AU9817365

Determination of 222Rn in Groundwater

- Recent Applications for the Investigation of River Bank Infiltration
K FREYER, H C TREUTLER
UFZ - Centre for Environmental Research Leipzig-Halle Ltd.
Permoserstr. 15, D-04318 Leipzig, Germany

J DEHNERT, W NESTLER
Hochschule fur Technik und Wirtschaft Dresden
P.O.B. 120701, D-01008 Dresden, Germany
/222 T
SUMMARY. Radon-222 ( Rn) can be used as a tracer to investigate exchange processes between
surface water and groundwater. A reliable method for the representative collection of groundwater
samples and the determination of 222Rn in water samples using liquid scintillation spectrometry was
developed for this purpose. The total error of the method is < ± 10 %. Groundwater measuring
profiles near a waterworks extracting bank-filtered water near Torgau on the Elbe were used to study
processes of exfiltration and infiltration. The passage of a flood wave across the bank was observed.
The infiltration velocity of river water was determined.

1. INTRODUCTION quality must also be taken into account. A

The noble gas radon-222 (hereafter simply
referred to as radon) occurs naturally in all
groundwater. With a half-life of 3.8 days, it
makes an ideal tracer for tackling a whole
range of problems in environmental research
and hydrology. One important issue is
describing the exchange processes taking place
between surface waters and aquifers. For
example, studies of how rivers are fed by
groundwater have been performed by Ellins et
al (1), Genereux and Hemond (2), Yoneda et al
(3), and Hamada and Komae (4).
In 1989 Hoehn and von Gunten (5) proposed
using the increase in radon activity
concentration of infiltrating river water as it
passes through the river bank to determine
flow time. Similar studies have also been
described by authors such as Bertin and Bourg
(6), and Wilme et al (7).
The speed of infiltration is required in order to
determine infiltration rates - which in turn are
needed to size drinking water catchment
facilities with artificial groundwater
recharging and to study the alluvial meadow
ecosystem. When dimensioning well intakes
for the extraction of bank-filtered water, the
infiltration speed is important because water
significant share of the pollutant degradation
process takes place in the silting zone - a
biologically highly active area which is just a
few centimetres deep, and whose efficiency
also depends on the infiltration speed.
In all cases, the radon activity concentration in
the water must be determined using a suitable
method. Radon's low activity concentrations
in Pleistocene aquifers require high measuring
accuracy. Owing to radon's high mobility,
sampling itself must also be carried out
carefully.
This paper describes a reliable sampling and
measuring technique which uses liquid
scintillation spectrometry for the fast,
reproducible determination of the radon
activity concentration in groundwater. The
method also includes using radon to optimise
the pumping times of groundwater gauging
stations as described by Dehnert et al (8). The
technique developed was used to study the
infiltration of a river into the aquifer in the
catchment area of a water works extracting
bank-filtered water.
 li
2. SAMPLING AND RADON-
MEASURING TECHNOLOGY
2.1 Sampling
Sampling techniques depend on the kind and
origin of the water sample. The representative
collection of a groundwater sample in the field
and its transportation to the laboratory are
particularly difficult as radon escapes very
easily. Suitable groundwater observation wells
must be used. Although the best approach is to
use "nests of observation wells", "bundles of
observation wells" represent a good
compromise between sample quality and
construction costs.
Different water sampling devices (submersible
pump, membrane pump, bailer) were
compared to check their applicability.
Assessed in terms of radon concentration, the
fluctuations registered were below 1%.
Furthermore it was found that a pumping rate
of a submersible pump in the range between
0.15 and 1.92 m3/h does not exercise any
significant impact.
The sample vessel and the filling technique
must take into account radon's tendency to
escape. Consequently, the sampling device
chosen was a glass mensure calibrated to one
litre featuring a narrow neck sealed with a
polyethylene stopper and a teflon gasket. First
of all, 20 ml toluene scintillation cocktail is
cpm
Rn222
(5.490 MeV)
•r-'T'rii'lT"11
1
10
Figure 1: Alpha spectrum of 222Rn and its daughter nuclides as shown by TRI-CARB
SPECTRAGRAPH software
poured into the mensure. The water is then
poured in through a glass tube specially made
for radon measurement by underlying the
scintillation cocktail without air contact.
Sampling must be carried out free of
turbulence and air bubbles. The filling
technique used was tried out in several series
of measurements. Radon loss after sampling
and transport was below 2%.
2.2 Radon measurement
In view of the advanced state of liquid
scintillation spectrometry, this measuring
technique was selected as being the most
suitable of the wide range of different options
for determining radon activity as described for
instance by Volkle and Borchardt (9).
The method detailed by Horiuchi and
Murakami (10) was adapted and modified to
determine activity concentration in water
samples. The radon dissolved in the water is
extracted by a toluene-based liquid scintillator
cocktail. After extraction the toluene phase is
separated from the water and the 222Rn activity
of this 20 ml cocktail is determined by liquid
scintillation spectrometry using a TRI-CARB
2550 TR/AB (Packard), which works in low-
level mode and has a special hardware and
software option for oc/p separation. The
optimum measuring time is 60 minutes. The
measuring error is < 3% for counting rates
between 40 and 8,000 cpm. Using comparison
measurement with gamma spectrometry, a
calibration factor of (8.3 ± 0.7) 10"3 Bq/cpm
was determined.
In order to check the correctness of the
I
,Po218
((6.003 MeV)
PO214
(7.687 MeV)
"I"1'!
100 1000 keV
measurements, the TRI-CARB 2550 TR/AB
also allows the alpha spectra measured to be
evaluated by means of SPECTRAGRAPH
software (Fig. 1). oc/p separation can be used
to calculate the oc/p ratio - which in the case of

222
Rn ought to be 60:40. The radon activity at
the time of sampling is calculated by
regression following multiple measurements
taken over a period of 6 days. The detection
limit, calculated as 3 standard deviations often
background measurements, is 0.05 Bq/1. The
total error of the measuring method is < ±10%.
3. INFILTRATION STUDIES
3.1 Method
The technique put forward by Hoehn and von
Gunten (5) is based on the fact that river water
only contains a small amount of radon because
a turbulent current causes the mobile noble gas
to escape. Radon is constantly emanated by the
aquifer's grain structure in different
concentrations (due to geogenous factors),
entering the liquid phase of the water either via
reaction effects during alpha decay or
diffusion from the grain surface (Wilkening
(11)). As infiltrating, low-radon river water
flows through the aquifer, it continuously
absorbs radon, so that an equilibrium between
radon uptake and radioactive decay is
established. If the equation for the increase of
activity of a daughter nuclide until the state of
equilibrium is reached is resolved in terms of
time, the retention time of the infiltrate in the
aquifer can be determined. Radon's half-life of
3.8 d enables this for a time window of up to
15 days after infiltration, but also requires
ascertaining the radon activity concentration at
the gauging stations once a state of equilibrium
has been reached.
3.2 Study area
Work was carried out at a battery of wells at
East Torgau Waterworks Ltd., where bank-
filtered water is extracted as part of the long-
distance water supply system covering the
Elbe water meadows and the eastern Harz. The
Pleistocene aquifer consists of fluviatile-
glacifluviatile sediments and Holocene Elbe
gravel, and is up to 55 m deep. The
meandering Elbe is in direct hydraulic contact
with the aquifer. In 1991/1992 two ground-
water observation measuring profiles were set
up for research purposes. The profiles, each of
which is 2 km long, cross the Elbe and run
along the central well of a battery of wells up
to the hinterland. They enable depth-dependent
sampling in up to 5 horizons using ground-
water gauging stations and sunken membrane
3
12
6*1*7
pumps. Moreover, three silting gauging
stations enable sampling below the submerged
river bed.
3.3 Optimising the pumping times of
groundwater gauging stations
When measuring radon activity concentration
in an aquifer, the optimum pumping time of
the groundwater gauging station before
sampling is of key importance. Owing to its
short half-life of 3.8 days, the radon in the still
water of the gauging station decays over 26
days to 1% of its original level. Incorrect
pumping times lead to much more inaccurate
results with radon than with other parameters.
To eliminate this problem, a new and generally
applicable technique for determining the
optimum pumping times of groundwater
gauging stations based on reaching constant
radon activity concentration during pumping
was developed and used (Dehnert et al (12)).
3.4 Results and discussion
The groundwater of the individual layers of the
aquifer have characteristic radon activity
concentrations of between 5 and 70 Bq/1. The
petrographic structure of the aquifer and radon
activity concentration in the groundwater
largely correlate with each other. Although
according to Hoehn and von Gunten (5), radon
emanation needs to be macroscopically
constant in order to determine durations of
stay, this is not the case underground.
Therefore the technique was modified such
that the radon equilibrium concentration was
determined for each gauging station by
sampling the entire measuring profile during
the slow exfiltration of the groundwater into
the Elbe. This took place during the passage of
a flood wave (Fig. 2). Further the passage of a
flood wave across the bank could be observed
by means of radon activity concentration.
Fig. 3 depicts the situation at the crest of the
flood. The sharp drop in radon activity
concentration at all gauging stations indicates
water infiltrated from the Elbe. Areas with
activity concentrations < 8 Bq/1 are shaded by
way of illustration.
The durations of stay and the velocities of
infiltration and flow can be calculated from the
change in radon activity concentrations at the
individual gauging stations over time
compared to the equilibrium concentration.
For example, an infiltrate retention time in the
 masl
85
Mean water
13.6.96

Water evel: 77.80 m asl

Coarse sand, fine gravel
13 9.0 8-1 79
+
75 11
+
13 16 14 14 16 23 15 +
14 +
+ + + + +
+
. Medium sand, coarse sand < 8 Bq/I
70
10 20 30 40 50 m

Figure 2: Radon activity concentrations in Bq/1 in the measuring profile during exfiltration

masl
85
Flood
12.7.96
Bottom clay •

— .—- — -. Elbe
T
- • —

80
Water level: 79.92 m asl
• — — .

47-- 0.2
Coarse sand, fine gravel •

. A Q -~.
+• ^ 0
8.9 3.3 5.1 1. 9 Ds —
7 8 5.6 4 .9 + 44 Hh
+ + ¥ + H
75 62
• 9.5 15 11 80 86 19 63 Hh

. Medium sand, coarse sand < 8 Bq/I

70
0 10 20 30 40 50 m

Figure 3: Radon activity concentration in Bq/1 in the measuring profile during infiltration

aquifer during the passage of a flood wave of speed of 1.7 m/d. This value corresponds
0.5 d was determined for the silting gauging closely to estimates of quality parameters and
stations under the Elbe bed at an infiltration geohydraulic calculations.
 IK-
4. SUMMARY AND CONCLUSIONS
If 222Rn is to be used as a tracer in water, a
reliable technique for measuring radon must be
used in view of its frequently low activity
concentrations. The method chosen must also
include a representative sampling technique,
especially in the case of groundwater. The
accuracy of the method described here with
radon determination by liquid scintillation
spectrometry has been proved and found to
have a total error of < ±10%. The precision of
radon measurement is further improved by
spectral analysis and a/(3 separation.
Extensive correlation was detected between
the radon activity concentration of the ground-
water and the petrographic structure of the
aquifer at the Torgau aquifer studied here.
Nevertheless, the existence of a radon activity
concentration in the groundwater sufficiently
representative of the flow path, which is
required to determine the retention times of
bank-filtered water, can probably not be
counted on. The representative radon activity
concentration in the equilibrium needed to
ascertain the retention time of the infiltrate
was determined separately for each gauging
station by sampling the entire infiltration
measuring profile during a phase of slow
groundwater flow. The passage of a flood
wave through the bank zone of an aquifer was
observed. The infiltration speed of the Elbe
could at first only be calculated for a silting
gauging station beneath the bed of the Elbe.
The results indicate that naturally occurring
radon can provide under critical consideration
of the boundary conditions a suitable para-
meter for describing infiltration by a river.
ACKNOWLEDGEMENTS
This work was carried out as part of the joint
research project "Water extraction in valley
aquifers in the Elbe catchment area"
(02WT9454) funded by the Federal Ministry
of Education, Science, Research and
Technology. The authors would like to thank
Mr Oese for his valuable assistance during
sampling. We are indebted to Ms Scholz, Ms
Petzold and Ms Bachmann for performing
LSC measurement at the UFZ. Thanks also go
to Fernwasserversorgung Elbaue-Ostharz Ltd.
for their kind assistance, especially Mr
Nitzsche (head of the Technical Office) and
Mr Heilmann (the manager of East Torgau
Waterworks).
REFERENCES
(l)Ellins K K, Roman-Mas A, Lee A: Using
222
Rn to Examine Groundwater/Surface
Discharge Interaction in the Rio Grande De
Manati, Puerto Rico. (1990), J.Hvdrology
115:319-341.
(2)Genereux D P, Hemond H F: Naturally
Occuring Radon 222 as a Tracer for
Streamflow Generation: Steady State
Methodology and Field Example. (1990),
Water Resour.Res. 26: 3065-3075.
(3)Yoneda M, Inoue Y, Takine N: Location of
Groundwater Seepage Points into a River
by Measurement of Rn Concentration in
Water Using Activated Charcoal Passive
Collectors. (1991), J.Hvdrology 124: 307-
316.
(4)Hamada H, Komae T: Analysis of
Exchange Between River Water and
Groundwater Using Radon-222. (1994)
XXV Congress of the International
Association of Hydrogeologists, 21-25
Nov., Adelaide, South Australia, In: Water
Down Under 94, Preprints of Papers 1: 115-
120, The Institution of Engineers; Barton,
Australia.
(5)Hoehn E, Von Gunten H R: Radon in
groundwater: A tool to assess infiltration
from surface waters to aquifers. (1989),
Water Resour.Res. 25: 1795-1803.
(6)Bertin C, Bourg A C M : Radon-222 and
Chloride as Natural Tracers of the
Infiltration of River Water into Alluvial
Aquifer in Which There Is Significant
River/Groundwater Mixing. (1994),
Environ.Sci.Technol. 28:794-798.
(7)Willme U, Schulte-Ebbert U, Hoehn E;
Bestimmung von Grundwasserverweil-
zeiten mit geogenem Rn bei kiinstlicher
Grundwasseranreicherung und Uferfiltra-
tion in einer Trinkwassergewinnungsan-
lage. (1995), Wasser + Abwasser 136: 234-
241.
(8)Dehnert J, Nestler W, Freyer K, Treutler H
C, Neitzel P, Walther W: Radon-222 - ein
neuer Leitkennwert zur Bestimmung opti-
maler Abpumpzeiten von Grundwasser-
 meBstellen. (1997), Grundwasser 2(1): 25-
33.
(9)V6lkle H, Borchardt D (ed.): Messung von
Radon und Radon-Folgeprodukten, (1991),
Publikationsreihe Fortschritte im Strahlen-
schutz. Verlag TUV Rheinland GmbH,
Koln.
(10) Horiuchi K, Murakami Y: A new
procedure for the determination of radium
in water by extraction of radon and
application of integral counting with a
liquid scintillation counter. (1981),
Int.J.Apol.Radiat.Isot. 32: 291-294.
(11) Wilkening H: Radon in the
Environment.- Studies in Environmental
Science 40, (1990), 137 p., Amsterdam-
Oxford-NewYork-Tokyo: Elsevier.
(12)Dehnert J, Nestler W, Freyer K, Treutler H
C, Neitzel P, Walther W: Bestimmung der
notwendigen Abpumpzeiten von Grund-
wasserbeobachtungsrohren mit Hilfe der
natiirlichen Radonaktivitatskonzentration.
(1996), in: Merkel B, Dietrich P G,
Struckmeier W & Lohnert E P (ed.):
Grundwasser und Rohstoffgewinnung.- 551
S.; Koln,GeoCongress 2, Verlag Sven von
Loga, 40-45.