAB SUCTION IN SLAB-ON-GRADE HOUSES
IN DAYTON
Interior/
House
_MJ.
1
15'
31
47a
58
70
72
77
85
Exterior
SSV
Exterior
Exterior
Exterior
Exterior*
Interior
Exterior
Exterior
Interior
Exterior
Interior
Interior
Exterior
Interior
Exterior
Exterior
Radon Concentrations (oCi/L)
System Off
(Site
(Site
(#D£
(#1 &
(#1 &
(#1 &
(Site
(Site
Db
2)
2)"
2}'
2)«
1)
2)
Range
14-32
6-42
If
7-27
n
20-39
19-43
rr
12-A4
n
12-30
9-23
n
3-34
n
Mean
20
17
11
14
II
29
27
tl
23
II
22
15
II
12
II
.2
.9
.7
.6
.5
.5
.1
.7
.9
System
Range
2-
0-
0-
0-
0-
0-
0-
0-
2-
0-0
1-
1-
1-
1-
0-
6
1
1
1
1
2
1
2
5
.5
3
4
3
5
5
On
Mean
2.
0.
0.
0.
0.
0.
0.
0.
3.
0.
1.
1.
1.
2.
1.
8
4
5
1
3
9
5
8
7
2
6
7
7
5
5
% Reduction
in Mean
86
98
97
99
98
97
98
97
84
99
93
89
89
81
88
Suction
fin. H,0}
N/Ae
1.4
1.2
0.85
0.6
1.2
0.55
0.7
N/A
0.5-0.85
1.2
0.25-0.3
0.65-0.8
1.2
1.0
Flow
(cfm)
100
N/A
43
64
69
69
140
63
111
104
12
114
148
45
42
Testing completed during Phase 1; these results have been reported in References 1 and 2.
Sites 1 and 2 refer to different positions around the house where the one exterior suction point
was located.
#1 and #2 refer to individual suction points in homes having two-point systems installed.
Two suction points operating simultaneously to achieve these best results. In all other cases,
only one point is operating.
N/A - not available.
Fan operating at high speed in all cases.
Conversion factors:
Bq/m3 - 37 x value in pCi/L.
Pa = 248 x value in in. H,0.
L/s •= 0.47 x value in cfm.
-------
TABLE 3. EFFECT OF ONE VS. TWO SUCTION POINTS IN SSV SYSTEMS IN SLAB-ON-GRADE HOUSES
Interior/ Mean Mean Performance After
House Exterior pCi/L Point #1 Only Point #2 Only Both Points
No. SSV Before pCi/L % Red. pCi/L % Red. pCi/L % Red.
31 Interior 14.7 0.3 98 2.7 82 1.0 93
70 Interior 23.5 3.0 87 1.8 92 0.2 99
77 Interior 15.7 6.7 57 7.7 51 1.7 89
Exterior 15.7 6.0 62 5.7 64 1.7 89
Locations of suction points:
House 31: #1 in furnace room (center of house); #2 in bath (center rear of house).
House 70: #1 in furnace room (center); #2 in bedroom closet (in right front quadrant).
House 77: interior - #1 in laundry room (center of left half of house); #2 in
bedroom closet (center of right half).
exterior - #1 on rear wall, center of left half of house; #2 on rear
wall, center of right half.
Fan at high speed in all cases reported here (best results). Both suction points
connect to a single fan, except for exterior system in House 77, where each exterior
point has its own fan.
Conversion factor: Bq/mJ - 37 x value in pCi/L
-------
TABLE ft. EFFECT OF FAN CAPACITY ON THE PERFORMANCE OF SSV SYTEMS
(SLAB-ON-GRADE HOUSES)
House
No.
1
151
31
47'
58
70
72
77
85
Interior/
Exterior
SSV
Exterior
Exterior
(Site 1)
Exterior'
Interior
Interior
Interior
Exterior
Exterior
Interior
Exterior
Interior
Interior
Interior
Interior
Exterior
Exterior
Exterior
Interior
Interior
Interior
Exterior
(Site 1)
Exterior
(Site 2)
Suction
Point
Number
0
b
D
#1
#2
#1 & 2
0
B
»
b
#1
#2
#1 & 2
b
#1
#2
#1 & 2
#1
#2
#1 & 2
b
&
Mean
pCi/L
Before
20.2
17.9
14.7
tl
it
11
29.6
27.5
"
23.5
II
II
11
22.1
15.7
n
11
11
n
11
12.9
n
Mean Performance After
Hi eh
pCi/L
2.8
0.4
0.1
0.3
2.7
1.0
0.9
0.5
0.8
3.7
3.0
1.8
0.2
1.6
6.0
5.7
1.7
6.7
7.7
1.7
2.5
1.5
Capacity
% Red.
86
98
99
98
82
93
97
98
97
84
87
92
99
93
62
64
89
57
51
89
81
88
Low Cacacitv
oCi/L
10.1
0.2
1.4
0.3
13.7
2.9
1.0
2.2
6.0
6.5
2.0
1.7
3.2
2.1
9.7
5.5
2.5
2.8
% Red.
50
99
90
98
7
90
96
92
74
72
91
93
86
87
38
65
81
78
Suction
Highc
N/A
1.4
0.85
0.6
0.8
0.4-0.55
1.2
0.55
0.7
N/A
0.75
0.95
0.5-0.85
1.2
N/A
N/A
0.65-0.8
0.4
0.35
0.25-0.3
1.2
1.0
fin. H..O)
Lovc
N/A"
0.8
0.5
0.3
0.1
0.45
0.3
N/A
N/A
N/A
0.35
—
0.75
0.4-0.8
0.06
N/A
0.35
0.5
Flow
Hlghc
100
N/A
64
69
49
80
69
140
63
111
39
74
104
12
98
125
148
96
83
114
45
42
rcfm)
Lowc
• J-—
37
N/A
N/A
33
14
25
82
23
64
N/A
32
9
132
24
N/A
16
27
* Testing completed during Phase 1; these results have been reported in References 1 and 2.
" Only one suction point was included in these systems.
c Suctions, flows in system piping shown for both high and low fan capacities.
" N/A = not available.
Conversion factors: Bq/mJ = 37 x value in pCi/L. Pa = 248 x value in in. H,0. L/s = 0.47 x value in cfm.
-------
TABLE 5. EFFECT OF PRESSURIZATION VS. SUCTION IN SSV SYSTEMS
(SLAB-ON-GRADE HOUSES)
Average Radon Concentration fpCi/L)
% Reduction
Post-Mitigation
Pre-Mitigation
17.9
14.7
29.6
27.5
23.5
15.7
12.9
Suction
0.4
0.4
2.9
0.5
0.2
1.7
1.6
Pressure
1.8
4.2
16.0
5.7
13.4
3.1
2.6
Suction Pressure
15' 17.9 0.4 1.8 98 90
31" 14.7 0.4 4.2 97 71
90 46
58 27.5 0.5 5.7 98 79
70" 23.5 0.2 13.4 99 43
77* 15.7 1.7 3.1 89 80
85" 12.9 1.6 2.6 88 80
' Testing completed during Phase 1; these results reported in References 1 and 2.
" Performance shown is for a two-point interior SSV system. All other results reported
here are for exterior SSV systems.
c Results obtained with the two-point exterior system.
d Results obtained with one-point exterior system, fan located at Site 2.
Conversion factor: Bq/m1 - 37 x value in pCi/L.
-------
VII-2
RADON MITIGATION EXPERIENCE IN HOUSES
WITH BASEMENTS AND ADJOINING CRAWL SPACES
by: Marc Messing
INFILTEC
Falls Church, VA 22041
D. Bruce Henschel
Air and Energy Engineering Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
ABSTRACT
Active soil depressurization systems were installed in four basement houses
with adjoining crawl spaces in Maryland. In addition, existing soil depressur-
ization systems were modified in two additional basement-plus-crawl-space houses.
These six houses were selected to include both good and poor communication
beneath the basement slab, and different degrees of importance of the crawl space
as a source of the indoor radon. The radon reduction effectiveness was compared
for: depressurization only under the basement slab; depressurization only under
a polyethylene liner over the unpaved crawl-space floor; and simultaneous
depressurization under both the basement slab and the crawl-space liner. The
objective of this testing was to identify under what conditions treatment of the
basement alone might provide sufficient radon reductions in houses of this
substructure, and what incremental benefits might be achieved by also treating
the crawl space.
The results suggest that, when there is excellent communication beneath
the basement slab, basement-plus-crawl-space houses can sometimes be treated by
basement sub-slab depressurization (SSD) alone, even when radon levels in the
crawl space indicate that the crawl space could be a source. With excellent
communication, the incremental benefit of including sub-liner depressurization
in the crawl space, in addition to basement SSD, can sometimes be limited
(although the two in conjunction may still be needed if the goal is to achieve
"ambient" levels indoors). However, when communication beneath the basement slab
is poor, and if the crawl space is an important source, then sub-liner
depressurization can provide potentially significant additional radon reductions
compared to basement SSD alone.
This paper has been reviewed in accordance with the U. S. Environmental
Protection Agency's peer and administrative review policies, and has been
approved for presentation and publication.
-------
INTRODUCTION
Twelve existing houses in Maryland were selected to study cost-effective
radon mitigation methods in three types of houses: five houses with basements
and adjoining slabs on grade; six houses with basements and adjoining crawl
spaces; and one house which was slab on grade with heating ducts beneath the
slab. All of these houses had previously reported radon levels between 10 and
70 pCi/L (370 and 2600 Bq/m3)*.
Active soil depressurization systems were installed in all of the houses.
For the basement houses with adjoining wings, one sub-slab depressurization (BSD)
pipe was installed beneath the basement slab, and a second suction pipe was
installed either beneath the adjoining slab or through polyethylene sheeting
placed over the floor of the adjoining crawl space. Radon concentrations were
measured in the basement and on the first floor with: none of the suction pipes
operating; SSD applied to the basement slab only; depressurization beneath the
crawl-space liner or beneath the adjoining slab on grade only; and depressuriza-
tion beneath both the lower slab and the crawl-space liner or upper slab. The
objective of this testing was to identify under what conditions treatment of the
basement alone might provide sufficient radon reductions in such basement-plus-
adjoining-wing houses, and what additional reductions might be achieved by also
treating the adjoining wing. Diagnostic testing (including sub-slab or sub-
liner radon concentrations, and sub-slab suction field extensions and air flows)
was conducted to help understand the conditions influencing the performance of
these systems.
This paper addresses the testing on the six basement-plus-crawl-space
houses. These six houses included four which had had no prior radon mitigation
system, and two which had existing systems which were not fully effective.
MEASUREMENT METHODS
INDOOR RADON MEASUREMENTS
For short-term measurements of radon mitigation system performance, radon
gas was measured using a Pylon Model AB-5 continuous radon monitor equipped with
either a 17.4 in.3 (285 mL) Lucas scintillation cell or a PRD-1 passive radon
detector. The Pylons were programmed to measure radon hourly. One Pylon was
placed in the basement, and one in the living area above the crawl space,
generally for about 2 weeks; during this time, the mitigation system would be
cycled through its range of operating conditions (system off, basement treatment
only, crawl-space treatment only, treatment of both wings), with the system being
left at a given condition for at least 48 hours.
For a long-term measurement of SSD performance, 3-month alpha-track
detector measurements are being repeated for at least 1 year with the systems
operating. These long-term results are not covered in this paper.
Radon in Bq/m3 - 37 x value in pCi/L.
-------
DIAGNOSTIC TESTING
In selecting houses for this study, it was desired to cover a range of the
key variables, including the communication under the basement slab, and the
relative importance of the basement vs. the crawl space as a radon source.
Knowledge of these parameters is also important for understanding the conditions
under which basement-only treatment is sufficient, and what additional benefit
can be derived by also treating the crawl space.
Sub-slab Pressure and Flow Measurements
To characterize communication under the basement slab, sub-slab pressure
and flow measurements were taken with a "Blower Floor," consisting of a high-
suction vacuum cleaner, a calibrated flow meter, and a digital micromanometer.
Where a sump existed, the sump cover was removed (if feasible) and a plate
with a 1.5-in. (4-cm) opening was secured over the sump with duct tape. The
vacuum cleaner nozzle was sealed into the 1.5-in. opening to depressurize the
sump. Measurements of induced air flow in the vacuum were taken as the suction
in the sump was measured using the micromanometer. The suction underneath the
basement slab was also measured through several 0.25-in. (0.6-cm) test holes
drilled through the slab at selected locations, usually in the basement corners.
In houses with no sump, a 0.5-in. (1.25-cm) suction hole was drilled
through the slab (at a location which appeared promising for the ultimate SSD
suction point). Suction was drawn on the sub-slab through this hole with the
vacuum cleaner. Flow in the vacuum cleaner was measured, as was the differential
pressure across the slab through the 0.25-in. (0.6-cm) test holes. One test hole
was located 4 in. (10 cm) from the suction hole, and several test holes were
usually located in the basement corners.
Radon Measurements Under Basement Slab
To assess the source strength under the basement, sub-slab radon
measurements were taken through 0.25-in. (0.6 cm) holes drilled in each corner
of the basement slab area. Sub-slab gas was pumped through a 17.4 in.3 (285 mL)
Lucas cell at approximately 0.03 cfm (1 L/min), and photomultiplier counts were
recorded at 1-minute intervals. Although this "sniffing" technique does not
allow the radon to reach equilibrium with its decay products, the sensitivity
of this non-equilibrium measurement system was determined experimentally to be
about 1.4 times the sensitivity of an equilibrium situation; this sensitivity
was used to estimate sub-slab values.
Radon Measurements in Crawl-Space Soil
As a rough estimate of the source strength in the crawl space, radon
measurements were made near the surface of the crawl-space soil. Where the
crawl-space floor was covered with a preexisting vapor barrier, a 0.25-in. (0.6
cm) hole was made through the vapor barrier (generally within arm's reach of the
access door to the crawl space), and sub-barrier samples were taken with the
sniffing technique described above. Where there was no existing vapor barrier
-------
in the crawl space, the air sampling tube was inserted approximately 6 in. into
the soil, or into the aggregate on top of the soil, to collect the soil gas
sample. Generally, no measurements were taken of radon levels in the breathable
area of the crawl space.
Post-Mitigation Diagnostics
To define the operation of the SSD system after installation, measurements
were made of the suction and flows in the system piping, and the extension of
the suction field created under the basement slab by the system.
DESCRIPTION OF STUDY HOUSES
The six houses included in this testing are summarized in Table 1. In
addition to certain house construction characteristics and the pre-mitigation
radon concentrations, the table summarizes results from the pre-mitigation
diagnostic testing.
In all cases, the crawl-space floor consisted of bare soil (often with an
existing vapor barrier) or of aggregate on top of soil {with no vapor barrier).
Only one house (House 208) was a split level, with a half-depth basement
adjoining a relatively low crawl space; the remainder of the houses were ranches.
The basements were somewhat larger in floor area than were the crawl spaces, by
factors ranging from 1.05 to 2.2.
Two of the houses (Houses 208 and 277) already had operating mitigation
systems which had reduced radon levels to the range of 4 - 10 pCi/L at the time
this work began. House 208 had a one-pipe basement SSD system. House 277 had
a one-pipe basement SSD system tied into a sub-liner depressurization system in
the crawl space, connected to a single fan.
House 208 is a split level house. The lower half is approximately 4 ft
below grade in the rear. The upper level is above a crawl space which is at the
same level as the basement slab. The crawl space and upper level are separated
from the lower level by the block foundation wall and an access door. There is
no sump. The total floor area of the basement slab and the crawl space (the
"house footprint") is 1375 ftz (130 m2) , with a basement/crawl area ratio of
1.05.
House 277 is a ranch-style house built over an L-shaped basement and an
interlocking L-shaped crawl space. The basement is unfinished and has no sump.
The crawl space is at the same grade as the basement floor. The crawl-space
floor is uneven and has been covered (as part of the earlier mitigation effort)
with a sealed plastic barrier. The foundation walls are hollow block.
No data are available on pre-mitigation sub-slab radon levels in these two
houses. As indicated in Table 1, both had poor airflow beneath the basement
slab. As part of this project, additional basement SSD pipes were installed;
these pipes improved the distribution of the suction field beneath the basement
-------
slab and resulted in effective mitigation, as discussed later, greatly improving
the performance of the original installations.
House 582 has a basement which is divided into two parts: a finished
family room; and an unfinished utility room with washer/dryer, furnace, and sump.
The crawl space is open to both parts of the basement. The house has a garage
on grade which is adjacent to the crawl space, and there are living spaces over
the basement and crawl space, but not over the garage. This house has excellent
sub-slab air flow. Of particular interest, higher radon concentrations were
measured in the crawl-space soil than were found beneath the basement slab, as
shown in Table 1, suggesting that the crawl space might be a major contributor
to the indoor radon. The total footprint is 825 ft2 (77 m2), and the basement/
crawl area ratio is 1.8.
Houses 1165 and 1248 were both characterized by relatively high sub-slab
radon levels for this area (>1,000 pCi/L) and excellent sub-slab flow. By
comparison, radon levels in the crawl-space soil were only moderate, suggesting
that the basement may be the primary source. House 1165 has a basement, of which
approximately 80% is finished living space and the remainder is a utility/work
room. There is no sump. There is also a large crawl space open to the basement,
and a garage on grade adjacent to the crawl space. There are living areas above
both the basement and the crawl space, although there is no living space above
the garage. There are both duct and plumbing penetrations between the crawl
space and the living area above it, and there is a full bath in the basement.
The footprint is 1375 ft2 (130 m2), and the area ratio 1.7.
House 1248 has a relatively small basement with poured concrete walls and
an adjoining crawl space which is open to the basement. There are living spaces
above both the crawl space and the basement, with forced air ducts in both the
basement and the crawl space. The basement is a deep basement (rather than a
walkout), and there is a sump on the side of the basement opposite the crawl
space. There are no adjacent slabs on grade. With a footprint of 900 ft2 (84 m2)
and an area ratio of 1.8, House 1248 is approximately 65% of the size of House
1165 (both in the area of the basement slab and in the area of the crawl space
surface).
House 1357 is a brick and block colonial, with a garage on grade, a deep
basement, and a crawl space between the garage and the basement. The foundation
walls are hollow cinder block. The crawl space has an uneven surface of packed
soil and is open to the basement; there is no sump. There is forced air heat,
and there are several plumbing penetrations through the basement slab (for a
washing machine and sink). House 1357 was characterized by high radon levels
under the basement slab, and poor sub-slab airflow; in this instance, there was
no measurable sub-slab airflow. Radon levels in the crawl-space soil were not
as high as at most of the other houses; however, concentrations in the crawl-
space air were once measured at 25 pCi/L at a time when the levels in the
basement were only 5 pCi/L, suggesting that the crawl space could still be a
contributor. The house footprint is 875 ft2 (81 m2), and the area ratio is 2.2.
-------
DESCRIPTION OF REMEDIATION SYSTEMS
In the four houses not having sumps, one or more SSD pipes were installed
through the basement slab. Houses 1165 and 1357 each received one-pipe systems;
House 208 received an additional pipe and fan to supplement the previous system,
thus winding up with a two-pipe system, each pipe having its own fan; House 277
received an additional two pipes tied into the previous system, to create a
three-pipe system. Typically, about 0.5 ft3 (15 L) of aggregate was excavated
from under the slab where the pipes penetrated. In the two houses which had
sumps (Houses 582 and 1248), the sump was fitted with a permanent radon sump
cover and a suction pipe. In all houses, the suction pipe was thin-walled 6-
in. (10-cm) PVC pipe. Except for House 208 (which had two independent SSD pipe/
stack/fan systems and no crawl-space sub-liner system), the SSD pipe(s) were T'd
together with the sub-liner suction pipe from the crawl space, and connected to
a single in-line fan mounted outside the house or in the garage attic, with the
exhaust released above the roof in all houses. The fans had a rated capacity
of 180 cfm (85 L/s) at 0.5 in. H20 (125 Pa). All basement cracks were caulked
along accessible floor/wall joints (i.e., not behind finished walls or other
inaccessible locations).
Sub-liner depressurization systems were installed in each house except
House 208. Except in House 277, the liner material was 8-mil (0.2-mm) high-
durability polyethylene sheeting, laid to form a barrier over the entire crawl-
space floor. In the final installations, the liner was sealed by overlapping
the individual sheets (by 6 to 12 in. , or 0.15 to 0.3 m), and sealing these seams
with urethane caulk; the junctions of the sheeting with the perimeter foundation
wall and with interior piers were sealed in the same manner. A single 4-in. (10-
cm) PVC suction pipe was inserted through the plastic liner at one location, with
a plywood "flange" around the pipe underneath the liner to support the pipe and
to lift the polyethylene off the soil at that location to aid in distribution
of the suction field, wherever possible, the measurements of mitigation system
performance included measurements with and without the sheeting sealed around
the perimeter wall, to evaluate the benefits of this sometimes-difficult
installation step.
In House 277, where the sub-liner system had been installed by an earlier
mitigator as part of the original mitigation effort, the liner was 20-mil (0.5-
mm) plastic, and a loop of perforated piping was located around the perimeter
of the crawl space under the plastic. Suction was drawn on this loop. The
plastic was already sealed at seams and around the perimeter, so that testing
without perimeter sealing was not possible during this project.
RESULTS
The radon reduction results from the testing in the six houses are
summarized in Table 2. These results represent at least 48 hours of measurements
at each test condition, using the continuous Pylon monitors.
-------
EFFECTS OF TREATING THE BASEMENT SLAB ONLY
Applying a vacuum beneath only the basement slab reduced the radon levels
both in the basement and in the living area above the crawl space in all of the
houses except House 1357. With the exception of that house, reductions in the
basement were consistently high (82 to 98%), reducing concentrations to below
1 pCi/L in these short-term measurements. The one house having less than 90%
reduction in the basement (House 582) had the relatively low percentage reduction
of 82%, at least in part because the pre-mitigation value was below 4 pCi/L.
(In all of the houses, pre-mitigation values at the time of these Pylon
measurements were much lower than they had been when the earlier, cold-weather
charcoal detector measurements had been made, even though the Pylon measurements
were largely completed during the month of February.) In the houses other than
1357 for which Pylon measurements were successfully completed in the living area
above the crawl space, radon reductions were about 80% in the living area with
basement-only treatment, reducing concentrations to 1.2 pCi/L and less.
In House 582 -- where the crawl space appears to be an important radon
source but where sub-slab communication is excellent, as discussed earlier --
basement-only treatment did about as well in reducing levels upstairs as it did
in House 1165, where the communication is also excellent but where the crawl
space seems less of a source. Basement-only treatment in House 582 appeared only
slightly poorer in reducing levels in the basement, compared to Houses 1165 and
1248 (also good communication, crawl space less of a source). Based upon these
fairly good results in House 582, it would appear that -- when communication is
very good -- basement-only treatment may be sufficient even when the crawl space
is an important radon source.
In House 1357, SSD under the basement slab reduced basement levels about
40%, much less than in the other houses. Concentrations in the living space
above the crawl actually appeared to increase from 2.5 to 3.5 pCi/L when basement
SSD was applied (although, given the variability of radon measurements, this 1
pCi/L differential is so small that it is not certain that any real change had
occurred at all). This relatively poor performance in House 1357 with basement
SSD only might intuitively be attributed to the very poor sub-slab communication
(and possibly the relative importance of the crawl space as a radon source, as
suggested by the 25 pCi/L once measured in the crawl-space air). House 208 also
had poor sub-slab communication (although not as poor as House 1357, which gave
no airflow during the pre-mitigation diagnostic testing); House 208 had still
had elevated radon levels with its original one-pipe SSD system, and achieved
the high removals shown in Table 2 only because of the additional suction pipe
and fan added as part of this project.
EFFECTS OF TREATING CRAWL SPACE ONLY
Sub-Liner Depressurizatton with the Liner Sealed
As shown in Table 2, sub-liner depressurization in the crawl spaces,
without any treatment of the basement, effectively reduced radon levels in both
the basements and the living areas of all five test houses in which this approach
was tested. (House 208 did not receive crawl-space treatment; House 277 was not
-------
tested in the crawl-space-only mode.) The results shown in the table are for
the case where the liner was sealed at seams and around the perimeter.
Reductions ranged from 38 to 73% in the basement, and from 15 to 88% in the
living area above the crawl space. Except for House 1357, reductions in the
basement from crawl-space-only treatment were not as great as the reductions
achieved by basement-only treatment, as might be expected. For the two houses
other than 1357 where measurements were successfully completed in the living area
with both mitigation approaches, crawl-space-only treatment gave living-area
radon reductions which were comparable to basement-only treatment in one house
(House 582, 88 vs. 84%), and substantially poorer in the other house (House 1165,
15 vs. 82%). The fact that crawl-space-only treatment is so much poorer than
basement-only treatment in House 1165 suggests that, in this house, either: a)
most of the radon in the living area was probably flowing upstairs from the
basement; or b) the basement SSD pipe was being very effective in achieving "site
ventilation" (i.e., in extending its suction field underneath the crawl-space
soil).
Crawl-space-only treatment in House 1357 gave significantly greater radon
reductions in the basement than did basement-only treatment (73 vs. 41%). This
result would appear to be consistent with the expectations that the crawl space
is an important source of the radon in the house, and that poor sub-slab
communication would reduce the effectiveness of a one-point SSD system in the
basement. The levels in the living area above the crawl space were reduced to
1.3 pCi/L with the crawl-space-only system, a 48% reduction from the pre-
mitigation level of 2.5 pCi/L; the basement-only system had given no apparent
reduction in the living area.
Sub-Liner Depressurization without Sealing the Liner
The results discussed above are for the final sub-liner depressurization
installation in each house, which involved sealing the liner at seams and around
the perimeter. Prior to that complete sealing, the sub-liner depressurization
system was tested without the liner's being sealed around the perimeter (although
it was still sealed between the sheets of polyethylene). Testing without the
liner sealed was conducted only for the crawl-space-only case, not in conjunction
with basement SSD. Comparison of the radon reductions achieved with the crawl-
space-only systems before and after liner sealing indicates that sealing improved
reductions in all four houses where before/after measurements could be made, both
in the basement and upstairs. Basement reductions increased from a pre-sealing
range of 0 to 60%, to the post-sealing range of 38 to 73%; upstairs reductions
increased from 0 - 56% to 15 - 88%.
It is emphasized that -- although the differences in the percentage
reductions may appear significant -- these differences often represent only a
few pCi/L. Given the variability in indoor radon levels with time, it is not
clear if the apparent differences reflect a real effect of liner sealing.
EFFECTS OF TREATING BOTH THE BASEMENT AND THE CRAWL SPACE
As shown in Table 2, simultaneous SSD in the basement and sub-liner
depressurization in the crawl space was clearly superior to basement-only
-------
treatment in House 1357. Radon reductions both in the basement and in the living
area above the crawl space are significantly higher with the combined treatment
compared to basement-only treatment (90 vs. 41% in the basement, 84 vs. 0%
upstairs). Again, this result is consistent with the diagnostic results
indicating that sub-slab communication is poor, and suggesting that the crawl
space might be an important source in that house. In one other house (House
582), which had excellent sub-slab communication, simultaneous treatment appears
somewhat superior to basement-only treatment (92 vs. 82% in the basement, 96 vs.
84% above the crawl space). However, the differential between the radon levels
for simultaneous vs. basement-only treatment is so small in House 582 (fractions
of a pCi/L), that measurement uncertainty and temporal variation in radon levels
could easily explain the differences; thus, it is not clear that there is truly
a significant difference in the performance of the two approaches in this house.
In the other two houses for which complete data are available (Houses 1165 and
1248) -- both of which had excellent communication -- basement-only treatment
appears at least as effective as simultaneous treatment in reducing the radon
levels both in the basement and upstairs.
In summary, in the one house where sub-slab communication is poor and the
crawl space might be an important source, treatment of both wings appears
necessary. (Additional suction pipes in the basement might also have improved
the performance of the basement-only system; post-mitigation suction field
extension measurements in this house confirm that the suction was not extending
underneath the entire slab.) In the three houses having excellent sub-slab
communication, treatment of the crawl space appears either unnecessary, or of
reduced incremental value.
Simultaneous treatment of the basement and the crawl space was always
superior to crawl-space-only treatment, providing greater reductions both in the
basement and upstairs. Reductions of 38 to 73% in the basement with crawl-space-
only treatment increased to 90 to 96% with simultaneous treatment; reductions
of 15 to 88% in the living area above the crawl space increased to 84 to 96%.
Therefore -- although crawl-space-only treatment was sufficient to reduce levels
in these houses below 4 pCi/L on both floors, due to the low pre-mitigation
values --it appears always beneficial to supplement a sub-liner depressurization
with SSD in the basement, regardless of the sub-slab communication or the
importance of the crawl space as a source.
In House 277 -- which has poor communication -- the three-pipe basement
SSD system, combined with the sub-liner depressurization system having a
perforated piping loop under the liner, is the most extensive of the systems
tested under this project. Yet this system gave the lowest percentage reductions
(82% in the basement, 84% upstairs) and the highest residual radon levels (1.7
pCi/L in the basement, 0.7 pCi/L upstairs) of any of the basement-plus-crawl-
space systems. Even though the residual levels are less than 2 pCi/L, these
results suggest that, with poor-communication houses, more extensive systems
(e.g., multi-pipe basement SSD installations, crawl-space sub-liner systems with
perforated piping under the liner) will sometimes be needed. The system at House
277 was not tested in the basement-only or the crawl-space-only mode, so that
it is not possible to assess the relative contributions of the basement and
crawl-space components of the system.
-------
CONCLUSIONS
The four conclusions listed below are based on the results from the six
basement-plus-crawl-space houses tested in this study. These conclusions must
be qualified, recognizing that: the number of houses tested was limited,
considering the range of house variables to be tested (e.g., communication,
importance of crawl space as a source); and the pre-mitigation radon concentra-
tions existing at the time of testing (February 1989 in most cases) were low,
so that changes in the radon concentrations caused by the mitigation systems were
small (in terms of pCi/L) and within the background variability created by
temporal radon variations and measurement uncertainty.
1) When communication under the basement slab is excellent (as defined in
Table 1), and when the crawl space is not a primary contributor to the
indoor radon levels: basement-plus-crawl-space houses can sometimes be
treated by basement SSD alone. Under these circumstances, the incremental
benefit of including sub-liner depressurization in the crawl space, in
addition to basement SSD, may sometimes be limited. This result can be
true even where diagnostic measurements indicate radon levels in the crawl
space several times higher than those in the living areas of the house.
However, even in houses with excellent communication, basement SSD may have
to be supplemented with sub-liner treatment in order to achieve "ambient"
radon concentrations indoors (i.e., a fraction of a pCi/L).
2) When communication under the basement slab is poor (as defined in Table
1), and when the crawl space might be a primary contributor to indoor
radon: sub-liner depressurization in the crawl space, in conjunction with
basement SSD, can provide potentially significant additional radon
reductions compared to basement SSD alone. This study did not definitively
address the potential for achieving these increased radon reductions
through an improved design of the basement SSD system (e.g., an increased
number of suction pipes), rather than by adding sub-liner depressurization.
3) Combined treatment of the basement and the crawl space always provided
greater radon reductions in this study than did crawl-space treatment
alone. Thus, it would not usually be desirable to install a sub-liner
depressurization system in the crawl space without also tying in basement
SSD pipes. In most cases, crawl-space treatment alone is not as effective
as is basement treatment alone; only in one house, which had poor sub-slab
communication in the basement and high radon concentrations in the crawl
space, did sub-liner depressurization in the crawl space prove more
effective in reducing indoor radon levels than did SSD applied to the
basement slab.
4) When applying sub-liner depressurization in a crawl space, with a single
suction pipe penetrating the liner at a central location, sealing the
polyethylene sheeting to the crawl space perimeter walls (as well as at
seams between sheets) appeared to provide roughly a 10 to 40 percentage
point improvement over the reductions achieved with the liner sealed only
at seams between sheets.
-------
If these conclusions from the six study houses can be extrapolated to the
general housing population, they are relevant to several different aspects of
commercial radon remediation. First, they clearly suggest that simple sub-slab
vacuums applied to the lowest levels of basement houses with adjoining crawl
space are likely to reduce the radon levels substantially, especially in houses
with good sub-slab flow beneath the basement slab. Therefore -- where the goal
of remediation is to reduce radon levels below 4 pCi/L or to achieve the most
cost-effective remediation (as has commonly been the case) -- then SSD beneath
the basement slab is likely to be an appropriate approach.
However, where the goal of remediation is to achieve ambient levels (a
fraction of a pCi/L) indoors, these findings suggest that, even under the best
of conditions, it may sometimes be necessary to supplement SSD systems with sub-
liner depressurization of exposed crawl space soil. Since installation of sub-
liner systems is generally difficult and labor intensive, adding crawl-space
treatment to a basement SSD system may be expected to increase the cost of
remediation by more than 100% in many instances.
It is important to note that the experience with House 1357 clearly
illustrates that, in some instances, the crawl space may be a primary source of
indoor radon. In such cases -- and when, in addition, communication under the
basement slab is poor -- the likelihood is increased that a sub-slab vacuum
beneath the basement slab will have to be supplemented by sub-liner
depressurization, even to achieve 4 pCi/L.
For the professional diagnostician and mitigator, therefore, it is
important to utilize appropriate diagnostic tests to assess when basement SSD
by itself will potentially not be sufficient. The diagnostic tests used in this
study suggest that sub-slab flow beneath the basement slab, together with radon
measurements in the crawl space air or soil gas, may be useful predictors of when
crawl-space treatment will also be desirable.
It is re-emphasized that additional testing -- on a larger number of
basement-plus-crawl-space houses, with more adjustments to the mitigation system
design/operating variables, and with higher pre-mitigation radon levels -- would
be necessary to confirm these conclusions.
ACKNOWLEDGEMENTS
The authors wish to express their appreciation to: the homeowners who made
their homes available for this testing; and the Maryland Department of the
Environment, which provided substantial support during the selection of houses
for testing.
-------
TABLE 1. SUMMARY OF BASEMENT-PLUS-CRAWL-SPACE HOUSES TESTED IN MARYLAND
House
No.
208
277
582
1165
1248
1357
Foundation
Block
Block
Poured
Block
Poured
Block
Pre-mit.
Indoor Rn
(pCi/D*
10.2-23.2
7.6-27.8
9.8-17.3
13.2-13.8
25.3-41.8
15.1-18.4
Source Concentration (pCi/L)
Sub-slab Soil
(Basement) (Crawl)
117- 225
N/A
138- 305
193-1570
1725-3674
600-1800
N/A
N/A
888
144
374
67
Basement
Sub-slab
Airflow**
Poor
Poor
Excellent
Excellent
Excellent
Poor
* Pre-mitigation indoor radon levels as determined by charcoal detector
measurements in the basement during cold weather. Measurements made by
homeowner prior to EPA study, and by EPA during house selection process.
** "Poor" airflow is defined here as less than 30 cfm (14 L/s) drawn by the
diagnostic vacuum cleaner through a 0.5-in. (1.25-cm) suction hole when 1.5
in. H20 (370 Pa) suction is being maintained under the slab at a test hole 4
in. (10 cm) away (when no sump present). "Excellent" airflow is defined as
greater than 50 cfm (24 L/s) under the same conditions. When sump is present,
suction is drawn through a 1.5-in. (4-cm) hole through sump cover, -0.1 in.
H20 in sump; sumps connecting to drain tiles usually result in excellent flow.
TABLE 2. SUMMARY OF RADON REDUCTIONS ACHIEVED (PYLON MEASUREMENTS)
House
No.
208
277
582
1165
1248
1357
System Off*
Avg. pCi/L**
Bsmt L.A.
Bsmt SSD Only
Avg. pCi/L
Bsmt L.A.
8.9
9.7
3.8
4.8
5.1
10.5
N/A
4.4
2.5
3.4
N/A
2.5
0.5
N/A
0.7
0.1
0.4
6.2
1.2
N/A
0.4
0.6
N/A
3.5
Crawl Sub-Liner Only
Avg. pCi/L
Bsmt
***
N/A
1.4
3.0
2.8
2.8
***
N/A
0.3
2.9
N/A
1.3
Bsmt + Crawl
Avg. PCi/L
Bsmt LA.
***
0.7
0.1
0.4
N/A
0.4
1.7
0.3
0.2
0.5
1.1
* " System-off values are Pylon measurements made back-to-back with the various
system-on tests reported in the table. These Pylon values differ from the
pre-mitigation charcoal detector levels reported in Table 1 probably due to
temporal effects.
** Radon levels are the average generally of at least 48 hours of hourly Pylon
measurements in the basement ("Bsmt") and in the living area ("L.A.") above
the crawl space. Upstairs Pylon malfunctioned in House 1248.
***House 208 did not include sub-liner suction in crawl space.
N/A - Not available.
Conversion factors: Bq/m3 - 37 x value in pCi/L; cm - 2.54 x value in in. ;
L/s - 0.47 x value in cfm; Pa - 248 x value in in. H20.
-------
VII-3
ENGINEERING DESIGN CRITERIA FOR SUB-SLAB DEPRESSURIZATION
SYSTEMS IN LOW PERMEABILITY SOILS
by: Charles S. Fowler, Ashley D. Williamson, Bobby
Fyle, Frank E. Belzer III, Raymond N. Coker
Southern Research Institute
Birmingham, AL 55305
David C. Sanchez
U.S. Environmental Protection Agency, AEERL
Research Triangle Park, NC 27711
Terry Brennan
Camroden Associates
Oriskany, NY 13424
ABSTRACT
Radon mitigation by sub-slab depressurization (SSD) in houses built over
soils with low permeability (less than 10"5 cm2) requires different techniques
than the more common case of slabs laid over permeable gravel beds. Over the
past 2 years 14 slab-on-grade houses in South Central Florida with compacted
soil fills and initial indoor radon concentrations between 400 and 4000 Bq/m3
(10-100 pCi/L) were mitigated using SSD systems. Studies of these installations
indicate that the completion and analysis of certain diagnostic tests can yield
parameters useful to the mitigator for the design and installation of
successful mitigation systems. Results from the study houses have been
combined into tables and graphs that can be used to help determine recommended
numbers and placement criteria of the suction holes. Fan and pipe size
selection is assisted by other tabulated or derived information. Installation
techniques are suggested to enhance the system operation and effectiveness.
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
Sub-slab depressurization (SSD) is generally the most common and most
effective radon mitigation strategy employed in basement and slab-on-grade
houses. In many areas of the country, the standard building practice is to
place a layer of coarse gravel directly beneath a vapor barrier before pouring
the slab. When this has been done, an SSD system is usually quite effective
-------
because of the good permeability and communication afforded by the gravel
layer. However, many older houses were built before using gravel became a
common practice and in some areas of the country gravel is not readily
available. In these houses the slabs are poured over either the native soil or
a fill soil that has been compacted to some degree to prevent settling away
from the slab once the concrete has hardened. Host of the time such a soil
fill has much lower permeability to air flow (usually less than 10"5 cm2). In
such instances an SSD system will not operate as effectively as it would over a
coarse aggregate bed. Since much of the literature (1-4) about SSD systems
addresses slabs poured over gravel, guidance in the installation of SSD systems
over low permeability soils has generally been lacking. Ericson et al. (5) in
Sweden and other researchers (6) have reported cases of low permeability
beneath the slabs and either have made some generic observations about the
average slab area affected by given suction holes or have offered unique
remedies found to work in specific houses. However, no uniform guidance
uniquely addressing design and installation strategies for solving this problem
seems to exist.
The South Central Florida (Folk County) area is one area in the U.S. where
coarse gravel is not readily available. The customary building practice is to
prepare a base of fill soil, overlay it with a vapor barrier, and then pour the
slab. This practice usually produces somewhat low, but fairly uniform,
communication in sub-slab gas flow. From December 1987 to September 1989, 14
single-story slab-on-grade houses with living areas of about 120-240 m2 (1300-
2600 ft2) and initial indoor radon concentrations of 400-4,000 Bq/m3 (10-100
pCi/L) have been mitigated with (SSD) systems. The systems have ranged from
central- and perimeter-located single suction hole systems to up to four
central and/or five perimeter suction holes, with a variety of combinations.
Suction pits ranged from no pits to up to 0.05-0.06 m3 (12-15 gal) in size.
Different sizes of fans and pipes have been installed. Suction holes were
drilled through the slab and through stem walls under the slab. Fans have been
located in attics and outside the houses.
There is continuing research being conducted relevant to design criteria
for similar sub-slab mitigation systems in the same and other areas of Florida,
across the U.S., and in other parts of the world. The University of Florida,
in particular, is contributing much complementary research to houses in a
different part of the state. The purpose of this report is to outline some
aids in design and installation of SSD radon mitigation systems for use by
mitigators working on houses where slabs are laid over uniform low-permeability
soils.
BACKGROUND INFORMATION
PROBLEM ASSESSMENT
Before a mitigator or homeowner starts to design a radon mitigation
system, it should of course be established that there is an indoor radon
problem. It is reasonable and ethical for a mitigator to communicate to the
homeowner that EPA has published guidance (7) for making reproducible
measurements of radon concentrations in residences, including recommendations
-------
for using the results to make well-informed decisions about the need for
additional measurements or remedial action. Other complimentary publications
that give more detail and updated information on the specific use of
measurement techniques are also available (8,9).
Once it is determined that the house in fact does have elevated radon
concentrations, before any other action is taken, certain basic house
information needs to be obtained. Some of the most crucial elements to note
include the area of the slab(s), the sub-slab media or aggregate, the floor and
ceiling covering, the existence and characteristics of any interior load-
bearing walls, and the existence of interior footings. Any information that
can be determined about the slab/wall interface is important, as is the
existence of any slab cracks and utility penetrations including plumbing lines.
The type of heating and air conditioning and the location of the duct work and
returns are also helpful to know. Some of this information can be obtained
from the homeowner, from either existing knowledge or plans, documents, or
pictures taken during construction or renovation. The rest may be visually
noted or measured during a visit to the house.
A visit and visual inspection provides an excellent opportunity to check
for potential radon entry points into the building shell. The cracks and
utility penetrations noted above are likely candidates. One technique for
detecting radon gas almost instantaneously is called the radon "sniff." It
consists of pulling the gas sample through a scintillation cell while counting
the alpha disintegrations detected each minute. This type of investigation is
strictly a diagnostic tool and has no set EPA protocol. Such a device and
procedure tests the candidate entry points for higher radon concentrations than
the ambient room air.
SUB-SLAB COMMUNICATION AND PERMEABILITY
All of the information described above in the problem assessment process
is useful regardless of the type of mitigation plan to be employed. Before
attempting to design an SSD system, one other diagnostic test needs to be run.
The diagnostic sub-slab communications and permeability measurement involves
drilling at least one 30-40 mm (1.25-1.5 in.) hole just penetrating the slab in
the corner of some closet or other space designated by the homeowner near the
center of the slab and drilling several 10-13 mm (0.375-0.5 in.) pressure and
velocity sample holes (6-10) at various distances (1-5 m or 3-15 ft) in several
directions from the suction hole. A variable speed/suction vacuum cleaner is
used to depressurize the volume beneath the slab at the suction hole.
Instruments capable of measuring pressures in the 500-5000 Pa (2-20 in. we)
range and low flows (0.5-20 L/s or 1-40 cfm) are needed to make the sub-slab
permeability measurements, and measurements down to 0.2 Pa (0.001 in. we) are
needed for the pressure field extension (communication) test. Figure 1 is a
floor plan of a house in which one suction hole was drilled in a back bedroom
closet and nine test holes were drilled in available corners of rooms and
closets. The resulting approximate pressure contours have been drawn.
-------
SUCTION HOLE DETERMINATION
DETERMINING THE NUMBER OF SUCTION HOLES
Once, the decision has been made to install an SSD system for radon
mitigation purposes, the most critical questions to answer are how many suction
holes will be needed to remedy the problem and where to put them. If the house
has more than one slab then, for determining the number of suction holes, each
slab is treated separately. The single most useful diagnostic tool to use as
input in this determination is the sub-slab pressure field extension
measurement. From it, the mitigator should have a reasonable feel for what
types of communications are present under the slab. The recommended procedure
calls for a small test hole to be placed about 300 mm (12 in.) from the vacuum
cleaner suction hole. With the vacuum cleaner set to produce about a 375-500
Pa (1.5-2 in. we) pressure differential at that test hole, which is about what
a standard mitigation fan would produce, the pressure field measurements should
be taken at the various pressure and velocity sample holes. At most of the
nearby holes some differential pressure may be measured, but at some of the
more remote ones, more than likely no consistent reading will be possible. It
is important to remember that in low-permeability soils sufficient time must be
allowed for the pressure field to be established (3-5 minutes for nearby holes
and successively longer times for the more remote ones). The greatest distance
from the suction hole at which a definite pressure is recorded before the first
zero pressure is read should be taken as the effective radius of extension, r,
of the pressure field from a suction hole in that location. The analysis to
follow assumes that the sub-slab medium is fairly uniform. If there seems to
be reason to doubt the validity of this assumption, an additional suction
hole(s) may be recommended.
Once the effective radius of extension from the suction hole is determined,
the next input required is the approximate area of the slab being considered.
Figure 2 is a graph in which the effective radius of extension is plotted on
the x-axis (from right to left) and the area of the slab is plotted on the y-
axis. The diagonal lines divide the regions of the effective coverage area of
the indicated number of suction holes. Find the effective radius, of extension
r, that was determined, go straight up parallel with the y-axis until you find
the area of the slab. The region between the diagonals where the radius and
area intersect indicates the approximate minimum number of suction holes
required by that slab. This number may need to be increased if features such
as interior footings, sunken slab areas, sub-slab obstructions, or geometrical
shapes of the slab, seem to limit sub-slab communication. Erratic or
discontinuous results of the communication test will indicate the possibility
of such a condition. If the diagnostic test was made when the sub-slab soil
was unusually dry, then the soil permeability and the pressure field extension
determined will most probably be greater than those that would have been
measured during a wetter season. Generally in low-permeability soils, there is
little likelihood in producing too great a flow for the depressurizing fan, so
when in doubt, an extra hole is a better option than not having enough.
-------
DETERMINING THE SUCTION HOLE PLACEMENT
If the mitigation system is being installed in an unfinished space such as
a basement wbere there are no restrictions on the placement of the suction
holes, then the mitigator has a great amount of flexibility in making his
decision. A floor plan drawn to scale, perhaps one on which the sub-slab
communication is plotted, is very useful at this point. Sketching in the
effective areas of pressure field extension from various suction hole
placements will give an idea of the optimum configuration to ensure the best
coverage of the slab. Geometry suggests that holes located about 1 effective
radius, r, from the nearest exterior wall(s) will give the widest coverage.
However, in practice, sometimes the soil near the edge of a slab has not been
compacted as well as that near the center, producing either a possible settling
space between the top of the soil and the bottom of the slab or else just a
more permeable trench near the perimeter of the slab. If the diagnostic
communication test indicated that a greater pressure field extension resulted
from a near-perimeter suction hole than from a central suction hole without a
much larger flow, then the placement of the suction holes nearer the perimeter
is recommended. If, however, the communication test showed much greater flows
from perimeter holes without much greater pressure field extension, then slab
cracks or other leakage is probably limiting the pressure field extension, and
perimeter suction holes should be avoided.
When the slab being mitigated is predominantly under finished space, such
as a finished basement or a slab-on-grade house, practical locations are
usually far more restricted. In such a circumstance a floor or house plan is
very helpful. The finished basement scenario is probably the more difficult
system to design. Usually the best locations from the homeowners' viewpoint
are corners of closets because there the installations will be less noticeable
and obtrusive. However, quite often closets will not be spaced to give full or
adequate pressure field coverage. If that is the case, one may consider
placing the suction hole in the corner of a room and then perhaps "boxing off"
that corner if the homeowner does not want the pipe to show. Boxing off can be
used for more central locations as well. The added difficulty with finished
basement installations involves finding a place or places for the pipes to
penetrate the basement ceiling which will line up with an acceptable first
story floor penetration.
Slab-on-grade houses usually also have most, if not all, of the area to be
mitigated as finished space. Closets may be spaced more advantageously than
are often found in finished basements. Moreover, there may be a pantry or
other location where a suction hole may be concealed. There may still be large
areas that cannot be affected by near-closet suction holes. These are most
typically open living room/dining room/kitchen/den areas. Quite likely there
would be more resistance from the homeowner to placing any interior piping,
even concealed, in such spaces. One possibility to pursue in such a occurrence
would be an exterior suction hole penetrating horizontally through a stem wall
beneath the slab rather than vertically through the slab in an interior space.
What is required for such an exterior penetration to succeed is that the stem
wall must be accessible from outside the house (no porches, patios, or concrete
or paving directly adjacent to the outside wall where the penetration is
-------
proposed) and that the penetration can be Installed without losing the pressure
field to slab cracks and stem wall leakage as was mentioned earlier with near-
perimeter placement in slab-on-grade houses. If the footing is on expansive
soils or there seem to be foundation or structural weaknesses near the stem
wall in question, a suction hole should not be placed there.
One other possible suction point location in some slab-on-grade houses is
through an attached garage area. Some garages actually have a portion of the
house slab exposed at one end of the space. Even if not, other garages are a
few steps down from the house floor level. In such an instance, the house stem
wall may form the lower course or two of the interior wall of the garage. Then
a horizontal penetration through the stem wall beneath the slab could be a good
suction point. Even if the garage is just a small step down from the house
slab, it may be possible to penetrate the garage slab and extend the system
depressurization under the house. A potential problem with using a garage
penetration is that often the garage slab has settled and/or cracked, leaving
possible by-passes where garage air may be drawn into the system, reducing the
effective suction head and limiting the effectiveness of the system.
FAN SELECTION
COMPARING SUB-SLAB FLOW CURVES WITH VARIOUS CURVES
While the pressure field extension measurements of the sub-slab communica-
tion diagnostic give a good approximation of an effective dspressurization
radius, the pressure and flow measurements are indicators of the sub-slab
permeability. Specifically, recommended procedures call for the simultaneous
measurement of the suction at the scaling baseline hole and flow from the 30-40
mm (1.25-1.5 in.) suction hole at suctions of at least 0.5, 2.0, and 5.0 kPa
(2, 8, and 20 in. we) produced by a vacuum cleaner. When these measured values
are plotted on an x-y axis, such as in Figure 3 for the highest permeability
(10"5 cm2) and one of the more typical (10'7 cm2) encountered in the Polk
County, Florida, study houses, one obtains a flow curve for the sub-slab fill
material.
Also plotted in figure 3 are fan performance curves taken from the EPA
Training Course Manual (10) and from other published fan company figures. The
RDS and R-150/K-6 are inline centrifugal fans designed with radon mitigation in
mind. The radial and vortex blowers are higher suction instruments that may be
adapted for use in mitigation systems. On such a simultaneous plotting, the
intersections of the soil curves with the fan curves indicate about where the
system will operate. Figure 3 suggests that, for both soils, but especially
the one with the lower permeability, the system will tend to operate near the
high suction/low flow end of the fan curve for the RDS, R-150/K6, or the radial
blower. The fan curve data for the vortex blower did not extend further than
the 1.5 kPa (6 in. we) suction in the plot, but it obviously intersects both
soil curves at higher suctions and higher flows.
FAN CHOICE CONSIDERING OTHER FACTORS
Because the mitigation field experience in low-permeability soils is still
in an early phase, it is not clear what the durability of a fan will be when it
-------
is operated at low flows and relatively high suctions. Some indications
suggest that fan failure nay occur sooner in such an operating environment. The
fans are often placed in attics which will be quite hot during the cooling
season. High heat with low flows through the fans may lower the durability of
the fans. Research is currently underway to determine if the system
deteriorates with time or if it is maintained fairly constant until some type
of failure occurs abruptly. Harrje et al. (11) at Princeton University have
developed a diagnostic which investigates the durability of the fan as well as
the mitigation system as a whole.
The inline centrifugal fans, since they have been designed for radon
mitigation situations, have been kept fairly lightweight and affordable.
The blowers that produce the higher suctions are generally built for industrial
applications and therefore are somewhat heavier and more costly. The power
requirements to operate these various fans may differ quite widely. The
lightweight inline fans are designed to perform in the 100-200 W power range.
The heavier, higher suction blowers will sometimes have several times this
power requirement. Therefore, the operating costs may vary greatly with the
choice of mitigating fan. Since research data have not been collected for a
long enough time in this area, it is not clear how to predict the long-term
costs of these various systems. If the inline fans have too short a lifetime,
replacement costs may make this system more expensive. If their durability is
long enough, then their lower initial cost and operating costs may make them
the more cost-effective system.
The inline centrifugal fans are designed to run very quietly (less than 6
sones) and, according to most reports, receive very little, if any, criticism
from homeowners in this regard. However, the larger, more powerful blowers,
especially if designed for industrial applications, characteristically produce
quite a bit more noise, often a steady, high-pitched whine. This noise factor
usually is dealt with by installing the fan as far from the living space as
possible and Including soundproofing when the system is first installed. Both
of these options may increase the initial installation costs, and a remote fan
placement will require longer piping runs which may reduce the system
effectiveness. Even with the additional precautions to limit the noise output,
some people sensitive to noise may still object to the larger fans.
Decisions on interior versus exterior suction holes and piping definitely
have a bearing on fan selection. If the exhaust pipe from suction holes in a
basement is routed through a rim joist to the outdoors, or if a suction hole in
a slab-on-grade house is through an exterior stem wall, then the fan will
probably be placed somewhere outside the house. Such a fan must be rated for
exterior applications. In some model lines these fans are more expensive than
Interior fans. The ease of handling and weight of the units with the supports
required are other factors to consider in the fan selection process.
PIPE SELECTION
Generally most mitigators use PVC pipe when installing SSD systems. It is
lightweight, easy to cut and handle, convenient for fittings and accessories,
strong in its glueing characteristics, noncorrosive, and smooth so as to offer
-------
low resistance to air movement. For permeable sub-slab environments conducive
to high volumes of airflow, 100 mm (4-in.) or larger FVC pipes are generally
used. For the low flows resulting from the low permeability soils addressed in
this document, 100 mm (4-in.) or smaller PVC pipes are usually adequate. The
smaller piping has the added advantages of being lighter and easier to handle,
less obtrusive to the homeowner, easier to conceal if desired, and usually less
expensive for the pipe, fittings, and accessories. Therefore, an important
determination is what size of pipe is the best to use for the given mitigation
project.
AIRFLOW VERSUS APPLIED SUCTION
The choice of pipe size is most directly governed by the volume rate of
flow (or velocity) expected to move through the pipe. The larger the volume
moving, the greater the resistance if the cross sectional area is kept
constant. However, increasing the pipe size will lower the frictional loss for
a given volume rate of movement. Therefore, the best inputs for estimating the
optimum pipe size for a mitigation system again come from the sub-slab
communication diagnostic pressure/flow measurements. If the fan has been
selected that is expected to be used in the mitigation system, then the point
of intersection of the fan curve with the sub-slab flow curve will give a good
approximation of the airflow that can be expected in the system.
From the airflow estimate, one may use a chart such as Figure 4 to
estimate the friction loss in various sizes of pipes or ducts. This chart,
like most of the available documentation on airflow through pipes or duct work
(12), is calculated for "average" pipe, which is usually some type of iron pipe
with a given smoothness and joints estimated to be present at some regular
frequency. PVC pipe is quite a bit less resistive to air movement because of
its greater smoothness. Therefore, these approximations usually overestimate
the friction loss that would actually be found in FVC pipes. If the fan
selected is one in which the sub-slab flow curve intersection with the fan
curve is in the 375-500 Pa (1.5-2 in. we) range, then one would probably want
to keep the friction loss to 0.8-1.5 Pa/m of pipe. If the fan curve intersects
the sub-slab curve at something greater than 1 kPa (4 in. we), then a friction
loss of 3-5 Pa/m of pipe could be tolerated.
To use a chart such as Figure 4, find on the y (vertical) axis the airflow
determined from the sub-slab/fan curve intersection. Go across horizontally
until you are in the friction loss range (x-axis) you determined as above. The
closest pipe size diagonal (those rising from left to right) would be approxi-
mately the best pipe size to achieve your goal. It is advantageous from the
perspective of friction loss to go with the larger pipe, but if other factors
such as expense, ease of handling, or homeowner preference indicate otherwise,
the smaller pipe would probably still be a safe choice, especially in light of
the lower friction of PVC pipe discussed previously.
The friction loss in straight pipes is only part of the loss of suction
head that is experienced in a system. Usually the next most significant
features contributing to friction loss are the bends or tees in the system. A
-------
90" elbow or tee in a pipe usually contributes the greatest pressure drop
potential of any of these features. A 45* elbow has slightly over half the
friction loss of a 90" elbow, and a 30° elbow has less than half that of a 90°
one. Table 1 lists the approximate length of pipe that produces the same
friction loss as each of several of the more commonly used connectors.
TABLE 1. APPROXIMATE FRICTION EQUIVALENCIES FOR VARIOUS FITTINGS
Equivalent Run of Pipe (m)
Type of Fitting
Tee
90* Elbow
45' Elbow
30s Elbow
Pipe Diameter (mm)
37.5
0.5
0.3
0.25
0.15
50
0.6
0.5
0.3
0.25
75
1
0.6
0.5
0.3
100
1.5
1
0.6
0.5
To determine the friction loss in Pascals (Fa) for a system, determine the
total length of pipe and the number and kinds of fittings for each pipe size.
Multiply the number of fittings for a pipe size by the equivalency from Table 1
for that fitting and pipe. Add the total equivalent meters so determined to
the actual length of pipe to be used to get the adjusted total length of pipe.
Then use the friction loss factor determined from Figure 4 to multiply by that
adjusted total. Repeat the calculation for each pipe size and add the total
together for the whole system.
APPLICABILITY AND AVAILABILITY
If the above calculation indicates a larger pipe size than is feasible or
desired by the homeowner, then perhaps a fan that can draw a larger suction at
lower flows is called for. If, however, a certain pipe and fitting size is
determined that is acceptable, then local supply stores should be investigated
to ensure that enough pipe, fittings, and accessories are easily available.
PVC pipe comes in a variety of thicknesses (sometimes called schedules). The
thicker walls are for high-pressure applications; consequently that PVC is
heavier and more expensive. The applications described here require no extra
thickening, so the thinnest-walled PVC pipe is usually adequate and preferred
for its weight, ease of cutting, and cost. However, some of the fittings and
couplings for one schedule will not fit properly or tightly on the same size
pipe of a different schedule. So a crucial part of the pipe selection process
is that there be an adequate supply of fittings and accessories for the size
and schedule of the PVC selected. Other couplings, reducers, bushings, etc.,
should be investigated at this phase of the process to ensure complete
compatibility and availability for the system. These are used chiefly at the
various interfaces--pipe to slab, pipe to fan, and fan to exhaust.
-------
REFERENCES
1. Henschel, D. B. Radon Reduction Techniques for Detached Houses: Technical
Guidance (second edition). EPA-625/5-87-019, U. S. Environmental
Protection Agency, Cincinnati, Ohio, 1988. 245 pp.
2. Osborae, M. C., T. Brennan, and L. D. Michaels. Radon Mitigation in 10
Clinton, New Jersey, Houses: A Case History. EPA-600/D-87-164 (NTIS PB87-
191847), Presented at Second APCA International Conference on Indoor
Radon, Cherry Hill, NJ, April 6-10, 1987. 12 pp.
3. Michaels, L. D., T. Brennan, A. S. Viner, A. Mattes, and W. Turner.
Development and Demonstration of Indoor Radon Reduction Measures for 10
Homes in Clinton, New Jersey. EPA-600/8-87-027 (NTIS PB87-215356), July
1987. 166 pp.
4. Findlay, V. 0., A. Robertson, and A. G. Scott. Testing of Indoor Radon
Reduction Techniques in Central Ohio Houses: Phase 1 (Winter 1987-1988).
EPA-600/8-89-071 (NTIS PB89-219984), July 1989. 301 pp.
S. Ericson, S. 0., H. Schmied, and B. Clavensjo. Modified Technology in
New Constructions and Cost Effective Remedial Action in Existing
Structures, to Prevent Infiltration of Soil Gas Carrying Radon.
Radiation Protection Dosimetry 7:224-225, 1984.
6. Scott, A. G., A. Robertson, and V. 0. Findlay. Installation and Testing
of Indoor Radon Reduction Techniques in 40 Eastern Pennsylvania Houses.
EPA-600/8-88-002 (NTIS PB88-156617), January 1988. 388 pp.
7. Ronca-Battista, M., P. Magno, and P. Nyberg. Interim Protocols for
Screening and Follow-up Radon and Radon Decay Product Measurements.
EPA-520/1-86-014, U.S. Environmental Protection Agency, Washington, D.C.,
1987. 22 pp.
8. U.S. Environmental Protection Agency. Indoor Radon and Radon Decay
Product Measurement Protocols. EPA-520-1/89-009, Washington, D.C., 1989.
102 pp.
9. U.S. Environmental Protection Agency. A Citizen's Guide Co Radon.
OPA-86-004, Washington, D.C., 1986. 13pp.
10. U.S. Environmental Protection Agency. Reducing Radon in Structures (2nd
Edition). Washington, D.C., 1989. 356pp.
11. Harrje, D. T., K. J. Gadsby, and D. C. Sanchez. Long Tern Durability and
Performance of Radon Mitigation Subslab Depressurization Systems.
Presented at 1990 International Symposium on Radon and Radon Reduction
Technology, Atlanta, GA. February 1990. 14 pp.
12. American Society of Heating, Refrigerating and Air-Conditioning Engineers,
Inc. ASHRAE Handbook 1981 Fundamentals. Atlanta, GA. Chapters 33 and 34.
-------
1 m
Figure 1. Approximate pressure contours from a suction hole in a representative
house plan (5 kPa suction at suction hole).
-------
100
CM
CO
-------
High-permeability soil
Low-permeability soil
RDSian
Radial blower
Vortex blower
R-150/K-6fan
20
50
0
1000
1500
Suction (Pa)
Figure 3. Sub-slab curves for two permeabilities plotted with fan curves for four
different kinds of fans.
-------
100
.£•
•i-*
6
10
0.1
Friction Loss (Pa/m)
10
100
Rgure 4. Friction chart for average pipes.
-------
VII-4
RADON MITIGATION TECHNIQUES FOR BASEMENT HOUSES
WITH POOR SUB-SLAB COMMUNICATION
by: Bobby E. Pyle
Southern Research Institute
Birmingham, AL 35255
Michael C. Osborne
AEERL, U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
ABSTRACT
While the sub-slab depressurization (SSD) technique is widely used to
mitigate radon levels in basement houses, there are instances for which
this method is not a viable option. For example, in some houses the slab
is poured directly on the soil resulting in poor-to-nonexistent
communication under the slab. To apply SSD methods requires elaborate
plumbing and multiple suction holes in the slab. In an effort to develop
viable alternatives, EPA has funded research to explore other radon
mitigation options. Case studies will be presented that include: basement
pressurization with and without heat extraction, and filtration using
charcoal. In the first, air from the upper portion of the house was forced
into the basement producing a pressure barrier at the slab/soil interface.
In one case heat was extracted from the upstairs air using a heat pump to
supply hot water for the occupants. The air filtration method used a bed
of charcoal to remove the radon gas. The charcoal bed was flushed with
outdoor air to extract the radon before it decayed completely to radon
daughters.
This paper has been reviewed in accordance with the United States
Environmental Protection Agency's peer and administrative review policies
and approved for presentation and publication.
-------
INTRODUCTION
Phase I of this EPA Radon Mitigation Demonstration, in Nashville,
Tennessee, identified many basement houses which had poor sub-slab
communication. In attempting to implement a sub-slab depressurization
(SSD) system in houses such as these, special attention had to be taken to
ensure an adequate means to prevent radon entry into the basement. In an
effort to determine if less expensive techniques were available that might
have a higher probability of success, the mitigation methods described
below were investigated. These techniques may be applicable in other areas
such as Central Florida where the lack of communication under the slab
increases the complexity of SSD systems.
DESCRIPTION OF HOUSES
Nine houses were selected for the Phase II demonstration in Nashville,
TN. Of these, six are basement houses and three are crawl space houses.
Of the six basement houses only four have been mitigated as of the date of
this paper. A brief description of these four houses follows. A final'EPA
report describing all of the mitigation efforts in Nashville should be
available by the summer of 1990.
Dtt 01 is a one story frame house with brick veneer upstairs. The
basement is a daylight walkout at the back of the house with walls of
concrete block. The basement area is approximately 70% finished with a
bedroom and den in the front (below-grade) part of the basement and an
unfinished laundry room and workshop at the rear. The heating system for
the house includes a central heat pump which is seldom used and two wood
stoves, one upstairs and one in the finished basement. There are no other
combustion appliances in the house. Charcoal measurements made in the
winter of 1987-88 in the basement by the homeowner showed the radon level
to be 21.4 pCi/L. Results of blower door infiltration measurements at 4 Pa
showed 2 ACH (air changes per hour) for the basement and 1.4 ACH for the
whole house.
DW 35 is a two story frame house with brick veneer on the exterior.
The basement is almost totally below grade with walls of cut stone. The
basement is semi-finished but unused because of water problems. A central
gas fired furnace in the basement has numerous leaky joints around
attempted seals with asbestos tape. Also located in the basement are a gas
fired clothes dryer and water heater. Visual inspection showed numerous
thermal bypasses from the basement to upstairs. A seldom-to-never used
fireplace is located in one room of the basement. The flue has been closed
loosely with sheet metal but still shows air exfiltration. A previous
radon measurement in the basement was 18 pCi/L. Infiltration measurements
at 4 Pa showed that the basement was very leaky: 6.3 ACH for the basement,
and 4.4 ACH for the whole house.
-------
DW 38 is a one story brick veneer house with a walk-out daylight
basement. The walls of the basement are concrete block with three of the
walls about half below-grade by about 3-4 ft (0.92 - 1.22 m). The basement
is 50% finished with a den in one part and a laundry room and work shop in
the other. The heating and air-conditioning (HAG) system is a central unit
in the basement (heat pump) with very leaky ductwork (both supply and
return). There were no combustion appliances in the basement. Previous
measurements in this house were 11.1 pCi/L in the basement. Infiltration
measurements at 4 Pa resulted in values of 2.8 ACH for the basement and 1.6
ACH for the whole house.
DW 47 is a single story frame/brick veneer house with a walk-out
basement. The homeowner was in the process of finishing the basement when
mitigation was undertaken. The walls of the basement are concrete block
with three sides below grade. The house is totally electric with the air
handler located upstairs. The basement contained a roughed-out opening for
a fireplace with a large amount of bypass area to the upstairs flue box.
Also located in the basement is a heat pump water heater using basement
heat. Previous radon measurements by the homeowner were 13.1 and 16.9
pCi/L upstairs and 21.1 pCi/L in the basement. Infiltration measurements
at 4 Pa gave values of 2.4 ACH for the basement and 1.3 ACH for the whole
house.
MITIGATION SYSTEMS
In three of the houses (DW 01, DW 38, and DW 47), radon levels were
mitigated by pressurizing the basement with air withdrawn from upstairs.
This was done in two of the houses (DW 01 and DW 38) by using an in-line
blower with maximum flow capability of approximately 400 cfm at 0 in. H20
(11.35 m3/min at 0 Pa) (T3A Turbo 8, R.B. Kanalflakt, Inc., Sarasota, FL).
In DW 01 the air inlet was located in the central hallway of the house and
in DW 38 it was located in the floor of the kitchen.
In the third house (DW 47) a heat pump water heater (HPV-80, Therma-
Vent™, Therma-Stor Products Group, DEC International, Inc., Madison, WI)
was used to withdraw air from the upstairs area, down through the heat
pump, and then exhaust the air into the basement. The heat extracted from
the air was used to provide hot water for the homeowner. In DW 47 the air
inlet was located in the central hallway.
Several parameters of each of these three houses were monitored and
recorded using a datalogging device. The parameters were: air temperatures
upstairs, in the basement, and outdoors; air pressure differentials between
the basement and upstairs, between the basement and outdoors, and between
the basement and the sub-slab region; wind speed and direction; rainfall
amounts; HAG operation; and continuous radon levels both in the basement
and on the first floor. The monitoring period was quite long, ranging from
several weeks to months in duration. Due to space limitations, the bulk of
this monitoring data cannot be listed and explained in this paper.
A major point considered for this type of mitigation system was the
amount of sealing required in the basement to achieve sufficient
-------
overpressurization while keeping the airflow rate down to less than 300-
400 cfm (8.51 - 11.35 m3/min). Also considered were the effects of such a
system on the air infiltration of the house. Air infiltration was measured
to determine how much of an energy penalty the system imposes on the house.
Results of these studies will be included in the EPA final report.
In the remaining basement house (DW 35) an off-the-shelf activated
charcoal air cleaning unit (Radsorb-222, RAd Systems, Inc., Westborough,
MA) was used to remove the radon gas from the basement air. In this device
the radon gas adsorbed by the charcoal is removed by reverse flushing to
the outdoors before the radon completely decays to its daughter products
and remains trapped in the charcoal bed. The radon levels in both the
basement and the first floor were continuously monitored over a period of
several months.
HOUSE MITIGATION RESULTS
These houses were mitigated during 1989. Because of the length of
monitoring time, the mitigations extended from January through November of
1989. Prior to mitigation, long term alpha track detectors (ATDs) were
installed in the houses. ATDs were obtained from Tech-Ops Landauer, Inc.
(Glenwood, IL). The results of these pre-mitigation measurements are shown
in Table 1. The radon levels in the basements ranged from 9.4 to 24.4
pCi/L and the levels on the first floor ranged from 4.9 to 10.4 pCi/L.
Short term (3 day), closed house, charcoal canister (CC) measurements were
carried out just prior to mitigation. Charcoal canisters were provided and
analyzed by Scientific Analysis, Inc. (Montgomery, AL). These results are
shown in Table 2. The agreements between the ATD results and the CC values
are not as good as might be expected. The differences are attributed to
differences in the time of year and to weather. The average levels in the
basement and on the first floor for both the CC and the ATD measurements
are summarized in Figure 1.
House DW 47
The first house to be mitigated was DW 47 using the HPV-80 device.
The radon levels in the basement and on the first floor along with the
house parameters were continuously monitored beginning on 1/31/89. The
continuous radon monitor (CRM) values during the pre-mitigation period
averaged 13.1 pCi/L in the living room and 18.1 pCl/L in the basement. The
HPV-80 system was installed on 3/15/89, and the openings between the
basement and upstairs were sealed the following day. The major openings
that were closed were the bypass to the chimney flue, the water and sewer
pipe penetrations, and an opening in the sub-floor below the HAC. Since
the basement was largely unfinished, sealing these openings was simple and
effective. The HPV-80 fan was turned on at 6:40 pm on 3/15/89. The radon
levels began to decline within 6 hours, and within 12 hours the levels were
less than 2 pCi/L in both the basement and on the first floor. The CRMs
remained in the house until 4/13/89. Over this post-mitigation period the
average radon levels were 1.8 pCi/L in the basement and 1.7 pCi/L on the
first floor.
Prior to turning on the HPV system, the basement pressure was
averaging about 2.5 Pa below ambient outdoor pressure. With the system on,
-------
the basement pressure was approximately 2.5 Fa above ambient outdoor
pressure. The HFV system flowrate was 141 cfm (4.00 m3/nin) measured at
the exhaust outlet in the basement. On 4/4/89 the HPV system was shut down
for approximately 48 hours and then was turned back on. The pressure
changes over this period are shown in Figure 2, and the changes in the
radon levels are shown in Figure 3. During the time the HPV system was on,
the sub-slab pressure was averaging about 2 Pa below the basement pressure,
and the basement was about 2.5 Pa above the outdoor pressure. With the HFV
system off, the sub-slab pressure increased to 1 Fa or less (below the
basement pressure) and the basement pressure decreased to about -3 Pa
relative to the outdoors. The radon levels in both the basement and on the
first floor increased by factors of 5 or more. When the system was turned
back on, the levels decreased sharply as seen in Figure 3. Overall the
reduction in radon levels was approximately 92% in the basement and 83% on
the first floor.
House DW 38
The second house to be mitigated using basement pressurization was DV
38. In this house an in-line duct fan was used to pull air from the
kitchen upstairs. The air was then distributed in the basement in two
zones, the finished den and the unfinished workshop. The mitigation system
was installed and energized on 4/25/89. Prior to mitigation, the radon
levels and house parameters were monitored beginning on 4/13/89. The
average pre-mitigation radon levels measured with the CRMs were 10.6 pCi/L
on the first floor and 14.6 pCi/L in the basement. Also over this period,
the average basement pressure was approximately 1 to 2 Fa below the outdoor
ambient pressure, and the sub-slab area pressure wa& approximately 2 Pa
more positive than the basement. During the initial mitigation a small fan
[158 cfm at 0 in. H20 (4.48 m3/min at 0 Pa), model R125, Fantech, Inc.,
Sarasota, FL] was used. It was soon apparent that the fan could not supply
sufficient airflow to overcome the leaks to the upstairs and to the
outdoors. A larger fan was installed [410 cfm at 0 in. H20 (11.63 ms/nin
at 0 Pa), model T3A Turbo 8, R.B. Kanalflakt. Inc., Sarasota, FL] to
pressurize the basement.
Sealing this house was difficult because of limited access to the
overhead from the basement. The finished basement ceiling had acoustic
tile attached directly to the floor joists. Also, the homeowner had made
several changes in the electrical and plumbing systems over the years and
consequently left many openings. The MAC system located in the unfinished
part of the basement not only had many leaks in the cold air return but
also had a return vent in the finished area of the basement that had to be
sealed off. To prevent the pressurization air from leaking back upstairs
through the return duct, a back-draft damper was constructed at the return
opening upstairs. Also, to prevent back-flow through the supply vents in
the basement, back-draft flaps of vinyl were attached to the outside of the
supply grills. Also, a direct path to the attic in the basement stairwell
required sealing. In summary, this house was extremely difficult to seal.
The basement and the upstairs were never completely isolated.
Following installation of the larger pressurization fan and sealing of
as many of the openings as possible, the pressure difference between the
basement and the sub-slab region was on the average reduced to zero. The
-------
basement was never pressurized above the outdoor ambient conditions.
However, by achieving a neutral pressure difference relative to the sub-
slab, radon gas flow into the basement was significantly reduced. No
condensation problems were observed with the kitchen air pulled down into
the basement. The radon levels in the basement, as measured with the CRM,
averaged 1.3 pCi/L. These levels were confirmed with CC measurements
carried out on 8/22-24/89. The CC results, shown in Table 2, were 1.2
pCi/L in the basement and 0.9 pCi/L on the first floor.
House DW 01
Radon in this house was mitigated over the period from 10/2/89 to
10/6/89. Prior to mitigation the datalogging system was installed on
9/13/89 and remained until 11/6/89. The pre-mitigation radon levels in
this house were 19.5 pCi/L in the basement and 9.8 pCi/L on the first floor
as measured with CCs and shown in Table 2. The average radon levels
recorded with the CRMs (model R210F radon monitor, Femto-Tech, Inc.,
Carlisle, OH) were 22.5 pCi/L in the basement and 8.5 pCi/L on the first
floor. The basement pressurization mitigation system was installed and
turned on at approximately 4:00 p.m. on 10/3/89. Sealing of the basement
openings to the upstairs and to the outdoors was delayed until 10/5/89.
With only the pressurization fan installed and no additional sealing, the
radon levels dropped within 6 hours to 2.8 pCi/L upstairs and 4.2 pCi/L in
the basement. This is shown in Figure 4. The average pressure in the
basement relative to: the outdoors changed from -0.4 to -0.2 Pa, the
upstairs from +0.03 to +2.30 Pa, and relative to the sub-slab area from
-0.1 Pa to zero. With a limited amount of sealing the radon levels dropped
further to 1.6 pCi/L upstairs and 1.9 pCi/L in the basement. The average
pressures in the basement increased to +1.1 Pa relative to the outdoors,
+2.4 Pa above the upstairs, and +0.14 Pa above that under the slab.
On one occasion, the homeowner inadvertently left the door between the
upstairs and basement open for most of a day. During this time the
pressurization was lost in the basement and the radon levels in both the
basement and the first floor increased sharply. These results are shown in
Figures 5 and 6, respectively.
The radon levels and the house parameters were monitored until
11/6/89. The radon levels in the basement averaged 1.2 pCi/L and on the
first floor 1.1 pCi/L. These levels were confirmed with CC measurements
carried out over the period 11/6-8/89. These tests gave the results (shown
in Table 2) of 0.2 pCi/L in the basement and upstairs.
House DW 35
This house was mitigated with the charcoal adsorption unit (Radsorb-
222). The unit was installed on 5/30/89. Continuous radon monitors (model
R210F radon monitor, Femto-Tech, Inc., Carlisle, OH) were installed prior
to installation on 5/18/89. Pre-mitigation radon levels averaged 11.2
pCi/L in the basement and 5.4 pCi/L on the first floor. The charcoal
-------
adsorption unit was operated more-or-less continually for the next 6
months. Monitoring of the radon levels in the basement and upstairs
continued throughout most of this period. Monitoring was interrupted on
several occasions by failure of the CRM devices. The failures were due to
excessive moisture levels in the basement. Over the last period for
which data were available for this paper, the average levels measured by
CRMs (model AB5 with PRD1 passive cells, Pylon Electronics Development Co.,
Inc., Ottawa, Canada) were 5.5 pCi/L in the basement and 4.2 pCi/L
upstairs. The CRMs also indicate that during heavy rains the levels can
increase drastically. During one such rain storm on 10/17/89 the levels
increased to about 54 pCi/L in the basement and 25 pCi/L upstairs.
Following the rain storm, the charcoal device reduced the levels to 6.4
pCi/L in the basement and 3.5 pCi/L upstairs within 24 hours. The radon
levels in both the basement and the living area over this period are shown
in Figure 7.
Post-mitigation CC measurements were carried out over the period 11/6-
8/89 as shown in Table 2. These levels are higher than those measured with
the CRMs over a much longer period. However, the CC measurements were
begun shortly (about 6 hours) after a storm dropped approximately 1.2 in.
(3.05 cm) of rain in the area. CRMs will be used to monitor this house for
some additional period of time. Also, long term ATDs have been installed
both in the basement and upstairs.
CONCLUSIONS
The basement pressurization technique has been applied to three houses
of differing age and construction type. In one house heat was extracted
from the upstairs air before pressurizing the basement. The degree of
success in reducing the radon levels with this method is determined by how
tight the basement can be made to air leaks to the outdoors and to the
upstairs. Changes in house tightness are currently being quantified. For
houses in which the HAC system incorporates continuous ductwork in good
condition, the sealing is not too difficult. For houses in which the floor
joists are used as the sides of cold air return ducts, sealing can be very
difficult and time consuming. Each house has to be evaluated on its own.
Use of diagnostic tests such as blower door measurements can aid in the
evaluation prior to mitigation system design. The charcoal adsorption
technique is currently under test in one house. The device is capable of
reducing moderate levels (10 to 15 pCi/L) of radon to levels that approach
4 pCi/L. However, the device cannot control rapidly increasing levels as
produced during water capping of the surrounding soils. The long-term
effectiveness of the device is currently under study.
-------
TABLE 1. ALPHA TRACK DETECTOR MEASUREMENTS FOR
NASHVILLE PHASE II BASEMENT HOUSES
Location Date Date Days Radon
Code Installed Removed Exposed Level
(pCi/L)
Comments
DW 01-B* 09/22/88
DW 01-B 09/22/88
DW 01-L* 09/22/88
DW 01-L 09/22/88
DW 35-B 09/20/88
DW 35-B 09/20/88
DW 35-L 09/20/88
DW 35-L 09/20/88
DW 38-B 09/21/88
DW 38-B 09/21/88
DW 38-L 09/21/88
DW 38-L 09/21/88
DW 47-B 09/21/88
DW 47-B 09/21/88
DW 47-L 09/21/88
DW 47-L 09/21/88
DW 47-B 04/13/89
DW 47-B 04/13/89
DW 47-L 04/13/89
DW 47-L 04/13/89
09/14/89
09/14/89
09/14/89
09/14/89
05/30/89
05/30/89
05/30/89
05/30/89
05/04/89
05/04/89
05/04/89
05/04/89
03/16/89
03/16/89
03/16/89
03/16/89
10/03/89
10/03/89
10/03/89
10/03/89
357
357
357
357
252
252
252
252
225
225
225
225
176
176
176
176
173
173
173
173
12.6
16.2
5.4
5.0
9.4
9.6
5.4
4.9
11.1
11.2
9.9
9.3
12.5
24.4
8.1
10.4
1.6
1.5
1.7
1.5
Pre-Mitigation
Pre-Mitigation
Pre-Mitigatlon
Pre-Mitigation
Pre-Mitigation
Pre-Mitigation
Pre-Mitigation
Pre-Mitigation
Pre-Mitigation
Pre-Mitigation
Pre-Mitigation
Pre-Mitigation
Pre-Mitigation
Pre-Mitigation
Pre-Mitigation
Pre-Mitigation
Post-Mitigation
Post-Mitigation
Post-Mitigation
Post-Mitigation
* B - Basement, L - 1st Habitable Level Above Basement
TABLE 2. CHARCOAL CANISTER MEASUREMENTS FOR
NASHVILLE PHASE II BASEMENT HOUSES
Radon Level (pCi/L)
1st Habitable
Level
Test 1 Test 2
House
ID
DW 01
DW 01
Date
Started
9/13/89
11/6/89
Date
Stopped 1
9/15/89
11/8/89
Basem
test 1
19.0
0.3
ent
Test 2
20.0
0.1
9.6
0.1
9.9
0.2
Comments
Pre-Mitigation
Post-Mitigation
DW 35* 11/6/89 11/8/89 9.6 9.2
DW 38 4/4/89 4/6/89 17.5 17.8
DW 38 8/22/89 8/24/89 1.3 1.1
DW 47** 3/14/89 3/16/89 4.6 4.7
6.3 6.1 Post-Mitigation
13.6 13.6
1.0 0.8
1.8 1.9
Pre-Mitigation
Post-Mitigation
Pre-Mitigation
* Time constraints did not permit pre-mitigation charcoal measurements
** Post-mitigation charcoal measurements could not be made due to homeowner's
constraints
-------
O
O
25-
20-
10-
5-
n.
CUATD-BSMT raATD-1 FLR ^CC-BSMT ZZ2CC-1 FLR
I.
t
'
i
i
i
i
i
;
1
>
X
1
DW 01 DW 35 DW 38 DW 47
HOUSE ID
Figure 1. Pre-mitigation AID and CC results (no CCs at DW 35).
10
o
a_
o
•z.
LJ
a:
o
LJ
a:
0
-5
-10
— Basement Pressure Relative to Outdoors
—• Basement Pressure Relative to the Sub—slab
Fan On
Fan Off
Fan On
4/7/89
4/3/89 4/5/89
DATE (Midnight)
Figure 2. AP changes as the fan was turned off then back on in DW 47.
-------
0
Q.
20
15
10
O
<
0
Fan Turned Off
Fan Back On
•—'Basement
— First Floor
4/3/B9 4/5/B9
DATE (Midnight)
4/7/89
Rgure 3. Radon level changes as the fan was turned off then back
on in DW 47.
o—o First Floor
— Basement
10/3/89
10/7/89
10/5/89
DATE (Midnight)
Figure 4. Changes fn radon levels as the fan was turned on
and as the basement was sealed in DW 01.
10
-------
—o Basement Relative to Outdoors
Basement Relative to First Floor
— • Basement Relative to Sub-slab
Door Closed
Door Closed
10/24/89 10/25/89 10/26/89
DATE (Midnight)
Figure 5. Effects of pressure loss in basement through
door left open in DW 01.
o — o First Floor
— Basement
10/24/89 10/25/89 10/26/89
DATE (Midnight)
Figure 6. Effects of pressure loss on radon levels in DW 01
11
-------
60
50
£40
LJ
30
§ 20
10
10/16/89
•—'First Floor
— Basement
• Rainfall
10/18/89
10/17/89
DATE (Midnight)
Figure 7. Effects of heavy rainfall on radon levels in house
DW 35 mitigated with charcoal adsorption device.
12
-------
Paper VH- 5
ONE-YEAR FOLLOW-UP STUDY OF PERFORMANCE OF
RADON MITIGATION SYSTEMS INSTALLED IN
TENNESSEE VALLEY HOUSES
C S. Dudney, D. L Wilson, R. J. Saultz and T. G. Matthews
Health and Safety Research Division
Oak Ridge National Laboratory
Oak Ridge, TN 37831-6113
ABSTRACT
Subbarrier depressurization systems were installed for radon mitigation in two basement ranchers
in Oak Ridge, TN, and in two ranchers with partial basements in Huntsville, AL. System performance
parameters, including pressure field extension, subslab permeability, and indoor radon concentrations, were
followed in each house for a year or longer. In general, these performance factors were stable over the
year. In one house built on a slab without underlying aggregate, the subslab permeability has increased in
two of four suction pits. In another house, the extension of the field of depressurization is markedly
anisotropic. In three out of the four houses, mitigation measures that do not consume electrical power seem
to have provided 30 to 60% reductions in indoor radon concentrations.
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
Radon is widely perceived as one of the most significant health risks present in indoor air (1).
According to recent state surveys, radon levels in excess of the Environmental Protection Agency (EPA)
guideline of ISO Bq m'3 are widely distributed throughout the US. (2). As a result, research programs have
been supported on a regional basis to more quickly develop mitigation technologies required for variant
macro-geologic and house-construction characteristics. A radon research program in the Tennessee Valley
has been sponsored by multiple agencies to:
1. better understand the physical basis of factors underlying radon entry and its control,
2. refine diagnostic methods for implementation of effective radon control methods, and
3. improve radon control technologies while systematically reducing radon levels in study
houses.
In addition, a study of performance of radon mitigation systems for one year following installation was made
in two houses in Tennessee and two houses in Alabama.
Several research groups have evaluated mitigation system performance over extended periods of time.
The principal means of evaluation to date have included post-mitigation radon monitoring with alpha track
devices and follow-up visits to inspect the system for operational integrity. Nitschke (3) discovered
numerous design flaws in early subsurface depressurization systems installed in New York houses. Prill (4)
discovered in six houses in Washington and Idaho that performance of subsurface pressurization systems
declined due to flow obstruction at the outlet and due to occupant interference with vents and fans.
Because of the substantial seasonal variation of indoor radon concentrations that we have seen in
the Huntsville study houses (5), the pressure and flow characteristics of the systems were followed closely
to monitor system performance. In each of 4 houses with basements, 12 or more sampling holes were
installed in the concrete slab and repeated measurements of radon concentration and air pressure were taken
at these sites. In addition, flow and subslab permeability in the system were followed. This paper will
discuss what changes were seen in the pressure field and indoor radon concentrations during the first year
of operation.
DESCRIPTION OF STUDY HOUSES
The study houses were selected based on: (1) the above research goals, (2) indoor radon survey
data, (3) the availability of tightly clustered houses with similar construction characteristics, and (4) predicted
homeowner cooperation with a 1-2 year, highly instrumented, research program. The Oak Ridge/Knoxville
area of eastern Tennessee, and the Huntsville area of Madison County, AL, were selected because of their
comparatively high indoor radon levels in state surveys (6), different underlying geologies, and similar
housing stock. A data base of prospective houses was generated in both locales from radon surveys of
approximately 500 houses using charcoal monitors, and door-to-door screening using continuous radon
monitors.
For the larger study of radon entry and control, four basement ranchers in Oak Ridge, TN, and four
crawlspace houses in the Garth Mountain area of Huntsville, AL, were selected. All eight study houses had
substructural (i.e., crawlspace and/or basement) and superstructural levels of equal floor area to minimize
variability in radon entry, transport, and exhaust properties. The Huntsville houses (HU11-HU14) and Oak
-------
Ridge houses (OR1S-OR18) were clustered within a 0.4 and 1.2 km radius, respectively, to minimize the
variability in underlying geology and the time dependence of meieorologic events. The four houses included
in this study were:
House HU13 This is a single-story house, built around 1957 and rebuilt after a fire in 1978. The house
has a substructure composed of crawlspace and walkout basement areas below the living
area. There was no vapor barrier in the crawlspace and a continuous wall between
crawlspace and basement areas with two doors. There are two single garage doors opening
into the basement area. There is an attached parking area with slab-on-grade construction.
An electric furnace with forced-air distribution system is located in the basement. An
electric central air conditioning unit is located beside the rear wall. The occupants routinely
use a wood stove in the upstairs living area for heat during the winter months. The house
is built on flat land with few trees in the backyard. There are three adult occupants, one
of whom was home most of the time.
House HU14: This is a single-story house, built around 1962, with a split foyer entrance and a substructure
composed of crawlspace and walkout basement areas below the living area. There was no
vapor barrier in the crawlspace. There is a continuous wall, with a door, between the
crawlspace and basement areas. There a double garage door that opens into the basement
and is opened twice per day for brief periods. An electric heat pump with forced-air
distribution system is located in the crawlspace. An electric central air conditioning unit
is located beside the rear wall. The occupant routinely uses an upstairs wood-burning
fireplace in the wintertime. The house is built on sloping land with many trees in the
backyard. There is one adult occupant who was home evenings and weekends.
House OR15: This is a single-story house, built in 1963, with a full walkout basement below the living
area. Three-quarters of the basement area is finished with panelled walls and tiled floors.
There is an attached carport with slab-on-grade construction. Beneath the carport slab is
a fallout shelter that can be entered from the basement. An electric furnace with forced-
air distribution system is located in the basement. An electric central air conditioning unit
is located beside the side wall. The house is built on sloping land with few trees in the
backyard. There are two adult occupants, one of whom was home most of the time.
House OR16: This is a single-story house, built in 1963, with a full walkout basement below the living
area. Three-quarters of the basement area is finished with panelled walls and tiled floors.
There is an attached carport with slab-on-grade construction. An electric furnace with
forced-air distribution system is located in the basement. An electric central air conditioning
unit is located beside the rear wall. The house is built on sloping land with many trees in
the backyard. There are two adult occupants, both of whom were home during evenings
and weekends.
DESCRIPTION OF MITIGATION MEASURES
WEATHERIZATION, HAC SEALING, AND SLAB SEALING
The first stage of radon mitigation performed in the Huntsville houses was simple weatherization
measures and sealing of the heating and air conditioning (HAC) system. All houses were weatherized to
a similar degree, consistent with standard Tennessee Valley Authority (TVA) recommendations and practice.
In both Huntsville houses the return-side ductwork extended into all substructure! compartments.
Prior to house weatherization, inspection of the HAC systems revealed leaky ductwork. For the return duct,
-------
leak testing was performed by two methods, physical inspection (i.e., for large holes or gaps) and by utilizing
a smoke pencil. For small leaks, latex caulk, fingertip caulk, and duct tape were applied to seal the leaks.
In cases where joint leakage was due to nonsupport of the duct, sheet metal support bands were added.
In more severe cases (i.e. large holes or missing connections), heavy gauge sheet metal form connectors,
support cuffs,or cover plates were fabricated and permanently attached with sheet metal screws. These joints
were sealed with latex caulk as well. In both houses, the pan-type return ducts were renailed to the floor
joist and the edges sealed with latex caulk In addition, all other accessible metal to metal joints in the
return ducting were sealed with latex caulk. Confirmation of sealing was made by using a smoke pencil.
In both Huntsville houses, the central HAC unit was located in the crawlspace. Smoke pencil
testing of the units revealed that the units were very leaky and drew crawlspace air into the HAC system.
Since 100% sealing of the units was impossible, user-removable service panels were sealed with a
combination of modeling clay, fingertip caulk, and duct tape. Similar seals were fashioned for all inter-panel
joints and gaps between the chassis and user-removable panels. Other non-removable panels or non-
serviceable areas were permanently sealed with latex or silicon caulk. Effectiveness of the sealing effort was
judged by smoke stick testing. After all direct surface seals were completed, the entire cold-air return duct
system was permanently covered with a vapor-seal/fiberglass duct-jacketing.
In all of these houses, the slabs were in reasonably good condition with most deficiencies being
small "hair line" cracks. However, in HU14 there were some large (0.5 cm) cracks. For houses OR1S and
OR16 access to the slab was limited since most of the basement area was finished. All stress cracks in the
slabs were expanded to a V-shaped channel, 0.5 to 1.5 cm in width and 1.5 cm in depth using a
percussion-type bit and filled with flowable urethane sealant. If needed, an application of latex caulk or
liquid adhesive was used to fill large gaps to prevent bleed-through of the flowable sealant. After the first
application of the flowable sealant had aged for 3 days, an additional layer was applied to fill any
depressions or gaps which had occurred during the curing process. Cracks and perforations in the block
walls were sealed with latex caulk or expansion foam. Wall outlets were sealed with a combination of latex
and fingertip caulk. Slab-wall expansion joints were filled with backer rod followed by flowable urethane
caulk. In House OR1S, the fallout shelter was isolated from the basement with a removable, weather
stripped plywood door.
SUBSLAB VENTILATION SYSTEMS
All pits for subslab ventilation were square in design, approximately 60 x 60 cm in size, and with
a depth chosen according to permeability test results. Excavation depths of each pit depended upon the
subslab permeability and the depth permeability profile data. Standard 10 cm (4") or IS cm (6") schedule
40 PVC pipe was inserted within the pit to a depth of 30 cm beneath the bottom of the slab and filled with
gravel (1 to 2 cm). A 60 x 60 cm cardboard shield was then inserted around the pipe and concrete was
poured and finished. The pit pipe was then attached to the main 10 cm (4") PVC pipe trunk equipped with
a Kanalflakt pipe fan. Exhaust locations for HU13 and HU14 were through the side of the house and
OR15 and OR16 were through the roof. A detailed description of the system installed at each house is
given below.
House HU13 The substructure consists of a basement/crawlspace combination with a hollow block wall
separating the two compartments. The basement construction consists of hollow concrete
block walls with interior walls built over the slab. A thick IS cm layer of aggregate was
found under the slab, and excellent communication and permeability (140 to 190 x 10'7 cm2)
were observed between diagnostic holes and slab suction points within the basement.
Subslab to wall communication was not detected in the basement during subslab diagnostic
tests. Due to the excellent communication and permeability found under the slab, only one
50 cm deep mitigation pit was installed near the center of the basement slab.
-------
House HU14 The substructure consists of a basement/crawlspace combination with a hollow block wall
separating the two compartments. The basement construction consists of hollow concrete
block walls with interior walls built over the slab. Some (~8 cm) aggregate was found
under the slab. Subslab permeability was reasonable ranging from 88 to 128 x 10*7 cm2
between diagnostic holes and slab suction points in the basement. Subslab to wall
communication was not detected in the basement during the subslab diagnostics experiments.
Due to the relatively good communication among test holes in the slab, only one 65 cm
deep subslab mitigation pit was installed in the basement.
House OR1S The substructure consists of a mostly finished basement with hollow block wall construction.
All interior walls are built over the slab. Varying layers of aggregate were found under the
slab (0 to IS cm) with permeability ranging from 4 to 123 x Iff7 cm1 at different sites
under the slab. Subslab to wall communication was not detected in the basement during
subslab diagnostic experiments. Due to the varying depths of gravel found under the slab
and wide ranges of permeability, two mitigation pits were installed. Pit 1 was 40 cm deep
and Pit 2 was 120 cm deep. Post mitigation radon measurements indicated that the
structure had not been sufficiently mitigated although diagnostic tests revealed an excellent
pressure field extending over 90% of the slab area. Based on collected grab sample data,
an additional source of radon within the structure was determined to be the attached fallout
shelter (in excess of 3700 Bq m°). Attempts were made to isolate the shelter from the
structure by the sealing of the basement/shelter wall and the entrance door. When these
measures failed, an 8 cm (3") PVC pipe was connected from the air exhaust of the shelter
to the subslab mitigation system to depressurize the shelter. A control valve was installed
in the 3" line to allow the room to be maintained at a constant -10 Pa. This action
reduced the radon levels within the occupied structure to less than 100 Bq m°.
House OR 16 The substructure consists of a mostly finished basement with hollow block wall construction.
AH interior walls are built over the footing with a poured concrete floor. Varying layers
of aggregate were found under the slab (0 to 5 cm) with permeability ranging from 0.1 to
1.7 x 107 cm1 throughout the slab. Subslab to wall communication was not detected in the
basement during subslab diagnostic experiments. Due to the varying depths of gravel found
under the slab and the overall poor permeability, a total of four mitigation pits of 120 cm
in depth were installed.
CRAWLSPACE VENTILATION AND MODIFICATIONS
Passive ventilation of the crawlspace areas in HU13 and HU14 was attempted by the installation
of standard 20 cm (8') x 40 cm (16") vents. The addition or replacement of the vents did require some
modifications to the substructures. Where below grade vents were required, gravel lined brick-walled wells
were installed. Three below grade crawlspace vents were added on the side of the crawlspace opposite the
basement in House HU13. Two additional above grade crawlspace vents were added to the front of the
house. Three below grade crawlspace vents on the side of the crawlspace opposite the basement and one
below grade vent on the north, rear corner of the house were added in House HU14. Two additional above
grade vents were added to the front of House HU14.
Vapor barriers and active subbarrier ventilation systems were installed in the crawlspaces of the
houses. The crawlspace walls were prepared for the adhesion of the barriers by scraping and then painting
the inner walls and upper floor supports with gloss or semi-gloss masonry enamel. A straight length of
8 cm (3*) schedule 40 PVC pipe was laid roughly along the center length of the crawlspace floor with a
8 cm (3") T placed approximately in the middle of the pipe. The ends of the main 8 cm (3') trunk were
-------
then capped and 2.5 cm (1*) horizontal holes were drilled in the side of the trunk at staggered 2 to 3 m
intervals. Perforated 2.5 cm (1*) PVC pipe (0.6 cm hole per 20 cm of pipe) was then inserted into the
trunk and cemented in place with epoxy and PVC cement. The barriers were then placed over the piping
system with 90 cm overlaps. Care was taken to custom trim the barrier to conform to the various supports
and obstacles in the crawlspace. Sealing of vapor barrier material to itself and to the painted portions of
the walls was accomplished with double sided tape. From the central trunk T, a series of adaptations were
made to connect the 8 cm (3") trunk to the 15 cm (6*) pipe fan. Flexible ducting was then attached to the
exhaust side of the fan and run to an existing vent to exhaust outdoors. The flexible duct was hung from
the ceiling using permanent galvanized band-strapping.
EXPERIMENTAL METHODS
The study design included a variety of radiological, environmental, and house dynamic measurements
with near-continuous and episodic data acquisition schedules. Continuous measurements of radon,
temperature, and relative humidity were performed in substructure! (Le., crawlspace and basement) and
superstructural (Le., upstairs) zones. Differential pressures were monitored between: (1) substructure! and
superstructure! zones, (2) substructure! and outdoor zones, and (3) basement and subslab zones. These
monitors were also used on an episodic basis to evaluate the impact of clothes dryers and HAC systems on
differential pressures across various building membranes. Meteorologic measurements of wind speed, wind
direction, and rainfall were made in two locations per housing group; barometric pressure was measured in
one location. Episodic measurements included subslab radon concentration and permeability, and pressure
difference across the slab. These techniques have been described by Dudney et aL (7).
Subslab permeability was determined by applying a 10-fold or greater range of pressures to a hole
in the slab and measuring the resultant flow of exhausted air (8). In premitigation studies, suction was
applied with a variable speed shop vacuum to a 3.8 cm ID hole in the slab. Airflow was measured with
a hotwire anemometer inserted into a 2.6 cm ED PVC exhaust pipe. In several houses, the permeabilities
of the subslab soil and soil/gravel mixtures were investigated as a function of depth by coring the 2.6 cm
ID holes to depths of approximately 10, 40, 80 and 120 cm. In post-mitigation studies, suction was applied
with a variable speed mitigation blower through 10.2 and 15.2 cm holes in the slab. These holes connected
directly to SO x 50 cm gravel-filled pits having typically 40 and 150 cm depths for houses with and without
subslab gravel layers, respectively. Airflow was monitored with a hotwire anemometer inserted into a 13.9
or 15.2 cm ID exhaust pipe. Induced differential pressure was monitored at the point of entry into the slab
with an electronic vacuum gauge.
The extension of the subslab pressure field was determine by applying controlled pressures to pre-
mitigation holes or post-mitigation pits and monitoring the induced pressure through 1 cm ID holes m the
slab at several remote locations. In the pre-mitigation diagnostic phase, negative pressures of 0.5, 2.0. and
5 0 kPa were used whenever possible to enhance comparisons between different houses and time periods (e.g.,
pre- and post-mitigation phases). After mitigation, pressures were limited to about 0.35 kPa due to the
blower capacity.
Figure 1 shows scale drawings of the basements and crawlspaces of the four houses described above.
The locations of pits for the subsurface depressurization systems and of all sampling holes are indicated.
RESULTS
In a manner similar to Fowler (9), we monitored pressure field extension for nearly a year in these
houses. Figure 2 shows the results from the houses. There was no discernible change over time in the
induced pressure fields when the mitigation systems were operated normally. What is most noteworthy is
the clear difference in pressure field extension under different sectors of the slab. In House HU13, the
-------
sampling sites closest to the driveway (1, 2,7, and 17) showed a gradual drop in pressure with increasing
distance from the pit The sampling sites on the crawlspace side of the pit (3, 9,10,13, IS, 16 and 18)
showed a much steeper decline. At a distance of 7 m, there was measurable pressure on the driveway side
but not on the crawlspace side. Because of the lack of a close sampling site on the crawlspace side of the
slab at House HU14, it is unclear whether the pressure field extension varies significantly with direction.
Because multiple pits were installed at the Oak Ridge houses, only sampling sites at the ends of the houses
were included in analysis of pressure field extension. Distances were computed from the nearest pit For
'House OR16, distances were computed from Pits 2 and 3, which are those on the uphill side of the house.
At House OR1S, the pressure field is easily measurable at a distance of 5 m. It is dear from the data that
there is an enhancement of pressure extension on the side of the house where the fallout shelter is located.
Since a -10 Pa depressurization was applied to the shelter, the enhanced extension may result from
communication between shelter and subslab zones. In House OR16, the pressure field falls off rapidly on
the end by Pit 3 but not so rapidly on the end by Pit 2.
For both Huntsville houses (HU13 and HU14), data from sampling sites closest to pit showed a
decline in induced pressure during the first several months of operation (see Table 1). This effect was only
seen in the holes where the induced pressure was largest and easiest to measure. This decline in pressure
following initial start-up may reflect the removal of fine paniculate material from the flow paths and the
consequent absence of some flow restrictions and pressure drops near the point of applied depressurization.
Using the techniques described by Matthews (8), permeability was monitored at all suction pits on
a recurring basis. Detailed analysis has not been completed on all available data, but there is no clear
evidence of change in subslab permeability except at House OR16 (see Table 2). The slab at this house
was poured directly onto packed clay. We suspect that severe drought conditions during 1988 may have
contributed to shifting of the soil beneath the slab, resulting in the opening of channels leading to the
suction pits. Continued monitoring during 1989 (data not shown) has revealed continued increase in the
permeability in Pits 1 and 4. The 1989 results and results from tracer gas experiments will be reported
separately.
Radon reduction efficiency was quite good in these houses following initial installation and
adjustment of the systems (see Table 3). Reductions ranged from 89 to 97% in 1988. During 1989, there
was an extended period during which the electric power to the systems was cycled on for a week followed
by a week without power. In Figure 3, the data from the period around Julian Dates 80 to 160 (1989)
include most of the period during which the power was cycled on and off. The comparison of power-off
radon concentration to a priori radon concentrations suggests that, except for House HU13, from 40 to 80%
of the reduction may be due to the passive effects of modifications that were made to the houses. The
reader should note that this comparison is suspect due to strong seasonal variations that can occur in this
area. This effect is easily seen (see Figure 3) in the changes in indoor radon that occurred in House HUM
during the first 50 days of monitoring.
In summary, this study has shown that there is generally very good stability in the performance of
radon mitigation systems that depressurize the soil beneath concrete slabs in three of these existing houses.
One house built on a slab without aggregate (OR16) was a notable exception. There was evidence of
domains beneath the slab of House HU13 with distinctly different extensions of the field of depressurization.
There is evidence present that in three of these houses passive measures, such as slab sealing, can yield
considerable radon reduction.
-------
ACKNOWLEDGEMENTS
The authors are very grateful to the homeowners in whose houses these data were collected.
Without their patience, this study would not have been possible. This research was sponsored by the Office
of Health and Environmental Research of the U.S. Department of Energy, by the Tennessee Valley
Authority under Interagency Agreement No. 40-1602-85, and by the U.S. Environmental Protection Agency
under Interagency Agreement No. 1824-1709-A1 under Martin Marietta Energy Systems, Inc., contract DE-
AC05-84OR21400 with the U.S. Department of Energy.
us
torn of Ha
«h« te fc ». to
-------
REFERENCES
1. National Research Council. Health risks of radon and other internally deposited alpha-emitters.
Washington. DC: National Academy Press; 1988.
2. U.S. Environmental Protection Agency and Centers for Disease Control A citizen's guide to
radon: What it is and what to do about it Report #OPA-86-004 (GPO 055-000-00258-4). 1986.
3. Nitschke, L A.; Clarion, M. E.; Brennan, T.; Rizzuto, J. E>; "Long-term assessment of residential
radon-mitigation systems.* Paper #107.5 in Proceedings of 1988 meeting of the Air Pollution
Control Association; 1988.
4. Prill, R. J.; Fisk, W. J.; Turk, B. R; "Monitoring and evaluation of radon mitigation systems over
a two-year period.1 Report #LBL-25909 (OTIS DE 89 00291). 1988.
5. Gammage, R. B.; Wilson, D. L; Dudney, C. S.; Sanltz, R. 14 Summertime elevation of "Rn levels
in Huntsville, AL; submitted to Health Physics. 1990.
6. Ronca-Battista, M.; Moon, M.; Bergsten, J.; White, S. B.; Alexander, B.; Holt, N. Radon 222
concentrations in the United States • Results of sample surveys in five states. Radiation Protection
Dosimetry 24:307-311; 1988.
7. Dudney, C S.; Hubbard, L M.; Matthews, T. G.; Socolow, R. R; Hawthorne, A R.; Gadsby, K. J.;
Harrje, D. T.; Bohac, D. L; Wilson, D. L Investigation of radon entrv and effectiveness of
mitigation measures in seven houses in New Jersey. Report 0ORNL-6487. 1989.
8. Matthews, T. G.; Wilson, D. L.; TerKonda. P. K.; Saultz, R. J.; Goolsby, G.; Burns, S. E.;
Haas, J. W. Radon diagnostics: Subslab communication and permeability measurements. In
Proceedings of the 1988 Symposium on Radon and Radon Reduction Technology, Volume 1, EPA-
600/9-89-0063 (NT1S PB89-167480),p. 6-45, March. 1989.
9. Fowler, C S.; Williamson, A D.; Pyle, B. E; Belzer, F. E.; Sanchez, D. C; Brennan, T.; 'Sub-slab
depressurization demonstration in Polk County, Florida, slab-on-grade houses.* In Proceedings of
the 1988 Symposium on Radon and Radon Reduction Technology, Volume 1, EPA-600/9-89-006a
(NT1S PB89-167480),p. 7-65, March, 1989.
-------
HIM 3
*1
Garage ,*
f 2
Garage
7
Scale: -^
::*:::
*T*
SSO Pit J
© J
^ V
3 L
* j
,18 t
•• r
Slab Soil w/ Barrier
*
10
*
13
*
15
5m
HU14
*
3
*
2
+ (
Scale:
5 fn
SSDP
* ©
5
3arage
*
6
t
*
4
*
.••'
7
8
...j.
I
..J.
'•• 1 *
1 9
1
Crawlspace
^ Soil w/ Barrier
OR15
,,* (£
L_
I
\m
Scale:
4m
*
i
OR16
Scale:
* I*
6 20
5
4
, ,*,,-./
* 1R
-* J^
*8
rF"11
Pit 2
TITf
1 Pit 4
2
^3*
V
Vs
* 1O
12 *
11
*
13f-
Figure 1. Schematic of substructures indicating depressurization points and sampling locations.
-------
HU13
Data from Sites:
1.Z7.17
Data from Sites:
3. 9.10. ia 15.16.18
HUM
OR15
OR16
Data from Sites:
8.9.10. 11.12
Data from Sites:
1.Z3.4
Data from Sites:
4.5.6.8.16.20
Data from Sites:
7.10.11. 1£ 13,14. 15
Figure 2. Variation of pressure with distance under slab for four houses.
-------
cr
m
(0
oc
Mt M
1987
1989
Julian Date
OR15
"?9B7
1 _______ •
at M at
Julian Date
1989
Julian Date
X) X3 M
1987
Figure 3. Daily average radon concentrations in four houses.
Julian Date
-------
Table 1. DIFFERENTIAL PRESSURE (Pa)
AFTER INSTALLATION.
House/Sampling Site
Date 13/3 14/5
8/88 -28 (113%*) -43 (118%)
10/88 -25 (103%) -39 (106%)
11/88 -23 ( 96%) -35 ( 97%)
12/88 -22 ( 89%) -36 ( 98%)
1/89 -23 ( 93%) -38 (105%)
2/89 -22 ( 89%) -35 ( 97%)
7/89 -24 (100%) -37 (100%)
' Fraction of 7/89 value, expressed
as a percentage.
Table 2. VARIATIONS IN SUBSLAB PERMEABILITY (x 107 cm'2)
AT OR16 IN 1988.
Date Pit 1 Pit 2 Pit 3 Pit 4
5/88 4.0 ± 0.5 7.4 ± 1.8 5.1 ± 1.3
6/88 — — — 10.3 ± 1.2
11/88 16.7 ± 1.0 5.5 ± 0.5 5.1 ± 1.3 20.0 ± 1.1
12/88 14.6 ± 0.9 5.2 ± 0.5 5.1 ± 0.8 18.8 ± 2.0
-------
Table 3. RADON REDUCTION EFFICIENCY: 1988 AND 1989.
1988 1988 1989 1989
House Pre-Mitipation Post-Mitipation System Off System On
HU13 191 Bqm4 21 235 32
HU14 958 30 220 63
OR15 585 26 218 35
OR16 724 78 315 100
-------
VII-6
A COST-EFFECTIVENESS COMPARISON OF PRIVATE-SECTOR RADON REMEDIATION
WITH TRADITIONAL RADIATION PROTECTION ACTIVITIES
by: Daniel J. Strom
William D. Ulicny
Department of Radiation Health
Graduate School of Public Health
University of Pittsburgh
Pittsburgh, PA 15261
John B. Mallon, Jr.
Richard W. Benchoff
Radon Detection & Control
P.O. Box 419
South Heights, PA 15081
ABSTRACT
Private-sector radon remediation can be well over 100 times more cost-effective than the
minimum expenditure of $1000 to avert a whole-body person-rem mandated by the U.S. Nuclear
Regulatory Commission for marginal control of exposures to radiation due to effluents from
nuclear power plants (10 CFR 50 Appendix I). Using data from the Pennsylvania Department
of Environmental Resources (DER) and from Radon Detection & Control (RDC; a commercial
radon remediation firm) spanning more than two years, we relate differences in radon
concentrations measured before and after remedial action to capital and operating costs. Simple
demographic assumptions and concentration-to-collective dose equivalent conversions lead to a
quantitative demonstration of the dramatic imbalance in society's valuation of radiation
protection by nuclear utilities when compared to that in the home. The DER data show that an
effective person-rem can be averted for $3.01 in high-radon houses (104 houses; average, 189
pCi/L; median 112 pCi/L; GSD 2.75), while the RDC data show that an effective person-rem
can be averted for $9.41 in more typical houses (201 houses; average 28 pCi/L; median 19;
GSD 2.31).
-------
INTRODUCTION
A basic tenet of public health activities is that limited resources should be used in a way
that is most beneficial to the most people. Studies on the value of a human life have
demonstrated the great incongruities in how money is spent for increments of health and safety
(1,2). In this paper, we demonstrate that radon remediation activities can be 100 to 300 times
more cost effective than radiation protection activities at nuclear power plants required by U.S.
Federal regulations (3).
In the case of nuclear power, the risk is perceived as dread, unknown, uncontrollable,
imposed, and manageable with someone else's money. In the case of radon exposure in the
home, the risk is perceived as less dread, better known, more controllable, voluntarily accepted,
and manageable only with one's own money. The traditional perception of the home as a "safe
haven" from the ills of the outside world runs counter to the acceptance of a serious
radiological risk that may exist from natural (or, rather, technologically-enhanced natural) causes.
The public health perspective, however, dictates that society's resources should be refocused on
indoor radon, rather than on tiny increments of safety achieved at tremendous cost in controlling
effluents from nuclear power plants.
DATA
Two sets of data are used in this study. The fust is from the Pennsylvania Department
of Environmental Resources (DER) report summarizing the Pennsylvania Radon Research and
Demonstration Project (4). This report lists radon concentration data from 104 houses for which
measurements before and after remediation were available. The DER data arose from research
and development work, and the houses to be remediated had very high radon concentrations
(many over 200 pCi/L). These data are not a random sample of houses, and remediations were
done in the mid-1980s when the technology was in its infancy. Concentration measurements
were made using grab or continuous flow-through alpha scintillation measurements. There is no
evidence that measurements were made using integrating detectors such as charcoal, diffusion
barrier charcoal, or alpha track.
The second set of data are from Radon Detection & Control (RDC). These data
included 201 houses remediated between 1987 and 1989, mostly in the Allegheny and Beaver
County areas of western Pennsylvania. The RDC data represent a more typical set of radon
remediations in that they were done for real estate transactions, for relocation companies, or for
health concerns of the occupants. The RDC radon measurements were made using activated
charcoal screening detectors' exposed for two to four days and processed commercially. In
some cases, the concentrations are averages of two or more measurements. The radon
concentrations were measured at the same location before and after remediation.
ANALYSIS AND DISCUSSION
The DER data are summarized in Table la, while the RDC data are summarized in
Table Ib. From left to right, the columns are the initial or "Before" remediation radon
'Key Technology, Inc., P.O. Box 562, Jonestown, PA 17038; or Air Chek. Inc., P.O. Box 2000, Ardcn,
NC 28704.
-------
concentration; the final or "After" remediation concentration; the remediation cost; the percent
reduction of radon concentration; the percent reduction in radon concentration per $1000 spent
on remediation; the change in concentration; the change in concentration per $1000 spent on
remediation; the ratio of the "Before" to the "After" concentrations; and the ratio reduction
achieved per $1000 spent on remediation. Note that minimum and maximum entries in different
columns may be from different houses; these values were the minima and maxima in the entire
data set.
Table la. Summaries of the Pennsylvania D.E.R.data (Weston & Simon 1988).
Radon
Cone.
Before
% Re-
duction
Cone.
Change
Reduct.
Factor
Cone. Cone. Reined- duction Cone. Change Factor
Before After iation %Re- per 1000$ Change per 1000$ ReducL per 1000S
(pCi/L) (pCi/L) Cost($) duction %/k$ (pCi/L) (pCi/L)/k$ Factor (/k$)
No. of houses
Minimum
Maximum
Mean
Std.Dev.
Geo. Mean
G.S.D.
r*2 (lognorm.)
104
11.6
1653.2
189.1
240.2
112.3
2.75
0.992
104
0.3
178.8
10.6
27.7
2.5
4.43
0.930
104
1301
16185
3729
2156
3324
1.57
0.966
104
-53.2
99.9
90.4
23.2
92.7
1.21
104
-13.0
73.7
29.6
14.2
27.8
1.59
104
-62.1
1576.2
178.5
231.3
107.4
2.91
104
-15.1
312.3
47.4
49.7
32.3
2.61
104
0.65
893.0
112.4
145.0
44.8
4.96
0.972
104
0.16
222.9
33.3
40.7
13.5
4,98
Table Ib. Summaries of Radon Detection & Control data, 1987-89.
Radon Radon % Re-
Conc. Cone. Remed- duction
Before After iation %Re- per 1000$
(pCi/L) (pCi/L) Cost($) duciion %/k$
Cone.
Cone. Change
Change per 1000$
(pCi/L) (pCi/L)/k$
Rcduct.
Factor
Reduct. per 1000$
Factor (/k$)
No. of houses
Minimum
Maximum
Mean
Std.Dev.
Gco. Mean
G.S.D.
r*2 (lognorm.)
201
4.2
156.0
28.1
29.0
19.2
2.31
0.969
201
0.0
18.3
2.0
2.4
1.3
2.53
0.993
201
150
2905
1030
325
957
1.55
0.862
201
-2.7
100.0
87.3
15.8
86.1
1.24
201
-1.4
634.4
99.0
63.4
90.3
1.47
201
-0.3
155.6
26.1
28.4
16.6
2.59
201
-0.2
380.7
27.7
37.3
17.4
2.50
199
1.0
390.0
29.3
46.8
14.5
3.21
0.997
201
0.0
421.1
30.3
52.9
15.4
2.97
Under each column are the number of houses used in that column; the minimum,
maximum, arithmetic mean, and standard deviations of each quantity; and the geometric mean
and geometric standard deviations (GSD; dimensionless) of the quantities. For the DER data,
the geometric means and GSDs for the difference, % reduction, concentration change per $1000,
and percent per $1000 were calculated from 101 data pairs, omitting the 3 negative values. For
the RDC data, geometric means and GSDs were calculated from 197 to 199 non-negative values
-------
for those columns with negative entries, or, in the case of the ratio, with zero final
concentration.
The DER houses had larger average initial ("Before" remediation) concentrations than did
the RDC houses, with arithmetic means of 189 ± 240 and 28 ± 29 pCi/L, respectively. This is
due to the different selection processes, and to the fact that many of the DER houses were in
the Reading Prong area of southeastern Pennsylvania, a region with exceptionally high indoor
radon levels. Both data sets showed lognormal distributions, as seen in Figures 1 and 2, with
geometric means of 112 I 2.75 (r2 = 0.992) and 19 ; 2.31 (r7 = 0.969). Lognormal fits were
done by the LPROBIT code using Finney's weighting method (5).
The maximum in the DER data set, 1653 pCi/L, was more than ten times higher than
the maximum in the RDC data set, 156 pCi/L. The minimum "Before" value (i.e., 4.2 pCi/L)
in the RDC data may indicate the force that the real estate and relocation companies have on
radon remediation decisions.
Final ("After" remediation) concentrations averaged 10.6 and 2.0, respectively, for the
DER and RDC data. The lower final concentration for RDC is understandable in light of the
fact that they usually guarantee a concentration below 4 pCi/L, and usually continue to
remediate until the concentration is below this level. Again, both data sets showed lognormal
distributions, with geometric means of 2.5 ; 4.43 (r2 = 0.930) and 1.3 : 2.53 (r2 = 0.993). The
wide GSD, 4.43, for the DER data, along with its moderate deviation from lognormality may
reflect the varying conditions applied to remediation during the DER's research and development
process.
The DER data show significantly higher average remediation costs than the RDC data:
$3729 ± $2156 vs. $1030 ± $325. The geometric means were $3324 : 1.57 (r2 = 0.97) vs.
$957 I 1.55 (r* = 0.86). These numbers are understandable in light of the differing nature of
the two endeavors: the DER project was a research and development activity, done to develop
technologies; the RDC work is done commercially in a competitive atmosphere, and benefits
from well-established practices in remediation.
We chose three ways of looking at the data: percent reduction, concentration change,
and the reduction factor. The percentage reduction is the difference in radon concentrations
divided by the "Before" concentration, expressed as a percent; the concentration change is
simply the difference between "Before" and "After" concentrations; and the reduction factor is
the ratio of the "Before" to "After" concentrations. All three methods are somewhat flawed
because no concurrent long-term background concentration measurements were available. Since
indoor radon levels cannot be reduced below ambient outdoor levels using any of the
remediation methods employed for DER or by RDC, both percent reduction and reduction factor
have values limited by background concentration. The limits are unimportant for high initial
concentrations, but can restrict measures of success when starting from lower concentrations.
The percentage reduction summaries are given for each data set in Table 1. The means
are 90.4% for the DER data, and 87.3% for the RDC data. The mean percentage reduction is a
somewhat misleading statistic, since its maximum value is 100, with many houses very near this
value. The mean percentage reduction per $1000 spent on remediation is 29.6%/$1000 for DER
vs. 99.0%/$1000 for RDC, which primarily reflects the different average remediation costs.
-------
Pennsylvania DER Data
With Finney Regression Predictions
o
^ft
ti
o
•l-t
-p
cd
o
a
o
u
ti
o
-d
cd
1000
100
10
0.1
Pre
0.13
2.28
15.87 50 84.13 97.72 99.87
Percentile
Figure 1. Log-probability :plot of "Before" and "After" radon concentrations in 201 houses
remediated by Radon Detection & Control, with Finney-weighted regression lines.
Radon Detection & Control Data
With Finney Regression Predictions
Avg. Geo.Mean GSD
Before 28.1 19.2 2.31
After 2.0 1.3 2.53
2.28 15.87 50 84.13 97.72 99.87
Percentile
Figure 2. Log-probability plot of "Before" and "After" radon concentrations in 201 houses
remediated by Radon Detection & Control, with Finney-weighted regression lines.
-------
The concentration change summaries show arithmetic means of 179 ± 231.3 pCi/L for
the DER data and 26.1 ± 28.4 pCi/L for the RDC data, the difference largely reflecting the fact
that the DER houses had higher concentrations to begin with. The geometric means were 107
pCi/L ; 2.91 and 16.6 pCi/L I 2.59, respectively. The mean concentration change per $1000
was 47 ± 49 [pCi/L]/k$ and 28 ± 37 [pCi/L]/k$ for the DER and RDC data, respectively,
indicating that, while the concentration changes in the DER work averaged nearly 7 times
greater than the RDC work, the remediation cost per unit concentration change averaged only
1.7 times greater for the DER work. Note that this analysis does not include operating and
maintenance costs.
Perhaps more interesting than the percentage reduction is the reduction factor. Reduction
factors as great as 893 were achieved in the DER program, and 390 in the RDC work.
Average reduction factors were 112 and 29, respectively, with geometric means of 45 I 4.96 (r2
= 0.972) and 29 ; 3.21 (r2 = 0.997), indicating both more vigorous remediation efforts in the
DER demonstrations and the limiting effects of ambient radon concentrations when remediating
houses with initially low radon levels. Of great interest are the mean reduction factors per
$1000 spent on remediation: these were 33/S1000 and 30/S1000 for DER and RDC, respectively,
with geometric means of 13.5/S1000 and 15.4/S1000, indicating that the work was quite
comparable in terms of cost-effectiveness when expressed in terms of reduction factors!
The dose equivalent averted through remediation is directly proportional to the
concentration change. To evaluate the overall cost-effectiveness of the two remediation projects,
we summed the "Before" and "After" concentrations, as well as the costs, as shown in Table 2.
These concentration sums can be considered as "Collective Concentration" by analogy with
"Collective Dose" and "Collective Dose Rate" as defined by the International Commission on
Radiological Protection (6).
Table 2. Sums of concentrations and costs, with analysis statistics for PA D.E.R. and RD&C data.
Sum of Sum of Overall Overall Overall
Radon Radon Sum of %Re- Overall Cone. Reduce
Cone. Cone. Remed- Overall duction Cone. Change Overall Factor
Before After iation %Re- per 1000$ Change per 1000$ Reduct per 1000$
(pCi/L) (pCi/L) Costs ($) duction %/k$ (pCi/L) (pCi/L)/kS Factor (/k$)
PAD.E.R. 19664 1102 387797 94.4 25.3 18562 47.9 17.8 4.8
R.D.&C. 5649 402 206936 92.9 90.2 5248 25.4 14.1 13.7
The columns in Table 2 that begin with the word "Overall" differ from the mean or
geometric mean values presented in Table 1 because they are computed only from the sums
given in Table 2, rather than from the actual distributions. The overall percent reduction in
each data set (94.4% and 92.9%, respectively, for DER and RDC) is greater than the mean
because greater reductions were achieved in high initial concentration houses than in low
concentration houses. The percent/Si 000 values (25% and 90%) are based on the mean
remediation costs in Table 1, and are lower than the means in that table.
-------
The collective dose equivalent averted by remediation is proportional to the overall
concentration change, which was 18562 pCi/L for the DER work and 5248 pCi/L for the RDC
remediations. While the DER overall concentration change is 3.5 times higher than the RDC
change, the cost-effectiveness value for DER is only 1.9 times that for RDC (47.9 vs. 25.4
[pCi/L]/$1000) based on average remediation costs from Table 1.
The overall reduction factors were 17.8 and 14.1, respectively, for DER and RDC, while
the reduction factor per $1000 values were 4.8/$1000 and 13.7/$1000, indicating an almost
three-fold better investment with the later work when viewed in these terms.
Rather than dealing with risk, as the DER report does, we compute the effective dose
equivalent averted by remedial actions, and compare the cost per unit effective dose equivalent
averted by remedial actions. The many assumptions needed to do this are summarized in Table
3. We calculate 1.93 lung-rems (to an adult) from a 1-year exposure to a concentration of 1
pCi/L of radon from the product (1 WL/[200 pCi/L])(8766 hours/environmental year)(l
occupational month/170 hours)(0.75 fraction of time spent at home)(0.5 rad/WLM)(20 rems/rad =
Quality Factor for alpha radiation). Effective dose equivalent rate per unit concentration (0.23
effective rems/([pCi/L]-year) is obtained by multiplying by the ICRP weighting factor for the
lung (6), i.e., w,., = 0.12.
Table 3. Assumptions used in calculating the expenditure per person-rem.
200 [pCi/L]/WL, i.e., 50% equilibrium
0.5 rad/WLM (ref. 7)
20 Q (rem/rad; ref. 6)
0.12 w,,,,,,, i.e., effective rem per lung rem (ref. 6)
0.75 fraction of time spent at home
8766 hours/environmental year
170 hours/occupational month
4 persons per house
10 years life expectancy for radon control system
125 $/year energy penalty (ref. 11, modified)
29 Watt fan'
0.10 S/kWru*
1.93 lung rems/([pCi/L]-year)
0.23 effective rems/([pCi/L]-year)
62.7 effective microsieverts/([Bq/m3]-year)
25.42 $/year, fan operation cost
1 Dayton Electric Mfg. Co., 5959 W. Howard St., Chicago IL 60648; Model 4C720.
b Duquesne Light Co., Pittsburgh, PA; 1989
The value of 0.5 rad/WLM is taken from Harley and Cohen (7), and is probably an
underestimate of the conversion for children (8, 9), for which estimates as high as 1.2 rad/WLM
are available (8). Additionally, the value of 0.75 of one's time in the house is somewhat higher
-------
than the values of 0.64 to 0.73 used by some at the EPA (10); however, the EPA work is based
on data which dramatically under-represent infants, small children, and the elderly.
Additionally, we assume that there are 4 persons per house, and a ten-year life
expectancy for the remediation system. We further assume $125/year for operating costs and
the cost of loss of conditioned air. Due to the use of a 29 W fan by RDC, rather than the
more common 90 W fan, we have lowered our estimate of annual costs below the $150
estimated by Henschel (11). It is important to point out that the annual cost is not well known,
and is limited at the lower end by the fan operation cost ($25.42), but may be much greater
than our estimate in some cases.
A critical assumption in our calculations is that the change in radon concentrations are
representative of the air that house occupants actually breathe. We do not know the proportion
of houses for which this is the case. In some cases, the radon measurements are simply
screening values measured in the lowest livable area of a house; in other cases, they are
representative measurements of air that the occupants breathe. If a ratio for basement
concentration to first floor concentration of 2.5 is used (12), then our calculated average cost per
(unit collective effective dose equivalent [CEDE] averted) could be multiplied by no more than
this factor.
The average cost per unit CEDE averted is tabulated in Table 4. The remediation cost
plus the operation and maintenance costs are added to yield totals of $4979 for the DER work
and $2280 for the RDC remediations. Average concentration changes are multiplied by 0.23
effective rem per ([pCi/L]-year) and multiplied by 4 persons and 10 years to get the 10-year
CEDE averted values of 1657 person-rems per house for the DER houses and 242 person-rems
for the RDC houses. These values, divided by the total 10-year cost, give the average cost per
unit CEDE averted, in dollars per person-rem. The values for the DER and RDC houses are
$3.01/person-rem and $9.41/person-rem, respectively.
Table 4. Average cost per unit collective effective dose equivalent averted in radon
remediation and as required of the nuclear power industry.
Average
Cost of Average 10-year CEDE* Cost per unit
10-year Rn Cone.
Energy Capital Change
Penalty Outlay Total (pCi/L)
$1250 $3729 $4979 178.5
$1250 $1030 $2280 26.1
PA D.E.R.
This Study
10 CFR 50"
10 CFR 50?
10-year CEDE'
Averted
(person-rems
/house)
1657
242
CEDE'
Averted
($/person-rem)
$3.01
$9.41
$1000.00
$33333.33
1 Collective Effective Dose Equivalent
b (App.I Sec.H.D.; whole body)
c (App.I Sec.H.D.; thyroid)
The U.S. Nuclear Regulatory Commission has set values of $1000 per whole body rem
and $1000 per thyroid rem as the minimum amounts that must be expended to limit offsite
-------
releases of radioactivity by nuclear power plants (3). The latter value can be converted to
effective dose equivalent by dividing by the ICRP risk-based weighting factor for the thyroid,
i.e., 0.03, giving $33,333.33. The NRC cost per unit CEDE averted values are roughly 300 and
10,000 times greater than the DER values, and 100 to 3,000 times the more realistic RDC
values. Clearly, these numbers are inconsistent.
If a cancer fatality rate of 2 x 10"* per person-rem is assumed, dollar values to save a
human life can be computed. These are $15,000 and $47,000 for the DER and RDC
remediations, and $5,000,000 and $167,000,000 for the NRC release limits. The radon
remediation costs are in line with other common societal expenditures for life saving (1,2), but
the NRC values are not.
CONCLUSIONS
By analyzing two sets of radon remediation data, we have rhown that remediation can
be a cost-effective activity when compared to other life-saving measures in our society.
Expenditures in other areas of radiation protection, such as those required of nuclear power
plants by the U.S. Nuclear Regulatory Commission, can be hundreds or even thousands of times
less cost-effective.
REFERENCES
1 Cohen, BJL. Society's valuation of life saving in radiation protection and other contexts. Health
Phys. 38:33-51, 1979.
2 Graham, J.D., and Vaupel, J.W. Value of a life: What difference does it make? Risk Analysis
1:89-95, 1981.
3. U.S. Nuclear Regulatory Commission. Title 10, Code of Federal Regulations, Part 50. Appendix I.
Originally published in 40 FR 40816, but still in effect in 1990.
4. Roy F. Weston, Inc., and R.F. Simon Company, Inc. Final report of the Pennsylvania radon
research and demonstration project. Pennsylvania Department of Environmental Resources, bureau
of Radiation Protection, Harrisburg, PA, 1988.
5. Strom, D.J. LPROBIT. Health Phys. 57:VII-VIII. July, 1989.
6. International Commission on Radiological Protection. Recommendations of the International
Commission on Radiological Protection. ICRP Publication No. 26. Oxford: Pergamon Press. 1977.
7. Harley, N.H. and Cohen, B.S. Updating radon daughter bronchial dosimetry. In: P. Hopke (ed.),
Indoor Radon. American Chemical Society Symposium Series 331. Washington, DC: American
Chemical Society, 1987.
8. National Council on Radiation Protection and Measurements. Evaluation of occupational and
environmental exposures to radon and radon daughters in the United States. NCRP Report No. 78.
Bethesda, MD: NCRP Publications, 1984.
-------
9. Hofmann, W., Steinhausler, F., Pohl, E. Dose calculations for (he respiratory tract from inhaled
natural radioactive nuclides as a function of age - 1. Health Phys. 37:517-532, 1979.
10. Exposure Assessment Group, U.S. Environmental Protection Agency. Exposure Factors Handbook.
EPA/600/8-89/043. Washington, DC: United States Government Printing Office, 1989.
11. Henschel, D.B. Radon reduction techniques for detached houses. Technical guidance. 2nd ed.
EPA/625/5-87/019. Washington, DC: United States Government Printing Office. 1988.
12. Cohen, B.L. Variation of radon levels in U.S. homes with various factors. Journal of the Air
Pollution Control Association 38:129-134, 1988.
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.
-------
THE EFFECTIVENESS OF RADON REDUCTION IN NEW JERSEY VII-7
By: Nick DePierro, Tonalee Key and Jennifer Moon
New Jersey Department of Environmental Protection
CN 415, Trenton, New Jersey 08625
ABSTRACT
The results of 473 homeowner-funded radon mitigations conducted
in New Jersey from April, 1988 to November, 1989 are reported and
compared to similar results previously reported for the period
January, 1986 to March 1988 (1988 Study). Data presented is
compiled from the New Jersey Department of Environmental
Protection's (DEP) post remediation testing program. A review of
the data compiled from the post remediation testing program
indicates that soil depressurization is still the predominant
mitigation method employed by certified firms, while soil
depressurization and sealing only methods tie for predominant
mitigation method employed by homeowners and non-certified firms.
Lowest floor radon concentrations following mitigation were
below 4 picocuries per liter (pCi/1) in 56% of the homes, compared
to only 36% in the 1988 study. This increase may be explained by
the fact that radon mitigators are more experienced and knowledgeable
in radon diagnostics and mitigation design. The average cost of
mitigation, as determined from reports submitted by mitigation
firms, was $1,100 compared to $1,300 in the 1988 study.
Introduction
In order to monitor and evaluate effective radon reduction
techniques for homes, the DEP has developed a program to track radon
mitigation efforts. Mitigation techniques monitored through this
program include sealing only, soil depressurization, and forced air
and crawl space ventilation. The latest procedures and guidelines
for this technology have been provided to the public and private
sectors of New Jersey through the dissemination of The United States
Environmental Protection Agency (EPA) documents, and phone
consultations provided by DEP technical staff.
The DEP instituted a voluntary mitigation certification
program in 1986. Successful program applicants are listed in a
brochure which is distributed to the public. One requirement of
this program is the successful completion of a DEP approved
mitigation training course, which, also serves as a means of
technology transfer. A mandatory certification program will be
instituted during 1990, and will require all individuals or firms
who test for or reduce radon levels in the State to be certified by
the DEP.
The DEP has compiled data on homeowner funded radon reduction
efforts since January, 1986. Evaluation of this data has enabled
the DEP to determine (1) what mitigation methods are being selected
-------
by New Jersey residents, (2) who is performing the mitigation work,
(3) the effectiveness of mitigation methods installed by both
homeowners and voluntarily certified firms and, (4) the average cost
of each mitigation technique.
The purpose of this paper is to report the results of data
collected on 473 mitigations performed in New Jersey from April,
1988 to November 1989 and to compare these results with previously
reported data collected from January, 1986 to March, 1988.
Evaluation of the effectiveness of mitigation techniques will help
the DEP to identify and target those areas where more training and
information transfer may be needed.
Use of the term "certified firm" in this paper means a radon
mitigation firm which has successfully participated in the DEP
Voluntary Certification Program. "Non-certified firm" refers to
those firms and individuals that either elected not be participate
in the DEP voluntary program or failed to meet minimum requirements.
"Effectiveness" means those post remediation radon levels in the
lowest liveable area which are less than four picocuries, per liter
(4 pCi/1).
Sources of Data
The DEP's post remediation testing program is designed to track
homeowner-funded radon reduction efforts in the State. The purposes
of this program are to provide homeowners, at no cost, with radon
testing to assess the effectiveness of mitigation measures installed
and to monitor the performance of the mitigation industry in the
State. Testing in the program was initially conducted through field
visits by DEP staff to remediated homes. Two carbon canisters were
deployed on the two lowest liveable levels of the home and a
post-remediation survey form was completed.
Since March, 1988, post remediation testing has been provided
through a mail-out program. Accordingly, homeowners are encouraged
to contact the DEP and arrange for post-remediation testing. DEP
staff complete phone information forms which include the resident's
name, address, pre-mitigation test results, if available, mitigation
technique installed, and the person or firm who installed the
system. The resident is then required to submit to the DEP, a copy
of his or her mitigation contract or a description of the system
installed. Upon receiving this information, the homeowner is mailed
two charcoal canisters with specific instructions on how to deploy
and return the devices. A mitigation installation is determined to
be effective when post remediation radon concentrations, under
closed house conditions, are less than 4 pCi/1 in the lowest
"liveable" floor. The review of the data which follows has been
derived from 473 post remediation surveys conducted by DEP from
April, 1988 to November, 1989. Comparisons are made with similar
data collected on 716 remediations from January, 1986 through March,
1988 (the 1988 Study).
-------
A second source of data on mitigation activities was collected
through the DEP's voluntary mitigation certification program. Firms
participating in the program are required to submit quarterly
survey reports on all radon remediations conducted in New Jersey.
In addition to the information obtained through the DEP
post remediation testing program, firms are required to include the
cost of each mitigation system installed. Data has been compiled
from firm reports submitted to DEP for 3,197 homes mitigated from
April, 1988 to October, 1989. Because evaluation of this data
relies on accuracy and the honesty of firm reporting, it is used
only as a comparison to DEP post remediation program data and in
determining mitigation costs.
Mitigation Methods
Methods employed to reduce indoor radon levels in the 473 homes
monitored through the DEP post remediation testing program have been
grouped into four major categories which include (1) sealing only
techniques, (2) soil depressurization methods, (3) air to air heat
exchangers and, (4) other methods.
Sealing only techniques typically consist of covering or
filling in the sump pump pits and perimeter french drains, and
sealing of floor/wall cracks and utility openings. The extent to
which sealing measures were used varied considerably, especially
with techniques and materials. According to phone conversations,
homeowners/non certified firms were less likely to follow guidance
measures outlined in EPA documents. Cracks were often not properly
prepared before applying sealants and in some cases ordinary cement
was used as the sealant rather than non-shrink urethane caulks.
Soil depressurization methods include the use of active and
passive sub-slab ventilation, drain tile, sump pump, and block wall
suction. Active sub-slab ventilation was the most widely employed
soil depressurization method and was generally accompanied by
sealing of major radon entry routes. However, after examining
mitigation contracts and design descriptions, it became apparent
that some certified firms and homeowner/non-certified firms did not
always follow EPA installation guidelines. The locations of exhaust
fans and exhaust ports was the most often noted non-conforming
practice. Basement mounted fans often leak and improperly located
exhaust ports may allow radon gas to re-enter a home. This could
explain the ineffectiveness of some systems. Also, some homeowners
and non-certified firms did not properly seal radon entry routes,
thus reducing system effectiveness.
Air to air exchangers include those systems with heat recovery
only. The principle of this design is to increase the ventilation
rate in the house thus diluting the indoor radon concentration while
decreasing the natural exfiltration of indoor air and reducing the
pressure differential which causes soil gas to be sucked into the
home. Only 14 such mitigation attempts were observed in this study.
This may be explained by the fact that the installation of air to
-------
air heat exchangers is often expensive and requires professional
services for proper sizing and installation. Additionally, this
type of mitigation is less effective than soil depressurization in
reducing radon levels and is more costly to operate and maintain.
The last mitigation type considered in this study, "other"
methods, includes forced air and crawl space ventilation. This type
of mitigation includes passive venting or installing fans or blowers
in basement and crawl space walls and windows. These measures
usually require a minimal amount of technical skill and money, which
may explain why they were most often installed by homeowners. Only
18 such mitigation attempts were reported.
Selection and installation of mitigation methods were
influenced by who was performing the work. Homeowners/non-certified
firms with some knowledge of buildings construction often performed
their own mitigation work employing mainly sealing techniques.
Certified firms overwhelmingly selected depressurization techniques
as a means of remediation (166 out of 227 installations). Table 1
shows the homes remediated for each of the 4 mitigation methods
discussed above. As shown, the most widely employed technique
selected by homeowners/non-certified firms and certified firms was
soil depressurization (58% of homes). Sealing only as a mitigation
technique was used in 35% of the homes, air to air exchangers were
installed in 3% of the homes and other techniques in 4% of the
homes.
When compared to Table 2, which shows the 716 homes mitigated
in the DEP's 1988 Study, some obvious conclusions can be drawn.
Installation of soil depressurization systems remained relatively
unchanged with 58% of the homes in this study and 59% of the homes
in the 1988 study selecting this method. Sealing only was employed
in 35% of the homes in this study and in 28% of the homes in the
1988 study, and was again the dominate method employed by
homeowner/non-certified firms. Certified firms installed mitigation
systems in 48% of the homes in this study and as in the 1988 study
selected soil depressurization techniques most frequently,
comprising 73% of all work which they performed. An interesting
observation is in the difference in the percentage of homes that
were mitigated by homeowner/non-certified firms and DEP certified
firms. In the 1988 study, 47% of the homes were mitigated by
homeowners/non-certified firms, and 53% by certified firms. In this
study, 52% of the homes were mitigated by homeowner/non-certified
firms and 48% by certified firms. Perhaps this is a result of
further dissemination of radon reduction information to the public
sector.
Effectiveness of Mitigation Systems
The performance of the mitigation systems installed in the 473
homes in this study was evaluated using post-mitigation radon levels
measured by the DEP. Two canisters were provided to the homeowner
with placement instructions in accordance with DEP/EPA testing
protocols. Mitigated homes with radon concentrations less than 4
pCi/1 in the lowest liveable level were considered effective
-------
remediations. It should be emphasized that post remediation testing
was conducted using short term charcoal canisters and therefore,
these results are not indicative of annual averages.
The performance of all mitigation methods installed in this
study is compared to similar data on the 716 homes surveyed in 1988
and is shown in Figure 1. The histograms show the distribution of
post remediation radon levels in homes mitigated by homeowner/
non-certified firms and certified firms. The distribution of post
remediation radon levels shows that in this study, 56% of the homes
mitigated had radon levels below 4 pCi/1. This was a significant
increase when compared with the 1988 study which shows only 36% of
the mitigated homes with post remediation levels less than 4 pCi/1.
A closer look at the homes mitigated by certified firms only
reveals an even greater increase in the number of homes mitigated
below 4 pCi/1. Figure 2 compares levels in 225 homes mitigated by
certified firms in this study with 382 homes mitigated by certified
firms in the 1988 study. The distribution of radon levels shows
that 78% of the homes in this study had post remediation levels less
than 4 pCi/1 compared to only 48% in the 1988 study. This shows an
increase in mitigation effectiveness of 30%. This higher success
rate may be attributed to the industry's increased experience in
diagnosing, designing and installing proper mitigation systems.
A comparison of each mitigation method installed by homeowner/
non-certified firms and certified firms is presented in Figure 3.
As can be seen, certified firms were significantly more effective in
reducing indoor radon levels to below 4 pCi/1 for all methods.
Horaeowner/non-certified firms were only 27% effective using sealing
measures only while certified firms were 65% effective in reducing
radon levels. Certified firms were even more successful in
installing effective soil depressurization systems, with 81% of the
systems reducing radon levels below 4 pCi/1. This is a somewhat
greater performance rate when compared to all methods used by
certified firms (78%). This is encouraging information since
certified firms are generally employed by homeowners with high radon
levels or when do-it-yourself attempts have failed. A conclusion
that can be drawn from this information is that properly trained
certified firms have had a consistently higher success rate than
homeowner/non-certified firms for all mitigation methods in reducing
radon levels below 4 pCi/1.
To further investigate the effectiveness of mitigation systems
to reduce lowest floor radon levels to below 4 pCi/1, a comparison
of the distribution of post remediation radon levels for each
mitigation technique is presented in Figures 4, 5, 6 and 7. Sealing
methods only is illustrated in Figure 4. A dramatic increase in
performance can be seen as the percentage of homes with post
remediation radon levels less than 4 pCi/1 more than doubled from
15% in the 1988 study to 37% in this study. The continued high
failure rate of sealing methods only, 63% in this study, might be
attributed to the fact that this method is most often selected by
inexperienced homeowner/non-certified firms who may not know how to
properly seal or know when such techniques alone will be successful.
-------
The effectiveness of air to air exchangers is shown in Figure
5. The distribution of post remediation results indicate an
increase in performance in homes remediated from 23% in the 1988
study to 57% in this study. One point should be emphasized in
explaining this significant increase. In this study only 14 air to
air heat exchangers were examined while in the 1988 study 57
installations were examined. Of the 57 installations, 23 were of
one particular design which often failed, thus accounting for 77% of
the homes having levels greater than 4 pCi/1. This particular
system was not noted in any of the designs surveyed in 1989.
The distribution of radon levels in homes in which soil
depressurization systems were installed is shown in Figure 6.
Although the performance of these systems has moderately increased
from 50% in the 1988 study to 67% in this study, it is still not
near the 90% effectiveness reported by firms which participated in
EPA's pilot survey of private sector firms. One explanation for
the 33% failure rate may be in the high number of soil
depressurization attempts by inexperienced mitigators. Forty
percent of all soil depressurization systems surveyed in this study
were installed by homeowner/non-certified persons (see Table 1).
Since these methods require diagnostic surveys and sealing measures
to maximize performance, they are more likely to be improperly
installed by inexperienced personnel.
Figure 7 shows the distribution of post remediation radon
levels in homes utilizing "other" methods. In the 1988 survey only
20% of the methods were effective. In this study, 55% of the
"other" methods reported post remediation radon levels less than 4
pCi/1. Although this may be impressive it should be emphasized that
these techniques included passive and forced ventilation which
usually cost the homeowner an additional heating expense.
Consequently, these methods are usually temporary measures and may
explain why they were only used in 18 of 473 remediations.
The performance of mitigation systems installed by certified
firms was also evaluated from data compiled from quarterly
mitigation reports submitted to the DEP by these same firms.
Mitigation firms participating in the DEP voluntary certification
program are required to report all radon mitigation work performed
in New Jersey on a quarterly basis. Pre and post remediation radon
levels, type cf mitigation design installed and the total cost of
mitigation system are some of the data reported. A total of 3,197
residential remediations were reported to the DEP from April, 1988
to October, 1989.
The percentage of homes reported as having post remediation
radon levels less than 4 pCi/1 was significantly greater than that
presented by the DEP. Certified firms reported effective
remediations in 96% of all mitigation methods attempted compared to
only 78% reported in DEP's post remediation data. This discrepancy
in the two data sets may be explained when considering the following
factors: various radon testing techniques were used by firms when
performing post remediation tests; firm and DEP post remediation
testing were often performed during different periods of the year;
-------
the possibility that firms might inaccurately report results, and
that firms may have not reported all post remediation radon levels
greater than 4 pCi/1 to the DEP.
Average Cost of Mitigation
The average cost of radon remediation charged to homeowners by
certified firms was also computed from the quarterly survey reports
submitted to the DEP. Table 3 compares the average costs of each
mitigation method in this study to the 1988 study. The average cost
of remediation for all mitigation methods was $1,100. This is
appreciatively less than the $1,300 reported in the 1988 study where
942 remediations were evaluated. The average cost of sealing only
methods in this study was $601 compared to $700 in the 1988 study.
Soil depressurization methods decreased by more than 10% to $1,120
in this study from $1,270 in the 1988 study. Air to air heat
exchangers decreased by about 20%. Their average cost in this study
was $1,620 compared to $2,000 per installation in the 1988 study.
Finally, remediation by "other" methods also cost less in this
study. The average cost was only $695 compared to $1,900 in the
1988 study.
Conclusions
The proportion of homes effectively remediated for elevated
indoor radon levels is increasing. Of the homes surveyed in this
study compared to the 1988 study, radon remediation effectiveness by
certified firms has increased from 48% to 78%. The proportion of
soil depressurization systems installed is about the same, however,
their effectiveness in reducing radon has increased significantly,
and this method has remained the most effective method in reducing
indoor radon levels below 4 pCi/1. Sealing only techniques were
chosen most often by homeowners who performed their own mitigation
work. Air to air heat exchangers were the least employed type of
mitigation performed in this study. This may be because of the high
costs of installation and maintenance along with their actual
ineffectiveness in lowering indoor radon levels.
All mitigation methods decreased by 15% in cost from the
previous year. The average cost of mitigation charged to New Jersey
residents by certified firms in this study was $1,100.
Although this survey is encouraging for both the radon industry
and New Jersey residents it also demonstrates the need for a program
of continuing education and training. State and federal training
programs will enable the radon industry to stay abreast of the
latest scientific and technological research as well as changes in
state and federal radon legislation. Support and participation in
such programs will enable the radon industry to become more
knowledgeable and effective in their remediation efforts and thus
offer the residents of New Jersey the quality workmanship they
deserve.
-------
This work described in this paper was not funded by the United
States Environmental Protection Agency, therefore, the contents do
not necessarily reflect the views of the agency and no endorsement
should be inferred.
Bibliography
1. USEPA, Radon Reduction Techniques for Detached Houses,
EPA/625/5086/019, 1986
2. USEPA, Application of Radon Reduction Methods,
EPA/625/5-88/024, August, 1988
3. DePierro, N. and Cahill, M. "Radon Reduction Efforts in New
Jersey", NJDEP, October, 1988
4. Cohen, S.A., "Results of a Pilot Survey of Radon Prevention and
Mitigation Firms," USEPA, Contract No. 68-02-4375, November,
1987
ndl!3089
-------
Table 1 Homeowner/Non-Certified Firm and DEP Certified Firm
Mitigation Installations in 473 Homes Surveyed in 1989
Sealing Soil Air to Air
Only Depressurization Exchanger Other Total
Homeowner/
Non-Certified
Firms
DEP Certified
Firms
Total
119
46
165[35]
110
166
276^58]
12
14 [3]
18[4]
246[52]
227[48]
473(100]
Bracketed numbers represent percent of total homes.
Table 2 Homeowner/Non-Certified Firm and DEP Certified Firm Mitigation
Installations in 716 Homes Surveys in 1988.
Sealing Soil Air to Air
Only Depressurization Exchanger Other Total
Homeowner/
Non-Certified
Firms
DEP Certified
Firms
Total
176
27
203[28]
95
326
421[59]
43
20
14 15
57[8] 35[5]
334[47]
382[53]
716[100]
Bracketed numbers represent percent of total homes.
ndl!3089
-------
Table 3^ Average Cost of Radon Mitigation ($)
Mitigation
Technicrue
Sealing
Soil Depressurization
Air to Air Exchangers
Other
Overall Average
1989 Study
(3197 Homes)
601
1120
1620
695
1100
1988 Study
(942 Homes)
730
1270
2000
1900
1300
ndl!3089
-------
FIGURE 1
Post-Remediation Radon Levels in Homes
Mitigated by Homeowners/Non-Certified Firms
and Certified Firms for All Methods
9OT
SO--
7O--
co 6O
CD
e
o
1C 50
-------
FIGURE 2
Post-Remediation Levels In Homes
Mitigated by DEP Certified Firms
for All Methods
9O-r
LEGEND
1988 Study
1989 Study
<4 4-8 8-20 20-5O 50-20O
Post Mitigation Radon Levels
pCi/l
-------
FIGURE 3
Effectiveness of Mitigation Measures
9OT
LEGEND
Owner/Non-Certified
Sealing Soil Depr.
Other
-------
FIGURE 4
Post-Remediation Radon Levels in Homes
Mitigated Using Sealing Only
(DEP Certified and Homeowner/Non-Certified)
9OT
80--
70--
to 60--
o>
E
o
m 50+
40--
o
a>
O_
3O--
20--
10--
LEGEND
1988 Study
1989 Study
<4 4-8 8-2O 20-5O 5O-2OO
Post Mitigation Radon Levels
pCi/l
-------
FIGURE 5
Post-Remediation Radon Levels in Homes
Mitigated Using Air to Air Heat Exchangers
(DEP Certified and Homeowner/Non-Certified)
LEGEND
1988 Study
1989 Study
<4 4-8 8-2O 20-50 50-2OO
Post Mitigation Radon Levels
pCi/l
-------
FIGURE 6
Post-Remediation Radon Levels in Homes
Mitigated Using Soil Depressurization
(DEP Certified and Homeowner/Non-Certified)
90 T
LEGEND
1988 Study
1989 Study
<4 4-8 8-2O 20-50 50-2OO
Post Mitigation Radon Levels
pCi/l
-------
FIGURE 7
Post-Remediation Radon Levels in Homes
Mitigated Using Other Methods
(DEP Certified and Homeowner/Non-Certified)
90 T
LEGEND
1988 Study
1989 Study
en
<4 4-8 8-20 20-50 5O-2OO
Post Mitigation Radon Levels
pCi/l
-------
VII-8
LONG-TERM PERFORMANCE AND DURABILITY OF ACTIVE RADON MITIGATION
SYSTEMS IN EASTERN PENNSYLVANIA HOUSES
by: A. G. Scott
A. Robertson
AMERICAN ATCON, INC.
Wilmington, DE 19899
ABSTRACT
Indoor radon reduction techniques were installed in 40 houses in the Reading
Prong region of eastern Pennsylvania over the period 1985 to 1987. Active soil
ventilation systems were installed in 36 of the houses; 3 had heat recovery
ventilators (air-to-air heat exchangers), and 2 installations included carbon
adsorption units to remove radon from well water. Follow-up measurements of the
post-mitigation radon concentrations in these houses were conducted with alpha-
track detectors during the winters of 1986/87, 1987/88, and 1988/89. The systems
were inspected and the owners interviewed during each installation and collection
visit.
In those houses where the radon reduction system was in continuous operation
during the monitoring periods, the measured average radon concentration in each
house compared well from year to year. High radon concentrations were measured
if the system did not operate during a monitoring period. Overall, no general
degradation in system performance was observed.
One manufacturer supplied 34 fans with the same motor and impeller assembly
which were used in the active soil ventilation systems. There have been five fan
failures in this population to date. The causes were capacitor failure in four
fans, and noisy bearings leading to shutdown in one fan. Assuming an exponential
lifetime distribution, the mean life of these fans (between repairs/replacements)
is calculated as 15 years, based upon this experience. There have been no
failures among three other fans from another manufacturer.
Of the three installed heat recovery ventilators, one fan bearing failed in
early 1989. This experience suggests a mean life between repairs of 8 years.
The performance of one charcoal adsorption unit has deteriorated from <95%
to 65% removal over 3 years. The second unit, which contains charcoal specially
selected for radon removal, has maintained a 97% removal efficiency over the same
period.
Five houses from this group of 39 houses (after 1 early dropout) have been
sold over the period of the study, suggesting a mean occupancy life of 29 years
in this study group.
This paper has been reviewed in accordance with the U. S. Environmental
Protection Agency's peer and administrative review policies, and has been
approved for presentation and publication.
-------
INTRODUCTION
Indoor radon reduction techniques were installed and tested in a total of 40
homes in communities on the Reading Prong in eastern Pennsylvania over the period
of June 1985 to June 1987 as part of EPA Contract 68-02-4203. The primary
mitigation method in 36 of these houses was active soil ventilation, with 3
houses receiving heat-recovery ventilators (air-to-air heat exchangers) and 1
house receiving just a radon-in-water removal unit. This project was reported
in detail in EPA-600/8-88-002 (Reference 1).
Follow-up alpha-track-detector (ATD) radon concentration measurements have
been made annually in 38 of the 40 demonstration homes during the heating seasons
of 1987/88 (Reference 2) and 1988/89 (Reference 3). Two of the houses were no
longer available for monitoring - one house had been removed from the original
site, and the owner of a second house has discontinued participation in the
project.
This paper examines the results in the 38 demonstration homes to make an
assessment of long-term system performance and equipment durability.
MEASUREMENT PROCEDURE
Average radon concentrations are measured by Terradex 'Track-Etch1 detectors
exposed for 4 months during the heating season. The forces that urge radon from
the soil into the home are highest during this time, and therefore this season
presents the greatest challenge to radon mitigation systems.
Quality assurance and control (QA/QC) measures were instituted to ensure that
the acquired data were of high quality. Unexposed detectors (blanks) were
included in each set sent to Terradex for analysis, and a number of detectors
from each batch were exposed to known concentrations in the radon chamber
operated by the U.S. Department of Energy's Environmental Measurements Laboratory
(EML) in New York. All these detectors were given fictitious house identifying
numbers, so that they were "blind" samples to the processor. A "zero
correction", equivalent to the mean reported exposure of the blanks, was
subtracted from each ATD result reported by Terradex. A "gain correction"
derived from the spiked samples (EML Exposure/Zero-Corrected Reported Terradex
Exposure) was used as a multiplier to adjust the reported house concentrations
to conformity with the EML.
As an additional QA/QC measure, detectors were exposed in groups of three
(1986/87 and 1987/88) or two (1988/89) at each measurement location, and the
results averaged to reduce the measurement uncertainties.
The demonstration houses were visited during December of each year to install
Terradex "Type SF" Track-Etch detectors in accordance with the guidance given
in the EPA Measurements Protocol (Reference 4). Detectors were placed in both
the main living area and the basement, hung together in groups of three (two in
1988/89) from an interior wall or ceiling in the living area and a central joist
in the basement. Each detector was marked with the installation date and the
-------
house identification code. That information, plus the detector numbers and
their locations, were recorded in a Track-Etch Record Book kept specifically for
that purpose. During these visits the soil ventilation fans and the air-to-air
heat exchangers were checked to see that they were operating. The houses were
visited about 16 weeks later, at the end of the heating season, to remove the
detectors. An ATD retrieval rate of 100% was achieved. The homeowners were
interviewed to discover any events that might have affected the system
performance during the monitoring period. Retrieved detectors were placed in
pairs into the manufacturer's envelopes, which were then sealed by folding and
taping in a low-radon atmosphere. The envelopes were returned to Terradex in
one batch.
QA/QC RESULTS
The 1986/87 and 1987/88 blanks did not indicate a need for zero correction.
The results from the QA/QC blanks in 1988/89 led to a zero correction of 10 Bq/m3
(0.26
The detectors exposed at EML had known exposures which were always greater
than the reported exposures, and so gain corrections were applied of 1.16 for
1986/87; 1.30 for 1987/88; and 1.21 for 1988/89.
• One outlier was identified in a group of three detectors in the 1986/87
reported results and was discarded. No outlier was discovered in the 1987/88
results. In 1988/89 there were two probable outliers; these were of limited
significance, since the measurements were both from houses where the mitigation
system did not operate during the entire measurement period.
RESULTS
RADON REDUCTION PERFORMANCE
Table I shows the results by year for the 27 houses where the mitigation
systems operated for the entire 3-year period and were monitored each winter,
thus providing a fair measure of long-term mitigation system performance.
Excluded from the table are the results from five houses where the system
equipment failed during the 3 years, three houses in which the system was turned
off during the monitoring period, three houses with measurements only in 1987
and 1988, and one house where the system was incomplete because the owner refused
further work.
The average radon concentration in the basement of these 27 houses was 4.7,
4.7, and 4.9 pCi/* in 1987, 1988, and 1989, respectively; in the living area,
the average was 4.0, 3.9, and 3.9 pCi/l. The small differences in the group
average from year to year are without statistical significance. The constancy
of the group average with time shows that there has been no general deterioration
of system effectiveness in 3 years of operation.
-------
Table I.
ALPHA-TRACK MONITORING RESULTS
POST-MITIGATION RADON RESULTS foCi/i)
ID
NO.
2
3
5
6
8
9
10
12
13
16
18
20
21
22
24
25
26
27
28
29
30
32
33
34
35
36
37
PREMIT
1989
(oCi/jn B*
413
350
110
60
183
533
626
11
64
395
12
210
172
24
66
122
89
21
21
61
17
6
82
470
144
300
87
Average - 172
5
3
5
3
2
12
10
1
2
4
12
8
1
10
4
7
0
5
3
2
3
0
11
5
2
0
0
4
%RDEV* - 104
.5
.0
.0
.2
.9
.4
.4
.6
.7
.8
.7
.3
.9
.8
.3
.2
.6
.6
.6
.1
.9
.5
.2
.1
.4
.8
.9
.9
76
LA*
8.7
1.9
4.4
2.7
1.0
17.1
8.9
2.1
2.8
1.1
5.1
9.3
2.7
4.0
3.7
5.3
1.1
2.1
5.1
2.3
2.1
3.2
0.7
5.5
1.0
0.7
0.5
3.9
94
B
4.8
3.5
5.0
4.1
3.5
10.4
15.2
2.2
2.6
5.7
13.5
6.5
2.0
8.6
3.6
7.7
1.1
4.0
4.1
1.6
4.0
1.2
3.5
5.4
1.0
1.1
1.2
4.7
78
B=Basement ; LA-Living Area ;
Average ;
%RDEV-Relative
Conversion factor:
Bq/m3 -
1988
LA
6.7
2.3
4.4
3.2
1.5
12.9
9.9
2.2
3.9
2.5
3.4
10.0
2.7
4.4
3.8
6.0
1.6
2.2
4.4
2.0
1.6
4.4
1.2
5.5
0.9
1.0
0.7
3.9
78
1987
B
3.0
4.1
5.0
3.8
4.5
13.5
10.4
4.3
2.7
6.3
10.2
6.7
3.6
8.8
5.0
6.3
2.4
4.4
2.8
2.2
3.5
1.2
2.6
6.4
0.9
1.9
0.7
4.7
66
BAVE-Basement
Standard Deviation
37 x value in
pd/Ji
LA
6.0
2.4
5.0
5.7
2.1
16.8
7.5
2.9
2.3
2.0
2.4
11.5
3.0
3.1
5.3
3.5
1.7
2.6
6.1
1.6
1.5
3.7
1.3
4.3
0.8
0.8
2.0
4.0
87
Average;
(in %).
•
SAVE*
4.4
3.5
5.0
3.7
3.6
12.1
12.0
2.7
2.7
5.6
12.1
7.2
2.5
9.4
4.3
7.1
1.4
4.7
3.5
2.0
3.8
1.0
5.8
5.6
1.4
1.3
0.9
4.8
70
LAVE'
7
2
4
3
1
15
8
2
3
1
3
10
2
3
4
4
1
2
5
2
1
3
1
5
0
0
1
3
.1
.2
.6
.9
.5
.6
.8
.4
.0
.9
.6
.3
.8
.8
.3
.9
.5
.3
.2
.0
.7
.8
.1
.1
.9
.8
.1
.9
84
LAVE-Living Area
Table I shows the average radon concentrations in the basement and the living
area of the individual houses. In general, the average concentration does not
vary greatly from year to year in a given house, but House 33 is an exception.
The radon concentration in 1989 in the basement increased by over 300% while
concentrations in the living area remained similar to previous years.
A number of houses (2, 9, 20, 28, and 32) have higher radon concentrations
upstairs than in the basement. This is more clearly seen in Table II, where the
ratios of radon concentration in the living area to that in the basement are
-------
Table II. RATIO OF RADON CONCENTRATIONS IN LIVING AREAS TO BASEMENTS
ID
No.
2
3
5
6
8
9
10
12
13
16
18
20
21
22
24
25
26
27
28
29
30
32
33
34
35
36
37
Average-
%RDEV*=
* %RDEV = Relative
LIVING AREA/BASEMENT RADON CONCENTRATION
1987
2.00
0.59
1.00
1.50
0.47
1.24
0.72
0.67
0.85
0.32
0.24
1.72
0.83
0.35
1.06
0.56
0.71
0.59
2.18
0.73
0.43
3.08
0.50
0.67
0.89
0.42
2.86
1.01
75
standard
1988
1.40
0.66
0.88
0.78
0.43
1.24
0.65
1.00
1.50
0.44
0.25
1.54
1.35
0.51
1.06
0.78
1.45
0.55
1.07
1.25
0.40
3.67
0.34
1.02
0.90
0.91
0.58
0.99
67
deviation (in %).
1989
1.58
0.63
0.88
0.84
0.34
1.38
0.86
1.31
1.04
0.23
0.40
1.12
1.42
0.37
0.86
0.74
1.83
0.38
1.42
1.10
0.54
6.40
0.06
1.08
0.42
0.88
0.56
1.06
109
RATIOS
Averaee
1.66
0.63
0.92
1.04
0.41
1.29
0.74
1.00
1.13
0.33
0.30
1.46
1.20
0.41
0.99
0.69
1.33
0.51
1.56
1.02
0.46
4.38
0.30
0.92
0.74
0.74
1.33
1.02
77
shown for individual houses, over a 3 year period. The ratios are consistently
above unity in those five houses. Houses 9, 20, and 32 are known to have radon
levels in well water above 20,000 pCi/X, which could be contributing
preferentially, through water-usage pattern, to higher concentrations upstairs.
The concentration of radon in water at House 28, which has a well, is not known.
In addition, Houses 2, 9, and 28 have block fireplace structures which could
provide a direct soil-gas entry route upstairs without entering the basement.
Houses 2 and 9 also have block wall pressurization systems which could force
radon-laden air upstairs. House 20 has an adjoining paved crawl space which may
not be fully treated. Although these are perhaps special cases, there are enough
of them to indicate that basement measurements do not always provide a
conservative estimate of the radon exposure to the occupants.
-------
INOPERATIVE SYSTEMS
Some mitigation systems were inoperative during the monitoring periods, and
are listed in Table III, where the post-mitigation radon results obtained during
all monitoring periods are provided along with brief comments as to why the
systems were not functioning. House 1, which was demolished and moved from its
site before post-mitigation monitoring could begin, and House 11, where the owner
refused further participation, have been included for completeness.
Table III. SYSTEMS INOPERATIVE DURING ONE OF THE ATD MONITORING PERIODS
POST -MITIGATION RADON RESULTS foCi/ii
ID
NO.
i
2
4
7
11
14
15
17
31
39
PREMIT.
(oCi/i)
U6
413
25
402
49
36
18
9
485
111
1989 1988 1987 ADDITIONAL
B** LA** B LA B
House moved from site
5.5 8.7 4.8* 6.7* 3.0
1.2 1.0 7.3* 3.1* 0.8
103.1* 24.1* 4.9 3.8 4.8
Owner no longer participating
10.8* 8.0* 1.1 1.4 0.6
1.3 1.3 19.7* 11.0* 1.3
8.5* 4.9* 8.2 6.4 8.8
252.7* 93.3* 2.8 8.3 2.1
7.5 1.8 46.1* 17.5*
* Asterisk marks monitoring period system was
B - Basement; LA - Living Area. *" -- -
LA COMMENTS
6.0 1. Capacitor failure;
repaired prior to
monitoring.
2 . Fan turned off by
homeowner.
0.9 Capacitor failure.
3.2 Capacitor failure.
0.8 Fan turned off.
1.2 Bearing failure.
4.8 HRV bearing failure .
6.6 Capacitor failure.
..*** Fan turned off.
inoperative .
No measurement during 1987.
In three houses, the soil ventilation fan was turned off by the homeowner,
either deliberately or inadvertently. Reportedly, the fans at House 2 were
turned off for 12 days during the 1988 measurement period while the family was
away on vacation. However no increase in radon concentration was observed,
indicating that the concentrations did not return promptly to the very high pre-
mitigation levels. The results were similar to 1989 when the owners took care
to have the fans running constantly throughout the monitoring period. Perhaps
-------
only one of the two fans involved in this system was turned off. Likewise, at
House 39, the system fan was turned off for an extended period while the owners
were away. At House 14, the owner unintentionally disconnected the fan from the
electrical outlet for some time during the 1989 monitoring period.
Besides the above three examples of soil ventilation systems being rendered
inoperative due to owner intervention, there have been five instances of soil
ventilation systems not working due to mechanical failure of the system fans
- Houses 2, 4, 7, 15, and 31. The first failure occurred at House 2, and was
corrected before the start of the 1987 monitoring season, the second two during
the 1988 monitoring period (Houses 4 and 15), and two more during the 1989 period
(Houses 7 and 31). In addition to these five mechanical failures in soil
ventilation systems, there was one failure of a HRV fan (House 17).
When the fans did not operate during monitoring periods in these houses, the
average radon concentrations were high, as would be expected. Levels at Houses
2, 4 and 15 returned to previous post-mitigation values after system repairs.
The systems have been repaired at Houses 7 and 31, and await further monitoring.
FAILURE EXPERIENCE WITH SOIL VENTILATION FANS
Three different types of soil ventilation fans, from two different manufac-
turers, were used in the active soil ventilation systems installed in the
demonstration houses in Pennsylvania. One type, from the first manufacturer,
was installed in three of the houses; there have been no failures to date among
these three fans.
Two types of fans, from the second manufacturer, use the same motor and
impeller assembly in different housings. A total of 34 fans from this
manufacturer have been installed in 31 houses as part of the Pennsylvania
project. These 34 fans have suffered 5 failures (Houses 2, 4, 7, 15, and 31)
since installation 2 to 4 years ago. The number of fans is greater than the
number of houses because three houses each had two fans in their soil ventilation
systems.
Four of the fan failures were due to failure of the electrolytic capacitor
used in the split-phase electric motor. When the capacitor fails, the fan cannot
be started. If the fan is operating when the capacitor fails, the fan will
continue to operate for some time, but at significantly reduced performance.
One fan, at House 15, developed noisy bearings and was turned off by the
owner while awaiting replacement.
In all instances the fans were repaired or replaced under warranty and are
functioning properly once more.
If we assume an exponential probability distribution for fan lifetimes, we
can write the probability for a failure after time, t, as
P(t)dt - A exp(-At)dt
-------
where A is the average failure rate. The mean lifetime between events requiring
fan repair or replacement is given by T-l/A. If we take experience to date as
extending from the time of installation until the end of the latest monitoring
period (April 1989), then the accumulated experience is 897 fan-months or 75 fan-
years. This sets the average failure rate, A, at 5 fan failures in 75 fan-years,
and therefore indicates a mean lifetime, T, between repairs/replacements of 75/5
= 15 years.
This lifetime can be used to estimate the number of fan failures to be
expected year by year, by integrating the probability above over time. The
expectation value for the number of fans, N, failed by the end of time, t, is
given by
= N(0)[l-exp(-t/T)],
where T is 15 years and N(0) stands for the number of fans installed, 34 in this
instance. The expected number of failures can then be calculated as 2.2 after
1 year, 4.2 after 2 years, 6.2 after 3 years, and 8.0 after 4 years. These are
in keeping with our findings to date.
HRV EXPERIENCE
The system breakdown at House 17 is the first failure of an HRV unit
installed at three of the demonstration houses. The owner reported that the
bearings seized on the motor which drives both the intake and exhaust fans, and
the rotary heat exchange wheel, in this HRV design. The homeowner suspects that
the fan shaft may have been bent by the contractor during the original
installation, potentially causing the eventual deterioration of the bearings.
The fan has since been replaced by the vendor, and the unit is now operating.
To date, then, our experience with HRV units indicates 1 failure requiring
repair in 100 unit-months of operation. Although this failure apparently
resulted from installation problems rather than from normal long-term wear, the
installation problem is felt to be representative of the factors which can
influence system lifetime; thus, this failure can be considered in the
calculation of equipment lifetime between repairs. If one assumes an exponential
distribution of lifetimes to failure, the mean lifetime, between required
repairs, can be estimated as 100/1-100 months or 8 years. The expected number
of failures after 3 years of operation would be 0.94 and after 9 years that would
have grown to 2.0 failures.
WELL WATER TREATMENT BY CHARCOAL ADSORPTION
Adsorption on activated charcoal was chosen for removal of radon from well
water at House 2, where it supplemented an active soil-ventilation mitigation
system, and at House 30, where it was the sole means of reducing household radon.
Two different suppliers for charcoal units were used. The first unit came from
a local water purification supplier and contained charcoal of unknown origin,
probably selected for organics removal. This was installed in House 2, where
radon concentrations in the water were only moderately elevated at >200,000 Bq/m3
(>50,000 pCi/i). The initial radon removal efficiency was about 95%.
-------
This system was checked in January 1988, and removal efficiency was 79%. By
December 1988, the removal efficiency was 65%. This steady deterioration in
performance is not altogether surprising, since the charcoal used was not
specially selected for radon removal from water.
The second unit came from a firm in Maine which specialized in water
treatment systems for radon removal. This unit contained coconut charcoal
specifically selected to remove radon. This was installed in House 30 where
radon concentrations in water were high - about 800,000 Bq/m3 (200,000 pCi/i).
The initial removal efficiency ranged from 95% to 99%.
This second system was checked in 1987, and the removal efficiency was 99%
at that time. In December 1988, the removal efficiency was 97%. This is
consistent with a constant removal efficiency. There is no degradation in
removal efficiency evident with this unit despite the high input concentration
and the high water usage associated with a family with two young children.
Perhaps this is attributable to the specially selected charcoal.
TURNOVER OF HOUSING STOCK
During the period of this study, starting from the initial selection of
houses in April 1985 through the latest monitoring period in April 1989, 5
homeowners out of total of 39 have moved. (The 40th homeowner dropped out of
the program within the first year, and is thus not included in this calculation.)
Since the total occupancy to date is 146 owner-years, the average rate of moving
during the study is 5/146 per year. Assuming an exponential distribution of
occupancy times, the mean occupancy life, T, is 146/5 - 29 years.
CONCLUSIONS
For the 27 houses in which the mitigation systems operated continually during
all three monitoring periods from 1986 to 1989, the post-mitigation ATD results
did not change greatly from year to year. Only at House 33 was there a
significant increase in radon concentration in 1989, and then only in the
basement. There does not appear to be any general degradation in mitigation
system performance with time alone.
There are 34 fans from one manufacturer used in the soil ventilation systems,
and 5 failures (requiring repair or fan replacement) have occurred over the past
2 to 4 years. Four of these failures have been due to capacitor failure, and
one due to bearing noise. Based upon this experience, these fans are estimated
to have a mean lifetime of 15 years before repair/replacement.
There are three HRV units installed, and one of these units had a bearing
seizure requiring repair after 3 years. The estimated mean HRV lifetime between
such repairs, based upon this experience, is 8 years.
There are two charcoal adsorption units for removal of radon from well water.
The removal efficiency of one has fallen from 95% to 65% over a period of 28
-------
months from installation. The removal efficiency of the other unit, containing
specially chosen charcoal, has remained constant at 97% over the same period.
There are 39 homeowners in the study group, and 5 of them have moved over the
period. A mean occupancy period of 29 years is estimated.
REFERENCES
1. Scott A.G., Robertson A., and Findlay W.O., "Installation and Testing of
Indoor Radon Reduction Techniques in 40 Eastern Pennsylvania Houses," report
prepared for U.S. Environmental Protection Agency by American ATCON,
EPA-600/8-88-002 (NTIS PB88-156617), Research Triangle Park, NC, January
1988.
2. Scott A.G. and Robertson A. , "Follow-up Alpha-Track Monitoring in 40 Eastern
Pennsylvania Houses with Indoor Radon Reduction Systems (Winter 1987-88),"
report prepared for U.S. Environmental Protection Agency by American ATCON,
.EPA-600/8-88-098 (NTIS PB89-110035), Research Triangle Park, NC, September
1988.
3. Scott A.G. and Robertson A. , "Follow-up Alpha-Track Monitoring in 40 Eastern
Pennsylvania Houses with Indoor Radon Reduction Systems (Winter 1988-89),"
report prepared for U. S. Environmental Protection Agency by American ATCON,
EPA-600/8-89-083 (in press), Research Triangle Park, NC, October 1989.
4. U.S. Environmental Protection Agency, "Interim Indoor Radon and Radon Decay
Product Measurement Protocols," EPA-520/1-86-04 (NTIS PB86-215258),
Washington, D.C., February 1986.
-------
Session D-VII:
Radon Reduction Methods—POSTERS
-------
D-VII-1
ENERGY PENALTIES ASSOCIATED WITH THE USE OF A SUB-SLAB
DEPRESSURIZATION SYSTEM
by : Mike Clarkin
Terry Brennan
Camroden Associates, Inc.
Oriskany, NY 13424
Michael c. Osborne
USEPA/AEERL
Research Triangle Park, NC 27711
ABSTRACT
One of the primary radon mitigation techniques used to reduce
indoor radon concentrations in houses is a sub-slab
depressurization system. In this type of system, a fan removes soil
gases containing radon from beneath the floor slab and exhausts the
gases to the outdoors by creating a pressure field beneath the slab
that is negative relative to the basement air pressure. Because of
this negative pressure, indoor conditioned air can be drawn through
the floor penetrations and exhausted outdoors. In order to
determine the amount of conditioned air that is being lost, a
series of experiments utilizing tracer gases were performed in
three houses. This paper presents the results of these experiments.
INTRODUCTION
One of the most accurate ways of determining a building's air
exchange rate is through the use of a tracer gas. A gas is injected
into the building's interior. Specialized eguipment is used to
monitor the concentration of the tracer gas. The rate of decay of
the tracer gas can then be used to determine the building's air
exchange rate. This procedure is commonly utilized to determine the
building's overall air exchange rate. For the purpose of
determining the amount of conditioned air being lost due to the
This paper has been reviewed in accordance with the U. S.
Environmental Protection Agency peer and administrative review
policies and approved for presentation and publication.
-------
sub-slab depressurization system, the methods normally used were
modified because the method described above would include the air
lost from the basement to the superstructure of the house through
thermal bypasses and other penetrations, and the air lost to the
outdoors through above grade basement penetrations. What was
required was a method that would disclose the amount of air that
the sub-slab depressurization system was actually drawing out of
the basement through the below grade penetrations due to the
negative pressure field being developed by the sub-slab
depressurization system.
GENERAL PROCEDURES
A series of tests utilizing sulfur hexafluoride (SF6) as a
tracer gas and Miran 101 specific gas analyzers to monitor the
tracer gas concentrations were performed in each house. One
.analyzer would monitor the basement SF( concentration, and one
^analyzer would monitor the SF6 concentration within the sub-slab
depressurization system exhaust pipe. Figure 1 is a simplified
schematic of the equipment setup. An amount of tracer gas was
injected into the basement. The gas and basement air were mixed
using several small fans. The tracer gas was continuously seeded
until the basement and the exhaust pipe concentrations were stable.
System airflow rates were measured, and the ratio of basement to
exhaust pipe SF6 concentrations at that airflow rate was then used
to determine the amount of indoor conditioned air that was being
lost due to the sub-slab depressurization system.
-------
Some of the air being exhausted by the sub-slab
depressurization system is being drawn through
penetrations by the negative pressure field being
developed. The rest of the exhausted air is coming
from the soil or bedrock, the amount of air drawn from
each source depends on how easily air can be drawn
through the foundation penetrations and through the
soil.
One analyzer monitors basement SF6 concentrations and the other
monitors the exhaust pipe concentration. A constant concentration
is maintained in the basement. The ratio of the two concentrations
reveals the amount of basement air being exhausted by the sub-slab
depressurization system.
Figure l. Equipment layout of tracer gas experiments.
-------
Figure 2 illustrates the data gathered during a typical test.
For this experiment, the sub-slab system was in the active mode.
Data were collected for 5 minutes prior to injecting the SF, as a
check on the stability of the analyzers. At the end of the first
5 minute period, SF4 was seeded and the injection rate was adjusted
until a stable concentration in the basement and exhaust pipe was
obtained.
50
— 40-,
30-
20 H
10 -I
Basement concentration
Pipe concentration
started tracer gas seed
10
—I-
20
30 40
ELAPSED TIME (min.)
50
60
70
Figure 2. Tracer gas experiment performed in House VA2 to
determine the energy penalty associated with the use of a sub-slab
depressurization system.
Figure 3 illustrates a test performed at house VA2 to
determine the effect that penetrations through the slab have on
the sub-slab system. For this test the sub-slab system was in the
active mode. SF, was continuously injected into the basement
throughout the entire monitoring period. For the first 45 minutes
of the test, the floor slab was left intact. At time 45, a 1/2-
in.* diameter (0.2 sq in.) hole was drilled through the
* Readers more familiar with metric units may use the conversion
factors at the end of this paper.
-------
floor slab, and the SFt concentration in the exhaust pipe and
basement air allowed to stabilize. At time 85, another 1/2 - in.
diameter hole was drilled in the floor slab and the SFC
concentrations again allowed to stabilize. At time 110 a 3.1 sq
in. hole was opened in the sump cover, and the concentration
stabilized. For the final test hole, at time 150, the 3.1 sg in.
hole in the sump cover was reduced to 1 sg in., and again, the SFC
concentration allowed to stabilize.
a
z
o
H
W
8
«
CO
I 4 sq in.
slab leakage
area
3 5 sq in
slab leakage
area
0 4 sq in.
slab leakage
area
60 90
ELAPSED TIME (min.)
120
150
160
Figure 3. Tracer gas experiment performed in House VA2 to determine
the effect slab penetrations have on the sub-slab depressurization
system.
The same type of experiment was performed in House PA1. SF, was
injected into the basement . A number of penetrations were made in
the slab. The ratio of the basement to exhaust pipe SF6
concentration, along with the volume of air being exhausted by the
sub-slab system, can be used to determine the additional energy
-------
costs incurred by the use of a sub-slab depressurization system.
The data collected during this experiment are illustrated in Figure
4.
0 2 sq in.
slab
leakage
area
0 3 sq in
slab
leakage
area
01 sq in.
slab
leakage
area
Basement concentration
Pipe concentration
30
60
90 120 150
ELAPSED TIME (min.)
Figure 4. Tracer gas tests performed in House PA1 to determine the
effect slab penetrations have on the performance of a sub-slab
depressurization system.
RESULTS
The results of the tests for the energy losses performed
in the three houses are presented in Table l. The estimated energy
costs were determined by using the degree day method. A furnace
efficiency of 90% was used for Houses VA1 and VA2 and 75% for House
PA1.
-------
TABLE 1. APPROXIMATE ADDITIONAL ENERGY COSTS DUE TO LOSS OF CONDITIONED AIR
THROUGH THE USE OF A SUB-SLAB DEPRESSURIZATION SYSTEM
HOUSE
NO.
VA1
VA2
PA1
% CONDITIONED
AIR BEING EXHAUSTED
BY SYSTEM
43
47
21
AIR VOLUME
LOSS (cfm)
34
11
3
HEATING
DEGREE
DAYS
5010
5010
5827
EST. ADDITIONAL
ENERGY USAGE
(10s Btu)
4.9
1.6
0.6
The results of the tests performed to determine the effect barrier
penetrations have on the performance of the sub-slab system in House VA2 are
presented in Table 2.
TABLE 2. SYSTEM PERFORMANCE AS EFFECTED BY FLOOR PENETRATIONS IN HOUSE VA2.
LEAKAGE AREA
(sq in.)
AIR VOLUME AT
EXHAUST (cfm)
PRESSURE DIFFERENTIAL
PIPE TO BASEMENT (in. H,0)
SLAB INTACT
0.8
1.6
2.6
4.7
24
25
27
30
30
0.212 ±
0.194 ±
0.180 ±
0.115 ±
0.071 ±
0.005
0.004
0.005
0.004
0.005
-------
Equation 1 is used to determine the fraction of basement air being
exhausted by the sub-slab depressurization system:
Cs = f(Cb
+ fCa x
(1)
Qt
where:
Cs = SF6 concentration in the exhaust pipe stack
Cb = SF6 concentration in the basement
Qbs = Volume of air flowing from the basement to the
stack
Ca = SF« concentration in the outdoor air getting back
inside the stack
Qas = Volume of outdoor air getting back inside the
stack
Qt = Total volume of air being exhausted by the system
Factoring this equation results in Equation 2 where:
Qt x Cs = fCb x Obs) + (Ca x Oasl
Cb Cb
(2)
Assuming that Ca is zero, which is not quite true but close enough
considering the miniscule amount of SF6 being exhausted compared to
the volume of outdoor air, and factoring again results in Equation
3:
££
Cb
(3)
Qt
Equation 3 translates to: the ratio of the SF« concentration in the
stack to the concentration in the basement equals the ratio of the
volume of air flowing from the basement into the stack to the total
airflow out of the stack. Therefore, if the SF« concentration in the
stack is 31 ppm and the SF, concentration in the basement is 72 ppm,
applying Equation 3 results in a value of 0.43 or 43% of the air
being exhausted by the sub-slab depressurization system is coming
from the basement.
Equation 4, used for determining the additional cost of energy
associated with the use of the sub-slab depressurization system,
-------
is known as the degree day method where:
Ht = V X 60 X 24 X 0.018 X HDD X Eff (4)
where:
HL = Heat loss due to exfiltration of air by
sub-slab system for the heating season
V = Volume of air being exhausted by system
in cubic feet per minute
60 = Number of minutes in an hour
24 = Number of hours in a day
0.018 = Heat capacity of air in British thermal
units per cubic foot per degree
fahrenheit. This means the number of
British thermal units required to heat
each 1 cu ft of air 1°F.
HDD = Heating degree days. This is the average
daily outdoor temperature in degrees
fahrenheit, subtracted from the indoor
temperature setpoint of 65 °F. This daily
temperature difference is totalled for the
entire heating season. HDDs should be
thought of as the number of degrees
fahrenheit for the heating season that
outdoor air entering the house must be
raised to reach the indoor temperature
of 65 °F.
Eff = Furnace efficiency. In this case 90%
Therefore, if the sub-slab depressurization system is exhausting
basement air at a volume of 34 cu ft/min., as in House VA1, the
volume of air being exhausted each day is 48,960 cu ft. To find the
number of British thermal units required to heat the exhausted air
1°F, multiply the volume of air by the heat capacity of 1 cu ft of
air. This results in a heat capacity of 881.3 Btu for each 1°F that
the air must be heated. Finally, by multiplying the heat capacity
of the exhausted air by the number of HDDs per heating season at
the location of the house (for this house, 5010 HDD) and
considering the efficiency of the furnace, the added amount of
energy required to heat the air being exhausted by the sub-slab
depressurization system is 4.9 million Btu.
-------
Table 3 presents the amount of energy lost in House VA2 with the various
size holes in the floor slab. Again, a heating degree day value of 5010 HDD
and a furnace efficiency of 90% are used.
TABLE 3. APPROXIMATE ADDITIONAL ENERGY COSTS DUE TO LOSS OF CONDITIONED AIR
IN HOUSE VA2 WITH VARIOUS SIZE FLOOR SLAB PENETRATIONS.
LEAKAGE
AREA
(sq in.)
SLAB INTACT
0.8
1.6
2.6
4.7
% CONDITIONED
AIR BEING EXHAUSTED
BY SYSTEM
47
52
57
75
77
AIR VOLUME
LOSS (Cfm)
11
13
15
23
23
EST. ADDITIONAL
ENERGY USAGE
(10* Btu)
1.6
1.9
2.1
3.3
3.3
The results of the tests performed to determine the effect barrier
penetrations have on the performance of the sub-slab system in House PA1 are
presented in Table 4.
TABLE 4. SUB-SLAB DEPRESSURIZATION SYSTEM PERFORMANCE AS AFFECTED BY FLOOR
PENETRATIONS IN HOUSE PA1
LEAKAGE AREA
(sq in.)
SLAB INTACT
0.4
0.8
1.2
AIR VOLUME AT
EXHAUST (cfm)
15
16
19
20
PRESSURE DIFFERENTIAL
PIPE TO BASEMENT (in. H,0)
0.80 ±
0.81 ±
0.78 ±
0.77 ±
0.01
0.01
0.02
0.01
-------
Table 5 presents the amount of energy lost in House PA1 with the various
size holes in the floor slab. Again, a heating degree day value of 5827 HDD
and a furnace efficiency of 75% are used
TABLE 5. APPROXIMATE ADDITIONAL ENERGY COSTS DUE TO LOSS OF CONDITIONED AIR
IN HOUSE PA1 WITH VARIOUS SIZE FLOOR SLAB PENETRATIONS
LEAKAGE
AREA
(sq in.)
SLAB INTACT
0.4
0.8
1.2
% CONDITIONED
AIR BEING EXHAUSTED
BY SYSTEM
21
29
36
41
AIR VOLUME
LOSS (Cfm)
3
5
7
8
EST. ADDITIONAL
ENERGY USAGE
(10* Btu)
0.6
0.9
1.3
1.5
CONCLUSIONS
Although a large percentage of the air being exhausted by the sub-slab
depressurization system can come from inside the building and not from
beneath the slab, the amount of air that this percentage constitutes is
relatively small. The estimated annual additional costs associated with the
use of the sub-slab depressurization system for the three houses tested (not
including electrical costs to run the fan) are illustrated in Figure 5. For
the dollar costs, a value of $0.65 per therm (10* Btu) of gas is used.
The effect on the performance of the sub-slab system of even such small
holes in the floor slab that were tested here seems to be a greater concern
than the additional costs of energy. For example, at House VA2, a leakage
area of 4.7 sq in. resulted in a drop in the negative pressure at the exhaust
pipe of 67%. A loss of this much pressure could have serious effects on a
marginally operating system.
METRIC CONVERSION FACTORS
Readers more familiar with metric units may use the following to convert
to that system:
Nonmetric Multiply by Yields Metric
Btu 1.055 kJ
°F 5/9 (°F-32) *C
ft1 28.3 L
in. 2.54 cm
in. H,0 6.45 cm'
in. H,0 2.5 Pa
therm 105,506 kJ
11
-------
40
30 -
! 1
s »
_; H
$31.85
VA1
VA2
HOUSE NO.
PA1
Figure 5. Estimated additional annual energy costs associated with
the use of a sub-slab depressurization system. Electrical
consumption of the fan in each system is not included.
12
-------
D-VII-2
EXHAUST OF CONDITIONED AIR BY RADON MITIGATION SYSTEMS IN HOMES
by: R. J. Saultz
Oak Ridge National Labs
Oak Ridge, Tennessee
WITHDRAWN BY AUTHOR
-------
D-VII-3
LONG TERM DURABILITY AND PERFORMANCE OF RADON MITIGATION SUBS1AB
DEPRESSURIZATION SYSTEMS
David T. Harrje and Kenneth J. Gadsby
Center for Energy and Environmental Studies
Princeton University, Princeton, NJ 08544
David C. Sanchez, Air and Energy Engineering Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
ABSTRACT
Subslab depressurization (SSD) systems have become the most popular
radon mitigation technique used by professional mitigators. Readily
available supplies and equipment from other trades are commonly utilized in
the installation of these systems. What is not currently known is the long-
term performance of some of the caulks or sealants and the mechanical
devices such as fans or controls. The operating environment for the fans is
often different from that for which they were designed. Specifically, this
involves studying whether the fans are still operating as installed, whether
there has been an increase or decrease in the radon levels after several
years of operation, and whether the house occupants are operating the
systems the way the systems were originally designed.
To answer these questions, we have tested eight of the Piedmont Study
houses as veil as houses in which post-mitigation measurements by the New
Jersey Department of Environmental Protection (NJDEP) have shown readings
that were above the U.S. EPA guideline of 4 pCi/L. Quarterly visits to each
house involved measurement of flows and pressures in the mitigation system,
taking radon grab samples in the basement and of the mitigation system
exhaust, interviews with the occupants, retrieving and installing track-etch
detectors, and a visual inspection of the mitigation system and house
substructure. This paper presents the findings of this research and
discusses some of the durability aspects of SSD systems and their
interaction with the house substructure.
This paper has been reviewed in accordance with the U.S. EPA peer and
administrative review policies and approved for presentation and publica-
tion. Mention of trade names does not constitute endorsement by the Agency.
-------
INTRODUCTION AND BACKGROUND
There is increasing evidence that the health risks in those houses
with significant levels of radon gas (above the EPA guideline of 4 pCi/L)
may constitute the most serious indoor air quality problem in the United
States. Radon gas intrusion is often pictured as a seasonal phenomenon,
with stack effect and other pressure-driven factors influencing soil gas
entry to building substructures. A number of solutions have been proposed.
These approaches involve energy use as well as indoor air quality concerns.
The proposed solutions must be tailored for the specific nature of the radon
source. If the radon enters the house via veil water, one approach is
necessary; radon in building materials may suggest other strategies. In
this report our attention will be focused on radon entry with soil gas
through the building substructure, and what that mechanism implies for radon
mitigation.
Even limiting our scope to soil gas entry, many solutions exist to
reducing radon gas concentrations in the house. Radon removal is one method
and is often the preferred approach. Local exhaust, after radon entry into
the living space, is another strategy. Ventilation used to dilute the radon
gas concentrations is still another option. Each approach must be matched
to a given radon condition in the individual building. This report will
consider only the method of radon removal from under the slab in basements
or crawl spaces. Known as subslab depressurization, SSD, this mitigation
approach has been proven to be very effective, often removing enough soil
gas containing radon so that substructure measurements of radon
concentrations drop by 90% or more following mitigation.
1. Theory of Operation of SSD Systems for Radon Mitigation
The theory of operation for the SSD system is that, by penetrating the
concrete floor slab with an exhaust pipe, one gains access to the area
beneath the slab. The area, often a gravel bed, serves as a collection site
for the soil gas containing radon. The negative pressure provided by the
exhaust pipe causes the soil gas to be removed from the gravel bed and, if
communication exists between the subslab volume and the walls of the
building, soil gas will simultaneously be exhausted from the walls. The
exhaust mechanism can be passive which implies that suction pressures
beneath the slab will vary on a seasonal basis, with the greatest suction
occurring during the coldest weather due to increased buoyancy of the air in
the vertical exhaust stack. In the systems tested in this research, exhaust
fans were used. These "active systems" were shown to maintain near constant
suction pressures under the slabs all year long.
The key point to remember in the merits of year round radon removal is
that there is no guarantee that radon problems will not be present even in
the summer months. The radon levels found in individual houses are a
complex result of radon source strength, soil transport, weather, and the
way the house is operated. To be certain of maintaining low radon levels in
the house normally requires that a SSD mitigation system work properly 24
hours per day, 365 days per year. It is for this reason that durability and
system performance are such important considerations. Performance level
goals are for 100% on-time operation for the life of the building. Thus we
require excellent durability of system components and a reliable means for
-------
determining whether the system is fully operational at all times.
The lack of long-term data on SSD systems is a major stumbling block in
determining whether or not the SSD systems perform adequately. This project
has been directed toward gathering such data from eight research houses that
were part of the Piedmont Study(l), as well as houses tested by the NJDEP as
a follow-up to mitigation activities.
2. Operational Environment
The question of durability of the mitigation system arises not only
from the need for a lifetime operation in the house, but concerns about the
environment to which the SSD system is subjected(2). Soil gas is often very
moist, causing condensation problems in the piping and at the fan of the
mitigation system. Also, particles can be drawn from the gravel bed; they
in turn may line the pipes and deposit on the fan or possibly interfere with
the fan bearings.
The moisture removed from the subslab can be very substantial, and
could amount to many gallons of water per day(2). Unless the piping design
allows for that water to drain back into the soil, the water could clog the
piping or interfere with the fan operation. Evidence of the moisture and
other debris has also been found in the staining of roofs near the exhaust
pipes of the SSD systems. Drying soil could result in slab cracking.
The amount of sand and other particles sucked from the soil must be
viewed as a possible cause of bearing failure or of the generation of
bearing noise (such effects can also be caused by the moisture). Noise can
directly influence the occupant to shut down the SSD system.
Another environmental effect that should not be overlooked is the
amount of airflow through the fan. To remain at an appropriate operating
temperature requires sufficient air to remove fan motor heat. Operating the
fan in a way that the blades are stalled can result in high bearing loads
and invites early fan failure.
OBJECTIVES
The following are our objectives for this research:
1. Our first objective has been to document the ability of the SSD
radon mitigation system to maintain houses at radon concentration levels
below the EPA guideline of 4 pCi/L. In these measurements we hoped to
observe the influence of such parameters as seasonal factors. Local weather
effects (e.g., rain storms) are covered in more detailed radon
monitoring(2).
2. A second objective was to observe the long-term characteristics of
radon levels in the SSD system exhaust. Source strength and transport
properties of the soil may be determined from these measurements.
3. A third objective was to evaluate the long-term influences of the
SSD system operation on the house substructure. Since our objective is
long-tern) operation, we need to know of any associated consequences.
-------
4. A fourth objective of the study is to determine critical parameters that
can degrade SSD performance and recommend ways to minimize such degradation.
THE TWO-PRONGED APPROACH FOR DURABILITY TESTING
Our approach to evaluation of durability is based upon our own
experience as to what might happen over time as well as on the experiences
of others; e.g., New York State Energy Research and Development Authority's
efforts to quantify durability(3), Lawrence Berkeley Laboratory research(4),
and Swedish studies which were able to look at houses after 5 years of
operation(S). Three short data sheets have been developed. One emphasizes
the house and mitigation system as observed by the house occupants; the
other data sheets involve a series of diagnostic tests to determine if the
mitigation system is achieving the necessary radon mitigation goals.
1. The house occupant questionnaire
Reviewing the data sheet, Radon Durability Diagnostics - I. the
emphasis is initially placed on whether the system has been running
steadily. Swedish studies have pointed to this problem as an explanation of
increasing radon concentrations(5). Our own experience is that occupants
hate to admit shutting off the system although system noise, radio
interference, and conserving electricity during the summer have been offered
as reasons to turn off the fan.
The second question concerns noise perceived by the house occupant. If
the system is becoming noisy, the occupant fears that the fan may be "on its
last leg," or that any noise may prompt an occupant to shut down the system.
The third question involves moisture. Ve are seeking to gain insight
into condensation, collection of water in the mitigation piping, or moisture
related events taking place at the roof exhaust. Water in the piping can
directly influence the amount of exhaust airflow possible. Condensation can
be another cause for occupants to turn off the mitigation system.
The fourth question is aimed at finding out about possible power
outages, construction in the house, or other events that could account for
higher than expected radon levels.
Question five involves the house occupants' perception of the system
and whether or not they have any questions about the way it functions.
2. Diagnostics and measurements
The second data sheet. Radon Durability Diagnostics - II. emphasizes
diagnostic procedures employed by the visiting inspection team. The
observations are initially visual: the house type, special features,
location of the SSD system, etc. On subsequent visits the visual
observations concentrate on location of cracks, sealed areas, etc. For
example, are there new cracks or places through which radon might enter the
house substructure? The inspection also concentrates on what sealing was
done previously and how well it has held up.
The second item involves noise generation. A stethoscope is used in
-------
order Co better detect early signs of bearing squeal which is an indication
that fan lifetime may be short.
Item three is a diagnostic check of the airflow in the mitigation
system piping. A heated-wire anemometer inserted through an inspection hole
to the center of the plastic pipe is used to measure the velocity. (In
houses where durability tests are in progress, openings in the piping are
sealed with duct tape to provide easy access during the study.) Care must
be taken that the airflow probe is sealed where it enters the pipe in order
to prevent erroneous airflow readings.
Item four makes use of those same openings in the piping to evaluate
pressure differentials. In this case, a digital readout micromanometer is
used. For this instrument, using the scale "inches of water," the readout
is the most sensitive (0.004 in. of water - 1 Pa).
Item five deals with measurement of radon levels in the exhaust of the
mitigation system. Over the long term, questions arise as to whether the
radon is being depleted or whether the soil is drying out; the latter could
lead to the system exhausting more distant radon gas. The radon levels in
the system exhaust would be an indication of such changes. Both pumped grab
samples and evacuated grab samples with Lucas cells are used in this
testing.
The final item is general observations. This is an opportunity to note
what a general inspection has revealed, and what may have influenced or
interfered with the radon mitigation process.
We also use the data sheet, Radon Durability Diagnostics - III. We
have been noting on this data sheet the serial numbers of alpha track
detectors that were in place as well as the replacement detectors. Normally
detectors have been placed in the basement/crawl space areas as well as on
the first floor.
RESULTS FROM THE NEW JERSEY PIEDMONT HOUSES
As previously discussed, there are four objectives of the study on
which we must focus attention. Perhaps the easiest way to review the
results is to plot the radon levels measured over time for each of the
Piedmont houses. Typical plots are shown in Figures 1-2 and summarized in
Table 1.
1. Occupant Effects
Immediately evident (in looking at Table 1) is that, while the majority
of the houses show more-or-less constant radon levels over time, two houses
(Houses 3 and 5 in Figure 1) show major variations in radon levels. These
variations include radon concentrations above the EPA guidelines as well as
House 3's return to premitigation levels.
On checking the occupant questionnaire it was noted that the occupant
of House 5 had turned off the SSD system because of radio interference and
because the occupant felt that, under summer conditions with open basement
windows, it was wasteful in terms of energy to operate the SSD system(6).
-------
The occupant of House 3 gave no hint on the questionnaire that the SSD
system wasn't working 100% of the time. Only when we informed the occupant
of the return to pre-mitigation radon values did the occupant remember that
the system had been turned off during a dinner party (room adjacent to the
fan) when mitigation fan noise was annoying and they had forgotten to turn
it back on for an extended period. We pursued this point further and found
that the noise was the result of vibration between the SSD fan motor and the
wood-framing in the attic above the garage. A small modification in the
mounting eliminated the problem. Similar vibrational problems have also
been experienced in one of our current research houses. The annoyance of
the vibration had resulted in the system being turned off. The importance
of avoiding such problems should be emphasized with the mitigator.
One point was very clear from even this very limited number of test
houses. The SSD system should not be turned off even for short periods of
time without immediately impacting the radon level. One occupant explained
that the system was only turned off for radio weather broadcasts to avoid
the static. The lesson is that the static should not be present through the
use of a higher quality speed controller, and careful checking of the wiring
arrangement would avoid interaction with sensitive electrical equipment.
2. Seasonal Variability
In order to look for such effects as seasonal variability, we must
focus our attention on houses where the occupants have allowed the SSD
systems to operate 100% of the time. Figure 2 (also see Table 1) indicates
this type of operation. A very simple evaluation of events over the
measurement period using Table 1 data (the basis for the figures) will
demonstrate the stability of radon concentrations over time. In this
exercise we have averaged readings for the first two periods and compared
them to the last two measurement periods. Altogether 10 measurements can be
compared if all basement and crawl space values are averaged equally. Thus
comparing October 1987 through May 1988 to November 1988 through June 1989
(a similar weather period) we find that in 7 out of 10 measurements radon
levels have dropped an average of 0.6 pCi/L. In 2 out of 10 houses, radon
levels have increased by an average of 0.1 pCi/L while one value has stayed
the same. Since the alpha track method of measuring radon levels has an
error band that would include these variations, our conclusion would be that
there is no significant change over the year.
Because of these same arguments of measurement error it is even more
difficult to look for seasonal effects. However, any significant seasonal
influences should be evident in these data. House 2, for example, indicaces
that winter readings are slightly higher than summer readings. But again
using October-February and November-March basement values we find a 1.6
pCi/L average versus 1.2 pCi/L for May-November combined with March-June.
This 0.4 pCi/L difference is well within the error band.
House 4 shows a general decline in radon concentrations with time and
no sign of seasonal fluctuations.
House 6 shows only a slight increase in the November-June reading and a
noticeable drop in radon level after the first October-February period.
-------
House 7 also shows a very small increase in the November-June period in
basement and crawl space radon concentrations. First floor concentrations
are basically background levels. Note that the substructure concentration
change is from approximately 0.5 to 1.0 pCi/L, again well within the error
band of the alpha track radon sensors.
Only in Houses 8 and 10 (Figure 2) do we see variations in the
November-March period that could be viewed as seasonal influences. In the
case of House B similar increased concentrations were measured during the
October-February period, the year prior. Averaging basement and first floor
values for House 8 (from Table 1) for the "winter periods," we obtain an
average value of 4.3 pCi/L, a value in excess of the EPA guideline. If we
average "summer periods" the value is only 1.8 pCi/L. This house
illustrates how seasonal influences can clearly affect SSD performance.
House 10 indicates only a slight increase in radon concentration in the
October-February period; i.e., 2.2 pCi/L average versus 1.5 pCi/L for the
"summer periods." However, the increase is more substantial in the
November-March period where values of 2.7 pCi/L are observed. This house
would also appear to exhibit seasonal effects that increase radon
concentrations by approximately 1 pCi/L.
3. Radon Levels in the Mitigation Exhaust System
One measurement of durability testing that is of interest is the
concentration of radon in the exhaust pipe from the subslab mitigation
system. When this information is combined with the velocity (or flow
measurement) at the same pipe location, we are in a position to calculate
the total flow of radon gas from the mitigation system(7).
One question to be resolved is: Can we compare the amount of radon
exhausted from any given house and understand fully the role played by the
mitigation system? One such comparison involves the natural flow of radon
through the same house. The physical meaning would be a radon flow rate
equal to the quantity of radon exhausting through a "perfect mitigation
pipe" through which the airflow rate is equal to air infiltration rate of
the house. To make the calculation of the natural flow requires a knowledge
of the average air infiltration rate for the house, the radon concentrations
upstairs and in the basement/crawl space, and the volumes of those zones.
The calculation is:
(Cu x Vu) + (Cb/c x Vb/c)
R - AI x 103 x - -
Vu + Vb/c
where R - radon flow, pCi/h
AI - air infiltration, m3/h
C - radon concentration, pCi/L
V - volume, or
u - upstairs
b/c - basement/crawl space
Comparative results for five Piedmont houses are shown in Table 2. The
ratio of the radon gas exhausted from each house via the mitigation system
is compared to the radon natural flow value through the house. Ratios vary
-------
from 1 to 9. The question that comes to mind is: "What is special about
House 4 which has the same flow of radon gas both through the mitigation
system or by natural means?" One answer could be, it is the house with the
least porous soil -- basically pure wet clay. It is a house where high
ventilation rates (such as using a blower door) depress the radon levels,
and then it takes many hours for the house to return to the previous
elevated levels. Such behavior has been interpreted as evidence of a
limited radon entry rate. However, just comparing the natural radon entry
rate with those of the other houses would not single out this house. In
fact, looking at natural radon entry rates, it is only House 3 that stands
out with a rate far above the others; i.e., 35 x 106 pCi/h versus 7.5 + 3 x
10° pCi/h for all the rest. House 3 has a soil condition of high porolity,
(i.e., stone flour roughly 1/8-in. (0.3 cm) diameter and a good gravel bed),
Just the opposite conditions of House 4.
The total amount of radon gas mechanically exhausted from the soil
varies by a factor of 7 in these houses. The lowest value is for House 4
with clay soil. The highest value of radon is for House 3 with very porous
soil. The two highest houses (Houses 3 and 5) were those with recurring
periods of above-guideline levels of radon. However, House 7 has clay soil
mixed with shale and ranks third.
The method used to analyze the radon concentrations from the mitigation
system exhaust involves Lucas cells. Analysis of the Lucas cells takes into
account the time elapsed from the time the sample was taken to the time it
was analyzed, the background level of the cell, and the efficiency of the
measurement equipment. The relationship used is:
(CPMread - CPMbgd) x e«>-00803 x At)
cactual ~
-------
period and House 3 in which real levels have increased because of airflow
variation. This will be discussed later.
In the second and third testing periods, when we measured exhaust
concentrations of radon (3/89 through 6/89, and 6/89 through 11/89), the
concentrations vary with the individual houses. Houses 3 and 10 show a
decreasing trend, while Houses 2, 4, 5, 7, and 8 indicate increasing levels.
Some seasonal effects may be present in that the lowest readings for the
majority of the houses were in months 15 and 18 which is the warmer time of
the year [the plots start at month 11 (November 1988); therefore, 15 is
March 1989 and 18 is June 1989]. Throughout these discussions, radon
concentration has been used since, hopefully, it provides the more physical
meaning, and because airflow rates (necessary to provide total radon flow as
shown in Table 2) tend to be constant in the houses. One exception to
constant flow is House 3 where, after the fan had been turned off by the
occupant and after radon levels had increased in the living space, the fan
speed was increased. Duct velocities changed from 2.65 m/s in November 1988
to 8.2 m/s in November 1989 (intermediate readings were 7.2 m/s). The trend
of falling concentration levels over time for House 3 does apply for the
last three periods, with a peak at month 15. Because of the lower fan
airflow rates present during month 11, radon flow rates are less than half
the month 15 peak. The profile of House 3 is opposite to the general trend
experienced in the majority of the test houses.
4. Substructure Changes
Based upon diagnostic team observation, only House 3 showed evidence of
physical changes. Two cracks appeared in the basement slab near the slab
edge extending toward the center of the room. Length of the cracks was
approximately 6-ft (1.8 m) and the width exceeded 1/16-in. (0.2 cm) at some
locations.
A noteworthy observation of what was happening in the basements or
beneath the slab was that conditions were noticeably drier. Several
occupants have stated that the need for summer dehumidification was
eliminated in their houses. Where observations were possible, water in
gravel beds was no longer visible. No quantitative measurements of relative
humidity have been made.
PRELIMINARY RESULTS FROM NJDEP HOUSES (MITIGATED BY PROFESSIONALS)
The concern for durability and performance of radon mitigation systems
was pointed out by Depierro and Cahill of the NJDEP(8). Based upon their
findings, 64% of the houses mitigated by occupants and professional
mitigators were not achieving the 4 pCi/L guideline. Even when only
professionally mitigated houses were assessed, the percentage of houses
failing to meet guidelines still exceeded 50%.
With this information as background, we undertook a program of
upgrading the radon information. Our test houses were taken from the larger
list of NJDEP-tested houses. From that list our criteria of selection were
houses less than half an hour's drive from Princeton, houses with
predominantly SSD systems installed, and houses with the highest post-
mitigation radon levels. Our approach was to question the occupant on the
-------
radon history in the house, to briefly inspect the mitigation system
installation, and to leave a charcoal canister in the house which 'after 3
days the occupant would mail in for analysis. Two examples are given.
Test house A. This house has a SSD system which uses three slab
penetrations routed to a fan mounted on a cinder block wall. Noise from the
fan caused the occupant to build an insulated box around the fan. The house
in early 1987 had measured radon levels from 40 to 120 pCi/L in the basement
and 13 to IS pCi/L upstairs. The values disagree between the state and the
private mitigator. Measurements 1 year later, after mitigation, indicated
levels reduced to less than the EPA guideline value. Our measurements in
October 1989 confirmed that the guideline was being met.
Test house G. The SSD system in this house has two floor penetrations
as well as a penetration at the wall under the crawl space. Early readings
in September 1987 were 143 pCi/L in the basement and 75 pCi/L in the living
room. Levels dropped to only 6-7 pCi/L in the basement and living room
until an improved fan, Kanalflakt K-6, replaced the original fan and dropped
levels to the 0.8 pCi/L level. Our tests in August 1989 indicated minimum
detectable concentrations of 0.54 pCi/L.
To briefly review these data, we found no cases where the substructure
radon levels were above the 4 pCi/L guideline. This was true even though we
chose the highest radon houses on the list supplied by NJDEP. One reason
for this was that occupants continued to have their systems checked and
improved with better fans and higher flow rates. Fan lifetime appears to be
short in several of the fans used, while the Kanalflakt fans continued to
perform satisfactorily. Similar to the Piedmont houses and our other test
houses, noisy fans can be a result of fan mounting, necessitating good
installation procedures if this critical problem is to be avoided.
CONCLUSIONS AND RECOMMENDATIONS
1. House occupant interaction significantly affected the operation of
the mitigation systems, based upon our survey.
One point that is very evident from the durability diagnostics program
is that increases in radon levels can often be traced back to occupant
intervention with the SSD radon mitigation system. In the majority of
cases, noise from the system caused annoyance and the system was shut off,
often for prolonged periods. Another instance was an occupant seeking to
limit operation of the system in the summer period when he erroneously
thought that open basement windows would suffice, and therefore the
electricity to operate the SSD system could be saved. Even in this
instance, radio interference --a different kind of noise -- instituted the
habit of this occupant to shut off the fan.
2. Installation and Building Anomalies
Every house is different from a radon standpoint. Even in the same
area, where the weather is the same, the soil porosity and radon content
often varies noticeably across an individual basement floor, wall systems
below grade are highly variable, equipment used to space condition the house
-------
can drastically alter interior pressures and that condition can vary from
room-to-room, the house design can have a major impact, etc. Given this
list of variables it is no wonder that any mitigation system must take these
factors into account. The installation must also consider moisture in the
soil gas. One of the Piedmont houses experienced major problems with water
in the mitigation piping. The problem was lack of attention to the slope of
the piping, causing condensation water to collect and not drain to the
gravel bed.
Anomalies often can be associated with the unknown conditions that
prevail beneath the basement and crawl space slabs. The amount of gravel,
hidden pipes, or partial concrete pours that separate the total floor area
into several zones requires that more than one slab penetration must be
used. Ve observed other houses with wall suction at the wall adjacent to
slab-on-grade construction. Here extra radon entry was expected and
basically taken care of by the wall penetration. In another house, very
similar when viewed casually, the SSD system may perform beautifully with
one penetration and remove radon from the walls at the same time.
Space conditioning cannot only circulate radon in the house but can
also directly influence the amount present in the house. Ve were aware that
inadequate returns in warm air systems resulted in under-pressurization of
the basement and more soil gas being drawn into the house.
3. Recommendations
Our recommendations to the installer of mitigation systems is to use
long-service-life components. The fan should be designed to operate for
long periods; the fewer replacements and the more trouble free, the safer
the occupants. Since wide variations in fan flow rates were found between
different regions of the country, the choice of fan must take this into
account to avoid fan overheating and/or bearing failure. The same criteria
of quality apply to sealants. This is a long-term installation and not the
place to use cheap caulk. Installations at the sump require careful design
so that the water-removal function of the sump pump does not compromise the
SSD system and vice versa, when done properly we observe little or no
degradation over time of these installations.
A clear perception of the characteristic of soil gas requires that the
mitigation system installer must properly slope all piping to drain any
condensing moisture into the gravel bed.
Electrical hookup of the SSD system should not generate electrical
noise in sensitive household equipment such as radios. The installer should
check with the occupant at the time of installation of the system that such
problems do not exist. Mechanical noise and vibration should also be elimi-
nated to prevent annoyance to the occupant and make certain noise is not a
cause for system shutdown. Several of the Piedmont and research house in-
stallations suffered from fan vibration interacting with nearby structures,
or with mitigation piping. These were easily corrected. Checking for such
noise problems at the time of installation is highly recommended.
The occupant not only is sensitive to SSD system noise but also is
normally very interested in maintaining the lowest possible radon levels.
-------
We recommend that the occupant regularly retest for radon. In the NJDE?
houses, retesting led directly to system improvements and system operation
better suited to the long-term (improved fans, sound suppression, etc.).
These houses lowered radon levels to values below EPA guidelines.
Installation of a pressure switch-activated indicator light to inform the
occupant when the system is not operating properly is highly recommended.
ACKNOWLEDGEMENTS
This work was funded by the U.S. Environmental Protection Agency under
Cooperative Agreement No. CR-814673 with Princeton University.
REFERENCES
1. Dudney, C.S., et al., Investigation of Radon Entry and Effectiveness of
Mitigation Measures in Seven Houses in New Jersey, Oak Ridge National
Laboratory Rpt. ORNL-6487, August 1989.
2. Harrje, D.T. and Hubbard, L.M., Proceedings of the Radon Diagnoseics
Workshop. April 13-14. 1987. EPA-600/9-B9-057 (NITS PB 89-207898), June
1989, also see Harrje, D.T., Hubbard, L.M., and Sanchez, D.C.. "Diagnostic
Approaches to Better Solutions of Radon IAQ Problems," Healthy Buildines '88
- Planning. Physics and Climate Technology for Healthier Buildings. Vol. 2,
Swedish Council for Building Research, Stockholm, Sweden, 020:1988 (ISBN* 91-
540-4933-4), pp. 143-152.
3. Nitschke, I., et al., "Preliminary Results from the New York State
Radon-Reduction Demonstration Program," Proceedings of the 1988 Symposium
on Radon and Radon Reduction Technology. Volume 1, EPA-600/9-89-006a (NIIS
PB89-167480), March 1989.
4. Prill R.J., Fisk, V.J.. and Turk. B.H., "Monitoring and Evaluation of
Radon Mitigation Systems Over a Two-Year Period," Ibid.
5. Nilsson, I. and Sandberg, P.I., "Radon in Residential Buildings -
Examples of Different Types of Structural Counter-Measures," Healthy
Buildings '88 - Planning. Phvslcs. and Climate Technology for Healthier
Buildings. Vol. 2, Swedish Council for Building Research, Stockholm, Sweden
020:1988 (ISBN 91-540-4933-4), pp. 163-172.
6. Harrje, D.T., et al., "The Effect of Radon Mitigation Systems on
Ventilation in Buildings." ASHRAE Transactions 1989, Vol. 95, Ft. 1.
7. Harrje, D.T. and Cadsby, K.J., "Airflow Measurement Techniques Applied
to Radon Mitigation Problems." Proceedings of the 10th AIVC Conference -
Progress and Trends in Air Infiltration and Ventilation Research. AIVC,
Coventry, UK, 1989.
8. Depierro, N. and Cahill, M., "Radon Reduction Efforts in New Jersey,"
Proceedings of the 1988 Symposium on Radon and Radon Reduction Technology.
Volume 1, EPA-600/9-89-006a (NIIS PB89-167480), March 1989.
-------
DURABILITY-HOUSE 3
<
-------
DURABILITY-HOUSE 8
u
o.
IUY-NOV
HUE (1987-1989}
C HAS A
HO V-MAR
KA3. ..'UH
o
o.
o
B
7
6
a
4-
s -
1 -
DURABILITY-HOUSE 10
c—
OCT-rt3
m-iur
MAY-MOV
HOV-KAR.
TIME (t9B7-19B«3
•0 BAS —ft 1-nMR
KAft-AJM
Figure 2 Radon levels in the interior spaces of Houses 8 and 10.
-------
1500
H2,Crci
II
H3.C«Tr
H3,Pum?
H4,Purr,p
HS.Grob
H5.Pump
Figure 3 Radon levels In the SSD system exhausts of Houses 2 through 5.
-------
TABLE 1
RADON' CONCENTRATIONS IN EIOHT HOUSES
Occ '87
11 Feb '88
11 Feb
20 Kay
'88
'88
23 Kay
16 Kov
'88
'88
16 Kov '£=
30 Kar '65
HOUSE LOCATION
33 Mar '£9
25 Juni '£9
2
3
4
5
6
7
8
10
Basesenc
Base=enC
Dining Rooa
Basemen:
Basesenc
Living Re OB
Base=enc
Living Rooa
Breezevay
Basesene
Kiddle BR
Crawl Space
Base=en:
Basesene
Living Rooa
Bas eseat
Crawl Space
Living Reea
Basesenc
Living ROOD
Basesenc
Living Rooa
2.1
2.6
6.S
9.9
4.8
3.1/3.0
2.8
•
11.6
8.4
6.8
4.8
5.1
2.6
1.0
0.8/0.6
0.6
5.6
3.6
2.4
1.9
1.91
1.3
1.1
1.2
0.6
2.3/2.6
3.1
-
0.7
0.8
2.8
1.9
2.5
1.6
0.3
0.6/0.3
0.3
1.9
1.7
2.2
1.9
1.4
1.0
0.6
7.6
8.4
4.7
2.8
2.7
•
9.8
6.0
2.2
1.7
2.3
1.3
0.5
0.2/0.4
0.3
0.9
1.3
1.8
1.7
1.8
2.0
1.5
53.2
40.6
27.0
2.6
1.8
2.1
12.9
12.4
-
2.7
2.2
1.8
1.2
1.1/0.8
0.3
3.6
4.6
2.6
2.7
1.3
1.1
0.4
2.5
3.6
0.8
•>
1.0
1.3
0.4
0.6
-
3.1
2.6
1.3
0.5
0.3
0.3
2.2
0.9
1.5
1.3
TABLE 2
COMPARISONS 07 RADON QUANTITIES EXHAUSTED BY MITIGATION
SYSTEMS AND BY NATURAL MEANS BASED ON FIVE HOUSES
Ho. IMIMIHJ vp*uir* Dumtnt | Uptuln
I
J
'
S
7
II
170
19
to
33
1}
TO
Si
IS
It
219
».
Ill
JT1
119
»t
161
•99
39*
391
i.i/si bft.utcKi/u issfa fgjsr"? 'aw "«/.:
19B 1S«
131 9«t
«3 M
135 <1S
IOJ SOI
J.JO
I.6S
I.M
«.SS
».»]
1S.TJ1.OM
TJ, 1*7, 000
10,901.000
ST,7t7,000
JJ.SBT.Oflfl
«, 971,000
n.sas.ooo
10.1TI.000
l,J»3,ooo
1. 180.000
t.Ii
l.ll
1.01
9.01
1.10
-------
TABLE 3
DURABILITY RADON SAMPLES
House Date Month Grab Pump
H2,Grab
H2,Pump
H3,Grab
H3.Pump
H4,Grap
H4,Pump
HS.Grab
H5,Pump
H6,Grab
H7,Grab
H7,Pump
H8,Grab
H8,Pump
10
HlO.Grab
HLO,Pump
11/88
3/89
6/89
11/88
3/89
6/89
11/89
11/88
3/89
6/89
11/88
3/89
6/89
11/89
11/88
3/89
6/89
11/88
3/89
6/89
11/89
11/88
3/89
6/89
11/89
11/88
3/89
6/89
11
15
18
11
15
18
23
11
15
18
11
15
18
23
11
15
18
11
15
18
23
11
15
18
23
11
15
18
146
65
224
808
810
601
299
28
50
84
360
70
99
136
N/A
N/A
458
382
116
179
342
N/A
N/A
485
978
N/A
542
305
163
65
268
1086
864
621
133
60
61
119
510
75
134
170
N/A
N/A
300
625
166
244
384
N/A
N/A
566
1050
N/A
614
360
-------
D-VII-4
RADON ABATEMENT SYSTEM ANCILLARY ITEM
by: Ronald F. Simon, President
RFSimon Company, Inc.
RD#2 Box 481
Barto, PA 19504
ABSTRACT
The task of the radon researcher was to develop approaches to reduce
indoor radon concentrations. As techniques were developed and tested, the
mitigation industry developed around these approaches. In most cases, the
research techniques were applied directly to the private sector. As the
number of installations grew and the technology was understood, the focus
changed from installation methods to ancillary items. These ancillary
items pertain to system installation techniques as they pertain to standing
building and fire codes. This paper deals with maintaining fire protection in
a soil depressurization system.
-------
RADON ABATEMENT SYSTEM ANCILLARY ITEM
The primary task of the early radon professional was to identify methods
to reduce the indoor radon concentrations. In some cases, he was not sure that
reductions could be made to the accepted 4 pCi/1 level. The task at hand was
to develop mitigation strategies that would control indoor radon concentrations.
Through research programs and field installations a variety of techniques
became known and generally accepted by an emerging mitigation industry. Soil
pressure manipulation soon became the work horse of the industry. In particular,
soil depressurization became the system of choice. As more was learned about
soil depressurization, installation techniques and material system choices were
modified in order to align with the level of effort required. The Intent of
this alignment was to make system installation less labor intense while still
controlling indoor radon.
Mitigation system installation standards have eluded the industry as a
whole. One of the contributing factors allowing this to take place is that the
industry has sprinted ahead with little regard for standardized installation
practices, unaware of possible unintended consequences. Typically the mitigation
installation focuses on a high profit/low cost installation. Standards also are
difficult to initiate and maintain possibly because of the installers limited
capabilities based on background, knowledge and skills. EFA's recommendation
for above eave exhaust location is a basic practice accepted in hazardous
material discharge, yet today, some installers still do not follow this guideline.
Within the frame work of the mitigation system evolution, little attention
has been placed on the appropriateness of the system with respect to the applicable
building codes. This paper is written to address the code issue and in particular
the fire barrier penetrations in both residential and commercial structures with
respect to radon soil pressure manipulation systems. Emerging EPA guidelines
recommend the installation of soil depressurization systems utilizing above
eave exhaust locations. To accomplish this, pipe systems typically penetrate
in place fire barriers within the structure both floor to floor in commercial
construction and garage to occupied space in residential construction. When
this takes place, the occupied quarters are placed at risk. The potential for
loss of life because of an installed abatement system and less important, the
liability issue, directed RFSimon Co., Inc., to identify installation techniques
and practices that would eliminate this risk. In 1988, 4,955 lives were lost due
to residential fires.
-------
The BOCA (Building Officials and Code Administrators) for commercial
and the CABO (Council of American Building Officials) for residential construction
addressed this issue with the following quote from the BOCA National Fire
Prevention Code: "The BOCA National Building Code was initially prepared and
has been maintained on the premise that all matters pertaining to the construction
of the building and built into it, either in its initial construction or through
subsequent alterations, repair or extension, should be covered by the building
code. This includes fire protection of the building elements as well as fire
separation walls or other precautions required for protection against specific
hazards of the particular use of the building."
F-105.1 General: Whenever the code official or the code official's designated
representative shall find in any structure or upon any premises dangerous or
hazardous conditions or materials as follows, the code official shall order such
dangerous conditions or materials to be removed or remedied in accordance with
the provisions of this code:
1. Dangerous conditions which are liable to cause or contribute to the
spread of fire in or on said premises, building or structure or endanger
the occupants thereof.
2. Hazardous conditions arising from defective or improperly used or
Installed electrical wiring, equipment or appliances.
To understand how a typical soil depressurization system would perform in
a fire situation, a full size model was fabricated to simulate a typical Radon
Vent Duct (RVD) routing. The model consisted of a floor deck utilizing 2x10 nom.
joist and collar or band joist, 5/8" plywood decking and a 2x4 nom. wall system
with wall insulation. The wall system and band joist were covered with fire
code dry wall. This model represents a typical wall/floor construction detail
in houses with attached garages. The building code typically requires the
garage floor be 8" min. below the house floor in order to contain gasoline
engine fume emissions. However, because of this required step, the garage/house
common wall represents a direct exit from the basement to the garage area.
Typical RVD systems are routed horizontally from the basement to the garage
area, then vertically to the attic and terminated outside of the building shell.
As part of the model, a 4" PVC pipe system was installed as described above.
See Sketch SK-1. A Fan-Tech Model R-150 fan was installed approximately 10'
above the horizontal pipe run. The pipe was insulated using a polyethylene
covered fiberglass duct insulation. The effluent end was covered with a
SIMON CAP while the inlet end was dampered to simulate expected operating
static pressures within the system.
A fire was set at the base of the vertical stack just at the transitional
90° el through the dry wall/band joist area. The model was built to confirm
what was thought would take place. The thought was that the PVC pipe would
melt open. The open pipe would draw the fire and smoke up the pipe as a result
of the force exerted by the fan thus spreading the fire and smoke to other
areas of the structure.
-------
The damper box is secured to the structure according to the manufacturer's
recommendations. The box is fastened so that it will remain in position and
function during a fire. Within the damper box rests a curtain type blade
closure mechanism. A fusible link holds the blade closure mechanism in the
open position. The blade closure mechanism is activated when the fusible
link is destroyed through exposure to elevated temperatures. Typically
the link will soften and release at 160° F. Vertical mount units depend
on gravitational force to close the curtain within the housing. Horizontal
mount units are fitted with a closing spring that draws the blades across the
opening. One concern when incorporating these units in the Radon Vent Duct
or RVD is the potential build up and continuing exposure to condensation.
These units are fabricated utilizing galvanized metal enclosure and curtain
dampers and the spring in the horizontal unit is stainless steel. However,
long term exposure to typical system conditions could perhaps render the
closing mechanism ineffective due to corrosion build-up, particularly in
the horizontal mount units.
Some structural work may be necessary to meet manufacuturer installation
requirements. The opening for the damper housing must be larger than the
assembly, typically 1/8" per ft. of damper housing plus 1" for the maximum
and a minimum of 1/4". These clearances are critical to allow for housing
expansion and free curtain damper mobility in elevated temperatures.
Mechanical fastening of the unit to the structure requires a metal sleeve
and angle supports mechanically fastened to the damper housing. These
supports are positioned on both sides of the penetration captivating the
structure while allowing the damper housing to expand freely. Sealing
between the RVD and the extended collar should be completed to maintain a
closed RVD system. The cost of these units vary depending on size and mounting
configuration. A 4" extended collar, horizontal mount unit cost is $40 and
the vertical mount equivalent is slightly more. Ancillary materials, i.e.,
sleeves, mounting angle, screws and caulk will be in the $10-$12 range.
Structural preparation will vary from site to site. The installation techniques
required for proper installation add significantly to the installed cost.
This type of installation would be utilized more in commercial RVD systems,
particularly in through floor to floor applications. The question of durability
and function will, however, require further study.
Another approach that was evaluated is the use of an intumescent fire stop
system. This technique incorporates some of the elements required for the
damper systems identified above. The installation requires 5 key elements
consisting of the intumescent wrap, foil faced tape, galvanized retaining
collar, adjustable stainless steel band and intumescent fire caulk. The
installation is completed at the fire wall in the unprotected area. The
procedure requires the application of the foil faced tape over the FVC pipe
system to protect the pipe from softening due to volatiles within the
intumescent. The preformed strips of intumescent are 2" x 20" x 1/4" thick
and are cut to length to fit snuggly around the RVD at the fire wall penetration.
Tape is used to temporarily secure the joint and hold the strip in place.
Additional layers are installed with the joints staggered. The number of
intumescent wraps depends on the pipe diameter; a 4" diameter RVD will require
3 bands of material.
-------
When the system was ignited the polyethylene vapor barrier was consumed
immediately in the area of the flame and heat. The duct insulation delayed
the impact of the flame on the PVC pipe because of its inherent properties.
Gradually the pipe was exposed to increased heat as indicated by the
increasing density of effluent smoke. As the flames and heat exposure to
the PVC pipe and elbow increased, the pipe system softened. The softening
of the pipe continued until the structural integrity of the pipe system was
below the level required to remain open. It began to collapse inward
responding to the system induced negative pressure.
The PVC pipe continued to close in on itself until it was completely
closed, causing the fan side of the system to be isolated from the fire source
and from further damage. The system was heat softened and sealed, subsequently,
operating under full negative pressure capacity of the fan. This was evidenced
by the gradual and eventual complete termination of smoke from the stack. This
event was in complete disagreement with anticipated results.
At this point, however, the flames had an open access through the in place
fire wall and into the basement. The PVC pipe system burned and created a
pathway for the fire to follow. The flames spread along the RVD into the
occupied side of the fire wall. Within minutes the band joist and plywood
sub-floor were ignited. Eventually, the model was consumed in flames.
Armed with this information the task ahead was to find an installation
process that would eliminate this from taking place. Several approaches were
investigated that would allow a mitigation system to overcome this potentially
lethal installation flaw.
An obvious approach would be to completely enclose the pipe system in a
fire code dry wall chase. There is, however, concern that this type of remedy
is not acceptable in that a chase is created. The chase would still have the
ability to transfer flames within itself and perhaps require special sealing
at the pipe penetrations within the chase. From an installation standpoint
this approach is labor intense and requires structural materials to complete.
Also in some installations this approach would be impractical or impossible
to implement.
Another approach would be to use a non-combustible pipe system through
the non-fire protected areas, i.e., garage. This could be accomplished using
a copper Drain, Waste, Vent or DWV pipe. The pipe would be connected to the
standard PVC RVD system using flexible couplings to make the transition to
PVC. A special intumescent sealant must be applied to stop flame spread
through the fire wall/pipe penetration. This will serve to inhibit flame
spread around the pipe through an irregular opening. Although this technique
is feasible, the expense of the copper pipe, its weight and pipe support
implications would make a significant contribution to the installation cost.
A mitigation system can be isolated from the occupied spaces utilizing
an in-line fire damper. This device is a sealed metal box with 4" diameter
extended collars at opposite ends.
-------
When this is completed, a pre-stamped galvanized metal retaining collar is
cut to length and folded to capture the intumescent and form attachment ears.
The retaining collar is secured around the intumescent with a stainless
steel adjustable band. The unit is then secured with screws through the
attachment ears to the structure. The final step is to seal the area
between the fire inhibiting wall or ceiling and the retaining collar. This
requires an intumescent caulk material. The intumescent material and caulk
system is designed to expand when exposed to elevated temperatures. This
expansion is directed toward the PVC RVD pipe system reacting against the
retaining collar and the stainless steel retainer band. The PVC RVD is
softened as the temperature increases facilitating its forced closure. The
intumescent forms a hard char material which, by design, remains intact
throughout the flame exposure. The materials cost for this type of
installation is $45. The number of installations will vary depending on
the number of fire barrier penetrations.
The issue of risk prevention is an important element within the frame
work of the mitigation installation. The approaches presented here have the
ability to significantly reduce the potential loss of life due to system
interaction with fire. The fire protection system chosen, as with many other
mitigation options, is based on the site specific influences. In some cases
the damper system may be appropriate while the intumescent system may be
applicable in others. The cost associated with the installation is minimal
when compared with the potential for loss of life and property. Professional
design and mitigation companies should develop skills in the installation of
fire protection systems recognizing this is not an option but a mandatory
ingredient for code compliance.
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.
-------
5K-1
GARAGE
SSV PIPING
CONCRETE SLAB
STUD WALL
FIRE CODE DRY WALL
SUB-SLAB AREA
BASEMENT
-------
Paper D-VII-5
LABORATORY STUDIES OF "BETWEEN THE ROOMS"
RADON DECAY PRODUCT REMOVAL UNITS
by: Dade W. Moeller, and
Xiaowei Yan
Harvard School of Public Health
Boston, MA 02115
ABSTRACT
A temporary wall was constructed to divide a large
laboratory chamber into two sections. Mounted in the wall
were two fan-positive ion generator units, each blowing in
opposite directions so as to maintain a balance in the
airflow between the two sections. Data showed that the
units provided PAEC reductions in the two sections of the
chamber in excess of 90%. Estimates of the overall
reductions in the lung doses were 86% (using the Jacobi-
Eisfeld model), 88% (using the Harley-Pasternak model), and
92% (using the James-Birchall model).
INTRODUCTION
A temporary, 4-m wide wall was constructed in the
Harvard Radon Chamber (1) to simulate two adjacent rooms in
a residence. The left section of the chamber had a volume
of 35.3 m3 and the right section had a volume of 43.2 m3.
Two radon decay product (RDP) removal units (fan and
positive ion generators) were installed 1.5 m apart in this
wall at a height of 2.0 m above the floor (shown
schematically in Figure 1). One fan directed an airflow of
approximately 50 L/s from one section of the chamber into
the other. The other fan directed an equal airflow in the
opposite direction so that pressures in the two sections
were equal. Positive air ions were dispersed into the
respective airflows.
LABORATORY TESTS AND RESULTS
The sequence of tests consisted of an initial time
period when the RDP removal unit ion generators were off
(only the fans were operating) and a second time period when
both the fans and ion generators were operating, with
measurements being made in one section of the chamber. This
-------
sequence was then repeated with measurements being made in
the other section of the chamber. Radon concentrations in
the laboratory chamber, as well as graphs of the potential
alpha energy concentrations with the fans on, and with the
fans and ion generators on, are shown in Figure 2. As may
be noted, the radon concentrations (upper curve) were
essentially constant throughout the studies, whereas the
fans plus ion generators reduced the potential alpha energy
concentrations by greater than 85%. Due to limitations on
the availability of monitoring equipment, the PAECs were
monitored in the left section of the chamber on the first
two days, and in the right section on the subsequent two
days.
Figure 1. Schematic representation of "between the rooms"
fan-ion generator airborne radon decay product
removal units.
-------
500
S
M
2 400
II
If
£l 200
^s
M
£ 100
Left Section
of Chanber
-•j-Fans Onn
Right Section
of Chaaber
Fans + Ion
Generators On
SO
10
Tim (DAYS)
Figure 2. Radon and potential alpha energy concentrations in
left and right sections of the laboratory chamber
with either the fans or the fan-ion generator
units in operation. Only the fans were in
operation on the first and fourth days; both the
fans and the ion generators were in operation
during the second and third days.
-------
Grab samples of radon and its decay products were
analyzed by standard methods (1). Single-screen samples
were analyzed for unattached decay products (2). These data
were used to calculate reductions in total PAECs, and
unattached PAECs which are summarized in Tables 1 and 2.
Reductions in total PAECs averaged 87% and reductions in
unattached PAECs averaged 56%. Estimates of the reductions
in lung dose to the basal cells of the tracheobronchial
region are summarized in Tables 3, 4 and 5. The percentage
reductions given in Tables 1-5 are based on the use of the
two "between rooms" RDP removal units (i.e., fans and ion
generators) as compared to the use of the identical devices
with the ion generator turned off (i.e., only the fans
operating). If the reductions had been compared to "no
treatment conditions" (i.e., both the fans and ion
generators off), the calculated percentage reductions would
have been greater. The range of probable reductions in the
lung doses attributable to the fans and ion generators,
based on linear forms of three dosimetric lung moels (3,4),
and their corresponding limits are shown below. Estimates
of the reductions, as compared to "no treatment conditions,"
were calculated on the assumption that the fans, alone,
would have reduced both the total and unattached PAECs by
50% (5).
Dosimetric Lung Models Percentage Percentage
and Model Limits Reduction Remaining
Minimum (i.e., effect on
unattached PAEC) 79% 21%
Lung Dose
Jacobi-Eisfeld Model (6) 86% 14%
Harley-Pasternak Model (7) 88% 12%
James-Birchall Model (8) 92% 8%
Maximum Effect (i.e., effect on 93% 7%
total PAEC)
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:
Assistance in the collection of the data reported in
these studies was provided by Captain Carl A. Curling, U.S.
Army, formerly a Doctoral Student, Department of
Environmental Science and Physiology, Harvard School of
Public Health.
-------
TABLE 1. EFFECTIVENESS OF FAN-ION GENERATORS IN REDUCING TOTAL PAEC*
Experimen
Number
1
2
3
Average
TABLE 2.
Experimen
Number
1
2
3
Average
Left Section
PAEC (mWL)
of Chamber
t Fan Fan + Ion Percent
On Generator
180 19.1
219 26.8
201 22.6
200 23
EFFECTIVENESS OF FAN
Left Section
PAEC (mWL)
t Fan Fan + Ion
On Generator
8.8 3.6
10.8 4.5
7.3 4.6
9.0 4.2
On Reduction
90
88
89
89
-ION GENERATORS
of Chamber
Percent
On Reduction
59
58
37
51
Fan
On
191
237
201
210
Right Section of
PAEC (mWLJ
Fan + Ion
Generator On
22.5
30.5
27.7
26.9
IN REDUCING UNATTACHED
Fan
On
12.4
13.5
9.4
11.8
Right Section of
PAEC (mWL)
Fan + Ion
Generator On
4.0
5.2
5.6
4.9
Chamber
Percent
Reduction
88
87
86
87
PAEC*
Chamber
Percent
Reduction
68
61
40
56
*External air exchange rate was 0.2/h.
-------
TABLE 3. EFFECTIVENESS OF FAN-ION GENERATORS IN REDUCING MEAN BRONCHIAL LUNG
DOSE EQUIVALENT RATE (James-Birchal1 Model)(8)
E
xperiment
Number
1
2
3
Average
1
(
1
1
1
1
Left Section
of
Chamber
reatment Conditions
Dose Rates in Sv/y)
Fan
On
.24
.50
.19
.31
Fan + Ion
Generator On
0.319
0.409
0.402
0.377
Percent
Reduction
75
73
67
72
1
(
1
1
1
1
Ri
ght Section of
Chamber
reatment Conditions
Dose Rates in Sv/y)
Fan
On
.53
.76
.35
.55
Fan + Ion
Generator On
0.360
0.470
0.489
0.439
Percen
Reducti
77
73
64
71
t
on
Notes :
1
2
The external
Percent
air exchange
Dose Rate (F
rate was 0.2/h.
an) - Dose Rate
(F
an +
Ion Generator)
( innm .
3.
Reduction
James-Birchal1 Model
Dose Rate (Fan)
Conversion Factor:
92 + 180 fn (mSv/WLM).
-------
TABLE 4. EFFECTIVENESS OF FAN-ION GENERATORS IN REDUCING MEAN BRONCHIAL LUNG
DOSE EQUIVALENT RATE (Har1ey-Pasternak Model)(7)
Left Section of
Chamber
Treatment Conditions
Experiment
Number
1
2
3
Average
(Dose
Fan
On
1.01
1.23
1.02
1.09
Rates in Sv/y)
Fan + Ion
Generator On
0.216
0.281
0.270
0.256
Percent
Reduction
79
77
74
77
R
ight Section of
Chamber
Treatment Conditions
(Dose
Fan
On
1.20
1.40
1.11
1.24
Rates in Sv/y)
Fan + Ion
Generator On
0.245
0.322
0.329
0.299
Percent
Reduction
80
77
70
76
Notes:
1. The external air exchange rate was 0.2/h.
2. Percent = Dose Rate (Fan) - Dose Rate (Fan + Ion Generator)
Reduction Dose Rate (Fan)
3. Harley-Pasternak Model Conversion Factor: 70 + 795 f (mSv/WLM).
(100%)
-------
TABLE 5. EFFECTIVENESS OF FAN-ION GENERATORS IN REDUCING MEAN BRONCHIAL LUNG
DOSE EQUIVALENT RATE (Jacobi-Eisfeld Model)(6)
Left Section
Treatment Conditi
(Dose Rates in Sv
Experiment
Number
1
2
3
Average
Fan
On
0.935
1.14
1.02
1.03
Notes:
1. The external
2. Percent
of Chamber
ons
A/1
Fan + Ion Percent
Generator On Reduction
0.124
0.169
0.150
0.148
air exchange
Dose Rate (
87
85
85
86
rate was 0.2/h.
Fan) - Dose Rate
R
iqht Section of
Treatment Conditions
(Dose Rates in Sv/y)
Fan
On
1.02
1.25
1.04
1.10
(Fan
Fan + Ion
Generator On
0.144
0.193
0.183
0.173
+ Ion Generator
Chamber
Percent
Reduction
86
85
82
84
.1 (1005M
3.
Reduction Dose Rate (Fan)
Jacobi-Eisfeld Model Conversion Factor: 66 + 1370
f (mSv/WLM).
-------
REFERENCES
1. Rudnick, S. N., Hinds, W. C., Maher, E. F., and First,
M.W., "Effect of plateout, air motion, and dust removal on
radon decay product concentration in a simulated residence,"
Health Physics. 45: 463, 1983.
2. George, Andreas C., "Measurement of the uncombined
fraction of radon daughters with wire screens," Health
Physics. 23: 390, 1972.
3. Maher, E. F., Rudnick, S. N.. and Moeller, D. W.,
"Effective removal of airborne 222Rn decay products inside
buildings," Health Physics. 51: 356, 1987.
4. Maher, Edward F., "The control and characterization of
radon decay products in residences," Doctoral Thesis, School
of Public Health, Harvard University, Boston, MA, 1985.
5. Rudnick, S. N. and Maher, E. F., "Surface deposition of
222Rn decay products with and without enhanced air motion,"
Health Physics. 51: 283, 1986.
6. Jacob! W. and Eisfeld K., "Dose to tissues and
effective dose equivalent by inhalation of radon-222, radon
220, and their short-lived daughters", Report Gesellschaft
fur Strahlen- und Umweltforachung GSF-S0626, Munich,
Germany, 1980.
7. Harley, N. H. and Pasternack, B. S., 1982, "Environ-
mental radon daughter alpha dose factors in a five-lobed
human lung", Health Physics. 42: 789, 1982.
8. James, A. C. "Dosimetric approaches to risk assessment
for indoor exposure to radon daughters", Radiation
Protection Dosimetry. 7: 353, 1984.
-------
D-V1I-6
RADON MITIGATION TECHNIQUES FOR NORWEGIAN HOUSES
l i 2
Bjom Lind , Terje Strand and J0m T.Brunsell
1 National Institute of Radiation Hygiene, Osteras, Norway
2 Norwegian Building Research Institute, Oslo, Norway
ABSTRACT
Results of an experimental study of different techniques to reduce the radon levels in
Norwegian houses are reported. A total sample of 30 detached houses, representing a broad range
of different types of Norwegian houses accordring to construction techniques and architecture, were
selected for the study. In this sample of houses, earlier measurements had shown average radon
concentrations ranging from about 500 Bq/m> to 15 000 Bo/m>. Passive measurements of radon
were performed at different stages in the process. In a few houses more extensive measurements
were carried out This included contineous measurements of radon, radon daughters and air
exchange rate, and grab sampling measurements of unattached fraction of radon daughters. Of the
different solutions studied, the methods based on sub-slab ventilation/depressurisation were found
to have the best reduction effect on the radon level in indoor air.
INTRODUCTION
In 1988, an extensive study on mitigation techniques for Norwegian houses was started. The
study was organised as a collaboration between the Norwegian Building Research Institute and the
National Institute of Radiation Hygiene. The aim of the study was to evaluate between different
possible mitigation methods/techniques to reduce the radon level for different types of Norwegian
houses and to give recommendations.
So far, most of the research work on mitigation methods/techniques in the Nordic countries
have been carried out in Sweden (1,2). Most of the recommendations in Norway have been based
on experience from there. However, the buidling stocks are quite different The most important
difference, as far as radon is concerned, is the fact that most detached houses in Sweden do not
have a basement, while probably more than 80% of detached houses in Norway do have.
-------
MATERIAL AND METHODS
SELECTION OF HOUSES
In Norway there is about 1.6 million dwellings. According to census data from the Central
Bureau of Statistics (3), about 85% of the Norwegian housing stock is detached and undetached
houses. In the most rural municipalities the percentage is nearly 100%. However, in Oslo, the
largest town in Norway, about 65% of the dwellings are in blocks of flats. Influx of radon from the
ground is found to be the most important source of indoor radon in detached and undetached
houses. In houses where the radon concentration is found to be considerably higher than normal,
the contribution from other sources can usually be neglected. For dwellings in blocks, on the other
hand, building materials are usually the main source of indoor radon. However, in most cases the
radon concentration is very low compared to detached and undetached houses. Therefore, blocks of
flats were not included in the experimental sample of our study.
Most of the houses were selected from earlier and recent measurements from the National
Institute of Radiation Hygiene. It was important to get a, as far as possible, representative sample
of typical houses according to construction technique, architecture, age of the house and geology. It
is assumed that the final sample of 30 houses meets these requirements fairly good. In table 1, the
houses is classified into three categories according to type of house and radon concentration. As
shown, most of the houses have basements.
TABLE 1. THE SAMPLE OF DETACHED HOUSES.
NUMBER OF HOUSES
1)
Radon concentration (Bq/m»)
< 1000 1000-5000 > 5000
With a basement
floor below 14 11
ground level
Concrete slab
directly on ground 2
On crawl space 2
^Representative averages of radon concentration for each house.
THE MEASUREMENTS
Passive measurements of radon were performed at different stages in he mitigation process in
all of the 30 houses. The ETB-method, which is a combination of activated charcoal and TLD (4),
were used in the measurements. The integration time in the measurements was between 7 and 10
days. In order to minimize the influence on the measurements from short term variations in the
radon level, it was essensial that the they were performed at about equal meteorological/-
climatical conditions. The uncertainties in the individual measurements were estimated to be about
15% at the 95% confidence level.
-------
In a few selected houses, more extensive measurement programs were performed. This
included contineous measurements of radon and radon daughters and measurements of the air
exchange rate. At the diagnostic stage in the process, grab sampling measurements of radon were
found to be especially valuable in characterising the source. The method and equipment has been
described in an earlier paper by Stranden and Berteig (5).
RESULTS AND DISCUSSION
Generally, mitigation techniques may be divided into four groups:
(1) removal of radon sources
(2) methods to reduce the radon influx from the ground
(3) methods to increase the air exchange rate
(4) removal of radon daughters by electric fields/electrostatic
filter
Methods involving a removal or exchange of source material from the building site and/or
near the house may have a considerable reduction effect on the radon level in future houses if it is
done at the site before the buidling work is started. For existing houses this kind of solution will
usually be too expensive and too disruptive. However, for houses with a basement floor below
ground level, where alum shale or other materials with a very high exhalation rate of radon has
been used as filler on the outside of the foundation wall, removal or exchange of the source may
be a successful solution.
If the air exchange rate is at a normal level, the relative reduction effect on the indoor level
of radon by increasing the air exchange rate will usually be very low. Energy saving aspects,
climatic conditions and indoor comfort limits the possibility to increase the air exchange rate in
most situations. Due to an increase in the energy costs, and the campaign for saving energy in the
70*s and 80"s, a large proportion of old houses have been retrofitted in the last decade. At the same
time period, modem houses have been made energy efficient. According to studies in Sweden (6),
this may have reduced the air exchange rate in a significant part of the housing stock. In tight
houses with no mechanical ventilation systems, average air exhange rates down to 1/10 of the
normal level have been reported. In such houses, the effect of an increase in the air exchange rate
to normal values, may lead to a reduction in the radon concentration by an order of magnitude. It
is, however, important that the ventilation systems do not increase the underpressure inside. If so,
then the influx of radon from the ground may increase.
For most houses in our study, diagnostic measurements showed that influx of radon from the
ground was the most important source of radon. In 28 out of the 30 houses, different methods to
reduce the influx from the ground were chosen. One of the main objectives of our study was to
quantify the effect of different low cost solutions. This of course, did have an impact on the
individual choice of solution. In several houses the mitigation process had to be carried out in
different steps, starting with the cheapest possible one. In table 2, a general view of the different
solutions is shown. In the following, a few examples from our study will be discussed in more
detail.
-------
TABLE 2. CLASSIFICATION OF THE DIFFERENT MITIGATION TECHNIQUES
IN OUR STUDY
Category of mitigation solution Number of houses
Sealing of cracks and openings
in the foundation wall and the slab 3
Changing the pressure differences
between indoor air and the ground 1) 20
Changing the ventilation conditions 2
A combination of
different solutions 2) 5
1) Including four different methods: pressurize basement and/or crawl
space (5 houss) sub-slab ventilation (13 houses), outdoor well (2 houses)
and wall ventilation (1 house).
2) Combinations of sealing and sub-slab ventilation.
In figure 1, the reduction effect of a step-wise solution is illustrated in a house with a
basement. The first step was based simply on sealing visable cracks and openings in the concrete
slab and the foundation wall. In one of the rooms in the cellar, the average radon level was
reduced from about 4,000 Bq/m' to 800 Bq/mV In this room a higly permeable wall of light
expanded clay aggregate blocks, with direct contact below the slab, were assumed to be an
important entry route. Most of the reduction effect can probably be explained by the plastering of
this wall. However, the radon level was still too high and further actions were found necessary. As
a second step, a sub-slab ventilation system (7) was installed. Due to the small base area of the
house (about 80 m3) and considerations about the permeability of the ground, one central exhaust
pipe was assumed to give a sufficient suction. The pipe was connected to a fan on the loft. The
flow rate was measured to 50-60 mVh. These type of solutions reduces the influx of soil gas by
reducing the pressure differential between the basement and the ground and by reducing the
concentration of radon below the slab by ventilation. After this second step, the radon
concentration was reduced to about 150 Bo/m» in the basement and to about 100 Bq/m* on the
first floor. On the average, in the four rooms where passive measurements were performed, the
radon concentration was reduced by more than 90%. The actions were concluded to be very
successful.
-------
Bq/m
4000 -i
3000
/I
^| Before action
| ';. | After action step 1
•1 After action step 2
2000-
1000
Room 1 Room 1 Room 2 Room 3
FIRST FLOOR BASEMENT
Figure 1. A two-step solution in a house with basement. Step 1: Sealing of cracks and openings in
the concrete slab and the foundation wall. Step 2: Sub-slab ventilation (one exhaust pipe).
Figure 2 illustrates the effect of a pure solution based on sub-slab ventilation/-
depressurisation. In this example, two exhaust pipes were connected to the fan on the loft The fan
is identical to the one in the previous case. For one of the rooms in the basement, the radon level
was reduced from about 2,700 Bq/m3 to below 100 Bq/m3. This is a reduction of nearly 97%. The
average radon concentration in the main living room on the first floor was reduced from
840 Bq/m' to 60 Bq/m3. The solution was concluded to be very successful.
Bq/m
3000
2000
1000-
Room 1
FIRST FLOOR
Room 1
Room 2
BASEMENT
Room 3
Figure 2. The effect of sub-slab ventilation/depressurisation (two exhaust pipes)
in a house with basement.
-------
It is important that the capacity of the fan is as low as possible. Not only because of energy
saving, but also due to problems in the cold winter months if large quantities of cold air are
beeing sucked beneath the basement floor. Outdoor temperatures below -30^2 in the coldest winter
months are not unusual in several parts of Norway. Our results have shown that fans with a very
low capacity (about 30 Watt/60 mVh at Ap about 100 Pa) may be sufficient to solve the radon
problem in many situations.
In one of the houses, the radon level was found to be very high. From contineous
measurements, short time concentrations of radon up to 190 000 Bq/m3 were found in the
basement. Grab sampling measurements of radon and radon daughters confirmed these results.
From passive measurements, the radon level in the basement was found to be between 10 000 and
15 000 Bq/m». Grab sampling measurements showed that an old and dry floor drain (without a
water trap) in the basement floor probably was one of the main entry points. This floor drain was
directly connected to alum shale rich soil. Since the floor drain was not in use any more, the first
step was then to glog it and try to seal all visable cracks and openings in the concrete slab and the
foundation wall. The results of this first step is illustrated in figure 3. In one of the rooms in the
basement the radon level was reduced by almost 80%. In the other room in basement, sealing of
the floor was found very difficult due to presence of numerios cracks and openings in the slab. In
this room the average radon concentration was found to increase by about 15%. In addition to
further sealing of the floor and the walls, an almost identical sub-slab ventilation/-
depressurisation system as the example in figure 2, was installed. As illustrated, the radon
concentration was reduced by nearly 95% in the basement and by about 90% on the first floor.
However, the radon level is still too high and further work is needed. One of the possibilities is to
increase the flow rate capacity of the fan.
Bq/m
20000 A
16000
12000
8000
4000
Before action
After action step 1
After action step 2
Room 1 Room 1 Room 2
FIRST FLOOR BASEMENT
Figure 3. A two-step solution in a house with a basement. Step 1: glogging of a floor drain
in the basement floor and sealing of different cracks and openings in the concrete
slab and the foundation wall. Step 2: Sub-slab ventilation (two pipes).
-------
The effect of sub-slab ventilation in the other houses where this type of technique as studied,
was generally found to be successful. However, in a couple of houses were the ground most
probably has a very low permeability, and the sealing of different cracks and openings in the slab
were found to be very difficult, there was only a very slight reduction in the radon level.
For two of the houses in our study, the effect of a soil gas ventilation well on the outside of
the foundation wall, were studied. The effect for one of these houses is ullustrated in figure 4. The
well was at 3m depth and 1m in diameter. The exhaust pipe (diameter and lenght of O.lm and 3m
respectively), which was placed in the centre of the well reaching below basement floor level, was
connected to a small fan (28 Watt) on the top. The flow rate was measured to about 50 m»/h. In
one of the rooms closest to the well in the basement, the radon concentration was reduced from
1300 Bq/ms to slightly below 350 Bq/ms. On the first floor where the living room, bedrooms and
kitchen were located, the radon level was reduced from 420 to 240 Bq/m».
Bq/rrT
1500
1200-
Room 1
FIRST FLOOR
Room 1 Room 2
BASEMENT
Figure 4. The effect of an outdoor soil gas ventilation well in a house with a basement.
In five houses, different techniques to pressurize the basement and/or the crawl space by
supply of fresh air from the outside, were investigated. Due to the climatic condtions it was found
necessary to include a heater unit in the fans, which of cource increases the energy costs.
Preliminary results from these exsperiments seems to show that these types of solutions may have a
relatively slight reduction effect on the radon level. In figure 5, the effect of pressurizing the
basement and the crawl space in a house on alum-shale ground is illustrated. The fan (99 Watt)
was installed in the foundation wall at the basement level. The flow rate was measured to
310 m'/h. The radon concentration in the basement and the first floor were reduced by 50% and
30%, respectively. One of the main explanation of this low reduction effect may be leakage
between the basement and the first floor.
-------
Bq/nrT
First floor
Basement
Crawl space
Figure 5. An example of a solution based on pressurizing the crawl space and the basement.
CONCLUSIONS
From the results and discussion of our study the following conclusions may be drawn:
1. Sealing of viseable cracks and openings in the slab and the foundation wall is recommended as
a first step in any mitigation process. However, the results of our study shows that sealing
alone in most situations will not be sufficient to solve the entire radon problem. This is
especially true in cases where the radon level is very high.
2. Sub-slab ventilation by exhaust pipes connected to a small fan were generally found to be the
most successful technique in our study. In most cases it is sufficient with only one or two
exhaust pipes and a very small fan (30 Watt/60 m'/h).
3. Preliminary results of our exprimental studies of techniques based on pressurizing crawl
space and/or basement by supply of outdoor air have shown to have a relatively low reduction
effect on the radon level.
The final results of our study will be presented later.
-------
ACKNOWLEDGEMENTS
This work was supported by the Royal Norwegian Council of Scientific and Industrial
Research (NTNF), the Ministry of Local Government and Labour, Ministry of Petroleum and Energy
and Ministry of Health and Social Affairs in Norway. The authours wish to thank Miss A.K.Kolstad
at the National Institute of Radiation Hygiene for her assistance during the measurements.
The work described in this paper was not funded by U.S. Environmental Protection Agency
and therefore the contents do not necessaryli reflect the views of the Agency and no official
endorsement should be inferred.
REFERENCES
1. National Institute fo Radiation Protection, National Board of Health and Welfare, National
Board of Physicsl Planning and Building. Radon in dwellings, interim report 1987. Report
SSI 87-17, National Institute of Radiation Protection, Sweden (in Swedish).
2. Ericson, S-0, Schmied, H. and Clavensjo, B. Modified technology in new constructions
and cost effective remedial action in existing structures to prevent infiltration of
soil gas carrying radon. Radiat.Prot.Dosim. 7(1/4): 223-226, 1984.
3. Central Bureau of Statistics, Census data, Personnal Communications 1986.
4. Stranden, E., Kolstad, A.K. and Lind, B. The ETB dosemeter, a passive integrating
radon dosemeter combining activated charcoal and TLD. Radiat.Prot.Dosim. 5(4): 241-245,
1983.
5. Stranden, E. and Berteig, L. Radon daughter equilibrium and unattached fraction in
mine atmospheres. Health Phys. 42: 479-487, 1982.
6. Swedjemark, GA The equilibrium factor F. Health Phys. 45(2): 453-462, 1983.
7. Sanchez, D.C. and Henschel, D.B. Radon reduction techniques for detached houses -
technical guidance. U.S. Environmental Protection Agency Report EPA/625/5-86/0019,
Center for Environmental Research Information, Cincinnati, OH, 1986.
-------
Session D-IX:
Radon in Schools and Large Buildings—POSTERS
-------
D-IX-1
PREDICTION OF MAXIMUM RADON CONCENTRATIONS IN SCHOOLS
USING PARTIAL SAMPLING METHODS
by: William Belanger, P.E.
U. S. Environmental Protection Agency, Region III
Philadelphia, Pennsylvania 19107
Michael Pyles
Pennsylvania Bureau of Radiation Protection
Harrisburg, Pennsylvania 17120
ABSTRACT
The Environmental Protection Agency published interim
guidance for radon measurements in schools in April of 1989.
That guidance recommends testing of radon in all rooms that are
frequently used. This paper describes a partial sampling
technique which might help identify schools with a potential
radon problem. Its use is not endorsed by EPA. It is an
analytical tool which might be used as an initial step in
facilitating implementation of the 1989 guidance.
Radon measurements were obtained for 198 schools in
Pennsylvania and Virginia. Schools in Pennsylvania were
generally located in the Reading Prong and were surveyed by the
Pennsylvania Department of Environmental Resources. In each
school, multiple rooms were measured using either charcoal
adsorption, alpha track or grab working level. Virginia schools
were surveyed by the Fairfax county, VA Public School System
using charcoal samplers. These data were analyzed using a Monte
Carlo method to determine the ability of partial sampling and
statistical methods to predict the highest room in the school. A
method was developed which yields reasonable success in detecting
the probable presence of elevated rooms (above 4 pCi/1) when ten
percent (or a minimum of ten) rooms are sampled. The confidence
in the prediction (type a and type b errors) is quantified as a
function of the parameters of the analysis. The implications of
potential errors induced by partial sampling are discussed.
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.
-------
PREDICTION OF MAXIMUM RADON CONCENTRATIONS IN SCHOOLS
USING PARTIAL SAMPLING METHODS
INTRODUCTION
On April 20, 1989, the Environmental Protection Agency (EPA)
published preliminary guidance for monitoring of schools in the
United States (1). This guidance recommends measurement of radon
in all rooms that are frequently used and are on or below ground
level. Frequently used rooms include classrooms, offices,
cafeterias, libraries and gymnasiums. Rooms used for storage,
broom closets, etc. would not be monitored. This recommendation
is based on observations in schools in an EPA pilot study(2).
While EPA clearly prefers measurement of every room, the
Agency recognizes that budgetary and other constraints may
prevent this, at least in the immediate future. School officials
who cannot measure all frequently used rooms are then encouraged
to test those rooms with the highest potential for elevated
radon. These include (1) basement classrooms, (2) occupied rooms
that are isolated from the central ventilation system or on
systems which only recirculate room air, (3) rooms on or near
structural joints such as adjacent slabs, (4) rooms with a large
floor/wall joint perimeter and (5) rooms that have floor slabs
with significant cracks. The intent of this guidance is to
maximize the probability of detecting rooms with elevated radon
concentrations.
This paper reports a technique which may be useful in
finding schools having rooms with elevated radon concentration
when it has been determined it is not possible to test every
room. The technique is not endorsed by the Environmental
Protection Agency. The method described here makes use of the
properties of the distribution of radon concentrations measured
in schools. This distribution is used to predict the
concentration in the room with maximum radon based on a partial
sample.
The method could be applied using a sample of rooms chosen
on a purely random basis, or it could be used to evaluate the
remainder of the rooms in a school after the rooms with highest
radon potential have been selected as described above. It could
also be used as a screening tool to decide the order in which
several schools should be tested more completely if it is not
possible to test all at once. The method yields an approximation
of the highest radon concentration in the school based on a
sampling of ten percent, or a minimum of ten rooms.
-------
DESCRIPTION OF THE METHOD
In radon screening procedures, one would prefer that there
be a low probability of failing to detect rooms with elevated
radon (false negatives) because potential health problems may go
uncorrected. Incorrectly classifying a school as having elevated
radon based on a preliminary screening procedure (false positive)
may be more acceptable for it provides a more conservative
approach in protecting public health.
The basis of the partial sampling technique is that the
properties of a population (for example the radon concentration
in the rooms of a school) can be estimated from a sample drawn
from the population. Given a sample of radon measurements taken
in some randomly selected rooms of a school, it should be
possible to devise a method to estimate the distribution of radon
in all the rooms within certain limits of uncertainty. It may
also be possible to arrive at an estimate of the maximum radon
level in the school based on empirical data. That is exactly
what is attempted here.
For populations where the property being measured is
normally distributed and the sample is chosen randomly, the mean
and standard deviation of a sample yield an unbiased estimate of
the mean and standard of the population. If the parameter is not
normally distributed, the mean and standard deviation of the
sample are not unbiased estimators of the population. Therefore,
in order to utilize the mean and standard deviation as
estimators, the population must be normally distributed. For the
radon in schools data used here, the radon concentrations have
been found not to be normally distributed. Most of these schools
show an acceptable (p=.05) fit to a lognormal distribution, so
lognormal statistics must be used. The most straightforward way
to do this is to transform the radon data by taking the log
(either natural or base 10) of the radon concentrations(3). The
resulting data yield an acceptable fit to a normal distribution.
When the distribution of radon concentrations is known, it
is straightforward to estimate the radon concentration at any
percentile of the distribution. Conversely, if a percentile is
chosen the radon concentration at this percentile can be
estimated. The confidence limits of these estimates are a
function of the degree of certainty with which the parameters of
the distribution have been estimated, and how well the
distribution conforms to the model (in this case lognormal). The
allowable confidence limits can also be used to determine the
necessary sample size.
However for the purpose of this paper it was desired to
estimate the radon concentration in the maximum room in each
school, not the radon at a fixed percentile. Since the number of
rooms is variable, there is not a fixed percentile of the
distribution that can be used. Instead it was decided to use an
empirical approach to relate the maximum observed radon
-------
concentration to the mean and standard deviation of the measured
radon concentrations in each of a large number of schools. A
similar empirical approach, described below, was used to
demonstrate the resulting confidence limits.
Similarly, the number of rooms in the sample was not derived
from sampling theory. It was based on the authors' estimate of
what is practical. It was felt that a minimum of ten rooms, or a
minimum of ten percent of the total number of basement and first
floor rooms, might yield an estimate of the mean and standard
deviation sufficient for the purpose at hand, an estimate of the
radon concentration in the maximum room in each school. This
approach is tested empirically using a Monte Carlo method and the
resulting confidence limits are shown for each school
individually. The number of rooms in the sample was varied and
found not to be an important parameter in the schools with the
"worst" predictions. The overview of all schools for which this
calculation was done demonstrates the strengths and weaknesses of
the method.
The mathematical form of the test is
(1) Xn + K sn < C pass
and (2) Xn + K sn > C fail
where Xn is the mean of the sample.
sn is the standard deviation of the sample.
K is a multiplier (number of standard
deviations) determined empirically to
yield the desired sensitivity.
The method proposed here makes use of the technique outlined
above, and has been structured to be of practical benefit. Many
choices in this candidate method are purely arbitrary and based
on practical considerations. For example a method requiring 90%
sampling is of little practical use since any cost savings would
be more than offset by the expense of confirmatory monitoring if
a problem is detected. A sampling rate of 10% was chosen to
yield a tangible benefit, and a 25% rate was also investigated
and found to yield little improvement over a 10% rate.
The environment within which the method would be applied is
as follows. If the estimate of the maximum radon is below 4
pCi/1 school officials may test further if they so decide, but
EPA believes it may be costly at this point to reduce the radon
below these levels. If additional monitoring were performed it
would be a relatively low priority compared to schools yielding a
higher estimate. An estimate of the maximum room above 4 pCi/1
would trigger confirmatory monitoring of the remaining rooms. It
would not be the basis for mitigation. There would be two ways
in which a radon problem could be "found". It could be directly
-------
measured if one of the radon tests came out above 4 pCi/1 or
could be inferred from the statistical calculation of the highest
room. This yields two ways in which problems can be found and
increases the power of the test. If rooms with a high radon
potential are also preselected for separate monitoring, the power
of the method is further improved.
The first condition that must be satisfied is normality of
the data distribution. Radon data in the schools here was found
to be approximately log-normal, and so can be normalized by
simply taking the logarithm. All data used in this analysis has
been transformed in this way. This allows the use of the sample
mean and standard deviation to estimate these parameters for the
population. It is unimportant whether natural or base 10
logarithms are used here since the difference is only a scaling
factor which cancels out in the later mathematics.
The second and more tenuous part of the analysis is to
estimate the radon concentration in the worst room. To do this
an empirical approach was selected. The logarithms of radon
concentrations were calculated from a subset of the schools
described below. The mean and standard deviation of the
logarithms were calculated for each school. The logarithm of the
maximum radon concentration was also calculated for each school.
The number of standard deviations from the log mean to the log of
the maximum observed radon concentration was then calculated for
each school. For convenience this resulting statistic will be
called K1 because of its use in selecting a value of K used in
estimating the maximum radon as described above. The values of
K1 were compiled and graphed separately for each method of
measurement used in Pennsylvania, for the Virginia charcoal data
and for data from the EPA phase I study of schools nationwide.
There was a remarkable similarity in the results in spite of
sampling methods and sample durations and radon levels which were
quite dissimilar.
The derived values of K1 are plotted on Figures 1 through 5.
It can be seen that there is a distribution of this statistic
with a maximum at K1 = 2 to 2.5 standard deviations from the
mean. This relationship appeared to hold well across the data
examined here. This is the basis for the choice of the value of K
in the predictive applications of the method. Note that the
maximum value of the statistic K1 is 4.5 standard deviations from
the mean and several schools yielded K' of 4 standard deviations.
This did not appear to be a function of the number of rooms in
the school, but was related to the presence of rooms with unique
characteristics that caused high radon. Given this distribution
of K1 the choice of K at 2.0 or 2.5 in the method above would
yield a central value in the prediction of the highest radon in
the school. The presence of schools with K1 as high as 4.5
guarantees that the method cannot predict all rooms above 4 pCi/1
all the time unless K is set at or above 4.5. The presence of
schools with K1 close to 1 will result in overprediction part of
-------
the time. The probability if overprediction and underprediction
at various values of K is the principal result of this paper.
NAT I ON A L S C HOOL S
EPA SURVEY
1 1.3 2 2.9 3 3.3
NUMBER OF LOG STD DEVIATIONS TO MAX
Figure l
e
e
7
6
9
4
3
2
VIRGINIA SCHOOLS
CHARCOAL
I
O.SO 1 DO 1.SO 2 OO
NUMBER OP LOG
2 SO
sro
3 00 3. SO
DEVIATIONS TO MAX
SO S.OO
Figure 2
-------
PENNSYLVANIA SCHOOLS
20
19 -
18 -
17 -
1G -
15 -
13 -
12 -
11 -
10 -
9 -
a -
7 -
B -
5 -
4 -
3 -
2 -
1 -
GRAB WORKING LEVEL
0 500 1 DOO 1 SOO 2 000 2 500 3 000 3 500 4 000 4 500 5 000
NUMBER OF LOG STD DEVIATIONS TO MAX
Figure 3
PENNSYLVANIA SCHOOLS
CHARCOAL
3 -
2 -
1 -
0 50 1 00 1 50 2 00 2 50 3 00 3 50
NUMBER OF LOG STD DEVIATIONS TO MAX
4 00
4 50
5 00
Figure 4
-------
PENNSYLVANIA SCHOOLS
ALPHA TRACK
5 -
3 -
2 -
0 50 1 00 1 50 2 00 2 50 3 00 3 SO 4 00 4 50 5 00
NUMBER OP LOG STD DEVIATIONS TO MAX
Figure 5
DATA USED FOR THE ANALYSIS
Radon measurements were obtained for eighty-seven schools in
Virginia which were measured by the Fairfax County Virginia
Public School System using bag-type charcoal detectors. The
Environmental Protection Agency Phase I school data(4) for nine
other areas of the country was also obtained. The radon in this
dataset was measured with EPA charcoal canisters. Radon data was
also obtained from the State of Pennsylvania for schools measured
by the State as a part of its Reading Prong investigation.
Seventy-nine schools were measured using three month alpha-track
detectors, thirty-two schools were measured using charcoal
detectors and fifty-five schools were measures with grab working
level samples. At the time when these measurements were made,
the grab working level was the principal tool for doing rapid
radon assessments. In schools where elevated working levels were
found in "high probability" rooms as defined by Pennsylvania
(this may differ from the later EPA definition), every room in
the school was tested. Where no elevated working levels were
found, only a few rooms were tested. This data sometimes
included boiler rooms which are infreguently inhabited. School
grab samples were also followed up with alpha track detectors, so
many of the working level and alpha track schools are common
between the datasets. Charcoal tests were used later in the
Pennsylvania program, and were made in a separate group of
schools. For this analysis the grab samples and alpha track
-------
results are treated as if they were not from the same schools
because it was desired to test the method with both measurements.
This did not affect the outcome of the analysis since the number
of rooms tested with alpha track detectors was below the lower
limit included in the Monte Carlo analysis but allowed the value
of K1 to be investigated for this measurement method.
A preliminary culling of data was performed to eliminate all
schools where less than ten rooms were sampled. This was done
because the first part of the analysis examined the mean and
standard deviation of the logs of the concentrations. It was
felt that the statistics would only be meaningful where there
were at least ten items of data. This also served to eliminate
schools where very few rooms were sampled. An additional
restriction placed on the data was that the maximum radon
concentration be at least 2.7 pCi/1 or .014 WL. This was done to
assure that the radon measurements were well above the lower
limit of detection so that there would be some confidence in the
radon measurements reported. After this first cut there were
twenty-nine schools in the Virginia charcoal dataset, eleven
Pennsylvania schools measured with alpha-track detectors, seven
Pennsylvania schools measured with charcoal samplers and thirty-
three Pennsylvania schools measured using a working level grab
sample. This also left 110 schools in the Phase I national
dataset which were measured with EPA charcoal canisters. Further
culling was done for the Monte Carlo analysis below.
MONTE CARLO ANALYSIS
Monte Carlo analysis is a technique that allows analysis of
data and distributions which are not regular or which are not
easily manipulated using theoretical techniques(5). At this
point it is important to recognize that the processes which
determine the radon concentration in a school room are, for the
most part, not random. Factors which determine the radon level
include the radon concentration in the soil gas, the ease with
which radon can move into the room, and the air pressure that
drives the radon. Because the dominant factors governing the
radon concentration are not random, there is no good reason to
expect that the radon concentrations will exactly fit any of the
commonly used statistical distributions. The observed fit to the
lognormal distribution allows the data to be characterized using
the geometric mean and geometric standard deviation, but does not
allow an exact analysis of the real-world situation. In this
case, Monte Carlo techniques allowed assessment of the
performance of the proposed radon analysis method using real
radon distributions rather than an idealized model.
The Monte Carlo simulation discussed below was restricted to
schools where more than fifty rooms were sampled. This was done
to improve the validity of the results, and because the
utilization of the method proposed here will most likely occur
where there are large buildings to measure. This selection of
data also removed the schools from the dataset where all rooms in
-------
the school were not sampled. This selection of schools resulted
in a Monte Carlo analysis of twenty-nine Virginia schools
measured with charcoal and four Pennsylvania schools measured for
grab working levels. No schools in the Pennsylvania dataset had
50 rooms measured with charcoal or alpha track detectors.
Because the Pennsylvania schools with all rooms measured were
selected because of high working level measurements, this allowed
examination of schools with very high radon concentrations as
well as those with only moderate levels.
The selected schools had radon measurements in at least 50
rooms. From these populations, samples of ten rooms each were
selected at random without replacement (no room could be selected
twice in the same sample). One hundred of these random samples
were generated for each school. This is as if the school had
actually been sampled 100 times, each time with a different
selection of rooms.
For each sample, the mean and standard deviation of the
logarithms of the radon concentrations were calculated. For each
sample, the mean was added to the standard deviation multiplied
by 0, 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, and 5.0. This
is in accord with the form of the test described above, and the
multipliers are the values of K to be tested. The antilog of the
result was taken for each sample and compared to 4 pCi/1 or .02
WL. As the number of standard deviations increases, there is a
greater probability that a result above 4 will be calculated.
Since there are 100 different samples for each value of K,
the number of times the result is above 4 pCi/1 or .02 WL
represents the probability, p, (in percent) of a predicted
maximum in excess of the selected limit at that school (a "hit").
This result is specific for that value of K. This analysis
utilizes the real observed radon distribution in the school, and
the result varies from school to school. For schools with no
actual concentrations above 4 pCi/1 or .02 WL, the number of
"hits" represents the probability of a false positive at that
school. For schools with measured rooms above 4 pCi/1 or .02 WL,
the complement of the probability of detection (1-p) represents
the probability of a false negative. This data is presented in
graphic form for a number of schools is in figures 6 through 8.
Figure 6 (school 4151) is typical of a school with no rooms
above 4 pCi/1, but selected to have at least one room above 2.7
pCi/1. Because this school had no rooms above 4 pCi/1, the chart
shows the probability of a false positive. Note that for K =
1.5 and below, none of the 100 trials resulted in a predicted
maximum greater than 4 pCi/1. For K = 2.5 there were 34 hits,
giving a 34% probability of a false positive for this school for
this value of K. As K is increased to 3.0, there is a 67%
probability of a false positive. Several similar schools yielded
similar results. This suggests that K be chosen below 3.0 if the
false positive rate is to be kept below 50% for schools with the
highest room just under 4 pCi/1.
-------
s
u.
=
i
ft
e
11O
1OO -
so -
eo -
70 -
oo -
so -
•40 -
ao -
2G -
1O -
O
SCHOOL MONTE-CARLO ANALYSIS
SCHOOL -»131 - NO ROOMS »• •*
O O O O
O OO O SO 1 OO 1 SO a OO 2 SO 3 OO
NUMBER OP LOC STO DEVIATIONS - LOC
3 30
MEAN
i
SO 9 00
Figure 6
SCHOOL MONTE- CARLO ANALYSIS
£
^
i
11D
1OO
SO
SO
70
OO
SO
aa
3D
10
O
SCHOOL SO BO - 3M OP ROOMS «-
ooo
o.oo o.so i oo t.so a.oo a so 3.00 3.90
NUMBER OP LOG STO DEVIATIONS » LOG MEAN
I
.30 9.OO
Figure 7
Figure 7 (school 2086) is typical of a school with only a
few rooms slightly above 4 pCi/1. Only 3% if the rooms in this
school are above 4 pCi/1. Because this school actually has rooms
above the critical radon concentration, the chart shows the
probability of correctly classifying the school. Subtracting
these probabilities from one yields the probability of a false
negative. In this case, a K of 2.0 gave a probability of
detection of 28%, or a probability of a false negative of 62%.
For K = 2.5 the probability of a false negative is 40%. For K =
3.0 it is 20%. Several similar schools yielded similar results.
This suggests K be selected above 2.0 if marginal cases like this
-------
are to be detected with a reasonable degree of certainty.
Whether 40% probability of a false negative is acceptable in this
case is a policy decision beyond the scope of this paper. The
highest radon measured in this school was 5.4 pCi/1.
SCHOOL
FBKBIT PROBABILITY GF POSITIVE FIIOIIG
j
-lUU&UOOCDCDC
IOOOODOODOC
1 Ill
a o
0 . OO 0.30
1O
^
MONTE- CARLO ANALYSIS
SCHOOL 332-7 - 13* OF ROOMS *• •*
03
I
Bf
1
I
1QO
I
1OO 1OO 1OQ 1OO
I
1.00 1.SO Z.OO 2 SO 3 OO 3 SO 1 OO 1 . 3O 3 . OO
NUMBER OP LOG STD DEVIATIONS » LOG MEAN
Figure 8
Figure 8 (school 3327) is clearly elevated in radon with 13%
of rooms above 4 pCi/1. The highest radon concentration is 6.9
pCi/1. In this case, K = 2.5 gives a 97% probability of
detecting this problem. There was another similar school which
yielded similar results. This suggests the choice of about 2.5
for K yields a good probability of detecting schools with similar
radon concentrations. Several Pennsylvania schools with much
higher radon concentrations were included in the analysis. For
these schools, the probability of detection is virtually 100% for
K = 2.5.
There were four schools in the dataset which gave misleading
results. One school with no rooms above 4 pCi/1 yielded a
probability of a false positive of 93% for K = 2.5. Decreasing K
below 2.0 made only a slight improvement. This problem appears
to be due to a wide spread in radon concentrations which gave a
large standard deviation. The maximum radon was 3.7 pCi/1. It
is possible that measurement at a different time would result in
a room above 4 pCi/1. Three schools with 6%, 2% and 2.4% rooms
above 4 pCi/1 yielded probabilities of a false negative of 53%,
82% and 69% for K = 2.5. For these three schools, increasing K
above 3 did not result in a great improvement. This appeared to
be due to a large number of rooms with a similar radon
concentration which yielded a small standard deviation. In each
case there was a room which did not appear to fit the
distribution. For all four of these schools, increasing the
sampling rate to 25% of rooms with a minimum of 25 did not yield
any improvement. None of these rooms exceeded 10 pCi/1. It is
-------
the opinion of the authors that these rooms could not be detected
except by measuring them. No system based on the distribution of
the other rooms will detect them. Preselection of rooms based on
high potential for radon entry gives the best chance of finding
these rooms short of measuring every room, but depends heavily on
the ability to identify such rooms based on known physical
features. It was not possible to examine the effect of
preselection in this study because the physical characteristics
of the rooms which would allow preselection were not reported in
the data used.
RESULTS
Examining a number of schools allows the pattern of false
negatives and false positives to be determined at each multiplier
of the standard deviation. If acceptable levels of error can be
chosen, then a constant can be selected to yield that level of
error. Figures 9, 10, and 11 are a summary of all schools
analyzed with the Monte Carlo method.
i
s
UJ
H
8
$
t
£
S
tt
90 -
ao -
70-!
60 -
50 -
40 -
30 -
20 -1
I
10 -j
.i
c
METHOD SUMMARY - MEAN + 2 STD DEV
GREATER OF 10 ROOMS OR 101 OF ROOMS
O
D
O
I
§
0 a
i
1 % a
i
1 a Q
2 4 G B 10 12 14 1
PERCENT OF ROOMS ABOVE 4 PCI/L
B
Figure 9
Figure 9 shows the performance of the method with K = 2.0.
Each data point represents an individual school. The abscissa is
the percentage of rooms above 4 pCi/1 actually measured in the
school. For the purpose of this analysis these radon
concentrations are taken as truth, though in fact there are
errors associated with the measurement methods and due to
temporal variations. The ordinate is the probability of a "hit"
-------
with K = 2.0. For schools on the Y axis, there were no rooms
above 4 pCi/1, so the Y value is the probability of a false
positive. For all other schools, there were rooms above 4, so
the Y value represents the probability of correctly classifying
the school. (The point at 0,72 is one of the exceptions
mentioned above.)
METHOD SUMMARY - MEAN + 25 STD DEV
100 -
30 -
BO -
B 70-
* 60-
9 '
a so .
ft
GREATER OF 10 ROOMS OR 10* OF ROOMS
0
I
a
S
a
l
D
I D °
I
^ 9
| 3Q 7 ° °
10 -
D -
i°° "
] 2 4 6 8 10 12 -M 16
PERCENT OF ROOMS ABOVE 4 PCI/L
Figure 10
Figure 10 shows the results for K = 2.5. Note the increase
in the number of false positives and the increase in confidence
in detecting schools with elevated radon. The school at 6,47 is
one of the exceptions discussed above. Except for this one
school, there is a high confidence in detecting schools with more
than 4% of the rooms elevated.
Figure 11 shows the results for K = 3.0. Note the still
higher occurrence of false positives, and the further increase in
confidence in detecting elevated rooms. The school at 6,50 is
the same exceptional school. The probability of false positives
becomes high enough to be cause for concern at this value of K,
but it must be kept in mind that the lowest school included had
radon above 2.7 pCi/1 in at least one room. The schools with
very low radon have all been excluded from the dataset.
Inclusion of these schools would yield a smaller proportion of
false positives.
-------
METHOD SUMMARY - MEAN + 3 STD DEV
100 -|
90 -
3 "'
Ul
= 70 1
> '
1 «,-
fe 30-1
t
=i -,0-j
i
20 -
in -
c
GREATER OF 10 ROOMS OR 10K OF ROOMS
a
a
D
a
a
a
D a
-a
a
a
a
' 2 ' 4 ' B ' B ' 10 12 1" IB
PERCENT OF ROOMS ABOVE •* PCI/L
Figure 11
APPLICATION OF THE METHOD
The method might be applied as follows:
1. Choose rooms with high potential for high radon
concentration and measure them. This step is optional but very
desirable. It will identify obvious problems but will not
identify rooms which are "high" in radon for hidden causes.
2. Select a random sample from the remaining rooms of the
school. Measure ten percent of the rooms or ten, whichever is
greater. Ten percent was chosen because it would result in a
substantial savings over complete testing. The minimum of ten
rooms avoids very small samples. Choices other than ten percent
or ten samples are possible but have not been investigated here
to quantify the uncertainties.
3. Measure the radon in the selected rooms. Examine the radon
results for unacceptable concentrations which clearly indicate a
problem. If a radon problem is found in this way there is no
need for the statistical analysis below.
4. Calculate the logarithm (natural or base 10) of each
concentration in the random sample. Do not include the rooms
selected for high radon potential. Calculate the mean and
standard deviation of the log radon concentrations of this
-------
sample. Add "K" standard deviations to the mean. Calculate the
antilog of this value. This is the expected radon concentration
of the highest room in the population from which the sample was
drawn. The choice of "K", the number of standard deviations,
determines the sensitivity if the test. K = 2.0 yields fewer
false positives at the risk of greater chance of a false
negative. K = 3.0 yields greater sensitivity at the expense of
more false positives. A value of 2.5 for K seemed to give a good
balance, but choice of K should be made based on the error
probabilities which are acceptable to the decision-maker.
DISCUSSION
It can be seen from the results above that partial sampling
does not give one hundred percent certainty of identifying
schools with rooms above 4 pCi/1. Such a result is to be
expected. Even when every room is sampled, the errors inherent
in measurements and the variability of radon in time prevent
positive determination in marginal cases. When measurements are
made, uncertainties cannot be avoided. The method investigated
here can identify most schools with rooms above 4 pCi/1 with
partial sampling. The errors of commission and omission are
quantified here and are available for use in deciding what choice
of rooms to be sampled is appropriate. It is the task of school
administrators and policy-makers to decide what constitutes an
acceptable uncertainties. The uncertainties associated with this
method are similar in magnitude to the uncertainties in short-
term measurements.
The uncertainties associated with application of the method
can be reduced by preselecting rooms with a potential for high
radon concentrations. In this case, the rooms not preselected
could be sampled and analyzed using this technique. Since the
EPA guidance allows monitoring of only the high radon potential
rooms as an acceptable but less preferred option, use of this
method on the remaining rooms would be expected to yield an
improved result and represents a middle ground. Its use is not
precluded in the guidance nor is it endorsed by EPA. It is also
possible that school districts may not have the funds to measure
all rooms in a number of large schools at the same time. Using
the method outlined here, it is possible to conduct partial
sampling on all schools with the object of prioritizing the more
complete measurements. Schools with suspected problems might be
measured first and others would be deferred until funds are
available for measuring every room. Partial sampling can
identify schools at which we suspect the presence of rooms above
4 pCi/1, but the only way to identify specific rooms is to test
them individually.
The choice of 4 pCi/1 was made within the context of EPA
guidance at the time. The method can be applied equally well for
a cut point at another radon concentration, or expressed in
another unit. Note that in this case the method did not need to
be changed to accommodate a 4 pCi/1 limit and a .02 WL limit.
-------
To the extent that schools represent other large buildings,
this analysis may be transferable. There would have to be some
investigation of the proper choice of K, and the tolerable errors
may be somewhat different.
ACKNOWLEDGEMENTS
The authors wish to thank Harry Chmelynski of the Washington
Consulting Group for development of much of the theory upon which
this work is based and for supplying the Fairfax Schools data and
Thomas Peake and Maureen Clifford of the Environmental Protection
agency for review of the manuscript.
REFERENCES
1. Radon Measurements in Schools, EPA 520/1-89-010, U.S.
Environmental Protection Agency, Washington, DC 20460, 1989.
2. MacWaters, J. et al, S. Cohen & Associates, Inc., Me Lean,
VA and Chmelynski, H. et al, The Washington Consulting Group,
Inc., "Radon Measurement in Schools", Contract 68-02-4375 for
U. S. Environmental Protection Agency, July, 1988.
3. Gilbert, R. 0., Statistical Methods for Environmental
Pollution Monitoring. Van Nostrand Reinhold Company, Inc., New
York, NY, 1987.
4. MacWaters, J. et al, S. Cohen & Associates, Inc., Me Lean,
VA and Chmelynski, H. et al, The Washington Consulting Group,
Inc., "Final Report - School Radon Protocol Development Study",
Contract 68-02-4375, for U. S. Environmental Protection Agency,
September, 1989.
5. Hammersiey, J. M. and Handscomb, D. C., Monte Carlo Methods.
John Wiley & Sons, Inc, New York, NY, 1964
-------
D-IX-2
COMMERCIAL MITIGATION TECHNIQUES USED IN REMEDIATING
A 2200 pCi/L PUBLIC BUILDING
by: James G. Davidson
Radon Detection Services, Inc.
Ringoes, New Jersey 08551
ABSTRACT
In March of 1989 EPA and Pa. DER officials were amazed to
discover a school in Pennsylvania with levels in its library of
2200 pCi/L. The library was a 30 year old, three story slab
on-grade structure more like a commercial building than a typical
school structure. It had three separate and complex HVAC systems.
Initial diagnostics indicated radon levels under the slab at over
80,000 pCi/L. Further investigations revealed major entry routes
and a HVAC system terribly out of balance.
Remediation consisted of installing a complex sub-slab
depressurization system with an exterior commercical fan unit,
major entry route sealing, and working closely with a
mechanical contractor to bring the HVAC systems back into balance.
Initial post remediation testing showed a 99% drop in radon levels,
Refinements to the system are still in progress.
-------
As the EPA and various states begin urging the testing of all
schools nationwide; and, as laws are passed mandating the testing
of commercial buildings, facility directors will ultimately have to
grapple with radon remediation in complex buildings. Unlike
residential remediation where mitigation techniques are becoming
standardized, little work has been done in schools with complex
HVAC systems or in commercial buildings with hundreds of thousands
of square feet of office space. In dealing with existing commercial
buildings or with commercial buildings under construction, new
mitigation techniques are still evolving, as is the retrofitting of
residential techniques for commercial applications. The following
presentation is the result of a recent school mitigation in
Pennsylvania in which this building had the highest radon levels
found to date in a school or commercial building.
In March of 1988, a student attending Westminster Theological
Seminary in Philadelphia, Pennsylvania placed a charcoal canister
on his desk on the ground floor of the school's library. To his
amazement, the levels came back at over 2,000 pCi/L. He alerted
school officials who conducted theri own test with similar results.
School officials immediately contacted the Pennsylvania Department
of Environmental Resources who in turn contacted the EPA. Very
quickly, EPA officials from the regional office, Washington and
the research division in North Carolina converged on the scene to
perform diagnostics and confirmatory testing using an AB-5 Pylon
Radon Monitor. Their testing showed levels over 2200 pCi/L. Gas
samples from under the slab, peaked out at over 80,000 pCi/L.
Students were evacuated from the ground floor of the building and as
word of the problem leaked, newspaper and T.V. crews arrived at the
school. The story hit the nightly news and was in every local paper
the next day.
THE BUILDING
The library is a three (3) story 30,000 square feet slab on
grade building constructed about 30 years ago. The first floor
is partially built into a bank so entry is on both the first and second
floors, most of the building is open as is typical of libraries. Each
floor has its own HVAC system. The HVAC unit for the first floor is
on the first floor while units for the second and third floors are
located in a penthouse on the roof. There is a large exhaust fan in
the penthouse which is ducted to all of the bathrooms for a bathroom
exhaust. This fan is on a timer with a two (2) hour on/off cycle.
DIAGNOSTIC TESTING
While residential testing has on occassion revealed radon
levels in the 2,000 - 3,000 pCi/L range, commercial buildings were
thought to pose less of a problem. Because of the large expanse
of the building area, the complexities of the buildings HVAC
-------
systems, and the problem of finding and sealing radon entry routes,
this mitigation, as with other commercial buildings, required more
building investigation, more diagnostic testing and, consequently
more sophisticated solutions then any residential dwelling.
While EPA testing data showed ambient first floor radon
levels in the library fluctuating between 1500-2200 pCi/L,
second floor measurements were about 100 pCi/L. Sub-slab
measurements of 80,000 pCi/L indicated we were dealing with an
extremely hot slab. Numerous floor cracks were observed as was a
large expansion pour joint in the first floor storage room. When
a section of the rubberoid baseboard was removed a 3/4" separation
between the wall and expansion joint was observed. Communication
testing1 revealed fair to marginal (.oo2" H^O) between the main
storage and mechanical room, meaning we had either little aggregate
under the slab or a very strong negative pressure. Building plans,
while very detailed, nowhere indicated the amount nor the size of
aggregate under the slab. Drilling at the communication sites
suggested 4" - 6" of coarse aggregate base. Micro-roentgen gamma
readings inside and outside the building while elevated, did not
indicate any unusual source.
The HVAC system was of considerable concern because the air
handling units were way out of balance with no fresh air coming
into the building. Over a 30 year period there was never a
maintainanee contract on the system but instead, whenever there
was a problem, a different mechanical contractor was hired. The
problem was obvious: a strong source under the slab, numerous
entry routes and a HVAC system totally out of balance exerting
a negative pressure on the building.
REMEDIATION PLAN
Any remediation plan had to take into account a major sealing
effort, a sub-slab depressurization system (SSD) to remove the high
radon concentrations under the slab, and a balanced HVAC system.
To achieve suitable depressurization under the slab, all
wall/floor joints in the lower level had to be sealed, as well as
any floor cracks and entry routes around electrical conduits, etc. on
the first floor. All exposed floor cracks were ground out and filled
with a flowable urethane. All 4" rubberoid baseboards were also
removed and the wall/slab expansion joint filled with a flowable
urethane, as well. Great care was taken to seal as many entry routes
as possible to improve the operation of the depressurization system
as well as limit the influx of the gas.
Communication Tests - Communication tests are used to determine
the ease or difficulty with which gas can move through the soil,
aggregate under the slab or within a cavity such as a block wall.
Typically two holes are drilled 20' apart and suction applied to one
hole while micromanometer measurements are made at the other.
-------
The principal of a sub-slab depressurization system (SSD) is to
create enough negative pressure under the slab to pull radon out
before the negative pressure of the building can pull the gas into
the structure. The keys to this system functioning properly are
1) an aggregate base that allows for the establishment of an adequate
pressure field and 2) a fan that will have sufficient CFM to draw
the gas from all sub-slab areas, up three stories and out of the
building.
The depressurization system designed for the building had four
suction points. The vertical piping from these holes was 4" PVC
manifolded into 6" PVC horizontal pipe located along the first floor
ceiling. The 6" PVC piping was then routed to a 4' x 6' pipe chase
behind the elevator shaft, and then up to the roof penthouse. Before
a fan was chosen, the 4" holes were cut in the slab to determine
the type of thickness of the aggregate. The building had 6" of
IV crushed stone which was a plus for the SSD. With this information,
a wall mounted fan (Penn Ventilator WAQ-24L-800 CFM/1.00: S.P.)
was installed on the north wall of the penthouse. By using a wall
mounted unit, the EDPM roofing did not have to be penetrated and its
guarantee was not breached.
The HVAC system, due for scheduled repairs, was only partially
corrected. The correction involved opening the fresh air intakes
and rough balancing the system by a new mechanical contractor. Fine
tuning of the system is still to be accomplished.
POST REMEDIATION TESTING
Installation of the system took 3 men/3 days. Much of this
time was spent answering questions of media personnel. Once the
system was operational we immediately retested all areas of the
building and found the first floor levels had dropped from 2200
to 10 pCi/L and the second floor from 100+ to 5 pCi/L. Subsequent
diagnostic measurements revealed good communication with the
under slab and the ground floor ambient air with pressure differntial
ranging from 4-28 pascals. Static pressure measurements between
the interior riser pipes ranged from 50-83 pascals with 325 pascals
at the fan. Fine tuning of the HVAC system should drop these still
slightly elevated levels below the EPA guideline.
IN CLOSING
The remediation techniques used in this mitigation were designed
and based on detailed information supplied from sophisticated diagnostic
measurements. While the sealing and SSD systems were somewhat
similar to residential remediation techniques the HVAC system was the
curve ball. As more commercial buildings and schools with complex HVAC
systems are shown to have radon problems, the radon mitigator in
partnership with the mechanical contractor will be required to be a
detective and problem solver all in one. Commercial mitigation is surely
going to be the new frontier of radon.
-------
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.
-------
D-IX-3
EPA'S PROTOCOL DEVELOPMENT STUDY
FOR RADON IN SCHOOLS
Anita Schmidt
U. S. Environmental Protection Agency
ABSTRACT
A year long radon measurement study is being conducted in 21
schools in 7 states geographically dispersed across the United
states. These schools were selected nonrandomly from 130 schools
previously screened for radon during the Winter of 1989 in the
first Phase of the study. The purpose of this study is to gather
seasonal radon data and compare measurement methods to update and
refine EPA's measurement guidance for schools.
Gomr>r.oha«,fJ^Sei ZI °£ ^he studv' initiated in June 1989,
comprehensive long- and short-term measurements are being made for
£.?^r«UA"!?-«i Tiet£ °f, devices- Long-term measurements are
being made with alpha track detectors and electret-ion chambers.
snort-term measurements, using charcoal canisters, electret-ion
chambers, and continuous monitors for radon and radon progeny, are
being made under different ventilation conditions. The results of
these different measurement methods will be compared for their
applicability in schools. The impact of school construction type
and ventilation systems on radon measurement will also be
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.
-------
D-IX-4
RADON LEVELS IN NON-RESIDENTIAL BUILDINGS IN NEW JERSEY
Karen Tuccillo
New Jersey Department of Environmental Protection
ABSTRACT
Radon testing in various structures has been conducted
throughout the State of New Jersey including non-residential
buildings. Non-residential buildings differ significantly in
construction from private homes, therefore radon levels seen in
non-residential buildings may not reflect radon levels seen in
surrounding residential structures. Testing has demonstrated
that radon/radon decay products can build up to unacceptable
levels in non-residential as well as in residential structures.
The concentration of radon was measured in over 120
non-residential buildings, including schools, hospitals, and
municipal buildings in New Jersey. All structures tested had
initial radon levels in excess of the USEPA's 4.0
picocuries/liter continuous exposure guideline for residential
structures.
The distribution of radon levels for various non-residential
structures is examined and the efficiency of remedial measures
taken described. Data indicate that radon levels measured in
non-residential buildings are generally lower than those found in
residential structures. Correlation is seen between geographic
areas and elevated radon levels.
-------
D-IX-5
ELECTRET ION CHAMBERS FOR
RADON MEASUREMENTS IN SCHOOLS DURING
OCCUPIED AND UNOCCUPIED PERIODS
by: Kenneth D. Niggers, Tom D. Bullets, and
Paul A. Zoske
American Radon Services, Ltd.
ISIS Center
Iowa State University Research Park
Ames, Iowa 50010
Kelly W. Leovic
U.S. Environmental Protection Agency
Air and Energy Engineering Research Laboratory
Research Triangle Park, North Carolina 27711
David W. Saum
Infiltec
P.O. Box 1533
Falls Church, Virginia 22041
ABSTRACT
Continuous radon measurements made in some school buildings have shown
large diurnal variations due to factors such as occupancy patterns, HVAC system
design and operation, and weather. Since many radon screening measurements are
made with relatively inexpensive detectors that integrate the daytime and
nighttime exposures, the test results may not reflect the actual exposure of
students and teachers during occupancy hours.
This paper discusses the results of radon measurements made in schools
with electret ion chambers (EICs) and continuous radon monitors (CRMs) during
both occupied and unoccupied periods. Ten school buildings in Iowa, Maryland,
and North Carolina were tested in this study. The results show that radon levels
in these schools during occupied and unoccupied periods may vary significantly
depending on the specific building (i.e., whether the building is pressurized
or depressurized during the given testing period) giving both false positives
and false negatives.
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
In April 1989, the Environmental Protection Agency (EPA) released an
interim report for radon testing in schools^ ? Two types of radon screening
measurement devices and two respective methods of radon testing during the winter
months are recommended: (1) 2-day exposure open-faced charcoal canister exposed
over a 2-day weekend with the school's mechanical system operated in the same
manner as it is operated when school is in session; and (2) an alpha track
detector exposed for 3 months. It is recommended that frequently used rooms on
and below ground-level be tested.
As an alternative to conducting radon screening measurements in schools
with charcoal canisters or alpha track detectors, radon measurements in this
study were made in 10 schools using electret ion chambers (EICs). Seven of the
school buildings are in Iowa, two are in Maryland, and one is in North Carolina.
The measurements in the Maryland and North Carolina schools were part of an EPA
research project on radon reduction methods in schools. A continuous radon
monitor (CRM) was used for comparison with the EICs in the measurements conducted
in the Maryland schools.
EICs were used for the radon screening measurements in this study since
they provide one method of measuring radon levels during occupied and unoccupied
periods without investing in an expensive CRM. Because the EIC measurements in
the three states were conducted independently, they are discussed separately.
BACKGROUND ON EICs
EICs require no power and function as true integrating detectors, measuring
the average concentration during the measurement period. EICs contain a
permanently charged electret which collects ions formed in the chamber by
radiation emitted from radon decay products. When the device is exposed, radon
diffuses into the chamber through filtered openings. Ions which are generated
continuously by the decay of radon and radon decay products are drawn to the
surface of the electret and reduce its surface voltage. The amount of voltage
reduction is directly related to the average radon concentration present during
the exposure period. Both short-term (2 to 7 day) and long-term (1 to 12 month)
EICs are currently marketed. The thickness of the electret affects the usable
measurement period. Short-term EICs were used in this study.
The electret must be removed from the canister and the electret voltage
must be measured with a special surface voltmeter both before and after exposure.
The difference between the initial and final voltage is divided first by a
calibration factor and then by the number of exposure days to determine the
average radon concentration during the exposure period. For a 7-day exposure
period using a short-term EIC, the lower level of detection is about 0.3
picocuries per liter, pCi/L (1 pCi/L - 37 becquerels per cubic meter, Bq/m3). (2)
-------
MEASUREMENTS IN IOWA SCHOOLS
METHODOLOGY AND RESULTS
Phase 1. EICs were used for all radon testing in the Iowa schools. Every
classroom in the six building Urbandale Community School system and all the
offices in the Administration building were tested for radon over two intervals:
(1) from a Friday after classes to late Sunday afternoon; and (2) from late
Sunday afternoon to Wednesday after classes. The EIC electret voltages were
measured on-site late Sunday afternoon; the EICs were left in place and picked
up Wednesday after classes and returned to the laboratory for final electret
voltage reading. Testing was conducted while school was in session in December
1988.
The first step in the analysis of the Urbandale radon concentration data
was to extract from the database values greater than 3.0 pCi/L. These data were
sorted into two groups: (1) weekday radon concentration greater than weekend
radon concentration, and (2) weekend radon concentration greater than weekday
radon concentration.
Figures 1 and 2 show the values greater than 3.0 pCi/L in bar graph form.
A building specificity for weekday radon concentration being greater than weekend
radon concentration or the converse is apparent in both figures. A mechanism
seems to function at the building level to determine the radon concentration for
the respective time periods.
Phase 2
The radon concentration was measured during the occupied time and during
the unoccupied time (approximately daytime and nighttime) in several of the
school buildings. Buildings were selected that represented both radon
concentrations greater on the weekday than on the weekend and radon
concentrations greater on the weekend than the weekday. Each frequently used
room of the Administration and Olmsted buildings was tested with EICs during the
occupied and the unoccupied time of the 5-day workweek. Two EICs were placed
side-by-side with a log sheet for recording opening and closure of the EICs.
One EIC was marked "day" (occupied time) and the other EIC was marked "night"
(unoccupied time). School personnel opened and closed the EICs and kept records
on a log sheet indicating dates and times of openings and closures of the
respective EICs. This part of Phase 2 was conducted in February and March 1989.
The testing in the Administration building, replicated in March 1989, showed
the pattern seen in Figure 3. The Olmstead building radon concentration data
are shown in Figure 4.
Phase 3
All rooms in the Ballard School System's Kelley Elementary School were
measured for radon concentration as described in Phase 2. These data are
presented in Figure 5.
-------
DISCUSSION
A total of 196 classrooms were tested in the Urbandale School System.
Figure 1 shows data from 13 rooms with radon concentrations greater than 4 pCi/L
during the weekdays but less than 4 pCi/L during the weekend. The weekend
operation of the mechanical system was the same as during the week, and the
classrooms in the schools are generally unoccupied during the weekend. The
weekday testing period includes night and day (24 hour weekdays) in this plot.
Students and teachers are in school during the weekdays; therefore, the weekday
testing period may better represent estimated exposure to radon concentration.
If the weekday testing period is more predictive of occupational exposure to
radon than the weekend testing period, 16 false negatives would have been
indicated by weekend testing only. Likewise, 6 false positives would have been
indicated by testing on the weekend only (see Figure 2). The data in Figures
1 and 2 (16 false negatives and 6 false positives, respectively) show the
uncertainty introduced when using either time period for predicting occupational
exposure to radon concentration.
Figure 3 shows data from the Administration building tested during the
night (unoccupied time) and day (occupied time). Four of seven rooms showed
substantially greater radon concentration during the occupied time than during
the unoccupied time. A replicate measurement showed a similar pattern of radon
concentration response.
Figure 4 shows data from the Olmsted classroom building tested during the
night (unoccupied time) and day (occupied time). Unoccupied time (night) radon
concentration values were greater than the occupied time (day) radon
concentration values in this school.
Figure 5 shows that five of six rooms had a greater radon concentration
during the occupied time than during the unoccupied time. (The sixth room was
a second floor 6th grade room that had its windows open during the day. Note
that typically second floor classrooms would not be tested; however, the 6th
grade science teacher wanted to test that room.) A replicate measurement of all
rooms conducted in April 1989 during relatively warm weather was similar but with
less exaggeration of the differences. The Kelley Elementary School has a boiler
located at one end of the building with a tunnel that extends under some of the
rooms. The temperature setback at the end of the day and the setup at the
beginning of the day may explain why the radon concentration increases during
the day (warm soil-gas laden air in the tunnel may rise into the classrooms
above).
MEASUREMENTS IN MARYLAND SCHOOLS
The radon measurements in the Maryland schools were made with EICs and a
CRM for comparison. One school is in Prince Georges County and the other is in
Washington County. Both schools are currently being researched as part of EPA's
radon mitigation research program in schools.
-------
METHODOLOGY
The radon measurements were made in one room in each of these two schools
and covered a 2-week period. A CRM collected hourly radon measurements over
the entire 2-week period; one EIC was exposed during the entire 2-week period;
one EIC was exposed only during the day on Monday through Friday; and finally,
one EIC was exposed only during the night on Monday through Thursday. School
personnel were instructed on opening and closing of the EICs and kept a log of
exposure times. The measurements were made in late May 1989.
RESULTS AND DISCUSSION
The results measured in the Prince Georges County classroom are displayed
in Figure 6. Note that radon levels rose sharply at night and decreased during
the daytime. The actual radon exposure during the daytime was 2.1 and 2.5 pCi/L
for the CRM and EIC, respectively. The nighttime exposure for both devices was
4.4 pCi/L. The average radon level for the entire 2-week period was 4.1 and
4.2 pCi/L for the CRM and EIC, respectively. These data show excellent
correlation between the CRM and EIC.
Comparing the daytime (only), nighttime (only), and cumulative (total)
exposures during the 2-week period indicates that a measurement device
integrating radon levels over 24-hour periods would overestimate actual radon
exposure (i.e., produce a false positive).
Figure 7 shows CRM and EIC data from an office in a Washington County
school. The radon level measured during the daytime periods was 2.1 and 6.2
pCi/L for the CRM and EIC, respectively. The nighttime measurements for the
CRM and EIC, respectively, were 2.0 and 4.1 pCi/L; and the average radon level
for the entire 2-week period was 2.0 and 4.1 pCi/L for the CRM and EIC,
respectively.
The daytime and nighttime comparisons for these data are less conclusive
than those shown in Figure 6, since the EIC measurements range from 0.3 to 4.1
pCi/L greater than the CRM averages. In any event, had the radon measurements
been collected with an integrating device over the 2-week period, a false
negative would have been produced.
MEASUREMENTS IN NORTH CAROLINA SCHOOL
A school in Forsyth County that is currently part of EPA's radon mitigation
research program was selected for the EIC measurements in North Carolina. The
classrooms in this school are located in three pods connected to a central media
area. There are eight classrooms in each pod. Initial charcoal canister
measurements during the summer of 1989 with the HVAC system off showed that all
of the 24 classrooms in the pod area ranged from 3.5 to 10 pCi/L. The building
has a central heating, ventilating, and air-conditioning (HVAC) system and
appears to operate at a very slight positive or neutral pressure relative to the
outdoors.
-------
METHODOLOGY
Four of the eight classrooms in each of the three pods were measured for
1 week. The data were collected in late October and early November 1989. One
EIC was exposed for the entire week; one EIC was exposed only during the day;
and finally, one EIC was exposed only during the night. School personnel were
instructed on opening and closing the EICs and kept a log of exposure times.
RESULTS AND DISCUSSION
Figure 8 shows the results of the integrated, daytime, and nighttime
measurements in each of the classrooms. (Note that data for 11 of the 12 rooms
are presented since the data were not recorded properly in the log sheet in one
of the classrooms.) The averaged data for each of the pods are presented in
Figure 7. The overall averages for the classrooms are relatively close during
all three measurement periods averaging 6.0, 5.4, and 5.8 pCi/L for the
integrated, daytime, and nighttime periods, respectively.
Although these averages are very close, some trends can be observed in
Figure 8. Looking at the daytime versus nighttime radon levels for 10 of the
rooms (they are nearly the same for Room 36), the nighttime level is higher than
the daytime level in 7 of the 10 rooms. The averaged data in Figure 9 also
follow this trend. As mentioned above, the nighttime average was 0.4 pCi/L
greater than the daytime average for all the classrooms. This is consistent with
the understanding that the building is operated under a slight positive pressure.
However, when radon levels build up overnight, they are not sufficiently diluted
during the daytime. Inspection of the HVAC system revealed that the outdoor air
controls for most of the multi-zone air handlers appeared to be operating in an
economizer mode with no minimum outdoor air position.
Except for one notable excursion (daytime levels in Room 25 were about 11
pCi/L above nighttime levels), the integrated, daytime, and nighttime
measurements in this school were relatively close. Based on these results, if
only integrated screening measurements had been made in this school, only one
false positive and no false negatives would have been produced.
CONCLUSIONS
The following conclusions are based on the EIC measurements made in 10
Iowa, Maryland, and North Carolina schools under occupied and unoccupied
conditions. Their applicability to other schools will depend on the unique
characteristics of each school.
(1) Although there was generally good cooperation from the school
personnel in the opening and closing of the EICs and logging of
times, this method is not flawless and introduces additional
uncertainty into the measurements. Depending on the school, it may
be better to have one person open and close all the EICs.
(2) Building-specific responses during occupied and unoccupied periods
were found in most cases. Some schools had greater radon
concentrations during the occupied time (day) than during the
-------
unoccupied time (night), and the converse was true of other
buildings. Consequently, radon measurements that integrate levels
over a 24-hour period have the potential to produce false negatives
or false positives.
(3) Operating the building for 48 hours as it is operated during school
hours as recommended by EPA should most closely correlate with
student exposure.
(4) Where a CRM was placed side-by-side with the EICs, the correlations
were generally good (Maryland schools). However, some of the EIC
measurements in Washington County were consistently higher than the
CRM measurements during the same period.
Some potential advantages and disadvantages of EICs are apparent from this
study and prior experience with EICs. Advantages include: (1) EICs are supposed
to be an integrating detector unlike charcoal which weights the later exposures
more than the earlier, and (2) EICs can be opened and closed to start and stop
their integration and enabling actual exposures to be estimated. Potential
disadvantages include: (1) Since EICs are calibrated based on an equilibrium
of radon and progeny, their response to radon concentrations is not instantaneous
(this lag is thought to be similar to that of the CRM: about 0.5 hr for
equilibrium with radon and about 4 hr for that radon to reach equilibrium with
progeny. (2) If the radon levels are varying significantly, the CRM might be
expected to follow changes better than an EIC that is opened and closed each day
and has a start-up period each day while the CRM is open all the time. (3) Can
school personnel be relied upon to record the daily openings accurately enough
to allow the EIC exposure average to be computed? (4) Will frequent handling
affect the EIC readings (this might be expected to increase the voltage losses
and lead to overestimates of radon levels)? (5) Are the changes in school radon
levels so rapid when school starts in the morning and the HVAC systems are turned
on that this type of measurement is invalidated (due to the decay delays in the
CRMs and EICs since they depend on an equilibrium calibration)? (6) If exposure
is to be estimated for occupied time, how many exposure days must be averaged
to achieve enough EIC voltage drop to give an accurate estimate for radon levels
around 4 pCi/L? Hopefully, these questions will be addressed in future research.
REFERENCES
1. United States Environmental Protection Agency, Office of Radon Programs,
"Radon Measurements in Schools - An Interim Report." Washington, D.C.
20460, EPA-520/1-89-010, NTIS PB89-189-419, March 1989.
2. United States Environmental Protection Agency, Office of Radiation
Programs, "Indoor Radon and Radon Decay Product Measurement Protocols,"
Washington, DC, February 1989.
-------
I
1
6 7 8 9 10 11 12
Urbandale school rooms
13 14
16
•
Weekend ^ Weekday
Weekend Weekday
Rn Rn
concentration concentration
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Admin. Office
Admin. Office
Admin. Office
Admin. Office
Admin. Office
Admin. Office
Admin. Office
Admin. Office
High School
High School
Jensen
Jensen
Jensen
Jensen
Karen Acres
Rolling Green
Room pCi/L
Account. & Pay. 3.1
Business Man. Off. 3.6
Dale Y. 2.7
Lounge 1 .7
Maintenance 2.0
Shelby Off. 1.5
Storage Room 3.6
Women's Bathroom 2 4
104 4.0
606 2.6
310 2.0
312 < 0.5
317
-------
1 2 3 4 5 67 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Urbandale school rooms
NO.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Building
High School
High School
High School
High School
High School
Middle School
Middle School
Olmsted
Olmsted
Olmsted
Olmsted
Olmsted
Olmsted
Olmsted
Rolling Green
Rolling Green
Rolling Green
Rolling Green
Rolling Green
Rolling Green
Rolling Green
Rolling Green
Rolling Green
• Weekend E
Room
101
106
305
602
614
Oust Break Rm
Vocal Music
22
23
Kitchen
Lounge
Marion S.
Mrs K
Music
Kitchen
Librarian
Lisa's Desk
McTaggert
Mrs. Ennen
Mrs 1.
Mrs O.
Mrs S.
Music Room
2 Weekday
Weekend
Rn
concentration
pCi/L
36
4.7
3.5
33
34
33
33
127
200
84
9.6
3.8
5.6
6.6
3.5
3.5
3.1
40
31
31
41
4.8
38
Weekday
Rn
concentration
pCi/L
1 7
1.3
09
21
30
06
<0.5
103
104
64
<0.5
31
21
1 5
16
25
1 3
<05
09
<05
09
1 t
1 0
Figure 2. Rooms from buildings in the Urbandale school system with
radon concentrations greater on weekends than weekdays.
-------
Urbandale administrative rooms
Night
Day
NO.
1
2
3
4
5
6
7
Room
Accounts Payable
Business Management Off ice
Lounge
Maintenance
Shelby's Office
Maintenance Office
Outer Maintenance Office
Nighttime
Rn
concentration
pCi/L
2.0
2.5
2.9
5.0
5.2
7.1
6.0
Daytime
Rn
concentration
pCi/L
5.2
7.9
6.1
7.7
5.2
1.3
4.8
Figure 3. Night (unoccupied time) and day (occupied time) radon
concentrations in the Urbandale school system Administration building.
-------
Olmsted school rooms
Night
E3 Day
No.
1
2
3
4
5
6
7
8
Room
22
23
Kitchen
Lounge
MarionS.
MarionS.
Mrs. K.
Music
Nighttime
Rn
concentration
pCi/L
6.2
11.2
9.3
3.0
2.7
1.9
3.3
2.3
Daytime
Rn
concentration
pCi/L
4.7
B.6
6.4
1.4
n.a.
n.a.
<0.5
1.3
Figure 4. Night (unoccupied time) and day (occupied time) radon
concentrations in the Urbandale school system's Olmsted building.
-------
Kelley elementary rooms
Night
Day
NO.
1
2
3
4
5
6
Room
1st Grade
3rd Grade
6th Grade
Kindergarten
Library
S.E. room
Nighttime
Rn
concentration
pCi/L
2.1
2.0
1.8
<0.5
1.4
1.0
Daytime
Rn
concentration
pCi/L
3.3
11.8
<0.5
3.8
5.5
1.8
Figure 5. Night (unoccupied time) and day (occupied time) radon
concentrations in the Ballard school system's Kelley Elementary building
-------
10
8 -
6 -
§
•o
4 -
2 -
Total Period:
Weekday Period:
Weeknight Period:
138
142
146
1989 JULIAN DAY
150
154
Weekdays when EIC was open
Weeknights when EIC was open
Figure r. CRM and EIC results for Washington County school
-------
15
12
hJ
•H
CX
I 6
Time
(days)
Total Period:
Weekday Period:
Weeknight Period:
14.1
4.0
4.0
CRM
(pCi/L)
4.1
2.1
4.4
EIC
(pCi/L)
4.2
2.5
4.4
139
142
145 148
1989 OULIAN DAY
151
154
Weekdays when EIC was open
Weeknights when EIC was open
Figure 6. CRM and EIC results for Prince Georges County school.
-------
i
27
31
SB
37
SB
INTEGRATED
ClassrooH Na.
DAYTIME
EZ3 NIGHTTIME
Figure 8. EIC measurements in North Carolina school under
occupied and unoccupied conditions.
i
INTEGRATED
Pod No.
DAYTIME
ALL ROOMS
V771 NIGHTTIME
Figure 9. Summary of EIC measurements in each pod of
North Carolina school.
-------
D-IX-6
MEASURING RADON IN THE WORKPLACE
Michael Boyd
U.S. Environmental Protection Agency
Office of Radiation Programs
Washington, DC 20460
Terry Inge, John MacWaters
Sandy Cohen and Associates
McLean, VA 22101
ABSTRACT
The Environmental Protection Agency (EPA) has issued
guidance for testing for radon in homes and interim guidance for
testing in schools. Information on testing for radon in the
workplace is the next initiative and this paper describes the
current status of this effort. The results of measurements made
in several buildings in the Washington, DC area are discussed. A
discussion of preliminary guidance on radon survey design that
has been offered to Federal agencies is presented .
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
The discovery of dangerous levels of radon gas in homes in
the early 1980's prompted the Environmental Protection Agency
(EPA) to issue guidance for testing and mitigation of this
serious health threat. Since most Americans spend a majority of
their time at home, EPA decided to concentrate its resources to
encourage people to test their homes for radon and to take
corrective action where appropriate. In 1988, EPA was able to
extend its radon program to address the hazard of radon in
schools. Since children may be at higher risk from exposure to
radon, this initiative was the most appropriate next step from a
public health standpoint. As our understanding of the risk from
indoor air pollution has grown, the predominance of radon as the
most serious indoor air pollutant has become obvious. It is for
this reason that EPA is now preparing guidance for testing for
radon in the workplace. Congress has included radon in the
workplace as an area of concern. Sections 118(k) and 403(b) of
the Superfund Amendments Reauthorization Act (SARA) instruct the
agency to undertake testing and research in this area.
In addition, Section 309 of the Indoor Radon Abatement Act
(Public Law 100-551) requires the heads of all Federal
departments and agencies that own buildings to test a
representative sample of those buildings for radon and report
their results to EPA's Administrator. A report to Congress on
these results is then due in October, 1990. Preliminary
information on the design of an effective radon survey has been
provided to Federal agencies to assist them in completing this
task. EPA expects that the information gained from these surveys
will significantly enhance the present understanding of radon
behavior in large buildings.
A study was undertaken in the winter of 1989 to perform
intensive measurements in several buildings in the Washington, DC
metropolitan area. Permission was obtained from the General
Services Administration (GSA) to use three of their buildings for
this study. The buildings were chosen to represent differing
ages, construction types, and sizes. This work provided
confirmation of several phenomena observed in homes including
diurnal cycling of radon concentrations and dependance of radon
levels on the operating parameters of the ventilation system.
Information on radon movement throughout upper floors of
buildings was not obtained because the observed radon
concentrations were too low. The three GSA buildings, all
located in downtown Washington, DC, had ground floor averages
below 1 picocurie per liter (pCi/L) which made detection on upper
floors impossible. This study is continuing this year.
-------
MATERIALS AND METHOD
Federal agencies were required to submit their radon survey
designs to EPA by July, 1989. The surveys are limited to radon
screening measurements. Although follow-up measurements and
mitigation are not addressed, these components will almost
certainly be undertaken when elevated concentrations of radon are
reported. Most of the agencies have chosen to perform 3-month
measurements using alpha track detectors placed during the winter
heating season. The next most popular device is the diffusion
barrier charcoal canister which may be deployed for 2 to 7 days.
Electret ion chambers and continuous monitors are also being
considered by some agencies. EPA's policy is that any
measurement device listed in the Indoor Radon and Radon Decay
Product Measurement Protocols (EPA 520/1-89-033, March 1989) is
acceptable for making screening measurements except that grab
sampling may not be used alone. Grab sampling can be a useful
confirmatory measurement made in conjunction with a measurement
from some other approved device.
As an independent initiative, EPA is investigating radon
behavior in large buildings in preparation for issuing interim
guidelines for measuring radon in the workplace. This effort
will include a variety of building types and sizes. The phase of
this project that was completed in 1989 included initial
screening measurements made with 2-day open faced charcoal
canisters. Continuous radon measurements were made using
continuous radon gas monitors. The data collected in the 3 GSA
buildings was supplemented with the results of extensive 3-raonth
alpha track measurements made by GSA as part of its radon program
begun in 1989. Other data collected included differential
pressure measurements, meteorological data, and information on
the layout, type, and operation of heating, ventilation, and air
conditioning (HVAC) systems.
This project is continuing this season. The prospect for
locating Federal buildings with elevated radon levels is improved
because of the availability of the data from GSA's radon program.
It is anticipated that seasonal and annual measurements will be
taken as part of this study.
-------
RESULTS AND DISCUSSION
EXPERIMENTAL DATA
Data from the 6SA buildings include charcoal canister
measurements and alpha track results from all 3 buildings and
continuous monitoring data from one building. A summary of the
ATD results is presented in Table 1.
Table 1
ATD Results in GSA Buildings
Radon Concentration Number of Readings at Each Level
fpCi/Ll Blda. 1 Blda. 2 Blda. 3
0 - 0.5 12 50 77
0.6 - 1.0 13 26 30
1.1-1.5 3 14 1
1.6 - 2.0 021
The charcoal canister results in the three buildings gave
readings that were for the most part in the 0-0.5 pCi/L range.
A few readings were obtained in the 0.6-2 pCi/L range with the
highest of these readings being located on the lowest level of
the building. One reading of 3.3 pCi/1 was obtained in a
crawlspace under one of the buildings, confirming the generally
low source potential for radon in the GSA buildings that were
tested.
Continuous monitoring of radon in Building 3 exhibited a
diurnal pattern of radon concentrations. During the weekdays, a
rise in radon concentration was noted beginning around 6 pro. By
2 am, the radon concentration was at its maximum where it
remained at this plateau or oscillated until 6 am. At 6 am on
weekdays, the radon concentration dropped rapidly to under 1
pCi/L by 10 am where it remained until 6 pm. This pattern is
explained by the operation of the HVAC system. The air handlers
in the area under study are turned on at 6 am and shut off at 6
pm on weekdays. In spite of a slight negative pressure created
by running the system, the air handlers are responsible for
decreasing radon concentration because of fresh air dilution and
mixing. On the weekends, when the HVAC system is not operating,
the peak concentration remains relatively stable until Monday
morning when the system is restarted.
-------
FEDERAL GUIDANCE
Based on the preliminary results of this study and input
from technical experts, information was distributed to Federal
agencies to assist them in preparing their radon survey designs
for complying with the Indoor Radon Abatement Act. It was
decided to recommend that all ground-contact occupiable rooms be
tested. In addition, it was suggested that a few detectors be
placed on upper levels in stairwells, near elevators, and beside
service shafts and ductwork. Radon is principally a soil gas
problem and testing ground contact rooms should indicate the
highest radon concentrations in the building. Movement of radon
throughout a building should be indicated by the upper level
locations that were specified. Private water supplies are also
required to be tested which increases our confidence that we can
identify a radon problem if one exists.
Federal agencies were advised to perform screening
measurements during the winter heating season. In warm climates
where there is no heating season, it is recommended that testing
be done during the period of minimal outside air intake to
minimize dilution flow. For colder climates, the stack effect is
assumed to be the principal driving force for radon entry. The
stack effect occurs during the colder months when the effect of
rising warm air creates a negative pressure in the lowest levels
of a building relative to the air outside the building. This
negative pressure causes radon-rich soil gas to be sucked into
the building. In commercial buildings, large negative pressures
can also result from unbalanced HVAC systems. The assumption
that winter is the time to test is being investigated as part of
EPA's ongoing research efforts.
The suggested time to test is during the work week under
normal building conditions. The HVAC system should be operated
normally. While screening conditions are chosen to produce
results higher than the annual average, there is no benefit to
testing under conditions or during periods that do not reflect
conditions present when the building is occupied.
In the situation where an agency decides not to test all of
its buildings, EPA's recommendation is to test fewer buildings
according to the above guidelines rather than to test more
buildings with only a few detectors per building. While the
latter option may produce a statistically valid estimate of the
percentage of Federal buildings with elevated levels of radon, it
cannot certify that any individual building is safe. Experience
has shown that even adjacent rooms can have dramatically
different radon levels.
Finally, all Federal departments and agencies were strongly
advised to put in place a quality assurance program with a strong
-------
quality control component. For radon measurements made with
passive detectors, a good quality control program will include an
appropriate number of duplicate and blank detectors as well as a
few spiked detectors. Duplicate detectors are collocated and
provide an indication of the precision of the analytical
laboratory. Blank detectors serve as a check on the background
of the detectors. Spiked detectors, those exposed to a known
concentration of radon, provide an indication of accuracy or
bias.
CONCLUSIONS
Ongoing research has tended to support assumptions about the
importance of building conditions to the potential for radon
entry. Research to confirm that winter testing will yield
maximum values is continuing. Experience has shown us that a
balanced HVAC system, particularly one that contributes to a
slightly positive differential pressure in rooms in ground
contact, will minimize or eliminate radon entry.
Yet to be determined is the extent of radon movement
vertically within large buildings. Testing has been encouraged
in stairwells, near elevators, and near service shafts. It is
hoped that the extent to which radon may be a problem on upper
levels of buildings can be determined from the current phase of
EPA's work in this area.
Because the Congress required Federal agencies to sample a
representative number of their buildings this year, EPA will have
a large amount of data to assist in confirming the preliminary
guidance for testing in the workplace. Interim guidelines for
measuring radon in the workplace will likely be available by the
end of this year.
-------
D-IX-7
THE SCHOOL EVALUATION PROGRAM
by: Eugene Fisher
Florence Blair
U. S. Environmental Protection Agency
Office of Radiation Programs
Washington, D. C. 20460
Terry Brennan
Camroden Associates
Oriskany, New York 13424
William Turner
Harriman Associates
Auburn, Maine 04210
ABSTRACT
A pilot program to provide classroom and field training to school
facility operators was implemented by the U. S. Environmental Protection
Agency's Office of Radiation Programs in 1989. This program consisted of two
phases. The first phase developed and delivered a three day workshop in
Nashville, Tennessee. As a result of the workshop, a second phase was
initiated. The second phase investigated several school buildings with
elevated indoor radon levels in the Western United States. Radon entry
mechanisms were identified. Measurements to evaluate soil depressurization as
a radon control method were made and HVAC systems were characterized.
Measurements were made to evaluate HVAC modification as a radon control
method. Building shell tightness measurements were made and information was
collected to judge the suitability of potential sites for additional EPA
sponsored "hands on" school training. Physical and institutional problem
areas were identified.
ft U.S. SOVEMfCNT PRINTING OFFICE: 1990 748-010/25004
------- |