# EVALUATION OF THE IMPACT OF THE DOMESTIC ENVIRONMENT ON THE POPULATION EXPOSURE TO RADON DAUGHTERS

**Project ID**: BI6*0112

**Financé au titre de**:

## EVALUATION OF THE IMPACT OF THE DOMESTIC ENVIRONMENT ON THE POPULATION EXPOSURE TO RADON DAUGHTERS

**De**1985-01-01

**à**1989-12-31

## Détails concernant le projet

### Coût total:

Non disponible### Contribution de l'UE:

Non disponible### Coordonné à/au(x)/en:

Belgium### Sujet(s):

### Régime de financement:

CSC - Cost-sharing contracts## Objectif

1. SYTEMATIC ANALYSIS OF THE RADON DAUGHTER EQUILIBRIUM IN HOUSES.

2. STUDY OF THE BEHAVIOUR AND NATURE OF RADON DAUGHTER IONS AND CLUSTERS.

3. INVESTIGATION OF THE RADON EXHALATION FROM BUILDING MATERIALS AND SOILS.

Apparatus for measuring the low concentrations of radon decay products, typically found indoors, involves drawing air at a constant rate through a filter and counting the activity by means of alpha spectroscopy during the sampling, and during a decay interval. The timing has been optimised by minimising the mean value of the minimum measurable radon daughter concentrations (MMC), defined as the concentration at which the relative standard deviation in the measurement due to counting statistics is 20%, assuming the flow rate and the detection efficiency to be 28 1 min{-1} and 0.127 respectively. It was found that a long sampling time and delay time longer than the minimum time needed to transfer the spectrum to a storage medium is favourable. The same kind of optimisation has been performed for the thoron daughters. In the calculations the sampling period was set at 30 min and the first decay time interval is started after the decay of the radon daughters (270 min). For a total measurement period of 16, the optimised MMC of lead-212 and bismuth-212 are 0.02 Bq m{-3} and 60 Bq m{-3} (270 to 370 min, 540 to 960 min) respectively. Better results for bismuth-212 are obtained with only 1 decay time interval and an estimation of the ratio lead-212 to bismuth-212 out of the removal processes (ventilation and deposition of the attached thoron daughters). The results of this analysis show that the influence of the removal rate is rather small and that a longer counting time leads to a substantial improvement in measurement precision. For a total measurement period of 16 (counting interval: 270 to 960 min), the MMC of lead-212 is about 0.014 Bq m{-3}.

Measuring the polonium-218 and polonium-214 alpha decay during sampling improves significantly the precision, in particular for polonium-218 and thus for the unattached fraction. However, at the same time part of the unattached activity is lost in the complex sampling head geometry. This loss has been determined in the 1 m{3} box as a function of the flow rate. At 28 I min{-1}, our standard flow rate in dwellings, the loss of the unattached daughters equals 0.22+/-0.02.

The numerical code AERO1A has been extended to describe the size distributions of the 3 relevant short lived decay products of radon polonium-218 (RaA), lead-214 (RaB), and polonium-214 (RaC'). Most of the relevant processes were taken into account, ie the physical ones (ventilation, attachment, deposition of the attached and unattached fraction and recoil) as well as the physicochemical ones (clustering and growth and neutralisation of the free radioactive ions); patterns of air circulation and degrees of turbulence, which may also affect the deposition of airborne material, are not considered. Application of the model in some selected conditions showed that the active size conditions showed that the active size distributions and the amount of airborne radioactivity is largely affected, not only by the aerosol loading of the atmosphere, but also by its chemical composition, as well as by the dielectric/conductive characteristics of the surfaces in the room. 2 extreme conditions were considered: a steel room for laboratory conditions and a normal living room, resulting in quite different size distributions. It was concluded that clustering and growth of the free radioactive ions are not too relevant in domestic environments because of important competitive removal processes like neutralisation by charge transfer and or attachment and or deposition by electrophoresis. On the other hand the charge properties of the radon decay products largely determine the amount of airborne activity in the domestic environment.

The model calculations give a better understanding of the many factors that determine the behaviour of radon decay products, and they explain why such a large range of values is being found of diffusion coefficients of the unattached fractions, of equilibrium constants, plate out rates etc.

The knowledge of radon exhalation rates for building materials is of fundamental importance in the assessment of possible radon daughters concentrations in dwellings. Measurements of exhalation rates are usually performed using small samples, often in closed can set ups. Since these laboratory conditions are very different from the actual walls, the mathematical problem of 3-dimensional steady state radon diffusion in rectangular samples was solved, and the relationship between exhalation from such samples and the 1-dimensional exhalation from walls was established. This implies that one has to solve the well known differential equation, in 3-dimensions, with the appropriate boundary conditions. The analytical solution of the problem is found by means of a triple fourier series.

Of more importance however is the exhalation rate. The results obtained from an exhalation measurement on a sample can be interpreted directly in terms of radon exhalation from walls (the latter being a 1-dimensional process) only if the calculated 1-dimensional exhalation rate from the sample agrees with the 3-dimensional experimental results. This ratio between 1- and 3-dimensional exhalation rates is found, and the discrepancy with increasing thickness is obvious. The exact analytical 3-dimensional solution is however very tedious, and therefore a modified 1-dimensional formalism was introduced. It is evident that the difference between 1- and 3 dimensional results with increasing sample thickness is due to the fact that in the 1-dimensional formalism all radon has to exhale through the x=+/- 1x planes, whereas in reality for an increasing part of the sample the radon diffuses to another more adjacent plane. Therefore and equivalent sample geometry was defined with a reduced thickness to take this effect into account, and then a pure 1-dimensional model was applied to this reshaped sample. It is evident that this reduced thickness 1'x will be chosen as 4 times the mean value over the s ample volume of the closest distance from a point in the sample to the surface.The exhalation E1dim, which will occur when using the material in a wall of the same thickness as the measured sample is calculated. This modified 1-dimensional formalism gives very accurate results, and allows for the interpretation of exhalation measurements on samples with geometries that can no longer be considered as 1-dimensional.

The radon daughter concentrations were periodically monitored at 4 different locations, with radon concentrations ranging from 7 to 111 Bq m{-3}. Together with radon daughter measurements, nearly continuous measurements of the ventilation rate were performed by means of the release of nitrogen oxide tracer gas and observation of its decay with and infrared spectrometer (Miran 101). Furthermore the aerosol concentration and size distribution were monitored every 20 to 30 min with an automated aerosol spectrometer.

83 radon daughter measurements were carried out on 18 different days. 11 measurements were rejected from the optimisation exercise because they were performed within 2 after the generation of a high aerosol concentration, so that no steady state was reached. High aerosol concentrations with different size distributions were produced by burning a joss stick, a bit of paper or by smoking or cooking. The attachment rate was calculated from the aerosol size distribution and has a systematic uncertainty of about 50%.

The measured radon daughter concentrations were fitted by the room model to optimise the deposition rate of the unattached daughters. Then the fraction of unattached daughters (f,p), the equilibrium factor (F) and the activity median diameter (AMD), has been assessed in each measurement. These parameters are important in the dosimetric models. Finally, the effective dose equivalent is computed with the Jacobi-Eisfeld model (J-E) and with the James-Birchall model(J-B).

The doses per unit radon concentration are plotted as a function of the equilibrium factor. The doses computed with the J-B model are up to a factor 2 larger than he ones computed with the J-E model, due to the higher dose conversion factor for the unattachment rates. As fixed parameters, the mean values of the measurements and 2 deposition rates of the unattached daughters are taken. The conversion factors of the International Commission on Radiological Protection (NEA) exper ts' report, (1983) and the (ICRP) (1984) are proportional to the equilibrium factor and show a very different relationship compared to the J-E model and the J-B model. The main reason for this is that a low equilibrium factor is connected with a high unattached fraction. With the J-B model the dose per unit radon concentration even decreases with an increasing equilibrium factor. Therefore it would be more adequate to introduce a constant conversion factor to the radon concentration.

The analysis results in a conversion factor per unit radon concentration of 5.6 (nSvh{-1}/Bq m{-3}) or 50 (nSvy{- 1}/Bq m{-3}). When this conversion factor is applied to the average radon concentration in Belgium (50 Bq m{-3}) and an occupancy factor of 80% is assumed an average dose equivalent to the population of 2.0 mSvy{-1} is obtained.

An intercomparison with the University of Goettingen and the National Radiological Protection Board (NRPB) has been carried out to investigate the differences between methodologies for determining the size distribution of the radon decay products. 21 measurements were performed during 3 days in a house with radon concentrations ranging from 700 Bq m{-3} to 1200 Bq m{-3}. The active size distribution and inactive size distrubtion were measured simultaneously with 3 screen diffusion batteries, with a Berner impactor and with a differential mobility particle sizer. The aerosol concentrations were changed by burning a gas stove, or by burning some candles, or by smouldering cigarettes.

Each participant performed radon measurements and radon daughter measurements. Simultaneously, nearly continuous measurements of the ventilation rate were performed by means of the release of nitrogen oxide tracer gas and observation of its decay with an infrared spectometer.

During the intercomparison, the ventilation rate in the room varied between 0.2 h{-1} and 0.4 h{-1}. The values noted during the night are at the high end, probably due to a greater difference between the outdoor and the indoor temperature.

The unattached fraction was evaluated for each measurement. A value of about 10% was found without aerosol sources in the room. The unattached fraction decreased below 5% in the presence of aerosol sources. The different physical parameters of the radon decay products (attachment rate, deposition rate of the attached and unattached daughters, etc) were also evaluated.

A model based on the Monte-Carlo technique has been developed to evaluate the uncertainties induced, on the 1 hand, by assuming constant daughter concentrations and a constant flow rate during sampling and, on the other hand, in calculating the different physical parameters of the radon decay products using the room model.

The model may be considered to consist of 3 parts.

First a realistic environment is simulated by varying the radon concentration and the parameters of the room model. Then the time evolution of the daughter concentrations is calculated from the dependent equations of the room model.

In the second part a radon daughter measurement is simulated. The number of polonium-218 and polonium-214 alpha counts during sampling and decay are numerically calculated. The air concentrations are computed from these count totals and compared to the steady state concentrations. In the last part of the model the deposition rate of the unattached daughters is fitted and compared to the input value.

The fluctuations of the parameters of the room model increase in uncertainty on the measured daughter concentrations. All the parameters are varied simultaneously around their mean value. The standard deviation on the measured concentrations at 0 variation (steady state) comes from counting statistics and from the variation on the flow rate (1%). These errors are quickly dominated by errors induced by the random fluctuations of the parameters of the room model.

The influence of a sudden change in the value of 1 or several parameters of the room model has also been investigated. Priority was given to realistic situations such as: smoking, cooking, opening or closing of a door or a window, etc. It is shown that, without special precautions to keep a steady state situation in the room, the error easily exceeds 50%.

Aerosol wall losses in a spherical batch reactor (2301) were investigated experimentally. The experiments were performed with a monodispersing sodium chloride aerosol with a diameter between 0.02 and 0.2 um.

Effects of irradiating the chamber and of initial particle charge distribution were investigated. It was found that the input charge distribution of the particles (neutral, singly charged or Fuchs-Boltzmann equilibrium) did not influence the deposition rate, which can only be explained by a small electric field or 0 electric field. In that case the time constant for establishing a charge equilibrium over the aerosol is small compared to the time constant for deposition of the charged particles.

Time dependent calculations showed that the observed deposition rates could not be explained by the presence of an electric field. Therefore, the general deposition formula of Crump and Seinfeld (1981) was used. Here, no electric field is taken into account and the eddy diffusivity is defined by De=KeX{n}, where Ke is a measure for the turbulence in the reactor and x is the distance from the wall. For n, values between 2 and 3 are found in the literature.

Error may occur in the assessment of gas deposition constants when the wrong value for n is adopted, based on aerosol experiments only highlighting the necessity of supplementing the aerosol measurements with data for (100% sticking) gaseous components which are also important for the modelling of aerosol formation in smog chambers. With this purpose, a new set of experiments on gas deposition was performed. The deposition constant of gaseous components was determined by measuring the plate out activity of polonium-218 on a well defined glass surface in the same vessel. Finally, the best fit to all measurements (gas + aerosol) was obtained by using Crump and Seinfeld's theory with Ke and n as adjustable parameters. We found Ke=3.46 E-2 and n=2.6. This value for n is very close to the value of n=2.7, found by Oku yama et al (1985) in a stirred reactor. Afterwards the result was confirmed by similar experiments performed by the Bureau of Mines United States of America. Deposition of aerosol particles and unattached radon daughters in a 4m{3} cylindrical chamber, stirred with propellers, was perfectly explained by Crumps' theory with n=2.6. It may be concluded that Crumps' theory with n=2.6 is capable to explain particle deposition over almost 3 orders of magnitude with respect to size as well as to turbulence.

The values of the deposition rate constant of the unattached radon decay products reported in literature, vary over 2 orders of magnitude. In order to classify some of this variability the influence of the degree of turbulence on the deposition rate constant of the unattached daughters was investigated. The experiments were carried out in a 1 m{3} radon chamber. The turbulence in the chamber was induced by closed circuit ventilation and/or by generating heat. The aerosol concentration was less than 10 particles cm{-3}, so that the attached fraction could be neglected. The deposition rate constants were derived from the radon concentration and the 3 decay product concentrations. The values for polonium-218 were found to be about 3 times the size of the values of lead-214, which means that the associated diameter of the ultrafine lead-214 particle is about twice as large as the diameter of the ultrafine lead-214 particle.

A modified form of the formula of Crump and Seinfeld (1981) was applied to compute the corresponding; diffusion coefficients. The results in the case of closed circuit ventilation are given as a function of the lifetime of the polonium-218 isotope and the lead-214 isotope in the radon chamber. The diffusion coefficient was expected to decrease with increasing lifetime. However, only a difference between the 2 nuclides was observed. The models for predicting the concentrations of the decay products in the indoor environment assume the deposition rate constants of polonium-218 and lead-214 to be equal. However a significant lower deposition rate constant for lead-214 would explain some of the difficulties found in fitting the room model to the data we collected during the case studies in houses. Indeed, when the deposition rate constant of the unattached lead-214 to the unattached polonium-218 was applied the radon concentrations were systematically underestimated.

In collaboration with the Belgian Building Research Institute (BBRI), a national survey was organised in 290 dwelling. The dwellings are representative for the population distribution and for the type of building material. They were monitored during 6 month periods. 1 detector was placed in the living room and in half of the houses a second detector was placed in a bedroom. The average radon concentration was found to be 48 Bq m{-3}, which is not significantly different from the 53 Bq m{-3} found in the pilot study.

The national survey indicated that the average radon concentration in dwellings in the geological Ardenned, including the Condroz and the Entre-Sambre-et-Meuse is much higher than in the rest of the country. Applying a dose conversion factor of 20Bq m{-3} radon concentration = 1 mSvy{-1} and assuming an occupancy factor of 80%, results in an average dose equivalent to the population of 3.05 mSvy{-1}. The other sources of radiation, including corrected cosmic exposure and population averaged exposure of patients to medical radiation are estimated at 1.45 mSvy{-1}. In the rest of the country the average exposure is estimated at 3.8 mSvy{-1}.

The range of doses within the population were calculated. The variation between extremes is for radon higher than for any other natural source of radiation. On the basis of the national survey it is estimated that 5 to 15% of the dwellings in 1 of the zones exceeds 150 Bq m{-3} or 6.0 mSvy{-1}, the lowest action level in the United States. Statistical considerations indicate that there are hundreds of houses where the average dose equivalent is higher than the occupational limit of 50 mSvy{-1} (1250 Bq m{-3}, 80% occupancy factor).

Infiltration of radon from the ground is suspected to be an important source of radon in houses. Evidence about this was found by means of a case control study. In 51 houses built on or close to an uranium anomaly a radon detector was placed in the living room and in about half of the houses a second detector was placed in the cellar. The houses were monitored for 1 year. 49 similar houses in a neighbouring village were monitored in the same way. The seasonal averages were calculated. The radon concentration in the houses on or close to the anomaly was significantly higher than in the control houses. In about 30% of these houses the radon concentration exceeds 150 Bq m{-3}, so that the area around the uranium anomaly has to be classified as a high risk area. Houses built in this area should be designed radon safe.

A second survey has been carried out in a municipality where high radon concentrations are expected because of the geological properties of the earth. In 54 houses, selected ad random out of the telephone book, a radon detector was placed in the living room and if possible a second detector was placed in the cellar. The houses were monitored for 1 year.

The seasonal averaged radon concentrations in the living rooms vary between 30 and 4000 Bq m{-3} and between 35 and 14500 Bq m{-3} in the cellars. The radon concentration is in 10% of the living rooms in excess of 150 Bq m{-3}.

The radon concentrations in 3 houses close to the house with 4000 Bq m{-3} in the living room, which corresponds to an effective dose equivalent of 160 mSvy{-1} (occupancy factor 80%) were measured. Concentrations of 3400 +/-200 Bq m{-3}, 360 +/- 20 Bq m{-3} and 165 +/- 15 Bq m{-3} respectively were found, so that the highest value of the ad random study is not an isolated case.

Mitigation measures, which mainly consist of a subfloor ventilation system, are installed in the houses with a high radon concentration. The installations are paid for by the municipality. The fir st results show a marked decrease in radon concentration (less than 20% left).

New types of phosphogypsum are sold as building material in Belgium because industry has switched over from Maroccan phosphate ore to South African ore. The radioactive properties of 2 of these new types of phosphogypsum have been tested. From each type 30 plates of 0.295x0.210 m by 0.005 m and 30 plates of 0.295x0.210 m by 0.020 m were made.

The specific radon-226 and radon-232 activities were measured by means of high resolution gamma spectroscopy and the thorium activity of the new types of phosphogypsum was funded to be much higher than the thorium activity of the phosphogypsum of Maroccan origin. For the radium activity the reverse is found.

The radon-222 exhalation rate was evaluated by closing the plates in airtight barrels and measuring the equilibrium concentration by means of the Lucas method. The radon-220 exhalation rate of the plates was determined by means of the radon-220 decay products concentration in an airtight 1 m{3} chamber. The chamber was ventilated in a controlled way with outside air, in order to obtain sufficient aerosol particles inside the chamber. Each sample was measured 3 to 4 times. The reproducibility of the equilibrium equivalent thoron concentration is about 25%, which is almost entirely due to fluctuations in the aerosol concentration in the chamber. The thoron exhalation is, as expected from the short half live of radon-220, independent of the thickness of the samples.

The effect of covering the entire surface of the plates with 2 layers of a common latex paint was also investigated. The thoron exhalation after painting was 10 to 20 times lower. In order to evaluate the radon-220 exhalation rates of the phosphogypsum plates, the radon chamber results were using convented using realistic assumptions. The effective dose equivalent received by an occupant due to the thoron exhalation of uncovered gypsum is 1.8 +/- 0.4 mSvy{-1} for type 1 and 0.9 +/- 0.2 mSvy{-1} type 2. Such a high dose, originating from 1 building material, can not be brought into agreement with the as low as reasonably achievable (ALARA) principle. Painting the gypsum, however, lowers the radon-220 dose by a factor of 10 to 20.

## Informations connexes

Related Information not available

### Open Access

### Open Access

## Coordinateur

9000 GENT

Belgium

See on map

**Contact administratif:**JANSSENS

Tél.: +32-91-228731

**Dernière mise à jour le**: 1994-01-23

**Numéro d'enregistrement**: 13433

**Dernière mise à jour le**1994-01-23

**Share this page**