EPA-600/R-94-175
September 1994
RADON GENERATION AND TRANSPORT
THROUGH CONCRETE FOUNDATIONS
By
V.C. Rogers, K.K. Nielson, M.A. Lehto, and R.B. Holt
Rogers and Associates Engineering Corporation
P. O. Box 330
Salt Lake City, UT 84110-0330
EPA Contract No. 68-DO-0097
Interagency Agreement RWFL 933783
DC A Project Officer: Mohammad Madani
State of Florida
2740 Centerview Drive
Tallahassee, FL 32399
EPA Project Officer: David C. Sanchez
U. S. Environmental Protection Agency
Air and Energy Engineering Research Laboratory
Research Triangle Park, NC 27711
Prepared for:
State of Florida
Department of Community Affairs
2740 Centerview Drive
Tallahassee, FL 32399
and
U. S. Environmental Protection Agency
Office of Research and Development
Washington, D. C. 20460
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before eomplet III IIIIII Mill 111II111 III 11 III
1. REPORT NO. 2.
EPA-600/R-94-17 5
3. ill milllllllIIIIIIIIIIIIIIII
" PB95-101218
4. TITLE AND SUBTITLE
Radon Generation and Transport Through Concrete
Foundations
5. REPORT DATE
September 1994
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
V. C. Rogers, K.K. Nielson, M. A, Lehto. and R. B.
Holt
8. PERFORMING ORGANIZATION REPORT NO.
RAE-9127/10-3R2
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Rogers and Associates Engineering Corporation
P. O. Box 330
Salt Lake City, Utah 84110-0330
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-DO-0097, WA 2-20
IAG RWFL933783
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Air and Energy Engineering Research Laboratory-
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Task Final; 3/91 - 10/92
14. SPONSORING AGENCY CODE
EPA/600/13
is. supplementary notes ^EERL project officer is David C. Sanchez, Mail Drop 54, 919/
541-2979.
is. abstractrep0rt gives results of an examination of radon generation and trans-
port through Florida residential concretes for their contribution to indoor radon con-
centrations. Radium concentrations in the 11 concretes tested were all < 2.5 pCi/g,
and radon emanation coefficients were all < 0,08. Measurements on the constituents
of four of the concretes revealed that radium concentrations > 1 pCi/g were gener-
ally due to elevated radium in the aggregate, but occurred occasionally from radium
in the cement. Because the aggregates tested generally had very low emanation co-
efficients, elevated radium in the aggregate had a lesser impact on indoor radon than
elevated radium in the cement component. Diffusion coefficients for Florida con-
crete samples range from 0. 00018 sq cm/sec to 0.0046 sq cm/sec, but
air permeability coefficients are generally < 10 to the minus 11th power sq cm.
Thus, advection through a concrete slab is negligible compared to diffusion. Finally,
the report presents simple correlations for diffusion and permeability coefficients
and for radon generation in concrete and entry into dwellings.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b.identifiers/open ended terms
c. COSATI Field/Group
Pollution
Radon
Radium
Residential Buildings
Concrete Slabs
Foundations
Pollution Control
Stationary Sources
Indoor Air
Florida
13B
07B
13 M
11B, 13 C
18. distribution statement
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
47
20. SECURITY CLASS (This pagej
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
-------
EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the contents necessarily
reflect the views and policy of the Agency, r»or does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
ACKNOWLEDGEMENT
This work was also supported by U.S. Environmental Protection Agency
IAG-RWFL933783. The phosphogypsum was supplied by Southern
Research Institute.
ABSTRACT
Radon generation and transport through Florida residential concretes are examined for their contribution
to indoor radon concentrations. Radium concentrations in the 11 concretes tested are all less than '2.5
pCi/g, and radon emanation coefficients are all less than 0.08. Measurements on the constituents of four
of the concretes reveal that when the radium concentrations are greater than 1 pCi/g, it is generally due
•to elevated radium in the aggregate, but may occasionally occur from radium in the cement. Because the
aggregates tested generally have very low emanation coefficients, elevated radium in the aggregate has
a lesser impact on indoor radon than elevated radium in the cement component.
Diffusion coefficients for Florida concrete samples generally range from 1.8 x 10~* cm2 sec"1 to 4.6 x 1G"3
cm2 sec"1. Air permeability coefficients are less than 10" cm2. Thus, advection through a concrete slab
is negligible compared to diffusion. Finally, simple correlations are presented for diffusion and
permeability coefficients and for radon generation in concrete and entry into dwellings.
ii
-------
CONTENTS
Page
Acknowledgement ii
Abstract ii
Figures v
Tables vi
1. INTRODUCTION 1-1
1.1 Scope 1-1
1.2 Background 1-1
1.3 Report Contents 1-4
2. LABORATORY TESTS OF RADON TRANSPORT PROPERTIES 2-1
2.1 Sample Description and Preparation 2-1
2.2 Diffusion Coefficient Measurements 2-2
2.3 Permeability Coefficient Measurements 2-4
2.4 Related Diffusion and Permeability Coefficient Measurements 2-7
3. LABORATORY MEASUREMENTS OF RADIUM AND RADON
EMANATION 3-1
3.1 Radium and Emanation Measurement Methods 3-1
3.2 Radium and Emanation for Concrete Samples 3-1
3.3 Radium and Emanation Measurements for Florida Concrete
Constituents 3-2
4. RADON TRANSPORT PROPERTIES OF CONCRETE CONTAINING
PHOSPHOGYPSUM 4-1
4.1 Background Information 4-1
4.2 Sample Preparation and Curing 4-2
4.3 Analytical Tests Performed 4-3
4.3.1 Diffusion Coefficients 4-3
4.3.2 Permeability Tests 4-3
4.3.3 Radium Content and Emanation 4-4
5. DATA INTERPRETATION AND MODELING 5-1
5.1 Estimate of Water-to-Cement Ratio 5-1
5.2 Estimate of D for Florida Concretes 5-1
5.3 Estimate of K for Florida Concretes 5-3
5.4 Indoor Radon Entry from Florida Concretes 5-3
iii
-------
CONTENTS
(Continued)
Paee
6. QUALITY ASSURANCE FOR CONCRETE ANALYSIS 6-1
6.1 Radium Concentration Measurements 6-1
6.2 Radon Emanation Measurements 6-5
6.3 Diffusion Coefficient Measurements 6-6
6.4 Permeability Coefficient Measurements 6-8
7. SUMMARY AND CONCLUSIONS 7-1
REFERENCES R-l
iv
-------
FIGURES
Number Page
2-1 Time dependent radon diffusion apparatus 2-3
2-2 Plot of pressure decrease with time for sample C002F 2-5
5-1 Regression of ambient-moisture radon diffusion measurements
on water/cement ratio of concrete 5-2
5-2 Regression of ambient-moisture air permeability measurements
on the dry bulk density of concrete 5-4
5-3 Diffusive contributions to indoor radon concentrations for varying
soil radon sources and five different radon diffusion coefficients 5-7
6-1 Relative uncertainties in radium determinations computed from
gamma ray counting statistics as a function of radium concentration 6-1
6-2 Individual radium measurements on the IPL standard in QC chart format 6-4
6-3 Relative uncertainties in radon emanation determinations by the
radon effluent method, as computed from gamma ray counting statistics 6-5
v
-------
TABLES
Number Page
1-1 Previous radium concentration and radon emanation coefficient
measurements in concrete 1-2
1-2 Previous diffusion coefficient measurements in concrete 1-3
2-1 Physical properties of concrete samples 2-1
2-2 Radon diffusion coefficients of Florida concrete samples 2-4
2-3 Air permeability coefficients of Florida concrete samples 2-6
2-4 D & K measurements by Snoddy as Part of FRRP 2-7
3-1 Ra and E values for Florida concrete samples 3-2
3-2 Relative amounts of constituents of dry Florida concrete mixes 3-2
3-3 Constituents for pozzolan concretes 3-3
3-4 Radium concentrations and emanation coefficient measurements
on concrete constituents 3-3
3-5 Radium and emanation measurements for samples M-5 and M-6 3-4
3-6 Radium concentration and emanation coefficient values for solid
concrete samples from dry mixes 3-5
4-1 Radiological properties of phosphogypsum concrete constituents 4-2
4-2 Diffusion coefficient measurement of phosphogypsum concrete 4-3
4-3 Radiological properties of phosphogypsum concrete 4-4
6-1 Comparison of duplicate radium assays to estimate analytical
precision 6-2
6-2 Analyses of standard reference material for ^Ra 6-3
6-3 Replicate analyses of a blank sample for "Ra 6-4
6-4 Comparison of duplicate radon emanation measurements 6-6
6-5 Comparison of duplicate radon diffusion coefficients 6-7
6-6 Diffusion measurements on standard reference materials 6-7
6-7 Comparison of duplicate air permeability coefficients 6-8
vi
-------
Section 1
INTRODUCTION
Indoor radon entry has been modeled most commonly as advective transport by pressure-driven air flow
from the soil through foundation openings or cracks. The flow is caused by the typically-negative indoor
pressure compared with that in the soil and the outdoor atmosphere. Radon generated in the concrete
floor and radon diffusion from the soil through the concrete floor has generally been ignored. Recently,
attention has been directed toward the importance of diffusion as a significant mechanism for radon entry.
In particular, Scott and Gordon (1) have identified radon diffusion through concrete as a possible
significant source of indoor radon, and Tanner (2) identified radon diffusion as the dominant entry
mechanism when foundation soil permeabilities are less than 7xl0"12 m2. Rogers and Nielson (3) also
identified diffusion through concrete floors and the contiguous soil as a significant mechanism for radon
entry for many soils under typical long-term average foundation pressure gradients. Loureiro et al, (4)
have compared theoretical diffusive and advective radon transport in soils to estimate conditions when
diffusion is insignificant.
While the diffusive radon flux through concrete floors is much smaller than the advective flux through
cracks in the floor, the predominance of the intact floor area over the crack area may compensate for the
difference in fluxes. Thus, it is desirable to examine the diffusive properties of concretes used in
dwelling floors to better assess this mode of radon entry. It also is instructive to characterize the relative
importance of radon generated within the concrete to determine whether aggregates or other concrete
components may contribute significantly to indoor radon concentrations. Very little relevant data on
concrete exist in the general literature.
1.1 SCOPE
This report characterizes the radon generating properties of Florida concretes. The work was conducted
by Rogers & Associates Engineering Corporation as part of the Florida Radon Research Program (FRRP)
cosponsored by the Florida State Department of Community Affairs and the U.S. Environmental
Protection Agency. The parameters measured are the radium concentrations and emanation coefficients
of Florida concretes and their constituents. The report also identifies the main properties of concrete that
influence radon migration from the subsoil into dwellings. The parameters characterizing radon transport
through concrete are the diffusion coefficient, the porosity, and the permeability coefficient. The report
then examines the relation of the measured properties to other physical properties of the concretes.
Finally, it examines the relative importance of the concrete properties, including radium concentrations,
to radon entry into dwellings. The radon entry correlations are based on the laboratory data, on a simple
indoor radon balance equation, and on a complete numerical analysis of combined diffusive-advective
radon entry.
1.2 BACKGROUND
The literature contains several references for the radium concentration (Ra) and radon emanation
coefficient (E) in concretes. In 1981, a group at the National Bureau of Standards published a
comprehensive review of relevant data prior to that time (5). Table 1-1 contains representative values
of Ra and E from the literature. The radium-226 concentration is generally less than 1 pCi/g and the
radon emanation coefficient is around 15 percent. Only the data in references 10 and 11 are for concretes
in the United States. References 10 and 17 also report Ra and E values for concrete which contains
phosphogypsum or phosphate slag. The last two entries of Table 1-1 contain the data for this type of
1-1
-------
concrete. Although Ra contents were not reported for the slag in Reference 10, they must be less than
1 pCi/g. Reference 17 reports the same Ra concentration in the phosphogypsum as the concrete, even
though the phosphogypsum comprises only 47 percent of the concrete.
TABLE 1-1. PREVIOUS RADIUM CONCENTRATION AND RADON EMANATION
COEFFICIENT MEASUREMENTS IN CONCRETE
Year
Ra (pCi/g)
E (percent)
Comment
References
1971
2.4
6
1971
2.0
7
1974
0.4
8
1979
5-25
No measurement details given
9
1983
0.49+ 0.19*
21+5*
114 measurements
10
1985
1.0
Compilation of U.S. values
11
1986
0.54
20
12
1988
0.3-2.2
10^0
Only ranges are given
13
1989
0.78+0.40"
Compilation of worldwide
values
14
1990
0.53-1.4
5.2-8.8
15
1991
0.54, 0.50
15, 4.2
16
1991
0.61 ± 0.50"
14±lfr
Concretes from several
Portland cement mixes
17
1983
0.20
16
Phosphate slag concrete
10
1991
19.2
9
Phosphogypsum concrete
17
a. Standard deviation of reported means.
1-2
-------
Previously reported values for the radon diffusion coefficients, D, of concretes are listed in Table 1-2.
Most of the data are reported as a diffusion length, L, where
L = v^D/X" <1_1>
and
L = diffusion length (cm)
D = diffusion coefficient (cm2 s"1)
X = radon decay constant (2.06x10* s"1).
Equation (1-1) was used to obtain the value of D when L was given in the literature. When available,
the porosity, p, of the samples is also presented in Table 1-2.
TABLE 1-2. PREVIOUS DIFFUSION COEFFICIENT MEASUREMENTS
IN CONCRETE
Year
Diffusion Coefficient
D (xlO4 cm*/s)
Porosity
p (percent)
Comments
Reference
1976
1.2-3.3
5-25
One combined pD value
measured, p given as a
range
18
1978
1.1
26
High density concrete
19
1983
3.3, 6.0
7, 32
High density concrete
20
1988
0.7-8.2
General review article
13
1988
0.7, 1.7
Low w/c concrete for
long-term vaults
21
The mean value of the measurements is 2.8x10"* cm2 s"1. This is probably lower than for typical
concretes presently used in U.S. construction.
Data for the permeability coefficient of air in concrete, K, are available mainly because many types of
concrete structures must be airtight under a specified internal pressure. Typical values of K for intact
concrete range from 10"14 to 10"12 cm2 (22,23).
1-3
-------
13 REPORT CONTENTS
Section 2 of this report contains the diffusion coefficient and permeability coefficient results as well as
a description of the concretes tested. This section also contains D and K measurements male by Acurex
(24) as part of the FRRP. Section 3 contains the radium concentration (Ra) and radon emanation
coefficient (E) measurements and results for the concrete and for concrete constituents. Measurements
on phosphogypsum concrete are reported in Section 4. The concrete correlations and supporting model
analyses are presented in Section 5. Hie quality assurance of data for all analysis is presented in Section
6, and Section 7 contains a summary and conclusions.
1-4
-------
Section 2
LABORATORY TESTS OF RADON TRANSPORT PROPERTIES
Radon diffusion (D) and permeability (K) coefficient tests were performed on samples from thirteen
cylindrical concrete test specimens from Florida. Duplicate concrete samples were obtained from two
of the test specimens, awl two samples, M-5 and M-6, were made from a concrete mix from Florida,
giving a total of 17 samples on which D and K measurements were made. Samples M-5 and M-6 have
a processed gypsum pozzolan additive. The processed gypsum comprises 8 and 15 weight-percent of the
mix, respectively. A detailed description of the preparation of samples M-5 and M-6 is given in Section
3.3.
2.1 SAMPLE DESCRIPTION AND PREPARATION
The physical properties of the samples are given in Table 2-1.
TABLE 2-1. PHYSICAL PROPERTIES OF CONCRETE SAMPLES
Sample
Water/Cement
Density
ID
Ratio
(g cm*^
Porosity
F-l
0.60
2.15
0.20
F-2
0.60
2.14
0.21
F-3
0.61
1.99
0.26
F-4
0.61
2.00
0.26
M-5
0.53
2.15
0.20
M-6
0.36
2.30
0.15
C002F
0.53
2.11
0.22
C003F
0.67
2.00
0.26
C004C
0.66
2.08
0.23
CQ05C
0.58
2.06
0.24
TC1-C
0.60
2.19
0.17
TC1-1
0.60
2.17
0.18
TC2-4
0.60
2.18
0.17
CC008
0.60
2.12
0.26
CC013
0.56
1.94
0.28
CC015
0.56
1.97
0.27
CC033
est 0.57
2.12
0.21
The water-to-cement ratios (W/C) for the F, C, and CC series samples were supplied by concrete plants
in Florida (24). The T series samples are from the radon entry test cells constructed and operated on the
property of the Florida Institute for Phosphate Research in Bartow. The CC series are samples from
2-1
-------
concrete slabs in Florida. Water-cement ratios were estimated for samples missing this information using
the correlation (25):
W/C - 1.93 - 0.64d , (2-D
where
d = concrete density (g/cm3).
If density values were not available, they were obtained from laboratory measurements of the mass and
volume of the samples. The porosities were calculated from the relationship
p = 1 - d/G , <2~2>
where
p = total sample porosity (citf/cm3)
G = solids density (assumed to be 2.7 g/cm3).
The concrete samples were 10 cm in diameter and ranged from about 5 to 10 cm in thickness.
To prepare the samples for the diffusion and permeability measurements, they were epoxied into standard
diffusion sample holders (26) using an epoxy which has negligibly low diffusion and permeability
coefficients. The air permeability measurements were also made with the concrete samples in the
diffusion sample holder to minimize disruptive handling of the samples.
2.2 DIFFUSION COEFFICIENT MEASUREMENTS
Figure 2-1 contains a sketch of the equipment. The sample is placed on a large radon source and an
alpha detector is placed on the top end of the sample. At zero time, the valve of the radon source is
opened and the radon diffuses upward through the concrete sample and into the detector. The time
dependence of the detector counts indicates the time dependence of the radon diffusion through the
sample. The resulting measured diffusion coefficients are given in Table 2-2. The diffusion coefficients
range from 1.8X10"4 cm2 s"1 to about 4.6xl0"3 cm2 s'1. Uncertainties associated with the measurements
range from 20 to 30 percent. Except for sample C002F, the D generally increases with increasing W/C
ratio.
2-2
-------
ALPHA
SCINTILLATION
DETECTOR
BIAS
SAMPLING
PORT
MIXING
FAN
MOIST
URANIUM
TAILINGS
RADON
SOURCE
PRE-AMP
SAMPLING
PORT
MULTI-CHANNEL
SCALER
TIMER
GATE.
VALVE\
SAMPLE
COLUMN
PRINTER
iii#
PERFORATED
TUBES
RAE-104545
Figure 2-1. Time dependent radon diffusion apparatus;
2-3
-------
TABLE 2-2. RADON DIFFUSION COEFFICIENTS OF FLORIDA
CONCRETE SAMPLES
D
Sample No.
(an1 sec*1)
W/C
F-l
1.0x10*
0.60
F-2
2.9xl0"3
0.60
F-3
1.3xl0"3
0.61
F-4
1.2x103
0.61
M-5
5.5x10^
0,53
M-6
2.9xl0~*
0.36
TC1-C
6.3X10"4
0.60
TC1-1
2.2x10"*
0.60
TC2-4
1.8x10"*
0.60
C002 F
4.6x103
0.53
C003 F
3.3x10*
0.67
C004 C
3.9xl0"3
0.66
€005 C
3.5xl03
0.58
CC008
3.1X10"4
0.60
CC013
1.0x103
0.56
CC015
l.OxlO'3
0.56
CC033
8.6x10"*
0.57
2.3 PERMEABILITY COEFFICIENT MEASUREMENTS
The permeability coefficients were measured using a procedure and equipment similar to that developed
by Snoddy et al. (24). Hie technique measures the decrease in pressure with time in a pressurized air
chamber due to air leakage through the concrete sample. The equipment consists of a small pressure
vessel with the sample in its holder forming one end of the pressure chamber. Pressures in the pressure
chamber were measured with a PX182-060GI pressure transducer and a DM160 data-logger with
compatible accessories.
As shown in Figure 2-2, the pressure decreases with time according to the expression
P(t) - P. exp(-t/T) , (2-3)
where
P(t)
T
2-4
initial pressure
pressure at time, t
relaxation time.
-------
0 .
120
160
200
TIME (min)
RAE-103755
Figure 2-2. Plot of pressure decrease with time for sample C002E
Resultant Permeability is 3.4 x 10"^ cm^
2-5
-------
Solution of the pressure balance equation reveals that (24):
P K A
(2-4)
or
K - (2-5)
P.TA
where
H = viscosity of air
P. = atmospheric pressure
Vc = enclosed volume under pressure
A = area of sample
x = concrete sample thickness.
This approach assumes a log-linear pressure drop throughout the concrete sample. Deviations from
log-linearity increase the uncertainty associated with the reported K values.
The results of the permeability measurements are given in Table 2-3. In general the Florida concrete
samples had very low permeabilities, with none having a value greater than 7xl012 cm2. These values
are consistent with the 1.0x1014 cm2 to 3.0xl0"12 cm2 values reported by Hansen et al. (23).
TABLE 2-3. AIR PERMEABILITY COEFFICIENTS OF FLORIDA
CONCRETE SAMPLES
Sample No.
K (cm1)
F-3
4.4xl0"12
M-5
5.0x10"
M-6
7.9x10-"
TC1-C
6.5xl0'12
TC1-1
8.0xl014
TC2-4
8.7xl014
C002 F
3.4xl012
C003 F
3.6x10-"
C004 C
4.0xl012
€005 C
4.1xl0"12
CC008
2.1xl0*13
CC013
IJxlO*12
CC015
3.3x10'12
CC033
8.8x10-"
2-6
-------
2.4 RELATED DIFFUSION AND PERMEABILITY COEFFICIENT MEASUREMENTS
The D ami K measurements by Snoddy (24), also part of the FRRP, were made with equipment similar
to that used in the measurements reported here. The results of Snoddy's measurements are given in Table
2-4. In general, Snoddy's values for D and K are within a factor of two of the present results, which
is within the experimental uncertainties of the measurements.
TABLE 2-4. D AND K MEASUREMENTS BY SNODDY AS PART OF FRRP
Sample No.
w/c
d (g/cm3)
P
pD
(xlO4 cm'/s)"
K
(xlO13 an1)
cooo
0.66
0.12
1.8
21
cooi
0.61
0.20
3.8
71
C002A
0.53
2.09
0.19
3.1
8,8xl0+3k
C002B
0.53
2.09
0.20
3.1
4.6xl0+3b
C003A
0.67
2.04
0.15
2.6
20
C003B
0.67
2.07
0.17
3.8
20
C004A
0.66
2.14
0.14
3.8
13
C004B
0.66
2.15
0.14
3.8
11
COOSA
0.58
1.96
0.17
2.6
21
C005B
0.58
2.04
0.14
2.1
9.4
C006
0.60
3.8
11
C007
0.60
4.6
6.9
C008
0.60
0.82
1.3
C010
0.56
4.6
10
C011
0.56
4.6
11
C012
0.56
9.9
25
C013
0.56
9.9
24
C014
0.56
3.8
11
C015
0.56
4.6
13
C016
0.56
6.6
14
CO 17
0.56
4.6
11
C018
0.56
5.6
7.2
C019
0.56
1.8
5.7
C020
0.56
4.6
9.1
C021
0.56
4.6
27
C022
0.56
3.8
23
C023
3.8
17
C024
8.2
16
C025
4.6
27
C026
4.6
23
C027
3.8
(continued)
2-7
-------
TABLE 2-4. (continued).
Sample No. W/C d (g/em3) p (xlQ4 cmVs)* (x!0u cm1)
C028
5.6
23
C030
3.8
10
C031
3.8
11
C032
2.6
11
C033
2.1
13
€034
6.8
17
a. Snoddy reports only pD for samples C006-C034, and only p,d for samples C000 through C005B.
b. Air passageways through these samples were sufficiently large that the airflow was audible.
2-8
-------
Section 3
LABORATORY MEASUREMENTS OF RADIUM AND RADON EMANATION
Concrete floors in buildings generate radon that can enter the dwelling in addition to the radon transmitted
from the underlying soils into the dwelling. The importance of the concrete floor and walls as an indoor
radon source depends mainly on the radium concentration and the radon emanation coefficient (E) in the
concrete, This section presents the results of radium-226 measurements and radon emanation coefficient
measurements in Florida concretes and in concrete constituents.
3.1 RADIUM AND EMANATION MEASUREMENT METHODS
Florida concrete and concrete constituents have relatively low Ra concentration and E. While the
methods in the FRRP Procedures Manual (26) may be sufficient for the low Ra values expected if long
counting times are used, they do not have sufficient sensitivity for the low E measurements.
Consequently, new methods were used to measure the Ra and E of the Florida concretes and their
constituents. These methods were developed for Ra and E measurements on the soils tested in the
Mapping Project and are documented in Reference 27,
Basically, the Ra is determined using the sealed can gamma method with a large 7,6 cm by 12.7 cm Nal
detector to increase counting efficiency and decrease sample counting time. The gamma counts under
the radium decay-chain peaks are corrected for interferences from the thorium decay-chain gammas.
The emanation coefficient is determined by extracting the free radon from the sealed can into a Lucas cell
and counting to determine the radon-222 concentration. Adjustments are made for pressure differences
that occur with this method. Comparison tests demonstrate the significantly improved sensitivity of this
method over the dual gamma counting method for measuring E.
3.2 RADIUM AND EMANATION FOR CONCRETE SAMPLES
The Ra and E measurements were made on some of the Florida concrete samples used for the D and K
measurements. The results are shown in Table 3-1. AH Ra values are given in terms of dry weight.
The Ra ranges from 1.0 pCi/g to about 2.4 pCi/g. These values are consistent with the values reported
in Table 1-1.
The E values are less than 0.08 which is less than the average values given in Table 1-1. One difficulty
occurred in measuring the E's for some of concrete samples, which calls into question the reported result
for sample C002F. The air extraction needle used to penetrate the sealed can would occasionally become
blocked by the surface of the concrete sample, thus restricting the air and radon flow into the Lucas flask.
This could potentially decrease the measured value of E, and it is likely the cause of the very low E value
for C002F.
3-1
-------
TABLE 3-1. Ra AND E VALUES FOR FLORIDA CONCRETE SAMPLES
Sample No.
Ra (pCi/g)
E
C002F
1.60 ± 0.24
0.019
1.60 ± 0.06
CO03F
1.18 ± 0.23
0.063 ± 0.012
1,44 ± 0.05
C004C
2.36 ± 0.06
0.057 ± 0.002
2.25 ± 0.22
C005C
1.69 ± 0.06
0.072 ± 0.003
1.56 ± 0.24
TC1-1
1.00 ± 0.10
0.075 ± 0.007
TC1-C
1.08 ± 0.09
0.078 ± 0.007
TC2-4
0.96 ± 0.10
0.070 ± 0.007
3 J RADIUM AND EMANATION MEASUREMENTS FOR FLORIDA CONCRETE
CONSTITUENTS
Dry concrete mixes were obtained from manufacturing facilities in the Jacksonville, Lakeland, Tampa,
and Pennsacola areas in order to measure the Ra and E of the constituents and compare them to the values
for the mixed concrete. The concrete mixes were sieved to separate the aggregate, sand, and cement
components. Hie relative compositions of the dry mixes are given in Table 3-2. Each component was
then sealed for die Ra and E measurements. In addition, water was added to the samples M-l through
M-4 to form solid concrete samples with water-to-cement ratios of 0.S0. Radium and radon emanation
measurements were also made on these samples.
TABLE 3-2. RELATIVE AMOUNTS OF CONSTITUENTS OF DRY
FLORIDA CONCRETE MIXES
Percentage
Sample No.
Location
Cement
Sand
Aggregate
M-l
Lakeland
16.4
45.0
38.6
M-2
Tampa
17.6
45.5
36.9
M-3
Jacksonville
14.6
77.3
8.1
M-4
Pensacola
21.7
43.5
34.8
3-2
-------
The effect of a processed gypsum pozzolan was investigated using the concrete constituents. The relative
amounts (weight percent) of constituents using the processed gypsum are given in Table 3-3. The Ra and
E values for the concrete constituents are given in Table 3-4. The cement powders have the highest
emanation coefficients, and, except for sample 1, the highest radium concentrations. Aggregates for
samples M-l and M-2 have radium contents exceeding 1 pCi/g; however, except for sample M-3, the
aggregates have very low E values. The low E's decrease the importance of the radium as an indoor
radon contributor. All the sands have low Ra and relatively low E.
TABLE 3-3. CONSTITUENTS FOR POZZOLAN CONCRETES
Constituents for Pozzolan Concretes (weight percent)
Sample No.
Cement
Processed
Gypsum
Sand
Aggregate
Water
M-5
M-6
10.8
13.5
7.2
13.5
45.1
31.5
27.0
31.5
9.9
9,9
The E values in Table 3-4 are reported for moist cement paste. This matches more closely the
environment in the concrete. Emanation coefficients were also measured for dry cement powder. These
measured emanation coefficients, given in the footnote to Table 3-4, also include the effects of radon
adsorption on the cement powder, causing the E values to be extremely low (8). For example, if the
adsorption coefficient for the dry cements is 5 cm3 g"1, then the actual E^y would increase by about a
factor of 6 over the reported values.
TABLE 3-4. RADIUM CONCENTRATIONS AND EMANATION COEFFICIENT
MEASUREMENTS ON CONCRETE CONSTITUENTS
Sample No.
Ra/® «
Cement
Sand
Aggregate
M-l
1.07 / (0.32)00
0.14 / (0.14)
1.82 / (0.04)
M-2
2.02 / (0.26)®
0.14 / (0.19)
1.29 / (0.03)
M-3
0.87 / (0.39)(b)
0.08 / (0.16)
0.10/(0.17)
M-4
0.98 / (0.29)«
0.20 I (0.08)
0.17/(0.04)
a. Radium 226 concentrations in pCi/g, emanation coefficient given in parentheses.
b. E values for moist cement paste. The respective E values for dry cement powder in samples M1,
M2, M3, and M4 are 0.024, 0.020, 0.041, and 0.011.
3-3
-------
The Ra and E values for the constituents of samples M-5 and M-6 are given in Table 3-5.
TABLE 3-5. RADIUM AND EMANATION MEASUREMENTS FOR SAMPLES
M-5 AND M-6
Ra/(E)*
Sample No.
Cement
Processed
Gypsum
Sand
Aggregate
M-5
1.3
0.55
0.14
0.52
(0.32)
(0.18)
(0.14)
(0.08)
M-6
1.3
0.55
0.14
0.52
(0.32)
(0.18)
(0.14)
(0.08)
a. Ra-226 concentrations in pCi/g, emanation coefficient given in parenthesis.
The Ra and E for the solid concrete samples are given in Table 3-6. Also shown in Table 3-6 are derived
Ra and E values based on the Ra and E values of the constituents. All Ra values are expressed in unite
of pCi per gram dry weight. However, in obtaining the derived values, the bound water in the solid
concrete samples is included in the computations. In general, the derived Ra values agree to within the
experimental uncertainties with the measured values; however, they average about 13 percent lower. The
E values also agree with the measured values to within experimental uncertainties. Except for M-5 and
M-6 the derived values average about 17 percent higher. It is expected that the emanation values derived
from constituents would be higher than the solid concrete sample values due to the chemical reactions and
binding processes that decrease the porosity and increase the fraction of radon recoil atoms that are
retained in adjacent particles.
3-4
-------
TABLE 3-6. RADIUM CONCENTRATION AND EMANATION COEFFICIENT
VALUES FOR SOLID CONCRETE SAMPLES FROM DRY MIXES
Derived
Sample No. Ra (pCi/g) E Ra E
M-l A
B
C
M-l Ave.
1.24 ± 0.17"
1.29 ± 0.16
1.18 ± 0.17
1.24 ± 0.17
0.087 ± 0.012*
0.087 ± 0.011
0.097 ± 0.014
0.090 ± 0.012
0.93 ± 0.24
0.098
M-2 A
B
C
M-2 Ave.
1.37 ± 0.17
0.94 ±0.11
0.86 ± 0.12
1.06 ± 0.14
0.081 ± 0.016
0.117 ± 0.014
0.178 ± 0.025
0.125 ± 0.019
0.89 ± 0.27
0.132
M-3 A
B
C
M-3 Ave.
0.47 ± 0.19
0.41 ± 0.15
0.61 ± 0.13
0.50 ± 0.16
0.151 ± 0.061
0.169 ± 0.063
0.120 ± 0.026
0.147 ± 0.053
0.20 ± 0.27
0.303
M-4 A
B
C
M-4 Ave.
0.19 ± 0.14
0.36 ± 0.17
0.15 ± 0.09
0.23 ± 0.14
0.407 ± 0.291
0.201 ± 0.094
0.562 ± 0.326
0.390 ± 0.258
0.35 ± 0.24
0.191
M-5
M-6
0.52 ± 0.11
0.63 ± 0.11
0.11 ± 0.21
0.17 ± 0.02
0.43
0.51
0.15
0.15
a. Quoted uncertainties are one-standard deviation from Poisson counting statistics.
3-5
-------
Section 4
RADON TRANSPORT PROPERTIES OF CONCRETE
CONTAINING PHOSPHOGYPSUM
One of the project's objectives is to determine the properties of and impacts from concretes that have
constituents elevated in radium. Phosphogypsum was selected as an additive to concrete constituents to
investigate this effect.
Phosphogypsum is a by-product of the phosphate fertilizer industry's wet acid production of phosphoric
acid. The main component of phosphogypsum is calcium sulfate dihydrate. According to Chang (28),
recent studies have concluded that central Florida has in excess of 500-million tons* of phosphogypsum
currently stockpiled, and that by the year 2000 the total may exceed 1-billion tons. In recent years,
efforts have been underway to find useful applications for this readily available and abundant resource.
One of the more promising potential uses for phosphogypsum is a cement additive. Preliminary
research reported by Chang & Mantell (29) has shown that concrete made with phosphogypsum exhibit
strength properties that are sufficient for many applications (such as road construction). This past
research has focused primarily on the mechanical properties of phosphogypsum containing concretes, but
relatively little work has been performed to examine the radiological properties of these concretes. Thus
the properties of concrete which affect gaseous migration through concrete, permeability and diffusion
coefficients, need to be characterized for concretes containing phosphogypsum. With the measured levels
of activity in phosphogypsum, reported by Chang & Mantell (29), between 17 pCi/g - 25 pCi/g, radon
generation and release from the concrete itself also needs to be addressed.
4.1 BACKGROUND INFORMATION
Research reported by Chang & Mantell (29) has shown that concretes containing phosphogypsum exhibit
very good mechanical properties when produced under high compaction pressures, and when the correct
amount of overall moisture is present. In order to test concretes which would be representative of the
real world, selections were based on data reported by Chang & Mantell (29), and on conversations with
Dr. W. F. Chang (30). Dr. Chang has conducted much of the initial research on phosphogypsum
concretes at the University of Miami. The percentages of sand, aggregate, cement, and phosphogypsum
were selected to yield a concrete with a compressive strength greater than 3,000 psi.* The selections also
provided concretes with different levels of phosphogypsum that will allow any differences in properties
due to phosphogypsum to be noted and addressed.
The constituents used for the phosphogypsum concrete are the same as those given in Table 3-3, except
the processed gypsum was replaced by phosphogypsum. The sets were chosen to represent realistic
percentages that could be used by industry to provide a concrete with good properties. The
phosphogypsum was supplied from the Southern Research Institute.
In order to produce a useful concrete containing phosphogypsum, the mixture must be compacted. The
typical method for preparing impact compacted concrete is the Modified Proctor Method (30). The
modified proctor method has been used in most of the initial characterization work on phosphogypsum
concrete, and was used to produce samples for this study.
(*) 1TM = 907 Kg; IPsi = 6.89 kPa.
4-1
-------
4.2 SAMPLE PREPARATION AND CURING
Table 4-1 shows the radium concentrations, and emanation coefficients for the concrete constituents.
TABLE 4-1. RADIOLOGICAL PROPERTIES OF PHOSPHOGYPSUM
CONCRETE CONSTITUENTS
Constituent Ra (pCi/g) E
Cement 1.32 0.32
Sand 0.14 0.14
Aggregate 0.52 0.08
Phosphogypsum 23.7 0.22
The phosphogypsum was tested for free moisture content. To prevent calcination of the phosphogypsum,
a sample was dried at 85°C for two days to drive off any free moisture. The free moisture content of
the phosphogypsum is 18.6 percent. The sand mixture was also tested for free moisture, and the results
indicated that there was no free moisture in the sand.
The following procedure, prescribed by Chang, was used to mix the concrete specimens:
1. The phosphogypsum should be ground to minimize lumps.
2. Combine the appropriate amount of phosphogypsum with sand. The sand will
help prevent the phosphogypsum from reforming lumps.
3. Mix in the correct amount of gravel to the mixture.
4. Add the cement powder to the mixture.
5. Add the correct amount of water to the mixture. The water should be added
slowly to ensure that the entire mix is uniformly wetted. Be sure to include the
free water and crystalline water in the calculation of the optimum moisture.
6. Compact the mixture according to the Modified Proctor Method (30).
7. Upon completion, wrap the samples in a plastic membrane to prevent water loss
and cure at room temperature (approximately 25 *C).
4-2
-------
Four samples from each set of parameters were produced. Three of the four samples were for diffusion
and permeability testing, and the fourth sample was used for radium content and emanation testing.
4.3 ANALYTICAL TESTS PERFORMED
4.3.1 Diffusion Coefficients
The samples, three from each set, were tested to determine the radon diffusion coefficient of the concrete.
The results of the diffusion coefficients measurements completed to date are presented in Table 4-2.
TABLE 4-2. DIFFUSION COEFFICIENT MEASUREMENT OF
FHOSFHOGYPSUM CONCRETE
Sample
Identification
Diffusion
Coefficient (xlO4
cm2/s)
Permeability
Coefficient (xlO
cm2)
SET 1, SAMPLE #1
3.5
7
SET 1, SAMPLE #2
8.6
7
SET 1, SAMPLE #3
4.9
9
SET 2, SAMPLE #1
18
200
SET 2, SAMPLE #2
35
210
SET 2, SAMPLE #3
10
200
The results for D from the phosphogypsum concrete fall within the range of measurements on regular
Florida concretes, given in Table 2-2. The phosphogypsum does not appear to have a significant impact
on die concrete's ability to hinder radon migration via diffusion.
4.3.2 Permeability Tests
The permeability tests were performed on three samples from each parameter set. The results of the
permeability tests are also given in Table 4-2.
The results for K from phosphogypsum concrete set 1 fall within the range of the previous tests.
However, the results for set 2 are about a factor of five higher than the upper range of the K values for
regular concretes given in Table 2-3. This is believed to be caused by small phosphogypsum clumps in
the concrete.
4-3
-------
4.3.3 Radium Content and Emanation
The radium content and emanation coefficient of the phosphogypsum concrete were measured. The
results of the measurements are provided in Table 4-3.
TABLE 4-3. RADIOLOGICAL PROPERTIES OF PHOSPHOGYPSUM CONCRETE
Radium Content Emanation
Sample (pCi/g) Coefficient
Set 1 2.3 0.18
Set 2 4.0 0.087
The radium content of the concrete is identical to the predicted values based on the constituent data of
2.3 pCi/g and 4.0 pCi/g for sets 1 and 2, respectively.
Therefore, the measurements for the concrete containing phosphogypsum yield radium concentrations
slightly higher than for normal concretes, D values that are consistent with and K values that slightly
exceed the ranges of values for normal concretes.
4-4
-------
Section 5
DATA INTERPRETATION AND MODELING
Several correlations and simple models can be obtained from application of the measured
data. This section identifies correlations for the water-to-cement ratio, the diffusion and
permeability coefficients, and the radon entry from concrete floors into structures. The radon
entry correlation is compared to radon entry from a concrete floor as calculated with the
RAETRAD code (3).
5.1 ESTIMATE OF WATER-TO-CEMENT RATIO
Frequently W/C of concrete is not known. For concretes that have about the same sand plus
aggregate fraction, W/C should be related to the dry density.
The following linear least-squares regression resulting W/C with d was presented previously
(25):
W/C = 1.93 - 0.64 d , (5-1)
where
W/C = water-cement ratio
d = bulk dry density of concrete (g/cm3).
The correlation coefficient associated with this expression is r=0.90. Equation (5-1) was used
to estimate the W/C for samples TC1-C, TC1-1, TC2-4, and CC033.
5.2 ESTIMATE OF D FOR FLORIDA CONCRETES
Nielson and Rogers (25) have previously given a correlation between the measured values of
D for concrete and then W/C ratio. The regression equation is
D = 1.5x10 "6 exp(11.4 W/C) . (5*2)
The data and correlation are shown in Figure 5-1. The correlation coefficient for the ln(D)
vs. W/C regression is r=0.82. Data from the literature (18,12) are also included in Figure 5-1.
The linear relationship between W/C and d, given in Equation (5-1) suggests that a useful
correlation also exists between D and d. Least squares regression correlating these variables
gives (25):
D - 2.6x104 exp(-8 d) , (5_3)
with r=0.77, where as before.the r applies to In (D) vs. d regression.
5-1
-------
10
-2
<0
CM
E
o
1cr
c
©
o
£
c
o
*55
3
3=
c
o
10"
10*"
a RAE
•
* Acurex
a
D
a
o Reference 18
~
¦ Reference 21
~
• /
*t
-
op^ a
-
a
•
~ >
a
¦
¦ ~
a
¦ *
a
m
¦
0.3S 0.40 0.45 0.50 0.55 0.60
Concrete Water/Cement Ratio
0.65 0.70
RAE-1Q4S31
Figure 5-1. Regression of ambient-moisture radon diffusion measurements
on the water/cement ratio of concrete.
5-2
-------
5.3 ESTIMATE OF K FOR FLORIDA CONCRETES
The permeability data do not exhibit the same definite trends with W/C as do the diffusion
coefficient data. Much of the scatter in the data are due to experimental errors and
uncertainties. Nevertheless, some benefit can be gained from a correlation between K and
W/C. The least squares analysis gives (25)
K = 1.6x10"16 exp(15 W/C) . <5'4>
The correlation coefficient associated with this lit is r=0.75.
A slightly better fit is obtained for the correlation between K and d. This expression is
K = 0.22 exp(~12.4 d) , <5-5)
with r=0.80. This correlation and associated data are shown in Figure 5-2.
5.4 INDOOR RADON ENTRY FROM FLORIDA CONCRETES
In general, the calculation'of radon generation and transport through soil and concrete into
dwellings is complex and involves multidimensional models such as RAETRAD (3). However,
for Florida concretes, advection is negligible and the total radon generation rate per unit area
is small compared to the radon generation rate per unit area in the subsoil. Under these
conditions, the radon flux from the concrete floor can be estimated separately and can be
added to the diffusive indoor flux from the subsoil (32). The flux from radon generated in the
concrete is estimated by
F = RadE^XD tanh
/
\
X
— xc
D ,
(5-6)
where
Ra = Ra-226 concentration (pCi/g)
X = radon decay constant (s*1)
xc = thickness of concrete slab (cm).
For the range of Florida concretes studied in the present work, d, E, D and xc can be
approximated by average values, so that Equation (5-6) can be expressed as
Qc = Ra As / 28 , (5-7)
where
5-3
-------
-11
-12
«M
E
o
>»
a
«
hm
l
~ RAE
~ ACUREX
<
1.9
2.0
2.1
2.2
2.3
2.4
Concrete Density (g/cm*3)
RAE-104532
Figure 5-2. Regression of ambient-moisture air permeability measurements
on the dry bulk density of concrete.
5-4
-------
Qc = radon entry rate from concrete slab (pCi/s)
Ag = area of concrete slab (m2)
28 = units conversion factor and constants (m2 s/g).
Use of Equation (5-7) is illustrated for the concrete sample C0Q4C. For Ra=2.31 pCi/g, and
a house area of 141 m2, the radon entry rate from radon generation in the concrete is 13
pCi/s. For comparison, comprehensive RAETRAD calculations yield an entry rate of 13 pCi/s
for radon generated in the concrete slab. This value is about 6 percent of the radon entry
rate from the subsoil, where the subsoil is a loamy sand with a radium concentration of
2 pCi/g.
A simple estimate of the radon diffusion through the slab from the subslab radon is obtained
in a similar manner as Equation (5-7)
Qs - Css As / 2000 , (5-8)
where
Qs = radon entry rate through concrete from subslab soil gas radon (pCi/s)
CS5 = subslab radon concentration (pCi/L)
2000 = units conversion factor and constants (s L"1 m+2).
The significance of indoor radon entry by diffusion through concrete floors can be estimated
from a simplified approximation of the indoor radon balance equation. The approximation
assumes that all indoor radon enters via the concrete foundation area, and that the indoor
volume is uniformly diluted at the continuous rate of ky with clean air having an insignificant
radon concentration:
(Qc * Qs) = CjnV^ (5-9)
where
Cin = steady-state indoor radon concentration (pCi/L )
V = indoor volume (L)
X/ = ventilation rate of indoor volume (s"1).
For a simple slab-on-grade house geometiy typical of Florida construction, Equation (5-9) can
be simplified further by introducing Equations (5-7) and (5-8) and setting the height of indoor
volume equal to 2.3 m. This leads to the expression:
Cin = [15.5 Ra * 0.22 C8S] / (lOOOXv) (5-10)
where
Xy = ventilation rate of indoor volume (ach).
5-5
-------
Based on the approximate separability of diffusive and advective radon entry into dwellings,
Equation (5-10) can be used directly to estimate the component of the indoor radon
concentration that results from diffusion through and from the concrete floor slab. The
diffusive radon flux through the slab was estimated by repeated analyses with the RAETRAN
code (33) in which a 10-cm slab separated an indoor radon concentration of 2 pCi/L from
5 m of sandy foundation soil that had varying source strengths corresponding to deep-soil
radon concentrations of 100 pCi/L to 10,000 pCi/L (Css = 75 pCi/L to 7,500 pCi/L ).
Various diffusion coefficients also were used for the slab, which had a fixed porosity of
p=0.23. The radium concentration in the concrete first was assumed to be 1 pCi/g , and to
have a radon emanation coefficient of 0.25.
The resulting diffusive component of the indoor concentrations, from the RAETRAN code
calculations, shown in Figure 5-3, start to exceed the 1 pCi/L level for elevated subslab
radon concentrations (several thousand pCi/L ) when concrete diffusion coefficients exceed
about IxlO"3 cmV1 or higher. The assumption of separability of diffusive and advective
entry generally made a difference of about 5 percent in the entry rates. Even if a 1-cm
perimeter floor crack is assumed for the house, diffusion rates through the intact part of the
slab are affected by less than 25 percent, mainly from altered gradients near the perimeter.
5-6
-------
7 5
100
Sub Slab Radon Concentration (pCi L*1)
750 7,500
Diffusion Coefficients
of Concrete (cm2 s*1)
1.E-2
O —
*-0
c a.
0)*-'
c c
o o
Qw55
E 2
o §
0) o
> c
*2> °
aO
£ c
Q-i
«
a
o
o
"D
c
10
0.1
Ra-226 in Concrete
= 1 pCi g*1
2.E-3
3.E-4 .
7.E-5
1.E-5
0.01
100
1000
10000
Radon Concentration in Soil Gas (pCi L*1)
RAE -104533
Figure 5-3. Diffusive contributions to indoor radon concentrations for
varying soil radon sources and five different radon diffusion
coefficients.
5-7
-------
Section 6
QUALITY ASSURANCE FOR CONCRETE ANALYSIS
The quality assurance of all analyses is determined by three data quality parameters: precision, accuracy,
and completeness. The following sections present the summary statistics of the analytical results in terms
of these data quality parameters for radium assays, radon emanation, radon diffusion, and permeability
coefficients. Completeness of the laboratory tests is estimated from the total number of measurements
compared to the total number of samples available for testing. On this basis, the completeness percentage
for all analyses was 100 percent.
6.1 RADIUM CONCENTRATION MEASUREMENTS
There are no numerical data quality indicators for radium concentration measurements in concrete due
to the lack of prior testing and of radium levels and variability in concrete. However, comparisons
between the data quality indicators for soils (27) and concrete illustrate the precision of this method. The
precision of the radium determinations is defined as the relative measurement uncertainty. Figure 6-1
presents all of the radium concentration measurements with their associated relative uncertainties. The
relative uncertainties are computed from one standard deviation Poisson gamma-ray counting statistics
and then divided by the measured radium concentration in order to express them in relative,
dimensionless units for plotting. Thus 10"1 in Figure 6-1 represents 10 percent standard deviation.
10'
o
E
3
T3
cO
ra
w
.£
c
3
O
O
05
©
O
c
0
,>
1 11 1 11
o
o
' A °
• A o°
>20% Uncertainty,, 0
0
o 2 pCi/g
-1
10*
1 0
Radium Concentration (pCi/g)
figure 6-1. Relative uncertainties in radium determinations computed from
gamma ray counting statistics as a function of radium concentration.
As Figure 6-1 illustrates, all the uncertainties fall below an uncertainty line of ±20 percent at > 2 pCi
g"1. As expected, numerous measurements are associated with higher uncertainties, but these are all in
a radium range low enough to approach the detection limit.
A second estimate of the precision of the radium determinations is based on comparing the results of
duplicate assays for a selected number of samples. Almost 34 percent duplicates were counted in the
6-1
-------
radium determinations, which corresponds to 12 samples with duplicate counts. Table 6-1 summarizes
these results. The final column lists the differences between the duplicate assay results and those given
in the report. The average difference, indicating net bias, is -0.06 with a sample standard deviation of
0.22, using a Student's t test. Thus, the data are consistent with the hypothesis of a zero bias. ITie
average absolute difference is 0.16 pCi g"1. The relative standard deviation between the duplicate
measurements is 14.6 percent if the entire set of measurements above the detection limit is included.
However, if only measurements above 1 pCi g"1 range are included, the relative standard deviation
between the duplicate pairs of measurements is only 6.0 percent. The relative standard deviation is
computed as (34):
RSDdufi - /2n E (x, - x2)2 / £ (xt + xj
where
RSD^ = relative standard deviation among duplicates
xt = first observation
x2 = second observation
n = number of pairs being compared.
TABLE 6-1. COMPARISON OF DUPLICATE RADIUM ASSAYS TO ESTIMATE
ANALYTICAL PRECISION
Sample No.
Duplicate Assay
Radium ± unc.*
(pCi g1)
Reference
Radium
(pCig1)
Difference
(pCig*)
C002F
1.60 ± 0.24
1.60
0.00
C003F
1.18 ± 0.23
1.44
-0.26
C004C
2.36 ± 0.06
2.25
0.11
C005C
1.69 ± 0.06
1.56
0.13
M-1B
1.29 ± 0.16
1.24
0.05
M-1C
1.18 ± 0.17
1.24
-0.06
M-2B
0.94 ± 0.11
1.37
-0.43
M-2C
0.86 ± 0.12
1.37
-0.51
M-3B
0.41 ± 0.15
0.47
-0.06
M-3C
0.61 ± 0.13
0.47
0.14
M-4B
0.36 ± 0.17
0.19
0.17
M-4C
0.15 + 0.09
0.19
-0.04
Average difference (pCi g"')
Average absolute difference (pCi g"1)
Relative standard deviation (all detected)
Relative standard deviation (> 1 pCi g*1)
-0.06
0.16
14.6 percent
6.0 percent
a. One standard deviation uncertainty based on gamma-ray counting statistics.
The agreement of similar analyses with standard reference material demonstrates the accuracy of the
radium concentration measurements. The Isotope Products Laboratory (IPL) reference material was
6-2
-------
spiked into several powdered quartz sand aliquots to prepare the standard at 1014 ± 16 pCi g"1. The
material was sealed in a can similar to those used for the concrete samples and was analyzed regularly
at the beginning and end of each batch of samples. Table 6-2 presents the results of these analyses.
TABLE 6-2, ANALYSES OF STANDARD REFERENCE MATERIAL FOR ^Ra
mRa ± S.D.
IPL Spike
(pci rl>
Difference from
Reference
mRa ± S.D.
Difference from
Reference
1020 ± 1
6
1025 ± 2
11
1020 ± 2
6
1019 ± 1
5
1024 ± 2
10
1024 ± 2
10
1016 ± 2
2
1023 ± 2
9
1018 ± 2
4
Average ± Standard Deviation:
1021 ± 3
7 ± 3
Average Relative Bias:
0.69 percent
As Table 6-2 indicates, the accuracy of the IPL standard measurement compares favorably with the actual
value ami uncertainty of the reference material. The positive bias is not statistically significant at the 90
percent confidence level. The positive bias was relatively consistent for all measurements. This is
probably the result of a small bias in either the calibration or reference radium concentration. An
estimate of the average relative bias for the standard is positive at 0.69 percent. The estimate of the
average relative bias is computed as:
Bias - 100[E(xm - xj / n] / xre, (6-2)
where
Bias = average relative bias (percent units)
x,,, = measured value (pCi g"1)
xref = reference value (pCi g"1)
n = number of measurements.
Figure 6-2 presents a plot of the IPL standard analyses in a control chart format. The control chart
illustrates the 3a computed confidence interval shifted slightly toward the upper portion of the uncertainty
range of the reference value for the standard. None of the individual measurements exceeded this
interval, suggesting good control of measurement system operation and variations.
6-3
-------
1040
5" 1030
I
C
¦B 1020
to
"E
o 1010
O
E
.2
j§ 1000
ec
990
0 2 4 6 8 10
Analysis Number
figure 6-2. Individual radium measurements on the IPL standard In QC chart
format.
Upper Confidence Limit: 1030 pCi/g
Measured Mean: 1021 pCi/g
¦
J V/ /
! Reference Value:
1014 ±16 pCi/g >
¦
- Lower Confidence Limit: 1012 pCi/g
—
-
Analyses of blanks utilized a 300 g aliquot of onyx rock that had previously been determined by extended
counting to contain negligible quantities of radium or thorium (<0.1 pCi g*1). The blank sample was
sealed in a can similar to those used for die concrete samples and was counted repeatedly during the
sample analyses. Table 6-3 presents the results of these counts. The average measured quantity of
radium in this sample is 0.0 ± 0.1, well within the measured standard deviation of 0.2 pCi g"1.
TABLE 6-3. REPLICATE ANALYSES OF A BLANK SAMPLE FOR ^Ra
Average Measured Value:
Average Uncertainties:
^Ra ± uncertainty
(pCi g'1)
-0.1
±
0.3
0.2
±
0.3
0.1
±
0.3
-0.1
±
0.1
-0.2
±
0.1
0.0
±
0.1
0.0
±
0.1
0.2
6-4
-------
6.2 RADON EMANATION MEASUREMENTS
As with radium concentration measurements, there are no numerical data quality indicators for radon
emanation measurements in concrete due to the lack of prior testing. However, comparisons between
the data quality indicators for soils (27) and concrete illustrate the precision of this method.
Measurements of radon emanation by the radon effluent method were analyzed for analytical precision
by plotting their standard deviations computed from counting statistics versus the radium concentration.
Standard deviations in the emanation measurements were expressed on a relative basis as AE/E
(uncertainty/mean). These precisions are plotted in Figure 6-3, illustrating the generally-successful
achievement of die precision goal (only two data points in the upper right quadrant). Most of the samples
that contained radium concentrations less than 2 pCi g'1 had relative standard deviations greater than IS
pa-cent due to the increased uncertainty in measuring the emanation coefficient. This is illustrated in the
figure below by observing the data points that appear in the upper left quadrant.
c
.2
co
c
05
E
til g
c *=
O «
*-5
co B
cc w
C CD
*"* .5
¦25* c
C 3
-c o
a> c
o E
= 2
_j »*-
a>
>
JS
co
EC
10
10l
10
-1
M 10
-2
0
: o
o
o _
" 4 °
' ° X)
>15% RSD &
o ;
o
<15% RSD o wo ^
; * *
I o
<2 pCi/g
<
o '
>2 pCi/g
10" 1 1Qu 10
Sample Radium Concentration (pCi/g)
figure 6-3. Relative uncertainties in radon emanation determinations by the
radon effluent method, as computed from gamma ray counting
statistics.
Another estimate of the precision of the radon emanation determinations was demonstrated by replicate
tests conducted on samples M-l, M-2, M-3, and M-4. The samples wore divided into three separate
concrete plugs (A, B, and C) and were sealed in cans as described previously. These results are
presented in Table 6-4. The standard deviations for all samples, excluding set M-4, do not exceed 4
percent which indicates a high degree of precision between the duplicate measurements. Sample M-4,
6-5
-------
however, had a relative standard deviation of 14.8 percent which is due to the extremely low radium
content and thus a high associated relative uncertainty as shown previously in Table 6-1.
TABLE 6-4. COMPARISON OF DUPLICATE RADON EMANATION
MEASUREMENTS
Measured Emanation Relative
Coefficient Standard Deviation
Sample No. (percent)
M-1A
0.087
±
0.012
M-1B
0.087
±
0.011
M-1C
0.097
±
0.014
M-2A
0.081
±
0.016
M-2B
0.117
±
0.014
M-2C
0.178
±
0.025
M-3A
. 1^51
±
0.061
M-3B
0.169
±
0.063
M-3C
0.120
±
0.026
M-4A
0.407
±
0.291
M-4B
0.201
±
0.094
M-4C
0.562
±
0.326
The accuracy of the radon emanation measurements could not be directly evaluated because there are
no emanation standards for concrete materials. The accuracy of this method was found acceptable for
soil samples in a previous report (27).
63 DIFFUSION COEFFICIENT MEASUREMENTS
The precision of the transient radon diffusion measurement technique was directly evaluated by replicate
measurements performed on samples F-l, F-2, F-3, and F-4. Sample set F-l and F-2, as well as set
F-3 and F-4, originated from the same concrete core and were cut in half to permit duplicate testing.
Table 6-5 lists measurements with their associated uncertainties. The uncertainties are based on earlier
reported values that analyzed numerous soil sample replicates for diffusion coefficients with varying
6-6
-------
moisture saturations (35). The samples with moisture saturation levels below 50% were observed to have
uncertainties around 12 to 20%, Uncertainties increased to about a factor of two as full saturation was
reached. The uncertainties associated with the diffusion measurements range from 20 to 30%.
TABLE 6-5. COMPARISON OF DUPLICATE RADON DIFFUSION
COEFFICIENTS
Measured D
Sample No. (xIO3 an1 s"1)
_ l.o ± 0.3
F-2 2.9 ± 0.9
F-3 1.3 ± 0.4
F-4 1.2 ± 0.4
The results from the duplicate samples, F-3 and F-4, are well within estimated experimental error.
Comparing samples F-l and F-2, however, reveals that the diffusion coefficient for the latter is almost
a factor of three higher. Hi is difference may be influenced by settling of aggregate in the cement core
from which the specimen was taken.
In order to verify the accuracy of the diffusion coefficient measurements, two diffusion tests were
conducted using dry sand and one using uniform glass beads. These diffusion measurements allow
comparisons with theoretically-derived diffusion coefficients. As Table 6-6 indicates, the percent
difference between these standard measurements does not exceed 7%.
TABLE 6-6. DIFFUSION MEASUREMENTS ON STANDARD REFERENCE
MATERIALS
Sample No. Measured D Standard D Percent
(cm2 s'1) (cm1 s"1)* Difference
0.059 0.063 6.3
0.064 0.063 1.6
0.070 0.071 1.4
a. Reference 35.
Dry Sand
Dry Sand
Glass Beads
6-7
-------
6.4 PERMEABILITY COEFFICIENT MEASUREMENTS
Hie methods for determining air permeability coefficients are similar to those developed by Snoddy et
al. (24), Since these procedures are rather new, limits of precision and accuracy are still considered
experimental. In order to verify die precision of data, duplicate analyses were conducted on samples
CC013 and CC033. Table 6-7 summarizes these results.
TABLE 6-7, COMPARISON OF DUPLICATE AIR PERMEABILITY
COEFFICIENTS
Sample No. Measured K (cm2) Percent Difference
_____ 1,7x10" 5^6
CC013 lJxlO"12
CC033 8.8x10*" 3.4
CC033 Z.SxiQf13
The duplicate permeability coefficients demonstrate the precision of this technique with respect to the
reproducibility of its measurements. The differences in these values do not exceed 6% which signifies
a strong agreement between the duplicate tests.
The accuracy of the permeability data cannot be directly measured due to the lack of test standards for
this new procedure. The simple equation used to determine the permeability coefficient assumes that the
log of die pressure drop throughout the concrete sample is linear with respect to time. This is not
necessarily valid for some initial pressure ranges. However, using this assumption, deviations from
log-linearity increase the uncertainty associated with the reported K value. A correlation coefficient, r,
was calculated for each sample by comparing the pressure drop to a log-linear best fit curve as shown
in Figure 2-2. The values for r ranged from about 0.94 to 1.00, demonstrating the excellent correlation
between the data points for each measurement. The linearity of the pressure drop on a semi-log scale
gave an indication of any leakage in the measurement system that would suggest the need for retesting.
6-8
-------
Section 7
SUMMARY AND CONCLUSIONS
The Florida concretes tested generally have radium concentrations less than 3 pCi/g, and emanation
coefficients less than 0.08. These values are consistent with literature values for a wide variety of
concretes. When the radium concentration exceeds 1 pCi/g it may either be due to the radium in the
cement or the aggregate. However, the aggregate has very low E values, rendering its radium Iks
important than radium in the cement component.
The measured diffusion coefficients for the Florida concretes ranged from about 2x1ft4 to 5x103 cmV1.
These values are consistent with previous values in the literature, but extend the upper limit of the range
by about a factor of five. The measured air permeability coefficients ranged from about 9x10"" to 7xl0"12
cm2. The D and K values measured by Snoddy in a cooperative effort are consistent with the present
measured values.
The correlations that are presented for W/C, D, K, and Qc are useful and provide sufficient accuracy for
general scoping studies. Example calculations of radon entry into a dwelling indicate that concretes with
radium content less than about 2 pCi/g contribute less than 10 percent of the total radon entry.
7-1
-------
REFERENCES
1. Scott, A.G. and G.W.S. Gordon, "Radon Diffusion through Concrete," Health Physics Soc.
Conf. June 1978.
2. Tanner, A.B., "The Role of Diffusion in Radon Entry Into Houses," In: The 1990 International
Symposium on Radon and Radon Reduction Technology, Vol. 2, EPA/600/9-91-026b (NTIS
PB91-234450), U.S. Environmental Protection Agency, Research Triangle Park, NC, 1991.
3. Rogers, V.C. and K.K. Nielson, "Benchmark and Application of the RAETRAD Model," In:
The 1990 International Symposium on Radon and Radon Reduction Technology, Vol. 2,
EPA/600/9-91-026b (NTIS PB91-234450), U.S. Environmental Protection Agency, Research
Triangle Park, NC, 1991.
4. Loureiro, C.O., L.M. Abriola, J.E. Martin, and R. G. Sextro,"Three-Dimensional Simulation
of Radon Transport into Houses with Basements Under Constant Negative Pressure,"
Environmental Science and Technology 24: 1338-1348, 1990.
5. Colle, R., etal., "Radon Transport Through and Exhalation from Building Materials," NBS, TN
1139, September 1981.
6. Hamilton, E.I., "Relative Radioactivity in Building Materials," Journal Amer. Ind. Hyg. Assn.,
Vol 32, pg 398, 1971.
7. Krisiuk, E.M., et al., "A Study of Radioactivity in Building Materials," National Comm. on
Utilization of Atomic Energy, Leningrad, 1971.
8. Kolb, Ul., "Influence of Building Materials on the Radiation Dose to the Population,"
Kerneneigie and Offenlichkeit, Vol.4, pg 18, 1974.
9. Stranden, E., "A Simple Method for Measuring the Radon Diffusion Coefficient and Exhalation
Rate from Building Materials," Health Physics, Vol. 32, pg 242, 1979.
10. Ingersol, J.G., "A Survey of Radionuclide Contents and Radon Emanation Rates in Building
Materials Used in the U.S.," Health Physics, Vol. 45, pg 363, 1983.
11. New, A., et al., "Characterizing the Sources, Range, and Environmental Influence of Radon-222
and its Decay Products," Science of Total Environment, Vol. 45. PI 233, 1985.
12. Passchier, W.F., P.J. Klijn, and H. B. Van Der Heijde, "A Proposed Standard for Radioactivity
in Building Materials Used in the Netherlands," Health Physics, Vol. 51, pg 661, 1986.
13. Stranden, E., "Building Materials as a Source of Indoor Radon," in Radon and Its Decay
Products in Indoor Air. W.W. Nazaroff, A.V. Nero, Jr., eds, New York, John Wiley, 1988.
14. Quindos, L.S., G.J. Newton, and M.H. Wilkening, "Estimation of Indoor ZBRn from Concrete,"
Health Physics Vol. 56, pg 107, 1989.
R-l
-------
15. Battaglia, A., et al., "Radon Exhalation Rate from Ashes and Building Materials in Italy," Proc.,
Twenty Ninth Hanford Symposium on Health and the Environment," Richland, WA, October 16-
19, 1990.
16. Schuler, R. Creamer i, and W. Burkart, "Assessment of the Indoor Rn Contribution of Swiss
Building Materials," Health Physics Vol. pg 447, 1991.
17. Van Dijk, W. and P. deJong, "Determining the 222Rn Exhalation Rate of Building Materials
Using Liquid Scintillation Counting," Health Physics M, 501, 1991.
18. Culot, M.J.V., H.G. Olson, and K. J. Schiager, "Effective Diffusion Coefficient of Radon in
Concrete, Theory, and Method for Field Measurements," Health Physics 20: 263-270, 1976.
19. Jonassen, N. and J.P. McLaughlin, "Exhalation of Radon-222 from Building Materials and
Wells," National Radiation Environment III. CONF-78044, Vol. 2, pg 1211, 1978.
20. Zapaloac, G.H., "A Time-Dependent Method for Characterizing the Diffusion of 222Rn in
Concrete," Health Physics Vol. 45, Pf 377, 1983.
21. Shuman, R,, V.C. Rogers, and K.K. Nielson, "Measurements of Concrete Properties for Low-
Level Waste Disposal Facilities," Rogers and Associates Engineering Corporation report RAE-
8716-3, Salt Lake City, UT, 1988.
22. Neville, A.M., Hardened Concrete: Physical and Mechanical Aspects. Am Cone. Inst. Monoy.
No. 6, Iowa St. U, Ames, IA, 1971.
23. Hansen, A.J., N.S. Ottosen, and C.G. Peterson, "Gas-Permeability of Concrete In Situ: Theory
and Practice," published by American Concrete Institute in In Situ/Nondestructive Testing of
Concrete. Sp 82, pp. 543 - 556, September 1984.
24. Snoddy, R., "Laboratory Assessment of the Permeability and Diffusion Characteristics of Florida
Concretes, Phase I, Methods Development and Testing," EPA-600/R-94-053 (NTIS PB94-
162781), U. S. Environmental Protection Agency, Research Triangle Park, NC, April 1994.
25. Nielson, K.K. and V.C. Rogers, "Radon Entry Into Dwellings Through Concrete Floors," In:
Hie 1991 International Symposium on Radon and Radon Reduction Technology, Vol 1, EPA-
600/9-9l-037a (NTIS PB92-115351), U.S. Environmental Protection Agency, Research Triangle
Park, NC, 1991.
26. Williamson, A.D. and I.M. Finkel, "Standard Measurement Protocols, Florida Radon Research
Program," EPA-600/8-91-212 (NTIS PB92-115294), U. S. Environmental Protection Agency,
Research Triangle Park, NC, 1991.
27. Nielson, K.K. and V.C. Rogers, "Development of a Prototype Map of the Soil Radon Potentials
in Alachua County, Florida," Rogers and Associate Engineering Corporation report RAE-
9127/3-1, Salt Lake City, UT, October 1991.
28. Chang, W.F., "Reclamation, Reconstruction and Reuse of Phosphogypsum for Building
Materials," University of Miami, Coral Gables, FL, Prepared for the Florida Institute of
Phosphate Research, January 1987.
R-2
-------
29. Chang, W.F. Mid M.I. Mantell, "Engineering Properties and Construction Applications of
Phosphogypsum," University of Miami Press, Coral Gables, FL, 1990.
30. "Standard Test Methods for Moisture - Density Relations of Soils and Soil Aggregate Mixtures
Using 10 lb (4.54 kg) Rammer + 18 in. (457 mm) Drop," ASTM Standard D-1557-78,
Reapproved 1990.
31. American Concrete Institute. Manual of Concrete Practices. Part 1. 6.3.7.1,1990.
32. Rogers, V.C. and K.K. Nielson, "A Handbook for the Determination of Radon Attenuation
Through Cover Materials," U.S. Nuclear Regulatory Commission report NUREG/CR-2340,
December 1981.
33. Rogers, V.C., K.K. Nielson, and G.B. Merrell, "Radon Generation, Adsorption, Absorption,
and Transport in Porous Media," DOE/ER/60664-1, U.S. Department of Energy, Washington,
D.C., 48 pp, 1989.
34. Natrella, M.G., Experimental Statistics. U.S. National Bureau of Standards, Handbook 91,
October 1966.
35. Nielson, K.K., D.C. Rich, and V.C. Rogers, "Comparison of Radon Diffusion Coefficients
Measured by Transient-Diffusion and Steady-State Laboratory Methods," U.S. Nuclear
Regulatory Commission report, NUREG/CR-2875, August 1982.
R-3
------- |