United States
  Environmental Protection
  Agency
 Risk Reduction
 Engineering Laboratory
 Cincinnati, OH 45268
  Research and Development
 EPA/600/S2-90/036  Nov.  1990
  Project  Summary

  Radon Removal Techniques for
  Small  Community  Public
  Water  Supplies
  Nancy E. Kinner, James P. Malley, Jr., Jonathan A. Clement,
  Peter A. Quern, and Gretchen S. Schell
   The  report  summarized  here
 presents the results of an evaluation
 of radon removal in small community
 water supplies with the use of full-
 scale granular activated  carbon ad-
 sorption, diffused  bubble aeration,
 and  packed tower  aeration.  Various
 low technology alternatives, such as
 loss in a distribution system  and
 addition of coarse bubble aeration to
 a  pilot-scale atmospheric storage
 tank, were  also evaluated.* The full
 report  discusses each of the treat-
 ment alternatives  with respect to
 their radon removal  efficiency, poten-
 tial problems (i.e.,  waste disposal,
 radiation exposure, and  intermedia
 pollution), and economics in small
 community  applications. In addition,
 several  sampling methods, storage
 times, scintillation cocktails, and ex-
 traction procedures  currently  used in
 the liquid scintillation technique for
 analysis of radon in water were
 compared.
   This Project  Summary  was
developed by EPA's Risk Reduction
Engineering  Laboratory, Cincinnati,
OH, to  announce  key findings of the
research project that is fully
documented  in a separate report of
the same title (see Project  Report
ordering information  at back).

Introduction
  As part of  the  1986  amendments to
Ihe Safe Drinking Water Act, the U.S.
 * Environmental Research Brief EPA/600/M-
  87/031 also outlines the results of the low
  technology study
 Environmental  Protection Agency (EPA)
 is to propose a rule  for  Maximum
 Contaminant Level Goals (MCLGs)  and
 National Primary Drinking Water Regula-
 tions (NPDWR) including Maximum Con-
 taminant Levels (MCLs) for radionuclides
 in drinking  water. One of  the  radio-
 nuclides that will be regulated under the
 rule  is  radon-222  (radon). EPA is
 considering setting the MCL for radon in
 the range of 200 to 2,000 pCi/L. Data on
 the distribution of radon  in groundwater
 supplies in the United States indicate that
 a large number of individual and public
 water supplies will be affected  by  an
 MCL in that range. In addition, many of
 these public water supplies will be ones
 serving small communities (<76 m3/day).
   The  rule  will  also  contain rec-
 ommendations with respect to Best Avail-
 able  Technologies (BATs) and analytical
 methods. Three conventional water treat-
 ment technologies (granular activated
 carbon  (GAG),  diffused bubble aeration,
 and packed  tower aeration)  have been
 used to remove  radon  from  drinking
 water. The GAG process, which has been
 used  to  treat point-of-entry  and small
 community water supplies, relies on  the
 ability of radon to adsorb to the carbon.
 One unique aspect of  this process is that
 the breakthrough/exhaustion profile typ-
 ically seen when  GAC is treating con-
 servative (nondecaying) contaminants is
 not exhibited during radon removal.
  The  aeration  methods  are  used
because  radon is a  highly  volatile gas
with a relatively large Henry's constant
(2.80  atm • m3H2O/m3 air at 100°C) that
can be  easily transferred from water to
air. Aeration methods  have been used in

-------
individual, small community,  and other
public water supplies.
   The primary purpose of this study was
to evaluate the  performance of full-scale
GAG and diffused bubble and packed
tower aeration  systems  when  treating
small community water supplies contain-
ing  radon.  In  addition,  several low
technology alternatives and various mod-
ifications of the  liquid  scintillation
counting  technique used for analysis of
radon in water were evaluated.
   The specific objectives of  the study
were to:
•  Evaluate full-scale  GAC   systems
   operating at two small communities in
   New  Hampshire,  by monitoring them
   for  changes  in  radon  removal,
   radiation emissions, and general water
   quality parameters  (e.g.,  pH, iron,
   turbidity, microbial numbers);
•  Conduct several  specific  monitoring
   events of the GAC systems to assess
   the  effect  on  GAC  performance  of
   diurnal variations in water flowrate and
   raw  water quality, high water flowrate,
   and  backwashing;
•  Core the GAC after several  months of
   operation  to  determine  if  iron,
   manganese,  microorganisms,  radio-
   nuclides,  or  all  of  them  were
   accumulating in the units;
•  Evaluate full-scale diffused bubble and
   packed tower aeration  systems
   operating in small communities in New
   Hampshire  by monitoring  them  for
   radon  removal, general water quality
   parameters, and off-gas emissions of
   radon;
•  Operate the  aeration  systems over  a
   range of volumetric air to water (A:W)
   ratios  at  two water flowrates  to
   determine the effect of  these param-
   eters on radon removal;
•  Evaluate three  randomly  packed
   plastic  media in  the tower  aeration
   system with respect to radon removal
   efficiency;
•  Evaluate the  radon removal efficiency
   of several  low technology  modifica-
   tions (e.g., free fall vs.  bottom  entry,
   spray  nozzle entry, venturi entry,
   coarse  bubble  aeration) retrofitted  to
   an existing  small  community atmos-
   pheric storage tank;
•  Assess the  effect  of   sampling
   techniques  (e.g.,  free fall,  hose
   connector,  direct syringe  collection,
   volatile organic analysis (VGA) bottle
   collection), storage time, scintillation
   cocktail, and  extraction via shaking on
   the liquid scintillation analytical tech-
   nique for analysis of radon in water.
Analytical Methods

   Standard  Methods and EPA methods
were used to determine  radon; gamma/
beta emissions; the  activity  of total
uranium,  radium-226,  and  lead-210;
microbial counts;  pH;   temperature;
dissolved oxygen; turbidity; total iron; and
manganese.  The  University  of New
Hampshire  and  the  State  of  New
Hampshire conducted the analyses; when
commercial equipment was used, it was
calibrated  according  to the manu-
facturer's directions.
Granular Activated Carbon
   Downflow  GAC systems were installed
at two mobile home  parks  located  in
Amherst  and  Mont  Vernon,  NH.  The
Amherst water system serves 56 homes
at an  average daily flow of 59  ±4
m3/day. The water  is obtained from one
well containing an average radon activity
of 49,500  ±11,200  pCi/L.  The  Mont
Vernon system supplies 40 homes  at an
average flow of 31   ±11 m3/day. Water is
obtained from two wells with an average
radon activity of 222,000  ±  52,000 pCi/L.
The  well  water in  both  systems  was
pumped to unpressurized (atmospheric)
storage  tanks  and,  subsequently,
pumped through the GAC systems upon
demand from the community.
   The systems were  evaluated accord-
ing to  the study  objectives. Radon and
several genera! water quality  parameters
(alkalinity,  turbidity, dissolved  oxygen,
temperature,  pH, iron, manganese, and
bacterial  numbers)  were monitored  at
each site  Uranium  and radium were also
monitored  in the water supply. The GAC
was cored and analyzed for accumulation
of uranium-238,  uranium-235,  radium-
226, lead-210,  iron,  manganese,  and
microbial  numbers.  Gamma  radiation
measurements were taken at the surface
of the  units  and at locations  inside and
outside of  the pumphouses to determine
whether exposure   presented  significant
health ant! safety problems. The effects
of diurnal variation  in  loading  and
backwashing were also evaluated.
   The GAC  used  in both systems was
Barneby  Cheney Type  1002.*  At
Amherst,  the system  consisted  of  one
91.4-cm-diameter fiberglass tank contain-
ing 0.85 m3 of GAC At Mont Vernon, the
system consisted  of  two contactors  in
series: a  76.2-cm-diameter  contactor
 * Mention of trade names or commercial products
  does  nol  constitute endorsement or recom-
  mendation tor use
(GAC #1) containing 0.57  m.3 of  GAC
followed by a 91.4-cm-diameter contactor
(GAC #2) containing 0.76 m.3 of GAC.
The system at  Mont  Vernon  was
designed with two units because of the
high influent  radon  activity.  Taps for
collecting  water  samples were installed
on the influent and effluent lines  and at
10 intervals in each GAC system.
   The GAC  systems  were  monitored
daily for 3 to  4 days and then every 2 to
5 days for approximately 1 month during
initial operation  (Phase  I).  Thereafter,
during  Phase  II,  they  were  monitored
weekly, biweekly, and then monthly. The
system at Amherst was monitored  for 122
days, and  the system  at Mont  Vernon
was monitored for 478 days.
   During the first few days of operation
at Mont Vernon,  all of the radon present
in the water was removed in GAC #1. The
radon removal front moved through the
units  over time, eventually  breaking
through  into the effluent  by Day  25. A
similar pattern was observed with the
Amherst system.
   The average  influent  radon activity
(210,491  ± 41,384 pCi/L) at Mont  Vernon
remained higher  than the design  influent
activity of  155,000  pCi/L. The effluent
quality during this period  varied  from
4,750 to  68,400 pCi/L (Figure  1).  In
addition, the GAC system at Mont Vernon
had  a higher average  water flowrate
x"=36 ±  12 m3/day) than it was designed
to handle (25 m3/day).  As a  result, the
overall radon loading  applied  to the
system  was  usually  higher  than
anticipated, which may have  accounted
for some of the increase in effluent radon
activity.  Radon  removal  followed an
exponential pattern as a function  of bed
volume  (Figure 2). Similar  results  were
obtained  at the Amherst site.  During the
final  3.5  months  of the study at  Mont
Vernon,  when a  new well was operating
(influent radon activity =  68,900 ± 1,400
pCi/L), the data showed a steep (almost
straight) pattern and an overall decrease
in removal efficiency  (e.g.,  Day  477).
Although  this  decrease  in  removal
efficiency  corresponded  to changes  in
other water  quality  parameters, the
effects of raw water quality  on  radon
adsorption   by  GAC  are  poorly
understood,  and predictions  about
removal are difficult to make.
   The  gamrna/beta  emissions  data
obtained at both  sites were in the 10° to
101 mR/hr  range at the units' surfaces.
This range is considerably  greater than
the background  values of 0.03 to 0.06
mR/hr  and  the   National  Council on
Radiation Protection guidelines of an 8-hr
maximum exposure  in  residences  of

-------
                                350.0
                                          50   700   750   200   250   300   350
                                                            Time (days)

                                                        a. GAC #1 Contactor
                                       400   450   500
                             O
                             d.
                                725.0
                                700.0
                                 75.0--
                                50.0- -
                                25.0--
                                    0    50    700   150  200   250   300   350  400   450   500
                                                            Time (days)

                                                        b. GAC #2 Contactor

                 Figure 1.  Phase II GAC System - Mont Vernon, NH. Radon activity through GAC #7(a) and #2(b) through
                          477 days operation. Note scale difference between a and b. (* New well began operating.)
0.058  mR/hr.  Though  exposures are
highly  dependent on  distance from the
source, they would need to be minimised
to meet accepted occupational safety and
health  standards,  perhaps by  using
shielding.
   Several  potential  problems  were
observed  with the  GAC,  including
accumulation of iron,  uranium-238 and
-235, radium-226,  and  lead-210,  and
release of bacteria.  The GAC provides a
good surface for bacterial attachment and
concentrates the nutrients  needed  by
microorganisms. At Mont Vernon, the
effluent contained as many 104 CFU/mL,
probably, in part, because of high influent
numbers; at Amherst the effluent ranged
from  200 to  400 CFU/mL. Coring data
indicated that  iron  precipitates were
retained in the GAC units. This occurred
even when there was little change in the
influent iron  concentration as compared
with that of the  effluent because  of the
large  volume of water passing through
the units over time.
   Uranium profiles  obtained from Mont
Vernon  during Phase II and  from core
samples indicated  that  uranium was
removed exponentially through the GAC
system. This  contrasted with the  lack of
uranium  removal  observed at  Amherst.
The discrepancies may be explained by
the difference in  pH of the raw waters at
the two sites  (Amherst =  8.03  ± 0.14;
Mont Vernon = 6.5 ± 0.2). The predom-
inant uranium species between pH 7 and
8  are  soluble anionic carbonate  com-
plexes in  natural waters,  whereas at pH
<6.8,  the  neutral  U02CO3  species
predominates. It is hypothesized that the
poorly  adsorbed anionic species predom-
inated  at Amherst,  whereas the favorably
adsorbed  neutral  species predominated
at  Mont Vernon.
   At both sites, radium-226  was accu-
mulated,  as  determined by coring. Since
radium adsorption  to GAC is  considered
unfavorable, it  is hypothesized  that
radium was removed by adsorption or ion
exchange  reactions occurring  with either
solid phases (e.g.,  Fe(OH3), Mg(CO3)) or

-------
                                                                          	• Day 71
                                                                         A	A Day 99
                                                                         A	A Day 119
                                                                        O	O Day 154
                                                                         D	D Day 245
                                                                         ~    ~ Day 477
                                           50.0
                                                      100.0
                                                                  150.0
                                                                              200.0
                                                                                         250.0
                                                   Radon Activity (x 10J pCi/L)
                       Figure 2.  Phase II GAC System - Mont Vernon, NH. Representative profiles  of radon
                                activity through the GAC system (" New well began operating.)
organic matter deposited  in the GAC.  A
comparison of the theoretical  lead-210
adsorbed (as a function of  radon removal)
and the lead-210 coring results indicated
that most, if not  all, of the  radon progeny
were  retained  by  the  GAC. Lead-210
accumulations on the  GAC at  the  sites
ranged from  105 to 106 pCi/kg,  with the
total beds containing between 107 to 108
pCi.  The GAC used at Mont  Vernon and
Amherst exceeded  the  State  of  New
Hampshire Radiological Health Program
de minimus regulations for uranium-238
(58,410 pCi/kg;  2.5  x  10-5  Ci/rn3)  and
radium-226 (44.39 pCi/kg;  1.4  x   108
Ci/m3). There  are no regulations in  New
Hampshire for lead-210. As a result,  all of
the GAC used in this study was  classified
as a low  level radioactive waste. It should
be  noted  that  the  State  of   New
Hampshire  has  stringent de  minimus
standards with respect to uranium  and
radium, and that in many other slates the
material  would  not be  considered a low
level  radioactive  waste. However,  other
states may regulate lead-210, which  was
present at much  higher activities.
   A  detailed  economic  evaluation,
including only  costs related to installation
and operation  of  the GAC  systems,  was
performed.  Direct capital  cosls  were
determined from expenditures made
during the project, and indirect capital
costs were  calculated based  on  fixed
percentages developed  by the  EPA
Office  of  Drinking Water.  Annual  costs
were developed  by adding  the amortized
(9%  interest over 20 yr) total  capital cost
to the annual operation and maintenance
cost. Production costs were calculated by
dividing the annual cost by  the  annual
design flow. All cost figures were updated
to second quarter  1989 dollars (ENR  =
426).  Costs  are presented  for  com-
parative purposes only, since actual sys-
tem costs will vary and be site specific.
   The  production  costs  for  the GAC
systems, including pretreatment  for iron
and manganese and disposal of the spent
GAC as a regulated low-level radioactive
waste,  were estimated to be $2.15/1,000
gal for Amherst  and $2.64/1,000 gal for
Mont Vernon.
Diffused Bubble Aeration
   A diffused  bubble  aeration  system,
consisting of three polyethylene tanks
aligned  in  series  (holding capacity  of
each  =  1022L),  was installed in  a small
community public water  supply in Derry,
NH.  An  air  blower with  a 2.66  m3/min
capacity that  forced  outdoor  air  into
diffusers provided aeration. The  diffusers
consisted of  1 90-cm-diameter coiled
plastic tubes with 0.038-cm holes drilled
in their underside (spacing between holes
=  0.5 to 1.6  cnrv). The diffusers were
located  79 cm below the water's surface
and  36  cm  above  the  bottom  of each
tank. The  radon  stripped from the water
was  vented outside the  building housing
the units.
   The  system  allowed  removal
efficiencies to be compared over a wide
range of influent  radon activities  (Tank 1:
60,843 to 86,355  pCi/L radon;  Tank  2:
10,096  to  80,271  pCi/L; and Tank 3:
1,767 to 74,112 pCi/L). Two flow ranges
were obtained by manually operating one
(low flow  =  0.047  ± 0.00053 m3  H2O/
min) or two wells (high flow  =  0.10 ±
0.0019  m3/min).  Radon  activities aver-
aged  65,487 ±  5,657 pCi/L during high
water flow and  78,385   ±  6,120 pCi/L
during low  water  flow.  The  two water
flowrates and the two  radon  activities
resulted  in  applied radon loading  rates
averaging 6,819 ± 548 nCi (103 pCi) per
minute (nCi/min) for high water flow and
3,639 ±  295 nCi/min for low water flow.
A:W ratios of 2:1, 3:1,  5:1, 7:1,  10.5:1,
15:1, and 20:1 were tested for  both water
flowrates. The tanks were drained at the
end  of  each run and  refilled with  raw
water immediately before the  start of  a
new run.
   Figure  3 shows the  overall  percent
removals of  radon  loading  versus  A:W
ratios  for  the  high  and  low  water
flowrates. These data were obtained after
steady state conditions were achieved in
the diffused bubble system.  As the A:W
ratio increased  from 2:1  to 5:1 for  each
flowrate,  there  was a sharp increase in
radon removal. Above 5:1, however, there
was much less improvement in efficiency
with large increases in A:W ratio.
   When operating  at A:W ratios of 5:1
and greater (at both high and  low water
flowrates),  the  overall  radon removal
efficiency  ranged from  90.0%  to
>99.6%;  the greatest  efficiency  was
obtained at A:W ratios of 15:1  and 20:1.
At A:W ratios of 10.5:1 and greater for the
low flowrate and 15:1 and greater for the

-------
                                 700
                                  90--
                                  80-.-
                                  70--
                                  60--
                                  50--
                                  40
                     -• Low Water Flowrate
                     •A High Water Flowrate
                                                   >             W            15
                                                           AirWater (X:1)
                                                   Removal as Function of A:W Ratio
                                                 20

Removec
c:
•§
03
o:
a5
p
Q.

90-
80-

70-

60-

50-
40-
0.
• rx^~
/ / • 	 • Low Water Flowrate
. I r A 	 A High Water Flowrate
/ A
- /
/
*


00 0.50 1.00 1.50 2.00 2.1
                                                        Air Flowrate (m /mm)
                                                   Removal as Function of Air Flowrate
                        Figure 3.  Diffused bubble aeration - Derry, NH. Percent radon removal as a function of
                                 (a) A:W ratios and (b) air flowrate for the low and high water flowrate.
high flowrate, there was  no significant
difference in removal efficiency (a = 0.05
and 0.01, Analysis of Variance (ANOVA)).
Hence,  for the  diffused bubble system
tested  at  the  given conditions  of  radon
loading, the  lowest  A:W  ratios  to  yield
(statistically significant) maximum  radon
removal for low  and high water flowrates
were in the range 7:1 to 10.5:1 and  10.5:1
to 15:1, respectively. From  the  practical
viewpoint of designing a diffused bubble
system, the  A:W ratio should  also  be
based on  the  most cost effective blower
size and mode of operation.
   An  evaluation  of  individual  tank
performance indicated that radon removal
in the  diffused  bubble  system was a
function of mass transfer. As the  radon
activity became  progressively lower
through the series of tanks,  the  driving
force  (Radon  Activitywater  -  Radon
ActivityAjr)  decreased,  limiting removal.
Though mass transfer may make it more
difficult to achieve low effluent activities,
the  diffused  bubble system  tested
produced  water  with  radon  activities  of
1,849  to   280 pCi/L  for  A:W  ratios
> 10.5:1. These data suggest that it may
be possible to meet an MCL of 200 pCi/L,
if air flowrate is  high, or  there is a long
contact time, or both.
   Stack emissions were monitored  to
determine  if the off-gas  radon activities
from the diffused  bubble system  could
affect the  air quality in the  surrounding
environment. For the system tested, the
off-gas activities (3,361  to  18,356 pCi/L)
would need to be diluted 104 to 105 times
to be similar to  radon activities found  in
the ambient air  at the site (0.1  to 0.15
pCi/L).
   An economic evaluation, similar to that
done  for  the GAG systems, was per-
formed for the  diffused bubble system.
The total  production  cost,  including pre-
treatment  for iron and  manganese  and
assuming  no required treatment of the
off-gas, was estimated to  be $2.14/1000
gal.


Packed  Tower Aeration
   The packed  tower  aeration  system
was installed at  the  same  mobile home

-------
  park in Mont Vernon, NH,  used in the
  GAG study. The system consisted  of  a
  5.49-m-tall. 0.30-m-diameter stainless
  steel tower containing randomly packed
  plastic  media. Raw water was pumped to
  the top of the tower and distributed  by  a
  nozzle  located 15.2 cm above the top of
  the media. Air entered the tower 0.15 m
  below the  media.
     The major  focus of the  study was  a
  series  of separate 3-hr runs  designed to
  determine the tower's radon  removal
  efficiency  for a variety of operating  con-
  ditions. Parameters varied included pack-
  in9 type,  packing  height,  liquid  loading
  rate, and  volumetric A:W  flow ratios.
  Though fluctuations in water flowrate and
  radon activity at the site during the study
  prevented  comparisons of the effects of
  A:W ratio  and packing types, there  was
  relatively  little difference in  the  overall
  percent radon  removal observed  (92.7%
  to  99.8%)  among the conditions tested.
  This  was  surprising  considering  the
  variation in water flowrate (0.18  to  2.6
  m3/hr),  influent radon  activity  (115,225 to
  278,488 pCi/L), and packing  type.  The
  resilience  of the  tower  system is
  encouraging, considering that  many small
  communities may experience variations
  in   water  flowrate  and  radon activity
  similar  to  those  observed  at the Mont
  Vernon  site.  It is hypothesized that  the
  consistently higher removals occurred
  because radon is a highly  volatile  gas
 and because  the packing  height used
 (~3.7 m)  was great  enough  to  com-
 pensate for large variations in loading.
    Most of  the  radon removal  (Frigure 4)
 occurred in the top 0.3 m of the tower,
 probably because of end effects of free
 fall, liquid  distribution, and  turbulence.
 The data  indicate  that  mass transfer
 limitations  may be  a major  factor  in
 designing  towers to achieve  very   low
 radon activities.  Therefore, to meet  an
 MCL of  200 pCi/L  with water  supplies
 containing   moderate  to high  radon
 activities,  towers may  be impractical
 because of the extremely  large packing
 heights required.
   As observed with the diffused bubble
 system,  radon  activities  in  the  stack
 emissions from the tower were extremely
 high and required  dilutions of  104 to 105
 to approach the  ambient air activities (0 1
 to 0.15 pCi/L) at the site.
  An economic evaluation, similar to that
 done  for   the  GAG   systems,  was
 performed for the packed tower. The total
 production  cost,  including  pretreatment
 for iron and  manganese  and assuming  no
 required  off-gas treatment, is  estimated
to be $2.10/1,000 gal.
  Liquid Scintillation Technique

     The  effects of sampling technique
  (direct collection using a syringe versus
  filling a VGA bottle),  storage (up to 21
  days), and  choice of scintillation cocktail
  (toluene- and mineral-oil-based cocktails
  and Opti-Fluor 0) on the liquid scintillation
  (LSC) analytical  technique  for radon in
  water were  evaluated.  Other experiments
  were conducted to determine the sources
  of variability in  the  method  (field,
  preparation,  instrument)  and the validity
  of the extraction via  shaking procedure
  currently defined by EPA.
     Numerically,  the  direct  syringe
  sampling technique always  yielded  the
  highest radon activities, whereas the VGA
  bottle filled with free falling water yielded
  the lowest. This is not surprising because
  less sample handling before injection  into
  the scintillation vial and  less agitation
  during sampling should result in less loss
  of  radon. Though  the VGA  collection
  techniques involve extra  handling of  the
  sample,  they can,  however, produce
  results statistically similar to the  syringe
  method,  especially  if a  universal  hose
  connector is  used.
    The data  from the storage  experiment
  indicated  that loss  of  radon  from  VGA
  bottles could  be  a factor  in  some
  situations, but the  loss  resulting  from
  radioactive   decay has  the  greatest
  potential effect on storage time. Within 4
 days, 50% of the radon originally present
 in  a  sample  will  be  lost because  of
 decay. For  samples containing  high
 levels of radon, permissible storage times
 could be  substantial provided that  the
 amount remaining  at the time of analysis
 is above the  practical quantification level.
 (For example, a  sample  containing
 approximately 7,670 pCi/L could be held
 up to 10 days, even with a 20% loss due
 to leakage, and still contain 1,000  pCi/L).
 However,  the amount of radon in a
 sample  is often  not  known, so the
 maximum  storage  time  sufficient  to
 obtain a  valid measurement must  be
 based on the  MCL.
   A hierarchical  experiment was
 conducted  using  the  direct syringe
 sampling technique and collection in VGA
 bottles with  subsequent  laboratory
 analysis. The  total variability  associated
 with the  two  methods  was  not
 significantly different (a  <  0.10, F test).
 The total  variation of 4%  to  6% as  a
 result of sample handling and  instrument
 variation  was   not high considering  the
 volatile  nature of  radon.  Most of  the
variability in the direct syringe  technique
(92.1% of total variance) was  due to  a
combination  of  sample  handling and
  instrumentation,  whereas  the  variations
  due to sample handling and instrumenta-
  tion for the VGA bottle technique were
  55.5% and 44.5%, respectively.
     An  ANOVA  showed that the  mean
  count rates for the mineral oil-based and
  Opti-Fluor 0  cocktails  were not signifi-
  cantly  different (a <  0.10); however,  the
  mean  count  rate of  the  toluene-based
  cocktail was  significantly  less  than both
  of these  (a  <  0.05).  In  addition, the
  percent relative  standard  deviation  (%
  RSD)  associated with the  toluene-based
  cocktail (3.22%)  was greater than  those
  of mineral  oil-based  (0.90%) and  Opti-
  Fluor 0 (1.34%) cocktails. The choice of a
  scintillation  cocktail may  also  be
  influenced by other factors such as cost,
  disposal, and mailing restrictions.
    The EPA  procedure for  analyzing
  radon   in water  requires  that the vial
  containing  the  sample be  shaken  to
  speed  the extraction  of radon  from the
  water   into the  cocktail  while  other
  radionuclides remain in  the  aqueous
  phase.  Both  shaken and nonshaken
  radium-226 standards  were  analyzed in a
  series of experiments. The data indicated
  that  transfer of the radon to the cocktail is
  continuous  and  does  not  require the
  extraction via shaking procedure, with the
  possible exception of samples containing
  very low  radon activities.  Use  of
  efficiency factors, which account for time
  between extraction and counting,  could
  lead to an underestimation of  actual
 radon activities, especially for samples at
 or  near the MCL.

 Conclusions and
 Recommendations
    When designing a treatment system to
 remove  radon  from a small community
 water  supply, good  data  on  water
 flowrates and influent  radon  activities at
 the site are essential.  These  data are
 major inputs into the design models for
 the aeration  and GAG systems. As
 observed  in  this study,  for  small
 community supplies, variations in flowrate
 and influent  activity may be substantial
 and  will, if  underestimated,  lead to
 inadequate  system design and  effluent
 radon activities that exceed  the design
 goal.  Variations in water quality  at each
 site may require  that  pilot-scale testing
 be  performed to  determine  the appro-
 priate design and pre-  or post-treatment
 requirements.
   For GAG  systems,  the steady  state
 adsorption-decay  constant, a  critical
 component of the  design model, varied
 over time and was site specific. The GAG
 systems  tested  in  this study had average
effluent  radon  activities of  12,000 to

-------

10 —
20 —

"§ 30--
i.
o
| 40- -i
a:
c
o 50--
to
CC
- 60--
c

>



!

t

^jjj§ 0-0.3 m
ESS 0.3-7.8 m
\ \ 1.8-3.7 m









C
c
c
f

I
-* .—^—i 	 1 	





















                              0.0
                                       w.o
                 30.0      40.0

           Air Flowrate (m3/hr)
                                                                               50.0
                                                                                         60.0
                       Figure 4. Packed tower aeration  - Mont Vernon, NH. Percent radon removal as a
                                function of air flowrate within the tower for one packing type (mini rings).
24,000 pC\/L.  The  GAG units accumu-
lated  iron,  manganese  and  particulates
(turbidity),  and  there were  significant
numbers of  bacteria  growing  on the
carbon.  As a result, GAG systems  may
require periodic  backwashing to  prevent
significant  headless  development.  In
some cases,  pretreatment  is  recom-
mended  to  decrease the backwashing
frequency. Gamma/beta emissions meas-
ured at the surface of the GAG units were
substantially greater than background
measurements possibly requiring shield-
ing to  lower them  to acceptable levels.
The data indicated  that retention  of
uranium-238 and -235,  radium-226, and
lead-210, which  appear to be related  to
water quality (e.g., pH  and  alkalinity),
may cause the GAG to be classified as a
low level radioactive waste.
   The diffused  bubble aeration  system
tested at A:W ratios > 5:1 yielded overall
radon removal efficiencies from 90.0%  to
>99.6%. The radon removal  efficiencies
for the  packed  tower aeration  system
ranged from 92.5% to 99.8%, in  spite  of
variations in  water flowrate, influent radon
activity,  and packing type. Extrapolations
of performance data obtained at  one site
with either aeration system  should not,
however, be made to  systems with other
configurations, process equipment, or low
influent activities, or to those  required  to
meet a  more  stringent  MCL.  Both
aeration  systems  had  off-gas radon
activities that were  104 to  105 times
higher than those of the ambient air, and
their effect on  the  environment  would
need  to be  considered. As with  all
aeration systems,  precipitation of  iron
and manganese can occur and result in
operational problems. Therefore,  raw
water  quality should  be monitored to
determine whether  pretreatment is
required.
   Several  recommendations can  be
made concerning the liquid  scintillation
analytical  iechnique for  radon in  water,
During  sample  collection, the universal
hose connector should  be  used  to fill
VGA bottles. Maximum storage times for
samples collected in VOA bottles should
be  established based  on   the  MCL,
practical quantification lever,  radioactive
decay,  and leakage. Opti-Fluor 0 yielded
the best  results  of the  scintillation
cocktails tested  with  respect to  count
rates,  variability, and  cost and  is,
therefore,  lecommended. The extraction
procedure should  not be  used to
calculate the  efficiency  factor; however,
samples should  be shaken,  especially
those with low activities, to ensure rapid
transfer ot radon to the  scintillation
cocktail.
   The  full  report  was submitted in
fulfillment of Cooperative Agreement CR-
812602 by  the University of  New
Hampshire  Environmental   Research
Group under the sponsorship of the U.S.
Environmental Protection Agency.

-------
    A/. £ Kinner, J.P.  Matey, Jr., J. A. Clement. P. A. Quern and G S
                          .
    K/m R Fox /s Me EPX\ Pro/ecf Officer (see below)
    The comP'^e report, entitled "Radon Removal Techniques for Small Community
          Pubhc  Water Suppl.es,"  (Order No.  PB  90-257 809/AS;  Cost: $3100
          sub/ect to change) will be available only from:                   w-w,
             National Technical Information Service
             5285 Port Royal Road
             Springfield, VA 22161
             Telephone: 703-487-4650
    The EPA Project Officer can be contacted at:
             Risk Reduction Engineering  Laboratory
             U.S. Environmental Protection Agency
             Cincinnati, OH 45268
United States
Environmental Protection
Agency
Official Business
Penalty for Private Use $300
Center for Environmental Research
Information
Cincinnati OH 45268
      BULK RATE
POSTAGE & FEES PAID
         EPA
   PERMIT No. G-35
EPA/600/S2-90/036

-------