United States National HisK Management
Environmental Protection Research Laboratory
Agency Cincinnati, OH 45268
Research and Development EPA/600/SR-96/005 March 1996
& EPA Project Summary
Residential Radon Resistant
Construction Feature Selection
System
Kirk K, Nielson, Rodger B. Holt, and Vern C. Rogers
The Florida Department of Commu-
nity Affairs (DCA) has proposed stan-
dards for radon-resistant features in
residential construction. The features
consist of engineered barriers to re-
duce radon entry and accumulation in-
doors. The proposed standards require
radon-resistant features in proportion
to regional soil radon potentials, which
are defined from a statewide radon po-
tential map. The report describes the
basis and development of the system
for selecting radon control features for
new house construction in different re-
gions according to their mapped soil
radon potentials.
The effectiveness of different radon
control features was estimated from
new laboratory measurements, analy-
ses of new and previous house stud-
ies, and mathematical model simula-
tions. The laboratory measurements
characterized five brands of polyeth-
ylene sub-slab membranes to have
equivalent radon diffusion coefficients
of 3.4x10-7 cm2 s*1 ± 6.3x10* cm2 s1. The
geometric mean air permeability of the
membranes was 6.5x10" cm2 with a
geometric standard deviation of 8.4.
New house studies included 14
houses characterized by Southern Re-
search Institute (SRI) and 10 houses
characterized by University of Florida
(UF). The analyses showed that both
monolithic-slab (Mono) and slab-in-
stem-wall (SSW) foundation designs
can passively control indoor/sub-slab
radon ratios to average levels of Cm/
Cw() = 3.3x10"4 to 4.2x10-4. These ratios
are slightly lower than measurements
in other houses the previous year, and
two to four times lower than ratios from
earlier studies. The SRI ratios are 1.4
to 3.7 times lower than values from a
lumped-parameter model, primarily due
to improved sealing of slab penetra-
tions. The UF ratios are within a factor
of 1.74 of calculated ratios. The geo-
metric mean of all measured Cn#/C
ratios for Mono houses is 5.5x10"*
(GSD=3.14, n=43), and the geometric
mean of ratios for SSW houses is
1.1 x10"3 (GSD=3.02, n=52). The Mono
design offers approximately twice as
much passive radon resistance as SSW
designs.
Radon Emanation and Transport into
Dwellings (RAETRAD) model simula-
tions estimated the numerical effective-
ness of SSW, Mono, and floating-slab
foundations in connection with other
radon controls. The controls were
ranked by decreasing effectiveness as:
(1) active sub-slab ventilation system,
(2) vapor membrane placement, (3) en-
hanced ventilation, (4) improved foun-
dation design (SSW or Mono), (5) 10-
cm slump concrete, (6) 15-cm slump
concrete, (7) sealing of slab openings
and cracks, (8) sealing of slab penetra-
tions, (9) sealing of openings and only
large cracks, (10) use of a passive sub-
slab ventilation system, (11) compac-
tion of fill soil, (12) elimination of slab
reinforcement, and (13) reinforcement
of re-entrant corners. From these
-------
rankings, a passive group (items 4, 6,
8, and 9) was selected to reduce radon
by a factor of 2.1. Active control by
item (1) increased radon control to a
factor of 9.3.
A Florida radon protection map was
developed to show where the passive
group was needed to keep radon be-
low 4 pCi L1 and where the active fea-
ture was also needed. The map shows
three categories. The green category
denotes regions where less than 5% of
the area should exceed 4 pCi L1 in a
reference house. Other regions, where
the top 5% of the computed radon lev-
els fell between 4 and 8.3 pCi L1, were
assigned to the yellow category, indi-
cating a need for passive radon con-
trols. Regions where the top 5% of the
computed radon levels exceeded 8.3
pCi L"1 were assigned to the red cat-
egory, indicating additional need for
active radon controls.
This project summary was developed
by National Risk Management and Re-
search Laboratory's Air Pollution Pre-
vention and Control Division, Research
Triangle Park, NC, to announce key
findings of the research project that is
fully documented in a separate report
of the same title (see Project Report
ordering information at back).
Introduction
Radon (222Rn) gas generated from ra-
dium (226Ra) decay in soils can enter
houses through foundation openings. With
elevated entry and inadequate ventilation,
radon can accumulate to levels that pose
significant risks of lung cancer with chronic
exposure. The Florida DCA has proposed
construction standards to protect public
health by requiring radon-resistant build-
ing features in areas of elevated soil ra-
don potential. The report describes a sys-
tem for selecting different radon resistant
features for new house construction in
different regions according to their soil
radon potentials.
The proposed DCA standards reduce
radon entry into houses using different
combinations of improved foundation de-
signs, understructure sealing, altered air
pressures, and other engineered features
developed under their Florida Radon Re-
search Program (FRRP). The proposed
standards seek to minimize radon levels
without undue cost by matching the radon
resistance of required features to the soil
radon potential of each geographic re-
gion. The radon potentials of different re-
gions are estimated from a statewide map
and data base developed by DCA specifi-
cally for this targeted use of radon-resis-
tance features.
This report characterizes the radon re-
sistance of different radon control features
from new laboratory measurements on
plastic membranes, analyses of new and
previous house studies with lumped pa-
rameters, and simulations with the
RAETRAD numerical model. The most
cost-effective features are ranked and
grouped into active and passive catego-
ries. From the combined effectiveness of
each category and the mapped radon
potentials of 3,919 Florida regions, each
region is assigned to one of three radon
protection categories to achieve a 4 pCi
L1 indoor radon average. The categories
show where no supplementary radon con-
trols are required, where passive controls
are required, and where active controls
are also required. A radon protection map
is developed from the respective regional
categories using green, yellow, and red to
show the minimum regional radon-protec-
tive construction requirements.
Radon Resistance of
Polyethylene Vapor Barrier
Membranes
The radon transport properties of poly-
ethylene vapor barrier membranes are re-
quired to correctly model radon entry
through floor slabs. Although commonly
used under Florida slabs, the membranes
have generally been ignored in modeling
because their radon transport properties
were unknown. Radon diffusion and air
permeability coefficients of polyethylene
membranes were measured to fill the data
gap and thereby provide for more accu-
rate modeling of radon entry.
Five brands of 0.015-cm (6-mil) poly-
ethylene membranes were purchased from
different commercial suppliers in Central
Florida. Actual membrane thicknesses av-
eraged 0.012 ± 0.002 cm. Three 10-cm
diameter circles were cut from each brand
for radon diffusion measurements by the
time-dependent method. Three 1.5 x 3.0
m sheets were cut from each brand for air
permeability measurements. Each sheet
was folded and sealed to form a 1.5-m
square envelope, which was inflated,
weighted by a 12-kg mass, and monitored
for air pressure and volume changes.
The 15 radon diffusion measurements
averaged 3.36x10 7 ± 6.3x106 cm2 s"1.
Analysis of variance showed no signifi-
cant differences (p < 0.25) among the five
brands. The 15 air permeability measure-
ments were log-normally distributed, with
an overall geometric mean of 6.5x1015
cm2 and a geometric standard deviation of
8.4. Analysis of variance showed signifi-
cant differences among the different
brands (p < 0.025).
Lumped-Parameter Model
Estimates of Feature
Effectiveness
The effectiveness of different features
in resisting radon entry was studied in
several demonstration projects under the
FRRP. These included two test cells and
20 houses studied by Geomet Technolo-
gies, Inc., 27 houses studied by Florida
Solar Energy Center (FSEC), 30 houses
studied by SRI, and 14 houses studied by
UF. Fourteen additional houses have now
been studied by SRI, and 12 additional
houses by UF.
Previous Analyses
The effectiveness of different house con-
struction features was compared using the
net indoor/subslab radon concentration
ratio (CJC ). Measured CJC^ ratios
for both SSW houses and Mono slab
houses were highest for houses with no
sub-slab ventilation (SSV) system, were
slightly lower for houses with a passive
SSV system, and were significantly lower
for houses with active SSV systems. Mea-
sured C /C . ratios for SSW houses av-
net sub
eraged 2.7x10'3 for houses without SSV
systems, 4.9x10 3 for houses with capped
SSV systems, 1.7x103 for houses with
passive SSV systems, and 4.3x10"4 for
houses with fan-activated SSV systems.
Measured C JO „ ratios for monolithic-
ndf sub
slab houses averaged 2.3x10-3 for houses
without SSV systems, 6.2x10"4 for houses
with capped SSV systems, 2.2x103 for
houses with passive SSV systems, and
4.4x10"4 for houses with fan-activated SSV
systems.
The results of a second semi-empirical
study suggested that SSW construction,
when completed in accordance with the
FRRP standard, reduces indoor radon to
about 9x10"4 of the sub-slab concentration
(with an uncertainty factor of 2.2). Mono
slab construction may improve radon re-
sistance by approximately 33%, reducing
indoor radon levels by a factor of 0.67
compared to SSW construction. Activa-
tion of SSV systems with exhaust fans
reduces the C /Csil) ratio by approximately
70%, reducing indoor radon levels to about
0.3 times the levels that occur when the
SSV system is in the passive or capped
mode. The measurements on active SSV
systems are sparse and uncertain, how-
ever, due to the small number of houses
where the SSV systems were activated.
New Data
Data from the 1993 FRRP New House
Evaluation Program (NHEP) were com-
piled in terms of the lumped-parameter
model parameters or their surrogates. The
2
-------
data were measured by SRI and UF. The
SRI set contained 14 houses, eight of
which were Mono slab-stem wall houses
and six of which had slabs poured into
SSW. The UF set contained 12 houses,
10 of which were Mono and two of which
were SSW designs. The building shell and
SSV systems constituted the major differ-
ences between the SRI and UF data sets.
The SRI houses utilized hollow-block con-
crete walls for all but one house, while all
12 of the UF houses were built with frame
walls. The SRI houses utilized ventilation
mats for SSV systems, while the UF
houses utilized suction pits in their SSV
systems. Except for two UF houses, none
of the SSV systems were activated in
either group.
House air leakage was compared un-
der passive pressure conditions. The SRI
houses had an average natural ventilation
rate of only 0.19 air changes per hour
(ach), while the UF houses averaged 0.29
ach (possibly due to frame versus ma-
sonry construction, or occupancy and pro-
tocol differences). The average of the ob-
served floor crack areas for the SRI houses
was more than twice as high as the aver-
age for the UF houses. However these
averages come from variable data, and
additional cracks may be concealed. The
SRI mean is dominated by the SSW
houses, which averaged three to five times
higher than the other groups.
The geometric means of the Cne/Csub
ratios for the passive-control SRI measure-
ments averaged 1.6x104 for the Mono
houses and 4.2x10~4 for the SSW houses.
The geometric means of the Cno/Csub ra-
tios for the passive-control UF measure-
ments averaged 1.0x10"3 for the Mono
houses and 4.1x10'4 for the SSW houses.
The single active-system Mono house had
a C /C . ratio of 6.2x10-4, and the active-
sub '
system SSV house had a C JC b ratio of
1.7x10"4.
Lumped-Parameter Model
Comparisons
The measured C^/C^ ratios for the
SRI and UF houses were compared with
corresponding C/C ratios that were
calculated using "the lumped parameter
model as reported previously. The lumped
parameters used in the comparisons were
defined primarily from site-specific mea-
surements or surrogates.
The measured ratios for the SRI houses
are lower than the calculated values by
factors of 3.7 for the eight Mono houses
and 1.4 for the six SSW houses. The
lower measured ratios are attributed to
improved sealing of slab penetrations by
SRI in the present study. Since the Mono
houses have virtually no advective radon
entry routes other than slab penetrations,
the greatest difference was noted for the
SRI Mono houses. The SSW houses, de-
spite improved sealing, still have perme-
able channels at the stem walls where
advective radon entry can occur.
The 10 UF Mono houses show closer
agreement between measured and calcu-
lated radon ratios, differing by an average
factor of 1.74. The closer agreement sug-
gests better consistency with the previous
radon resistance effectiveness on which
the lumped-parameter model is based. The
two UF SSW houses have significantly
lower measured radon ratios than the cal-
culated values (by an average factor of
nearly 2.7). This lower ratio may reflect
better construction technique; however it
is more uncertain because only two houses
are being compared.
Comparisons with Prior
Measurements
The present NHEP measurements of
C ,/C ub ratios can be interpreted better
with tfie perspective of the two prior sets
of NHEP house studies by the present
contractors and others. The prior studies,
primarily covered the FY-91 and FY-92
budget periods. Two prior FSEC sets of
Mono houses are intermediate between
the Geomet '91 and SRI '91 sets and the
UF '92 set and the present SRI data set.
The geometric mean of all of the passive
Mono houses is 5.5x10"1 (GSD = 3.14).
The present C /C^,, ratios for passive
SSW houses are only slightly lower than
the FSEC '92 and UF '92 data. However,
the present data are well below the range
of the FSEC '91, the Geomet '91, and the
SRI '91 studies. The geometric mean of
all of the passive SSW houses is 1.1x103
(GSD = 3.02). The 52 SSW houses com-
prising this estimate therefore are approxi-
mately half as resistant to radon entry as
the 43 Mono houses. The chronological
trend in C /Ctub ratios is shown by time
averages in the full report. The improve-
ment in radon control may potentially be
attributed to increased experience in build-
ing the passive radon control features. At
least some of the 1991 studies included
houses built before the radon standard or
allowed builders to select their own pas-
sive features from earlier alternatives. Sub-
sequently, closer surveillance and training
by FRRP researchers ensured that most
or all of the desired radon-resistant fea-
tures were actually incorporated.
Data for estimating the effectiveness of
active SSV systems are limited. The geo-
metric mean of seven Mono houses with
active SSV systems is 3.3x104
(GSD = 1.51). The geometric mean of the
17 SSW houses with active SSV systems
is nearly identical at 3.4x10"4 (GSD = 2.58).
RAETRAD Model Estimates of
Feature Effectiveness
The radon resistance effectiveness of
different construction features was also
evaluated by numerical simulations with
the RAETRAD model. The RAETRAD
model uses multiphase calculations of ra-
don generation and transport, including
moisture effects on the radon entry simu-
lations. The simulations utilized reference
houses and soil profiles to compare in-
door radon levels with and without the
different building features. The effective-
ness of each feature was defined as the
ratio of the reference indoor radon con-
centration (without the feature) to the in-
door radon concentration with the feature.
For interactive features, several reference
houses are defined so that the feature
effectiveness can be evaluated with the
different interactions. The features evalu-
ated include the effects of fill soil compac-
tion, sub-slab vapor barriers, slab rein-
forcement, reinforcement of re-entrant slab
corners, sealing of slab penetrations, clo-
sure and sealing of large slab openings,
reduction of water/cement ratio (as esti-
mated by reduced concrete slump), and
use of different slab edge details.
The houses were modeled on a 4.3-m
soil profile with a water table at 2.5 m
depth and its resulting moisture profile.
Three reference houses were defined for
the floating slab, SSW, and Mono cases.
The reference houses had a 8.6 x 16.5 m
footprint, and an interior height of 2.4 m.
The indoor air pressure and outdoor air
exchange rate were -2.4 Pa and 0.25 lr\
respectively, consistent with previous
analyses.
Ranking and Groupings of
Construction Features
The various features were ranked in
descending order of effectiveness. The
reference case was defined as the com-
mon floating slab house with no SSV sys-
tem. The reference house had a sub-slab
vapor barrier, since this is commonly used
or required by building codes. The rank-
ordered list of radon resistance factors is
shown in Table 1. Several factors were
grouped with an average effectiveness
because of their dependence on slab edge
details. Where the most radon-resistant
approach was already standard building
practice, the factors are less than unity. In
these cases, the feature ranking is listed
as the reciprocal of the factor, but the
factor is not included in the composite
because it is already being utilized. Fea-
3
-------
Table 1. Ranking of Residential Construction Features by Average Radon Resistance Effectiveness
Effectiveness Summary Rank
Construction Feature
^ro/^loatufo
Relative to
Effectiveness
1. Active SSV system
4.45
No SSV
4.45
2a. No vapor barrier - floating slab
0.57
6 mill v. barrier*
0.48(2.1)
2b. No vapor barrier - SSW
0.47
6 mil v. barrier
"
2c. No vapor barrier - monolithic
0.40
6 mil v. barrier
"
3. Enhanced ventilation
2
0.25 ach
2
4a. Monolithic slab & stem wall
1.76
Floating Slab
1.62
4b. Slab poured into stem wall
1.47
Floating Slab
"
5a. 10-cm concr. slump - floating slab
1.17
20-cm slump
1.33
5b. 10-cm concr. slump - SSW
1.26
20-cm slump
"
5c. 10-cm concr. slump - monolithic
1.40
20-cm slump
"
6a. 15-cm concr. slump - floating slab
1.08
20-cm slump
1.15
6b. 15-cm concr. slump - SSW
1.12
20-cm slump
"
6c. 15-cm concr. slump - monolithic
1.17
20-cm slump
"
7. Seal slab openings & cracks
1.15
Unsealed
1.15
8. Seal slab penetrations
1.13
Unsealed
1.13
9. Seal openings
-------
The radon protection map was com-
pared numerically with 9,038 indoor radon
measurements from three data sets. The
middle 95% of the map range included
95.4% of the 2,952 Geomet land-based
data points, which best represent Florida,
with 1.9% below and 2.7% above the mid-
range, compared to 2.5% expected for
each. The 2,095 measurements in the
Geomet population-based data set aver-
aged slightly lower, with 4.0% below and
1.5% above the 95% mid-range, compared
to 2.5% expected for each. The 3,938
measurements in the Florida Health and
Rehabilitative Services residential data set
were slightly high, with 0.7% below and
4.7% above the 95% mid-range, compared
to 2.5% expected for each.
Over 250 houses with the greatest dif-
ferences between measured and predicted
indoor radon concentrations showed trends
that offer further explanations. Houses
above the 95% mid-range were about 25
times more likely to use slab-on-grade
construction than to have crawl spaces,
while the opposite trend was seen for
houses below the mid-range. Similarly,
houses above the 95% mid-range were
about 50% more likely to use hollow-block
construction than frame construction, and
the opposite trend was also seen for
houses below the mid-range. These trends
are consistent with model predictions. Con-
sidering the variations in both measure-
ments and map calculations, the mea-
surements give excellent overall statewide
validation of the radon protection map.
5
-------
K. Nielson, R. Holt, and V. Rogers are with Rogers and Associates Engineering
Corp., Salt Lake City, UT 84110.
David C. Sanchez is the EPA Project Officer (see below).
The complete report, entitled "Residential Radon Resistant Construction Feature
Selection System," (Order No. PB96-153473; Cost: $19.50, subject to change)
will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Air Pollution Prevention and Control Division
National Risk Management Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
United States
Environmental Protection Agency
National Risk Management Research Laboratory (G-72)
Cincinnati, OH 45268
Official Business
Penalty for Private Use $300
BULK RATE
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
EPA/600/SR-96/005
------- |