How to measure radon in water


GRAB SAMPLES
Liquid scintillation counting (LSC)
Radon is considerably more soluble in many organic solvents than in water. It thus can be extracted with a high efficiency from the aqueous phase to the organic phase. If the organic phase contains a scintillator, one can measure radon in water with any scintillation counter. If there is already a liquid scintillation counter with a sample changer at hand, e.g. for tritium determination, LSC is the easiest way to measure large batches. With a 10 ml sample one typically gets a detection limit below 1 Bq/l ( units )after 1h counting time. However, these instruments are very expensive and it would not be worth to buy one just for radon measurements. There are affordable portable instruments now, without sample changer, that achieve detection limits of around 10 Bq/l ( units ) after a counting time of 10 min. Depending on the scintillator used one can measure either radon or radon daughters. With scintillators miscible with water, one can start measurements immediately after preparation. They form stable emulsions and thus radon daughters are counted even if they are not in the organic phase. If the scintillator is immiscible with water one has to wait for the daughters to build up in the organic phase. An advantage in this case is that one really measures radon.

Radon daughter adsorption on glass fiber filters
Some glass fiber filters show a nearly quantitative radon daughter adsorption [ 7 ]. The filters can then be measured with a beta counter. The processes leading to this strong adsorption are not well understood, but it works.


Radon daughter adsorption on ZnS layers
A detector originally developed to measure artificial alpha activity continuously in water [ 8 ] measures radon daughters as well. A sample volume of some 100 ml is limited on one side by a light guide covered with ZnS. Scintillations in the ZnS are detected by a large surface photomultiplier. It seems that radon daughter products produced in some 10 ml close to the ZnS layer are adsorbed on this layer or on its protective cover.

Radon daughter adsorption on MnO 2 layers
Polonium adsorbs quickly onto MnO 2 layers. A substrate, e.g. a polyamide covered with MnO 2 is exposed to the water sample. After quick drying, this disk can be measured by alpha spectrometry.
An experimental setup where polonium adsorption on MnO 2 is used to measure Rn continuously is shown below .

Radon daughters gamma measurements
A fast and easy method, but one does not always know where daughter products are located. They may be adsorbed to the walls or may co-precipitate with small particles (hydroxides, carbonate particles, organic material).
click to enlarge

Outgassing
A very old, but still widely used method is to bubble air through the water sample and to measure radon in the air circuit [ 9 ]. Advantages of this method are that one does not have to wait for radon daughters built up in the water and that there is a large number of radon-in-air monitors available. A detailed discussion of the method can be found in [ 10 ]
click to enlarge

Instead of a bubbler one can use a membrane tube, a tube that is watertight but lets pass gases easily . When using "Accurel" [ 11 ] tubes one gets an equilibrium between radon in the water phase and radon in the gas phase within minutes. Diffusion through silicone tubes is slower, but they are far cheaper than "Accurel" tubes. Both types of tubes are "watertight", but they let water vapour pass through. In a closed air circuit humidity may soon reach saturation. If the radon monitor used does not support high humidity (most don't), or if there is a risk of condensation at some colder part in the air circuit, one needs a desiccant. Silica gel and most molecular sieves cannot be used for they adsorb radon. We recommend using dehydrated Ca-Sulfate (Anhydrite).
click to enlarge

How the methods compare
For radioprotection applications, a detection limit of 10 Bq/l ( units ) is all you need. Outgassing in water treatment plants and in spas may lead to indoor air radon problems above some 10 Bq/l in the water. For hydrogeologic studies however it is worth having a lower detection limit, around 1 Bq/l ( units ). When deciding which method to use, an important point is how long it takes to reach a certain precision. Assuming that a signal is significantly different from the background if it is 4 sigma above the background and if one wants to have a statistical 1 sigma uncertainty that is less than 20 % one gets the curves in the figure on the left. This figure shows that all methods described allow measuring radon concentration above 5 Bq/l ( units ) with a precision of 20% within 10 minutes. That is largely sufficient for radioprotection. Thus, one can concentrate on other parameters. An important one is whether there is time to wait for radon daughter buildup in the water sample.
click to enlarge
click to enlarge

The top figure on the left shows some Rn data for Swiss springs and groundwaters, grouped by cantons. The data look as if there were regional, possibly geological differences. However, as can be seen from the second figure, temporal variations can be as large as regional variations. As long as sampling frequency is not adapted to the dynamics of the corresponding aquifer any correlation between Rn concentration and geology may just be an artifact. Unfortunately one rarely knows how fast an aquifer will react to changing environmental conditions, e.g. to heavy precipitation. Extreme cases are karst springs. Discharges can increase by two orders of magnitude within hours after a storm. Even springs emerging from an aquifer containing very old water may react quickly to precipitation because of of a change in hydraulic pressures. One needs continuous montoring not to miss anything.

CONTINUOUS MEASUREMENTS
click to enlarge

We thus have built continuous radon-in-water monitors, having a temporal resolution on the order of an hour. There are several continuous monitor generations now in use but most of our instruments are based on the same principle : radon is measured in a closed air circuit coupled to the water. The coupling first was done by bubbling air through the water [ 12 ]. Later we replaced the bubblers with diffusion tubes [ 10 ]. Using 4 m of an "Accurel" tube and a Lucas cell as a detector (being notoriously slow but cheap) one gets a temporal resolution slightly below 1h at 1 Bq/l ( units ). By combining more expensive 218 Po detectors having an internal volume of less than 200 ml with longer tubes (10m), it should be possible to get temporal resolutions below 10 min. To get a good temporal resolution it is essential to have a high water flow. If not, there is either a depleted or enriched zone around the diffusion tube, not having the same radon concentration as the water flowing in. A radon monitor with a completely different design is shown just below.
click to enlarge
click to enlarge
click to enlarge

This experimental setup shows how adsorption on an MnO 2 thin film can be used to measure radon in water continuously. Originally it has been developed to study radium adsorption dynamics. It turned out that radon daughters are adsorbed as well or even better than radium. Radon daughter adsorption is measured on-line by placing an alpha detector behind a thin (12 micro-meter) stained polymer foil in contact with the water. The film is formed on a special polyethylen-terephthalate (PET) foil by exposing the foil to a hot KMnO 4 solution. This film thickness still allows for some spectrometry as can bee seen from the second figure on the left. 214 Po is well visible on films exposed to a sample containing radon whereas 218 Po shows only as traces. Transfer form solution to thin film happens with a rate that corresponds to a half live of about 40 min. Steady state concentration for 214 Po (in equilibrium with 214 Bi) on the film is about 10% of the activity contained in a100ml sample. 218 Po is at a far lower level because of its large decay rate compared to the adsorption rate. With a 150 mm 2 detector placed at some mm from the foil one gets about 1 count/h in the 214 Po window for every Bq/l of 214 Po in the 100 ml sample. With a larger detector and a larger sample volume 10 counts/h per Bq/l may be realistic. Its clearly not a method to detect fast changes at low levels, but it's a simple method to monitor radon daughter product variations at the 10 Bq/l level with a temporal resolution of an hour. With an intermittent flow through the measurement cell one can determine not only radon daughters but also radon. This may be of some interest to determine residence time of groundwater in a water distribution system. Radon daughters are normally not in equilibrium with radon in freshly pumped groundwater and build up during transport and storage. A more important application of this type of adsorption experiments may be to study adsorption dynamics. There is clear evidence for a different chemical behavior of nuclides that had time to "cool down" after production from their mother nuclide with respect to those freshly produced.(Full paper 50kB, *.pdf)
click to enlarge

When we tested our continuously measuring devices with tap water in our lab, we noted an application for radon as a natural tracer : to study flow patterns in a public water supply system. The city of Fribourg's base water demand is covered by springs having a radon concentration around 8 Bq/l ( units ). At times of higher consumption well water having a low radon concentration is added .
click to enlarge

Our first field measurements are shown in this figure.There is a clear positive correlation between radon concentration and discharge, but why ? We are still looking for an answer. The problem turned out to be too complex because of a mixture of water from different origins (tertiary, quaternary, surface) [ 13 ].
click to enlarge

Data for this spring are easier to explain [ 6 ]. Here we could count on numerous hydrogeologic studies already done for this karst spring. In addition, the radon source is clearly located. Local limestone contains only little radium, but soil in the watershed area is rich in radium, it contains up to some 100 Bq/kg ( units ). After a storm, the discharge first increases without a noticeable change either of the temperature or the radon concentration, but with a conductivity increase lasting several hours. This conductivity peak corresponds to water pushed out from the less permeable zones. Radon levels stay low, reflecting the low radium concentration in local limestone. Radon concentration starts rising only with the temperature drop announcing the arrival of fresh water (during the period shown, temperature in the catchment area was below the annual average). This water has taken up radon during its passage through the soil, which is rich in radium. Radon concentrations in the spring water decrease slower than the discharge. This points to a recharge of low permeability aquifer zones with high-radon water during peak flow and a subsequent release to the spring.








REFERENCES
[1] Lauffenburger, T., Auf der Maur, A., In : Proc. 6th Int. Cong. IRPA, 1984
[2] Surbeck, H., Radiat.Prot.Dosim. 1988, 24, 431 - 434
[3] Surbeck, H., In : Proc. 1992 Int. Symp. on Radon and Radon Reduction Technology, EPA-600/R-93-0836, U.S. EPA, Washington, DC, 1993
[4] Von Gunten, H.R., Surbeck, H. and Rössler, E., Environ.Sci. & Technol., 30/4, 1268 - 1274, 1996
[5] Surbeck, H., Medici, F., In : Proc. Int Conf. on Water Resources in Mountainous Regions, Lausanne, switzerland, 1990, Int. Assoc. of Hydrological Sci., Wallingford, 1990, IAHS-AISH Publication, 193 - 194
[6] Eisenlohr, L. and Surbeck, H., Radon as a natural tracer to study transport processes in a karst system. An example in the Swiss Jura. C.R.Acad.Sci.Paris,321,série IIa, p, 761-767, 1995.
[7] von Philipsborn, H., Schnellmessung von Radonfolgeprodukten (Radon) im Wasser, In : Proc. Strahlenbiologie und Strahlenschutz, 28. Jahrestagung Fachverband für Strahlenschutz, 23.-25.Oktober 1996, Hannover, Verlag TUeV Rheinland, Köln, Germany
[8] TUFF130AWM-H, Alpha Liquid Monitor, Ris Corp., represented in Europe by FCI Frenzel Consulting & Instruments, Straubenhardt, Germany
[9] Gockel, A., Die Radioaktivität von Boden und Quellen, Friedr. Vieweg & Sohn, Braunschweig, Germany, 1914
[10] Surbeck, H., A radon-in-water monitor based on fast gas transfer membranes, In : Proc. Int. Conf on Technologically Enhanced Natural Radioactivity, 16-19 Sept. 1996, Szczyrk, Poland.
[11] Akzo Nobel Faser AG, Membrana, Deutschland.
[12] Surbeck, H., Radon monitoring in soils and water. Nucl.Tracks Radiat.Meas., 22, 1-4, p. 463-468, (1993)
[13] Surbeck, H. and Eisenlohr, L., Radon as a tracer in hydrogeology; a case study, In : Proc. 2nd CIGG, Colloque Internat. sur la Géochimie des gaz, July 5-9, Besançon, France
[14] Ivanovich, M. & Harmon, R.S. (Eds.) 1992 : Uranium Series Disequilibrium: Application to Earth, Marine, and Environmental Sciences, Claredon Press, Oxford, (1992)
[15] McDowell, W.J. and McDowell, B., Liquid Scintillation Alpha Spectrometry, CRC Press, (1994), ISBN 0-8493-5288-6. Spectrometer und extractive scintillators available at ORDELA/ORTEC.
[16] Moore, W.S. and Reid, D.F., Extraction of Radium from Natural Waters Using Manganese Impregnated Acrylic Fibers, J. Geophys. Res. 78 (1973), 8880- 8886
[17] Glöbel, B. and Berlich, J., Eine einfache und schnelle Methode zur Bestimmung von 226Ra in wässrigen Proben, In : Proc. Fachgespräch Ueberwachung der Umweltradioaktivität, 22-24 March 1983, Karlsruhe, Der Bundesminister des Innern.
[18] Surbeck, H. , Piller, G. and Ferreri, G. , Die Suche nach Radonquellen, In: Tagungsbericht "Radon und die Strahlungsbelastung der Lunge", Crameri, R. and Burkart, W. (Eds.), PSI-Bericht Nr.22, Villigen, Switzerland, (1989)
[19] Surbeck, H., Determination of natural radionuclides in drinking water; a tentative protocol. Sci.Total Environment 173/174, p. 91-99, (1995)











AMERICAN AND SI UNITS
1 pCi = 37 mBq 10 mBq = 0.27 pCi
2 pCi =74 mBq 20 mBq = 0.54 pCi
5 pCi = 185 mBq 50 mBq = 1.35 pCi
10 pCi = 370 mBq 100 mBq = 2.7 pCi
20 pCi = 740 mBq 200 mBq = 5.4 pCi
50 pCi = 1.85 Bq 500 mBq = 13.5 pCi
100 pCi = 3.7 Bq 1 Bq = 27 pCi
200 pCi = 7.4 Bq 2 Bq = 54 pCi
500 pCi = 18.5 Bq 5 Bq = 135 pCi
1 nCi = 37 Bq 10 Bq = 270 pCi
10 nCi = 370 Bq 100 Bq = 2.7 nCi