Journal of Environmental Radioactivity 140 (2015) 16e24

Contents lists available at ScienceDirect

Journal of Environmental Radioactivity journal homepage: www.elsevier.com/locate/jenvrad

Research note

Determining the radon exhalation rate from a gold mine tailings dump by measuring the gamma radiation Joash N. Ongori a, Robert Lindsay a, *, Richard T. Newman b, Peane P. Maleka c a

Department of Physics, University of the Western Cape, Private Bag X17, Bellville 7535, South Africa Department of Physics, Stellenbosch University, Private Bag X1, Matieland, Stellenbosch 7602, South Africa c Department of Nuclear Physics, iThemba Laboratory for Accelerator Based Sciences (LABS), P. O. Box 722, Somerset West 7129, South Africa b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 13 June 2014 Received in revised form 6 October 2014 Accepted 19 October 2014 Available online

The mining activities taking place in Gauteng province, South Africa have caused millions of tons of rocks to be taken from underground to be milled and processed to extract gold. The uranium bearing tailings are placed in an estimated 250 dumps covering a total area of about 7000 ha. These tailings dumps contain considerable amounts of radium and have therefore been identified as large sources of radon. The size of these dumps make traditional radon exhalation measurements time consuming and it is difficult to get representative measurements for the whole dump. In this work radon exhalation measurements from the non-operational Kloof mine dump have been performed by measuring the gamma radiation from the dump fairly accurately over an area of more than 1 km2. Radon exhalation from the mine dump have been inferred from this by laboratory-based and insitu gamma measurements. Thirty four soil samples were collected at depths of 30 cm and 50 cm. The weighted average activity concentrations in the soil samples were 308 ± 7 Bq kg
1, 255 ± 5 Bq kg
1 and 18 ± 1 Bq kg
1 for 238U, 40K and 232Th, respectively. The MEDUSA (Multi-Element Detector for Underwater Sediment Activity) g-ray detection system was used for field measurements. The radium concentrations were then used with soil parameters to obtain the radon flux using different approaches such as the IAEA (International Atomic Energy Agency) formula. Another technique the MEDUSA Laboratory Technique (MELT) was developed to map radon exhalation based on (1) recognising that radon exhalation does not affect 40K and 232Th activity concentrations and (2) that the ratio of the activity concentration of the field (MEDUSA) to the laboratory (HPGe) for 238U and 40K or 238U and 232 Th will give a measure of the radon exhalation at a particular location in the dump. The average, normalised radon flux was found to be 0.12 ± 0.02 Bq m
2 s
1 for the mine dump. © 2014 Elsevier Ltd. All rights reserved.

Keywords: Radon Radon exhalation Gamma-ray spectrometry Gold mine tailings South Africa

1. Introduction Radon (herein referring to 222Rn) gas has attracted the attention of various researchers ever since it was understood that radon and its progeny are harmful to human health. The short-lived daughter products of radon, 218Po, 214Pb, 214Bi and 214Po have such short halflives that they probably decay through to the longer-lived 210Pb before the lung can clear them. Due to that, the alpha decays of 218 Po and 214Po give the epithelial layer of the bronchi a substantial radiation dose; consequently radon and its progeny pose risks to human health (Durrani and Ilic, 1997; BEIR VI, 1999; WHO, 2009). In the soil, radon gas is released by the decay of radium and afterwards some radon atoms get to the pore spaces. The atoms

* Corresponding author. Tel.: þ27 (0) 21 959 2326/7. E-mail address: [email protected] (R. Lindsay). http://dx.doi.org/10.1016/j.jenvrad.2014.10.012 0265-931X/© 2014 Elsevier Ltd. All rights reserved.

become available to migrate over large distances within the earth and in the atmosphere through the diffusion process as described by Fick's Law (Crank, 1975) or by advection as described by Darcy's Law. Radon transport equations and numerical models have been developed to explain the underlying processes involved in radon gas transport (Nazaroff et al., 1988; Loureiro, 1990; Owczarski, 1990; Rogers and Nielson, 1991; Kohl, 1994; Van der Spoel, 1997, 1998, 1999; Andersen, 2000, 2001; Schubert, 2002; Antonopoulos-Domis, 2009). In-situ measurements have also been carried out to investigate the processes involved in radon gas transport (Bigu, 1984; Åkerblom and Mellander, 1997; Tsela and Brits, 1998; Lindsay et al. (2004a, 2004b, 2008); AntonopoulosDomis, 2009). Radon releases to the atmosphere have been measured from materials such as soil, cement, building materials, granite, uranium tailings and concrete (UNSCEAR, 1982; Ingersoll, 1983; Schery,

J.N. Ongori et al. / Journal of Environmental Radioactivity 140 (2015) 16e24

17

1989; Graustein, 1990; Aldenkamp et al., 1992; Chen, 1993; De Jong, 1996; Escobar, 1999; Arafa, 2004; Kovler, 2005; Mudd, 2008a,b; Lawrence, 2009; Sahoo, 2010). Some of the techniques that measure radon releases either in the laboratory or in the field include the accumulation can method, the flow-through method, adsorption method, vertical profile method, soil concentration gradient method and the diffusion tube method (Clements, 1974; Fleischer, 1980; NCRP, 1988; Wilkening, 1990; IAEA, 1992, 2013; Durrani and Ilic, 1997; Hosoda et al., 2007). These techniques require direct measurements and for most they turn out to be labour-intensive. In recent years, concerns over the shortcomings of some techniques have been raised since they are prone to a number of uncertainties, like back-diffusion and uncontrollable advection due to pressure differences induced by temperatures or wind. As a result, some of these techniques have been critiqued (Samuelsson and Pettersson, 1984). On a large scale, radon flux maps can be generated using indirect methods (see Table 1) given that the knowledge about parameters related to radon flux has improved. The main objective of this paper is to investigate an alternative technique for mapping radon exhalation from a gold mine tailings dump by measuring the gamma radiation from the dump fairly accurately over an area of more than 1 km2. The mine tailings dumps are considered to have potentially high radon release rate especially during uranium mining, milling and tailings disposal operations (IAEA, 1992). The technique obviates the need for collecting hundreds of samples or making many individual direct flux measurements.

2. Materials and methods 2.1. Study area The mining activities taking place in Gauteng Province, South Africa have caused millions of tons of rocks to be taken from underground to be milled and processed to extract gold. Uranium is extracted by some mines as a by-product of gold. However, most of the mines do not do this as it is often not economically viable. The uranium bearing tailings are placed on large dumps. Even if uranium is extracted, the dumps will still contain radium, the parent of radon, with a half life of 1600 years. The dumps are usually one or more kilometres in diameter in the environment. It is estimated that there are approximately 250 gold mine tailings dumps covering a total area of about 7000 ha (Lindsay et al., 2004a). These gold mine tailings dumps contain considerable amounts of radium (226Ra) and have therefore been identified as potentially large sources of radon (222Rn). Considering the above statements, the non-operational Kloof mine dump (see Fig. 1) which belongs to the Carletonville Gold Field was considered to be suitable to apply the alternative

Fig. 1. (Top) A bird's eye view of Kloof mine dump. (Bottom) Pictures of the vegetation on the mine dump. N indicates north.

technique for mapping radon exhalation rates by measuring the gamma radiation from the mine dump. Kloof mine dump lies on the north-western edge of the Witwatersrand basin, 35 km west of Johannesburg, South Africa. The Witwatersrand gold repository is estimated to have yielded over 47 000 tonnes of gold between 1886 and 2002 which represents between 33% and 40% of all gold ever produced (GDACE, 2008; Hartnady, 2009). 2.2. In-field gamma-ray measurements 2.2.1. Gamma-ray detection The MEDUSA (Multi Element Detector for Underwater Sediment Activity) detector system was used for field measurements. The

Table 1 A summary of some of the indirect methods used to generate radon flux maps. Method

Region

222

Rn flux (mBq m
2 s
1)

Reference

Radon transport model (TRACHGEO) Automatic monitoring of Gamma Dose Rate (GDR) A radon exhalation rate distribution model A model for estimating seasonal and annual radon flux densities A model utilising data from national gamma-ray aerial surveys Radon flux densities estimation using the Bayesian inversion technique Continuous and integrated measurement techniques Terrestrial gamma radiation used to calculate radon flux density A numerical model used to calculate radon exhalation rate of European soils

France Finland East Asia China Australia India Spain Netherlands & Europe Europe

52.0 27.7 18.0 29.7 24.1 33.0 11.1e25.0

Ielsch et al., 2002 Szegvary et al., 2007 Goto et al., 2008 Zhuo et al., 2008 Griffiths et al., 2010 Hirao et al., 2010 Grossi et al., 2011 Manohar et al., 2013 Lopez-Coto et al., 2013

18

J.N. Ongori et al. / Journal of Environmental Radioactivity 140 (2015) 16e24

MEDUSA system was designed and developed by the Nuclear Geophysics Division (NGD) of the Kernfysisch Versneller Instituut (KVI) at the University of Groningen (RUG) in the Netherlands in collaboration with the British Geological Survey (De Meijer, 1998). The system was initially used to measure the activity concentrations of the natural radionuclides 40K, 232Th series and 238U series in underwater environments (De Meijer et al., 1997; De Meijer, 1998; Hendriks et al., 2001; Venema and de Meijer, 2001). The MEDUSA system available at iThemba LABS, Somerset West South Africa, was utilised at the Kloof mine dump incorporates amongst others a gamma-ray detector (CsI (Na)) and a global positioning system (GPS-Garmin 76 model) device. The CsI (Na) crystal with 70 mm diameter and 150 mm length was integrated into the system mainly because of its stability in light output at variable temperatures. The detector is sensitive up to a depth of 30 cm as a result of the attenuation of the gamma rays by the soil. The MEDUSA detector system was mounted approximately 60 cm off the ground at the front of a 4  4 vehicle (see Fig. 2) and accessible parts of the mine dump were traversed at approximately 2 m s
1. The gamma-ray spectra are recorded every 2 s including the locations as logged by the GPS device (mounted directly above the g-ray detector crystal). More than 4000 data points were measured while the 4  4 vehicle was mobile. Data were also measured using the stationary mode which implies that data were acquired with the detector system mounted as in Fig. 2 and the vehicle stationary at a particular spot for a period of 30e60 min. Stationary gamma ray measurements were carried out at five spots on different days. 2.2.2. Data analyses The data collected using the MEDUSA system were synchronised and analysed using the commercially available MEDUSA software packages. The Medusa Post Analysis (MPA) software is used for the determination of activity concentration from gamma-ray spectra. MPA is a Full Spectrum Analysis (FSA) based software that takes into account almost the entire energy spectrum as opposed to the traditional Window Analysis method which concentrates on single peaks in the gamma-ray spectrum. The FSA applies a Chi-squared fitting algorithm to fit a set of “standard spectra” to the measured spectrum and a measured background spectrum. The standard spectra refer to the response of the detector per unit time in a given

Fig. 2. A photo of the iThemba LABS MEDUSA system mounted on a 4  4 vehicle approximately 60 cm off the ground.

geometry to activity concentration of 1 Bq kg
1 of a given radionuclide (De Meijer, 1998; Hendriks et al., 2001). The activity concentrations extracted using the FSA method were exported to Microsoft Excel where they were multiplied by normalisation factors. The normalisation factors were determined using the average activity concentrations of soil samples measured using the Hyper Pure Germanium detector (as described in Section 2.3 below) divided by the activity concentrations measured using the MEDUSA system. The normalisation factors used were 0.54 ± 0.05, 0.36 ± 0.03 and 0.21 ± 0.01 for 238U, 40K and 232Th, respectively. As a result, the average activity concentrations for the field measurements were computed and found to be 259 ± 75 Bq kg
1 for 40K, 309 ± 40 Bq kg
1 for 238U and 18 ± 5 Bq kg
1 for 232Th. The reported uncertainty for 40K, 238U and 232Th represent the expanded uncertainty which provides a coverage probability of approximately 90% (ISO, 1992; Gilmore, 2008). 2.3. Laboratory-based measurements 2.3.1. Soil sample preparation Thirty four soil samples from the mine dump were collected at a depth of up to 30 cm and 50 cm from five spots where the MEDUSA stationary data were captured. From each spot, a soil sample was collected directly beneath the detector system and four more samples roughly 80 cm from the detector system at north, south, east and west. The soil samples were dried in an oven (Labotec EcoTherm) overnight at 105  C to remove moisture. The moisture content (S) for each soil sample was determined; S ¼ 1
ðMdry =Mwet Þ where Mwet and Mdry is the wet and dry mass of each soil sample respectively. The moisture content was found to be between 4.3% and 14.4%. To get homogeneous dry soil samples a pestle and mortar was used to crush large grains of soil and then sieved through a mesh of 1-mm diameter holes to remove any remaining stones and organic material. The dry soil samples were transferred to Marinelli beakers and stored for at least 21 days to attain near secular equilibrium between the g-emitters; 226Ra, 214Bi and 214Pb in the uranium decay series and between 228Ac, 208Tl and 212 Pb in the thorium decay series. 2.3.2. Gamma-ray spectra analyses The low-background Hyper Pure Germanium (HPGe) g-ray detector system available at iThemba LABS was used for radiometric analysis of the dry soil samples. This detector system is a Canberra GC4520, p-type detector with 45% relative efficiency and 2 keV FWHM resolution at 1.33 MeV. The detector is encased in a 10 cm thick lead castle fitted with a 2.0 mm thick copper inner lining in order to reduce the background in the sample spectra (Newman, 2008). Standard nuclear electronics were used to process the detector signals. This detector system is energy and detection efficiency calibrated regularly with certified 232Th, 238U, 40K ore (IAEA/ RGTh-1, /RGU-1, /RGK-1, respectively). The traditional Window Analysis method was used to estimate the activity concentrations of radionuclides in the dry soil samples from the gamma-ray energies at 295, 351, 1238, 1378, 2204 keV for 238 U series, 238, 338, 911 keV for 232Th series and 1461 keV for 40K (Newman, 2008). It is important to note that the Windows Analysis method requires the associated absolute photopeak detection efficiency response which was computed by using the relative efficiency approach described by Croft (1999) and used by Newman (2008). The absolute photopeak detection efficiency was also determined using the reference materials from the International Atomic Energy Agency (IAEA), RGU-1 and RGTh-1 and potassium chloride powder KCl-99.5%. The self-absorption corrections required for samples when density is taken into account were not

J.N. Ongori et al. / Journal of Environmental Radioactivity 140 (2015) 16e24

considered because the average density of the soil samples and the IAEA reference materials were almost similar (at about 1.4 g cm
3). The activity concentration for eighteen soil samples taken from the air-soil interface up to a depth of 30 cm varies from 270 to 350 Bq kg
1 and for the other remaining sixteen soil samples taken from a depth of 40 cm up to 50 cm on the same spot, the activity concentration was found to be between 250 and 350 Bq kg
1 for the 238 U series. Overall, the weighted average activity concentrations of the thirty four soil samples were found to be 308 ± 7 Bq kg
1 for the 238 U series, 18 ± 1 Bq kg
1 for the 232Th series and 255 ± 5 Bq kg
1 for 40K. 2.3.3. Dry bulk density (rb) and soil particle density (rs) The dry bulk density is described as the total volume of the solid particles (mineral and organic) together with the pore spaces where air and water are found in the soil. The dry bulk density of two oven-dried solid cylindrical soil samples measuring about 6.4 cm in height and 8.0 cm in diameter with mass about 488 g was determined by dividing the mass (m) of the dry sample by the total volume (v) of soil and air; rb ¼ m/v. The average value for the dry bulk density found was 1.50 ± 0.16 g cm
3. The soil particle density is considered to be the density of the soil particles only. It was measured by the submersion method which is based on the difference in weight between a volume of water and that same volume with some of the water displaced by a known weight of soil (Blake, G.R. & Hartge K.H., 1986). Then the ratio of the mass of the soil (Msoil) to the volume of soil (Vsoil) gave the soil particle density; rs ¼ Msoil/Vsoil. The average value for the soil particle density calculated was 2.4 ± 0.3 g cm
3. 2.3.4. Porosity (ε) Porosity can be described as the fraction of pore space of the soil which is usually occupied by fluids (water and air). The porosity of a material was obtained when the ratio of the dry bulk density to the soil particle density was subtracted from unity; ε ¼ 1
ðrb =rs Þ. Therefore with a value of 1.50 ± 0.16 g cm
3 for the dry bulk density and 2.4 ± 0.3 g cm
3 for the soil particle density, a value of 0.38 ± 0.06 for the porosity was calculated.

2.3.5. Emanation coefficient (E) The emanation coefficient, also known as the emanation fraction or the emanating power, is defined as the fraction of radon atoms generated which escape the solid phase in which they are formed and enter the pore space and are free to diffuse through the bulk medium. The emanation coefficient was determined using a sealable glass jar, an Electret ion Chamber (S-Chamber), a shortterm electret and the solid cylindrical Kloof soil sample described in Section 2.3.3 above, as shown in Fig. 3. The radium concentration (CRa) for the soil sample and the radon escaping from 226Ra bearing soil are required to estimate the radon emanation coefficient. An average radium concentration value of 308 Bq kg
1 was used while the emanating part of radon was determined using the Radon Emanating 226Ra Concentration , 1995; (RnERaC) method (Kotrappa, 1994; Rad Elec Inc., 1994; Colle Kotrappa and Stieff, 2009). The ratio of the RnERaC to the CRa gives the radon emanation coefficient that is, E ¼ RnERaC/CRa. Six other smaller samples from different parts of the mine dump were also measured using this technique. An average radon emanation coefficient value of 0.28 ± 0.03 was estimated using this technique. 3. Results and discussions 3.1. Radon exhalation results using different techniques Physical factors such as the 226Ra activity concentration, soil porosity, soil moisture; atmospheric pressure, wind speed, rainfall and temperature influence the radon exhalation from the soil surface interface. This study was done during a dry period when the temperature was between 24.7  C and 31.8  C, the highest wind speed recorded was 2.4 m/s using a Kestrel 4000 pocket-sized weather station and the moisture content for soil samples collected at up to 30 cm from the ground-air surface was less than 10%. Our aim in this work was primarily to investigate measurement methods to obtain a good estimate of the radon exhalation based on the radium concentration measurements and not to measure the annual radon exhalation. The first method used to obtain the radon flux used the in-situ gamma radiation measurements by using the approximation, IAEA (1992, 2013).

pffiffiffiffiffiffiffi F ¼ CRa rb E l D ¼ CRa rb E[l

Fig. 3. A set-up for measuring the emanation coefficient of radon using a solid cylindrical Kloof soil sample in a sealable glass jar.

19

(1)

where, CRa(in Bq kg
1) is the radium concentration in the soil, rb(in kg m
3) is the dry bulk density of the soil, E (dimensionless) is the radon emanation coefficient,pand ffiffiffiffiffiffiffiffiffi[ is the diffusion length for radon in the soil which is equal to D=l where l (s
1) is the radon decay constant (2.1  10
6 s
1), and D (in m2 s
1) is the effective radon diffusion coefficient. This equation is derived in Appendix Ae see Eq. (A.9). The MEDUSA detector (or any other in-situ gamma detector system) provides an efficient method to obtain the radium concentration in the soil as described in Sections 2.2.1 and 2.3.2. All parameters in Eq. (1) are known except the effective radon diffusion coefficient, or equivalently, the diffusion length. The effective radon diffusion coefficient of a material is defined as the ratio of the effective flux density of radon activity across the pore area, to the gradient of the radon activity concentration in the pore or interstitial space. One common way of estimating the effective radon diffusion coefficient value is the phenomenological formula of Rogers and Nielson (1991). The moisture and porosity values found in this work, gives a diffusion coefficient that varies between 2.8  10
6 to 8.5  10
7 m s
2 In our work we preferred to use an average diffusion length value of around 0.4 m (corresponding to a diffusion coefficient of 3.4  10
7 m s
2) that was extracted from

20

J.N. Ongori et al. / Journal of Environmental Radioactivity 140 (2015) 16e24

the radon soil gas concentration depth profile measured at five points on the dump using the radon-in-air monitor, the RAD7 (Durridge, 2000) and by modelling (Ongori, 2013). The uncertainty in this extraction process and the differences at different points leads to uncertainties in the diffusion which is the major contribution to the uncertainty in the radon exhalation determination when the IAEA formula is used. Fig. 4 presents an interpolated radon flux map obtained using Eq. (1) and a diffusion length of 0.4 m which gives an average radon exhalation value of 0.12 ± 0.02 Bq m
2 s
1. The interpolated radon flux map of the mine dump was generated using the Natural Neighbour gridding method of the Golden Software Surfer® 8 (Golden Software, 2002). In this study, an alternative method to obtain radon flux values was investigated. The two g-ray detector systems, that is the HPGe and the MEDUSA, were used to estimate the 238U activity concentration via radon progenies (214Bi and 214Pb). The in-situ measurements by the MEDUSA system do not account to the escape of radon; hence the activity concentration of the 238U is expectedly lower than it should be after radon has escaped. Therefore, one can assume that this measurement obtained reflects the effective radon loss due to exhalation over the last hour or so, considering the halflives of 214Bi and 214Pb to be of the order of 20 min. In the field, both the 40K and 232Th activity concentrations are not affected by radon exhalation so that their concentrations are independent of the radon loss. In the decay series of 232Th, thoron (220Rn) has too short a half-life (55.6 s) for much of it to escape before decay, hence this series is considered to be near equilibrium. By contrast, in the 238U decay series the radon (222Rn) has a relative long half-life (3.82 d), which allows radon to diffuse to the surface and possibly escape before it decays. As a consequence of radon loss an expression for mapping the exhalation of radon from a particular location in the mine dump can be derived as follows. If one considers, I, as the counts as recorded by MEDUSA system, and compare those with the counts that would have been recorded by the system if there was no radon escape, IH, then the radon flux will clearly be proportional to the difference between these two values,

    I Jf IH
I ¼ IH $ 1
IH

(2)

The counts in the field are known whereas the equilibrium counts are not measured. However, the follow-up measurements in the laboratory setting with the sealed samples that were left to attain equilibrium, these equilibrium counts can be determined. The activity concentration of 40K, 238U series and 232Th series are extracted from the measured data in the field using the Full Spectrum Analysis (FSA) method (De Meijer, 1998; Hendriks et al., 2001). It should be noted that if the standard spectra used during the FSA is perfect for 238U and 40K, then it can be shown (see Eq. (A.19) in Appendix A) that,

238

  Em[ UH ¼ 238 UH $ 1
1 þ m[

(3)

while for the potassium 40

KM ¼

40

KH

(4)

where, 238 UM represents uranium disequilibrium activity concentration determined in the field, 238 UH the uranium equilibrium activity concentration determined in the laboratory, 40 KM the potassium activity concentration from the field measurements, 40 KH the potassium activity concentration from the laboratory

measurements, E the emanation coefficient, [ the radon diffusion length and m the gamma ray attenuation coefficient. If the standard spectra are not absolutely normalised but all require a similar normalisation factor (Nf), then 238

40

UM ¼ 238 UH 

KM ¼

  1 Em[ $ 1
Nf 1 þ m[

(5)

1  40 KH Nf

(6)

Combining Eqs. (5) and (6) results in 238 U M 40 K M

¼

238 U H 40 K H

  Em[ $ 1
1 þ m[

(7)

and then,

1


238 U M 238 U H

=40 KM Em[ ¼ 1 þ m[ =40 KH

(8)

In this study, it was found that due to the normalisation of the standard spectra, the individual normalisation factors for 238U series, 232Th series and 40K varies. Thus, 238

40

UM ¼ 238 UH 

KM ¼ 40 KH 

  1 Em[ $ 1
1 þ m[ NfU

(9)

1 NfK

(10)

where, NfU and NfK represent 238U and 40K normalisation factors respectively. Combining Eq. (9) with (10) results in

1


238 U M 238 U H

U =40 KM Nf Em[ $ ¼ 1 þ m[ =40 KH NfK

(11)

Now, since the counts are proportional to the radon concentrations, Eqs. (2), (3) and (11) can be rewritten (Ongori, 2013),

" Jf

238

UH $ 1


U =40 KM Nf $ K 238 U =40 K H H Nf

238 U

M

!# (12)

In Eq. (12) above, the 40K values are used in order to compensate for the normalisation differences. The corrections suggested by De Groot (2009) for moisture corrections can also be incorporated in Eq. (12). This equation presents an alternative method of determining radon exhalation using ratios by combining field and laboratory measurements. This technique is herein referred to as MEDUSA Laboratory Technique (MELT). The implementation of the MELT (Eq. (12)) has the advantage that the potassium measurement at each point is used as an additional input in the results. This would help to take differences in the absorption of gamma rays at different points on the dump into account. The simplest method is to apply to the final normalised MEDUSA activity concentrations for 238U, 232 Th and 40K resulting in unnormalised radon flux values. The normalised radon flux values were obtained by multiplying the unnormalised radon flux values by an appropriate radon flux normalisation constant. An average radon flux normalisation constant of 6.4  10
4 kg m
2 s
1 was found when the radon flux values obtained using Eq. (1) were divided by the unnormalised radon flux values calculated using Eq. (12).

J.N. Ongori et al. / Journal of Environmental Radioactivity 140 (2015) 16e24

The results herein are presented in interpolated tri-colour map (Fig. 5). The interpolated radon flux map was generated as mentioned earlier. The blue, yellow and red colours (in web version) indicate low, average and high radon flux values, respectively. The average radon flux values were found to be 0.12 ± 0.02 Bq m
2 s
1 without including the uncertainty contribution due to the diffusion length. Another practical method to calculate the exhalation is to use the measurement of radon exhalation measurements using a traditional, labour intensive radon exhalation method at a few spots and then use those points to normalise the calculations at all points. The E-PERM flux measurement system was used at the 5 stationary spots of the MEDUSA system. The E-PERM systems are used globally for measuring short-term, long-term radon/thoron concentrations in air, radon in water, radon in soil and radon flux from surfaces and mill tailings (Kotrappa, 1988, 1994, 1996, 2000). The radon flux values were extracted from the modified H-chambers loaded with short-term electrets (Kotrappa, 1990; Rad Elec Inc., 1994). The radon flux values extracted were between 0.06 and 0.08 Bq m
2 s
1, slightly lower than the flux values computed using the IAEA method. These lower values could be attributed to the systematic uncertainties associated with the devices (adapted Hchambers loaded with short-term electrets) given that they are placed on the tailings surface but then cover only a small area of about 215 cm2. Also air-tightness around the base of the chamber is always a problem on hard uneven surfaces. If these results are assumed to be correct, then the diffusion length used to calculate the fluxes in Fig. 4 is higher than predicted, hence the results would have to be scaled down by a factor of around 20%. Finally, Manatunge (2002), used an E-PERM dynamic system to compute radon exhaling from the mine tailings surface. The system monitored radon releases for a period of 24 h. The dynamic system set-up consisted of the E-PERM system with S-chambers loaded with long term electrets. The radon flux values ranged between 0.04 and 0.20 Bq m
2 s
1 with an average value of 0.12 ± 0.05 Bq m
2 s
1, in close agreement with this study. This method would work well if the diffusion length is fairly constant over the mine dump and if the major change is a difference in radium concentration. That is probably a good approximation in the present case since the dump is a result of a consistent treatment in the extraction process whereas the radium concentration is not constant for different parts of the mine.

21

Fig. 5. Interpolated map generated showing radon flux values after utilising Eq. (12).

4. Conclusions The measurement of radon flux from the large mine tailings dumps present a major challenge. The size of the dumps leads to very time consuming and labour intensive procedures if the usual methods (IAEA, 1992, 2013) are used. In this work we have investigated the radon exhalation from the Kloof mine tailings dump by measuring the activity concentrations of the primordial radionuclides namely; the 238 U series, the 232Th series and 40K. This was realised by utilising the mobile MEDUSA g-ray detector system which has a high detection efficiency. The MEDUSA Post Analysis software which uses the Full Spectrum Analysis method was used to extract the activity concentrations over the whole dump within a relatively short measurement time. The flux was then calculated using the IAEA formula, Eq. (1) (IAEA, 1992, 2013) or the MELT, Eq. (12) using the measured radium concentrations. In this study it was possible with the MEDUSA technology to extract the 214Bi/214Pb activity concentration which is used as a proxy for radium concentration in the uranium series for all locations on the mine dump traversed. After normalising the activity concentrations measured by the MEDUSA system with laboratory measurements of soil samples it was possible to generate radon flux maps for the mine dump. The other parameters that are required in the IAEA formula were measured and are expected to be fairly constant on the mine dump. An alternative technique using the ratio of the 238U series relative to the 40K values at each point, was also investigated. This technique provided very similar results to the IAEA formula, but requires a normalisation factor which can be provided by the IAEA value or any other traditional method used at a few points. The techniques provide a quick means through which the radon flux values can be assessed because many points are analysed for a particular area covered and it also reduces the need for collecting several hundred samples in order to be measured in the laboratory or making many individual direct flux measurements. Only a few samples from the area of study are needed for laboratory measurements to be used for calibration purposes.

Acknowledgements

Fig. 4. An interpolated map showing radon flux results derived from the IAEA formula.

Funding from the ESKOM TESP programme, the NRF (grant 41900020) and the NNR is acknowledged. Discussions with Prof

22

J.N. Ongori et al. / Journal of Environmental Radioactivity 140 (2015) 16e24

Rob de Meijer on the work in Appendix A are acknowledged, as well as support on the measurements from Israel Hlatshwayo and Sive Noncolela.

C∞ ¼

E C r ε Ra b

(A.8)

The exhalation per unit of surface is usually defined as (IAEA, 2013)

Appendix A Exhalation of radon from a mine tailings dump by measuring the gamma radiation from the dump

J ¼ εD

Consider a semi infinite homogeneous layer of consolidated dry sand. The sand with specific density of rs (which can be approximated fairly accurately by the density of sand made of SiO2 that equals 2.6 g cm
3) contains CRa Bq kg
1 of 238U (as measured in the HPGe at iThemba LABS) which is in secular equilibrium with 226Ra. A fraction ε per unit volume is filled with air (porosity ε), hence the grains occupy (1
ε) of the volume. Ignoring the mass of the air, the bulk density of the layer is given by

J¼

rb ¼ ð1
εÞrs

(A.1)

For every m of layer, CRarb Bq of Ra is present. (rb was measured at the University of the Western Cape based on the mass of material brought from the dump). Due to nuclear decay, 226Ra will turn into 222Rn (radon). A fraction, E, will leave the grains and enter the air-filled void space. Since each Bq of 226Ra produces one Rn atom, per second, the number of radon atoms entering the void space per second per m3 equals 3

226

NRn ¼ ECRa rb

(A.2)

In secular equilibrium, the total number of radon atoms in the pore space per m3 of layer follows from the condition that production equals decay: 3 ECRa rb ¼ lRn N ¼ Aair Rn per m of layer

vC at z ¼ 0 vz

εD ∞ ED C ¼ C r [ [ Ra b

(A.9)

from Eq. (A.7). This is similar to the IAEA equation (Eq. (1)) and the discussion above is essentially a derivation of this equation. Radon in the soil decays further into the gamma-ray producing nuclei 214Pb and 214Bi which are assumed to be in equilibrium with respect to each other and to the 222Rn concentration. At depth z, the volumetric concentration of the A ¼ 214 nuclei follows from the radon concentration in the grains, which is independent of z, and the z dependent radon concentration in the air filled pore space

    grain air C 214 z ¼ 1
ε CRn þ εCRn     ¼ 1
E CRa rb þ ECRa rb 1
e
z=[   h i C 214 z ¼ CRa rb 1
Ee
z=[

(A.10)

Consider again a semi infinite layer of homogeneous material with density r and activity concentration of C (Bq/kg). Assume that the gamma-rays are emitted isotropically.

(A.3)

The activity concentration of Rn in the air-filled pore space becomes: air CRn ¼

ECRa rb Bq ε m3

(A.4)

In the grain filled part of the layer, the radon production is (1
E)CRarb atoms per second and hence the volumetric concentration in the grains becomes grain

CRn

¼

ð1
EÞCRa rb Bq ð1
εÞ m3

(A.5)

Note: The total radon concentration under secular equilibrium conditions (only radon is disappearing by nuclear decay)

    grain total air CRn ¼ εCRn þ 1
ε CRn ¼ ECRa rb þ 1
E CRa rb ¼ CRa rb

(A.6) as to be expected. So at great depth, we know the partitioning between radon in the grains and radon in the pore space. Near the surface, radon will diffuse to the surface and exhale into the air. We assume that the radon concentration profile for the air-filled pore space can be written as (IAEA, 2013)

    C z ¼ C ∞ 1
e
z=[

(A.7)

where C∞ is the volumetricpconcentration of radon in the pore ffiffiffiffiffiffiffiffiffiffiffiffiffi space at large depth and [ ¼ D=lRn is the diffusion length with D the effective diffusion constant for radon and lRn is the radon decay constant. According to Eq. (A.4):

Fig. A.1: A sketch showing the MEDUSA detector above the ground detecting gammarays emitted from the ground.

We simplify the picture by assuming that all gamma-rays are emitted perpendicular to the surface. On top of the surface we place a detector with surface Adet and efficiency 1. The correction for the real efficiency will be made later. The number of gamma-rays emitted towards the detector per unit of volume is set to I0, hence for a layer with thickness dz the number is given by

J.N. Ongori et al. / Journal of Environmental Radioactivity 140 (2015) 16e24

1 I0 dz ¼ Crdz 2

(A.11)

At the detector, due to absorption one has a count rate due to this layer

  I z dz ¼ Adet I0 e
ðm=rÞ r z dz

(A.12)

where m/r is the mass-attenuation coefficient and Adet is the area of the detector. If C(z) is constant, the number of gamma-rays at the detector follows from the integration:

I tot ¼

1 A C r 2 det Ra b

Z∞

e
ðm=rÞ r z dz ¼

1 CRa rb A 2 det m

(A.13)

0

This result can be applied to the case for radon. So in the case that no radon escapes from the soil, we may write

IH ¼

1 CRa rb A 2 det m

(A.14)

In case of radon exhalation by the surface, I0 is no longer a constant and Eq. (A.11) becomes

    A I0 z dz ¼ det C z rdz 2

(A.15)

where C(z) is given by Eq. (A.10). This results in

    A I0 z dz ¼ det CRa rb 1
Ee
z=[ dz 2

(A.16)

The total number of gamma counts in the detector then becomes, if Eq. (A.12) is used.

Z∞ I¼

2 ∞ 1 3   Z  Adet
ðm=rÞ r z
ðmþ1=[Þ z A 5 4 e C r I0 z dz ¼
Ee dz 2 Ra b

0

0

(A.16a) I¼

  Adet 1 E CRa rb
m m þ 1=[ 2

(A.17)

The first term in the bracket results from the radon in the grains and the second term from the decay of radon in the pore space. From Eqs. (A.17) and (A.14),

I IH

 ¼

1


Em[ m[ þ 1

 (A.18)

This equation implies that if E is measured in the laboratory, the value of [ (and hence the exhalation rate) follows from the ratio I/IH. Since I is linearly dependent on the concentration of radium, CRa,

I IH

¼

M CRa H CRa

(A.19)

H values The ratio in Eq. (A.19) can be obtained as follows; the CRa are found from the sample measurements in the laboratory and M Þ from the measurements in the field. ðCRa

References Åkerblom, G., Mellander, H., 1997. Geology and radon. In: Durrani, S.A., Ili c, R. (Eds.), Radon Measurements by Etched Track Detectors: Applications in Radiation

23

Protection, Earth Sciences, and the Environment. World Scientific Publishing Co. Pty. Ltd., pp. 21e49. Aldenkamp, F.J., de Meijer, R.J., Put, L.W., Stoop, P., 1992. An assessment of in situ radon exhalation measurement, and the relation between free and bound exhalation rates. Radiat. Prot. Dosim. 45 (1e4), 449e453. Andersen, C.E., 2000. Radon Transport Modelling: User's Guide to RnMod3d. Risø National Laboratory, Roskilde, Denmark. Andersen, C.E., 2001. Numerical modelling of radon-222 entry into houses: an outline of techniques and results. Sci. Total Environ. 272, 33e42. Antonopoulos-Domis, M., Clouvas, A., Alifrangis, D., 2009. Experimental and theoretical study of radon distribution in soil. Health Phys. 97 (4 (October)), 322e331. Arafa, W., 2004. Specific activity and hazards of granite samples collected from the Eastern Desert of Egypt. J. Environ. Radioact. 75, 315e327. BEIR VI (Committee on Health Risks of Exposure to Radon), 1999. Health Effects of Exposure to Radon: BEIR VI (Free Executive Summary) National Research Council. Bigu, J., 1984. Study of radon concentration, surface radon flux and other radiation variables from uranium mine tailings areas. Uranium 1, 257e277. Blake, G.R., Hartge, K.H., 1986. Particle density. In: Klute, A. (Ed.), Methods of Soil Analysis. Part 1-Physical and Mineralogical Methods, second ed. American Society of Agronomy, Madison WI. Chen, C.J., Weng, P.S., Chu, T.C., 1993. Radon exhalation rate from various building materials. Health Phys. 64 (6 (June)), 613e619. Clements, W.E., Wilkening, M.H., 1974. Atmospheric pressure effects on 222Rn transport across the earth-air interface. J. Geophys. Res. 79 (33), 5025e5029. , R., Kotrappa, P., Hutchinson, J.M.R., 1995. Calibration of electret-based integral Colle radon monitors using NIST polyethylene-encapsulated 226Ra/222Rn emanation (PERE) standards. J. Res. Natl. Inst. Stand. Technol. 100 (6, (NovembereDecember)), 629e639. Crank, J., 1975. The Mathematics of Diffusion. Clarendon Press, Oxford. Croft, S., Hutchinson, I.G., 1999. The measurement of U, Th and K concentrations in building materials. Appl. Radiat. Isot. 51, 483e492. De Jong, P., et al., 1996. The effect of the composition and production process of concrete on the 222Rn exhalation rate. Environ. Int. 22 (Suppl. 1), S287eS293. De Meijer, R.J., 1998. Heavy minerals: from ‘Edelstein’ to Einstein. J. Geochem. Explor. 62, 81e103. De Meijer, R.J., et al., 1997. Improved and new uses of natural radioactivity in mineral exploration and processing. Explor. Min. Geol. 6 (1), 105e117. Durrani, S.A., Ili c, R. (Eds.), 1997. Radon Measurements by Etched Track Detectors: Applications in Radiation Protection, Earth Sciences, and the Environment.. World Scientific Publishing Co. Pte. Ltd, London. Durridge Company Inc, 2000. Reference Manual Version 6.0.1, RAD7™ Electronic Radon Detector. Escobar, V.G., Tome, F.V., Lozano, J.C., 1999. Procedures for the determination of 222 Rn exhalation and effective 226Ra activity in soil samples. Appl. Radiat. Isot. 50, 1039e1047. Fleischer, R.L., 1980. Radon flux from the earth: methods of measurement by the nuclear track technique. J. Geophys. Res. 85 (C12), 7553e7556. GDACE, 2008. Mining and Environmental Impact Guide. Produced by staff of Digby Wells and Associates, Growth Lab and the Council for Geoscience for the Gauteng Department of Agriculture, Environment and Conservation. Gilmore, G.R., 2008. Practical Gamma-ray Spectrometry. John Wiley & Sons, Ltd, England. Golden Software, Inc, 2002. Surfer8 User's Guide, Contouring and 3D Surface Mapping for Scientist and Engineers, Colorado, USA. Goto, M., et al., 2008. Estimation of global radon exhalation rate distribution. In: Paschoa, A.S. (Ed.), The Natural Radiation Environment e 8th International Symposium. Graustein, W.C., Turekian, K.K., 1990. Radon fluxes from soils to the atmosphere measured by 210Pbe226Ra disequilibrium in soils. Geophys. Res. Lett.17 (6), 841e844. Griffiths, A.D., Zahorowski, W., Element, A., Werczynski, S., 2010. A map of radon flux at the Australian land surface. Atmos. Chem. Phys. 10, 8969e8982. De Groot, A.V., van der Graaf, E.R., de Meijer, R.J., Maucec, M., 2009. Sensitivity of insitu g-ray spectra to soil density and water content. Nucl. Instrum. Methods Phys. Res. A 600, 519e523. Grossi, C., et al., 2011. Inter-comparison of different direct and indirect methods to determine radon flux from soil. Radiat. Meas. 46, 112e118. Hartnady, C.J.H., 2009. South Africa's gold production and reserves. South Afr. J. Sci. 105 (9/10), 328e329. Hendriks, P.H.G.M., Limburg, J., de Meijer, R.J., 2001. Full-spectrum analysis of natural g-ray spectra. J. Environ. Radioact. 53, 365e380. Hirao, S., Yamazawa, H., Moriizumi, J., 2010. Inverse modeling of Asian 222Rn flux using surface air 222Rn concentration. J. Environ. Radioact. 101, 974e984. Hosoda, M., et al., 2007. Effect of soil moisture content on radon and thoron exhalation. J. Nucl. Sci. Technol. 44 (4), 664e672. IAEA, 1992. Measurement and Calculation of Radon Release from Uranium Mill Tailings. Technical Report Series No. 333. International Atomic Energy Agency, Vienna, Austria. IAEA, 2013. Measurement and Calculation of Radon Releases from NORM Residues. Technical Report Series No. 474. International Atomic Energy Agency, Vienna, Austria. Ielsch, G., Ferry, C., Tymen, G., Robe, M.C., 2002. Study of a predictive methodology for quantification and mapping of the radon-222 exhalation rate. J. Environ. Radioact. 63, 15e33.

24

J.N. Ongori et al. / Journal of Environmental Radioactivity 140 (2015) 16e24

Ingersoll, J.G., 1983. A survey of radionuclide contents and radon emanation rates in building materials used in the U.S. Health Phys. 45 (2 (August)), 363e368. ISO, 1992. Guide to the Expression of Uncertainty in Measurement. International Organization for Standardization, Switzerland. Kohl, T., Medici, F., Rybach, L., 1994. Numerical simulation of radon transport from subsurface to buildings. J. Appl. Geophys. 31, 145e152. Kotrappa, P., 2000. Review of E-PERM passive integrating electret ionization chambers for measuring radon in air, thoron in air, radon in water and thoron flux from surfaces and mill tailings. In: International Radon Symposium. Kotrappa, P., Stieff, F., 2009. Radon exhalation rates from building materials using Electret Ion Chamber radon monitors in accumulators. Health Phys. 97, 163e166. Kotrappa, P., Dempsey, J.C., Hickey, J.R., Stieff, L.R., 1988. An electret passive environmental 222Rn monitor based on ionization measurement. Health Phys. 54 (1 (January)), 47e56. Kotrappa, P., Dempsey, J.C., Ramsey, W., Stieff, L.R., 1990. A practical E-PERMTM (Electret Passive Environmental Radon Monitor) system for indoor 222Rn measurement. Health Phys. 58 (4 (April)), 461e467. Kotrappa, P., et al., 1994. Measurement of the radon surface flux from undisturbed soil using electret ion chambers. In: International Radon Symposium III. Kotrappa, P., Stieff, L.R., Bigu, J., 1996. Passive E-PERM radon flux monitors for measuring undisturbed radon flux from the ground. In: International Radon Symposium II. Kovler, K., Perevalov, A., Steiner, V., Metzger, L.A., 2005. Radon exhalation of cementitious materials made with coal fly ash: part 1-scientific background and testing of the cement and fly ash emanation. J. Environ. Radioact. 82, 321e334. Lawrence, C.E., et al., 2009. Radon-222 exhalation from open ground on and around a uranium mine in the wet-dry tropics. J. Environ. Radioact. 100, 1e8. Lindsay, R., et al., 2004a. Monitoring the radon flux from gold-mine dumps by g-ray mapping. Nucl. Instrum. Methods Phys. Res. B 213, 775e778. Lindsay, R., et al., 2004b. Measurement of radon exhalation from a gold-mine tailings dam by g-ray mapping. Radiat. Phys. Chem. 71, 797e798. Lindsay, R., Newman, R.T., Speelman, W.J., 2008. A study of airborne radon levels in Paarl houses (South Africa) and associated source terms, using electret ion chambers and gamma-ray spectrometry. Appl. Radiat. Isot. 66, 1611e1614. pez-Cotoa, I., Mas, J.L., Bolivar, J.P., 2013. A 40-year retrospective European radon Lo flux inventory including climatological variability. Atmos. Environ. 73, 22e33. Loureiro, C.O., Abriola, M.L., Martin, J.E., Sextro, R.G., 1990. Three-Dimensional simulation of radon transport into houses with basements under constant negative pressure. Environ. Sci. Technol. 24 (9), 1338e1348. Manatunge, S., 2002. The Impact of Atmospheric Pathways on Public Exposure. MSc thesis. University of North West, South Africa. Manohar, S.N., Meijer, H.A.J., Herber, M.A., 2013. Radon flux maps for the Netherlands and Europe using terrestrial gamma radiation derived from soil radionuclides. Atmos. Environ. 81, 399e412. Mudd, G.M., 2008a. Radon releases from Australian uranium mining and milling projects: assessing the UNSCEAR approach. J. Environ. Radioact. 99, 288e315. Mudd, G.M., 2008b. Radon sources and impacts: a review of mining and nonmining issues. Rev. Environ. Sci. Biotechnol. 7, 325e353.

Nazaroff, W.W., Moed, B.A., Sextro, R.G., 1988. Soil as a source of indoor radon: generation, migration, and entry. In: Nazaroff, W.W., Nero, A.V. (Eds.), Radon and its Decay Products in Indoor Air. John Wiley and Sons, Inc., pp. 57e112. NCRP, 1988. National Council on Radiation Protection and Measurements. Measurement of radon and radon daughters in air, NCRP Report No. 97. National Council on Radiation protection and Measurements, Bethesda, Maryland. Newman, R.T., et al., 2008. Determination of soil, sand and ore primordial radionuclide concentrations by full-spectrum analyses of high-purity germanium detector spectra. Appl. Radiat. Isot. 66, 855e859. Ongori, J.N., 2013. In-situ Measurements and Calculation of Radon Concentration and Exhalation at a Mine Dump Site. PhD thesis. University of the Western Cape, South Africa. Owczarski, P.C., Holford, D.J., Gee, G.W., Freeman, H.D., Burk, K.W., 1990. Radon Transport from the Subsurface: the Role of Certain Boundary Conditions at Subsurface/Environment Boundaries. Pacific Northwest Laboratory. Rad Elec Inc, 1994. Reference Manual, Radon and Radiation Measurements. Rogers, V.C., Nielson, K.K., 1991. Multiphase radon generation and transport in porous materials. Health Phys. 60 (6 (June)), 807e815. Sahoo, B.K., 2010. Radon exhalation studies in an Indian uranium tailings pile. Radiat. Meas. 45, 237e241. Samuelsson, C., Pettersson, H., 1984. Exhalation of 222Rn from porous materials. Radiat. Prot. Dosim. 7 (1e4), 95e100. Schery, S.D., Whittlestone, S., Hill, S.E., 1989. The flux of radon and thoron from Australian soils. J. Geophys. Res. 94 (D6 (June)), 8567e8576. Schubert, M., Schulz, H., 2002. Diurnal radon variations in the upper soil layers and at the soil-air interface related to meteorological parameters. Health Phys. 83 (1 (July)), 91e96. Szegvary, T., Leuenberger, M.C., Conen, F., 2007. Predicting terrestrial 222Rn flux using gamma dose rate as a proxy. Atmos. Chem. Phys. 7, 2789e2795. Tsela, A.S., Brits, R.J.N., 1998. A Theoretical Study of the Effect of Various Parameters on Radon Exhalation from Mine Tailings. SARPA Conference. UNSCEAR, 1982. Ionizing Radiation: Sources and Biological Effects Annex D. United Scientific Committee on the Effects of Atomic Radiation. 1982 Report to the General Assembly, with annexes. United Nations publications. Van der Spoel, W.H., van der Graaf, E.R., de Meijer, R.J., 1997. Diffusive transport of radon in a homogeneous column of dry sand. Health Phys. 72 (5 (May)), 766e778. Van der Spoel, W.H., van der Graaf, E.R., de Meijer, R.J., 1998. Foil coverage of a crawl-space floor: measurements and modelling of radon entry. Health Phys. 74 (5 (May)), 581e593. Van der Spoel, W.H., van der Graaf, E.R., de Meijer, R.J., 1999. Diffusive transport of radon in a column of moisturized sand. Health Phys. 77 (2 (August)), 163e177. Venema, L.B., de Meijer, R.J., 2001. Natural radionuclides as tracers of the dispersal of dredge spoil dumped at sea. J. Environ. Radioact. 55, 221e239. WHO, 2009. World Health Organization. WHO Handbook on Indoor Radon: a Public Health Perspective. WHO press, Geneva. Wilkening, M., 1990. Radon in the Environment. Elsevier Science Publishers B.V., Amsterdam, The Netherlands. Zhuo, W., Guo, Q., Chen, B., Cheng, G., 2008. Estimating the amount and distribution of radon flux density from the soil surface in China. J. Environ. Radioact. 99, 1143e1148.