RADON REMOVAL USING POINT-OF-ENTRY
WATER TREATMENT TECHNIQUES
by
Nancy E. Kinner
James P. Malley, Jr.
Jonathan A. Clement
Environmental Research Group
University of New Hampshire
Durham, New Hampshire 03824
Cooperative Agreement CR 812602-01-0
Project Officer
Kim R. Fox
Drinking Water Research Division
Risk Reduction Engineering Laboratory
Cincinnati, Ohio 45268
RISK REDUCTION ENGINEERING LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
The information in this document has been funded wholly or in part by the
United States Environmental Protection Agency under assistance agreement
number CR 812602-01-0 to the New Hampshire Department of Environmental
Services. It has been subject to the Agency's peer and administrative review
and has been approved for publication as an EPA document. Mention of trade
names or commercial products does not constitute endorsement or recommendation
for use.
ii
-------
FOREWORD
Today's rapidly developing and changing technologies and industrial
products and practices frequently carry with them the increased generation of
materials that, if improperly dealt with, can threaten both public health and
the environment. The U.S. Environmental Protection Agency is charged by
Congress with protecting the Nation's land, air, and water resources. Under a
mandate of national environmental laws, the agency strives to formulate and
implement actions leading to a compatible balance between human activities and
the ability of natural systems to support and nurture life. These laws direct
the EPA to perform research to define our environmental problems, measure the
impacts, and search for solutions.
The Risk Reduction Engineering Laboratory is responsible for planning,
implementing, and managing research, development, and demonstration programs
to provide an authoritative, defensible engineering basis in support of the
policies, programs, and regulations of the EPA with respect to drinking water,
wastewater, pesticides, toxic substances, solid and hazardous wastes, and
superfund-related activities. This publication is one of the products of that
research and provides a vital communication link between the researcher and
the user community.
This report presents the results of an evaluation, performed by the
University of New Hampshire - Environmental Research Group (ERG), of radon
removal by 3 point-of-entry water treatment techniques: granular activated
carbon, diffused bubble aeration and bubble plate aeration. This research is
a follow-up to a larger study of radon removal techniques for small community
water supplies. The report discusses each of the treatment alternatives with
respect to their radon removal efficiency, potential problems (i.e., waste
disposal, radiation exposure and intermedia pollution), and economics in
point-of-entry applications.
E. Timothy Oppelt, Director
Risk Reduction Engineering Laboratory
iii
-------
ABSTRACT
Radon is one of the naturally-occurring radionuclides which will be
regulated as part of the Safe Drinking Water Act Amendments of 1986. Several
studies on the distribution of radon in ground water supplies have indicated
that there will be a large number of small public water supplies impacted if a
Maximum Contaminant Level (MCL) is set for radon in the proposed range of 200-
2,000 pCi/L. Although the regulations do not apply to private water supplies
(i.e., individual wells), states normally adopt the Federal regulations and
apply them to individual well water supply- systems. These private supplies
will require point-of-entry (POE) treatment. The purpose of this EPA
Cooperative Agreement was to evaluate the performance of POE granular
activated carbon (GAC), and diffused bubble and bubble plate aeration systems
treating a ground water supply containing radon (35,620 ± 6,727 pCi/L). The
pattern of loading to the units was designed to simulate daily demand in a
household. Each of the systems was evaluated with respect to three primary
factors: radon removal efficiency, potential problems, and economics.
The radon removal efficiency of the POE GAC units gradually deteriorated
over time from 99.7% to 79% for the GAC without pretreatment and from 99.7% to
85% for the unit preceded by ion exchange. It appeared that iron
precipitation in the GAC without ion exchange pretreatment impeded radon
sorption by fouling the GAC surface or causing channeling of water through the
bed. The gamma emissions data indicated that the zone of radon removal slowly
moved down the GAC bed. However, in POE applications where the water has low
to moderate activity, excess GAC can be supplied with little extra cost, so
that the unit may be able to sustain >70% removal of radon for long periods of
time, even with variations in radon loading. At higher radon loadings, the
GAC units may not be able to dampen variations in influent activity and flow
without some increase in effluent activity.
Most GAC systems will require either pretreatment and/or backwashing to
prevent fouling resulting from accumulation of particulates, metal
precipitates, bacteria or organic matter. If frequent backwashing can control
this problem, ion exchange should be avoided because its resin, backwash water
and regenerant brine become contaminated with long-lived radionuclides and may
require special handling and disposal.
GAC POE treatment systems were found to accumulate long-lived
radionuclides, particularly lead-210, which may significantly increase
handling and disposal costs. In addition-, GAC units receiving high radon
loadings will produce gamma emissions that exceed recommended health standards
even when water jacket shields are provided.
The bubble plate and diffused bubble POE units were very efficient (>99%)
at removing radon from the water. They should be able to meet an MCL in the
iv
-------
range 200-2,000 pCi/L in most cases, even with significant variation in radon
mass loading. This resilience is primarily due to the high air to water
ratios supplied by the aeration blowers. One major problem associated with
the aeration techniques is iron oxidation/precipitation. Iron treatment will
be required in almost all POE applications which use aeration. The aeration
units are also prone to operational problems because they rely on blowers and
pumps to function properly. As a result, they require frequent maintenance.
In addition, the off-gas stack should be vented above the roofline to prevent
any radon from re-entering the home.
The total production costs for the GAC units were $9.88/1,000 gallons
without pretreatment and $13.40/1,000 gallons with pretreatment including
disposal costs (assuming the GAC and ion exchange resin are low level
radioactive wastes). The total production costs for the aeration systems were
$22.58/1,000 gallons (diffused bubble) and $26.74/1,000 gallons (bubble
plate). The aeration costs do not include auxiliary blowers for venting, or
air treatment. The productions costs for the POE radon removal systems are
high relative to public water supplies because (i) there is no economy of
scale, and (ii) it is assumed individual homeowners will be purchasing single
POE units unlike utilities which could purchase large quantities of POE units
for a service area. Therefore, there will not be a quantity discount on the
POE equipment for the individual well water supplies.
At low influent radon activities (<5,000 pCi/L), GAC systems are the
preferred treatment technique in POE applications where there are not
excessive concentrations of long-lived radionuclides in the water. GAC is
recommended because it (i) requires the least owner maintenance, (ii) is the
easiest system to operate, and (iii) is the least expensive with respect to
capital and operation and maintenance costs, even considering disposal costs.
At low mass loadings, achieving a high percent removal efficiency is less
critical, and the GAC can therefore meet an MCL in the range 200-2,000 pCi/L.
In addition, at these loadings gamma emissions are minimized. At higher mass
loading rates, aeration techniques would be required to meet an MCL of
200-2,000 pCi/L.
This report was submitted in partial fulfillment of Cooperative Research
Agreement CR812602-01-0 by the Environmental Research Group of the University
of New Hampshire under the sponsorship of the U.S. Environmental Protection
Agency in conjunction with the State of New Hampshire Department of
Environmental Services. This report covers the period January 1989 to January
1990 and work was completed as of- June 1990.
-------
CONTENTS
Foreword iii
Abstract ., iv
Figures viii
Tables xii
Acknowledgments xiii
1. Introduction 1
Research Objectives 2
Fundamentals of Radioactivity 3
2. Conclusions and Recommendations 6
Granular Activated Carbon 6
Aeration Units 7
Overall Evaluation of POE Systems 8
3. Methods and Materials 10
Description of Treatment Units 10
Overall System Design ,. 12
Sampling. Events 14
Sampling and Analytical Techniques 13
Data Analysis 26
4. Results and Discussion 29
Water Flowrate 29
Temperature 30
pH 30
Alkalinity 30
Calcium 36
Turbidity- 39
Iron 39
Manganese 44
Microbial Numbers 54
NPDOC 56
Radionuclide Removal 56
Economics 93
Overall Evaluation of the POE Systems 96
References 101
vi
-------
Appendices „ 105
A. Summary of Quality Control Data 105
B. Summary of Liquid Scintillation Counter Program 113
C. Summary of Raw Water Characteristics at the
Derry, NH POE Site „ 114
D. Methods of Calculation of Lead-210 Adsorption . „ 115
vii
-------
FIGURES
Schematics of the diffused bubble unit and the
bubble plate chamber 13
2. Schematic of the POE systems 15
3. Schematic showing the locations of the alpha track
detectors 19
4. Schematic shoving the locations of the gamma emissions
measurements 27
5. .Water flowrate for the POE treatment systems 31
6. Temperature of the influent water 33
7. pH of the influent water and effluents from the
GAC systems and aeration systems 34
8. Alkalinity for the GAC without pretreatment
and the GAC with pretreatment 35
9. Alkalinity for the bubble plate unit and diffused
bubble unit 37
10. Calcium for the GAC without pretreatment, the GAC
with pretreatment, and aeration systems 38
11. Turbidity for the GAC without pretreatment and
the GAC wi th pretreatment 40
12. Turbidity for the bubble plate unit and the
diffused bubble unit 41
13. Turbidity as a function of total iron for the bubble
plate unit and diffused bubble unit 42
14. Total and soluble iron for.the GAC without
pretreatment 43
15. Total and soluble iron for the GAC with
pretreatment 45
viii
-------
Figures (continued)
Page
Total and soluble iron for the bubble plate
unit 47
17. Total and soluble iron for the diffused
bubble uni t 48
18. Total and soluble manganese for the GAC
system without pretreatment 49
19. Total and soluble manganese for the GAC with
pretreatment 50
20. Total and soluble manganese for the bubble
plate unit 51
21.. Total and soluble manganese for the diffused
bubble unit 52
22. Microbial numbers for the GAC without pretreatment
and the GAC with pretreatment 55
23. Microbial numbers for the aeration systems 57
24. NPDOC for the GAC without pretreatment and the GAC
with pretreatment 58
25. NPDOC breakthrough curves for the GAC units 59
26. Influent radon activity for the POE treatment
systems 60
27. Effluent radon activities for the GAC systems for
the first 4 months and for the entire study 62
28. Radon mass loading rate to the GAC systems 63
29. Radon mass removed as a function of mass applied for
the GAC without pretreatment and the GAC with
pretreatment 64
30. Cumulative radon mass removed as a function of
cumulative mass applied for the GAC without
pretreatment and the GAC with pretreatment 65
ix
-------
Figures (continued)
Number . Page
31. Radon mass removed as a function of mass applied for
the first 6 months and the latter 6 months of the
study for the GAG without pretreatment and the
GAC with pretreatment 66
32. Total radon activity adsorbed over time for the GAC
units 67
33. Radon breakthrough (Ce/Co) as a function of NPDOC
breakthrough (Ce/Co) for the GAC without pretreatment
and the GAC with pretreatment. 69
34. Gamma emissions measurements taken at the unit's
surface for the GAC without pretreatment and the
GAC with pretreatment 70
35. Radon activities for the diurnal study for the GAC
systems 73
36. Ratio of radon mass removed:mass applied for the high
loading rate study for the GAC systems 74
37. Ratio of radon mass removed: mass applied for the
high loading rate study for the GAC systems. 76
38. Ratio of radon mass removed: mass applied for the
high loading rate study for the GAC systems 77
39. Radon activities for the influent and effluent
the GAC systems for the period of August through
November 78
40. Radon mass removal as a function of the mass applied
for the high loading rate condition and for the normal
conditions for the GAC without pretreatment and the
GAC wi th pretreatment 79
41. Influent and effluent radon activities after backwashing
for the GAC without pretreatment and the GAC with
pretreatment 80
42. Gamma emissions measurements for the GAC systems at
1.5 m away from the units'-surfaces 81
43. Uranium for the GAC without pretreatment and the GAC
wi th pretreatment 84
-------
Figures (continued)
Number Pagji
44. Radium for the GAG without pretreatment and the GAG
vi th pretreatment 87
45. Radon activities for the influent and the effluent
from the aeration uni ts 92
xi
'''--~''f•''''- '••' "™
-------
TABLES
Numbers Page
1. Abbreviated uranium (4n + 2) decay series 4
2. Regeneration program for the ion exchange unit 11
3. Daily operating schedule for the POE units 16
4. Sampling materials preparation and preservation 20
5. Locations of survey meter measurements on surface
of GAC units 28
6. Mean water flowrates for the POE treatment units 32
7. Total iron coring data for the GAC units 46
8. Total manganese coring data for the GAC units 53
9. Comparison of predicted lead-210 removal in the GAC
uni ts to the coring data 72
10. Comparison of the predicted uranium adsorbed to the
coring data for the GAC uni ts 85
11. Uranium coring data for the ion exchange unit 86
12. Radium coring data for the GAC and ion exchange
units 88
13. Total uranium and radium in the brine exiting the
ion exchange unit during brine regeneration 91
14. Cost estimates for the GAC POE treatment systems
at Derry, NH 95
15. Cost estimates for the aeration POE treatment systems
at Derry, NH 97
xii
-------
ACKNOWLEDGMENTS
The assistance and guidance of Mr. Kim R. Fox, Project Officer; and Mr.
Thomas J. Sorg both with the Drinking Water Research Division; U.S.
Environmental Protection Agency; Cincinnati, Ohio is very gratefully
acknowledged. Ve also thank Mr. Richard Thayer of the New Hampshire
Department of Environmental Services (NHDES) who served as project manager.
Mr. Charles Larson and Dr. Stan Rydell of U.S. EPA (Boston, MA) served as
liaisons with the Region I office. Our thanks also go to Mr. Peter A. Quern
who worked on the project as a graduate student during the initial phases. Ve
thank Lowry Engineering (Unity, ME) and North East Environmental Products
(Lebanon, NH) for their assistance with the design and Smith Pump (Manchester,
NH) for their assistance with construction at the study site. Ve appreciate
the radionuclide analyses done by the NHDES Laboratory staff, and Dr. Jack
Dibb of the UNH Department of Earth Sciences. Ve thank Southern New Hampshire
Vater Company, Hudson, NH for their cooperation and participation. Finally,
Ms. Claire Simmons and Ms. Elayne Ketel are thanked for their help with
administration and Ms. Alice Greenleaf, Ms. Nan Collins and Ms. Linda Andrews
are thanked for their help with manuscript preparation.
xiii
-------
SECTION 1
INTRODUCTION
As part of the 1986 amendments to the Safe Drinking Water Act, the United
States Environmental Protection Agency (EPA) will propose a rule for Maximum
Contaminant Level Goals (MCLGs) and National Primary Drinking Vater
Regulations (NPDWR) including Maximum Contaminant Levels (MCLs) for
radionuclides in drinking water (Federal Register, 1986). One of the ft
radionuclides that will be regulated under, the proposed rule is radon-222
(hereafter referenced as radon in the text). In a recent status report, EPA
(1989a) indicated that it is considering setting the MCL for radon in the
range 200 to 2,000 pCi/L. There have been several studies of the distribution
of radon in ground water supplies in the U.S. (Hess et al., 1985; Longtin,
1988; Dixon and Lee, 1988; Vitz, 1988), all of which indicate that there will
be a large number of private wells and small (<76 m /day) public water
supplies impacted by an MCL in that range.
The rule will also contain recommendations with respect to Best Available
Technologies (BATs) and analytical methods. As a result, EPA established a
Cooperative Agreement to evaluate the performance of full-scale systems
treating small community ground water supplies containing radon (Kinner et
al., 1990). Though the rule will not directly affect POE water supplies, EPA
also wanted to obtain data on several small-scale radon removal systems, so
that information could be available with respect to Best Available
Technologies (BATs) for those applications. Granular activated carbon (GAC)
adsorption and aeration are the primary processes that have been used to
remove radon from drinking water. The GAC process relies on radon's ability
to adsorb to the carbon. One unique aspect of GAC treatment is that the
breakthrough/exhaustion profile typically seen when GAC is treating
conservative contaminants has not been observed during radon removal. GAC
units have been used to remove radon from drinking water in several POE
applications (Lowry and Brandow, 1985; Lowry et al., 1987; Lowry and Lowry,
1987).
The aeration methods are predicated on the fact that radon is a highly
3 '}
volatile gas with a relatively high Henry's constant (2.80 atm • m H-O/nT air
at 10°C) and therefore can be easily transferred from water to air. Both
diffused bubble and bubble plate aeration have been used in POE applications
(Lowry et al., 1987; LaMarre, 1988 and 1989). Packed tower aeration is rarely
used in these conditions because the height of the tower is too great for
installation in most homes.
Though there are many radioactive isotopes of radon, radon-222 is the most
common.
-------
RESEARCH OBJECTIVES
The purpose of this EPA Cooperative Agreement was to evaluate the
performance of POE GAC, and diffused bubble and bubble plate aeration systems
while treating a ground water supply containing radon. In this study, direct
comparison could be made among the individual POE systems because they were
operated in a parallel flow configuration and were each receiving the same
influent water. The raw water came from an abandoned small community ground
water supply and contained 22,837 to 54,765 pCi/L of radon. Each treatment
system treated 1.02 m /day of water applied in six 18 min intervals and one 30
min interval distributed throughout the day. The pattern of application was
designed to simulate daily demand for a POE system (Anderson and Watson,
1967). Each of these systems (GAC, diffused bubble aeration, bubble plate
aeration) was evaluated with respect to three primary factors: radon removal
efficiency, potential problems (i.e., waste disposal, radiation exposure,
equipment malfunctions, intermedia pollution) and economics. The study was
not intended to develop design equations for these systems, but rather to
evaluate their performance over one year of constant operation.
The specific objectives of the study were to:
Evaluate POE GAC systems, operating with and without ion
exchange pretreatment, monitoring them for changes in radon
removal, radiation emissions, general water quality
parameters (e.g., pH, alkalinity, iron, manganese, turbidity,
microbial numbers, and non-purgeable dissolved organic carbon
(NPDOC));
Conduct several special monitoring events of the GAC systems
to assess the impact of diurnal variations in water flowrate
and raw water quality (i.e., loading rate), and backwashing
on GAC performance;
Core the GAC and ion exchange units after one year of
operation to determine if iron, manganese, microorganisms,
and/or radionuclides were accumulating in the units;
Evaluate POE diffused bubble and bubble plate aeration
systems monitoring them for radon removal, general water
quality parameters and off-gas emissions of radon.
The report is divided into four sections including introduction
(Section 1), conclusions and recommendations (Section 2), methods and
materials (Section 3), and results and discussion (Section 4). Section 4
includes a discussion of (i) the radon removal efficiency of the systems, (ii)
baseline data on other water quality parameters, (iii) potential problems with
their operation, (iv) suggestions for design modifications, and operation and
maintenance requirements, and (v) an economic analysis of each system. Data
from each of the units is discussed and compared in each subsection, where
appropriate. Recommendations were made regarding the use of each system in
POE applications.
-------
FUNDAMENTALS OF RADIONUCLIDES
Radionuclides are different from the majority of drinking water
contaminants regulated by EPA because of the instability of their nucleus and
their ability to decay. This often leads to confusion resulting from the
dissimilarity between the units used to describe radioactivity and those used
for other water contaminants; as well as the physical meaning of these units.
Therefore, a brief synopsis of some of the fundamental principles of
radionuclides important to this project is given below.
Radon-222 is part of the uranium (4n+2) decay series (Table 1) which
originates with uranium-238. Radon and its four short-lived progeny
(polonium-218; lead-214, bismuth-214 and polonium-214) have some of the
shortest half-lives (less than 27 min) of any of the radionuclides in the
series. Radon is the only gas in the decay series and is a member of the
Periodic Chart Group 0 (noble) gases including helium, argon and other
chemically inert monatomic gases. Radionuclides are described in units of
activity (Curies) which indicate the rate at which a radioactive atom is
undergoing decay (i.e., nuclear disintegrations). By convention, one
picocurie (pCi) means that 2.22 atoms of that radionuclide are decaying per
min (disintegrations per min m dpm). A radionuclide's half-life (t,/2) is
the time interval during which 1/2 of the atoms decay (t-,/2 s ln 2/^ wnere X
= decay constant, 0.0075 hr~ for radon). For radon and its progeny, the
half-lives are very short varying from 3.82 days (radon-222) to 0.00016 sec
(polonium-214). The actual life of an individual atom, however, can range
from 0 to infinity (i.e., decay occurs randomly).
Though the units specified for radionuclides in water are in pCi/L (i.e.,
activity per unit volume) and their range in ground water is 10 -106 pCi/L,
the mass and number of atoms present is usually extremely small and can be
determined using the equation: A = NX where A = activity (pCi or dpm) and N =
number of atoms. For example, 1,000 pCi of radon (2,220 dpm) in a liter of
water is only 1.76 X 10 atoms of radon or a mass of 6.49 X 10~15g (radon m
222g/mole) or 6,490 pg/L of water.
Therefore, when designing treatment systems to remove radon from water,
though the activities may be substantial (and have important public health
significance), the numbers of atoms and mass to be removed is very small.
It should also be noted that the mass concentration of radon in ground water
(ex., 1,000 pCi/L • 6.49X10" g/L) is much greater than its concentration in
air (0.1 pCi/L • 6.49X10~19g/D-
Detection capabilities for most drinking water contaminants are in the
—8 —3
10 to 10~ g/L range. The instrumentation used in radon analysis detects
light flashes created when the radioactive particles emitted during decay
strike fluorescing compounds and therefore, is able to measure lower
concentrations (32.5 x 10 g/L = <500 pCi/L ). The instrument's detector
records the number of counts (scintillations) per minute or cpm. The
-------
TABLE 1. ABBREVIATED* URANIUM (4n + 2) DECAY SERIES
Chemical
Notation
238 „
92 '
234 I*,
90?
234 Fa
91
234 ^
92 4,
230 Th
90 4
226 Ra
88
222
^" Rn
86 4,
218 Po
84 4.
214 Pb
82 4.
214 Bi
83 4-
214
^* Po
84 4.
210 Pb
82 4-
210 Bi
83 4,
210 Po
84 4,
206 Pb
82
Historical
Name
Uranium
Thorium
Protactinium
Uranium
Thorium
Radium
Radon
Polonium
Lead
Bismuth
Polonium
Lead
Bismuth
Polonium
Lead
Half-life Radiation Emitted Upon Decay'1*
4.5 x 109 yrs.
24.1
1.17
2.47
days
ndn
x XT yrs.
a
ftY
ftY
«,Y
8.0 x 104 yrs. «,Y
1602
3.82
3.05
26.8
19.7
yrs.
days
ndn
ndn
ndn
164ysec
21 yrs.
5.01
days
138.4 days
Stable
«,Y
a
a
ftY
B,Y
a
ftY
0
a
Protactinium-234 (Uranium Z), Astatine-218, ThaUium-210 and 206 are not
shown. These represent radionuclides created in <0.2% of the decays of
Protactinium-234 (Uranium Xg), Polonium-218, Bismuth-214 and Bismith-210,
respectively.
+ Only listed if % emission is >0.1%.
4
-------
efficiency of the instrument is usually not 100%, so an efficiency factor must
be determined using standards to relate the cpra obtained to the dpm or pCi
actually present.
Atoms of the various radionuclides in a decay series are constantly being
produced and are, in turn, decaying to form other unstable nuclides. For
example, radon is created by the decay of radium-226 (t» = 1602 yr) and
decays to polonium-218 (radon t« = 3.82 days). Because of this continuing
pattern of formation and decay, it is important to understand the concept of
secular equilibrium. Vhen a radionuclide is in secular equilibrium with
another radionuclide or its progeny, there is a constant amount of that
radionuclide present which is a function of the rate at which it is being
created and the rate of its decay. The mathematical relationships are best
explained by Evans (1969). Hence, at secular equilibrium between radon and
its progeny or between radium and radon, the activities of the parent and its
progeny are equal. For example, when 1,000 pCi/L of radium is in secular
equilibrium with radon, the activity of radon will also be 1,000 pCi/L.
-------
SECTION 2
CONCLUSIONS AND RECOMMENDATIONS
GRANULAR ACTIVATED CARBON
1. The radon removal efficiency of the POE GAC units gradually
deteriorated over time from 99.7% to 79% for the unit without
pretreatment and from 99.7% to 85% for the unit with ion exchange
pretreatment. The decreased efficiency did not appear to be
correlated with increases in radon mass loading, or retention of
manganese, NPDOC, uranium or radium. The effect of bacterial growth
on the GAC was not studied. The data suggested that iron
precipitation in the GAC without pretreatment impeded radon sorption
by fouling the GAC surface or causing channeling of water in the top
of the bed. The gamma emissions data indicated that the zone of
radon removal moved down the GAC bed over time. As a result, use of
GAC systems in POE applications should be limited to raw water where
the influent radon activity is low <<5,000 pCi/L). Under these
conditions, the GAC should be able to produce effluent in the range
of the proposed MCL (200-2,000 pCi/L).
2. When GAC is used for radon removal in small community applications,
accounting for variations in water flowrate and raw water radon
activity is crucial in design. In POE applications, these variables
are somewhat less important, especially where the raw water has low
to moderate radon activity because excess GAC can be supplied for
the units with little extra cost. At higher loadings, however, the
POE GAC units may not be able to dampen excess radon applied without
some increase in effluent activity.
3. GAC systems in most POE applications will require either
pretreatment and/or backwashing to prevent fouling resulting from
accumulation of particulates, metal precipitates, bacteria or
organic matter. Ion exchange pretreatment will be essential where
iron and manganese concentrations in the raw water are high.
However, there are serious problems with ion exchange because the
resin, backwash water and regenerant brine may become contaminated
with long-lived radionuclides and may require specialized handling
and disposal. If frequent backwashing of the GAC unit and use of a
sediment filter can control the accumulation of particulates and
metal precipitates in the bed, ion exchange pretreatment should be
avoided.
4. The GAC may retain longer-lived radionuclides, such as uranium-238,
radium-226 and lead-210, in quantities which require disposal in'a
stabilized landfill or a low level radioactive waste facility. As a
result, GAC may not be a viable radon removal technique for raw
6
-------
water with high concentrations of uranium, radium or radon. A more
thorough risk assessment and economic analysis should be conducted
for applications where the raw water has low radon activities to
determine whether the problems associated with disposal are
outweighed by the benefits of GAC (e.g., cost, ease of operation and
maintenance).
5. Gamma emissions from GAC beds were in the range of 2-3 mR/hr at the
surface of the unit. Water jacket shielding only reduced the
measured dose by 17%. However, Rydell et al. (1989) have shown that
in POE applications where the influent radon activity is <5,000
pCi/L, the gamma emissions from the GAC will be greatly reduced to
acceptable levels. In these applications, use of water jackets
could be avoided making access to the unit easier.
AERATION UNITS
1. The bubble plate and diffused bubble POE units were very efficient
(>99%) at removing radon from the water even at the relatively high
activities used in this study (35,620 + 6,727 pCi/L). They should
be able to meet an MCL of 200-2,000 pCi/L in most cases, primarily
because of the high air to water (A:V) ratios used. In addition,
variations in influent radon activity and water flowrate can be
handled without a significant increase in effluent radon activity.
Though the A:V ratios are excessive for most POE applications, the
blowers used in the units are the most cost effective based on
overall efficiency.
2. One major problem associated with the aeration techniques is iron
oxidation and precipitation. Even at concentrations <1.0 mg/L,
precipitates can form and accumulate in the units and/or be released
in pulses to the residence when the unit is started to meet demand.
In the diffused bubble unit, iron accumulation on the diffusers can
cause a decrease in air flow. As a result, iron treatment will be
required in almost all POE applications which use aeration
techniques. Pretreatment with ion exchange creates problems with
radionuclide retention and disposal. Post filtration with sand has
been proposed, but needs to be evaluated before it is used to insure
it is effective and does not have problems with radionuclide
retention, and to determine the frequency of backwashing required.
3. The aeration units are more prone to having operational problems
than the GAC units because they have more mechanical parts (e.g.,
pumps and blowers). These will require frequent maintenance to
insure proper operation. In addition, the blowers and pumps are
very noisy during operation and the low power cut-off feature which
protects the secondary (jet) pump will require manual restarting in
the event of pressure loss (e.g., due to power failure). These
features may be viewed by the owners as a nuisance and lead them to
bypass or disconnect the unit.
• 4. Frequent maintenance is essential to the proper, operation of the
aeration units. The air intake filters to the blowers should be
cleaned frequently to maintain adequate air flow. There should be
7
-------
gauges which the owner can monitor to insure the proper air
pressure/flovrate is maintained. Addition of audible alarms to warn
the owner of potential low air flow or pump failure is necessary.
Soundproofing or placement in a remote location may be required to
reduce the noise during operation. Directions for restarting in the
event of a power failure should be clearly displayed. Finally, the
units should provide easy access to the owner, so that they can be
properly serviced as directed in the operation and maintenance
manual.
5. Proper venting of the off-gas from the aeration units is important.
As demonstrated in this study, the plume is diluted fairly rapidly.
However, the off-gas stack should be vented above the roofline to
prevent any radon from re-entering the home. Placement of auxiliary
blowers may be required to produce adequate flow in these vent
pipes.
OVERALL EVALUATION OF POE SYSTEMS
1. The total production costs for the GAC units were $9.88/1000 gallons
without pretreatment and $13.40/1,000 gallons with pretreatment
including disposal costs (assuming the GAC and ion exchange resin
are low level radioactive wastes). The total production costs for
the aeration systems were $22.58/1000 gallons (diffused bubble) and
$26.74/1000 gallons (bubble plate). The aeration costs do not
include auxiliary blowers for venting or air treatment,. All
production costs were derived for a flowrate of 1.02 m /day and
-35,000 pCi/L influent radon activity. The production costs for the
POE radon removal systems are high relative to public water supplies
because (i) there is no economy of scale, and (ii) it is assumed
individual homeowners will be purchasing single POE units unlike
utilities which could purchase large quantities of POE units for a
service area. Therefore, there will not be a quantity discount on
the POE equipment for these individual well water supplies.
2. There were elevated levels of bacteria in the effluent from all of
the units tested. Depending on EPA and state regulations, it may be
necessary to disinfect the treated water before consumption.
3. Frequent monitoring of the effluent quality from any of the POE
units should be stressed to the owner. Because radon cannot be
readily detected by the senses, there is a potential for the water
to be untreated if a POE unit fails without the homeowner being
aware of the problem. Water analyses at the tap should be conducted
frequently to avoid this problem.
4. At low influent radon activities (<5,000 pCi/L), GAC is the
preferred treatment technique in POE applications where there are
not excessive concentrations of long-lived radionuclides in the
water. GAC is recommended because (i) it requires the least owner
maintenance, (ii) is the easiest system to operate, and (iii) is the
least expensive with respect to capital and operation and
maintenance costs even considering disposal costs. At low mass
loadings, achieving a high percent removal efficiency is less
8
-------
critical, and the GAC can therefore meet an MCL in the range 200-
2,000 pCi/L. In addition, at these levels gamma emissions are
minimized. At higher mass loading rates, aeration techniques would
be required to meet an MCL of 200-2,000 pCi/L.
-------
SECTION 3
METHODS AND MATERIALS
This section describes the treatment systems and outlines the sampling
and analytical methods used in the project. A summary of all data from the
quality control checks made throughout the project is provided in Appendix A.
DESCRIPTION OF TREATMENT UNITS
Granular Activated Carbon FOE Systems
The GAC systems vere designed by Lovry Engineering (Unity, ME). Each GAC
unit consisted of a cylindrical fiberglass contactor (I.D. » 24.8 cm, Height =
137 cm). A three way valve situated on the top of the unit controlled
backwashing and bypassing. Barneby Cheney (Columbus, OH) Type 1002 activated
carbon was used in both GAC units. The unit preceded by ion exchange
pretreatment contained 0.0465 m of carbon and the unit without pretr-eatment
3
contained 0.0470 m of carbon. The units were initially backwashed for 13 min
to remove fines. Raw water entered the top of each unit and was distributed
across the surface of the GAC. The treated water was collected at the bottom
of each unit and flowed up a central FVC pipe and out of the unit.
Both GAC units were preceded by sediment filters designed to remove
particulates from the raw water. These were manufactured by OMNI (Hammond,
IA), and contained 24.5 cm long pleated paper filters (O.D. 6.5 cm) also
manufactured by OMNI. The paper filters were replaced every 2-3 weeks.
An ion exchange unit (45,000 grains) manufactured by Lancaster Water
Softener (Lancaster, PA) was installed between the sediment trap and the
3
second GAC unit. The unit contained 0.042 m of a strong cationic resin
(Cybron C-249-sodium polystyrene sulfonate). It was designed to remove iron
and manganese from the raw water before it entered the GAC. The unit was a
24.8 cm diameter (I.D.) and 119 cm tall fiberglass cylinder and was operated
in a downflow mode. Regeneration with sodium chloride (brine) occurred every
2 weeks in the sequence shown in Table 2. The regeneration cycle was
automated and recommended by the manufacturer.
Diffused Bubble Aeration POE System
The diffused bubble unit was designed by Lowry Engineering (Unity, ME)
and was constructed of molded polyethylene. The unit consisted of three 46 cm
(length) x 40 cm (width) x 24 cm (water depth) tanks in series, each of which
contained an internal diffuser fed from a common header. All air piping was
10
-------
TABLE 2. REGENERATION PROGRAM FOR THE ION EXCHANGE UNIT
Event
Backwash
Brine Regeneration
Rapid Rinse
Brine Refill
Time (min)
10
90
10
16
Flowrate (L/rain)
9.5
3.8
9.5
3.8
Brine was 180 g/L of NaCl.
11
-------
constructed from PVC. The diffusers contained variably-spaced 0.64 mm
3
diameter holes. Air to the diffusers was supplied by a 1.58 m /min capacity
2
blower (Cast; Benton Harbour, MI) with a 1.5 ft polyester cloth intake
filter. A pressure gauge located just downstream of the blower was monitored
to indicate whether the air pressure exceeded the value recommended by the
manufacturer (<26 in. ^0). [N.B., Greater air pressures indicated air flow
out of the diffusers was restricted.] The exhaust collected in the top of the
unit was vented outside using a 5 cm diameter PVC pipe fitted with a wire mesh
to prevent animals from entering the exhaust pipe. Raw water was pumped into
the first chamber of the diffused bubble unit directly from the well. The
water flowed through the unit by simple displacement and was pumped (Myers;
Ashland, OH; jet pump; 32 L/min) from the third compartment to a 76 L
hydropneumatic tank. The solenoid valve, controlling the operation of the
unit, was regulated by high and low water float switches located in the third
chamber. A timer connected to the blower was set to insure that air flow
continued 10 min after cessation of water flow to insure that radon was
removed from all water remaining in the diffused bubble unit. A schematic of
the diffused bubble unit is shown in Figure 1.
Bubble Plate Aeration System
The bubble plate aeration unit was designed by North East Environmental
Products (Vest Lebanon, NH). It was housed in a 60 cm long x 38 cm wide x 23
cm high molded plastic casing with a central 10 cm diameter PVC vent pipe.
Raw water entered the unit directly from the well and was distributed by a
0.95 cm spiral diffuser into a 7.6 cm wide channel constructed of
polyethylene. The bottom of the channel was perforated by 4.8 mm diameter
•3
holes spaced 1.9 cm apart. During operation, air from a'535 m /hr capacity
EBM Industries (Unionville, CT) blower was forced up through these holes. The
blower had a 200 ym polyethylene foam filter in the air intake. The air
caused the raw water column to rise to a total height of 17 cm. The off gas
was vented out the top of the bubble plate chamber via the PVC exhaust pipe.
The flow of water through the channels in the bubble plate unit is shown in
Figure 1. At the end of the channel (length = 270 cm), the treated water
flowed over a 7.6 cm weir and into a 39 cm (height) x 57 cm (length) x 57 cm
(width) polyethylene holding tank located below the bubble plate. Water was
pumped from the holding tank to a 76 L hydropneumatic pressure tank by a
Jacuzzi 1/2 hp jet pump. The flow of water out of the bubble plate was
regulated by a float switch in the holding tank connected to a solenoid valve.
A timer connected to the blower was set to insure that air flow continued 1
min after cessation of water flow.
OVERALL SYSTEM DESIGN
The entire project was conducted at the Scobie Pond ground water well and
pumphouse facility (Derry, NH) owned by Southern New Hampshire Water Company
(Hudson, NH). The facility no longer serves the community so the raw water
could be used for our study. Treated water was discharged to a nearby wetland
area.
12
-------
OIFFUIIR
POUTBT.HYLCNE TANK
T
TO HOUSEHOLD
HOZZLI
FROM WELL
HOLI MTTIKN
TYPICAL
38 om
NOT TO SCALE
TO LOW CM
HOLDIHQ TANK
Figure 1. Schematics of the diffused bubble unit (a) and the bubble plate
chamber (b). Not to scale.
13
-------
At the beginning of the study, water was pumped from the wells with
•3
submersible pumps into a 0.95 m polyethylene equalization tank. A 3/4 hp jet
pump (Myers; Ashland, OH) supplied the raw water to the units (Figure 2). The
tank was designed as an equalization basin for the raw water to dampen
short-term changes in water quality. Use of this tank was discontinued after
2 months because iron was precipitating out of solution during storage. Two
Amtrol (W. Warwick, RI) Well-X-Trol household hydropneumatic tanks, (76 L and
121 L capacity) were installed in place of the tank to equalize the flow to
the units. Copper piping (1.9 cm ID) was used to plumb the entire system.
Samples were collected from seven different locations (Figure 2).
Samples were collected from 1.8 cm brass valves using 1.27 cm (I.D.) clear
polyethylene flexible tubing. All water flowed to a sampling board to
facilitate rapid and efficient collection. Brass valves (1 cm I.D.) located
on the sampling board were used to regulate flow into the 0.64 cm polyethylene
tubing from which the samples were directly collected.
; Approximately 1,022 L of water was distributed to the units each day in
one 30 min and six 18 min sampling intervals using a Dayton Time Switch
(Chicago, IL) (Table 3). The flowrate was adjusted to approximately 7.6 L/min
using a CAL-Q-FLO rotometer (maximum capacity = 38 L/min) (Blue and White
Industries; Westminster, CA) located on the influent lines to the GAC units
and the effluent lines of the aeration units. These operating periods and
flowrate conditions were chosen to simulate POE applications (Anderson and
Watson, 1967). The volume of water treated during a sampling event and
between sampling events was recorded using Hersey (Dedham, MA) MVR water
meters, located on the influent lines for both GAC units and the diffused
bubble unit and the effluent line for the bubble plate unit (Figure 2).
SAMPLING EVENTS
All systems were started on January 7, 1989. The units were monitored
every 24 hr for the first three days of the study. The GAC units were
analyzed every 3 to 4 days for the first month, weekly for the next 5 months,
and biweekly for the remainder of the study. The aeration units were
monitored weekly for the first 6 months and biweekly thereafter. During
routine sampling of all systems, aliquots of water were collected and analyzed
for radon, pH, alkalinity, turbidity, iron, manganese, calcium, microbial
numbers, and temperature. The GAC systems were also sampled for uranium,
radium and NPDOC. -Data collection was completed on January 3, 1990.
Diurnal Variation Studies
The GAC systems were monitored during each operating interval for one day
in February and one day in August to determine the effects of natural diurnal
variations in loading on system performance.'
Loading Rate Studies
A series of studies were conducted to evaluate the efficiency of the GAC
systems during periods of increased loading. In the first study, the flowrate
was held constant at 7.6 L/min, while the total daily throughput was increased
14
-------
HYDRONEAUMATIC
TANKS
I I
i
WELL
SUBMERSIBLE
PUMP
BUBBLE PLATE
SAMPLE
PORT
NFLOW
METER
DIFFUSED BUBBLE
PAPER
FILTER
GAG.
FILTER
PAPER
FILTER
ION GAG
EXCHANGE FILTER
Figure 2. Schematic of the POE systems (not to scale).
15
'= v ..,'*• " ' •,. - " yf-.~- ""^'^•..jX',"-**". ,***•» •*'• • -'•• '.- ''.''" ""Ji'-.^r''';"— ' ~ i,";.!™~ ' '.. ' •' ' ^. . '.
-------
TABLE 3. DAILY OPERATING SCHEDULE FOR THE POE UNITS
Starting Time
6:30 a.m.
7:30 a.m.
8:30 a.m.
12:00 noon'
5:30 p.m.
7:00 p.m.
9:10 p.m.
Operating Time (min)
18
18
30
18
18
18
18
16
~.
S=v^^s^f«g^y.^sav&iJ.£^;.;/s^;, st-.^ffif*
-------
to 1,890 L. This was achieved by increasing the frequency of the 18 min
loading intervals from 6 to 12. Two other studies were conducted by
increasing the flovrate to -20 L/min with a total daily throughput of 1,022 L
(normal conditions) and 1,890 L, respectively. The loading rate tests were
conducted one week apart, during the month of August and repeated again during
October. Samples were collected at each operating interval throughout the day
for all water quality parameters except uranium and radium during the August
trials and for radon, pH, alkalinity, turbidity and total iron and manganese
during October.
Backvashing
A study was conducted to determine if backvashing affected the
performance of the GAC units. The manufacturer (Lovry Engineering)
recommended a backwash flowrate of 19 L/min be used. However, this flowrate
caused a loss of carbon from the unit, so the flowrate was lowered to
approximately 7.6 L/min. This flowrate was maintained for 15 min, as
recommended by the manufacturer, without noticeable loss of carbon. The GAC
without pretreatment was backwashed during the last week of November and the
GAC system with pretreatment was backwashed during the second week of
December.
Samples for radon, turbidity, bacterial numbers, and total and soluble
iron and manganese were taken before backwashing and 1, 3, 24 and 48 hours
after normal operation resumed.
GAC and Ion Exchange Coring Study
Carbon samples were taken from the GAC units at the termination of study
on January 3, 1990. Samples were collected at the top of the bed and at
depths of 43 and 93 cm below the GAC surface for the unit without
pretreatment. For the GAC preceded by pretreatment, samples were collected at
the top of the GAC and at depths of 41 and 93 cm below the surface. Resin
samples from the ion exchange unit were collected at the top of the resin and
at depths of 38 and 76 cm below the surface. Solid carbon and resin samples
were removed by cutting through the units at these depths with a reciprocating
saw and sampling material not exposed to the blade. The samples were analyzed
for iron, manganese, calcium, uranium-238, uranium-235, radium-226, lead-210,
moisture content, and microbial numbers.
Gamma Radiation Shielding Study
At the end of the study, a water jacket, supplied by the GAC system's
manufacturer, was placed around the GAC unit with pretreatment to investigate
the effectiveness of water shielding in reducing gamma emissions. The jacket
was composed of a polyethylene tank (61 cm diameter, 136 cm height) and
provided a layer of water 30.5 cm thick around the entire GAC unit. Gamma
emissions measurements were taken 30.5 cm and 150 cm away from the surface of
the GAC unit before and after the jacket was filled with water. The
measurements were taken at the top of the GAC bed and 41 and 93 cm deep within
the bed.
17
-------
Air Monitoring
Alpha track detectors, type DP (Terradex Corp.; Glenwood, IL), were
placed at various locations inside and outside of the pumphouse to measure
potential increases in the radon activity in the air as a result of the
release of radon from the aeration units. Four detectors were placed 135 cm
beneath the (downward curved) vents from the aeration units. Four detectors
were placed on each outside corner of the pumphouse roof. Ten detectors were
placed at various locations in the wooded area surrounding the pumphouse
(Figure 3) at heights of approximately 1.5 m off the ground. Two detectors
were placed outside a residence approximately 5 km from the site to determine
background levels of radon in the air in Derry, NH.
The detectors were placed inside 0.5 to 1.0 L inverted cans to protect:
them from precipitation. The cans were not sealed so the detector was
directly exposed to the air. The detectors were installed on September 19,
1989 and removed on January 9, 1990. The alpha track detectors were analyzed
in January 1990 by the manufacturer.
SAMPLING AND ANALYTICAL TECHNIQUES
Sampling Methods
Water samples for radon, temperature, pH, alkalinity, turbidity, total
iron and manganese, soluble iron and manganese, calcium, NPDOC, microbial
numbers, total uranium, and radium-226 were collected from the treatment
systems at seven locations (Figure 2). Duplicate samples were collected for
all analyses except total uranium, radium-226, and microbial numbers.
Water samples were collected directly from the plastic tubing connected
to the sampling valves located in the FOE treatment system. The flow exiting
the tubing was controlled using brass valves attached to the sampling board.
The entire FOE system and all plastic tubing was allowed to flush for 10 min
prior to sample collection, to insure that water which had been stagnant in
the system prior to operation was removed.
Sample containers for temperature, pH, turbidity, alkalinity, dissolved
oxygen, total iron and manganese, soluble iron and manganese, calcium, and
NPDOC were cleaned by soaking in soapy water for 20 min, rinsing three times
with deionized water, soaking for a minimum of 20 min in a primary cleaning
solution (Table 4),* rinsing three times with deionized water and once with
double deionized water. The sample containers and caps were rinsed three
times with sample water before collection.
Blanks were taken for all analyses by transporting a 7.57 L jug into the
field. Blanks for microbial numbers were taken by transporting a 250 mL
container of sterilized water into the field. Double deionized water was
obtained by passing deionized water through a. Milli-Q 4 bowl water system
(Millipore; New Bedford, MA).
18
-------
SAMPLE
LOCATION
AERATION EXHAUSTS
NOT TO SCALE
Figure 3. Schematic showing the locations of the alpha track detectors.
19
-------
TABLE 4. SAMPLING MATERIALS PREPARATION AND PRESERVATION
Analysis
Radon
Temperature
NPDOC
Container
Volume Material
(mL)
20
250
20
G*
P"
G*
Preparation
(Primary
Cleaning
Solution)
10% ExtranR
(VWR Scientific)
50% HC1
25% Chromic Acid
Preservation
None
None
HNO,
Hold
Time
4-12 hr+
(in growth)
None
2 wk
pH,
Alkalinity
and Turbidity
250
50% HC1
0.15% (v/v)
4°C (iced) 24 hr
Iron, 60 P*
Manganese ,
and Calcium
Uranium, 3785 P*
Radium
Microbial 60 P*
Numbers
50% HNO-
3
50% HNO-
J
Sterilized
Autoclaved
121°C @ 15 psig
HNO-
0.15% (v/v)
HNO-
to pH<2
4°C (iced)
6 mo
6 mo
24 hr
G • Glass.
"""After extraction via shaking.
P » Polyethylene plastic.
20
-------
Radon
Aqueous samples for radon were drawn from the polyethylene sampling tubes
j^
using a 10 mL Hamilton gastight glass syringe fitted with a 13 gauge needle.
.The syringe was rinsed three times with the sample water before collection. A
10 mL sample, devoid of air bubbles, was slowly injected into a 20 mL glass
low absorbance scintillation vial (Wheaton; Millville, NJ) below the surface
of 10 mL of a toluene-based liquid scintillation cocktail [40.3 mL Liqufluor
(Dupont; Wilmington, DE) per 1 L of scintillation grade toluene].
The samples were analyzed after radon came to secular equilibrium with
its short-lived progeny (preliminary experiments indicated we should wait 4
hrs after collection) and before 12 hr had elapsed between collection and
analysis. Radon samples were analyzed using a Beckman (Fullerton, CA) LS 7000
liquid scintillation counter (LSC). The manufacturer's program 6 was used for
all analysis (Appendix B). A counting time of 10 min was used for all
analyses except for the effluent samples taken from the aeration units. The
counting time was increased to 50 min on these samples because of their low
radon activities. Automatic quench compensation was used with a cesium-137
source. The instrument was calibrated with Beckman tritium, carbon-14 and
background sealed standards. The efficiency factor (E) for the instrument was
determined daily using 10 radium-226 standards ranging in activity from 6,500
to 65,000 pCi/L. The data obtained from these standards was used to make a
linear standard curve. The efficiency factor (E) was calculated from the
slope of the line-of-best fit for the data through the origin. The LSC
provided results in counts per minute (cpm) for two windows: Channel 1=0-
397 (corresponding to a full tritium channel) and a Channel 2 = 397 - 940 (a
tritium - phosphorus-32 channel). The majority of the counts from radon-222
and its progeny were located in the Channel 2 pulse heights. The total counts
(Channel 1 and 2) for each sample were then converted to pCi/L using the
equation:
(Cg - Cfa) x 1000 mL/L
(eq* 1}
where: A » radon activity (pCi/L)
C = cpm of the sample
s
C, = background cpm of field blank (cpm)
E = efficiency factor (cpm/pCi)
D = decay correction factor
V s sample volume (mL)
The decay correction was determined from:
D = e-0.693(T)/V (eq
where: T * time from sampling to counting (days)
t » half -life of radon (3.82 days)
21
-------
Duplicate samples were analyzed every tenth sample to determine precision.
The radium-226 standards and field blanks were counted at the beginning of
each analytical run.
The radium-226 standards were prepared from National Bureau of Standards
(NBS) or EPA (Environmental Monitoring Systems Laboratory; Las Vegas, NV)
primary standards. The standards, which came in glass ampules, typically
contained 5 g of an HC1 solution with radium-226 activities ranging from 4 to
5 nCi/g. The contents of the ampule were emptied into a 50% HNO^ washed and
air dried, glass beaker to facilitate weighing in a Mettler AC-100 analytical
balance. The contents of the beaker were then transferred to a 200 mL
volumetric flask (cleaned using 10% ExtranR, 50% HN03 and double deionized
water) and the beaker was rinsed several times with 0.05M HNOj to insure all
radium-226 was transferred. The stock radium-226 solution was prepared in a
balance of 0.05M HNO-. It was stored in the wax-sealed 200 mL volumetric
flask at 4°C. The stock solution usually contained 117.5 pCi/mL (260.85
dpm/raL) of radium-226. The appropriate volume of the stock solution required
to make the standards was added to a clean scintillation vial using a
volumetric pipette and was diluted with 5 mL of 0.05M HNO-j. The contents of
the vial were purged for 20 min with fine bubbles of laboratory-grade nitrogen
gas. Subsequently, 10 mL of the toluene-based scintillation cocktail were
slowly added to the standard. The scintillation vial was sealed and counted
daily on the LS 7000 to determine when the radium-226 was in secular
equilibrium with radon-222 (-20-25 days). New standards were prepared every 3
months. Once the standards were in secular equilibrium, they were used for ~4
months. If the cpra of the radium standards began to change before that time,
they were no longer used.
Temperature
Temperature was obtained by using Fisher (Boston, MA) Model 14-983-17A
mercury thermometers (Range -20 to 110°C). The thermometer was immersed into
250 mL bottles containing sample water for NPDOC and soluble metals
immediately after collection. The reading was taken after the thermometer
equilibrated (-1-3 min).
The accuracy of the field thermometers was checked monthly against an
ASTM-approved mercury thermometer using boiling water and ice baths. Readings
on the field thermometer were corrected, as needed.
Samples for pH were collected in 250 mL polyethylene bottles. The pH was
measured immediately after arrival at the. laboratory with a Beckman (San
Ramon, CA) Altex 71 pH meter connected to a combination gel pH electrode and a
temperature compensation probe. The meter was calibrated with two VWR
Scientific (Boston, MA) pH standards (pH = 4 and 7), as outlined in Section
423 of Standard Methods (1985). The drift of the instrument was checked by
analyzing the standards every 20 samples or every hour, whichever came first.
If one or both of the standards did not read within 0.1 pH units of the
original value, the meter was recalibrated and the samples were reanalyzed.
22
-------
The precision of the meter was checked every fifth sample. The accuracy of
the meter and calibration method were checked monthly with two NBS pH
standards (pH = 4 and 6.8). Action was taken to repair the meter or replace
the working standards if the measured value differed by more than 0.1 pH units
from the NBS standards.
Alkalinity
Alkalinity was determined potentiometrically as outlined in Section 403
of Standard Methods (1985). The initial pH of the sample was recorded before
the sample was titrated against 0.02 N HjSO^. The titration was continued
until the pH of the sample dropped into the pH range of 4.4 - 4.6. The final
pH of the sample was recorded after the titration was completed. Alkalinity
(mg/L as CaCO,) was determined from the following equation:
^30,000, (eq.3)
where A is the volume (mL) of titrant used and N is the normality of the
titrant. The normality of the titrant was determined monthly using a Na2C03
primary standard. The precision of the analysis was checked every fifth
sample. If the precision exceeded the control limit of +5 mg/L as CaCOj, all
previous analyses were repeated.
Turbidity
R
Samples for turbidity were analyzed on a Hach (Loveland, CO) Model 2100A
turbidimeter, using the method outlined in Section 214 of Standard Methods
(1985). The instrument was calibrated with secondary-sealed standards
supplied by Advanced Polymer Systems (Redwood City, CA). The calibration was
checked between each sample. Precision quality control charts were developed
prior to the beginning of the project using the method outlined by Standard
Methods (1985). Duplicate analyses were performed every fifth sample to check
precision. If the value of precision was unacceptable (as based on QA/QC
criteria - see Appendix A), all previous samples were reanalyzed. Turbidity
samples from the EPA water supply performance evaluation study were analyzed
to determine the accuracy of the analysis.
NPDOC
Aliquots for NPDOC were taken from the 250 mL polyethylene bottles
n
containing the sample using a 60 mL Luer-Lok plastic syringe (VWR
Scientific). The samples were filtered through water-washed GF/C filters into
20 mL glass VOA vials fitted with teflon septa. The syringes were cleaned in
the same manner as the glassware (Table 4), except the syringes were rinsed
with the primary cleaning solution (chromic acid) and not soaked.
The NPDOC samples were analyzed on a Dohrmann (Rosemont Analytical; Santa
Clara, CA) Model DC-80 total organic carbon analyzer, using the UV-persulfate
oxidation method outlined in Section 505b of Standard Methods (1985). Every
sample was analyzed in duplicate after it had been purged for 5 min. Before
each analysis, 5 potassium hydrogen phthalate standards were prepared to
23
-------
develop a calibration curve for the instrument (range = 1-5 mg/L). A 1 mg/L
EDTA standard was analyzed during each analysis to determine the accuracy of
the calibration.
Metals
Samples for total iron, total manganese and total calcium were collected
in 125 mL polyethylene bottles. Immediately after collection, the samples
were acidified with 180 uL of HNOj (final concentration - 0.15% (v/v) HN03).
Samples for soluble iron and soluble manganese were collected in 250 mL
polyethylene bottles. Approximately 60 mL was drawn from these containers
using 60 mL VWR plastic syringes, and filtered through 0.22 urn filters into
125 mL polyethylene bottles. The same filter was used for a duplicate set of
samples. The 125 mL bottles containing 60 mL of filtrate were acidified to a
final concentration of 0.15% (v/v) HNO-j.
An influent sample and a sample of blank water were spiked in the field
for both total and soluble metals to check accuracy. Fisher Scientific 1000
ppm stock solutions were used to spike samples for iron and manganese. A
100,000 mg/L stock solution was prepared from calcium chloride, to spike
samples for calcium analyses.
The metals were analyzed on a Perkin Elmer (Norwalk, CT) Model 2380 flame
atomic absorption spectrophotometer (AAS) using an air acetylene flame and a
10 cm burner head. The samples were analyzed according to the procedure
outlined in Section 303A of Standard Methods (1985). A calibration curve was
developed, using 4 to 5 standards spanning the linear range of each metal.
The standards were prepared daily from Fisher Scientific 1000 ppm stock
solutions. Initial experimentation indicated that neither the method of
standard additions nor the addition of calcium solution to the iron and
manganese samples (Standard Methods, 1985) was required to overcome matrix
problems or reduce ionization interferences. A lanthanum oxide solution
(Standard Methods, 1985) was added to all samples being analyzed for calcium.
Calibration standards were analyzed every tenth sample to check the drift of
the instrument. The field spikes were analyzed every ninth sample to
determine percent recovery and were compared with control chart
specifications. If the value exceeded the control limits, all previous
samples were reanalyzed. Precision was monitored every fifth sample. At the
end of each set of analyses, the calibration standards were reanalyzed.
GAC and resin samples collected for metals analyses during the coring
experiment were dried to a constant weight at 60°C, Two subsamples (2g) were
placed in tared 125 mL polyethylene bottles and weighed on a Mettler AC-100
analytical balance. The samples were then digested in 50 mL of 10% HNO^ in a
water bath at 60°C for 24 hrs (Lessard, 1987). The supernatant was decanted
and filtered through double deionized washed GF/C filters and diluted to 100
mL in a volumetric flask. The liquid was analyzed using the AAS according; to
the methods outlined above.
24
-------
Microbial Numbers
Water samples for subsequent plating were collected in 60 mL polyethylene
bottles. The bottles were cleaned by soaking in a low phosphate soap for 20
min, rinsing three times with deionized water and autoclaving for 15 min at
121°C and 15 psig. Sterilized double deionized water was transported to the
field in a sterilized 250 mL polyethylene bottle. A sample of this sterilized
water was analyzed as a control.
Microbial numbers were determined using the heterotrophic plate count
technique outlined in Section 907 of Standard Methods (1985). Aseptic
techniques were followed throughout the entire procedure (Methods 907 and
907B). A 0.1 mL aliquot was pipetted onto a low nutrient R2A agar (Difco?
Detroit, MI) and incubated at 10°C for 11 days. All samples were plated in
12 -2
triplicate. Initially, dilutions of 10 and 10~ were used, but 10 was
later found to be unnecessary. The incubation temperature and time were
determined in preliminary experiments to yield the maximum number of
organisms. The number of colonies (between 20 and 200) were recorded and the
results were converted to colony forming units (CPU) per 100 mL. Sterile
controls were run with every set of samples to check aseptic technique.
GAC and resin samples for microbial numbers from the coring experiment
were stored for 20 hr at 4°C prior to analysis. Aliquots of the solid matter
(5g) were placed in tared, sterilized beakers and weighed on a Mettler AC-100
balance. A 45 mL portion of sterilized 0.1% sodium pyrophosphate was added to
each of the beakers. The samples were then sonicated for 30 min. Serial
dilutions of the supernatant (10~3 to 10 ) were plated in triplate on R2A
agar and analyzed as outlined above.
Uranium, Radium and Lead
Aqueous samples for uranium and radium analyses were collected in 3.79 L
polyethylene containers and immediately acidified to a pH <2 with concentrated
HNO-. The samples were analyzed for total uranium and radium-226 by the State
of New Hampshire's Environmental Services Laboratory (Concord, NH). Total
uranium alpha activity was measured by the coprecipitation technique (Method
908.0) outlined by EPA (1980) using a Ludlum (Sweetwater, TX) 1000 sealer and
a Ludlum 43-10 detector. Radium-226 activity was measured by the radon
emanation technique (EPA, 1980} Method 903.1), using a Lucas scintillation
cell and a Ludlum 1000 sealer equipped with a M-182 detector. A trace spike
of barium-133 was added to all radium samples for determination of percent
recovery.
Solid GAC and resin samples from the coring equipment were dried to a
constant weight at 60°C. Aliquots of 2-3. g were placed in tared, plastic
containers and weighed on a Mettler AC-100 analytical balance. The samples
were counted on a Canberra high resolution gamma spectrometer with a coaxial
Ge diode well detector (Model GCW 1022-7500). The readings were processed
with a multi-channel-analyzer (Model 35). Standards to calibrate the
instrument were made with lead-210 and virgin GAC and resin samples. Aliquots
of Virgin GAC and resin were counted to determine background readings for the
instrument. Data was reduced to activities of uranium-238, uranium-235,
25
-------
radium-226 and lead-210 per kg by the method presented by Larsen and Cutshall
(1981).
Gamma/Beta Radiation Emissions
Gamma/beta emissions were taken by placing the detector of a Ludlum
(Sweetwater, TX) Model #9 Micro R survey meter on the surface of the treatment
units or holding it exposed to the air. The meter was set to record mR dose
rate'per hr. Measurements were taken at various locations in and around the
pumphouse (Figure 4). Measurements were taken on the surface of the GAC
contactors at four different locations (Table 5). The meter was calibrated
every 6 months with a cesium-137 source by the University's radiation safety
officer.
Percent Moisture Content
GAC and resin samples collected during the coring experiment were dried
to constant weight at 60°C to determine percent moisture. Samples were placed
in tared aluminum dishes and reweighed after cooling in a dessicator until a
constant weight was attained. The percent moisture was calculated by dividing
the final mass by the initial mass. The percent moisture was used to adjust
wet weight to dry weight.
DATA ANALYSIS
All data were analyzed using Student's t tests (Miller and Miller, 1984)
at a values of 0.10, 0.05 and 0.01. The most rigorous a value of the three at
which the stated trends occurred is reported for each data set tested.
26
-------
GAG W/PRETREATMENT
GAG
30 a
D
NOT TO SCALE
Figure 4. Schematic of the pumphouse showing the locations of the gamma
emissions measurements (A,B,C, and D are locations where
measurements were taken).
27
•..- -.^..-i..—.-, -••JsiijSr -:'- ••^••••--•••^•••A'-'i'/-:^-'^ ^:A~-^:^ .
-------
TABLE 5. LOCATIONS OF SURVEY METER MEASUREMENTS ON THE SURFACE OF GAC UNITS
Sample
Depth Below GAC Surface (cm)
GAC Without Pretreatment GAC With Pretreatment
Top
Middle
Bottom
1.5
43
93
0
41
93
28
-------
SECTION 4
RESULTS AND DISCUSSION
All of the treatment systems vere put on-line on January 7, 1989 treating
an influent resulting from a combination of ground vater from 2 wells. The
raw water to the treatment systems was found to vary frequently with respect
to radon and other water quality parameters such as iron. This variation
resulted from the differences between the raw water quality of the 2 wells and
because of iron precipitation occurring during storage in the polyethylene
tank. Therefore, in late April, the system was replumbed so that only 1 well
was used and the water flowed directly from the well to 2 hydropneumatic tanks
instead of the atmospheric storage basin. These modifications minimized
differences in the raw water quality going to the individual treatment
systems. A summary of all of the raw water characteristics is found in
Appendix C. All of the systems were operated until December 1989 without
major interruption in service. During December, there were several very cold
days and on two occasions the plumbing froze and flow was stopped. In
addition, during this period flow to the GAC system with pretreatment was
stopped for 5 days while a treatment system for another project was installed.
During the year of operation, there were several problems with the bubble
plate's mechanical systems and plumbing. As a result, it was inoperable on a
few occasions. In addition, the diffused bubble unit had a mechanical failure
in late November and was not repaired because the study was almost complete.
.Sampling was conducted weekly from January through June and biweekly
thereafter. All special sampling events on the GAC systems (backwashing,
coring, loading rate studies) were conducted during the latter half of the
year. Air monitoring with the alpha track detectors was conducted from
September through January.
The data obtained for each unit were compared with respect to the
following parameters: water flowrate, temperature, pH, alkalinity, calcium,
turbidity, iron, manganese, microbial numbers, NPDOC, radon, uranium, and
radium. Analysis of the non-radioactive constituents allowed us to determine
how the units were operating compared to similar systems treating other ground
water contaminants.
WATER FLOWRATE
The Derry FOE radon treatment system'was constructed so that each unit
produced enough water to meet a demand of 7.6 L/min at six 18 min intervals
and one 30 min interval over the day. It is important to note that the
flovrate into the two aeration units was in the range of 19 to 26 L/min, so
that these units operated intermittently to meet the 7.6 L/min demand. Each
unit had its own flow regulator. As a result, the actual flowrate through
29
-------
each unit differed slightly at each sampling period (Figure 5). However, over
the course of the study, there was no significant difference (a = 0.10) in the
water flowrates treated by each unit. The overall flowrates for the units
during the study are shown in Table 6 and were similar to the design value of
7.6 L/min typical of the demand in a residence (Anderson and Watson, 1967).
TEMPERATURE
The temperature of the raw water followed a seasonal pattern increasing
in the spring and summer months and decreasing in the late fall and winter
(Figure 6). This was similar to the results obtained in the small community
radon removal project (Kinner et al., 1990). The temperatures of the raw
water remained in a relatively small range (8°-17eC), typical of northern New
England ground water (Johnson, 1975).
PH
The influent pH to the units averaged 6.24 + 0.19 during the first 4
months of the study. In late April, it decreased to 5.25 when the system was
replumbed with one well and continued at that level for the remainder of the
project averaging 5.84 + 0.33. The pH of the water did not change
significantly as it passed through the GAC units (Figure 7). During the study
of GAC systems treating small community ground waters contaminated with radon
(Kinner, et al., 1990), a dramatic increase in pH was observed in the first
several days of operation. The increase in pH was also observed for the FOE
GAC units and was probably caused by dissolution of CaCO,(s) from the coconut
shell-based carbon. However, because of the small amount of GAC used in the
FOE units, alkaline material leached out (solubilized) rapidly and after the
first few days of operation, effluent pH was no longer affected by the GAC,,
The pfls of the effluents from the aeration units (Diffused Bubble = 7.27
+ 0.38; Bubble Plate = 7.24 + 0.36) were not significantly different from each
other (a * 0.10). However, The effluents from both units were significantly
higher than the influent (a » 0.01) by approximately 1 pH unit. This same
level of increase was observed consistently even when the influent pH
decreased in late April after the system was replumbed. A similar increase
was observed in pH through the packed tower and diffused bubble systems during
the small community radon treatment project (Kinner et al., 1990). The
aeration process in the FOE bubble plate and diffused bubble units probably
increased the pH by. removing CO- from the water. Even though the pH was
raised significantly, the resulting effluent pH was still in the acceptable
range (6.5-8.5) for drinking water.
ALKALINITY
During the period from January to May, there were large variations in the
alkalinity of the raw water. This occurred because each of the 2 -wells used
during this period had a different alkalinity. However, the influent
alkalinity was still highly variable when only one well was used, especially
during the period from August through October (Figure 8).
30
-------
c
1
LJ
20
15-
10-
5--
-A GAC
-• GAC W/ION EXCHANGE
-T DIFFUSED BUBBLE
-+ BUBBLE PLATE
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
TIME (months)
Figure 5. Water flowrate for the POE treatment systems,
31
' - i *"-::fe»v;r<-3^?*I£Xl;A^'?'-^J:'^
-------
TABLE 6. MEAN WATER FLOWRATES FOR THE POE TREATMENT UNITS
Treatment Unit
Water Flowrate (L/min)
Mean Standard Deviation
GAC without Pretreatment
GAG with Pretreatment
Bubble Plate
Diffused Bubble
8.13
8.62
8.08
9.01
1.59
1.53
2.33
1.77
32
-------
o^
UJ
EC
33
h-
UJ
0.
UJ
h~
£.*J -
20-
15-
10-
5-
0-
/%v-^— V\
0.» »-•-«
«rtlx^- *~~* •
••
1 1 1 1 1 1 1 1 1 1 1
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 6. Temperature of the influent water.
33
-------
10--
9--
8-.
7--
6-.
5--
INFLUENT
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
10
9--
8--
7"
6--
5--
—A GAC
• GAC WITH ION EXCHANGE
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
10--
9--
8--
7--
6-r
5..
* - * BUBBLE PLATE
• - • DIFFUSED BUBBLE
H 1-
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 7. pB of the influent water (a), and effluents from the
GAC systems (b) and aeration systems (c).
34
--V^^-^H:y^K^ ••-"„- -". ' •.'.--"'-'-'" •': ••
-------
100.
n
8
3
m
a
|
£
• INFLUENT
A SED FILTER
A GAC
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
100
• • INFLUENT
D D ION EXCHANGE
•—• GAC
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 8. Alkalinity for the GAC without pretreatraent (a) and the GAC
with pretreatment (b).
35
-------
Prior to June 1989, there was no change in the alkalinity of the water as
it passed through the GAC systems (Figure 8). The one exception to this
occurred during the first day of operation when the alkalinity in the GAC
effluents was dramatically increased. This increase probably resulted from
the dissolution of the CaC03(s) from the GAC similar to the effect observed
for the pH. An increase in alkalinity also occurred initially in the small
community radon treatment study (Kinner et al., 1990).
During the period from mid-July until December, the water exiting both
GAC systems was lower in alkalinity than the influent though not significantly
(a - 0.10) when all of the data was pooled. The decrease in alkalinity may
have resulted from release of C02 by microorganisms growing within these units
or by sorption of calcium species to the GAC. A similar decrease in
alkalinity has been reported by Hubele and Sontheimer (1984) and Eberhardt
(1976) for biological activated carbon. This hypothesis is supported by the
microbial enumeration data which indicated bacteria were present in the GAC
systems and by the calcium data which decreased through the GAC unit without
pretreatment. However, calcium was not detected (detection limit = 14 mg/kg
dry weight GAC) on the GAC obtained during the coring experiments. It is
doubtful that microbial CO, evolution would have resulted in a decrease in pH
in these units because they were fairly well buffered (mean alkalinity = 28-34
mg/L).
There was no significant difference (a - 0.10) between the influent and
effluent alkalinity for the bubble plate unit (Figure 9a). A similar trend
was observed for the diffused bubble unit (Figure 9b). However, during the
last half of the study, there was a small decrease in the alkalinity of the
water as it was treated by the diffused bubble unit. During this same period,
there was substantial iron precipitation and microbial activity. Either of
these conditions may have increased the acidity of the water in the unit
causing the decrease in alkalinity observed. The water was probably still
buffered well enough, however, to prevent any significant change in pH.
CALCIUM
During the first 4 months of operation, the calcium concentration in the
raw water averaged 15.5 + 1.15 mg/L as CaC03 (Figure 10). Calcium
concentrations did not vary as greatly as the alkalinity data did over the
same period. During the last 6 months of the study, the raw water calcium and
alkalinity concentrations showed similar trends.
A small amount of calcium was removed by the GAC unit without
pretreatment (Figure lOa) from July through November and this corresponded to
the removal of alkalinity occurring in that unit. The adsorption of calcium
by the GAC is consistent with other studies that report calcium accumulation
in GAC columns (Alben et al., 1983). The ion exchange unit removed all of the
calcium from the raw water and thus the subsequent GAC unit did not receive
any measurable input of calcium (Figure lOb). Analysis of the carbon samples
from the coring experiment indicated that no detectable amount of calcium was
retained in either GAC unit (analytical detection limit = 14 mg/kg dry weight
of GAC).
36
..-#"•'••'
-------
m
a
100
80
60-.
40-
20-
••—• INFLUENT
*—* BUBBLE PLATE
H t-
•< f
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
100
•—• INFLUENT
» DIFFUSED BUBBLE
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 9. Alkalinity for the bubble plate unit (a) and diffused
bubble unit (b).
37
-------
o
INRUEN1
SEDIMENT FILTER
GAG
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
• INFLUEN1
Q Q ION EXCHANGE
• • GAC
,_ i-
JAN ^B MAR APR MAY JUN JUL AUG SEP OCT NCV DEC
50
s§
2
20-
10-
• ^« INFLUENT
T * DIFFUSED BUBBLE
* » BUBBLE PLATE
•^ FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 10. Calcium for the GAC without pretreatment (a), the GAC
with pretreatment (b), and aeration systems (c).
38
-------
There was no significant change in the calcium concentration of the water
when it was treated in the two aeration units (a » 0.10) (Figure lOc).
TURBIDITY
The raw water at the site varied between 0.13 and 4.1 NTU over the course
of the study (Figure 11). During the initial months of operation, the
variable, but relatively high level of turbidity was probably the result of
particulate iron in the influent water. During this period, raw water was
held in an atmospheric storage tank where iron precipitation was occurring.
When the system was replumbed by replacing the storage tank with 2
hydropneumatic pressure tanks in late April, there was a decrease in turbidity
in the influent even though particulate iron was still present in the
influent. During the later months of operation, turbidity in the raw water
increased dramatically on 2 sampling days. The cause for these increases is
unknown, but one event occurred on a day in October when microbial numbers
were also very high (59,000 CFU/100 mL). When only one well was used, the raw
water had a turbidity <1 NTU except for these 2 sampling days near the end of
the study.
The turbidity exiting the sediment filter was sometimes significantly
less (a » 0.10, not significant at a = 0.05 and 0.01) than the influent
concentration by an average of 32 + 94% (Figure 11). The ion exchange unit
removed most of the remaining turbidity, so that the GAC unit with
pretreatment received very little turbidity loading. The GAC without
pretreatment removed turbidity from the water throughout the study. In a
survey of water treatment plants that use GAC in the U.S., influent
turbidities of 1.1 to 7.0 NTU were decreased to 0.1 to 0.5 NTU through the
units (Graese et. al., 1987). Although turbidity was reduced through the GAC
and ion exchange units, headless was not observed. The ion exchange unit was
backwashed every 2 weeks, so this result was expected, however, the GAC
without pretreatment was not backwashed until late November. The data
suggests that headless development from retention of particulates in the GAC
unit without pretreatment was negligible in this study. This may not hold
true for GAC units that receive a higher turbidity loading.
The turbidity in the bubble plate and diffused bubble effluents was
significantly greater (a » 0.01) than the influent concentration (Figure 12).
The increase in turbidity in the aeration units was probably the result of
iron oxidation and precipitate formation (see subsequent discussion of iron)
during aeration (Figure 13).
IRON
The total iron in the raw water was typically less than 0.6 mg/L and in
most samples it was less than or equal to 0.3 mg/L except for 4 days over the
course of the study. An average of BOX of this iron was soluble. The
sediment filter removed an average of 24% of the total iron (Figure 14a) and
usually negligible amounts of the soluble iron (Figure 14b), though some
soluble iron removal was observed on-a few occasions. It is hypothesized that
the soluble iron was removed by the filter as a result of deposition of
precipitate forms and/or sorption of complexed forms. The GAC unit without
39
-------
o
CD
•—» INFLUENT
—A SED. FILTER
GAC
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
CD
5
• • INFLUENT
a a ION EXCHANGE
• • GAC
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 11. Turbidity for the GAC without pretreatment (a) and
the GAC with pretreatment (b).
-------
o
CD
10
9
8--
7-
6
• • INFLUENT
*—* BUBBLE PLATE
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
10
o
m
9--
S-.
7-
6--
5"'
4--.
3-
• • INFLUENT
*—* DIFFUSED BUBBLE
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 12. Turbidity for the bubble plate unit (a) and the
diffused bubble unit (b).
-------
|
IRBIDITY
Pi
10
9
8
7
6-
5-
4-
3
2-
1-
0-
o.c
a
• .
"'• ••
•
• J *
>0 0.50 1.00 1.50 90
TOTAL IRON (mg/L)
x-s
E
t
a
m
tr
?
9
8
7-
6-
5'
4-
3-
2-
1-
0-
b
»
•
•
• e
• • •
* • **** ••
I \*
-*-^ ; — i — 1 1 ;
0.50 1.00 1.50
TOTAL IRON (mg/L)
2.00
Figure 13. Turbidity as a function of total iron for the bubble
plate unit (a) and diffused bubble unit (b).
42
-------
•—* INFLUENT
A SEDIMENT FILTER
A A GAC
0.00
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
1.50
a
m
o
in
• • INFLUENT
A—A SEDIMENT FILTER
A GAC
0.00
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 14. Total (a) and soluble iron (b) for the GAC without
pretreatment. (detection limit • 0.06 mg/L).
43
-------
pretreatment (Figure 14) removed all detectable iron (detection limt = 0.06
mg/L) from the water. A similar trend was observed in the small community
radon removal studies for one of the GAG systems. (Kinner et al., 1990). The
ion exchange unit removed all detectable iron from the water (Figure 15),
hence no iron was applied to the downstream 6AC.
The coring experiment data indicated that most of the iron removed by the
GAG unit without pretreatment was retained in the top of bed (Table 7), with a
marked decrease in concentration with depth. As expected, little iron was
accumulated in the GAG with pretreatment. Visual inspection of the GAG bed in
the unit without pretreatment during coring showed that a crust of orange-red
precipitate was prevalent in the top 15.2 cm of the carbon. Below this depth,
the GAG appeared unchanged. The carbon in the GAG with pretreatment had no
visible fouling through the bed.
The effluent total iron concentrations for the bubble plate (Figure 16a)
(0.43 + 0.24 mg/L) and diffused bubble (Figure 17a) (0.53 + 0.30 mg/L) units
were somewhat higher than the influent (0.40 + 0.27 mg/L). However, these
differences were not usually statistically significant (a - 0.10 or 0.05).
Conversely, the concentrations of soluble iron did significantly (a = 0.01)
decrease through these units (Figures 16b and 17b). The data indicates that
the aeration processes were causing iron to be oxidized and precipitated. The
increase in total iron in the effluents probably resulted from the release of
particulate iron accumulated in the units. Iron accumulations were observed
in the water stored in both units between periods of operation. This was
particularly true for the diffused bubble unit. Each unit is designed to
aerate the water for a period of time (Bubble Plate = 1 min; Diffused Bubble =
10 min) after the influent flow stops. During this period, iron precipitates
are probably formed in the units and subsequently released along with the
effluent when a new cycle is started.
MANGANESE
The influent total manganese concentration in the raw water varied
between 0.08 and 0.68 mg/L (X = 0.36 + 0.12 mg/L) during the study, 95% of
which was soluble (X - 0.35 + 0.11 mg?L) (Figure 18). Manganese was not
significantly (a = 0.10) changed in any of the units (Figures 18, 19, 20 and
21) during treatment except in the ion exchanger which lowered it below
detection (0.02 mg/L). This was expected because the resin had a strong
affinity for cations. In all other cases, manganese removal was not expected
since manganese remains soluble below pH = 11 and oxidation/precipitation via
aeration requires periods of >1 hr.
The coring data (Table 8) indicated that little manganese was retained in
the GAG units. The concentration at the bottom of the GAG without
pretreatment was the highest of the three depths sampled. It appeared that
dissolved manganese entering the unit, was undergoing oxidation as it passed
through the upper part of the GAG and being deposited as an insoluble complex
44
-------
1.500
« INFLUENT
0 ION EXCHANGE
m GAC
0.000
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
1.500
• « INFLUENT
Q Q ION EXCHANGE
• • GAC
0.000
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 15. Total (a) and soluble iron (b) for the GAC with
pretreatment, (detection limit » 0.06 mg/L).
- . r.-.-.-••;. - ,,-:• - > -. -7- • -;1- .• \~< • ••?;-,-.,.-/'• t-^f^.'f'^^ •',:• "—••.••, - -
W^-;^ .•*-'.:>.-\S--&'^*-^:^^^ ,--''^' ' •'••
-------
TABLE 7. TOTAL IRON CORING DATA IN THE GAC UNITS
Location in Bed"1"
Top
Middle
Bottom
Total Iron
GAC without Pre treatment
13,520
1,110
640
(rag/kg )
GAC with Pre treatment
425
ND*
ND
All data reported on a dry weight basis obtained from the percent moisture
analyses.
+In the GAC without pretreatment, coring depths were top, middle = 43 cm,
bottom « 93 cm. In the GAC with pretreatment, coring depths were top, middle
» 41 cm, bottom = 93 cm.
*Not detectable. Detection limit = 3 mg/kg.
46
-------
1.50
0.00
•—« INFLUENT
*—* BUBBLE PLATE
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
1.50
G> 1-20-k
0.00
• • INFLUENT
» * BUBBLE PLATE
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 16. Total (a) and soluble iron (b) for the bubble plate unit,
(detection limit = 0.06 mg/L).
47
-------
1.50
1.20 +o
i
J, 0.90 •
O
g£
i
0.60 ••
0.30 ••
0.00
• • INFLUENT
* * DIFFUSED BUBBLE
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
1.50
• « INFLUENT
DIFFUSED BUBBLE .
0.00
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 17. Total (a) and soluble iron (b) for the diffused
bubble unit, (detection limit = 0.06 mg/L).
48
-------
1.00
0.80"
0.40-
0.20-
0.00
• » INFLUENT
A A SEDIMENT FILTER
A—A GAC
Vy*
H 1 1-
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
1.00
0.80-
0.60 •
0.00
• • INFLUENT
A A SEDIMENT FILTER
A A GAC
H 1 H
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 18. Total (a) and soluble manganese (b) for the GAC system without
pretreatment. (detection limit = 0.02 mg/L).
49
' ' " "-.•' - •• • -....• • J.--^,' ' -^-^''-.^V-^;::*"'^^?"*"^^"^*?'^^"^^ •'•-ife'-'.1. .
-------
1.00
O.SO
0.00
• INFLUENT
Q ION EXCHANGE
—« GAG
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
UJ
• • INFLUENT
0 0 ION EXCHANGE
• • GAC
0.20-
0.00
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 19. Total (a) and soluble manganese (b) for the GAC with
pretreatment. (detection limit = 0.02 mg/L).
50
-------
1.00
< 0.80
O»
£
a 0.60
i 0.40
0.20-
0.00
« • INFLUENT
•—» BUBBLE PLATE
H 1
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
ui
m
V)
1.00
0.80
0.60
0.00
• « INFLUENT
* * BUBBLE PLATE
H 1
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 20. Total (a) and soluble manganese (b) for the bubble
plate unit (detection limit » 0.02 mg/L).
51
-------
1.50
8
0.00
•—• INFLUENT
* « BUBBLE PLATE
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
CD
1.50
^ 1.20-•
O*
r 0.90-I-
0.60 4-n
0.30-
0.00
• « INFLUENT
v v DIFFUSED BUBBLE
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 21. Total (a) and soluble manganese (b) for the diffused
bubble unit, (detection limit = 0.02 mg/L).
52
-------
TABLE 8. TOTAL MANGANESE CORING DATA IN THE GAC UNITS
Total Manganese (mg/kg )
Location in Bed"1" GAC without Pretreatment GAC with Pretreatment
Top
Middle
Bottom
1.45
0.30
6.35
ND**
ND
ND
All data reported on a dry weight basis obtained from the percent moisture
analyses.
+ In the GAC without pretreatment, coring depths were top, middle = 43 cm,
bottom * 93 cm. In the GAC with pretreatment, coring depths were top,
middle - 41 cm, bottom = 93 cm.
**.
Not detectable. Detection limit = 1 mg/kg.
53
-------
in the bottom of the bed. A similar pattern was observed in the GAC units
used in the small community radon removal study (Kinner et al., 1990).
MICROBIAL NUMBERS
During the period April through September, the influent raw water (Figure
22) at the site usually had no detectable bacterial contamination (<20,000
CFU/100 mL) as assessed using R2A heterotrophic plate counting (N.B.,
microbial enumeration was not started until April). However, in late
September and thereafter, the raw water contained bacteria in the range 20,000
to 59,000 CFU/lOOmL. This corresponded to the time when the raw water also
had variable concentrations of alkalinity, calcium, iron and turbidity. It
appears that during the late summer and fall, the ground water supplying the
site was more influenced by changes in the water table and surface recharge.
Such variability in water quality was observed in the previous long-term
monitoring during the small community radon removal study (Kinner et al.,
1990).
Neither the sediment filter (Figure 22a) nor the ion exchange unit
(Figure 22b) significantly (a - 0.10) increased bacterial numbers. However,
on certain days, the effluent from the ion exchange unit contained higher
numbers than the raw water. Flemming (1987) noted that microbial growth in
ion exchange units is common and can lead to their release into the finished'
water. In addition, the effluent from both GAC units (Figure 22) contained
large numbers of bacteria (GAC without pretreatment » 32,400 + 33,600 CFU/100
mL* GAC with pretreatment = 34,200 + 22,500 CFU/100 mL), often greater than
the concentration exiting the sediment filter and ion exchange unit, but
overall not significantly different (a = 0.10) from the influent. These
concentrations tended to fluctuate widely over time (range = 20,000 to 200,000
CFU/100 mL). A similar phenomenon was observed during the small community
radon removal study (Kinner et al., 1990) and by other researchers (Wilcox et
al., 1983). It has been well documented that GAC filters are capable of
supporting microbial populations (Wilcox et al., 1983; Camper et al., 1985,
1986, 1987; Graese et al., 1987). As with the ion exchange unit, the GAC
provides a good surface for attachment and concentrates the nutrients as well
as the organic matter necessary for microbial maintenance and growth. The
numbers observed in this study were within the ranges reported in the
literature [0-30,000 (Wilcox et al., 1983) and 0-60,000 colonies/100 mL
(Bourbigot, 1981)J.
During the coring study in January 1990, samples of carbon collected from
the two GAC units in January 1990 were analyzed for microbial numbers. The
concentrations ranged from <1.80 x 10 to 2.46 x 10 CFU/g. In the GAC
without pretreatment, the highest concentration was at the top of the bed
(2.46 x 105 CFU/g) and this decreased to <1.80 x 105 CFU/g in the samples
taken at 43 and 93 cm depths within the bed. In the GAC with pretreatment,
there was no clear profile within the bed {Top < 1.80 x 10 CFU/g, 41 cm ==
2.07 x 105 CFU/g, 97 cm < 1.80 x 105 CFU/g]. These data confirm that the GAC
54
-------
OTO
'S§ 140 +
• • INFLUENT
A A SEDIMENT FILTER
A GAC
20
MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
• • INFLUENT
D a ION EXCHANGE
• • GAC
20
MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 22. Microbial numbers for the GAC without pretreatment (a)
and the GAC with pretreatment (b).
55
-------
was fouled with microorganisms which could potentially be released into the
water passing through the units.
The bacterial numbers exiting the aeration units (Figure 23) were
sometimes greater than the influent concentrations. However, the average
effluent bacterial concentrations were not significantly different than those
in the influent (a = 0.10) nor those in the effluent from the GAG units (a =
0.10).
The data from all of the treatment systems indicated that bacterial
contamination of the treated water will be a problem. The problem may be
exacerbated by periodic increases in the microbial concentration in the raw
water and by seasonal increases in temperature. Though regeneration and
backwashing may lower bacterial numbers in the filter beds, they will not
eliminate the contamination (Flemming, 1987). Consequently, all of the
treatment units may need supplemental disinfection to insure that bacterial
numbers in the finished water are reduced to acceptable levels.
NPDOC
In the small community radon removal study, the GAC's ability to remove
radon from the water was less than predicted. One explanation proposed for
this decreased efficiency was the accumulation of natural organic matter in
the unit which could have changed the GAC's affinity to adsorb radon (Kinner
et al., 1990). Hence, in the FOE study, the GAC units were monitored for
NPDOC.
The NPDOC concentration in the raw water ranged from 0.94 to 1.98 mg/L
(X » 1.30 ± 0.28 mg/L). This was similar to the average TOC (1.03 ± 0.76
mg/L) reported for Kansas groundwaters (Miller et al., 1990). NPDOC was
effectively removed by both GAC units (Figure 24), but became increasingly
more prevalent in the effluent over time. Breakthrough occurred in the period
of July and August (Figure 25) and thereafter, the influent concentration was
approximately equal to or less than the effluent. The data indicated that:
when the units were started after a quiescent period, there was a potential to
have greater concentrations of NPDOC in the effluent. It is possible that
NPDOC was desorbed from the GAC into the water flowing through the unit during
these periods. Similar accumulation and breakthrough of NPDOC has been
observed in pilot- and full-scale GAC water treatment systems studied by other
researchers (Roberts and Summers, 1982; Veber et al., 1983).
RADIONUCLIDE REMOVAL
GAC
Radon —
The radon activity in the raw water at the Derry site ranged from 22,837
to 54,765 pCi/L, averaging 35,620 + 6,727 pCi/L (Figure 26). Variation in the
influent during the first 6 months of the study was partially due to operating
with water from 2 wells and to natural variation in the source. Variation in
56
-------
180
coo
0:0
UJO
!<- 140
^o
fflO
INFLUENT
BUBBLE PLATE
DIFFUSED
BUBBLE
100
MAY JUN JUL AUG SEP OCT NOV DEC
Figure 23. Microbial numbers for the aeration systems.
57
-------
O
o
O
Q_
SEDIMENT FILTER
GAC
APR MAY JUN JUL AUG SEP OCT NOV DEC
o>
o
o
o
Q.
a ION EXCHANGE
GAC
APR MAY JUN JUL AUG SEP OCT NOV DEC
TIME (months)
Figure 24. NPDOC for the GAC without pretreatment (a) and the GAC with
pretreatment (b).
58
-------
c
6
o
o
ID
O
iS
Ct:
CD
O
O
Q
Q_
» GAC WITH ION EXCHANGE
APR MAY JUN JUL AUG SEP OCT NOV DEC JAN
Figure 25. NPDOC breakthrough curve for the GAC
units.
59
-------
60
50
b
Q.
O
<-
X
O
Q
o;
40 J
304
20--
10--
n
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 26. Influent radon activity for the POE treatment systems.
60
-------
the latter half of the year was due to a single veil and was similar to that
observed in the small community radon removal study (Kinner et al., 1990).
The sediment filter and ion exchange unit did not remove any detectable amount
of radon from the water (a = 0.05). The effluent radon activity pattern from
the two GAC systems (Figure 27a) was similar to that observed by Lowry and
Lowry (1987) during the first 4 months of operation. The activity in the
effluent was negligible for several days and then increased rapidly to an
average of 881 + 662 pCi/L for the GAC without pretreatment and 660 + 204
pCi/L for the GAC with pretreatment where it remained until April. During the
remaining 8 months of operation, the effluent activity from both GAC units
generally increased, however, there was a large degree of variability in the
data (Figure 27b). In some cases, there was a difference in the effluent
quality of the two systems. These discrepancies were usually accounted for by
the differences in the radon mass loading to the units (Figure 28) (GAC
without pretreatment = 299,219 + 85,942 pCi/min; GAC with treatment = 310,204
+ 89,522 pCi/min), mostly as a function of water flowrate.
Using a radon adsorption model presented in Lowry and Lowry (1987) with
the influent and effluent radon activities and water flowrates from the FOE
GAC units, the steady state radon removal rate constant (K ) was determined
ss
for comparative purposes. The mean K values for the units were 2.98 + 1.52
-1 -1 ~
hr (GAC without pretreatment) and 3.29 + 1^29 hr (GAC with pretreatment).
These values compare favorably to the K of 3.02 hr~ at 6-10°C obtained in
SS
other tests with this type of GAC (Lowry and Lowry, 1987). The range of K
_1 ss
values over the year was 1.17 to 7.39 hr . Differences in the quality of the
water treated and the column hydraulics may have accounted for the differences
in K observed.
. SS
Most reports on GAC units used in FOE applications have observed radon
removals in the range 95 to >99% (Lowry and Lowry, 1987; Lowry et al., 1987;
Lowry, 1988), whereas in our study the removal gradually deteriorated over
time from'99.7 to 79% for the GAC without pretreatment and from 99.7 to 85%
for the GAC with pretreatment. There are several factors which may explain
such a deterioration in effluent quality. The most likely is a change in
influent loading to the GAC. The mass loading of radon at the Derry site was
within the range 130,000 to 653,888 pCi/min throughout the study, becoming
somewhat less variable during the last 3 months of operation (Figure 28).
However, it was during the latter period that the effluent quality
deteriorated the most. The mass of radon applied was linearly correlated to
the mass removed over the entire range of loadings (Figure 29) and a similar
relationship was found for the cumulative mass relationships (Figure 30) for
both units. When the data sets obtained for each unit were examined with
respect to time, there were significant decreases (a = 0.01) in the slopes of
the removal lines between the periods January through June as compared with
July through December, especially for the GAC unit without pretreatment
(Figure 31). However, there was no significant difference (a = 0.10) between
the two units with respect to mass radon removal at any time. Further, a plot
of the total radon activity adsorbed over time for each GAC unit showed their
similarity, but appeared to be somewhat non-linear (Figure 32).
61
-------
o
CL
Z
O
I
JAN
GAC
GAG WITH ION EXCHANGE
FEB
MAR
APR
O
Q.
1*3
O
O
5
GAC
GAC WITH ION EXCHANGE
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 27. Effluent radon activities for the GAC systems for the first
4 months (a) and for the entire study (b).
(note scale difference).
62
-------
c
E
700
600
CJ
Gi-
ro
O
X
O
2
Q
g
CO
<
GAC
» GAC WITH ION EXCHANGE
JAN .FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 28. Radon mass loading rate to the GAC systems.
63
-------
c
'£
o
CL
O
X
Q
I
LU
ce
500
400
300
200
100
100
500
200 300 400
MASS APPUED ( X 103 pCi/min)
500
200 300 400
MASS APPLIED ( X 10^ pCi/min)
500
Figure 29. Radon mass removed as a function of mass applied for the GAG
without pretreatment (a) and the GAG with pretreatment (b).
-------
o
Q.
05
o
Q
I
L±J
ct:
CO
GO
13
O
5 10 15
CUMULATIVE MASS APPLIED ( x 10 9 pCi )
Figure 30. Cumulative radon mass removed as a function of cumulative
mass applied for the GAG without pretreatment (a) and
the GAG with pretreatment (b).
65
-------
c
E
o
Q.
O
X
I
LU
ce
to
500
400-•
200 300 400
•MASS APPLIED ( X 10^ pCi/min)
500
E
fO
o
o
1
01
500
400-•
300-
JAN-JUN
JUL -NOV
200 300 400
MASS APPLIED ( X 10^ pCi/min)
500
Figure 31. Radon mass removed as a function of mass applied for the
first 6 months and the latter 6 months of the study for
the GAC without pretreatment (a) and the GAC with
pretreatment (b).
66
^<^-^-'-^-r^-."^:™ V.:fe^-. ^•-'•i^v-'.v.:.- •S^w",-£:^H:^:^:^::S?*v0^ ^-*;" -"•' ''*''•'
-------
o
Q.
oo
o
X
Q
.UJ
CD
O
O
£
>
o
<
150
120--
90--
30-
0
GAC
GAC WITH ION EXCHANGE
J
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 32. Total radon activity adsorbed over time for the GAC units.
67
-------
The data suggest that the reduction in removal over time may be a
function of the adsorptive capabilities of the GAG bed. Traditionally, it has
been thought that GAG beds will not approach exhaustion with respect to radon
for very long periods of time (Lowry and Brandow, 1985). This is expected
because the mass of radon removed is very small even though the reduction in
activity is high. Therefore, it was hypothesized that some other parameter in
the raw water was sorbing to the GAG and reducing its capacity for radon.
Initially, NPDOC was suspected because its breakthrough occurred in
July/August when there was an increase in the radon in the effluent. Though
there was an increase in the radon and NPDOC remaining in the effluent for
both units (Figure 33), the correlation between them was not strong. Another
possible candidate was iron, however, the GAG with pretreatment, whose radon
removal decreased in a pattern similar to the GAG without pretreatment, did
not receive any iron loading.
Unfortunately, in this study (unlike the small community radon project),
we did not have sampling ports within the GAG bed. As a result, the radon
removal profile could not be determined directly. However, gamma/beta
emissions measurements were taken at the surface of the GAG tanks and these
data show an interesting pattern (Figure 34). Initially in both units, the
greatest gamma emissions occurred at the top of the bed. This was expected
because the gamma emissions result from Bi and Po, which are both
short-lived progeny of radon and high energy gamma emitters. The emissions at
mid-depths (41-43 cm) were initially lower than those at the top of the bed
because less .radon was available to be removed at that depth. The bottom of
the bed (93 cm) emitted much less gamma energy because there was little radon
left in the water at that depth to be removed. Unlike small community GAG
systems, POE units usually contain a large amount of excess carbon with
respect to the amount required because of the relatively low cost of
3
purchasing 0.028 to 0.084 m of GAG. This excess GAG serves to dampen the
effect of potential overloads to the system.
From April through the rest of the study, a greater amount of gamma
emissions came from the mid-depth (43 cm) of the GAG without pretreatment than
from the mid-depth (41 cm) of the GAG with pretreatment. This dose was also
significantly (a = 0.01) higher (2.9 + 0.77 mR/hr) than the emissions from the
top of the bed (1.9 + 0.57 mR/hr). Concurrently, there was slight increase in
the gamma emissions from the bottom (93 cm) of this GAG bed over time.
Initially, in-the GAG.unit with pretreatment, the greatest gamma
emissions were from the top of the bed. However, there was no significant (a
- 0.10) difference between the gamma emissions from the top and middle (41 cm)
of the bed from July until December. In this unit, there was a smaller
increase in the gamma emissions from the bottom (93 cm) of the unit.
The gamma emissions data imply that the zone of radon removal in the
units moved down the bed over time. This was especially true in the GAG unit
without pretreatment. The effluent radon activity measured was not
significantly different between the two GAG units. This lack of difference
may have been caused by the excess amount of GAG present in both POE units.
68
-------
V
o
o
1
0.20
0.15--
0.10-
0.05-
0.00
• BEFORE BREAKTHROUGH
OAFTER BREAKTHROUGH
O
0.0
0.5 1.0
NPDOC(Ce/Co)
O
1.5
0.20
W
O
O
O
0.15-
0.10-
0.05-
0.00
o
o
•
o
0.0
0.5 1.0
NPDOC(Ce/Co)
1.5
Figure 33. Radon breakthrough (Ce/Co) as a function of NPDOC breakthrough
(Ce/Co) for the GAC without pretreatment (a) and the GAC with
pretreatment (b).
69
-------
o:
E
• • 1.5 cm
43
* 93 cm
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
TIME (months)
Figure 34. Gamma emissions measurements taken at the unit's surface
for the GAC without pretreatment (a) and the GAC with
pretreatment (b).
70
;;^-Vi T31^.^v^r^x;is^'v-f""^*-'^"-.'•':fr*'y""'^'.;-'—"^^•V'^-^'y^s" • ^;-;r.'^. •f-,. • ,. -
f'£~-^%z^Z^>^£'~^^ ?&i .->j::-:.X
-------
The units were cored in January 1990 to determine the activity of
lead-210 retained in the GAG. Rubin and Mercer (1981) reported that GAC
adsorption capacities for non-radioactive lead range from 6.2 x 10 to 1.9 x
106 yg/kg at pH 6.5 and from 2.1 x 106 to 1.8 x 10 ug/kg at pH 8.0 (based on
Langmiur adsorption isotherms). Comparison of these data to the lead-210
coring results (Table 9) indicated that the GAC's capacity for lead-210 was at
least 107 to 1011 times higher than the actual and predicted lead-210 loading.
This supports the hypothesis that lead-210 should be retained by the GAC unit.
This hypothesis was confirmed by the comparison of the predicted lead-210
adsorbed based on radon mass removal to' the actual lead-210 found from the
coring data (Table 9).
The lead-210 profiles in both beds suggested that the greatest radon
retention occurred at the top of each bed. However, while the radon loadings
to the two GAC beds were not significantly different (a « 0.10), the lead-210
accumulations on the top of the GAC bed without pretreatment were
significantly less (a = 0.05) than those in the GAC with pretreatment. At the
mid-depth, the two beds were not significantly different (a « 0.10) in
lead-210 adsorbed. While at the bottom, the lead-210 was significantly
greater (a => 0.10) in the GAC without pretreatment.
The lead-210 retention data and the gamma emissions data from the survey
meter indicated that radon removal was occurring deeper within the bed for the
GAC without pretreatment. The iron data from the coring experiment of the GAC
without pretreatment showed that there was significant accumulation (a = 0.01)
of iron precipitates in the top of the bed as compared to the GAC with
pretreatment. Visually, the greatest red-orange precipitate accumulations
were observed in the top 15 cm of the bed. It is possible that the iron
precipitation in the GAC impeded radon sorption by fouling the GAC surface or
causing short-circuiting (channeling) of water in the top of the bed. As a
result, the radon removal front appeared to be moving further down the bed in
the GAC without pretreatment, so that the volume of excess carbon available
for polishing was reduced.
Loading Rate Experiments — The radon removal over the course of a day was
monitored twice (February and August) when the flowrate was -7.6 L/min and the
total daily throughput was -1,022 L/day (i.e., a normal cycle). The effluent
activity was not significantly different (a = 0.10) over the course of the
day. Similar results were obtained during the small community radon removal
study. The data indicated that both GAC units were able to dampen short term
changes in radon activity (Figure 35).
A series of 3 loading experiments were conducted one week apart during 2
months (August and October). When the flowrate remained between 7.7 to 8.9
L/min, but the total throughput was increased to 1,900-2,100 L/day, there was
a significant (a = 0.01) decrease in the ratio of radon mass removal to mass
loading over time during the second run (Figure 36). The other two loading
conditions tested had flowrates in the range 20.4 to 22.6 L/min, and total
throughputs of 1,264 to 1,400 L/day, and 1,904 to 2,139 L/day, respectively.
In both of these cases, although the effluent radon activities were similar,
71
-------
TABLE 9. COMPARISON OF PREDICTED LEAD-210 REMOVAL IN THE GAC UNITS
TO THE CORING DATA
Location in Bed
Lead-210 Measured (ug/kg GAC)+
GAC without Pretreatment GAC with Pretreatraent
Top
Middle
Bottom
X in Bed
Predicted in Bed
0.00457 (370 x 10°)
0.00323 (262 x 103)
0.000946 (73 x 103)
0.0029 (235 x 103)
0.00457 (371 x 103)
0.00549 (445 x 10J)
0.0033 (265 x 103)
0.00080 (65 x 103)
0.0030 (246 x 103)
0.00494 (402 x 103)
See Appendix D for methods of calculation.
+ All data reported on a dry weight basis obtained from the percent moisture
analyses.
In the GAC without pretreatment, coring depths were top, middle = 43 cm,
bottom = 93 cm. In the GAC with pretreatment, coring depths were top,
middle - 41 cm, bottom = 93 cm.
**.
Values in parentheses are the lead-210 activity measured in pCi/kg dry
weight.
72
-------
50
40 +
to 30f •
o
20-
10-
—• INFLUENT
A GAC
—• GAC WITH ION EXCHANGE
5
TIME ( Hrs)
10
50
•3 40-
10 30+ •'•
o
20-
10-
Q
INFLUENT
GAC
GAC WITH ION EXCHANGE
TIME ( Hre)
10
Figure 35. Radon activities for the diurnal study for the GAC systems.
(February = a; August = b).
73
-------
Q
LU
CO
CO
I
LU
CO
2
1.000
0.900
0.800 -
0.700-
0.600
A - A GAC
— • GAC WITH ION EXCHANGE
TIME4(hrs)
8
1.000
Q
UJ
to
LJ
LU
to
CO
0.900 •-
0.800-
0.700-
0.600
A GAC
GAC WITH ION EXCHANGE
0
TIME (brs)
8
Figure 36.
Ratio of "don mass removed: mass applied for the high loading
rate study for the GAC systems. (August = a; October = b).
(t-Lowrate - 7.6 L/rain; throughput * 1,900 L/day).
74
-------
the mass removal of radon was significantly less (a = 0.01) in the GAC without
pretreatment (Figures 37-38). On the days following these events, the
effluent radon activity from the GAC units was sometimes higher than on the
previous sampling event (Figure 39) indicating that the increased loading may
have adversely affected radon removal for a period of time after loading
returned to normal. Figure 40 also showed that the mass removal during the 20
L/min flowrate was significantly (a = 0.01) lower for both GAC units than that
achieved at the normal 7.6 L/min flow. These data, in conjunction with the
gamma emissions and lead-210 data indicated the GAC may become saturated with
respect to radon even though it is in very low concentrations (10~ to 10"
mg/L). It appears that over time, the ability of the GAC to dampen mass
loading variations may be reduced. Therefore, in POE applications, effluent
radon activities greater than the proposed MCL of 200-2,000 pCi/L may result.
Furthermore, it appears that a GAC system's loss of resilience may be
exacerbated by accumulation of iron in the top of the bed.
Backwashing Experiment -- Each bed was backwashed with raw water for 12 to 14
min, as recommended by the manufacturer, until GAC appeared in the backwash
water. Backwashing did not change the radon removal observed in either GAC
unit (Figure 41) nor did it affect gamma profiles or any of the other water
quality parameters except for a small increase within the few hours of
operation. This initial decrease in water quality, after restarting a
backwashed filter, is typically observed in granular media systems
(Amirtharajah, 1988). Interestingly, the backwash water did not contain a
continuous release of iron precipitate as was observed in the small community
GAC backwash water. Only small bursts of iron precipitate were released.
Gamma Exposure Dose/Shielding Event —
The maximum survey meter readings (mR/hr), taken at the surface of the
tank, were divided by the influent radon activities (pCi/L) to determine the
correlation between dose and influent radon activity. The average values
(10,658 and 10,581 mR/hr/pCi/L GAC without and with pretreatment,
respectively) were in the range reported by Lowry and Brandow (1985) and Lowry
(1988). However, as noted in the small community radon study (Kinner et ail.,
1990) and by Rydell et al. (1989), these relationships are affected by the
geometry, the type of detector used and the proximity to the source.
Survey meter measurements were also taken 1.5 m from the GAC tanks
(Figure 42). Though the two GAC systems were in separate sections of the
Derry pumphouse, there may have been some interferences. It was also
difficult to get readings completely around the GAC tanks because of the
restricted space in the building. The data collected were within 80% to 100%
of the values obtained using the Carbdose 2.0 program developed by Keene and
Rydell (1989) when an average influent radon activity of 35,620 pCi/L, water
flowrate of 1,122 L/day, and 90% removal were used. The Carbdose 2.0 model
also predicts the distance at which the probable exposure dose is less than
0.058 mR/hr based on the National Council on Radiation Protection residential
guideline value (Keene and Rydell, 1989). The model predicted that for the
Oerry water supply the required distance was 189.5 cm.
75
-------
S
CO
o
UJ
CO
1.000
0.900
0.800
0.700-
0.600
^ A GAG
I • GAG WITH ION EXCHANGE
0
2 4
TIME (hrs)
Q
LJ
1.000
0.900
0.800
I
UJ
CO
0.700-
0.600
GAG
GAG WITH ION EXCHANGE
0
245
TIME (hrs)
8
8
Figure 37. Ratio of radon mass removed: mass applied for the high loading
rate study for the GAC systems. (August = a; October = b).
(flowrate ~ 20 L/min; throughput = 1,050 L/day).
76
-------
LJ
_l
Q_
CO
co
1.000
0.900 -•
0.800-
0.700-
0.600
A-
A A GAC
• • GAC WITH ION EXCHANGE
2 4 . g
TIME (hrs)
8
1.000
Q
UJ
CO
I
UJ
Q=
CO
0.900 -•
0.800 --
0.700 +
0.600
-A GAC
-• GAC WITH ION EXCHANGE
-i f— 1-
246
TIME (hrs)
8
Figure 38. Ratio of radon mass removed! mass applied for the high loading
rate study for the GAC systems. (August = a; October = b).
(flowrate ~ 20 L/rain; throughput = 1,900 L/day).
77
-------
u
a.
60
50
40
30
20-
10-
0-
AUG
SEP
OCT
NOV
o
a.
o
x
» GAG WITH ION EXCHANGE
SEP
OCT
NOV
Figure 39. Radon activities for the influent (a) and effluent (b) of the
GAC systems for the period of August through November, (note
scale difference). (* denotes sampling event following high
loading experiment).
78
-------
1100
JUL-NOV( 7.6 L/min)
o HIGH FLOW (20 L/min)
100
100
300 500 700 900
MASS APPLIED ( X 1()3 pCi/min)
1100
• JUL-NOV (7.6 L/min)
o HIGH FLOW (20 L/min)
100
300 500 700 900 1100
MASS APPLIED ( X 1fl3 pCi/min)
Figure 40. Radon mass removal as a function of the mass applied for the
fof ti^rl?* "J6 condltlons and f°r ^e normal conditions
nrf^S , vithout pretreatment (a) and the GAG with
pretreatment (b).
79
-------
o
0.
o
I
• • INFLUENT
A A GAC
10
20
30
40
50
o
a.
fO
o
o
o
10-
• « INFLUENT
• • GAC WITH ION EXCHANGE
10 20 30
TIME (hrs)
40
50
Figure 41. Influent and effluent radon activities after backwashing
for the GAC without pretreatment (a) and the GAC with
pretreatment (b).
80
-------
s_^
CO
0.40
0.30
ft 0-20
CO
0.10-
0.00
o GAC
• GAC WITH ION EXCHANGE
BACKGROUND
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 42. Gamma emissions measurements for the GAC systems at 1.5 m
away from the-units' surfaces.
81
-------
Water jackets are often proposed to reduce the gamma exposure dose in the
area surrounding the GAC tank. At the end of the study, the GAC unit with
pretreatment was placed in the center of a 61 cm diameter, 136 cm high
polyethylene plastic tank filled with water. Survey meter measurements were
taken at the surface of the plastic tank surrounding the GAC unit (i.e., 30.5
cm away from the GAC tank itself) and 150 cm away from the GAC tank before and
after the plastic tank was filled with water. These data indicated that the
water jacket attenuated the gamma radiation by 14-17* at 30.5 cm. There was
no measurable attenuation at 150 cm. Though the 14-17% decrease in potential
gamma dose is substantial, it is not sufficient to eliminate the problems
associated with residential radiation exposure when POE GAC units are used to
treat ground water for radon, especially when the influent radon activity is
high.
Uranium/Radium/Lead-210 —
At the end of the study, the GAC and the ion exchange units were cut open
and samples of GAC were collected for high resolution gamma spectrometer
analysis. The retention of radionuclides in these units was a concern because
during the small community radon removal study (Kinner et al., 1990) it became
apparent that the GAC was contaminated with substantial quantities of
radium-226, lead-210 and uranium-238, and required special disposal practices.
All coring data were calculated in kg dry weight of GAC (based on percent
moisture content) and reflect a composite of 2 subsamples of the GAC and ion
exchange resin obtained at particular depths within the beds. Samples of
virgin GAC and resin were also analyzed to assess background levels and the
values obtained were subtracted out of all data presented. Radioactive
standards were prepared using virgin GAC and resin spiked with lead-210.
Comparisons were made, when possible, between the predicted retention
(based on average influent activity and average flowrate for the entire bed)
and the coring data obtained for each radionuclide. The estimated adsorption
usually corresponded fairly well to the coring data considering these
estimates were based on average influent activities and total flow through the
units and did not take into account the actual variations in these parameters
occurring at the site. It is not possible to do an accurate mass balance on
any of the elements sampled because of the natural variation in influent water
quality and flowrate occurring in ground water supplies.
Uranium was removed from the water as it passed through the GAC without
pretreatment (80 + 32% removal), and through the ion exchange (73 + 23%
removal) and GAC with pretreatment (35 + 39% removal). At the pH of the Oerry
water (6.03 + 0".34 removal, range 5.25-6.86), the predominant uranium species
was the neutral UO-CO, (Sorg, 1988). Since researchers (Bean et al., 1964)
have demonstrated that the surface of GAC is primarily nonpolar with a net
negative charge, adsorption should favor positively charged species. However,
work by Weber (1972) has shown that positive species tend to be hydrophillic
and are poorly adsorbed. Therefore, adsorption would be a maximum for neutral
species. Based on these findings, it was expected that some uranium
adsorption would occur in the GAC and ion exchange units. There was uranium
removal observed in the ion exchange and GAC units during treatment (Figure
82
-------
43) and there was good agreement between the coring data and the predicted
uranium removal (Tables 10 and 11). IN.B., Excellent correlations were not
expected between the predicted adsorption and the coring data because of the
variation in influent radionuclide activities and the degree of extrapolation
required when only 3 samples were taken over the entire depth of the filter
beds.] Interestingly, more uranium was retained in the GAG without
pretreatment, especially in the top of the bed. This can be attributed, in
part, to the higher uranium loading to this GAC unit (i.e., the ion exchange
removed some uranium from the water, so the load to the GAC with pretreatment
was reduced). However, it is possible that some uranium may have been removed
by co-precipitation with the iron in the water entering the GAC without
pretreatment (S. Rydell; USEPA Region I; Boston, MA; 1990). The uranium
retention correlated closely with the iron accumulation data (see discussion
of iron).
The removal profile indicated that uranium was removed exponentially
through the bed and as in the small community study would probably eventually
approach exhaustion. Presently, there is no MCL set for uranium, but it is
likely to be in the range 20 to 40 pCi/L (EPA, 1989a). Though the GAC has
some capacity for adsorbing uranium, it seems to be related to pH, alkalinity
and iron deposition and is not considered a viable treatment technique.
Based on the uranium coring data from the GAC unit without pretreatment,
the carbon potentially exceeds the State of New Hampshire's de-minimus
standards for (State of New Hampshire Radiological Health Program, 1983)
uranium-238, but not for uranium-235 (both regulations are 58,410 pCi/kg, 2.5
x 10 Ci/m ). However, to confirm this, samples would need to be taken of
the bed after it was completely homogenized to get a representative activity
per unit mass. The data indicated that neither the ion exchange resin nor the
GAC with pretreatment would probably exceed the de-minimus standards for
uranium-238 or 235. According to the draft EPA Guidelines for Disposal of
Drinking Water Treatment Plant Residues Containing Naturally-Occurring
Radionuclides (1989b), the solid residues would fall in the range 30-300 pCi/g
and could be placed in a stabilized landfill, or could be diluted with
uncontaminated materials to an activity <30 pCi/g for use as construction fill
or an agricultural soil conditioner or for deposition in a landfill.
A small amount of radium removal was detected through the GAC without
pretreatment (Figure 44), while the GAC with pretreatment did not receive any
detectable (<0.10 pCi/L) input of radium because of the efficiency of the
cationic exchange resin (radium-226 removal >99%). As a result, it was
difficult to make comparisons between the predicted radium-226 adsorbed and
the coring data. The coring data indicated that the GAC did adsorb some
radium-226 (Table 12), but there was no significant change in activity with
depth in the bed (a = 0.10). The ion exchange unit adsorbed and retained
substantial amounts of radium-226 even after the unit was regenerated (Table
12). This was not surprising since the resin was designed for adsorption of
cations and a certain degree of irreversible (chemical) adsorption would be
expected.
83
-------
INFLUENT
A—A SEDIMENT FILTER
A—A GAC
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
3 15-.
i
••—• INFLUENT
A A ION EXCHANGE
A-—A GAC
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 43. Uranium for the GAC without pretreatment (a) and the GAC
with pretreatment (b).
84
-------
TABLE 10. COMPARISON OF PREDICTED URANIUM ADSORBED TO THE CORING DATA FOR
THE GAC UNITS
Uraniutn-238 (pCi/kg*)
Location in Bed* GAC Without Pretreatment GAC With Pretreatment
Top
Middle
Bottom
Top
Middle
Bottom
178,854
59,034
25,956
Uranium-235 (pCi/kg+)
7,118
3,034
£,184
64,938
30,400
11,475
2,433
1,057
1,124
X of U-238 + U-235
in Bed
Predicted U=238 + U-235
in Bed*
79,684
214,414
33,216
59,657
Predicted calculation based on total uranium as assessed by EPA Method 908.0
(EPA, 1980). Calculated using average total uranium activity and water
flowrate measured during operating period.
+pCi/kg dry weight of GAC, based on percent moisture data.
In GAC without pretreatment, coring depths were top, middle = 43 cm, bottom
= 93 cm. In the GAC with pretreatment, the coring depths were top, middle =
41 cm, bottom = 93 cm.
85
A .. . : 1-.-. ..-••,;.-;.; '.' ...-': >.»;~i-'-:i.y '^^^^]f'l,"-^^.^&^y^4^!g^^^^^^^f»-i'x^~^^^if'^^''.\i^.'j: •»•••'•' ' f:3fr~. "',;', '' -•
-------
TABLE 11. URANIUM CORING DATA FOR THE ION EXCHANGE UNIT*
Location in Bed
Top
Middle
Bottom
X
Uranium-238
19,221
8,126
8,614
11,987
(pCi/kg+)
Uranium-235
1,866
1,944
2,416
2,075
Because of frequent regeneration, predicted values could not be determined.
+pCi/kg dry weight of resin, based on percent moisture data.
In ion exchange unit, coring depths were top, middle - 38 cm, bottom - 76 cm.
86
-------
1
12-r
10 •
8- •
6-.
4..
2-
• • INFLUENT
A A SEDIMENT FILTER
A A GAC
-\ h
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
12
10-
£ 8-.
6-.
4- •
2--
• • INFLUENT
n—a ION EXCHANGE
• • GAC
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 44. Radium for the GAC without pretreatment (a) the GAC with
pretreatment (b).
87
-------
TABLE 12. RADIUM CORING DATA FOR THE GAC AND ION EXCHANGE UNITS
Radium-226 (pCi/kg )
Location in Bed"1"
Top
Middle
Bottom
X in Bed
GAC Without
Pretreatment
715
651
786
703
Ion Exchange
10,189
21,027
21,773
17,663
GAC With
Pretreatment
240
534
379
427
pCi/kg dry weight of resin, based on percent moisture data.
"*" In GAC without pretreatment, coring depths yere top, middle = 43 cm, bottom
» 93 cm. In ion exchange unit, coring depths were top, middle = 38 cm,
bottom = 76 cm. In GAC with pretreatment, coring depths vere top, middle =
41 cm, bottom = 93 cm.
88
-------
It is well documented that radium is poorly adsorbed by GAC (Sorg and
Logsdon, 1978) since it has the greatest ionic and electropositive nature of
all Group II elements (Cotton and Wilkinson, 1980) rendering it extremely
hydrophillic. Therefore, removal of radium by GAC adsorption alone is
considered unfavorable (Sorg and Logsdon, 1978). It was hypothesized that the
radium found in the GAC cores may have been removed by adsorption or ion
exchange reactions occurring with other solid phases (e.g., Fe(OH)3, MgC03) or
organic matter deposited on the GAC. However, reactions with Fe(OH)~ are
considered unlikely because radium-226 activities were not significantly
higher in the top of the GAC without pretreatment when compared to GAC samples
deeper in the bed (see Iron Coring Data - Table 7).
While the GAC removed some radium from the water, the MCL for total
radium is 5 pCi/L and GAC sorption should not be considered an alternative
treatment technique. Based on the regulations established by the State of New
Hampshire Radiological Health Program (1983), the GAC from the unit without
pretreatment and the ion exchange resin would be considered low level
radioactive waste with respect to radium-226 because they exceeded the de
8 3
minimus standard of 44.39 pCi/kg (1.9 x 10 Ci/m ). The average data from
the GAC unit with pretreatment was below this level. In the draft EPA
Guidelines for Disposal of Drinking Water Treatment Plant Residues Containing
Naturally - Occurring Radionuclides (1989b), only solid wastes containing
>2000 pCi/g of radium-226 must be disposed of at a low level radioactive waste
facility. The POE GAC generated in this study with radium-226 activities <3
pCi/g would not require special treatment. However, the ion exchange resin
(10-22 pCi/g) would require disposal in a stabilized landfill.
The highest activities on the GAC came for lead-210, the long-lived
progeny of radon (Table 9). The State of New Hampshire does not regulate
lead-210, however, the draft EPA guidelines (1989b) have specified special
treatment for activity levels of 30-2000 pCi/g. As with uranium, the carbon
from both units would be in this range (GAC without pretreatment = 235 pCi/g;
GAC with pretreatment = 246 pCi/g). Unless the GAC could be diluted with
uncontaminated materials, it would need to be placed in a stabilized landfill.
The Carbdose 2.0 model (Keene and Rydell, 1989) predicts that the
lead-210 adsorbed on the GAC (using an average influent radon activity =
35,620 pCi/L, a water flowrate = 1,122 L/day and a 902 removal efficiency)
should be 956 pCi/g (wet carbon). Using the daily radon activity and water
flowrate data collected over the course of the study, the predicted lead-210
adsorption (992 and 1,072 pCi/g (wet carbon) for the GAC without pretreatment
and the GAC with pretreatment, respectively) was very similar to the Carbdose
2.0 model. However, the actual lead-210 adsorbed to the GAC was 74 to 75%
less than the Carbdose 2.0 prediction. It is possible that some of the
lead-210 was not retained by the GAC, but.this conclusion cannot be verified
because of the variability in radon loading and removal throughout the period
of operation.
In addition, the ion exchange resin was contaminated with many
radionuclides besides radium-226 and uranium-238 and 235. The cationic resin
89
-------
probably retained many of the naturally-occurring radionuclides present in the
raw water as they are usually cationic. It is also possible that some of
these may have been removed by adsorption or ion exchange reactions occurring
with other solid phases (e.g., Fe(OH),, MgCO.,) or organic matter in the ion
exchange unit. It is interesting to note that the coring samples for the ion
exchange unit were taken just after the unit had completed an entire
regeneration program (Table 2). It appears that normal regeneration with
brine will not completely remove the sorbed radionuclides. The water exiting
the ion exchange unit during the brine regeneration part of the regeneration
program was sampled on one day in November and contained substantial amounts
of total uranium and radium-226 (Table 13) even after the 90 min brine
regeneration process was completed. Disposal of this slurry should be
carefully evaluated to determine the appropriateness of any alternative,
including direct discharge into the residence's septic system or sewer.
Aeration Techniques
Radon —
The diffused bubble unit consistently produced effluents with radon
activities less than or equal to 200 pCi/L (Figure 45). This occurred even
when the air flowrate was restricted because the diffusers accumulated iron
precipitates (May-June, 1989). Removal efficiencies of this magnitude (>99%)
are expected with this technique considering that (i) the influent radon
activity was not exceptionally high, (ii) the blower provided an air to water
ratio (A:W) of approximately 119:1 assuming a water flowrate of 9 L/min
through the unit, and an air flowrate of 1.08 m /min, and (iii) the bubble
size was relatively small (diffuser hole diameter = 0.64 mm). The data agree
with results obtained by other researchers (Lowry et al., 1987; Lowry, 1988).
The bubble plate unit also produced a very high quality effluent
.radon removal) except on a few occasions. This is similar to results reported
by LaMarre (1988 and 1989). The unit experienced several mechanical problems
throughout the course of the study (e.g., solenoid valve failure, pump
failure, leaking plumbing). The most detrimental problem with respect to
radon removal was the clogging of the air intake filter for the blower. This
reduced the air flow to the bubble plate chamber and resulted in the higher
effluent radon activities in August through December. The manufacturer
suggested yearly inspection of the filter, however, that was insufficient in
our application. Once this problem was corrected, the unit began to produce
effluent containing < 200 pCi/L again. The A:W ratio for the bubble plate was
156:1 (assuming a water flowrate of 22.7 L/min and an air flowrate of 3.54
m /min). Unfortunately, the bubble plate unit used in this study was not
equipped with a device to inform the operator that the air flowrate was
restricted. Therefore, inadequate air flow occurred until water quality
monitoring results showed that the effluent radon activity in the water was
unacceptable.
It is not surprising that both of the aeration units worked well at
removing the amount of radon present in the raw water. They have high A:W
ratios compared to those found in other aeration units optimized for radon
90
-------
TABLE 13. TOTAL URANIUM AND RADIUM IN THE BRINE EXITING THE
ION EXCHANGE UNIT .DURING BRINE REGENERATION
Time in Brine Total Uranium"1" Radium-226
Regeneration Process (pCi/L) (pCi/L)
(min)
20
50
90
1,179
918
187
1,498
772
223
The brine regeneration process is 90 min long. It is one step in the entire
ion exchange unit's regeneration program. See Table 2.
+ Analyzed using EPA method 908.0 (EPA, 1980).
**Analyzed using EPA method 903.1 (EPA, 1980).
91
-------
60
50
_J
o
Q.
O
*~
X
o
1
40-
30--
20 1
10--
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
o
a.
o
X
O
Q
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 45. Radon activities for the influent (a) and the effluent from
the aeration units (b). (note scale difference).
92
-------
removal (Kinner et al., 1990). However, the blowers used in these POE units
are the most cost effective based on overall efficiency.
Air Monitoring —
The air monitoring study was conducted to determine if there was a
detectable effect of the exhaust from the aeration units on the ambient air
radon activities surrounding the building. Detectors located at a site "5 km
from the Derry pumphouse registered radon activities < 0.3 pCi/L (the
detection limit of the alpha-track monitors). The detectors located 1.5 to
24 m away from the pumphouse all registered <0.3 pCi/L. [N.B., All of these
detectors were located above the elevation of the vents from the aeration
units.] Monitors inside the puraphouse indicated that the airborne radon
activities were in the range 0.9 to 1.7 pCi/L. This is not atypical for a
building in this region with a dirt floor. Some of the activity may have
resulted from the treatment units, however?, it was not possible to determine
this. The detectors located 1.35 m directly under the vents for the aeration
units registered 1.4 to 2.1 pCi/L. Considering that the radon exiting to
2
vents should have been approximately 228-298 pCi/L this dilution of 10 is
expected. If the exhaust from POE units is vented above a home's roof-line
there should be sufficient dilution to get the radon activity in the plume
diluted to background levels very quickly.
ECONOMICS
A detailed economic evaluation of the technologies examined in this
project for removal of radon using POE treatment was performed using
conventional engineering economic practices. The economic evaluation assumed
all water system components (e.g., well, pressure tank, piping) would already
be in place and therefore, only the costs related to installation and
operation of radon removal systems are presented. Costs have been divided
into two major categories: capital costs and operation and maintenance (O&M)
costs. Capital costs include equipment and installation costs (obtained from
records of actual expenditures made during this project) plus 15% for
engineering and subcontractor fees and 15% for miscellaneous contingencies.
The latter percentages have been recommended by USEPA (1987) for use in
estimating POE treatment costs.
Operation and maintenance costs include power, labor, maintenance and
administrative costs. Power costs were calculated based on equipment
horsepower, operating efficiency, operation period and electrical power cost.
An electrical power cost of $0.10157/kW-hr was used based on a Public Service
of New Hampshire (PSNH) Class G rate scale for household usage below 5 ktf
(J.R. Prescott, Public Service of New Hampshire, personal communication,
1990). Labor and equipment maintenance costs were estimated based on a survey
of available POE service records for carbon and aeration units presently
operating in New Hampshire and Maine. Administrative costs were calculated as
the sum of 20% and 25% of the labor and maintenance costs, respectively
(Cummins, 1987).
93
-------
For comparative purposes, annual costs and production costs were
developed for each treatment system. The production costs were expressed in
conventional units of $/1,000 gallons to be consistent with those reported in
other studies. The annual cost was computed as the sum of the amortized
capital cost and the O&M cost. The amortized capital cost is the total
capital cost amortized over a 5 year time period at a 9% interest rate. A 5
year time period was chosen since it reflects a typical home equity loan
(i.e., typical means of financing FOE costs) payback period (D.L. Perry,
Internal Revenue Service, Consumer Information Service, personal
communication, 1990). The interest rate of 9% reflects the current U.S.
Treasury Bill (T-Bill) interest rate. Production costs were calculated by
dividing the annual cost by the volume of water treated per year assuming the
3
system was operated at a design flow of 1.02 m /day.
All cost figures were updated to first quarter 1990 dollars using the ENR
Construction Cost Index (CCI). The 1967 base year CCI index for first quarter
1990 is 435.
The cost estimates presented in this report are intended to give a
general indication of the economics of the POE radon removal technologies
studied. Specific treatment requirements will vary from site to site and may
not be accurately reflected by the systems or assumptions used in this .
economic analysis.
The cost estimates for the GAC POE systems with and without pretreatment
o
(ion exchange) are shown in Table 14. Based on a design flow of 1.02 m /day,
the production costs for GAC without and with pretreatment are $9.31/1,000
gallons and $12.25/1,000 gallons, respectively. The total production costs
including disposal of the GAC/resin are estimated to be $9.88/1,000 gallons
for GAC without pretreatment and $13.40/1,000 gallons for GAC with
pretreatment.
It should be noted that the O&M cost estimates assume power costs are
negligible and that backwash or ion exchange regenerant streams can be
discharged without additional treatment costs (e.g., to the septic system).
The cost of GAC/resin disposal assumes it must be handled as a regulated
low level radioactive waste and is based on a disposal cost estimate of
$4,400/m (1990 dollars) of GAC/resin (R.L. Kennedy, General Dynamics Service
Company, Reactor Plant Services, personal communication, 1990). It was
assumed that this cost would be incurred at the end of the 5 year life. Thus,
the cost was adjusted to a future cost using a 10% annual inflation rate and
this future cost was then amortized (9% for 5 years) to an annual cost of
•3
$340/m -yr or $57/yr for the GAC without pretreatment and $115/yr for GAC with
pretreatment. This annual cost represents an additional production cost of
$0.57/1,000 gallons for the GAC without pretreatment and $1.15/1,000 gallons
for the GAC with pretreatment. This cost will be higher than the actual
production costs if the GAC and resin can be disposed of in a stabilized
landfill.
94
-------
TABLE 14. COST ESTIMATES FOR THE GAC POE TREATMENT SYSTEMS AT DERRY, NH
Item
Total Cost
(1st Quarter 1990 Dollars)*
GAC Without Pretreatment GAC With Pretreatment
CAPITAL COSTS
Equipment
Installation
Total Direct Cost
Engr./Sub. Fees
Contingencies
TOTAL CAPITAL COST
AMORTIZED CAPITAL COST (Annual)
ANNUAL O&M COSTS
Power
Maintenance
Labor
Administrative
TOTAL ANNUAL O&M COST
TOTAL ANNUAL COST
TOTAL PRODUCTION COST
$ 785
1,200
$1,985
298
298
$2,581
$ 664
NA+
160
45
49
$ 254
$ 918
$9.31/1,000 gal
$1,500
1,400
$2,900
330
330
$3,560
$ 916
NA+
185
50
56
$ 291
$1,207
$12.25/1,000 gal
(1.02 m /day design flow)
GAC/RESIN DISPOSAL COST
TOTAL PRODUCTION COST
WITH DISPOSAL
,**
$0.57/1,000 gal
$9.88/1,000 gal
$1.15/1,000 gal
$13.40/1,000 gal
Based on ENR Construction Cost Index = 435, Base Year 1967 = 100.
Additional power costs incurred by operation of the GAC systems are
considered negligible.
Based on survey of existing carbon POE systems. Includes brine solution
regeneration of ion exchange unit.
** i
Based on estimated disposal cost of $340/m -yr (assuming it is classified as
a low level radioactive waste) and the design flow.
95
-------
Cost estimates for the diffused bubble and bubble plate aeration systems
are shown in Table 15. These data indicate that the total production costs
for the diffused bubble and bubble plate aeration systems were $22.58/1,000
gallons and $26.74/1,000 gallons, respectively. The major difference in cost
for the two aeration systems resulted from the differential in retail price of
the equipment.
A major assumption implicitly made in the economic analysis of the
aeration systems is that the radon present in the off gas will disperse in the
atmosphere and will not require additional treatment. Although the economic
analysis does account for the cost of piping the radon vent above the average
home roof line (i.e., 11 m), no additional costs for auxiliary blowers or air
treatment are included. If subsequent treatment is required to meet federal
or state air quality standards, then the cost estimates could be significantly
higher.
It should be noted that in comparison to total production costs for
public water supplies the POE costs are high. The production costs for the
individual POE systems are higher because (i) there is no economy of scale,
and (ii) there will not be a quantity discount on the equipment when purchased
for an individual well water supply.
OVERALL EVALUATION OF THE POE SYSTEMS
GAG
Of the three systems evaluated, the GAG requires the least owner
maintenance, is the easiest to operate, and is the least expensive with
respect to capital, and operation and maintenance costs (even when low level
radioactive waste disposal costs are included). However, neither GAG system
consistently produced effluent radon activities in the range of the proposed
MCL (200-2,000 pCi/L). The data obtained in our study indicate that the only
condition where POE GAG systems would produce effluent in this range would be
if the influent activity were low (e.g., < 5,000 pCi/L). Influent activities
<5,000 pCi/L would also result in lower gamma emissions from the GAG system as
noted by Rydell et al. (1989) and yield acceptable levels of exposure dose
(Rydell et al., 1989). Water jacket shielding could be used as an added
measure of safety, but it would make accessibility to the unit difficult if
repairs or maintenance were required. At such low influent loadings, the
small reduction in gamma exposure dose from the water jacket would have a
minimal effect on o'verall exposure.
As noted in the small community radon study, the operation of GAG systems
usually requires either pretreatment and/or backwashing to prevent fouling
resulting from accumulation of particulates and metal precipitates or fouling
by bacteria and organic matter. Ve also hypothesized from the small community
data that iron precipitates, bacteria or organics could limit the effective
life of the GAG bed. Comparing the data from the GAG units with and without
cationic pretreatment, it appears that retention of iron precipitates in the
bed could shorten the life of the GAG with respect to radon removal and make
it more susceptible to changes in loading. Both beds were similarly fouled
with bacteria and organic matter. Turbidity was removed by the ion exchange
96
-------
TABLE 15. COST ESTIMATES FOR AERATION POE TREATMENT SYSTEMS AT DERRY, NH
Item
Total Cost
(1st Quarter 1990 Dollars)
Diffused Bubble Bubble Plate
CAPITAL COSTS
Equipment
Installation
Total Direct Cost
Engr./Sub. Fees
Contingencies
TOTAL CAPITAL COST
AMORTIZED CAPITAL COST (Annual)
ANNUAL O&H COSTS
Power ^
Maintenance
Labor
Administrative
TOTAL ANNUAL O&M COST
TOTAL ANNUAL COST
TOTAL PRODUCTION COST
(1.02 m3/day design flow)
$2,215
955
$3,170
476
476
$4,122
$1,060
80+
345
545
195
$1,165
$2,225
$22.58/1,000 gal
$3,295
955
$4,250
638
638
$5,526
$1,421
54"
368
583
209
$1,214
$2,635
$26.74/1,000 gal
.Based on ENR Construction Cost Index = 435, Base Year 1967 = 100.
+ Based on input power of 0.5 Hp for blower operating 3.5 hr/day, 0.5 Hp for
pump operating 2.3 hr/day, and an electric rate of $0.10157/kW-hr.
Based on input power of 0.33 Hp for blower operating 2.4 hr/day, 0.5 Hp for
**.
pump operating 2.3 hr/day, and an electric rate of $0.10157/kW-hr.
Based on survey of existing service contracts for aeration POE systems in
Maine and New Hampshire.
97
-------
unit lowering that load to the subsequent GAC unit. Obviously, in FOE
applications where iron and manganese concentrations are high, ion exchange
pretreatment is essential. However, its use is not without problems;
specifically the contamination of the resin with long-lived radionuclides and
the disposal of the heavily-contaminated backwash water and regenerant brine.
If frequent backwashing of the GAC unit and use of a sediment filter can limit
the accumulation of particulates and metal precipitates.in the GAC bed, then
ion exchange pretreatment should be avoided.
One of the primary conclusions of the small community radon removal study
was the need to have sufficient data on variation in water flowrate and raw
water radon activity. Based on the loading experiments, this seems to be
somewhat less important for POE GAC units, especially if they are used to
treat water with <5,000 pCi/L of radon where excess GAC can be supplied for
little extra cost. At higher loadings, however, the GAC units may not be able
to dampen the excess radon applied without-some increase in effluent activity.
One of the ancillary problems associated with GAC treatment systems is
the retention of longer-lived radionuclides, such as uranium-238, radium-226
and l.ead-210, in quantities which may dictate that the GAC must be disposed of
in a special manner. This issue may prohibit the use of GAC in some
applications where uranium and radium are present in large quantities,
however, even if they are absent, lead-210 will always accumulate in the GAC
as it is the long-lived progeny of radon. A more thorough evaluation must be
made to determine whether for POE applications where the radon activity is
<5,000 pCi/L the benefits of GAC (e.g., cost, ease of operation and
maintenance) are outweighed by problems associated with specialized handling
and disposal.
Aeration Techniques
Both the bubble plate and the diffused bubble POE units are very
efficient (>99Z) at removing radon from the water even at relatively high
concentrations as demonstrated by this study and those of other researchers
(Lowry et al., 1987; Lowry, 1988; LaMarre, 1988 and 1989). It appears that
they will be able to meet an MCL of 200-2,000 pCi/L in most cases primarily
because of the high A:V ratios used. Variations in influent radon activity
and water flowrate should be handled without a significant increase in
effluent radon activity.
There are several problems associated with the aeration treatment
techniques. The primary issue is iron oxidation which occurs readily in these
units. Even at iron concentrations below 1.0 mg/L, iron precipitates can form
and accumulate in the units and/or be released in pulses to the residence when
the unit is started to meet demand. In the diffused bubble unit, iron
accumulation on the diffusers caused a gradual decrease in the air flowrate.
An attempt was made to clean the diffusers using a low pH solution (sodium
hydrosulfite, pH = 3.67) as recommended by the manufacturer. This did not
improve performance, so the diffusers were replaced in June after 6 months of
operation. The problem did not recur during the remainder of the study.
98
-------
In order to avoid problems with iron precipitation in the units or
problems within the residence, iron treatment will probably be required in
almost all POE applications using aeration techniques. If the raw water is
pretreated to remove iron before it enters the aeration device, precipitate
problems should be completely avoided. This may be necessary in the diffused
bubble units to prevent the diffusers from clogging. However, if pretreatment
involves ion exchange resins, the problems previously discussed, such as
radionuclide retention, must be solved. Sand filtration has been installed in
some POE applications following bubble plate aeration units. These filters
are designed to remove the iron precipitates formed during the aeration
process. Before this alternative is used widely, the chemistry/biology of the
filters should be examined to insure that they are effective and do not have
problems with radionuclide retention associated with organic matter or solid
phases {e.g., Fe(OH)- or MgCO,). Further, the required backvashing frequency
and related costs of these units should be thoroughly evaluated prior to
widespread use.
POE aeration units are more expensive with respect to capital and
operation and maintenance costs as compared to GAG units. In addition, they
require much more frequent maintenance to insure that an adequate air flowrate
is maintained and an extra pump is needed to repressurize the water system.
It is particularly important to maintain an adequate air flowrate and hence,.
the air intake filter to the blowers should be cleaned frequently, every 1-3
months or more in dusty locations. There should be gauges which the owner can
monitor to insure the proper air pressure/flowrate is maintained. This is not
a surrogate for radon monitoring, but is the only way a homeowner could check
that the air flowrate is correct. Addition of audible alarms to warn the
homeowner of potential low air flow or pump failure is necessary, as the units
will usually be in a remote location.
Proper venting of the aeration units is important. As demonstrated in
this study, the plume is diluted fairly rapidly. However, the off-gas stack
should be vented above the roofline to prevent any radon from re-entering the
home. Placement of external auxiliary blowers may be required to produce
adequate flow in these vent pipes.
During operation, the blowers and pumps associated with the aeration
units are very noisy. Either the units should be put in a remote location or
some soundproofing may be required, so that homeowners are not tempted to turn
the units off to avoid the noise. In addition, the secondary jet pump
required by the aeration systems will have a low pressure cut-off feature.
This feature protects the pump from damage in the event of air binding due to
the loss of system pressure (e.g., a power failure). Although the feature is
required to avoid costly pump replacement, it will be a nuisance to the owner
since it will require manual system restart. It is recommended that
directions for the proper restarting procedure be clearly displayed on the
unit. Despite these efforts, the low pressure cut-off feature may lead owners
to by-pass the treatment system.
99
-------
Summary Evaluation for POE Systems
There were problems with the bacterial numbers in the effluent from all
of the units tested. Depending on EPA and state regulations, it may be
necessary to disinfect the water before use to avoid public health problems.
It should be noted that none of the units were tested to evaluate its
performance at lower radon loading rates. In the small community radon
removal study, it was stressed that extrapolation of results in full-scale
application should not be made to systems with much lower influent activities.
This is probably not as important for the POE units because they are usually
overdesigned (e.g., excess GAC or high A:W ratios), but it should be
considered.
Finally, frequent monitoring of the effluent quality of any POE unit
should be stressed to the homeowner. Because radon cannot be detected by
taste, smell or any other human sense, there is a potential for it to be
reintroduced into the water supply without the homeowner knowing it if a POE
unit fails. Traditionally, conventional engineering wisdom has suggested that
using POE treatment to protect public health is only viable as a short term
solution due to lack of proper maintenance and monitoring by the homeowner.
Since widespread use of POE treatment units for radon removal is likely, it is
vital that water analyses of the effluent at the tap be conducted at least
every 3 months to insure risk reduction is maintained.
100
-------
REFERENCES
Alben, K. et al., Experimental Studies of Distribution Profiles of Organic and
Inorganic Substances Adsorbed on Fixed Beds of Granular Activated Carbon.
ACS Advances in Chemistry Series, No. 202 Treatment of Water by Granular
Activated Carbon (1983).
Amirtharajah, A., Some Theoretical and Conceptual Views of Filtration. Jour.
AWWA, 80:12:36 (1988).
Anderson, J.S., and Vatson, K.S., Patterns of Household Usage. Jour. AWWA,
79:10:1228 (1987).
Bean, J.R., et al., Zeta Potential Measurements in the Control of Coagulation
Chemical Doses. Jour. AWVA, 56:2:214 (1964).
Bourbigot, M.M., et al., Treitement Biologique de 1'Eau Potable. Mesure de
,1'Activite Materiaux Filtrants. Techniques et Sciences Municipales -
1'Eau, 12:639 (1989).
Camper, A.R., et al., Bacteria Associated with Granular Activated Carbon
Particles in Drinking Water. Appl. Environ. Microbiol., 52:3:434 (1986).
Camper, A.K., et al., Growth and Persistence of Pathogens on Granular
Activated Carbon Filters. Appl. Environ. Microbiol., 50:6:1378 (1985).
Camper, A.K., et al., Operational Variables and the Release of Colonized
Granular Activated Carbon Particles in Drinking Water. Jour. AWWA,
79:5:70 (1987).
Cotton, F.A. and Wilkinson, G., Advanced Inorganic Chemistry. 4th ed. Wiley
and Sons, New York (1980).
Cummins, M.D., Removal of Radon from Contaminated Ground Water by Packed
Column Air Stripping. USEPA Draft Report, Office of Drinking Water,
Technical Support Division. Cincinnati, OH (1987).
Dixon, K.L. and Lee, R.G., Radon Survey of the American Water Works System;
Radon in Ground Water. Lewis Publishers, Chelsea, MI (1987).
Dixon, R.L. and Lee, R.G., Occurrence of Radon in Well Supplies. Jour. AWA,
80:8:65 (1988).
Eberhardt, M. "Experience with the Use of Biologically Effective Activated
Carbon Filters." Translation of Reports on Special Problems in Water
Treatment Technology - Vol. 9, Adsorption, EPA 600/9-76-030, p. 312
(1976).
EPA. Prescribed Procedures for Measurement of Radioactivity in Drinking
Water. EPA-600/4-80-032 (August 1980).
101
-------
EPA. Technologies and Costs for the Removal of Radon from Potable Water
Supplies (Fourth Draft), Office of Drinking Water (January 1987).
EPA. Status of the Radionuclides Proposal. Office of Drinking Water. (January
1989a).
EPA. Draft Guidelines for the Disposal of Drinking Water Treatment Plant
Residues Containing Naturally Occurring Radionuclides Criteria and
Standards Division, ODW (WH-500D). (1986b).
Evans, R.D., Engineers Guide the Elementary Behavior of Radon Daughters.
Health Physics, 17:229 (1969).
Federal Register., Water Pollution Control; National Primary Drinking Water
Regulations; Radionuclideds; Advanced Notice of Proposed Rulemaking.
51:189. (September 1986).
Fleming, B.C., Microbial Growth on Ion Exchangers. Wat. Res., 21:7:745
(1987).
Graese, S.L., et al., Granular Activated Carbon Filter Adsorber Systems.
Jour. AWWA, 79:12:64 (1987).
Hess, C.T., et al., The Occurrence of Radioactivity in Public Water Supplies
in the United States. Health Physics, 48:5:553 (1985).
Hubele, C. and Sontheimer, H., "Adsorption and Biodegradation in Activated
Carbon Filters Treating Preozonated Humic Acid." in Proceedings of the
1984 ASCE Specialty Conference on Environmental Engineering, p. 376
(1984).
Johnson Division, UOP, Inc. Ground Water and Wells. Johnson Division, UOP,
Inc., Saint Paul, MN. (1975).
Keene, B. and Rydell, S., "CARBDOSE. BAS Version 2.0." USEPA Region I.
Boston, MA. (1989).
Kinner, N.E., et al., Radon Removal Techniques for Small Community Public
Water Supplies. U.S. EPA Risk Reduction Engineering Laboratory,
Cincinnati, OH. 248 p. (1990).
LaMarre, B.L., Removal of Radon from Residential Water Supplies by a Unique
Aeration Method. Proc. USEPA Symposium on Radon and Radon Reduction
Technology. Denver, CO. (October 1988).
LaMarre, B., World's "Hottest" Radon Well. Well Water Jour., June: 42 (1989).
Larsen, I.L. and Cutshall N.H., Direct Determination of Be-7 in Sediments.
Earth Planet. Sci Lett., 54:379 (1981).
102
-------
Lessard, C.E., An Evaluation of the Degree of Metal Contamination in a Small
River. Masters Thesis. Department of Civil Engineering. University of
New Hampshire. Durham, N.H. (1987).
Longtin, J.P., Occurrence of Radon, Radium, and Uranium in Groundwater. Jour.
AWWA, 80:7:84 (1988).
Lowry, J.D. and Brandow, J.E., Removal of Radon from Water Supplies. Jour.
Env. Eng. (ASCE), 111:4:511 (1985).
Lowry, J.D., et al., Point-of-Entry Removal of Radon from Drinking Water,
Jour. AWWA, 79:4:162 (1987).
Lowry, J.D. and Lowry, S.B., Modeling Point-of-Entry Removal by GAC. Jour.
AW A, 79:10:85 (1987).
Lowry, J.D., et al., New Developments and Considerations for Radon Removal
from Water Supplies. Proc. USEPA Symposium on Radon and Radon Reduction
Technology. Denver, CO (October 1988).
Miller, J.C. and Miler, J.N., Statistics for Analytical Chemistry. John Wiley
and Sons, New York, NY (1984).
Miller, R.E. et al., Organic Carbon and THM Formation Potential in Kansas
Groundwaters. Jour. AWWA, 82:3:49 (1990).
Roberts, P.V. and Summers, R.S., "Performance of Granular Activated Carbon for
Total Organic Carbon Removal." Jour. AWWA 74:2:107 (1982).
Rubin, A.J. and Mercer, D.L., Adsorption of Free and Complexed Metals from
Solution by Activated Carbon; Adsorption of Inorganics at Solid Liquid
Interfaces. Ann Arbor Science Publishers, Ann Arbor, MI (1981).
Rydell, S. et al., Granular Activated Carbon Water Treatment and Potential
Radiation Hazards. Jour. New England Water Works Assoc., 103:4:234
(1989).
Sorg, T.J. Methods for Removing Uranium from Drinking Water. Jour. AWWA,
80:7:105 (1988).
Sorg, T.J. and Logsdon, G.S., Treatment Technology to Meet Interim Primary
Drinking Water Regulations for Inorganics: Part 2. Jour. AWWA, 70:7:379
(1978).
Standard Methods for the Examination of Water and Wastewater. (16th edition)
APHA, Washington, D.C. (1985).
State of New Hampshire Radiological Health Program. New Hampshire Rules for
Control of Radiation. Division of Public Health Services. Concord, NH
(April 1983).
103
- • ''•'^-•^-ir.:»<.V-:w'i^iN^ —. . •••" . • ••
-------
Vitz, E., Preliminary Results of a Nationwide Waterborne Radon Survey. Proc.
USEPA Symposium on Radon and Radon Reduction Technology. Denver, CO
(October 1988).
Weber, W.J., Jr. PhyBiochemical Processes for Water Quality Control. John
Wiley and Sons, New York (1972).
Weber, W.J., Jr., et al., "Adsorption of Humic Substances: The Effects of
Heterogeneity and System Characteristics." Jour. AWWA, 75:12:612 (1983).
Wilcox, et al., Microbial Growth Associated with Granular Activated Carbon in
a Pilot Water Treatment Facility. Appl. Environ. Microbiol., 46: 2:406
(1983).
104
-------
APPENDIX A
SUMMARY OF QA/QC DATA
0.200
0.150
0.100
0.050-
0.000
OpH=4STD
• pH=7STD
CONTROL LEVEL
o o
«
o
o
•4 1-
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC JAN
0.200
¥
•S 0.1504-
Q
O
0.100-
^ 0.050-0
0.000
00
0
OBOOOOO
ooo?
o
0 <00 OODO °O° °° (
m ooaDaDO °
-------
0.025
0.015
*
O
0.020--O 0000 O OOO
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC JAN
10
o»
Q
CONTROL LEVEL - 5 mg/L
— — — • •^^"F • ^•^^•WNMMT^^I^^N
100 200 300 400
SAMPLE NUMBER
500
Figure A-2. QA/QC normality of titrant (a) and precision (b) for alkalinity
measurements. (* denotes new solution)
106
'''' ' ''"'"" ' '
-------
a
o
0.500
0.400 --
0.300 - •
0.200-•
0.100
0.000
LU
UJ
or
LU
o
or
LJ
a.
180
140 •-
100--
60-
20
CONTROL LEVEL
100 200
SAMPLE NUMBER
—o-
300
UPPER CONTROL
LOWER CONTROL
100 200
SAMPLE NUMBER
300
Figure A-3. QA/QC precision (a) and accuracy (b) for calcium measurements.
107
-------
o
§
a
Q
Q
0.400
0.350
0.300
0.250
0.200-
0.150-
0.100-
0.050-
0.000-
C
o
o
o
CONTROL LEVEL = 0.13
WARNING LEVEL = 0.09
°o o o oooo
00 °
100 200 300 400 50
0
SAMPLE NUMBER
Figure A-4. QA/QC precision for turbidity measurements.
108
-------
o>
g
o
o
0.100
0.080-•
0.060-•
0.040
0.020-
0.000
CONTROL LEVEL
•
• • o«
0 200 400 600 800 1000 1200
SAMPLE NUMBER
1 UW
£
2 120-
o
0
UJ
0£.
1—
2
g 80-
UJ
o_
40
b
UPPER CONTROL •
• * " ...
• ••• »•
• .. • • *
• *•• • •«. • •
• •• •••••••• •" • * *S
• ^B0* 0 • (
* • * a « « •• ** * •
• . • • • f
: • ••• •
• • •
» LOWER CONTROL . .
1 1 1 1 1 1_
0 200 400 600 800 1000 1200
SAMPLE NUMBER
Figure A-5. QA/QC precision (a) and accuracy (b) for iron measurements.
109
-------
0.100
< 0.080
j[
1 0.060
CO
0.040-
0.020-
0.000
CONTROL LEVEL
• ••
0 200 400 600 800 1000 1200
SAMPLE NUMBER
LU
LU
a.
160
120
80--
40
UPPER CONTROL
. .
. •
• w .
LOWER CONTROL
0 200 400 600 800 1000 1200
SAMPLE NUMBER
Figure A-6. QA/QC precision (a) and accuracy (b) for manganese measurements.
110
-------
Q
Q
0.400
0.300-•
0.200 --
0.100-
0.000
180
cr
o
LL)
o:
uu
o
o:
LL)
Q.
140--
100--
60--
20
50 100 150 200 250
SAMPLE NUMBER
0 50 100 150 200 250
SAMPLE NUMBER
Figure A-7. QA/QC precision (a) and accuracy (b) for NPDOC measurements.
Ill
-------
JL.
UJ
O
o:
UJ
Q_
o
s
UJ
Q
Q
ce
1
*™/ ^^ f ),_
0 200 400 600 800 ~ ~ 1 00
SAMPLE NUMBER
Figure A-8. QA/QC precision for radon measurements.
112
-------
APPENDIX B
SUMMARY OF LIQUID SCINTILLATION COUNTER PROGRAM
The program outlined below was obtained from the operations manual for
the Beckman (Fullerton, CA) LS 7000 liquid scintillation counter used in this
research project. It is library program tt6 for this instrument.
Unmodified Count Time 10 min for all samples except 50 rain for
effluent from aeration units
Channels Counted Lower Limit Upper Limit
1 2.00a%* 0 397
2 2.00o% 397 940
Sample Channels Ratio No
H Number Calibration 1 time (in channels 1 & 2)
Automatic Quench Compensation Yes
Data Printout Units cpm
2.009% - 200//N where N - total number of counts obtained at the time of
calculation.
The H-number quantifies, quench and indicates the effect of interferences on
the counting efficiency. Automatic quench compensation uses an internal
cesium-137 source to generate an H-number for each sample, which the liquid
scintillation counter uses to adjust the count rate generated in
correspondence to the quenching.
113
-------
APPENDIX C
SUMMARY OF THE RAW WATER CHARACTERISTICS
AT THE DERRY, NH POE SITE
Constituent
Concentration
Radon
pH
Temperature
Alkalinity
Calcium
Turbidity
Microbial Numbers
NPDOC"1"
Total Iron
Soluble Iron
Total Manganese
Soluble Manganese
Uranium
Radium
35,620 ± 6,727 pCi/L
6,24 ± 0.19
11.8 ± 2.5°C
34 ± 19 mg/L as CaCOj
15.5 ± 1.15 mg/L
1.04 ± 0.94 NTU
30,400 ± 29,500 CPU/100 ml.
1.30 ± 0.28 mg/L
0.40 ± 0.27 mg/L
0.32 ± 0.22 mg/L
0.36 ± 0.12 mg/L
0.35 ± 0.11 mg/L
14.4 ± 6.5 mg/L
3.5 ± 2.4 pCi/L
Mean ± Standard Deviation
+NPDOC = Non-Purgeable Dissolved Organic Carbon
114
-------
APPENDIX D
METHODS OF CALCULATION OF LEAD-210 ADSORPTION
I. Calculation of Lead-210 activity from gamma counter data using channels
89-96.
Net Peak Counts
(m) (E) (Ig)
where A' = activity (dpm/g GAC dry weight)
Net Peak Counts = (Adjusted Peak Counts - Adjusted Compton Counts)
Adjusted Peak = (Sample Peak - Background Peak)
Counts Counts Counts
Adjusted Compton = (Sample Compton - Background Compton)
Counts Counts Counts
t » time sample counted (min)
m - mass of sample (g)
E a efficiency factor based on counts obtained from Lead-210
standard sorbed to equivalent volume of virgin GAC (cpm/dpm)
I = 0.04 = amount of Lead-210 emissions which are gamma radiation
O
. 1.000 g
2.22dpm kg
. . _ pCi Lead-210 measured
wnere A - weight)
II. Calculation- of theoretical Lead-210 accumulated on the GAC.
Assumptions:
(i) All radon-222 removed by the GAC was completed retained.
(ii) All lead-210 accumulation originates from decay of the radon-222
progeny. Any direct adsorption of lead-210 from the water supply
or lead-210 resulting from the decay of other adsorbed species
such as radium-226 was considered negligible.
(iii) The measured water flowrate and radon removed by a given volume
of GAC were constant over a sampling interval.
115
-------
A. Calculation of pCi of Radon-222 adsorbed in a given volume of GAC
during the entire operating period (up to time of coring)
V' Vc**. - c<«iHl
where A. = pCi radon adsorbed during sampling interval i
Q = water flowrate during sampling interval i (L/d)
wi
C. , <= measured influent radon into volume of GAC during
infi sampling period i (pCi/L)
C ft - measured effluent radon out of volume of GAC during
effi sampling period i (pCi/L)
(C. -C..)=net change in radon activity through
infi e"i volume of GAC (pCi/L)
t. = length of time of sampling interval (d)
n
the
where A™, - pCi of radon adsorbed to a given volume of GAC during
Rn the entire operating period (up to time of coring)
.B. Calculation of mass of Radon-222 adsorbed to a given volume of
GAC
M A v 6.50 x 10"18g Rn
"Rn = ^Tjk X pCi Rn
where Mp » g of radon adsorbed to a given volume of GAC during the
operating period (up to the time of coring)
6.50 x 10"18g Rn 2.22 dpm _1_ v mole Rn 222gRn
pCiRn - IpCi * XRn * 6.023 x 1023 atoms mole Rn
In 2 0.693 , ,fi lo-4 . -1
tTT" = (3.82d)(24hr/d)(60min/hr) = 1'26 X 10 min
Rn
where X-^ » number of radon atoms disintegrating per total radon
atoms present per minute
t,/9 = half-life of Radon-222 = 3.82d
17 ^
116
-------
C. Calculation of atoms of Radon-222 absorbed to a given volume of GAC.
23
M M mole Rn 6.023 x 10 atoms Rn
Rn = Rn 222gRn mole Rn
where ND = number of atoms of radon adsorbed to a given volume of
to GAC
D. Calculation of mass of Lead-210 (Pb) adsorbed to a given volume
of GAC
NPb = NRn
where Np, = number of atoms of Lead-210 adsorbed to a given
volume of GAC
M M v mole Pb _ 210g Pb
Pb ' Pb 6.023 x 1023 atoms mole Pb
where MDK = g of Lead-210 adsorbed to a given volume of
Pb of GAC
E. Calculation of pCi of Lead-210 (Pb) adsorbed to a given volume of
GAC
A_ =M pCi Pb
Pb 1.24 x 10~14g Pb
PCi Pb PCi . 6.023 x 1023 atoms Pb mole Pb
1,24 x 10'14g Pb = 2<22 dpm b mole pb
where A_ = pCi Lead-210 adsorbed to a given volume of GAC
rPb
In 2 0.693 , 7R ln-8 m. -1
=
where Xp. =. number of lead-210 atoms disintegrating per total
lead-210 atoms disintegrating per total Pb atoms
present per minute
t1/7 . half-life of Lead-210 = 21 yr
lx
-------
TECHNICAL REPORT DATA
(Please read Instructions on the nveru before completing)
3. RECIPIENT'S ACCESSION NO.
. TITLE AND SUBTITLE
RADON REMOVAL USING POINT-OF-ENTRY WATER TREATMENT
TECHNIQUES
, REPORT DATE
8. PERFORMING ORGANIZATION COOE
IOR(S)
Nancy E. Kinner, James P. Malley, Jr. and
Johnathan A. Clement
8. PERFORMING ORGANIZATION REPORT NO.
BNC1A
G ORG>
IZATION NAME AND ADDRESS
Enyi ronmental Research Group
University of New Hampshire
Durham, New Hampshire
10. PROGRAM ELEMENT NO."
CR812602
11. CONTRACT/GRANT NO.
Final 10/88-6/90
12. SPONSORING AGENCY NAME AND ADDRESS
Risk Reduction Engineering Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
' Cincinnati, OH 45268
IS. SUPPLEMENTARY NOTES
Project Officer: Kim R. Fox (513) 569-7820
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY COOE
The purpose of this EPA Cooperative Agreement was to evaluate the performance of POE
granular activated carbon (GAC), and diffused bubble and bubble place aeration systems
treating a ground water supply containing radon (35.-620 ±.6,717 pCi/L. The pattern of
-4pading to the units was designed to simulate daily demand in a household. Each of the
systems was evaluated with respect to three primary factors: radon removal efficiency,
potential problems, and economics.
'The radon removal efficiencies of the POE GAC units gradually deteriorated over time
from 99.7% to 79% for the GAC without pretreatment and 99.7% to 85% for the units pre-
ceded by ion exchange. It appeared that iron sorption caused fouling of the GAC surface
or caused channeling of water through the bed. The gamma emissions data indicated that
the zone of radon removal slowly moved down the GAC bed. At higher radon loadings, the
GAC units may not be able to dampen variations in influenc activity and flow without
some increase in effluent activity.
The bubble plate and diffused bubble POE units were very efficient ( 99%) at removing
radon from the water. This resilience is primarly due to the high air to water ratios
supplied by the aeration blowers. One major problem associated with the aeration
techniques is iron oxidation/precipitation. Iron treatment will be required in most al
.PHF applirar-mne uih-.,-h ijea aoya-Mnn ~
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lOENTIFIERS/OPEN ENDED TERMS
COSAT1 Field/Group
DISTRIBUTION STATEMENT
RELEASE TO PUBL'IC
19. SECURITY CLASS {Tilts Report)
UNCLASSIFIED
21. NO. OF PAGES
2O. SECURITY CLASS (Tliispags)
UNCLASSIFIED
22. PRICE
EPA Form 2220—1 (R«». 4—77) PREVIOUS .EDITION is OBSOLETE
\
------- |