Relevance of radon progeny measurements for the assessment of inhalation doses in groundwater utilities

The high radon concentrations measured in the indoor air of groundwater facilities and the prevalence of the problem have been known for several years. Unlike in other workplaces, in groundwater plants, radon is released into the air from the water treatment processes. During the measurements of this study, the average radon concentrations varied from 500 to 8800 Bq m–3. In addition, the indoor air of the treatment plants is filtered and there are no significant internal aerosol sources. However, only a few published studies on groundwater plants have investigated the properties of the radon progeny aerosol, such as the equilibrium factor (F) or the size distribution of the aerosol, which are important for assessing the dose received by workers. Moreover, the International Commission on Radiological Protection has not provided generic aerosol parameter values for dose assessment in groundwater treatment facilities. In this study, radon and radon progeny measurements were carried out at three groundwater plants. The results indicate surprisingly high unattached fractions (fp = 0.27–0.58), suggesting a low aerosol concentration in indoor air. The corresponding F values were 0.09–0.42, well below those measured in previous studies. Based on a comparison of the effective dose rate calculations, either the determination of the fp or, with certain limitations, the measurement of radon is recommended. Dose rate calculation based on the potential alpha energy concentration alone proved unreliable.


Introduction
The linear relationship between radon exposure and lung cancer risk has been substantiated through several international epidemiological studies, and there is no established health-based threshold dose for the hazardousness of radon, 222 Rn (Darby et al 2004, Kelly-Reif et al 2023).
Radon, an inert noble gas, is partially released from soil aggregates when its parent, 226 Ra, decays.Radon is transported in the soil either dissolved in groundwater or as a gas via soil pores.The half-life of radon is 3.8232 ± 0.0008 d (LNHB 2011), and its radioactive decay produces a series of short-lived, solid and chemically reactive decay products.When radon decays, its first progeny, 218 Po, immediately attaches to trace gases and vapours in the air to form particles less than 5 nm in aerodynamic diameter (Porstendörfer 2001, ICRP 2017), known as unattached progeny.It then decays into 214 Pb, 214 Bi and 214 Po, during which time the decay products attach to larger aerosol particles, forming attached progeny.
The concentration of radon decay products is typically expressed as the potential alpha energy concentration (PAEC), most commonly in nanojoules per cubic metre (nJ m -3 ).Potential alpha energy is the energy released as radon progeny undergoes a full decay process, transforming into 210 Pb.The equilibrium factor (F) is defined as the ratio of the prevailing PAEC to the situation where the prevailing radon concentration would be in full secular equilibrium with its decay products.In other words, it indicates how much of the decay products are present in the air in relation to radon gas.The unattached fraction (f p ), on The first plant under investigation is a groundwater treatment plant, the second is an artificial groundwater treatment plant and the third concurrently treats both surface and groundwater.Detailed descriptions of the processes and key statistics for these plants are provided in the appendix.All of these water plants are equipped with mechanical supply and exhaust ventilation (balanced ventilation) systems equipped with supply air filtration, except for the limestone filtration room in the artificial groundwater plant before its renovation.The process areas in these water plants undergo regular cleaning, and there are no known significant sources of aerosols or pollutants within the plant premises.
At the groundwater plant, raw water (radon concentration 28-30 Bq l -1 ) is pumped through packed tower aeration (PTA) and subsequently passed through a limestone (CaCO 3 , 2-4 mm grains) filter, an anthracite filter and a bottom water tank.The water pumped into the distribution network undergoes UV disinfection, and monochloramine is dosed into it.The typical daily flow rate is 400 m 3 d -1 , with a maximum of 1000 m 3 d -1 .The starting and stopping of the plant are regulated by the water level of the water tower in the network.
In the case of the artificial groundwater treatment plant, raw water (radon concentration 27-50 Bq l -1 ) from its wells is pumped through an overflow weir, directed to a limestone (CaCO 3 , 4-16 mm grains) filtration system and subsequently transferred to a water tank before undergoing UV disinfection and transfer pumping.The daily flow rate varies between 10 000 and 15 000 m 3 d -1 , with a maximum of 20 000 m 3 d -1 .The plant operates continuously, but its process flow varies significantly depending on the demand.
In the case of the surface water plant, the radon problem in indoor air is limited to the chemical precipitation flotation-filtration room.The plant treats an equal ratio of surface and groundwater.The radon concentration is 38-44 Bq l -1 in groundwater and only 17-21 Bq l -1 in the treatment process before flotation-filtration.At the beginning of the process, the groundwater is ozonated to oxidize the dissolved iron and mixed with surface water, in which the radon concentration is very low (about 0.4 Bq l −1 ).Ozonation removes about 10% of the radon in the groundwater (Tyrväinen et al 2023), and mixing of the waters before chemical precipitation dilutes the radon concentration to about half of the original level.Rapid mixing of the precipitation chemical (ferric sulphate (PIX-322), Kemira Oyj, Finland) and distribution of the water to the pool lines has been implemented using overflow weirs.Flotation and sand filtration are utilized to separate the floc formed in mixers.The surface water plant operates continuously and steadily, the typical daily flow is 6000 m 3 d -1 , and the maximum is 10 000 m 3 d -1 .

Sampling and measurements
Measurements of radon concentrations in the air were carried out using AlphaGUARD radon monitors (models P30, PQ2000 and PQ2000 Pro by Bertin Technologies SAS, Montigny-le-Bretonneux, France).The instruments underwent calibration both before and after the measurements at the Finnish Radiation and Nuclear Safety Authority (STUK) laboratory in Vantaa, Finland.The laboratory is accredited and complies with standard EN ISO/IEC 17025:2017(FINAS 2021).Specific calibration coefficients and calibration uncertainties for each device were taken into account during the calculation of results.Possible sudden changes in air pressure were also monitored with an AlphaGUARD monitor (air pressure measurement was uncalibrated).
Tracerlab BWLM-PLUS-2 S dual channel radon/thoron progeny monitors (Tracerlab Engineering & Technology, Köln, Germany) were used to measure the concentration of radon progeny.There are two sampling lines and detectors in the monitor.One line collects all the decay products on a filter, based on which the total PAEC is calculated.In the other line, the unattached decay products are collected on a wire screen with a 50% cut-off at 4 nm.Calibration of these devices was carried out by Bundesamt für Strahlenschutz, Berlin, Germany, covering both the total and unattached PAEC of radon progeny.Their accreditation encompasses the total PAEC of radon progeny but not the unattached PAEC.As per the manufacturer's statement, when performing measurements in environments with very low aerosol concentrations, there may be increased uncertainties associated with the wire screen method.The limitations related to the wire screen method have been presented in detail by Reineking and Porstendörfer (1990).
Temperature and air humidity were obtained using a Rotronic Hydrolog HL-NT3 Humidity & Temperature Data Logger (Rotronic AG, Basserdorf, Switzerland).The calibration of the instrument is traceable to the national standard maintained by VTT MIKES, which is a signatory of the CIPM MRA (International Committee for Weights and Measures Mutual Recognition Arrangement).
Radon and radon progeny monitors were placed in pairs at the measurement sites.The continuous measurements were sampled with 1-hour intervals (duration of 1 h, table 1).
The water flow data utilized in this study were sourced from the automation system of the water utilities.b Disturbances of mechanical ventilation on 1-2 November 2022 were excluded from the result (see figure 3). a Breathing rates: sedentary = 0.86 m 3 h -1 ; substantial = 1.2 m 3 h -1 .

Calculations
The equilibrium factor, F, was calculated using a conversion coefficient of 5.56 • 10 -9 J Bq -1 for radon in full equilibrium: where PAEC potential alpha energy concentration of radon decay products in J m -3 , C Rn radon concentration in Bq m -3 .The unattached fraction, f p , was calculated as follows: where PAEC U is the unattached PAEC in J m -3 .When estimating effective doses from radon, measurement results related to radon decay products are typically unavailable.In such situations, the dose is estimated using generic values for the characteristics of the decay product aerosol (table 2).
In water treatment plants, workers are mostly on the move.Therefore, a dose conversion factor of 6 Sv (Jh m -3 ) -1 was used, which is based on the reference worker's breathing rate of 1.2 m 3 h -1 .Since data on the equilibrium factor in waterworks are sparse, an F value of 0.4 was used.The effective dose (E) would, hence, be assessed as follows: where C Rn is the radon concentration, F is 0.4 and t is the exposure time in hours.
Since the quality of the indoor air in a water treatment plant differs from the air in other indoor workplaces in terms of factors such as temperature, cleanliness and humidity, it is justified to conduct additional studies by measuring the PAEC of the indoor air.In this case, the dose was estimated as follows: (4) If unattached decay products are also measured, a more precise dose assessment method provided by ICRP ( 2017) can be used: (5) Effective doses were assessed using the assessment methods presented above.

Groundwater treatment plant
At the groundwater plant, the radon and PAEC measurements were carried out at a site near the PTA, which had been identified as the area where radon is released from the water into the air.
Figure 1 presents two nested temporal variations in both the radon concentration and concentrations of decay products.The most significant temporal variation is caused by the starting and stopping of pumping operations.The water level in the water tower reaches the upper limit in the early morning and the plant continues to operate again at noon.Weaker but more frequent temporal variation is due to the intermittent starting and stopping of the ventilation system.When the ventilation system is running at full capacity, the ventilation rate in the waterworks room is 0.8 h -1 .However, the exact ventilation rate at the measurement site is unknown because the room is partly open to other process rooms.
The measurement results were calculated for two situations: for the total measurement period and the maximum situation of the radon concentration during the operation periods of the process (>2 h from the start of the process).During the total measurement period, the radon concentration varied between 100-3000 Bq m -3 , F between 0.1 and 0.6, and f p between 0.2 and 0.7.During the operating periods of the process, changes in the radon concentration and the total and unattached PAEC occur at slightly different times, so that when pumping stops or starts, the process flow changes first, then the radon concentration and after this the PAEC.The f p value is high, up to 0.7, immediately after the plant starts up, but decreases when a stable radon concentration is reached during plant operation.A stable radon concentration is generally reached approximately 2 h after the pumps start, but the time depends on the ventilation rate and other local conditions.During the operation periods, the radon concentration varied from 1500-3000 Bq m -3 , F from 0.1-0.4 and f p from 0.2-0.6,when the disturbance caused by plant stoppages was removed from the results.Figure 7 presents the F-f p ratio of the measurement data during the operation period and for the entire measurement period, adapted from Marsh et al (2002) to the measurement data collected.The measurement data for the operation periods fit within the 95% error limits of the reference data, but data for the process stop and start periods do not.
The measurement data and the dose rates are presented in table 3 and example calculations of annual doses are in provided table 4.

Artificial groundwater treatment plant
Indoor air radon and PAEC measurements were carried out in the limestone filtration room and the pump room.This water plant is characterized by large temporal variation in the process flow, resulting in high variation in indoor radon and decay product concentrations (figures 2-4).At the time of the measurements, the water plant was undergoing a radon renovation of the mechanical ventilation system.
In the limestone filtration room (figures 2 and 3), the measurement instruments were placed in the centre of the room, away from the main radon source, which is the overflow weir at the end of the room.The distance between the overflow weir and the measurement site was approximately 18 metres and the ventilation rate was very low.The measurement site was located in a central work area.The first measurement was conducted before the renovation when the ventilation rate was close to zero (figure 2), and the indoor radon concentration varied from 3000-19 000 Bq m -3 .Changes in the radon concentration followed changes in the process flow with a clear delay.F varied between 0.2-0.5, f p was high (range 0.25-0.8)and was inversely correlated with F.
Figures 3 and 4 represent the results in the limestone filtration room and the pump room after the renovation of the ventilation system.Ventilation rates in both rooms were 1.2 h -1 from 8 am to 6 pm and 0.3 h -1 at other times.Due to defects in the ventilation unit during the trial use period, the measurement times were short.The results presented in figure 3 also show two short ventilation unit failures.
In the pump room, the instruments had been placed close to the ventilation outlet to allow the longest possible time for radon decay products to form.This was also the usual working area.The results mainly represent radon that has been in the air for a relatively long time, because radon gas is transferred from water to air in covered flow chambers that are gradually ventilated into the room air.In other words, radon and its decay products remain in the air much longer than the theoretical ventilation rate (1.2 h -1 ) would suggest.After ventilation repairs, the radon concentration varied from 1000-4000 Bq m -3 , F from 0.3-0.6 and f p from 0.2-0.4(figure 3, without failure times) in the limestone filtration room.In the pump room, radon varied from 1000-3000 Bq m -3 , F from 0.2-0.3 and f p from 0.3-0.5 (figure 4).
The measurement data in figure 7 present the F-f p ratio in the artificial groundwater plant.The data show a change in f p when ventilation has been implemented in the release area of the radon and mechanical ventilation has been implemented in the limestone filtration room.F slightly increased from 0.33 to 0.42 but f p concurrently decreased from 0.51 to 0.27.At the same time, in the pump room, the slightly lower F (0.24) and higher f p (0.37) in comparison to the limestone filtration room can be explained by the release of radon from the water into the air in a closed space, from where the radon and decay products move to the well-ventilated pump room only after a delay.

Surface and groundwater treatment plant
At the surface and groundwater treatment plant, indoor radon and decay product measurements were carried out in the flotation-filtration room at three different measurement sites.This water plant is characterized by a steady process flow, so the indoor radon concentration also remains relatively constant (figures 5 and 6).The measurement sites were chosen in such a way that S0 was close to the point-like radon release area and the area affected by the air conditioning exhaust valve.S1 and S2 were located in the area of the flotation basins and there was an indoor air flow connection between the measurement sites.All measurement sites were located in working areas.The recurrent spikes that stand out in the graphs of groundwater flow and radon concentration (sites S0 and S2) are due to the air-water backwashing of sand filters.
Figure 5 presents measurements from site S0, near the radon release area, where decomposition products have not had time to form and the ventilation discharges radon and decay products directly to the outdoor air.In this case, F is very low (0.09) and f p is very high (daily values range from 0.45-0.64).
The F-f p ratios in the surface and groundwater treatment plant are presented in figure 7.In the comparison of the measurement sites (S0, S1 and S2), the increase in the value of F indicates a temporal distance from the point of radon release.Correspondingly, the range of variation in f p is similar, regardless of the measurement site.Measurement sites S1 and S2 were located between the air inlet and exhaust valves of the ventilation unit so that there was an air flow connection between them.As illustrated in figure 6, the radon concentration slightly decreased between the measurement sites from 530 Bq m -3 -500 Bq m -3 and, correspondingly, the The ventilation rate was 0.9 h -1 .The drying function of the air conditioner started on 29 October, when the outdoor air humidity rose above 60%.
total PAEC increased from 440-660 nJ m -3 .The inverse correlation between F and f p can also be observed in the results between these measurement sites: F increased from 0.15 to 0.25 and f p decreased from 0.58 to 0.49.Detailed results with corresponding uncertainties are presented in table 3.

Annual doses
A water producer typically operates several water treatment plants and employees rotate between them during their working days.Usually, each visit is short and includes a visual inspection of the equipment and the operation of the process.Sometimes, maintenance work is carried out at a single plant, which can result in several hours of exposure per day, but if the plant is stopped completely for maintenance, the ventilation transports radon to outdoor air.
Table 4 presents the annual doses of workers calculated from working hours at the water plants included in this study.In reality, these working hours are distributed among several workers.The annual doses corresponding to the total working time (240 d per year) in the conditions of the water plants are also included in the table.

Discussion
In groundwater plants, the release rate of radon from water to air primarily depends on the radon concentration of raw water, the process flow rate and the radon release rate from water to air during aeration.The temporal and spatial variation of the radon concentration in indoor air is additionally due to the location of the radon release area and the effect of ventilation (Tyrväinen et al 2023).Reineking and Porstendörfer (1990), ICRP (2017) have described the formation processes of radon decay products in the air: (1) the formation of positive and neutral decay nuclides during the decay of radon gas, (2) the formation of molecule clusters (unattached progeny) and their attachment to aerosols (attached progeny) and removal from the air by (3) deposition on surfaces (plate-out) and ( 4) as a result of ventilation.In addition, an attached decay product may (5) form an unattached decay product due to recoil energy in radioactive decay.
Near the very highly ventilated (about 3-4 h -1 ) radon release area in the present study, the average F value was measured to be only 0.1.Only radon gas, without radon progeny, is released from groundwater into indoor air during water treatment, which explains the low F values measured near the release areas (figure 5). Figure 6 displays the results from two measurement sites in the same airflow.The differences in radon and decay product concentrations, as well as in F and f p measured between the measurement sites S1 and S2 indicate the decay of radon and the attachment of decay products to aerosols during the airflow between the measurement sites.Effective ventilation results in significant differences in radon and decay product concentrations in different parts of the room (measurement sites S0, S1 and S2).At normal ventilation rates (0.8-1.2 h -1 ), the F value varies in the range of 0.15-0.4.The F values are lower than the ICRP (2017) recommendation of 0.4 for indoor workplaces (table 2).With a very low ventilation rate (<0.1 h -1 ), F = 0.3-0.4 is close to the ICRP-recommended F value for workplaces.In Austria and Germany, higher F values (0.37-0.55) have been reported for groundwater plants in four areas (AGES 2006).
No particle number concentration (Z) measurements were conducted during this study, but an estimate of the conditions can be presented at a general level and a particle number concentration can be estimated

3.
Averages of radon decay product measurement periods, dose rates and environmental variables.The working hours of the year presented here are distributed among several workers.
Ė values calculated based on the unattached and attached PAEC ( Ėfp+PAEC ), the total PAEC ( ĖPAEC ) and the radon concentration ( ĖRn ) are presented in table 3, and the annual doses calculated for employees at each groundwater plant in table 4.
If the dose conversion factor calculated by the unattached fraction (equation 5) is taken as a benchmark and the dose estimate is considered, it is found that the dose estimated from the radon concentration alone is on average in the same order (−7%), but the differences between the methods increase at the extremes of the ventilation rates.When the ventilation rate is near zero, the assessment based on the radon concentration alone is only 50% of the benchmark estimate, and at the high ventilation rate (3-4 h −1 ), it is about 80% higher than the benchmark.
When dose assessment is based on the PAEC measurement alone (often called the working level measurement), the average assessment is 50% lower than the benchmark assessment.Here, the variation between facilities is much less pronounced.The smallest difference (−36%) was observed in the limestone filtration room of an artificial groundwater plant after radon renovation.The largest difference (−62%) was observed at measurement site S0 of the surface water plant, which was the most significant local source of radon in the plant and where the ventilation rate was very high.
It should be noted that this study did not use the best available dose calculation method published by the ICRP, which takes into account the activity concentration of each decay product in the unattached particles, nucleation mode and accumulation mode.In any case, it can be concluded that PAEC measurements are not a good basis for effective dose assessment in water plants due to the high f p of decay products.The so-called working level measurements would thus underestimate the doses in this study by an estimated 50%.If the applied measurement cannot separately detect unattached radon progeny, it is best to assess the dose based on the radon concentration.
When examining the results, it should be noted that the studied water plants have very high standards of cleanliness.The intake air is filtered, ventilation is efficient in most cases, and no dusty materials are handled in the water production rooms.Therefore, the results cannot be generalised to all water treatment plants.

Uncertainties of the study
The measurements of this study were taken with 1-3 fixed measuring device pairs placed in central work areas.As the radon release areas were confined to specific sites and the rooms were mechanically ventilated, different concentrations could be found in different parts of the room.Thus, the measurement results may not be representative of the average values for the entire workspace or the range.
As discussed in section 2.2, the calibration of the PAEC measurement of unattached radon progeny for the Tracerlab BWLM-PLUS-2 S instruments was not within the scope of accreditation.In addition, the calibration was performed under conditions corresponding to normal indoor air conditions (fp = 0.10, F = 0.36 and Z = 849 cm -3 ).When measuring clean room conditions, f p is high by default (Paul et al 1999).Therefore, the instruments were tested at very low particle concentrations.They were exposed for 46 h in a radon chamber at STUK (Vantaa, Finland), a 1.6 m 3 thermostated steel tank into which radon-containing air is introduced, which is first HEPA-filtered and humidified to the desired humidity, typically RH 40%.As the tank is frequently cleaned, few particles are present during operation.The instruments provided average f p values (and standard deviations) of 1.10 (0.05) and 1.19 (0.07).It is therefore possible that, especially at very low particle concentrations, the instrumentation may overestimate the PAEC of the unattached progeny and thus overestimate the f p value.The total PAEC measurement has been calibrated with accreditation, so its accuracy need not be questioned.Since the study measured f p values between 0.3 and 0.8, the overestimation of the f p value at 0.30 could be up to 0.03-0.05and at 0.80 up to 0.07-0.13.However, this would not affect our conclusions considering the uncertainties of the measurements.

6.
The conditions for measuring radon and decay products at groundwater and artificial groundwater plants differ significantly from the conditions of better-known locations such as homes, indoor workplaces, mines and tourist caves: (1) At water plants, the radon and its decay product concentrations in the indoor air can vary significantly both in different parts of the room and at different times of the day.An incorrectly placed or timed measurement may easily underestimate or overestimate the concentration to which workers are exposed.
(2) The proportion of f p of radon decay products was high at all the studied water plants, which indicates the cleanliness of the indoor air and the absence of significant aerosol sources.(3) Either the determination of f p or radon concentration measurement can be used to calculate the dose rate caused by radon and its decay products.Dose rate calculation based on the total PAEC alone proved unreliable, as well as calculation based on radon concentration measurement if the room ventilation rate was very low or high.
Thus far, few radon decay product measurements have been carried out at water plants.To supplement the present knowledge, variations in radon concentrations as well as the concentrations of radon decay products and unattached fractions should be measured at various types of groundwater plants representing different periods of building technology.At the same time, the particle concentrations in the air should be measured and possible aerosol sources assessed.

Figure 1 .
Figure 1.Groundwater flow and radon concentration (a), total and unattached PAEC (b), and F and fp (c) of a periodically running groundwater plant, as well as the relative humidity, temperature and air pressure in the indoor air (d) during the measurement.The ventilation rate was 0.8 h −1 .

Figure 2 .
Figure 2. Groundwater flow and radon concentration (a), total and unattached PAEC (b), and F and fp (c) before radon renovation in the limestone filtration room of the artificial groundwater treatment plant, as well as the relative humidity, temperature and air pressure in the indoor air (d) during the measurement.Installation work in progress and ventilation rate <0.1 h −1 .

Figure 3 .
Figure 3. (a) Groundwater flow and radon concentration (a), total and unattached PAEC (b), and F and fp (c) in the limestone filtration room of the artificial groundwater treatment plant.Ventilation rate 1.2 h -1 from 8 am to 6 pm and 0.3 h -1 at other times.Humidity 48%, temperature 12 • C and air pressure 1026 hPa.During 1.-2.12.2022, short interruptions of the ventilation unit occurred.

Figure 4 .
Figure 4. (a) Groundwater flow and radon concentration (a), total and unattached PAEC (b), and F and fp (c) in the pump room of the artificial groundwater treatment plant.Ventilation rate 1.2 h -1 from 8 am to 6 pm and 0.3 h -1 at other times.Humidity 48%, temperature 12 • C and air pressure 1026 hPa.

Figure 5 .
Figure 5. (a) Groundwater flow and radon concentration (a), total and unattached PAEC (b), and F and fp (c) in the flotation-filtration room of the surface and groundwater treatment plant, as well as the relative humidity, temperature and air pressure in the indoor air (d) during the measurement.The measurement site S0 was located near the area where radon is released into the air from overflow weirs and near the exhaust air valve of the ventilation unit.

Figure 6 .
Figure 6.Groundwater flow and radon concentration (a), total and unattached PAEC (b), and F and fp (c) in the flotation-filtration room of the surface and groundwater treatment plant, as well as the relative humidity, temperature and air pressure in the indoor air (d) during the measurement.The measurement sites S1 and S2 were located on the same airflow route.The ventilation rate was 0.9 h -1 .The drying function of the air conditioner started on 29 October, when the outdoor air humidity rose above 60%.

Figure 7 .
Figure 7.The F−fp averages of all measurement periods and sites of this study (a), and all F−fp measurement results from the groundwater plant (b), the artificial groundwater plant (c) and the surface and groundwater treatment plant (d).All figures include fp modelled as a function of F according to Marsh et al (2002) . Figure ventilation on 1-2 November 2022 were excluded from the result (see figure3).Table 4.Annual doses calculated for an example worker based on different measurement results.

Table 1 .
Radon and progeny sampling period and number of samples.The sampling interval was one hour.

Table 2 .
Generic parameters for dose assessment in different types of work (ICRP 2017).