8
Mitigation

Mitigation Of Radon In Indoor Air

Radon Entry into Buildings: A Brief Review

Radon is a ubiquitous constituent of soil gas as its radioactive parent, 226Ra, is widely distributed in the earth's crust. Typical soil-gas radon concentrations are around 30,000-300,000 Bq m-3, and values ranging from about 5,000 Bq m-3 to about 5,000,000 Bq m-3 have been reported. The principal mechanisms of radon transport in porous media (e.g., soil) are advection and diffusion; both are sources of radon entry into buildings, and they are described briefly in this section. More complete discussions of radon transport in soils and entry into buildings can be found in the literature (Sextro 1994; Nazaroff 1992; Nazaroff and others 1988).

Advection

Bulk flow of soil gas that contains radon is the main mechanism of radon entry into buildings. This flow occurs in response to pressure differences between the air in buildings and the air in the adjacent soil. These differences are established by the natural interaction between the building and the surrounding environment and in some cases by the operation of mechanical systems within the building.

The temperature difference between the air in a building and the air outside creates a pressure gradient across the building shell that varies with height along the shell. When the indoor air temperature is higher than the outdoor, the indoor air pressure in the lower parts of the building (e.g., in the basement or the region of the ground-contact floor) is slightly lower than the air pressure in the adjacent



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 141
--> 8 Mitigation Mitigation Of Radon In Indoor Air Radon Entry into Buildings: A Brief Review Radon is a ubiquitous constituent of soil gas as its radioactive parent, 226Ra, is widely distributed in the earth's crust. Typical soil-gas radon concentrations are around 30,000-300,000 Bq m-3, and values ranging from about 5,000 Bq m-3 to about 5,000,000 Bq m-3 have been reported. The principal mechanisms of radon transport in porous media (e.g., soil) are advection and diffusion; both are sources of radon entry into buildings, and they are described briefly in this section. More complete discussions of radon transport in soils and entry into buildings can be found in the literature (Sextro 1994; Nazaroff 1992; Nazaroff and others 1988). Advection Bulk flow of soil gas that contains radon is the main mechanism of radon entry into buildings. This flow occurs in response to pressure differences between the air in buildings and the air in the adjacent soil. These differences are established by the natural interaction between the building and the surrounding environment and in some cases by the operation of mechanical systems within the building. The temperature difference between the air in a building and the air outside creates a pressure gradient across the building shell that varies with height along the shell. When the indoor air temperature is higher than the outdoor, the indoor air pressure in the lower parts of the building (e.g., in the basement or the region of the ground-contact floor) is slightly lower than the air pressure in the adjacent

OCR for page 141
--> soil; conversely, at the top of the building, the pressure gradient across the building shell is reversed, so air flows out of the building. This thermal-stack effect is one of two principal mechanisms responsible for the natural ventilation of buildings; it is sometimes referred to as infiltration. Wind also creates pressure differences between the inside and the outside of a building. The pressure fields can be complex and depend on the size and shape of the building and the wind direction. The pressure fields also extend into the surrounding soil, increasing the pressure of the soil air on the upwind side and decreasing it on the downwind side of the building (Riley and others 1996). The net effect is usually an airflow out of the top of the building caused by the Bernoulli effect of the wind over the roof or by the reduction in air pressure on the leeward side of the building. In response to the slightly lower air pressure in the building, ''makeup" air flows in through openings in the building shell, some of which might provide direct contact with soil air. The effects of the thermal stack and the wind can independently result in indoor-outdoor pressure differences of about the same size at the lower portions of the building shell. For example, an indoor-outdoor temperature difference of 20 °C (a common wintertime temperature difference in many parts of the United States) results in an indoor-outdoor pressure difference of about-3 Pa at the bottom of the thermal stack (the basement or other ground-contact floor). Similarly, a wind speed of 4 m s-1 results in an indoor-outdoor pressure difference of about-2 Pa for a typical house (the relationship between the indoor-outdoor pressure difference and wind speed is quadratic, so doubling the wind speed increases the pressure difference by a factor of 4). Those are "steady-state" values. Temperature differences usually do not change very rapidly, but in the course of a day outdoor temperatures can change by 20 °C or more as part of the diurnal cycle. Wind speeds and directions are highly changeable, and this leads to substantial variation in "instantaneous" pressures. A more detailed discussion of the pressure gradients developed in buildings can be found in Liddament (1986). The operation of mechanical systems in a building can lower the pressure in a building, especially when the flows induced by these systems are unbalanced. Operation of an exhaust fan—such as a bathroom or kitchen fan, whole-house fan, or, in some cases, an attic fan—will result in lower indoor pressures. Just as in the case of infiltration, the "makeup" air flows into the house through leaks in the building shell, some of which provide a pathway for soil-air entry. Operation of a forced-air heating and cooling system can also lead to unbalanced flows and result in lower indoor pressures, depending on the locations of the supply and return ducts and their leakage characteristics. Diffusion Molecular diffusion, driven by the concentration difference between low-concentration regions, such as the interior of a building, and the higher-concen-

OCR for page 141
--> tration soil is another mechanism for radon entry into buildings. A key controlling variable—in addition to the radon concentration gradient—is the diffusivity of any material that separates the soil from the building interior, such as a concrete floor. In the case of an open soil floor, and in the absence of pressure differences, the radon flux density is the highest across this interface and is about the same as would be observed outdoors (soil moisture differences can have an effect on the diffusivity of soil). For soils with typical radium content, the radon flux density is 1–2 × 10-2 Bq m-2 s-1. The presence of a concrete floor can increase the concentration gradient over that found in open soil, but radon diffusivities are typically smaller in concrete than for soil. The concrete floor acts as a diffusion barrier; diffusive radon entry though such a floor is likely to be somewhat lower than that for open soil. For nominal values of the diffusivity of concrete and typical radon concentrations in soil gas adjacent to a building, the radon entry rate due to diffusion through a concrete floor is about 1 × 10-2 Bq m-2 s-1, which is about half the open-soil value. Most of this radon is from the soil itself, as opposed to the radon arising from radium in the concrete (Sextro 1994). This estimate is consistent with measurements of flux density conducted as part of extensive field experiments, where the average flux density was 1.3 × 10-2 Bq m-2 s-1 (Turk and others 1990). Building materials themselves—especially those with soil-based constituents, such as concrete, brick, and natural stone—contain radium and will thus be a source of radon diffusion into indoor air. In most cases, however, the amount of radium in such materials is small enough that, in combination with the diffusivity of the material and typical infiltration and ventilation rates of buildings, their overall contribution to indoor radon concentrations is modest. Other Sources Three other sources of radon are worth noting. The first is advective transport of soil gas driven by changes in atmospheric pressure. Although large changes in atmospheric pressure can result from changes in weather, they are relatively infrequent compared with the smaller diurnal and semidiurnal atmospheric pressure changes (Robinson and Sextro 1997). Overall, these effects are estimated to be small and to yield overall radon entry rates roughly the same as that due to the second source, infiltrating outdoor air. The latter, considered in more detail in chapter 2, provides an irreducible "baseline" indoor radon concentration. The third source is the topic of this report: indoor use of water that contains dissolved radon, which is the subject of detailed discussions elsewhere. In the context of other sources, the average contribution made by water to indoor-air radon concentrations is very modest, given that the average transfer coefficient is 10-4.

OCR for page 141
--> Radon Entry in Context Several sources of indoor radon have been described; from all but outdoor air, the resulting indoor concentrations (and hence exposures) depend on the combination of the source strength and the ventilation rate of the building. As noted earlier, the stack and wind effects are primarily responsible for the natural ventilation of buildings, in addition to providing a driving force for radon transport into buildings. It is useful to provide a context for these flows. The steady-state solution to the first-order differential equation that describes indoor radon concentrations illustrates the key variables: Where, Ci is the indoor concentration (Bq m-3), Co is the outdoor concentration (Bq m-3), S is the radon entry or production rate (Bq per unit time, t), V is the house volume (m3), R is the removal rate (t-1). Here R can account for any method of removal and is just the sum of the individual removal terms. In this case, R is the air-exchange rate (AER). A typical single-story house has a "footprint" of 120 m2 and a corresponding volume of about 300 m3. Annual average natural ventilation rates are about 0.9 h-1. Using those values for V and AER, the air flow rate (the product of V and AER) through this house is 270 m3 h-1. The radon entry rate corresponding to the Environmental Protection Agency (EPA) guideline concentration of 150 Bq m-3 can be estimated from equation 8.1. Neglecting any contribution from outdoor air and using the ventilation rate described above, the radon entry rate, S, is about 40,000 Bq h-1 (about 11 Bq s-1). That can be compared with the estimates of diffusive radon entry. Assuming a floor area of 120 m2, the entry rate due to diffusion is about 1 Bq s-1, a small fraction of what is needed to produce an indoor air concentration of 150 Bq m-3. Similarly, the soil-gas flow to produce this indoor concentration can be estimated. Assuming a typical value of about 40,000 Bq m-3 for the concentration of radon in soil gas, the soil-gas entry rate is about 1 m3 h-1, which is about 0.4% of the overall air flow rate into the house. Mitigation Methods for Existing Houses Conceptually, there are two approaches for mitigating indoor radon concentrations (or most other indoor pollutants, for that matter): source control and

OCR for page 141
--> concentration reduction. One can use equation 8.1 to provide some insight into the relative efficacy of the two approaches. Indoor radon concentration is directly proportional to the source term and (again neglecting outdoor air as a source) inversely proportional to the removal terms. Considering the latter first, removal can mean either increased ventilation or some other method of removing radon or radon decay products from indoor air. In any case, for the previous example, to decrease the radon concentration by a factor of 2 by ventilation alone, the AER will need to be increased to 1.8 h-1. Although that is not an excessive ventilation rate and is often achieved naturally when doors and windows are open, AER values of 2 h-1 commonly have comfort and energy penalties during colder seasons. Thus, this means of reducing radon concentration has some practical upper limits. In addition, forced ventilation can result in additional depressurization of a building and potentially increase the radon entry rate. Other nonventilation removal methods are possible, and two are described in more detail below. As with ventilation, substantial removal means processing indoor air at rates that are comparable with or greater than the ventilation rate (about 270 m3 h-1 in the example above). It also means that essentially the entire living space will need to be treated; this could require multiple single-room reduction devices (such as air cleaning, described below) or whole-house devices used in conjunction with a forced-air system. In the following sections, source-control methods are described first and then concentration-reduction methods. Source Control When high indoor radon concentrations in houses were found in various locations in North America in the middle 1970s, initial research on reduction methods was based on two key assumptions: that the source term was high concentrations of radium in soil materials derived from uranium mining and that the principal means of radon transport and entry was diffusion. Thus, initial attempts at source control focused on removal of the materials, typically uranium mill tailings used as back fill under floor slabs or adjacent to basement walls. In addition, several projects investigated the use of coatings and other sealants that would serve as an additional barrier to radon diffusion (see, for example, Culot and others 1978). Although removal of some of the high-radium-concentration materials had an effect, indoor radon concentrations in some cases were not reduced commensurately. As additional measurements of indoor radon concentration were conducted, houses were found with high indoor radon concentrations that had no known anthropogenically enhanced radon source (Sachs and others 1982). At the same time, mass-balance considerations (similar to equation 8.1 above) showed that diffusion alone had only a slight potential to produce the high indoor concentrations that were being observed (Bruno 1983). Although removal of high-radium-concentration source materials can be part of an overall radon-control method, it generally is not part of current practice,

OCR for page 141
--> because it is not necessary (except in rare cases). The vast majority of radon-mitigation systems now installed in existing houses rely on mechanically driven, or active, subslab depressurization (ASD) techniques (Henschel 1994). These methods seek to reverse the pressure gradient across the part of the building shell that is in contact with the soil. As noted earlier, this pressure difference drives the advective flow of radon-bearing soil gas into a building. The systems are sometimes referred to as subslab ventilation systems, but as a general rule that is a misnomer. When operated in a depressurized mode, the system does draw some outdoor air from the surface into the soil near the building. It also draws air into this region from the basement (reversing the flow of gas in the cracks and openings in the building shell). The flow of air may dilute the soil-gas radon concentration in the vicinity of the building somewhat, but the extent depends on the permeability of the soil. The key operating principle is still reversal of the indoor-outdoor pressure gradient. Operationally, a subslab system consists of one or more pipes that penetrate the floor slab. The pipes, typically 7–15 cm in diameter, run vertically through the house and terminate above the roof. A mechanical fan, usually an in-line axial fan designed specifically for this application, is installed in the pipe system where it passes through the attic or some other location outside the conditioned living space of the house. The fan operates at about 100–400 m3 h-1 at a pressure of up to a few hundred pascals (Henschel 1993). In an ASD system, the fan creates a low-pressure zone in the soil outside the building shell. A successful system will reverse—or at least reduce—the pressure gradient at all major building-shell penetrations that are in contact with the soil. An important entry pathway for soil gas in many basement structures is the expansion-contraction joint at the edge of the concrete floor slab where it abuts the wall. In some cases, there will also be openings or utility penetrations through the basement walls or, as in the case of walls constructed of hollow-core "cinder" or concrete block, the wall itself is permeable to air flow. To eliminate or reduce soil gas entry in these areas, the low-pressure zone must extend beyond the region of the floor and up the walls. Almost all the retrofitted ASD systems are successful in reducing indoor radon concentrations to less than 150 Bq m-3 and often concentrations are reduced to about 75 Bq m-3. In some cases when the basement walls are constructed of blocks, depressurization pipes are inserted into the hollow cores of the blocks themselves. Because these cores are typically interconnected, directly or through thin permeable concrete "webs," there is in effect a depressurized plenum within the walls themselves, thus largely eliminating any flow of soil gas across the wall and into the building interior. Over the last decade, a considerable amount of research and practical experience on the installation of these systems has been accumulated (Henschel 1993). Two elements aid the successful implementation of an ASD system. One is the presence of a high-permeability gravel layer below the floor slab. This layer essentially establishes a low-flow-resistance pressure plenum that enhances the

OCR for page 141
--> lateral extension of the pressure field, as described earlier. In some cases, the presence or continuity of the gravel layer cannot be easily determined. In these cases, a ~ 1 m diameter sump or pit is dug into the soil below the slab at the point where the ASD pipe extends below the slab. The pit helps to ensure that the pressure field created by the ASD extends as far as possible throughout the region of the soil-building interface. The second practical element in the implementation of an ASD system is sealing as many of the potential radon-entry locations as possible. Although sealing by itself is usually not effective in eliminating radon entry, sealing does enhance the effectiveness of an ASD system because it helps to reduce any short-circuiting of air flow from the building interior into the depressurized region below the floor slab. By reducing this air leakage, the low-pressure field created by the ASD system can be further extended laterally along the soil-building interface. One variant of the subslab system uses a fan to pressurize the region below the floor slab. In this case, the system is providing ventilation of soil gas, thus reducing the radon concentration in the soil region adjacent to the building. Rather than reducing or reversing the pressure gradient across the building shell, this method actually increases the interior-to-exterior pressure difference and so increases the flow of gas from the soil into the building. When successfully implemented, the reduction in radon concentration in the soil gas more than compensates for the increased flow. Careful studies have shown, however, that high soil permeability is key to the successful use of this technique, because it permits a larger dilution effect (Gadgil and others 1994; Turk and others 1991a; 1991b). Basement pressurization has also been used to control radon entry. This method uses the same principle for control as does an ASD system, but it pressurizes the entire basement volume to reverse the indoor-outdoor pressure gradient. Successful use of the technique in a research-house study provided strong empirical evidence that radon entry into buildings is dominated by advective transport. However, as a practical matter, use of the technique has been limited to basements that can be made very tight with respect to air leakage, particularly the membrane between the basement and first floor. Pressurization is done with conditioned air, usually drawn from the first floor. If flow rates are too large, a substantial energy (and in some cases comfort) penalty is associated with heating or cooling the extra "make-up" air as it infiltrates into the house. This method can also create backdrafting problems for fireplaces or other combustion appliances on the first floor (Turk and others 1991a; 1991b). The source-control methods described thus far for use in existing houses are all mechanically driven (that is, fan-powered), so-called "active" methods. Two other techniques—both passive—have been used. The first technique, sealing, has been noted earlier. Empirically, this method has not been found to reliably produce substantial reductions in radon entry, largely because it is often difficult

OCR for page 141
--> to find or satisfactorily seal all the leakage pathways. The problem is acute when part of the soil-structure interface has a low resistance to flow, as when there is a gravel layer below the floor or when the basement walls are constructed of hollow-core block. The second passive (nonmechanical) technique that has been used as a retrofit mitigation system with some (but not uniform) success is the passive thermal stack. Similar in some respects to ASD systems, it consists of a pipe system that is inserted through the floor, and passes through the house and out through the roof. It is important that the pipe pass through the heated portion of the house, because it relies on heat transfer from this conditioned space to heat the air column inside the stack, thus creating the thermal stack effect. There is a small amount of pressure loss at each bend in the pipe, so it is also important to minimize the number of bends in the pipe system as it passes through the house. For this system to be effective, the pressure field developed by the stack below the floor slab needs to be sufficient to reverse or at least substantially reduce, the pressure gradient between the soil and the building interior, which drives advective flow of gas from the soil into the building. The soil-to-building-interior pressure difference will be greatest when the inside-the-stack-to-outdoor temperature difference is the largest, for example, during the winter in cold or moderate climates. This is the same period when the advective transport of soil gas into the building is potentially the greatest. The influence of wind can complicate the behavior of a passive stack system. As described earlier, wind can depressurize the building interior, in addition to the depressurization caused by the stack effect. Wind can also affect the flows and pressures at the stack opening, depending on the wind direction with respect to the orientation of the roof. It is important to have a high-permeability zone below the floor to ensure that the pressure field created by the passive stack extends along the soil-building interface, especially inasmuch as the pressure field generated by the stack is typically 1–10 Pa less than the air pressure inside the building, compared with the 100-to 400-Pa pressure difference generated by an ASD system (Gilroy and Kaschak 1990). In an existing house, the presence of such a layer and the extent to which it is present throughout can be difficult to determine. Concentration Reduction Unlike source-control methods, which seek to limit radon entry, concentration-control methods are designed to reduce radon or radon decay-product concentrations in indoor air. Three concentration-control techniques will be described in this section. As mentioned earlier, increased ventilation can reduce both radon and radon decay-product concentrations, as long as it does not enhance the indoor-outdoor pressure difference. In one set of experiments conducted in a house, basement radon concentrations were observed to be lower when the basement windows were open. Measurements conducted with a tracer gas showed that basement

OCR for page 141
--> ventilation increased somewhat with the windows open but that the largest effect was due to reducing the indoor-outdoor pressure difference across the basement wall (Cavallo and others 1996). One method of increasing ventilation while avoiding some of the energy and comfort penalties noted earlier is to use an air-to-air heat-exchanger system, often referred to as a heat-recovery ventilation (HRV) system. In this approach, designed around commercially available HRV units, ventilation air is exhausted through a heat exchanger through which incoming unheated air also passes. The heat-exchange process is about 40–80% efficient thermally and substantially reduces the energy cost of increased ventilation (Turk and others 1991a; Fisk and Turiel 1983). Such systems have been used successfully for radon control, especially in houses with basements. Because most radon entry occurs through the basement floor and walls, basement radon concentrations are often higher than elsewhere in a house. Use of an HRV to reduce radon concentration in this space, as opposed to the whole house, means that the effective ventilation rate of the space is higher (which affords more control); by controlling basement concentrations, it also reduces radon levels throughout the house. One important ancillary benefit is that the HRV can be used to alter the basement pressure somewhat and will thus provide some additional radon-concentration reduction via source control (Turk and others 1991a). Another method that has had very limited use is based on the sorption of radon gas by activated carbon (Bocanegra and Hopke 1989; Brisk and Turk 1984). A commercially available device based on this approach consisted of two carbon beds; one removed radon from indoor air flowing through it while the other was being purged of accumulated radon by having outdoor air passed through it and exhausted to the outdoors. The two beds were switched periodically so that the freshly purged bed was used to accumulate radon and the bed used for sorption began to be purged (Wasiolek and others 1993). Overall performance of this method is limited by the rate of air flow through the device, which in turn helps to determine the charcoal bed thickness. Like the HRV system, this approach appears to have the greatest applicability in radon control for a basement. The third concentration-reduction approach is the use of air-cleaning to reduce radon decay-product concentrations. Unlike their chemically inert parent, the decay products 218Po, 214Pb, and 214Bi are metals and easily attach to the surfaces of any aerosols that are present (the ''attached" mode). Some decay-product atoms, particularly 218 Po, can also remain as ultrafine aerosols (the "unattached" mode, a few nanometers in diameter). Indoor air concentrations of both modes can be reduced by using an air cleaner designed to remove particles. There have been a number of evaluations of air-cleaning systems undertaken in test chambers or actual indoor environments (reviewed in Hopke and others 1990). Some of these systems can effectively remove radon decay products from indoor air. However, the reduction of 218Po is not as large as that of 214Pb and

OCR for page 141
--> 214Bi. At the same time, the particles are removed from the air. As a result, the unattached fraction of airborne activity increases, especially of 218Po. Because the unattached fractions of the radon progeny have been considered to be far more effective in depositing their radiation dose to lung tissue, concerns have been raised regarding the efficacy of air-cleaning as a means of mitigating the hazards arising from indoor radon. A major problem in the previous studies was that the systems used to measure radon progeny were not able to determine the full size distribution, especially in the size range below 10 nm. Estimates of the unattached fractions were made with systems that provide a poorly defined size segregation (Ramamurthi and Hopke 1989). In many cases, the size-measurement methods and results were not clearly stated. In 1990–1992, a research program supported initially by the New Jersey Department of Environmental Protection and then also by EPA undertook field studies to investigate the effects of room-air cleaners on radon progeny concentrations and activity-weighted size distributions (Hopke and others 1993; Wasiolek and others 1993; Li and Hopke 1992; 1991b) A unique, semicontinuous graded screen-array sampling system (Ramamurthi and Hopke 1991) was used to measure the radioactivity associated with indoor aerosol particles in the size range of 0.5–500 nm. In an early set of studies, particles were produced by a variety of activities, such as cooking, smoldering of a cigarette, burning a candle, and operating a vacuum cleaner. Aerosol behavior in the absence of an air cleaner was determined for each condition (Li and Hopke 1991a). The experiments were then repeated with a high-efficiency filter system operating (Li and Hopke 1992). It was found that the filtration unit reduced the airborne activity concentrations by removing particles, but the reductions in estimated dose were much smaller than the decrease in PAEC. Other experiments in normally occupied houses have involved the measurement of the effectiveness of the filtration unit and an electrostatic precipitator by comparing the cumulative frequency distributions of measurements made during a week while a particular cleaner was operating and measurements made during a background week in which no cleaner was being used (Li and Hopke 1991b). A similar experimental design was used to study the two cleaners and an ionization system in an occupied home (Hopke and others 1993). The results of the 1992 measurements in Parishville, NY, in which two ionizing units were measured along with two filtration units were described by (Hopke and others 1994). More detailed studies of the NO-RAD ionizer system under the controlled conditions of a room-sized chamber at the Lawrence Berkeley National Laboratory were performed, and there are several other ionizer-based cleaners for which there have not yet been field studies (Hopke 1997; Hopke and others 1995b). From the more recent studies on air cleaners and their effects on exposure to and dose from airborne radon decay products, several important conclusions can

OCR for page 141
--> be drawn. With the new dosimetric models that more accurately reflect nasal and oral deposition of ultrafine particles, it is extremely unlikely that an air cleaner can reduce exposure and increase dose as suggested by Maher and others (1987), Sextro and others (1986), and Rudnick and others (1983). Thus, there is no reasonable likelihood that the use of an air cleaner will increase the hazards posed by indoor radon. In studies of different types of air-cleaning devices, reductions in exposure have always exceeded reductions in dose. However, cases have been observed in which there has effectively been no reduction in dose. Thus, for many air cleaners, the clean-air delivery rate is insufficient to provide substantial protection from the radon decay-product hazard. The air cleaner might be effective in removing other contaminants—including cigarette smoke, dust, pollen, and spores—and thus provide a considerable benefit to an occupant without lowering the radon-progeny risk substantially and, more important, without raising that risk at all. The experiments with the newest systems suggest that the combination of substantial air movement and ionization could provide sufficient reduction in exposure and dose to be effective in reducing the radon-progeny risk at radon concentrations up to around 400 Bq m-3. If it is desirable to reduce the risk to that equivalent to the average dwelling in the United States, then such units would be useful only for lower 222Rn concentrations. There would also need to be multiple units in a home to provide complete room-to-room reduction. Mitigation Methods for New Construction Radon-mitigation methods for new buildings can be incorporated directly into the construction process and both enhance the performance of the system and reduce the cost of installation, compared with the cost of retrofit mitigation methods. Systems for controlling radon concentrations described earlier have essentially the same applicability whether their use is in existing or new buildings, and they will not be discussed further in this section. Some cost savings might be associated with installation of systems like an HRV during the construction process or with integrating such a system into the space-conditioning system of the building. In the following sections, the application of "existing-house" techniques for radon-entry control is discussed briefly and then a more systemic approach for making buildings radon-resistant as part of the construction process is discussed. Application of Existing-House Radon-Entry Control Methods As described earlier, one of the most widely used radon-mitigation techniques is ASD. Key to the successful implementation of these systems is reversal of the pressure gradient at all the major soil-gas entry points. Typically, this

OCR for page 141
--> relative to the small mass of them created from the decay of radon (Reed and Arunachalam 1994; Rubin and Mercer 1981) (for example, GAC can sorb lead at 6.2 × 104-1.9 × 106 µg kg-1 at a pH of 6.5). Some groundwater also contains radium and uranium in addition to radon. Uranium can sorb directly to the GAC, but its fate is a function of the pH of the water. At a pH greater than 7, the poorly sorbed, negatively charged carbonate species of uranium, UO2 (CO3)2-2, is predominant. At a pH lower than 6.8, the neutral species, UO2CO3, is predominant and can sorb to GAC (Sorg 1988). Radium is poorly sorbed to GAC (Kinner and others 1990; Clifford 1990; Kinner and others 1989; Sorg and Logsdon 1978) because it forms a hydrophilic species, RaSO4, in water. The pattern and rate of accumulation of uranium, radium, and 210Pb in a GAC unit can be quite different if iron is present. Cornwell and others (1999) found high concentrations of these radionuclides associated with iron-rich backwash residuals from GAC units. That is because radium readily associates with ferric hydroxide and negatively charged metal oxides and hydroxides (Clifford 1990). Uranium also reacts with iron (Clifford 1990; Sorg 1988). Operational Issues. During operation of a GAC unit, an equilibrium is established between the radioactivity of radon and its short-lived progeny sorbed to the carbon. The primary problem resulting from retention of radionuclides is worker exposure to gamma emissions from 214Bi and 214Pb. The maximum occupational accumulated dose equivalent per year recommended for radiation workers in the United States is 50 mSv (EPA 1987a). However, EPA has stated that "there is no need to allow" workers in water-treatment facilities that remove naturally occurring radionuclides from water to receive such high annual radiation doses. It further suggests that these workers' annual accumulated dose equivalent should be "well within the levels recommended for the general public" of 1,000 µSv. Hence, EPA has recommended a maximum annual administrative control level of 1,000 µSv until more experience with such situations is gained. Lowry and others (1988) measured the gamma-exposure fields surrounding 10 point-of-entry units treating water with radon at 96 to 28,074,000 Bq m-3 and achieving removal efficiencies of 83% to over 99%. The gamma exposure rates measured at about 1 m were considerable, in all but one case, because the radon concentration removal was very large (Cnet = 611,000 to 27,926,000 Bq m-3). Except for the 27,926,000-Bq m-3 case, the measurements are in agreement with the range of the calculations from the extended source model. The gamma exposure to workers can be decreased by using water or lead shields around the GAC units. Lowry and others (1991; 1988) studied the effect of water shielding and lead jackets on the point-of-entry units' gamma exposure fields. For example, at site 9 (table 8.3) (28,074,000 Bq m-3) the maximum gamma-exposure rate at the unit's surface was 73 mR h-1. With a 76.2-cm water shield, this was reduced to 8.0 mR h-1. A 61.0-cm water shield reduced a maximum surface gamma-exposure rate of 4.0 mR h-1 to 0.4 mR h-1 at site 5; and at

OCR for page 141
--> Table 8.3 Maximum Gamma-Exposure Rates and Equivalent Dose Rates from Some point-of-Entry GAC Units at a Distance of 1 ma Site Average Radon Cob (Bq m-3) Average Radon Ctc (Bq m-3) Exposure Rate (mR hr-1) Equivalent Dosed (µSv hr-1) 1 97,000 14,000 NAe NAe 2 613,000 2,200 0.40 4.0 5 1,989,000 165,000 0.186 1.86 5 (with 61-cm water shield) 1,989,000 165,000 0.040 0.40 9 28,104,000 175,000 1.73 17.3 a Lowry and others (1988). b Input water concentration. c Treated water concentration. d Equivalent doses were calculated by the committee and were not included in Lowry and other (1988). See appendix E for method of calculation. e NA= gamma radiation so low that it was impossible to measure with acceptable accuracy. about 1 m, there was also a significant reduction. Lead shielding (0.64 cm thick) also reduced gamma exposure. It is possible to estimate the equivalent dose in millisievert per hour from a GAC unit that removes radon from water if an extended-source model is used (see appendix E). This type of model calculates radiation doses in the vicinity of an extended source of radioactive material including self-shielding. This approach is used in the nuclear industry and in radiation protection to predict the radiation levels associated with a variety of radioactive materials (such as ion exchange units that treat nuclear-reactor cooling water and steam condensate). GAC units operating in the United States are treating 11-981 m3 of water per day (table 8.1). GAC is also used in point-of-entry applications (water flowrate of 1 m3 d-1). For the purposes of calculating the equivalent dose with the extended source model, the committee used the suggestion made by Rydell and others (1989) that GAC should be used only to treat water that contains radon at less than 185,000 Bq m-3. The unit would be required to meet the MCL, which is assumed to be 25,000 Bq m-3 for this example. (The committee makes no recommendation or endorsement of a specific value for the radon MCL and uses 25,000 Bq m-3 in order to provide a framework for the example.) The results (table 8.4) indicate that as the radon loading (becquerel applied per day) increases with increasing water flow rate over the range of 1 to 981 m3 d-1, the time until the 1,000 µSv maximum equivalent dose is reached decreases from about 7,000 h to about 150 h (1 m from the GAC tank surface). In many cases, it is unlikely that water-treatment plant personnel would need to spend hundreds of hours per year near the GAC units. In fact, the actual number of hours of exposure per worker would be different for each water supply. Certainly, the number of hours of

OCR for page 141
--> Table 8.4 Estimated Equivalent Gamma Dose for Workers at Water-Treatment Plants or in Point-of-Entry Applications Using GACa Water Flow (m3 d-1) Estimated Gamma Dose at 1 m (µSv hr-1) Hours until 1,000 µSv Dose 1 (point of entry) 0.14 7,143 11 (water plant) 0.75 1,333 981 (water plant, pressure driven) 7.0 143 981 (water plant, gravity driven) 6.4 156 a See appendix E for calculations. exposure would need to be monitored to ensure worker safety. If it were necessary to work on a GAC unit for a substantial number of hours, the gamma emissions could be reduced by first taking the unit off line for about 20 to 30 d to allow the radon responsible for the short-lived progeny to decay. The calculations also show that as radon loading increases, the dimensions of the tank increase, providing increased absorption of gamma radiation within the tank. Modeling gamma emissions from the tank as a point source will be satisfactory only for very small units (for example, point-of-entry applications). The equivalent gamma dose from a GAC system that removes radon from a public water supply should be modeled with an extended-source model that can be modified to the dimensions of the treatment units. It is clear that treating water that contains more radon (over 185,000 Bq m-3), where high removal efficiencies are required, or at high flow rates (high radon loading) will probably lead to unacceptable equivalent gamma doses to water-treatment plant personnel. Rydell and Keene (1993) have developed a computer program (CARBDOSE 3.0) that calculates the probable gamma-exposure dose with distance from a typical point-of-entry unit. The program "approximates a 25.4 cm diameter, 12.7 cm high cylindrical volume of GAC as a cylindrically-corrected 24 cm × 24 cm × 13 cm array of 1 cm3 sources using the 72 gamma energies reported for 214Bi and 214Pb and allowing for self absorption and build-up." Rydell and others (1989) reported that the CARBDOSE models' estimated gamma dose rates and the measured values for 10 point-of-entry GAC units were in "reasonably good agreement." They suggest that CARBDOSE can be used as a design tool to estimate the potential gamma radiation exposure during operation of the GAC unit. The committee notes that CARBDOSE should only be applied to GAC units that have very small dimensions (that is, ones that treat very small flows) and are similar to those used in developing the model. Equivalent gamma doses for larger GAC units should be predicted with an extended-source model that can address more-complex geometries. Disposal Issues. A few weeks after a GAC unit ceases operation, the major radionuclide remaining sorbed to the carbon is 210Pb because of its relatively long

OCR for page 141
--> half-life (22.3 y). In cases where iron is associated with the GAC or the raw water has a pH less than 6.8, uranium and radium can also be found. These species can pose problems for the long-term disposal of the GAC. No federal agency currently has legislative authority concerning the disposal of drinking-water treatment-plant residuals that contain naturally occurring radionuclides (Cornwell and others 1999). If the GAC is transported to a site for disposal, the Department of Transportation could regulate its shipment. EPA has published two guidelines that suggest how such wastes might be handled (EPA 1994c; 1990). However, the states are responsible for regulation of naturally occurring radioactive materials (NORM). There have been three detailed reviews of federal and state guidelines and regulations regarding NORM and how they might apply to disposal of GAC used to remove radon from water (Cornwell and others 1999; Drago 1998; McTigue and Cornwell 1994). The EPA (1994c) guidelines for disposal of water-treatment residuals are centered on the levels of uranium and radium present (for example, in spent GAC or backwash residuals) (table 8.5). Unlike its 1990 draft guidelines, EPA's 1994 version did not cite specific action levels for 210Pb. Instead, because of a lack of conclusive technical data, EPA recommended that the impact of 210Pb contamination be considered case by case. Most states also address NORM wastes on a case by case basis (Drago 1998); the exceptions are Illinois, Wisconsin, and New Hampshire, which have established disposal criteria (Cornwell and others 1999). The Conference of Radiation Control Program Directors published a draft set of suggested state regulations for technologically enhanced NORM (TENORM), naturally occurring radionuclides whose concentrations have been enhanced by technology (for example by such practices as water treatment). Lead-210 associated with GAC is not specifically addressed in this document, and materials with 226Ra or 228Ra at less than 0.19 Bq g-1 are exempt. The draft recommends flexibility Table 8.5 EPA Suggested Guidelines for Disposal of Naturally Occurring Radionuclides Associated with Drinking-Water Treatment Residuals Radionuclide Bq g-1 (dry weight) Suggested Disposal Site Radium <0.11 Landfill   0.11–1.85 Covered landfill   1.85–74 Possible RCRA facility (case by case review)   >74 Low-level radioactive-waste facility Uranium <1.11 Landfill   1.11–2.78 Covered landfill   2.78–27.8 Possible RCRA facility (case by case review)   >27.8 Low-level radioactive-waste facility 210pb — Caution and thorough state-agency review of water treatment and waste disposal plans   Source: EPA (1994c).

OCR for page 141
--> in regulating TENORM as long as members of the public receive less than 1 × 10-3 Gy y-1 from all licensed sources (including TENORM). If spent GAC from a water-treatment plant had enough 210Pb, radium, or uranium associated with it to warrant disposal at either a low-level radioactive waste site or a naturally occurring and accelerator produced radioactive materials (NARM) site, this could have a substantial impact on operation and maintenance costs for the water utility. Actual disposal costs have been estimated as $335 m-3 yr-1 (Kinner and others 1989), approximately $48,000 m-3 (McTigue and Cornwell 1994) and about $11,100 m-3 y-1 (Cornwell and others 1999). In addition, broker and transportation fees would likely be assessed. A typical broker would send trained personnel to the treatment plant to dewater the bed, load and seal the GAC in containers, and decontaminate the site. Cornwell and others (1999) estimated the broker fee at $5,000 (mostly associated with time and travel). Perhaps the biggest question surrounding GAC disposal is the availability of sites that will accept such radioactive material. Drago (1998) reported that two sites are operating (in Barnwell, SC, and Richland, WA). (Note: Clive, UT, receives only limited low-level and NARM wastes.) However, these facilities are not available to all states. Rather, the Low Level Radioactive Waste Disposal Policy Act (PL-99-240) enacted in 1980 and its amendments (1985) direct states to form compacts with their neighbors and designate a host low-level disposal site. There are nine compacts and one other pending, and five states, Washington, DC, and Puerto Rico are unaffiliated. Low-level disposal sites have been proposed by some of these compacts, but none has been built. The result is that low-level waste generators in all states except North Carolina have access to a disposal facility (Drago 1998), but new facilities are not likely to be readily available in the near future. Several ways of avoiding the need to dispose of the GAC at a low-level waste facility would not require changing legislation or regulations. Perhaps the easiest would be to dispose of the GAC before radionuclide accumulation necessitates special disposal. McTigue and Cornwell (1994) developed a model that allows operators to predict when a bed is reaching such a level with respect to 210Pb. The CARBDOSE model (Rydell and Keene 1993) makes a similar prediction for POE GAC units. These models are simple to use, and periodic measurements of the actual 210Pb accumulation on the GAC can be made to confirm their estimates. It should be noted that the models do not address the effect of GAC-associated-iron on the 210Pb accumulation (Cornwell and others 1999). If substantial amounts of iron were present in the raw water, such a prediction would be more difficult. Another alternative to disposal of the spent carbon is thermal regeneration of the GAC that Lowry and others (1990) showed was possible. Both 210Pb and its progeny are volatilized at 850 °C. It is not clear whether release of the 210Pb or 210Po to the atmosphere would be acceptable. If those radionuclides were collected in an air scrubber, they would potentially still present a radioactive-waste disposal problem with respect to the fly ash. Acid regeneration of the spent GAC

OCR for page 141
--> is also possible (Lowry and others 1990). In this case, 210Pb, like stable lead (Reed and Arunachalam 1994), would desorb from the GAC and enter the acid-regenerant solution. However, the spent acid could become a radioactive waste that requires special disposal. Several authors (Cornwell and others 1999; McTigue and Cornwell 1994) explored the possibility of the GAC's being returned to the vendor (an approach used for GAC used to treat VOCs or substances that impart taste and odor to water). However, the willingness of the manufacturers to do this with radioactively contaminated GAC is not clear, especially for small quantities of GAC (less than 9,100 kg). The best option overall with respect to disposal appears to be use of GAC in sites where the potential for 210Pb accumulation is minimized (that is, where the radon and iron concentrations in the raw water are low or the water flow rate is low. This would ensure fairly long operating times before the 210Pb reached a critical level likely to necessitate special disposal. The low radon loading would also result in lower risk of worker exposure to gamma radiation. Long EBCT and High Cost Bench-, pilot-and full-scale studies of GAC removal of radon have produced estimates of the KSS (adsorption-decay constant, see appendix C) for different carbons (Cornwell and others 1999; Kinner and others 1993; Lowry and others 1991; Lowry and Lowry 1987). Lowry and Lowry (1987) found that the best carbon for radon sorption was a coconut-based GAC (KSS = 3.02 h-1). This carbon has a larger percentage of micropores (0.002 µm) than other types of GAC. It is hypothesized that micropores are most effective for sorbing small molecules and atoms, such as radon gas (Drago 1998). The cost estimates for GAC treatment have used a KSS of less than 3.02 h-1 (for example, EPA 1987b, KSS = 2.09 h-1). Recent studies by Cornwell and others (1999). specifically designed to calculate KSS values for different carbons, found that for one groundwater with low iron and TOC concentrations, the KSS ranged from 3.5 to 5.2 h-1. These higher values suggest that GAC could be a much more cost-effective option at some sites than originally thought. For example, with a raw-water radon concentration of 111,000 Bq m-3, a flow of 39 m3 d-1, and a KSS of 4.5 h-1, the EBCT and amount of GAC needed to achieve an MCL of 11,000 Bq m-3 (90% removal) would be 31 min and 0.83 m3, respectively, compared with 66 min and 1.8 m3, respectively, if the KSS were 2.09 h-1. Assuming a cost of $883 m-3 of GAC (Cornwell and others 1999), this 54% reduction translates to an $848 savings. However, Cornwell and others (1999) also suggest that a pilot study must be conducted at each site where GAC is being considered, because the KSS will likely be different for each groundwater. For example, for a low-iron, low-TOC groundwater in New Hampshire, a lignite-coal based GAC had an

OCR for page 141
--> average KSS of 4.52 h-1. At a site in New Jersey that also had low iron and low TOC, the average KSS for the same carbon was 2.54 h-1. Microbial Risk GAC units often support microbial populations because they provide a good surface for attachment and concentration of organic carbon and nutrients (Camper and others 1987; Graese and others 1987; Camper and others 1986; 1985; Wilcox and others 1983). As a result, GAC units that treat radon have had high concentrations of heterotrophic bacteria in their effluent (Cornwell and others 1999; Kinner and others 1990; 1989). Because heterotrophic plate counts could periodically exceed 500 colony forming units per milliliter, which would violate the proposed GWDR, the treated water would need to be disinfected before distribution (EPA 1992c). Either chlorination or ultraviolet disinfection could be used. Unlike the situation with aeration systems, it is less likely that disinfection byproducts would be formed if chlorination were used after a GAC unit. Disinfection byproducts result from the reaction between chlorine-based disinfectants and naturally occurring organic matter in the water. GAC can sorb the low levels of naturally occurring organic matter in the groundwater (Cornwell and others 1999) until it becomes saturated (Kinner and others 1990). Before saturation, the risk posed by disinfection byproducts would be minimal. Thereafter, it could be similar to that of an aeration system that uses chlorine-based disinfection. Precipitate Formation In groundwaters that have high levels of iron, precipitates might accumulate on the top of the GAC bed. This reduces the hydraulic head and contaminant-removal efficiency of the GAC and makes it a poor choice for radon treatment for these types of raw water. Although some iron precipitation has been observed in field evaluations of GAC units that treat radon with low iron (Cornwell and others 1999; Kinner and others 1990; 1989), the problem is usually less acute than in aeration treatment; the water is not usually oxygenated and exposed to atmospheric conditions, so, much less oxidation of the iron occurs. If iron precipitates do form, they present the same problems outlined for aeration systems. Pretreatment to remove the iron before the water enters the GAC is unlikely to reduce the disposal issue unless sequestering agents are used to prevent precipitation. Pretreatment, such as with ion-exchange, would just accumulate the long-lived radionuclides on the resin, also presenting a disposal problem. In addition, Lowry and others (1990), Lowry and Brandow (1985) and Dixon and Lee (1988) have noted that backwashing releases sorbed radon to treated water, although this has not been observed in all cases (Cornwell and others 1999; Kinner and others 1990; 1989). Desorption of the nongaseous radon progeny has not been observed during backwashing (Lowry and others 1990). In this

OCR for page 141
--> case, disposing of the initial water produced after backwashing and not sending it to the consumer, could eliminate the risk of exposure from release of desorbed radon into the water supply. Water Storage Because radon has a relatively short half-life, it is possible to obtain some reduction in concentration by storing the water. It can be stored in separate storage tanks or in those normally used to provide water to a community during periods when demand exceeds the yield of wells. Over 24 h, radon reduction due to storage averages 20–40%, and 5–6 d of storage yields 80–90% losses (Kinner and others 1989). In two waterworks in Sweden, losses of 17–34% were documented during storage (Mjönes 1997). The small reductions mean that this method of treatment is effective only where the percentage removals required to meet the MCL are relatively low and the daily demand for water is small. For example, storage might be an adequate treatment for a school that uses a well to supply water (that is, is not served by a community supply). Repumping is usually required with storage systems used for radon reduction because they are typically operated at atmospheric pressure. Repumping can be avoided if the storage tank is elevated. Perhaps the biggest problem with this method is providing a reliable and consistent quality of water. As demand fluctuates, the retention time in the storage tank can change, potentially resulting in smaller radon reductions. To avoid that problem, the capacity of storage needs to be increased to ensure acceptable overall radon removal. The tanks usually are vented to the atmosphere, which would increase the risk of air emission as with aeration methods. However, this risk would probably be low because use of storage as a treatment method would be limited to very small water supplies that have relatively low radon concentrations. If radon loss is due solely to decay and not to losses to the atmosphere, there is also the risk of exposure from ingestion of radon progeny such as 210Pb. Because the storage tanks are typically vented to the atmosphere, they might require disinfection under the pending GWDR. As a result, there could be an increased risk of exposure to disinfection byproducts if chlorination is used. Simple Aeration During Storage Losses observed after water passes through a storage tank are often higher than those measured when the water resides in the tank undisturbed (Mjönes 1997; Kinner and others 1989). The increased loss has been attributed to aeration that occurs when the water splashes into the tanks. Indeed, the mode of entry is very important. Bottom entry below the water line yields removals similar to loss due to storage alone. Free-fall or entry via spray nozzle or splash box can increase removals to the range of 50–70% (Mose 1993; Kinner and others 1987). Adding a crude coarse bubble-aeration system to the tank can boost removal to the 90%

OCR for page 141
--> range (Kinner and others 1987). The issues associated with these systems are the same as those outlined for storage alone (acceptable percentage removals in spite of variable retention times) and for aeration systems (disinfection byproduct formation upon disinfection, off-gas emissions, and precipitate formation). Blending Some public supplies (for example, for some large communities) use a combination of surface water and groundwater. In this case, radon-free or very-low-radon water might be available to mix with the groundwater, which would result in reduction in concentration due to dilution. Blending has been documented as effective in some water supplies (Kennedy/Jenks 1991b; Dixon and Lee 1988), but reductions are usually low (20%) (Drago 1998). The use of blending depends on whether the water is monitored for compliance as it enters the distribution system or when it reaches the first tap (consumer). In the former case, both types of water would need to be available at the same location. In addition, blending can be effective as a best available technology only if mixing is complete. Reverse Osmosis In the 1991 proposed rule, EPA specified reverse osmosis (RO) as a best available technology for uranium, radium, and beta-and photon-emitters, but not for radon. RO systems use semipermeable membranes and pressure to separate dissolved species from water. In Sweden, Boox (1995) used an RO filter to treat water in two homes. The systems were installed to improve the taste of the water, but they concurrently reduced the radon content by 90%. RO systems have not been tested for radon removal in the United States, and their use in Sweden exclusively for treatment of radon is doubtful because of their low capacity and relatively high cost. In addition, RO membranes do not work well if turbidity-causing material or precipitates (for example, iron, manganese, silica, and calcium) foul them in the raw water. A brine that contains the contaminants removed from the water is created and must be disposed of. Because the brine is concentrated, the levels of radionuclides might be high, dictating special disposal (as outlined for GAC and aeration systems). Loss in the Water Distribution System There have been several studies of this method of radon removal from drinking water. The majority of the decrease in radon results from decay during transit or storage in the closed piping network. Most of the studies have documented losses in the range of 10–20% (Rand and others 1991; Kinner and others 1989; Dixon and Lee 1988). A study in Sweden of four waterworks (Mjönes 1997)

OCR for page 141
--> found higher losses (30–70%), but some of this could have been due to volatilization during pumping or mixing and agitation. If EPA requires that water meet the MCL when it enters the distribution system, this method of treatment would not be acceptable. Furthermore, it is doubtful that it would consistently produce water with the same radon content, because retention times and therefore losses differ on the basis of the distance of travel and the demand. This method probably would be used only where the radon levels entering the distribution network are relatively low. Field and others (1998; 1995) have shown that radon levels in a distribution system can actually increase when radon is released during decay of radium associated with iron-based pipe scale. A generation rate would have to be quantified if loss in the distribution system were considered as a treatment alternative in water supplies that have this type of scaling problem. It also has implications for where water samples are collected (at the origin of the distribution system or at the point-of-use) (Field and others 1995). Vacuum Deaeration and Hollow-Fiber Membrane Systems These are very new systems that have undergone only laboratory-or pilot-scale testing (Drago 1998; 1997). They have been developed to address the issue of off-gas emissions associated with aeration systems. In both technologies, the radon removed is trapped in a sidestream of water rather than being released to the air. Therefore, these systems have the potential to fill a niche where radon concentrations in the raw water are high (precluding use of GAC) and air emissions of radon are prohibited (precluding use of aeration systems alone). The sidestream water is passed through a GAC bed that sorbs the radon. The GAC is effective in these cases because the radon is dissolved in water, not in a vapor phase. Descriptions of vacuum deaeration and hollow-fiber membrane systems are found in appendix C. The issues of gamma emissions and disposal of the spent GAC used in the vacuum deaeration and hollow-fiber membrane systems are similar to those for GAC when it is used directly to remove radon from water. Though disinfection might be required to prevent biofilm development on the GAC, microbial and disinfection-byproduct risks are not applicable, because water from the GAC unit used in vacuum deaeration and hollow-fiber membrane systems will not be released to the consumer. The issues of precipitate accumulation and backwashing are also minimized because the sidestream water can be fairly clean. The low transfer efficiency of radon from the sidestream water to the GAC dictates a long EBCT. This, along with the complexity of the systems, would increase the costs of these systems.

OCR for page 141
--> Conclusions Several aeration methods exist or are being developed to remove radon from drinking water. Aeration was designated as the best available technology by EPA in its 1991 proposed radon rule; however, some issues and secondary effects are associated with these technologies. They discharge radon into the air (intermedia pollution), and the extent to which the off-gas emissions will be regulated is not clear. Removing radon from the off-gas is much more difficult and expensive than removing it from water. Although aeration is, in general, a straightforward technology to use, aeration of groundwater that contains dissolved species, such as iron, can lead to the formation of precipitates (scale) that can cause operational problems, aesthetic concerns, and disposal problems (for example, iron precipitates enriched with long-lived radionuclides). GAC was not listed as a best available technology by EPA in the 1991 proposed radon rule, but it might be an option for small systems that require only minor treatment to meet the MCL. Several issues and secondary effects are associated with GAC. The primary one is retention of radionuclides, which can lead to substantial gamma emissions from the unit and pose a potential problem of radiation exposure for plant operators. Equivalent gamma doses from GAC units that remove radon from public water supplies should be predicted with an extended-source model that can address more-complex geometries. Accumulation of radionuclides can also lead to potential disposal problems for the spent GAC. Both those issues are most problematic when radon loadings are high (that is when treated flows are high or large amounts of radon are retained on the GAC). The treated water leaving aeration and GAC water-treatment systems might require disinfection on the basis of the pending GWDR. If chlorination is used as the disinfection method, trihalomethanes, which are known to cause cancer in rats and mice, could be created. The committee estimated that the average lifetime risk associated with disinfection byproducts formed during chlorination of groundwater is on the order of 5 × 10-5. Storage, blending, loss in the water-distribution system, and other newly emerging technologies (such as vacuum deaeration and hollow-fiber membrane systems) are not likely to be major alternatives for removal of radon from drinking water. Storage, blending, and loss in the water distribution system would be limited to situations where only low removal would be required.