- Email: [email protected]
Studies of the radon concentration in drinking water from the horst Söderåsen in Southern Sweden
Journal of Environmental Radioactivity 53 (2001) 145–154
Studies of the radon concentration in drinking water from the horst So¨dera˚sen in Southern Sweden B. Erlandsson*, B. Jakobsson, G. Jo¨nsson Department of Physics, Lund University, P.O. Box 118, S-221 00 Lund, Sweden Received 1 October 1999; received in revised form 2 May 2000; accepted 9 May 2000
Abstract The radon activity concentration in ground water from drilled and dug wells on the horst So¨dera˚sen in Southern Sweden has been determined with two different methods, gamma activity measurements with a germanium HPGe detector and alpha activity measurements with plastic track detectors. The results are consistent. High activity concentration is connected to granite bedrock. Dug wells have low concentrations and no trivial correlation between concentration and depth of the well is found. Large local variations exist. Activity concentrations >700 Bq/l appear to be associated with leakage from layers of volcanic origin. The concentration from drilled wells is found to be quite constant over a 3 year period but short time variations appear to be significant. Evaporation from the open surface of a normal cooking vessel is slow with an activity gradient DA=A of about 0.1–0.2 per hour at room temperature whereas even modest heating of water in e.g. a coffee machine is very efficient and reduces the radon activity concentration by >90% in one process. # 2001 Elsevier Science Ltd All rights reserved. Keywords: Radon; Drinking water; Methodology; Decontamination; Sweden
1. Introduction Radon in water is a potential health risk when the water is used for consumption in the household, mainly because of the increase of the radon concentration in the air that is inhaled. In Sweden, it is found that high concentration of radon in water is restricted to small areas where the bedrock contains a very high concentration of *Corresponding author. Tel.: +46-462227645; fax: +46-462224709. E-mail address: [email protected] (B. Erlandsson). 0265-931X/00/$ - see front matter # 2001 Elsevier Science Ltd All rights reserved. PII: S 0 2 6 5 - 9 3 1 X ( 0 0 ) 0 0 1 1 9 - 3
146
B. Erlandsson et al. / J. Environ. Radioactivity 53 (2001) 145–154
uranium or radium. There are, however, large areas with bedrock of granite or alum shell where the concentration is above 100 Bq/l, the current limit recommended for the sake of precaution by the Swedish Radiation Protection Institute. A review of the problem of radon in ground water in Sweden has been presented by A˚kerblom and Lindgren (1997). A general presentation on radon in water has been given by Graves et al. (1987). A common method for measuring the radon activity concentration in drinking water is to tap water in bottles for measurement of the gamma activity from the radon daughters. Detectors used in such measurements are NaI scintillators, liquid scintillators and HPGe detectors. A special method for rapid analysis was presented by Freyer, Treutler, Dehnert and Nestler (1997) using liquid scintillation spectrometry and by Surbeck (1993) who presented a method for continuous monitoring of radon activity in water. It is also possible to use track detectors (Singh, Singh, Singh & Virk, 1986; Vasarhelyi, Csige, Hakl & Hunyadi 1997). In this case, it is the alpha activity that is measured. In this paper, we compare measurements made both with HPGe detectors and plastic film detectors on the radon activity in water. We have taken special precautions against efficiency losses from escaping radon gas during sampling and measurements. Our results, from samples collected on the horst So¨derasen in Southern Sweden, have been compared with results from an earlier investigation in the same area by the Institute for Geological Survey in Sweden (SGU) (Gustafsson, 1992). We also present a discussion of our observations of a number of effects that influence the activity concentration.
2. Experimental methods Groundwater sampling was always performed according to a strict protocol. The drinking-water tap was opened for 10 min before drawing the samples. For the HPGe method, 30 cl glass bottles were completely filled and immediately closed under water in order to avoid air bubbles. Samples were normally delivered for gamma activity measurements within 24 h. It is important that the bottles are made of glass since even 1 mm thick plastic bottles have shown a leakage of Rn gas of about 10% per day. For the plastic film method we filled 2 mm thick 10 l plastic buckets up to 2 l level and immediately afterwards the bucket was tightly covered with a plastic lid. In the centre of the inner surface of the lid a 21 cm2 plastic film strip was placed, facing the 1600 cm2 water surface. The film is of the type Kodak LR 115-II. These containers were kept closed for either 2,3 or 4 weeks, before the film was removed and isolated from further alpha irradiation. To measure the gamma activity concentration, the bottle was placed in front of an HPGe detector with 23% relative efficiency and a resolution of 1.9 keV FWHM at 1332 keV. The detector efficiency was calibrated with a standard 226 Ra solution from PTB (Physikalisches Technisches Bundesanstalt) in Braunsweig. The accuracy of the activity concentration of this Ra solution is 2% and the 226 Ra activity was in equilibrium with its decay products. The 351.9 keV gamma decay from 214 Pb and the
B. Erlandsson et al. / J. Environ. Radioactivity 53 (2001) 145–154
147
609.3, 1377.7 and 1764.5 keV gamma decays from 214 Bi were used for normalization of the actual activity measurements on 222 Rn from the water samples. All samples contained the same volume of water in the same kind of bottle. The gamma emission probabilities are taken from Scho¨tzig and Debretin (1983). It should be noticed that these probabilities are somewhat lower than those given by Chu, Ekstro¨m and Firestone (1999), resulting in a correspondingly higher radon concentration. This indicates that the systematic error may be of the order of 3–4%. The overall uncertainty in the 222 Rn activity is then determined to be 5%. We have measured the long-term stability of the calibration source and found variations of less than 1% over a 2-year period. The evaluation of the activity with the plastic film method in this investigation was made by counting the number of alpha hits per unit area in a film of Kodak type LR115-II with a 12 mm alpha-sensitive cellulose nitrate layer on thick backing. The film, which is of the same type as used for monitoring radon activity measurements both in indoor air and in soil air (Jo¨nsson, 1995), is etched in 10% NaOH for 2 h at a temperature of 598C. The sensitivity of the film to low-energy alpha bombardment at different angles as well as other details have been presented by Jo¨nsson and Hellborg (1992). The film is sensitive to alpha particles with energies between 1.5 and 4.5 MeV. In the measurements performed here, the maximum alpha energies are 5.5 (222 Rn), 6.0 (218 Po) and 7.7 MeV (214 Po). The holes in the film are counted at 400 times magnification of the microscope. The scanning is always continued until at least 200 holes are found. The background in the film can be neglected. The film thus registers the activity built up in the air above the water surface coming from the escaping radon gas and the daughters formed in this air. The total activity is given as the number of tracks per field of view (FOV) in the microscope.
3. Experimental results The water samples were collected on the horst So¨dera˚sen, 56800’–56809’N, 13810’– 13817’E, in the southernmost region of Scania in Sweden. Fig. 1 shows the position of the sampling sites. The bedrock consisted mainly of granite and gneiss. It is well known that in certain regions volcanic activity took place during the Tertiary era (Wikman, Bergstro¨m & Sivhed, 1993). Along the main direction of the horst, southeast–northwest, there are several diabase passages. Deep cracks in the bedrock could be expected along these passages. In our investigation, 21 wells with depths up to 103 m were examined. In the earlier, NaI-based measurements (Gustafsson, 1992), 20 wells were investigated and out of these seven could be identified and these were also included in our investigation. 3.1. Measurements with plastic film The time dependence of the alpha registration in plastic film was investigated by 2, 3 and 4 weeks of exposure. Fig. 2 shows the correlation between 2 and 4 weeks of
148
B. Erlandsson et al. / J. Environ. Radioactivity 53 (2001) 145–154
Fig. 1. Map of bedrock and sampling sites on the horst So¨dera˚sen.
Fig. 2. The number of alpha tracks per FOV after 2 weeks of exposure vs. the number after 4 weeks of exposure in the same samples.
B. Erlandsson et al. / J. Environ. Radioactivity 53 (2001) 145–154
149
exposure. The number of tracks per FOV is 1–15 which, according to comparisons with HPGe measurements of the same samples, correspond to a radon activity concentration between 30 and 400 Bq/l. No increase in the number of alpha tracks after 2 weeks is observed and the correlation slope is also consistent with unity. The conclusion is that the alpha contribution from 222 Rn or any of its daughters along the decay chain to 210 Pb is negligible after 2 weeks. This is reasonable since only the 222 Rn decay has a lifetime (half-life, t1=2 =3.82 d) comparable to this time and actually the activity has decreased in this case to 0.026 of the original value. The decay of 210 Po, with a half-life of 138.4 d, will effectively be blocked by the 22.3 y beta decay of 210 Pb. Even if the plastic in both bucket and lid is so thick that no leakage of radon gas is expected, we suggest a somewhat shorter exposure time, say 10–14 days. 3.2. Comparisons with the HPGe measurements Fig. 3 shows the comparison between the plastic film results and the Ge detector measurements from the same tapping. The correlation is well established and shows that the plastic film method can very well be used in practice if a proper absolute calibration is made. A least-squares fit to the linear correlation is shown. The negative value of the intersection with the y-axis indicates that the subtraction of the background in the plastic film method is overestimated by 1–2 tracks/FOV. The calibration function should thus read as Na ¼ 0:0399ARn ÿ 0:555½Bq=l;
ð1Þ
Fig. 3. The number of alpha tracks per FOV after 2 weeks of exposure vs. the radon activity measured by the Ge(Li) method.
150
B. Erlandsson et al. / J. Environ. Radioactivity 53 (2001) 145–154
where ARn stands for the activity concentration measured by the Ge(Li) method and Na is the number of alpha tracks per FOV. 3.3. Results on activity concentration The radon activity distribution for our 21 samples and the additional samples from SGU (Gustafsson, 1992) which have been corrected for the observed systematic discrepancy discussed below, is presented in Fig. 4. This distribution appears to be binomial or possibly composed of one Gaussian and one exponential distribution with an extended high concentration tail. The fact that the dug wells have lower concentrations indicates that the water in them is surface water (A˚kerblom & Lindgren, 1997). The concentration as a function of the depth of the well is plotted in Fig. 5. It is obviously difficult to extract a significant correlation in this case, which is confirmed by the significance level (w2 =1.39 per d.o.f.) if a linear relation is assumed. This may depend on higher concentrations of U and Ra in volcanic ash layers or strata with volcanic origin.
Fig. 4. Distribution of Rn activity concentration measured from drilled wells, dug wells and artesian wells.
B. Erlandsson et al. / J. Environ. Radioactivity 53 (2001) 145–154
151
Seven wells were measured both by us and by SGU (Gustafsson, 1992) and these activity concentrations are shown in Table 1. Thus one can observe: *
One case where the sample collecting must have been wrong in the previous investigation since it shows a very low concentration in an area where all neighbouring values are of the order of the present result. Probably, the water has been open to evaporation which shows the importance of clear instructions if the sample collection is carried out by the well owner.
Fig. 5. Rn activity concentration as a function of depth of well. The artesian wells also belong to ‘‘this measurement’’. Table 1 Depth of well and radon concentration in drilled wells investigated by Gustafsson (1992) and in this work Depth of well (m)
Act. conc. of Rn (Bq/l)
Act. conc. of Rn from SGU (Bq/l)
29 12 45 102 59 37 44
22020 24512 26510 35020 46015 65620 153275
29 410 575 699 746 743 1656
152
B. Erlandsson et al. / J. Environ. Radioactivity 53 (2001) 145–154
Systematically lower concentrations in our measurements. On average this difference is 200100 Bq/l in absolute concentration but its functional dependence is very difficult to determine from these limited statistics. The difference may indicate a problem in determining the background, which should be subtracted to obtain the correct Rn concentration.
*
One concentration value in Table 1 (depth 44 m) is definitely higher than any other. It is interesting to notice the remark from the SGU archive about this well: ‘‘Volcanic ashes were coming up from the 40 m level’’. Finally, we have studied the time dependence of the radon activity concentration in a well with medium high concentration (Table 2.). The average of the 14 measurements during 1995–1998 is 29540 Bq/l. The remarkably small long-term variation is partly a consequence of the careful sampling technique and partly due to the large depth, 72 m, of this well, drilled into the granite bedrock. Instead, we observe a short time variation (Fig. 6a) over a few days with a pattern which is not varying randomly. The standard deviation of all measurements is 40 Bq/l and this is much larger than the statistical errors of the individual measurements. No direct correlation between the non-statistical variation and the air pressure was found and furthermore we do not find any plausible reason for a variation in the pumping procedure. Table 2 also shows how the activity in water is decreasing with time when exposed to free evaporation after tapping. This loss is obviously non-linear,0 at least for the first 30 min, as shown in Fig. 6b. Although a free surface for evaporation reduces the activity, this is obviously a slow process and such water can, even after several hours, provide a non-negligible radioactive dose to babies. Any boiling process, even soft boiling in a coffee machine, makes the activity disappear completely as discussed earlier (A˚kerblom & Lindgren, 1997). For such a test we took from the same sample one subsample that was immediately isolated after tapping and one that was isolated
Table 2 The Rn activity concentration in water samples collected under identical conditions on different occasions Year, month, day, hour
Act. conc. of Rn (Bq/l)
Year, month, day, hour
Act. conc. of Rn (Bq/l)
950910 20.00 960306 12.00 960731 12.00 970608 12.00 evap a0.5 h evap a1 h evap a2 h
28810 33415 35818 32210 24110 2319 1998
971226 971227 971228 971229 971230 971231 980101 980102 980110 980111
30010 2625 2433 2355 2655 35810 31410 26110 28510 30910
a
12.00 12.30 12.00 12.00 12.30 04.00 12.00 12.00 14.00 12.00
open evaporation at room temperature, 208C. The errors given in the table contain both statistical and systematic contributions.
B. Erlandsson et al. / J. Environ. Radioactivity 53 (2001) 145–154
153
Fig. 6. Time variation of the Rn activity concentration from the test well (a) and after evaporation from an open surface (b). Dates are given for the first five points in (a) and for the last ten points there is a difference of 1 day.
after processing in a normal coffee machine. This was repeated a second time. The first sample had a decrease of radon activity from 322 to 34 Bq/l and the second from 285 to 32 Bq/l by this process. Naturally, if this method is to be used, effective ventilation over the boiling water must be installed in order to avoid the inhalation risk.
4. Concluding remarks It has been shown in this report that plastic film is useful to measure radon activity concentrations in water if the sampling is controlled and absolute calibration of the activity can be performed by another method, like Ge detector measurements of gamma activity. The Ge samples must be tapped in glass bottles at the same time as the plastic film samples are tapped in thick enough buckets to prevent diffusion of radon gas. These buckets must be well sealed for 10–14 days.
154
B. Erlandsson et al. / J. Environ. Radioactivity 53 (2001) 145–154
The measured radon activity in samples from the horst So¨dera˚sen confirms (A˚kerblom & Lindgren, 1997) that dug wells have low concentrations. There is no trivial depth dependence in the drilled wells and high concentrations may instead be connected to radium and uranium deposited on surfaces in cracks in the bedrock. It was found that layers of volcanic ash contribute to high radon activity concentration in water from drilled wells. It is confirmed (A˚kerblom & Lindgren, 1997) that the radon activity concentration in a water sample is reduced by at least 90% if processed in a coffee machine. From evaporation at room temperature the activity decrease is slow and exponential due to initial turbulence in the water. The radon activity concentration from a deep drilled well is remarkably constant over a period of 3 years, although short-time (daily) non-statistical variations appear.
Acknowledgements Discussions with Gustaf A˚kerblom, Thomas Arnstro¨m and Ove Gustafsson are acknowledged and Gustaf A˚kerblom is also thanked for allowing us to use partly unpublished data. Jessika Jakobsson is thanked for collecting the plastic film water samples and performing measurements on them.
References A˚kerblom, G., & Lindgren, J. (1997). Mapping of ground water radon potential. European Geologist, 5, 13–22. Chu, S. Y. F., Ekstro¨m, L. P., & Firestone, R. B. (1999). www Table of Isotopes, database version 2/28/99 from URL http:nucleardata.nuclear.lu.se/toi/. Freyer, K., Treutler, H. C., Dehnert, J., & Nestler, W. (1997). Sampling and measurement of radon-222 in water. Journal of Environmental Radioactivity, 37, 327–337. Graves, B., et al. (1987). Radon, Radium, and Other Radioactivity in Ground Water. In Hydrological Impact and Application to Indoor Airborne Contamination. Chelsea, Mich., USA: Lewis Publishers. Gustafsson, O. (1992). The activity concentration of radon in granite areas in the Malmo¨hus county. The Geological Survey of Sweden Report 72 (1992). In Swedish. Jo¨nsson, G. (1995). Radon gas } where from and what to do? Radiation Measurements, 25, 537–546. Jo¨nsson, G., & Hellborg, R. (1992). The exposure of plastic films and their response to 1.3-5.5 MeV helium iones. Nuclear Instruments & Methods in Physics Research, B63, 399–406. Scho¨tzig, U., & Debertin, K. (1983). Photon emission probabilities from decay of 226Ra and 232Th in equilibrium with their daughter products. International Journal of Applied Radiation and Isotopes, 34, 533–538. Singh, M., Singh, N. P., Singh, S., & Virk, H. S. (1986). Calibration of radon detectors. Nuclear Tracks, 12, 739–742. Surbeck, H. (1993). Radon monitoring in soils and water. Nuclear Tracks Radiation Measurement, 22, 463–468. Vasarhelyi, A., Csige, I., Hakl, J., & Hunyadi, I. (1997). Spatial distribution of radon content of soil-gas and well-waters measured with etched track radon monitors. Radiation Measurement, 28, 685–690. Wikman, H., Bergstro¨m, J., Sivhed, U. (1993). Description of the bedrock map for Helsingborg SE, SGU series Af 180.
Studies of the radon concentration in drinking water from the horst So¨dera˚sen in Southern Sweden B. Erlandsson*, B. Jakobsson, G. Jo¨nsson Department of Physics, Lund University, P.O. Box 118, S-221 00 Lund, Sweden Received 1 October 1999; received in revised form 2 May 2000; accepted 9 May 2000
Abstract The radon activity concentration in ground water from drilled and dug wells on the horst So¨dera˚sen in Southern Sweden has been determined with two different methods, gamma activity measurements with a germanium HPGe detector and alpha activity measurements with plastic track detectors. The results are consistent. High activity concentration is connected to granite bedrock. Dug wells have low concentrations and no trivial correlation between concentration and depth of the well is found. Large local variations exist. Activity concentrations >700 Bq/l appear to be associated with leakage from layers of volcanic origin. The concentration from drilled wells is found to be quite constant over a 3 year period but short time variations appear to be significant. Evaporation from the open surface of a normal cooking vessel is slow with an activity gradient DA=A of about 0.1–0.2 per hour at room temperature whereas even modest heating of water in e.g. a coffee machine is very efficient and reduces the radon activity concentration by >90% in one process. # 2001 Elsevier Science Ltd All rights reserved. Keywords: Radon; Drinking water; Methodology; Decontamination; Sweden
1. Introduction Radon in water is a potential health risk when the water is used for consumption in the household, mainly because of the increase of the radon concentration in the air that is inhaled. In Sweden, it is found that high concentration of radon in water is restricted to small areas where the bedrock contains a very high concentration of *Corresponding author. Tel.: +46-462227645; fax: +46-462224709. E-mail address: [email protected] (B. Erlandsson). 0265-931X/00/$ - see front matter # 2001 Elsevier Science Ltd All rights reserved. PII: S 0 2 6 5 - 9 3 1 X ( 0 0 ) 0 0 1 1 9 - 3
146
B. Erlandsson et al. / J. Environ. Radioactivity 53 (2001) 145–154
uranium or radium. There are, however, large areas with bedrock of granite or alum shell where the concentration is above 100 Bq/l, the current limit recommended for the sake of precaution by the Swedish Radiation Protection Institute. A review of the problem of radon in ground water in Sweden has been presented by A˚kerblom and Lindgren (1997). A general presentation on radon in water has been given by Graves et al. (1987). A common method for measuring the radon activity concentration in drinking water is to tap water in bottles for measurement of the gamma activity from the radon daughters. Detectors used in such measurements are NaI scintillators, liquid scintillators and HPGe detectors. A special method for rapid analysis was presented by Freyer, Treutler, Dehnert and Nestler (1997) using liquid scintillation spectrometry and by Surbeck (1993) who presented a method for continuous monitoring of radon activity in water. It is also possible to use track detectors (Singh, Singh, Singh & Virk, 1986; Vasarhelyi, Csige, Hakl & Hunyadi 1997). In this case, it is the alpha activity that is measured. In this paper, we compare measurements made both with HPGe detectors and plastic film detectors on the radon activity in water. We have taken special precautions against efficiency losses from escaping radon gas during sampling and measurements. Our results, from samples collected on the horst So¨derasen in Southern Sweden, have been compared with results from an earlier investigation in the same area by the Institute for Geological Survey in Sweden (SGU) (Gustafsson, 1992). We also present a discussion of our observations of a number of effects that influence the activity concentration.
2. Experimental methods Groundwater sampling was always performed according to a strict protocol. The drinking-water tap was opened for 10 min before drawing the samples. For the HPGe method, 30 cl glass bottles were completely filled and immediately closed under water in order to avoid air bubbles. Samples were normally delivered for gamma activity measurements within 24 h. It is important that the bottles are made of glass since even 1 mm thick plastic bottles have shown a leakage of Rn gas of about 10% per day. For the plastic film method we filled 2 mm thick 10 l plastic buckets up to 2 l level and immediately afterwards the bucket was tightly covered with a plastic lid. In the centre of the inner surface of the lid a 21 cm2 plastic film strip was placed, facing the 1600 cm2 water surface. The film is of the type Kodak LR 115-II. These containers were kept closed for either 2,3 or 4 weeks, before the film was removed and isolated from further alpha irradiation. To measure the gamma activity concentration, the bottle was placed in front of an HPGe detector with 23% relative efficiency and a resolution of 1.9 keV FWHM at 1332 keV. The detector efficiency was calibrated with a standard 226 Ra solution from PTB (Physikalisches Technisches Bundesanstalt) in Braunsweig. The accuracy of the activity concentration of this Ra solution is 2% and the 226 Ra activity was in equilibrium with its decay products. The 351.9 keV gamma decay from 214 Pb and the
B. Erlandsson et al. / J. Environ. Radioactivity 53 (2001) 145–154
147
609.3, 1377.7 and 1764.5 keV gamma decays from 214 Bi were used for normalization of the actual activity measurements on 222 Rn from the water samples. All samples contained the same volume of water in the same kind of bottle. The gamma emission probabilities are taken from Scho¨tzig and Debretin (1983). It should be noticed that these probabilities are somewhat lower than those given by Chu, Ekstro¨m and Firestone (1999), resulting in a correspondingly higher radon concentration. This indicates that the systematic error may be of the order of 3–4%. The overall uncertainty in the 222 Rn activity is then determined to be 5%. We have measured the long-term stability of the calibration source and found variations of less than 1% over a 2-year period. The evaluation of the activity with the plastic film method in this investigation was made by counting the number of alpha hits per unit area in a film of Kodak type LR115-II with a 12 mm alpha-sensitive cellulose nitrate layer on thick backing. The film, which is of the same type as used for monitoring radon activity measurements both in indoor air and in soil air (Jo¨nsson, 1995), is etched in 10% NaOH for 2 h at a temperature of 598C. The sensitivity of the film to low-energy alpha bombardment at different angles as well as other details have been presented by Jo¨nsson and Hellborg (1992). The film is sensitive to alpha particles with energies between 1.5 and 4.5 MeV. In the measurements performed here, the maximum alpha energies are 5.5 (222 Rn), 6.0 (218 Po) and 7.7 MeV (214 Po). The holes in the film are counted at 400 times magnification of the microscope. The scanning is always continued until at least 200 holes are found. The background in the film can be neglected. The film thus registers the activity built up in the air above the water surface coming from the escaping radon gas and the daughters formed in this air. The total activity is given as the number of tracks per field of view (FOV) in the microscope.
3. Experimental results The water samples were collected on the horst So¨dera˚sen, 56800’–56809’N, 13810’– 13817’E, in the southernmost region of Scania in Sweden. Fig. 1 shows the position of the sampling sites. The bedrock consisted mainly of granite and gneiss. It is well known that in certain regions volcanic activity took place during the Tertiary era (Wikman, Bergstro¨m & Sivhed, 1993). Along the main direction of the horst, southeast–northwest, there are several diabase passages. Deep cracks in the bedrock could be expected along these passages. In our investigation, 21 wells with depths up to 103 m were examined. In the earlier, NaI-based measurements (Gustafsson, 1992), 20 wells were investigated and out of these seven could be identified and these were also included in our investigation. 3.1. Measurements with plastic film The time dependence of the alpha registration in plastic film was investigated by 2, 3 and 4 weeks of exposure. Fig. 2 shows the correlation between 2 and 4 weeks of
148
B. Erlandsson et al. / J. Environ. Radioactivity 53 (2001) 145–154
Fig. 1. Map of bedrock and sampling sites on the horst So¨dera˚sen.
Fig. 2. The number of alpha tracks per FOV after 2 weeks of exposure vs. the number after 4 weeks of exposure in the same samples.
B. Erlandsson et al. / J. Environ. Radioactivity 53 (2001) 145–154
149
exposure. The number of tracks per FOV is 1–15 which, according to comparisons with HPGe measurements of the same samples, correspond to a radon activity concentration between 30 and 400 Bq/l. No increase in the number of alpha tracks after 2 weeks is observed and the correlation slope is also consistent with unity. The conclusion is that the alpha contribution from 222 Rn or any of its daughters along the decay chain to 210 Pb is negligible after 2 weeks. This is reasonable since only the 222 Rn decay has a lifetime (half-life, t1=2 =3.82 d) comparable to this time and actually the activity has decreased in this case to 0.026 of the original value. The decay of 210 Po, with a half-life of 138.4 d, will effectively be blocked by the 22.3 y beta decay of 210 Pb. Even if the plastic in both bucket and lid is so thick that no leakage of radon gas is expected, we suggest a somewhat shorter exposure time, say 10–14 days. 3.2. Comparisons with the HPGe measurements Fig. 3 shows the comparison between the plastic film results and the Ge detector measurements from the same tapping. The correlation is well established and shows that the plastic film method can very well be used in practice if a proper absolute calibration is made. A least-squares fit to the linear correlation is shown. The negative value of the intersection with the y-axis indicates that the subtraction of the background in the plastic film method is overestimated by 1–2 tracks/FOV. The calibration function should thus read as Na ¼ 0:0399ARn ÿ 0:555½Bq=l;
ð1Þ
Fig. 3. The number of alpha tracks per FOV after 2 weeks of exposure vs. the radon activity measured by the Ge(Li) method.
150
B. Erlandsson et al. / J. Environ. Radioactivity 53 (2001) 145–154
where ARn stands for the activity concentration measured by the Ge(Li) method and Na is the number of alpha tracks per FOV. 3.3. Results on activity concentration The radon activity distribution for our 21 samples and the additional samples from SGU (Gustafsson, 1992) which have been corrected for the observed systematic discrepancy discussed below, is presented in Fig. 4. This distribution appears to be binomial or possibly composed of one Gaussian and one exponential distribution with an extended high concentration tail. The fact that the dug wells have lower concentrations indicates that the water in them is surface water (A˚kerblom & Lindgren, 1997). The concentration as a function of the depth of the well is plotted in Fig. 5. It is obviously difficult to extract a significant correlation in this case, which is confirmed by the significance level (w2 =1.39 per d.o.f.) if a linear relation is assumed. This may depend on higher concentrations of U and Ra in volcanic ash layers or strata with volcanic origin.
Fig. 4. Distribution of Rn activity concentration measured from drilled wells, dug wells and artesian wells.
B. Erlandsson et al. / J. Environ. Radioactivity 53 (2001) 145–154
151
Seven wells were measured both by us and by SGU (Gustafsson, 1992) and these activity concentrations are shown in Table 1. Thus one can observe: *
One case where the sample collecting must have been wrong in the previous investigation since it shows a very low concentration in an area where all neighbouring values are of the order of the present result. Probably, the water has been open to evaporation which shows the importance of clear instructions if the sample collection is carried out by the well owner.
Fig. 5. Rn activity concentration as a function of depth of well. The artesian wells also belong to ‘‘this measurement’’. Table 1 Depth of well and radon concentration in drilled wells investigated by Gustafsson (1992) and in this work Depth of well (m)
Act. conc. of Rn (Bq/l)
Act. conc. of Rn from SGU (Bq/l)
29 12 45 102 59 37 44
22020 24512 26510 35020 46015 65620 153275
29 410 575 699 746 743 1656
152
B. Erlandsson et al. / J. Environ. Radioactivity 53 (2001) 145–154
Systematically lower concentrations in our measurements. On average this difference is 200100 Bq/l in absolute concentration but its functional dependence is very difficult to determine from these limited statistics. The difference may indicate a problem in determining the background, which should be subtracted to obtain the correct Rn concentration.
*
One concentration value in Table 1 (depth 44 m) is definitely higher than any other. It is interesting to notice the remark from the SGU archive about this well: ‘‘Volcanic ashes were coming up from the 40 m level’’. Finally, we have studied the time dependence of the radon activity concentration in a well with medium high concentration (Table 2.). The average of the 14 measurements during 1995–1998 is 29540 Bq/l. The remarkably small long-term variation is partly a consequence of the careful sampling technique and partly due to the large depth, 72 m, of this well, drilled into the granite bedrock. Instead, we observe a short time variation (Fig. 6a) over a few days with a pattern which is not varying randomly. The standard deviation of all measurements is 40 Bq/l and this is much larger than the statistical errors of the individual measurements. No direct correlation between the non-statistical variation and the air pressure was found and furthermore we do not find any plausible reason for a variation in the pumping procedure. Table 2 also shows how the activity in water is decreasing with time when exposed to free evaporation after tapping. This loss is obviously non-linear,0 at least for the first 30 min, as shown in Fig. 6b. Although a free surface for evaporation reduces the activity, this is obviously a slow process and such water can, even after several hours, provide a non-negligible radioactive dose to babies. Any boiling process, even soft boiling in a coffee machine, makes the activity disappear completely as discussed earlier (A˚kerblom & Lindgren, 1997). For such a test we took from the same sample one subsample that was immediately isolated after tapping and one that was isolated
Table 2 The Rn activity concentration in water samples collected under identical conditions on different occasions Year, month, day, hour
Act. conc. of Rn (Bq/l)
Year, month, day, hour
Act. conc. of Rn (Bq/l)
950910 20.00 960306 12.00 960731 12.00 970608 12.00 evap a0.5 h evap a1 h evap a2 h
28810 33415 35818 32210 24110 2319 1998
971226 971227 971228 971229 971230 971231 980101 980102 980110 980111
30010 2625 2433 2355 2655 35810 31410 26110 28510 30910
a
12.00 12.30 12.00 12.00 12.30 04.00 12.00 12.00 14.00 12.00
open evaporation at room temperature, 208C. The errors given in the table contain both statistical and systematic contributions.
B. Erlandsson et al. / J. Environ. Radioactivity 53 (2001) 145–154
153
Fig. 6. Time variation of the Rn activity concentration from the test well (a) and after evaporation from an open surface (b). Dates are given for the first five points in (a) and for the last ten points there is a difference of 1 day.
after processing in a normal coffee machine. This was repeated a second time. The first sample had a decrease of radon activity from 322 to 34 Bq/l and the second from 285 to 32 Bq/l by this process. Naturally, if this method is to be used, effective ventilation over the boiling water must be installed in order to avoid the inhalation risk.
4. Concluding remarks It has been shown in this report that plastic film is useful to measure radon activity concentrations in water if the sampling is controlled and absolute calibration of the activity can be performed by another method, like Ge detector measurements of gamma activity. The Ge samples must be tapped in glass bottles at the same time as the plastic film samples are tapped in thick enough buckets to prevent diffusion of radon gas. These buckets must be well sealed for 10–14 days.
154
B. Erlandsson et al. / J. Environ. Radioactivity 53 (2001) 145–154
The measured radon activity in samples from the horst So¨dera˚sen confirms (A˚kerblom & Lindgren, 1997) that dug wells have low concentrations. There is no trivial depth dependence in the drilled wells and high concentrations may instead be connected to radium and uranium deposited on surfaces in cracks in the bedrock. It was found that layers of volcanic ash contribute to high radon activity concentration in water from drilled wells. It is confirmed (A˚kerblom & Lindgren, 1997) that the radon activity concentration in a water sample is reduced by at least 90% if processed in a coffee machine. From evaporation at room temperature the activity decrease is slow and exponential due to initial turbulence in the water. The radon activity concentration from a deep drilled well is remarkably constant over a period of 3 years, although short-time (daily) non-statistical variations appear.
Acknowledgements Discussions with Gustaf A˚kerblom, Thomas Arnstro¨m and Ove Gustafsson are acknowledged and Gustaf A˚kerblom is also thanked for allowing us to use partly unpublished data. Jessika Jakobsson is thanked for collecting the plastic film water samples and performing measurements on them.
References A˚kerblom, G., & Lindgren, J. (1997). Mapping of ground water radon potential. European Geologist, 5, 13–22. Chu, S. Y. F., Ekstro¨m, L. P., & Firestone, R. B. (1999). www Table of Isotopes, database version 2/28/99 from URL http:nucleardata.nuclear.lu.se/toi/. Freyer, K., Treutler, H. C., Dehnert, J., & Nestler, W. (1997). Sampling and measurement of radon-222 in water. Journal of Environmental Radioactivity, 37, 327–337. Graves, B., et al. (1987). Radon, Radium, and Other Radioactivity in Ground Water. In Hydrological Impact and Application to Indoor Airborne Contamination. Chelsea, Mich., USA: Lewis Publishers. Gustafsson, O. (1992). The activity concentration of radon in granite areas in the Malmo¨hus county. The Geological Survey of Sweden Report 72 (1992). In Swedish. Jo¨nsson, G. (1995). Radon gas } where from and what to do? Radiation Measurements, 25, 537–546. Jo¨nsson, G., & Hellborg, R. (1992). The exposure of plastic films and their response to 1.3-5.5 MeV helium iones. Nuclear Instruments & Methods in Physics Research, B63, 399–406. Scho¨tzig, U., & Debertin, K. (1983). Photon emission probabilities from decay of 226Ra and 232Th in equilibrium with their daughter products. International Journal of Applied Radiation and Isotopes, 34, 533–538. Singh, M., Singh, N. P., Singh, S., & Virk, H. S. (1986). Calibration of radon detectors. Nuclear Tracks, 12, 739–742. Surbeck, H. (1993). Radon monitoring in soils and water. Nuclear Tracks Radiation Measurement, 22, 463–468. Vasarhelyi, A., Csige, I., Hakl, J., & Hunyadi, I. (1997). Spatial distribution of radon content of soil-gas and well-waters measured with etched track radon monitors. Radiation Measurement, 28, 685–690. Wikman, H., Bergstro¨m, J., Sivhed, U. (1993). Description of the bedrock map for Helsingborg SE, SGU series Af 180.