NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Research Council (US) Safe Drinking Water Committee. Drinking Water and Health: Volume 4. Washington (DC): National Academies Press (US); 1982.

Cover of Drinking Water and Health

Drinking Water and Health: Volume 4.

Show details

IIIChemical Quality of Water in the Distribution System

Even if one could eliminate the causes of contamination associated with pipe breakages, cross-connections, back-siphonages, and other factors inherent in water distribution systems, there would still be changes in the physical, chemical, and biological properties of the water as the result of either chemical or biological activity.

Chemical activity producing changes in water quality within the distribution system is associated with corrosion, leaching, deposition, and reactions involving water treatment chemicals and their residuals. Each of these topics is discussed separately in this section.

The materials comprising pipes, pumps, storage reservoirs, and other system components can corrode through contact with water or may leach constituents in water over time. Solubility and kinetic factors will determine whether these constituents will deposit (precipitate) onto pipe walls or whether the materials used in the conveyance system will partially dissolve or corrode into the water. Chlorine and other treatment chemicals added at the water treatment plant or in the distribution system itself can continue to react with organic compounds in the water. Thus, the chemical content of water at the consumer's tap may be different from that of water leaving the treatment plant or other source as a result of its contact with materials in the distribution system and the time available for reactions to progress.

Chemical Water Quality Indexes

A number of water quality indexes have been used to predict whether water will corrode materials used in distribution systems or home plumbing units. In most cases, these indexes are used as a criteria for water treatment control, but they can also be used as guide to the selection of materials. Their principal advantage is simplicity, but they are not always perfect predictors. Long-term tests of materials are more costly to conduct, but provide more direct evidence of water quality and its potential to corrode given materials.

The oldest and most widely used index is the Langelier Index, which is based on the solubility of calcium carbonate and the potential of the water to deposit a scale that would protect the pipe. This index has been applied to both metal and asbestos-cement pipe. A simplified version of the Langelier Index, called the Aggressiveness Index, was developed especially for asbestos-cement pipe to predict whether the water will either deposit a protective scale or seek calcium carbonate saturation by dissolving the pipe's cement. A third index, the Saturation Index, is based on solubility characteristics of a number of compounds, not just calcium carbonate. Its potential application for asbestos-cement pipe is discussed below.

Langelier Index

The Langelier Index was developed in 1936 in order to investigate systematically the chemical relationships involved in the corrosion of iron or galvanized pipe (Langelier, 1936). It is sometimes referred to as the calcium carbonate saturation index or simply as the Saturation Index. (To avoid later confusion with the term Saturation Index, which is used for a number of constituents in addition to calcium carbonate, the term Langelier Index is used herein).

The Langelier Index (LI) can be defined as follows:

Image img00001.jpg

where

pHs = saturation pH, the pH at which water of the measured calcium and alkalinity concentration is in equilibrium with solid calcium carbonate and

pH = actual or measured pH of the water.

In its simplest form, which is applicable between pH 7.0 and pH 9.5, the equation for calculating pH5 is as follows:

Image img00002.jpg

where

pK2' = negative logarithm of second dissociation constant for carbonic acid (H2CO3),

K2' = Image img00003.jpg

pKs' = negative logarithm of the solubility product of calcium carbonate (CaCO3),

Ks' = [Ca2+][CO32-],

pCa2+ = negative logarithm of the molar concentration of calcium, and

pAlk = negative logarithm of the equivalents of alkalinity (titrable base), assuming that [Alk] = [HCO3-].

The terms K2' and Ks' are dependent upon temperature and ionic strength, which is a measure of ionic composition of the water. Corrections for temperature and ionic strength are made for each calculation.

The utility of the Langelier Index is that it predicts whether calcium carbonate will precipitate, dissolve, or be in equilibrium with solid calcium carbonate. If it precipitates, calcium carbonate can form a protective scale on pipes including asbestos-cement (A/C) or metal pipe. If calcium carbonate dissolves in water of a given quality, calcium carbonate scale, previously deposited at the water-pipe interface, will be removed, thus exposing the pipe surface to the corrosive effects of the water.

The Langelier Index is interpreted as follows:

When LI > 0, water is supersaturated with respect to solid calcium carbonate and will tend to precipitate and form a scale.

When LI = 0, water is at equilibrium.

When LI < 0, water is undersaturated with respect to solid calcium carbonate and protective calcium carbonate scales on the pipe may dissolve.

Aggressiveness Index

The A/C pipe industry developed the concept of an Aggressiveness Index for use as a guide in determining whether A/C pipe would be appropriate in a given situation. The original purpose of the index was to ensure the structural integrity of the pipe. More recently, it has been used to predict whether water quality degradation would occur from pipe dissolution. The Aggressiveness Index is a simplified form of the Langelier Index and has some shortcomings, which are noted below.

The Aggressiveness Index (Al) is defined as follows:

Image img00004.jpg

where

AI = Aggressiveness Index,

A = total alkalinity, mg/liter as calcium carbonate, and

H = calcium hardness, mg/liter as calcium carbonate.

The Aggressiveness Index does not incorporate the corrections for temperature and ionic strength. At a selected temperature (14°C) and ionic strength (0.01) and by converting to alkalinity and calcium concentrations in mg/liter, it can be shown that:

Image img00005.jpg

Application of the Aggressiveness Index to determine when A/C pipe should be used has been incorporated into standards published by the American Society for Testing and Materials (1976) and the American

Water Works Association (1975b, 1980). The need for water quality guidelines is also acknowledged by the A/C Pipe Producers Association (1980). The most recent standards apply the Aggressiveness and Langelier Indexes to relate water quality and the use of A/C pipe (Table III-1). These standards recommend that nonaggressive water (AI ≤ 12.0) be used with Type I (nonautoclaved) or Type II (autoclaved) A/C pipe. Type II pipe is recommended for moderately aggressive water (AI between 10 and 12). For highly aggressive water, ''the serviceability of pipe for such applications should be established by the purchaser in conjunction with the manufacturer'' (American Water Works Association, 1980). Recognizing the relationship between water quality and the use of A/C pipe, the U.S. Environmental Protection Agency (1979a) recently proposed that the Aggressiveness Index should be ≤ 12 for water transported through A/C pipe in order to prevent adverse effects.

TABLE III-1. The Relationship of Water Quality (Expressed as Aggressiveness and Langelier Indexes) to Asbestos-Cement (A/C) Pipe.

TABLE III-1

The Relationship of Water Quality (Expressed as Aggressiveness and Langelier Indexes) to Asbestos-Cement (A/C) Pipe.

Data published by Millette et al., (1979) provide a perspective on the typical quality of water in the United States as it pertains to the use of A/C pipe. Through a sampling of representative utilities throughout the

United States, they determined that 52% of the water supplies had water that was at least moderately aggressive (Aggressiveness Index between 10 and 12). Furthermore, 16.5% of the water supplies could be classified as very aggressive. They concluded that these data suggest that as many as 68.5% of the U.S. water systems carry water that is potentially capable of corroding A/C Type I pipe and that water supplies with very aggressive waters (< 10) may be significantly corrosive to any type of pipe. including cast iron, galvanized, and other types of pipes.

When using the Aggressiveness Index, one could assume that the mechanism for A/C pipe deterioration by aggressive waters is related to release of calcium from the cement portion of the pipe. If the water is in fact attacking the pipe, the cement could be dissolving into the water. This would leave the asbestos fibers unprotected or not encapsulated within the cement matrix. This would leave the fibers free to be released into the water. These fibers could be released individually or in bundles. Hallenbeck et al., (1978) theorized that once fibers are released into the water, they can be further broken down so that counts of asbestos fiber from the breakdown products are even higher. Thus, if A/C pipe is used. there is a potential for consumers to be exposed to significant concentrations of asbestos in some drinking water supplies.

The use of the Aggressiveness Index represents an advance over the original preconception that A/C pipe is not subject to the effects of water quality. As recently as a decade ago, Bean (1970) stated that A/C pipe does not require lining, even with soft water, which could be classified as aggressive water. Since that time, both manufacturers and pipe producers have acknowledged that it is not judicious to use A/C pipe with aggressive water. Thus, the Aggressiveness Index has been a means for alerting suppliers and users that A/C pipe cannot be used under all situations and that it is not resistant to corrosion in all cases. It is also simpler to calculate than the Langelier Index.

Since the Aggressiveness Index (as well as the Langelier Index) is based on calcium carbonate saturation, it should yield a fairly accurate prediction of "nonaggressiveness" provided by a protective calcium carbonate coating if water is oversaturated (Schock and Buelow, 1980). However, if the water is undersaturated with calcium carbonate, there is no reason to expect the Aggressiveness Index to predict with accuracy the dissolution of A/C pipe since calcium carbonate is only a minor constituent of the cement and calcium silicate is the predominant pipe component. Furthermore, the Aggressiveness Index does not account for temperature and ionic strength as does the Langelier Index. Finally, the Aggressiveness Index fails to account for protective chemical reactions in drinking water.

The Aggressiveness Index has been used for several years by pipe manufacturers and the water supply industry. Therefore, the majority of the data on water quality and A/C pipe deterioration contains information on the Aggressiveness Index, calcium, and alkalinity of the water.

In the absence of a better predictor of pipe performance, this index has been used extensively and is still a simple first approximation for predicting pipe performance.

Saturation Index

The Saturation Index has been proposed by Schock and Buelow (1980) for use in predicting performance of A/C pipe under given water quality conditions. In this approach, both the solubility of pipe components and the possible protective coating of constituents in the water are considered.

The cement matrix of A/C pipe is a complicated combination of more than 100 compounds and phases. Since electrochemical corrosion is not an issue, the corrosion of A/C pipe is governed by solubility considerations. Possible dissolution reactions in A/C pipe include:

Image img00006.jpg
Image img00007.jpg
Image img00008.jpg
Image img00009.jpg

where s indicates the solid phase.

The first constituent, Ca(OH)2, is lime, and the others are tricalcium silicate, dicalcium silicate, and tricalcium aluminate. Solubility constants for pure solids in Reactions 5, 6, and 7 are 10-5.20, 10 -8.6, and 10-16. For Reaction 8, it is not known. The actual solubility constants in pipe are difficult to estimate, since solids in pipe are highly substituted. Schock and Buelow (1980) concluded that these materials are soluble under typical water quality conditions, but that they dissolve slowly. Pipe dissolution by Reactions 5 through 8 would increase pH, calcium, and alkalinity of water in contact with the pipe. The Langelier Index or Aggressiveness Index would also increase. These phenomena have been observed in several studies that are described below.

Schock and Buelow (1980) have also used chemical equilibrium calculations to estimate whether calcium carbonate film would form to protect pipe. Protection by metal precipitation has also been modeled for iron, zinc, manganese, and silica since they could form dense solids.

Models were estimated using the aqueous chemical equilibrium computer program called REDEQL.EPAK (Schock and Buelow, 1980). The thermodynamic state of saturation was quantified by the Saturation Index (SI), defined as the logarithm of the ratio of the ion activity product (IAP) to the solubility product constant (Kso). For example, for hydroxyapatite, the equilibrium reaction is:

Image img00010.jpg

Assuming activity coefficients equal to unity, the Saturation Index (SI) would be:

Image img00011.jpg

If the solid and solution are in equilibrium, IAP = Kso and SI = 0. If the solution is supersaturated, the SI is >0, and undersaturation occurs if SI <0.

The results of SI calculations are shown in Figure III-1. Initial water quality is 1.0 mg/liter calcium, 24 mg/liter total carbonate, 0.24 mg/liter magnesium, 0.5 mg/liter zinc, 0 mg/liter iron, 0 mg/liter phosphate, 20 mg/liter sodium, and 11-33 mg/liter chlorine. Based on this model, zinc hydroxycarbonate [Zn5(CO3)2(OH)6] would precipitate if the pH was higher than 8. None of the other species would precipitate. Schock and Buelow (1980) suggested that zinc hydroxycarbonate, once precipitated, could be converted by reactions with silicates in the A/C pipe to a zinc silicate coating, which is hard and should provide good protection.

Figure III-1. Saturation Index diagram for model system.

Figure III-1

Saturation Index diagram for model system.

This approach to predicting pipe performance by modeling equilibrium characteristics of a number of protective solids in addition to calcium carbonate appears to contribute to the understanding of A/C pipe. Schock and Buelow (1980) have demonstrated the applicability of the Saturation Index to several model systems. Although it is more difficult to use than the Langelier Index, it is expected to produce more accurate predictions.

Corrosion

Uhlig (1971) defined corrosion as "the destructive attack of a metal by chemical or electrochemical reaction with its environment." He also noted that the term "rusting" applies to the corrosion of iron or iron-base alloys to form corrosion products consisting mostly of hydrous ferric oxides. Therefore, other metals can corrode, but not rust.

A principal concern about corrosion in water distribution systems is the possibility that its products will have an adverse impact on the health of consumers exposed to them. Moreover, materials introduced into this system to mitigate corrosion might themselves provide a source of potentially hazardous chemicals. For example, protective coatings on pipes could leach such hazardous constituents into the water, or chemicals added to the water to inhibit corrosion could be toxic. Before discussing all of these concerns, it is necessary to consider some of the mechanisms of corrosion, its inhibition, and measurement.

Although the economic impact of corrosion in water distribution systems is not of direct concern in this report, it is of some importance because it provides an incentive for reducing corrosion. Ultimately, this may have either a positive or negative effect on the generation of corrosion products to which the consumer is exposed. The reduction of metallic corrosion resulting from economic incentives is likely to benefit human health; however, corrosion-inhibiting additives or coatings selected without full awareness of their possibly toxic nature may be counterproductive.

The Corrosion Process

Corrosion is most often considered to be an electrochemical process. That is, electrons move through the corroding metal, and separate (but not necessarily distant) locations at the metal-water interface act as anodes and cathodes for the oxidation and reduction half cell reactions that occur. For example, as described by Larson (1971), the corrosion of an iron surface in contact with water can involve the following reactions:

Image img00012.jpg
Image img00013.jpg

This cathodic reaction will generally occur slowly, but a faster alternative one will occur in the presence of oxygen:

Image img00014.jpg

For both cathodic reactions, two hydroxide ions will be produced and an alkaline condition will result near the cathode. However, the ferrous ion can be further oxidized by oxygen and precipitate ferric hydroxide:

Image img00015.jpg

This clearly generates acid.

In neutral or near-neutral water, dissolved oxygen is necessary for appreciable corrosion of iron (Uhlig, 1971). The initial high rate of corrosion will diminish over a period of days as the rust film is formed and acts as a barrier to oxygen diffusion. The steady-state corrosion rate will be higher as the relative motion of the water increases with respect to the iron surface. Increased temperatures can also increase iron corrosion when it is controlled by diffusion of oxygen to the metal surface.

Because the rates of electrochemical processes are related to the electrochemical potential at the metal-solution interface, processes affecting potential can hasten or reduce the rate of corrosion. This applies particularly to "cathodic protection," which is an important approach to corrosion control. This process involves the external application of electric current, modifying the electrochemical potential at the metal-solution interface, thereby arresting the tendency for metal ions to enter solution. In some portions of the water systems, such as water tanks, a more easily corroded metal such as magnesium or zinc can be used as a sacrificial anode, and cathodic protection is achieved without the use of an impressed source of current.

Two principal types of electrochemical corrosion cells are of concern in water distribution systems (Larson, 1971; Uhlig, 1971). The first results from a galvanic cell, which is due to the contact of two different metals. The rate of the resultant corrosion is increased by greater differences in electrochemical potential between the two metals, as well as by increased mineralization of the water. For such a cell, the anodic metal corrodes, and the cathodic metal is, in effect, protected. Thus, when zinc-coated (galvanized) steel corrodes, the zinc will generally do so at the expense of the iron. Galvanic corrosion can be a problem when, for example, copper is in contact with iron, which will tend to corrode by galvanic action.

The other and often more important corrosion cell is the concentration cell. This cell involves a single metal, but different portions of the metal are exposed to different aqueous environments. Such a cell could be generated by one region of an iron surface exposed to oxygen and another one nearly protected from oxygen by rust or other surface coatings. Similarly, differences in pH, metal, or anion concentrations could generate such a concentration cell. As noted above for the corroding iron system, the anodic and cathodic reactions generate different corrosion products, which can enhance the ability of the concentration cell to cause corrosion.

Corrosion can also be classified with respect to the resulting outward appearance or altered physical properties of the piping (Uhlig, 1971). Uniform corrosion takes place at a generally equal rate over the surface. Pitting refers to a localized attack resulting, in some cases, in marked depressions. In water containing dissolved oxygen, oxide corrosion products can deposit at the pitting site and form tubercles. Dezincification is a corrosive reaction on zinc alloys (e.g., brass, which contains copper) in which the zinc corrodes preferentially and leaves behind a porous residue of copper and corrosion products. Soft waters high in carbon dioxide content may be particularly aggressive to brass. Erosion corrosion can result when the protective (often oxide) film is removed, such as by abrasion occurring in fast-moving waters. Normally, many metals in contact with water will form such a protective oxide coating. One example of erosion corrosion occurs near joints and elbows of copper pipes when water flows at high velocities.

It is apparent from the above discussion that the corrosion process is highly complex and is influenced by a large number of factors, including the nature of the corrodible materials, the physicochemical quality of the water, and the physical structure and hydrodynamics of the distribution system.

Biologically Mediated Corrosion

The role of microorganisms in the corrosion of metal pipe in the water distribution system has been recognized for some time (Hadley, 1948). Microorganisms may influence corrosion by affecting the rate of cathodic or anodic activity, producing corrosive end products and metabolites, creating electrolytic concentration cells on the metal surface, and disrupting or breaking down the protective film (natural or otherwise) at the metal surface. The microorganisms may be heterotrophic or autotrophic and may grow under aerobic or anaerobic conditions.

The pipe surface, joints, valves, and gates provide a wide variety of niches for the growth of many different microorganisms that can alter the chemical and physical habitat and produce conditions very different from those observed in the water passing through the pipe. Although water in the distribution network is generally well aerated, containing several milligrams of oxygen per liter, microenvironments without oxygen may occur in the pipe. Concentrations of organic matter promote the growth of aerobic microorganisms that deplete the oxygen and create anaerobic conditions. Tuberculation, sediments, and pipe joints can yield protected environments in which neither dissolved oxygen nor disinfectant residuals can penetrate.

Under anaerobic conditions, low oxidation-reduction potentials occur, and, in the presence of sulfate, sulfate-reducing bacteria may proliferate. Desulfovibrio desulfuricans can grow autotrophically under the above conditions, reduce the sulfate to sulfite, and oxidize the hydrogen. Uhlig (1971) suggested that an iron surface aids the process by which sulfatereducing bacteria function. These anaerobic bacteria generally possess hydrogenase enzymes that act on hydrogen and require ferrous iron (Booth and Tiller, 1960). Since the possible corrosion products of iron pipe are ferrous iron and hydrogen, the sulfate reducers may provide a mechanism for the continual removal of corrosion products, thereby influencing the equilibrium of the corrosion reaction (Lee and O'Connor, 1975). Tuovinen et al., (1980) conducted a series of experiments in which iron coupons (thin plates) were immersed under various aqueous conditions. He observed that the corrosive effects were less pronounced in sterile solutions than in solutions that contained ground tubercles or cultures of D. desulfuricans. These same microorganisms were isolated from cast-iron coupons in polyvinyl chloride and steel pipe loops. Other anaerobic microorganisms may also play a role in the corrosion in water distribution systems. More detailed discussions of microbiologically mediated corrosion can be found in reviews by Davis (1967), Iverson (1974), Miller and Tiller (1978), and Uhlig (1971).

Aerobic microorganisms have also been reported to contribute to corrosion. Gallionella, Sphaerotilus, and Leptothrix are the predominant members of the "iron bacteria" commonly found in water distribution systems. These microorganisms convert soluble ferrous iron to insoluble ferric iron, which often accumulates on their stalks and sheaths. Olsen and Sybalski (1949) suggested that the "iron bacteria" initiated tubercle formation on pipe walls and that this process was the critical factor for the corrosion of iron pipes. Other microorganisms can remove iron from solutions indirectly. Strains of Pseudomonas, Aerobacter (Enterobacter), and Mycobacterium strains can form precipitates containing iron (Macrae and Edwards, 1972). Sulfur-oxidizing bacteria (Thiobacillus, Beggiatoa) may also contribute to corrosion under aerobic conditions.

Although the activity of microorganisms in water distribution systems has been well documented, the extent and nature of the microbial activity in the internal corrosion of the pipe network require considerable elucidation. Following microbial populations on the pipe in situ has been difficult.

The Effect of Water Quality on Corrosion

Because of the electrochemical nature of the corrosion reaction, the quality of the water in contact with the corrodible surface has a substantial impact on the rate at which the reaction occurs. Constituents in the water are also important because they can form products that coat the surface, and these can similarly affect the corrosion process. Some combinations of water quality factors may lead to high rates of corrosion that may expose consumers to potentially dangerous levels of corrosion products. In such circumstances, alteration of water quality prior to its entry into the distribution system may reduce some risks.

Camp (1963) reported observations of some aspects of water quality that can effect the corrosion of iron. He noted that an increase in oxygen concentration initially resulted in an increased corrosion rate. Subsequently. it reduced corrosion by the formation of oxide films. Waters of low alkalinity and hardness are more corrosive than waters of high alkalinity and hardness, primarily due to their actions in the formation of calcium carbonate precipitate coatings. Other electrolytes that often promote corrosion include sulfates and chlorides. Corrosion is more rapid in acid solution, principally because potentially passivating layers are more soluble in such solutions.

Larson (1971) noted that chloride and sulfate salts increase the corrosion rate of mild steel at a pH below the range where pitting occurs in the presence of oxygen and in the absence of carbonates. Pitting can occur at local unprotected points of corrosion, under deposits of debris, or at the water line of surfaces exposed partly to air and partly to water. In the presence of oxygen, but no calcium, carbonate minerals inhibit corrosion by countering the acceleration effect of such salts as chlorides and sulfates. This inhibiting effect reaches a maximum when the alkalinity is more than 5 to 10 times the sum of the chloride and sulfate and the pH is 6.5 to 7.0. It is at a minimum at pH 8 to 9. When the alkalinity ratio to these other ions is less than 5, corrosion rates increase.

Finally, Larson (1971) noted that from the standpoint of corrosivity the most widely accepted criterion of water quality is the stability of its saturation by calcium carbonate. As discussed above, the Langelier Index can be used to measure this.

A water with a positive Langelier Index is oversaturated with calcium carbonate, which would tend to form a protective coating on the pipe, thereby reducing corrosion. A water with a negative Langelier Index would be undersaturated with calcium carbonate. Thus, a protective coating would not be formed, and corrosion would be more likely to occur. If a water had such a negative index, due to low calcium, pH, and/or carbonate-bicarbonate species, its Langelier Index could be increased by the addition of lime or soda ash. Uhlig (1971) noted that a Langelier Index of +0.5 is often considered to be satisfactory, and that higher values may cause scaling (excessive deposition of calcium carbonate), especially at elevated temperatures. The attainment of values less positive than +0.5 is a common goal in water treatment. However, Larson (1971) indicated that the Langelier Index may have to be as great as + 1.0 to + 1.3 to maintain protection in lime-softened waters with low alkalinity and calcium concentrations approaching 50 mg/liter. In these waters softened by high pH, complexes of calcium and magnesium can form with carbonate and bicarbonate, thereby decreasing the effective concentration of both. As noted by Millette et al., (1980), a substantial fraction of surveyed U.S. water supplies have chemical water quality indices and pH values that could lead to corrosion in the distribution systems. Clearly, this should be evaluated on a continuous basis to ascertain whether such corrosion does indeed occur and result in concentrations of corrosion products that could adversely affect human health.

Corrosion Inhibitors

As discussed by Uhlig (1971). corrosion may be reduced effectively by the addition of small concentrations of chemicals, called inhibitors. Of these. the passivators (usually inorganic oxidizing agents) act by shifting the electrochemical corrosion potential several tenths of a volt in the noble (corrosion-resistant) direction. They are often successful in reducing the corrosion rates to very low values. The nonpassivating inhibitors, usually organic substances, have only a slight effect on the corrosion potential and are. therefore, usually less efficient than the passivators.

Many substances used to inhibit corrosion in water supply systems are nonoxidizing alkaline inorganic chemicals that indirectly cause passivation of iron. Apparently. they facilitate the sorption of dissolved oxygen, which acts as the oxidizer and, hence, is the actual passivating agent. The excess oxygen may form a passive film at the surface of the iron, although other films, such as those of iron silicate or phosphate, could also form. Sodium silicate and sodium polyphosphate are examples of such indirect inhibitors. Uhlig (1971) noted that the addition of 2 mg/liter sodium polyphosphate to water supplies reduces the corrosion rate to a modest extent if the water is fully aerated and is not stagnant. but there is probably no practical benefit in stagnant parts of a distribution system. Larson (1971) reported that there are some data indicating that at low velocities polyphosphates may even cause a slight increase in corrosion. He noted that the effectiveness of sodium silicate appears to depend on its absorption in the hydroxide corrosion products of iron and zinc, that treatment must be relatively continuous, and that its effectiveness is favored by high velocity and generally low pH. Zinc-glassy phosphates have also been used to control corrosion in water distribution systems, presumably by forming a protective film at the pipe surfaces (Schweitzer. 1970).

Although most of the corrosion inhibitors that may be added to public water supplies are relatively innocuous inorganic chemicals that are unlikely to cause adverse human health effects at the concentrations used. there should be a continuing evaluation of any possible effects, especially taking into account the results of animal and other toxicity tests. Bull and Craun (1977) considered possible health effects from polyphosphates used in water treatment. They concluded that, although they are generally safe at the low concentrations applied in water treatment, they should be used with care because of their possible adverse effects on human health.

Studies of animals showed that manganese and stoichiometric equivalent amounts of sodium hexametaphosphate fed at high concentrations in water interfere with trace metal metabolism, as judged by their effect on growth. They concluded that the biological effects of the entire range of polyphosphate compounds used in water treatment and their various metal complexes should be studied further. Nevertheless, although they did express concern about the use of these polyphosphates, they did indicate that at least for sodium hexametaphosphate it could continue to be used with care at dosage rates in water resulting in concentrations not to exceed that required to complex manganese by more than 10%, i.e., no more than 1 mg/liter.

Pips and Linings

A variety of materials are available for use in both large and small pipes (Table III-2). Improper selection of piping materials may have an adverse influence on the quality of the distributed water.

TABLE III-2. Lined and Unlined Piping Used in Water Distribution Systems.

TABLE III-2

Lined and Unlined Piping Used in Water Distribution Systems.

Among the metals that can be introduced into the water from the corrosion of these materials are zinc, cadmium, copper, iron, and lead. The latter is probably of the greatest concern because of its known toxicity and the results of numerous studies indicating that it can be corroded readily and accumulate in concentrations higher than generally accepted health-related maximum contaminant levels, e.g., 0.05 mg/liter, which was specified in the Interim Primary Drinking Water Regulations (U.S. Environmental Protection Agency, 1979a, 1980).

Although the use of plastic piping and linings of bituminous coal tar may be effective in reducing or eliminating the adverse impacts of metallic corrosion of pipes, such materials need to be continuously evaluated with regard to the possible leachings of their constituents into the water distribution system. (See section entitled ''Leaching'' later in this chapter.)

Lane and Neff (1969) have discussed the need to select proper materials to reduce corrosion in hot and cold water distribution systems. Indeed, they regard this step as the first line of defense and believe that chemical treatment should serve only a supplementary role. Although they focused primarily on distribution systems in institutions, their experience is germane to this report because it provides an insight into practical corrosion problems, it is consistent with general principles for reducing corrosion, and it would be applicable to homes as well as to institutions. They noted that compatible materials should be installed in the same system; for example, copper-lined heaters should be used with copper tubing, and the same type of piping used throughout the system. The widespread practice of softening water in homes and institutions to very low hardness values can cause serious corrosion of galvanized steel and other metal piping. Lane and Neff (1969) pointed out that blending to a final hardness of 60 to 90 mg/liter (calcium carbonate hardness) has often reduced corrosion in domestic hot water piping, but at temperatures above 60°C, corrosion is more severe. They suggested that more corrosion-resistant material such as stainless steel should be used when higher temperatures may be required.

The role of water quality and its impact on the corrosion of various materials were also evaluated by Lane and Neff (1969). They categorized midwestern U.S. water supplies into five types, primarily related to their hardness and to their chloride and sulfate content. For example, they suggested that copper piping was preferable to galvanized piping for hard waters with high chloride and sulfate content. In addition, the water should be softened, the pH adjusted to levels between 8.0 and 8.4, and sodium silicate treatment might be required. In contrast, they noted that hard waters that are not aerated and contain low to moderate chloride and sulfate are compatible with galvanized steel piping, although softening may be important in domestic hot water systems. This study by Lane and Neff emphasizes the principles that should be used to reduce corrosion in water systems in homes and institutions, particularly when it they may lead to excessive trace metal concentrations at the tap.

Monitoring and Evaluating Corrosion and Its Control

A variety of methods can and have been used to monitor and evaluate corrosion in water distribution systems. To control such corrosion effectively these methods must be used to evaluate the success of specific steps that are undertaken.

Wagner (1970) has summarized the methods that have been used. These methods can be divided into three groups. In the first, the mass of material formed is measured visually, by weight determination, or by measurement of thickness. In the second, the production of specific corrosion products, such as the evolution of hydrogen, is measured. The third involves the measurement of electrical and electrochemical parameters. including electrical resistance, corrosion current, pH, and electrical potential. Many of these tests are standardized by such organizations as the American Society for Testing and Materials (ASTM) and the National Association of Corrosion Engineers (NACE).

A recent example in which such tests were used to monitor corrosion-reducing measures was reported by Mullen and Ritter (1980). The corrosion problem they encountered arose in a public water supply system in New Jersey after a hard well-water supply (260 mg/liter calcium carbonate hardness) used for many years was replaced by a much softer surface water source (68 mg/liter hardness). When the new supply was in use, many consumer complaints of discolored water were reported. Various chemical control measures were then undertaken. Primarily, they examined the rate of corrosion by measuring the weight loss of a mild steel coupon. This was determined both by laboratory tests in which specific chemical controls and additives in the water were introduced and by the periodic measurement of coupons placed in the distribution system. They evaluated pH adjustment by caustic addition, temperature, and the addition of two corrosion inhibitors, sodium zinc-glassy phosphate and zinc orthophosphate, each added in amounts to produce zinc concentrations of 0.5 mg/liter in the water. Initially, the laboratory bench studies showed that the sodium zinc phosphate offered little corrosion protection compared to that of pH adjustment alone to 7.8 or 8.1, while the zinc phosphate caused an average of 55% reduction. Over a temperature range of 4°C to 25°C, adjustment of pH and the use of zinc phosphate were more effective than pH adjustment alone. However, below approximately 13°C, the inhibitor was more effective without pH adjustment. They then used the coupon tests over a 5-year period to measure the effectiveness of their corrosion-reduction steps in the actual distribution system, using pH control and the addition of zinc phosphate. In the warm months, their zinc dosages were 0.43 to 0.49 mg/liter, and the corrosion rates ranged from 0.8 to 1.4 µm per year.

The example discussed above indicates how steps can be taken systematically to mitigate corrosion by the adjustment of water quality and the use of corrosion inhibitors and that the effects of such steps can be measured. Although the principal interest here is reduction of the possible adverse health effects from the corrosion products, it is highly likely that this will be achieved simultaneously, even though the evaluative process focuses on the integrity of the pipe material, rather than on the potentially harmful corrosion products.

Field Studies of Metallic Corrosion in Distribution Systems

In this discussion of metallic corrosion in distribution systems. the metals of particular interest are cadmium, copper. iron, lead, and zinc. To the extent that they may be present in lower concentrations in pipe or distribution fixtures, other metals such as nickel and chromium might also contribute to corrosion products in water. Corrosion is not the only source of these products in treatment and distribution systems. They could also be introduced as contaminants in chemical additives, such as lime.

In the first volume of Drinking Water and Health (National Academy of Sciences, 1977), it was reported that the high concentrations of several heavy metals in household tap water samples from Dallas. Texas. could be attributed to corrosion at various points in the distribution system. Among these metals were iron from the steel water mains, copper and zinc from household plumbing, and lead and nickel due to "local influences." In the Denver, Colorado, municipal system, there were both increases and decreases in heavy metal content between the finished water at the treatment plants and samples taken at the domestic taps. although frequently there were no significant changes (Andelman, 1974). Typically. the maximum decrease at the tap for most metals was approximately 50%, but in some cases they were larger. Iron increased by a factor as large as 25, zinc and copper by approximately 5, and manganese by 4. Corrosion did not necessarily account for all of the increases, but was probably playing a role.

McCabe (1970) reported increases in the trace metal content of the Chicago water treatment system. In this study, 550 samples taken at various points in the distribution system were compared to 2-week composite samples at the treatment plants. McCabe used the criterion that a distributed water sample had "picked up" metal if its concentration was higher than that of any of the composite samples at a treatment plant. On this basis, it can be seen in Table III-3 that a substantial proportion of the distributed samples increased their trace metal content.

TABLE III-3. Percent of Distributed Water Samples with Increased Metal Content in Chicago System.

TABLE III-3

Percent of Distributed Water Samples with Increased Metal Content in Chicago System.

Dangel (1975) reported the results of a U.S. Environmental Protection Agency (EPA) study of corrosion products in the Tolt River portion of the public water supply system of Seattle, Washington. That portion was studied because its relatively low pH (approximately 6), alkalinity, and hardness and the many complaints of "red water" indicate that it is a corrosive system. The hypothesis of the study was that increases in metal concentrations in the distribution system could arise from unprotected metal service connections and residential plumbing, since the water mains were predominantly lined with cement and bituminous coatings. A comparison was made for several constituents in running (30 seconds) and standing water samples in residences. The standing samples had been collected from the water that had been in contact with household piping for at least one night. The results, for constituents that appeared to differ substantially, and a comparison of raw water concentrations of those constituents, are shown in Table III-4. Conductivity and the inorganic ions normally present at higher concentrations than those of the trace metals did not vary significantly. These include chloride, fluoride, calcium, and magnesium. However, the concentrations of the six metals shown in Table III-4 all were higher in the standing water samples. Copper, iron, zinc, and lead concentrations were clearly higher in the running water samples than in the raw water. Dangel concluded that most of the increases in metal concentration occurred in the service lines and plumbing inside the buildings.

TABLE III-4. Concentrations of Trace Metals in Raw and Distributed Water in the Seattle Tolt River System.

TABLE III-4

Concentrations of Trace Metals in Raw and Distributed Water in the Seattle Tolt River System.

Subsequent to the EPA study of the Seattle Tolt River system, the Seattle Water Department in 1976 and 1977 conducted a similar but larger study, which included both its Tolt River and Cedar River supplies (Hoyt et al., 1979). Samples were taken by both the Seattle Water Metals Survey Committee (SWMS) and the National Heart and Lung Institute (NHLI). In both the EPA Tolt River study and the SWMS study, almost all of the distribution sample sites selected were likely to have corrosion problems. That is, they had either visible water quality deterioration or copper plumbing less than 3 years old. In contrast, the NHLI selected a larger number of samples to represent a statistically valid cross-section of Seattle families in order to evaluate typical tissue uptake of waterborne metals. A comparison of the preliminary results of the three surveys for standing water samples in the distribution system is shown in Table III-5. This table indicates the percentage of those samples exceeding maximum contaminant levels (MCL's) of the EPA Interim Primary or Secondary Drinking Water Regulations (U.S. Environmental Protection Agency, 1979a,b, 1980).

TABLE III-5. Distribution of Standing Water Samples Exceeding Maximum Contaminant Levels in Three Seattle Studies.

TABLE III-5

Distribution of Standing Water Samples Exceeding Maximum Contaminant Levels in Three Seattle Studies.

It is apparent that there are substantial differences in the percentage of samples exceeding the MCL's. The larger percentage for the Tolt River system may be attributed to its water being more aggressive. Hoyt et al., (1979) indicated that after the very soft water of the Tolt River was brought into the system, a large increase in the complaints of "red water" occurred. Except for copper, the relatively smaller fraction of samples exceeding the MCL's in the larger NHLI survey can probably be attributed to the fact the study did not focus on homes with known or likely water quality problems. Although the aggressive waters in the Seattle system are not typical of most U.S. public supplies, they do demonstrate the kinds of corrosion problems that can cause increased exposures of the public to heavy metals.

The corrosion of lead pipe resulting in high lead concentrations in the water in distribution systems is a well-known phenomenon in which the characteristics of water quality play a substantial role. Large concentrations of lead in distributed water have been found in Bennington. Vermont. McFarren et al., (1977) reported that the pH of raw water was 5.5 and that the alkalinity was 1.0 mg/liter, both unusually low. The lead concentration in the raw water was less than 0.005 mg/liter. In contrast. high lead concentrations were found in water in the street mains and the service lines and in water in contact with interior plumbing. The latter was defined as the "first water" collected in the morning. Table III-6 shows the results of the three types of samples collected from 10 homes in Bennington. It is clear that the water from the street main was high in lead content in that every sample was approximately equal to or greater than the 0.05 mg/liter EPA MCL. Moreover, in almost every instance there was a higher lead concentration in the service line water compared to that in the street main. In contrast, most of the water in the interior plumbing had either comparable or somewhat lower lead concentrations than the water in the service lines. Although there was limited discussion of these results by McFarren et al., (1977), it appears that both the water mains and service lines are the likely sources of the lead and that the water quality may be playing an important role in the corrosion.

TABLE III-6. Lead Concentrations (mg/liter) in Various Water Samples Servicing Several Residences in Bennington, Vermont.

TABLE III-6

Lead Concentrations (mg/liter) in Various Water Samples Servicing Several Residences in Bennington, Vermont.

Karalekas et al., (1976) reported a 1974 study of lead and other trace metals in drinking water in the Boston, Massachusetts, metropolitan area. They compared the trace metal concentrations, with emphasis on lead, in distribution samples from Cambridge with those from Boston and Somerville. The Cambridge water had a higher pH, hardness, chloride, sulfate, and total dissolved solid content than did the other two, which had a common reservoir source. Typical hardness of the Cambridge supply before distribution was 56 mg/liter, while that of the others was 14 mg/liter. Thus, the finished waters in the Boston metropolitan area can be considered quite soft. There was also a widespread existence of lead in water service lines in the three cities. Of the 383 households studied, approximately 60% had lead pipe.

A total of 936 standing, running, composite (three equal portions taken during the day), and early morning samples were taken. Samples of finished water were also taken prior to distribution. Of the 10 trace metals analyzed, 5 were found in a substantial portion of the samples (Table III-7). It is apparent that many (15.4%) of the water samples exceeded the EPA interim primary standard for lead.

TABLE III-7. Trace Metal Concentrations in Water Distribution Samples in Boston Area Study.

TABLE III-7

Trace Metal Concentrations in Water Distribution Samples in Boston Area Study.

Table III-8 compares the results of lead analyses of the samples from the three Boston-area cities. Boston and Somerville had a higher percentage of households with lead exceeding the drinking water standard. This may be attributable to the differences in water quality described above. The authors suggested that when the lead content was high in the absence of an indication that lead pipe was used, either the record may have been faulty or the source could have been brass or bronze pipe, which contain substantial amounts of lead.

TABLE III-8. A Comparison of Lead in the Distributed Water of Three Cities in the Boston Area.

TABLE III-8

A Comparison of Lead in the Distributed Water of Three Cities in the Boston Area.

The early morning and the daily composite samples were much higher in lead content than were the running and standing samples taken during the day. The highest contrast existed between the early morning samples (a mean lead concentration of 0.104 mg/liter, and 47% of the samples exceeded the standard) and the running samples (0.031 mg/liter, and 5.5%). The authors concluded that additional water treatment, such as raising the pH to 8.5, was needed to prevent corrosion of lead in these systems.

In 1976 and 1977, an attempt was made to reduce this lead corrosion in the Boston area. As reported by Karalekas et al., (1978), the application of zinc orthophosphate, a corrosion inhibitor, was not particularly successful in that the average lead concentration in 18 to 23 homes was not reduced below 0.05 mg/liter. However, it was found subsequently that raising the pH from 7.5 to 7.7 was somewhat effective in that the average lead concentration was substantially reduced. Nevertheless, approximately onethird of the samples in this high-lead group of homes still exceeded the 0.05 mg/liter standard.

Summary and Conclusions

It is apparent from the case studies discussed above that metallic corrosion products in tap water distributed to the consumer can originate from water mains, service lines, and interior plumbing. Of these products, lead is of particular concern because of its known effects on health. The quality of the water also clearly affects the corrosion process. Thus, the choice of the system materials and the treatment regime undertaken at the treatment plant are both possible points of attack in the mitigation of corrosion that may arise in distribution systems.

The chemical quality of the finished water at the treatment plant can affect the rates of corrosion, which can be mitigated by chemical additions, including corrosion inhibitors. Bacteria in the distribution system are also known to affect corrosion.

Although corrosion can be reduced by pipe linings and chemical additives, the latter must be used judiciously to minimize human exposures. Also, there is evidence that potentially hazardous chemicals can leach from some, but not all, materials used to line the pipes. The adverse effects of the chemicals need further evaluations.

Leaching

In contrast to corrosion, which is an electrochemical phenomenon, leaching is a process of dissolution governed by solubility and kinetic properties of the materials involved. Materials that were initially part of the water distribution system pipes, storage reservoirs, and other components may slowly dissolve into the water. For some of them, such as the components of A/C pipe, leaching is strongly influenced by water quality. For other materials, such as those in linings and plastic pipe, water quality does not appear to govern leaching. In this section, A/C pipe, lining materials, and plastic pipe are discussed.

Asbestos-Cement Pipe

A/C pipe was originally introduced to the water supply industry as a material that was believed to be resistant to deterioration. However, recent findings have shown that A/C pipe is similar to other pipe materials commonly used in water distribution systems in that it can be subject to corrosive action of water in certain situations.

The potential release of asbestos fibers from A/C pipe in the water distribution system is but one potential source of asbestos contamination in potable water supplies. Other sources include natural erosion of asbestos-containing minerals and dumping of asbestos-containing materials from industrial projects, such as those contributing asbestos fibers to the drinking water of Duluth, Minnesota, and other nearby communities using Lake Superior water supplies (Cook et al., 1974). These other sources of asbestos in drinking water are beyond the scope of this study, which emphasizes presence of asbestos fibers in potable water distribution systems resulting from deterioration of the A/C pipe itself.

The concern about asbestos in drinking water from all sources, including pipes, is relatively recent. The first report of asbestos fibers in drinking water was published by Canadian investigators in 1971 (Cunningham and Pontefract, 1971). In the United States, the first major concern arose in 1974 when asbestiform fibers were found in the drinking water of Duluth, Minnesota, as a result of discharges from an iron ore processing plant (Cook et al., 1974). The same year, the American Water Works Association (AWWA) Research Foundation (1974) began its own investigation of the asbestos in drinking water, including a review of the possible effects of A/C pipe. At that time, there was very little information regarding the release of asbestos fibers from A/C pipe. The AWWA concluded:

At present there are not adequate data for a description of the quantitative or qualitative role of asbestos-cement pipe as a contributor to asbestos fibers found in potable water distribution systems. However, the available data do indicate that additional studies need to be made and also afford a first approximation of the magnitude of the possible health hazard.

The AWWA also concluded that further research was necessary to answer the question, ''Does the use of asbestos-cement pipe in potable water systems constitute a health hazard?'' Among the recommended research priorities were determinations of the increment of asbestos fibers added to the water as it traverses asbestos-cement, metal, and plastic pipe systems under varying conditions. A more complete understanding of the health effects of ingested asbestos fibers was also suggested as a research need.

Since 1974, a considerable amount of work on health effects of asbestos, water quality impacts of A/C pipe, and control measures has been sponsored by the EPA's Health Effects Research Laboratory and Municipal Environmental Research Laboratory.

At the time of the publication of the first volume of Drinking Water and Health (National Academy of Sciences, 1977), the concern about the release of fibers from A/C pipe was noted, and the relationship between soft water and dissolution of the calcium carbonate in the pipe was mentioned. The report noted that EPA studies under way had produced tentative results suggesting that some fibers are emitted from A/C pipe in corrosive waters.

Since that time, there has been a considerable amount of new evidence regarding the release of asbestos fibers from A/C pipe under various conditions. Both laboratory tests and field investigations have shown that, under certain circumstances, asbestos fibers can be released from A/C pipe into water supplies.

The data concerning the health effects of asbestos ingested in drinking water are indecisive, but it is known to be a carcinogen when inhaled (National Academy of Sciences, 1977). Thus, there has been some concern over the use of A/C pipe and possible release of asbestos fibers into drinking water. The concern over the addition of asbestos fibers and potential health effects in drinking water supplies led to a proposal by the EPA that corrosion control be required for systems using A/C pipe. The Aggressiveness Index (discussed above under "Corrosion Indexes") was proposed as a criterion to determine suitability of water quality for A/C pipe use (U.S. Environmental Protection Agency, 1979a).

The principal issue regarding the use of A/C pipe and its possible deterioration within a distribution system is the quality of water transported in the pipe. A considerable body of evidence collected to date suggests that there is a link between aggressive water and degradation of the pipe. Many investigators have linked a low Aggressiveness Index with release of fibers. However, the Aggressiveness Index does not always predict where the fibers will be released.

Current understanding of the factors influencing behavior of A/C pipe under various conditions is discussed below. Data from laboratory studies and field investigations regarding the release of fibers from A/C pipe are also summarized. Finally, control measures for limiting the exposure of the population to asbestos fibers derived from A/C pipe are presented.

Use of Asbestos-Cement Pipe

Originally introduced in Italy during the early 1900's, A/C pipe is now widely used in potable water distribution systems throughout the world. The formulation of A/C pipe from Portland cement and asbestos fibers was originally developed in an effort to provide a corrosion-resistant material of sufficient strength to be used for transmission of water. From 1906 to 1913, a company (Societa Anonima Eternit Pietra Artificiale) in Genoa, Italy, combined asbestos fibers with cement to produce a reinforced pipe that would withstand the high pressure necessary to pump salt water up to the City of Genoa for its street-flushing system (Olsen, 1974). Subsequently, A/C pipe began to be used more widely in Europe and was in limited use 20 years later in the United States. Today, A/C pipe is used throughout North America and Europe as well as in other parts of the world (Craun and Millette, 1977).

A/C pipe was first introduced in the United States around 1930 (American Water Works Association Research Foundation, 1974, 1975b; Hallenbeck et al., 1978). By 1974, there were an estimated 2.4 million km of A/C pipe in service worldwide and approximately 320,000 km in use in the United States (Olsen, 1974). According to a recent survey by the A/C pipe industry, approximately 38% of U.S. cities with a population of 1,000 or more specify, purchase, or have in service A/C pipe (Craun and Millette, 1977). Thus, approximately 65 million people in the United States may be receiving water that has passed through an A/C pipe distribution system. As of 1978, approximately one-third of all water distribution pipe currently being sold in the United States was made from an asbestos-cement combination (Hallenbeck et al., 1978). In addition to A/C pipe, gaskets and insulation used in treatment and pumps also contain asbestos (Levine, 1978). Advantages of using A/C pipe have been noted by pipe manufacturers and others. Some of them have been listed by Olsen (1974):

  • resistance to corrosion, both internal and external;
  • strength sufficient to withstand internal forces imposed by water hammer and shock earthloads from earthquakes;
  • benefits to water quality, since A/C pipe does not rust, cause discolored water, or contain jute and other types of joints that serve as focal points for bacterial growth;
  • light in weight and easy for contractors to install, resulting in lower installation costs; and
  • a permanently smooth interior wall, leading to low pumping costs.

Levine (1978) cites advantages of A/C products over nonasbestos counterparts as follows: better tensile strength, strength to weight ratio, strength under heat stress, resistance to acid, and smoothness of finished surface (critical to ensure laminar flow in pipe used for transport of liquids).

As shown later, the purported advantage of corrosion resistance of A/C pipe has not always been found to be true, especially in aggressive waters.

Composition of A/C Pipe

A/C pipe is composed of a mixture of asbestos fibers, Portland cement, and inorganic hydrated silicates. In the manufacture of A/C pipe, asbestos fibers are mixed, either wet or dry, with Portland cement and silica in proportions ranging from 10% to 70% of the total material. Typically, the asbestos fibers comprise less than 20% of the A/C pipe (A-C Pipe Producers Association, 1980; American Water Works Association, 1978a,b).

For dry mixtures, the mixture is generally distributed in a flat layer onto an open surface, where the water is then applied by an overhead spray. The thin layer can then be wound onto mandrels in a spiral mat to produce pipe until the required thickness is built up. For wet mix products, a similar winding process can be used for the slurry or the mixture can be cast (Levine, 1978). After pipes are made, the water content is reduced by autoclaving and air drying.

The AWWA specifications for A/C pipe also include physical and chemical requirements for the pipe itself. For the pipe composition, it requires:

Asbestos-cement pipe shall be composed of an intimate mixture of either:

(1) Portland cement or Portland blast furnace slag cement and asbestos fiber with or without silica; (2) or Portland pozzolana cement in asbestos fibers. Both (1) and (2) can be with or without the addition of curing agents. The pipe shall be formed under pressure and cured. The finished pipe shall be free from organic materials (American Water Works Association, 1980).

The same specifications limit the amount of uncombined calcium hydroxide, presumably to curtail pipe dissolution: for Type I, there is no limit, and for Type II, 1.0% or less uncombined calcium hydroxide is permitted. Type I, which is not autoclaved, is no longer manufactured in the United States.

The asbestos portion of A/C pipe is composed of naturally occurring hydrated mineral silicates that possess a crystalline structure. There are four main types of asbestos, as described by Michaels and Chissick (1979):

1.

chrysotile (3MgO · 2SiO2 · 2H2O) or white asbestos, which occurs as fine silky flexible white fibers and is mined mainly in Canada. Russia. and Rhodesia;

2.

amosite [(FeMg)SiO3], a straight brittle fiber, light grey to pale brown in colour and found in South Africa;

3.

crocidolite [NaFe(SiO3)2 · FeSiO3 · H2O] or blue asbestos, which is found as a straight blue fibre in South Africa. Western Australia, and Bolivia; and

4.

anthophyllite [(MgFe)7 · Si8O22 · (OH)2], a brittle white fibre mined in Finland and Africa.

Other types of asbestos include tremolite [Ca2Mg5Si8O22 · (OH)2] and actinolite [CaO · 3(MgFe)O · 4SiO2].

The principal type of asbestos found in A/C pipe is chrysotile. According to the A-C Pipe Producers Association (1980), the chrysotile is added for reinforcement purposes. Another type of asbestos, crocidolite (or blue asbestos), is also used for reinforcement of the pipe manufacturing process.

Effect of Water Quality on Pipe Performance

In contrast to the original expectations that A/C pipe would not be attacked by corrosive water, it is now recognized that under certain circumstances A/C pipe can be attacked by aggressive water. If the pipe is exposed to aggressive water, the cement matrix constituents will dissolve, thereby exposing asbestos fibers and releasing some of them into the water. Concepts developed to predict pipe performance are described above under "Chemical Water Quality Indexes."

Laboratory Tests

Only a few laboratory studies have been conducted to investigate the influence of water quality on deterioration of A/C pipe. One A/C pipe manufacturer, Johns-Manville, developed a semiclosed recirculating system with 9.1 meters of A/C pipe (Transite) (American Water Works Association Research Foundation, 1974). Water ahead of the pipe was filtered in an attempt to remove all asbestos fibers prior to contact with the pipe. A series of runs at various levels of pH (4.9-7.4) and total hardness (12 to 105 mg/liter) were made from 1969 to 1971. The data do not supply sufficient information to calculate either the Langelier Index or the Aggressiveness Index. Fiber analyses were performed using particle and fiber counts from magnified electron microscope photographs combined with a radioactive tracer technique. They showed that asbestos fibers were present in the water leaving the pipe in average concentrations ranging from 0.37 to 4.44 X 10-5 μm/liter.

In a pilot study conducted in Seattle, Kennedy Engineers (1978) evaluated the aggressiveness of two water supplies, the Tolt River and Cedar River, and the effectiveness of various inhibitors in reducing the aggressiveness of water to A/C pipe. Corrosivity measures were weight loss, electron microscopy spectrum analysis, asbestos fiber pickup, and water quality analysis. The results confirmed the aggressiveness of both water supplies, which were low in pH, alkalinity, calcium, and total dissolved solids (TDS), to A/C pipe. For the Cedar River, a treatment with lime and soda ash appeared to be the inhibitor strategy most likely to provide protection to A/C pipe. For the Tolt River, lime plus zinc orthophosphate gave the best protection.

Since 1973, the EPA has been conducting a series of laboratory tests using A/C coupons in an experimental pipe loop system (Buelow et al., 1980). In one experiment, the EPA noted two important findings: (1) iron dissolved in water precipitates to provide a protective coating to A/C pipe, even under highly aggressive water quality conditions, and (2) waters indexed as moderately aggressive attacked Type II A/C pipe, even though this result was not predicted. Another important finding of this early work was that pipe drilling and tapping operations can greatly influence asbestos fiber counts in water.

Following the pilot-scale A/C pipe loop experiments, the EPA proceeded to conduct more rigorous tests with bench scale equipment that could be more carefully controlled. Water was recirculated through 6-in. coupons cut from 4-in. Type II A/C pipe. Findings of these experiments are summarized in Table III-9 (Buelow et al., 1980).

TABLE III-9. Summary of Individual EPA Bench Scale Tests of A/C Pipe.

TABLE III-9

Summary of Individual EPA Bench Scale Tests of A/C Pipe.

Schock and Buelow (1980) reported that results of several of these EPA experiments and other tests were compared with Saturation Indexes and precipitation diagrams calculated for model systems with equivalent water quality and treatment. A good correlation of experimental and predicted results was found in all cases.

The results of these investigations provided several pieces of information elucidating mechanisms of deterioration of A/C pipe. The principal findings are as follows:

  • Zinc can provide a protective coating, which prevents the surface from deteriorating even when Aggressiveness Indexes predict that the water will be moderately aggressive to A/C pipe. This protective action of zinc appears to be dependent on pH, higher pH's providing more effective protection.
  • Zinc orthophosphate and zinc chloride appear to provide equivalent protection. This suggests that the protective mechanism is related to zinc rather than to the anion or to the compound as a whole.
  • The Aggressiveness Index alone cannot be used as a single means of predicting performance of pipe under a given water quality condition. This observation is based on the fact that water with a high Aggressiveness Index, but not saturated with calcium carbonate, has attacked A/C pipe.
  • Control of calcium carbonate saturation may, under certain situations, prevent deterioration of A/C pipe.

Some of the more promising findings are being tested by the EPA in field applications for controlling pipe deterioration.

Field Tests

In recent years, concern that A/C pipe can deteriorate under certain water quality conditions has led to numerous studies, including field observation of water quality, pipe conditions, and/or other factors. These studies are summarized below chronologically.

Effects of A/C pipe on water quality were first observed in 1945 in Vermont. Tracy (1950) reported increases in pH, hardness, and alkalinity in water after exposure to A/C pipe. The tendency for water quality changes diminished after several years of exposure to the pipe. Another problem with A/C pipe was reported in 1971 (American Water Works Association, 1971). The investigators reported that the pH of water exposed to new A/C pipe would rise as high as 11.5, if the pipe was not used continuously.

Kay (1974) published a summary of asbestos concentrations found in distribution system samples from 22 cities in Ontario, Canada. Fiber counts by electron microscopy ranged from 136,000 to 3.87 million fibers per liter. The calculated mass concentrations ranged from a low of 0.93 to a high of 35.4 × 10-4 µg/liter. Although there was evidence that the distribution samples contained asbestos, it was not clear whether the fibers originated in the raw water supply or whether they were contributed by action of the water on A/C pipe.

Johns-Manville researchers studied the waters from nine cities using A/C pipe from 8 to 17 years (Olsen, 1974). The asbestos content of the water at the source and at a point in the distribution system following exposure to A/C pipe was tested with an electron microscope technique. Source concentrations varied between 0.26 and 1.32 µg/liter. Fiber content after exposure to A/C pipe ranged from 0.26 to 1.58 µg/liter. The differences do not appear to be significant.

Two municipalities using A/C pipe in their distribution systems were sampled from 1969 through 1970 (American Water Works Association Research Foundation, 1974). In Malvern, Pennsylvania, the content of fibers in well water prior to exposure to A/C pipe was 0.04 µg/liter. After exposure to water having an Aggressiveness Index of 11.3, the water in the distribution system had an average fiber level of 0.12 µg/liter. The water in Glendale, Arizona, had an Aggressiveness Index of 11.8 and an average initial fiber level of 0.006 µg/liter. After exposure to A/C pipe, the average fiber level was 0.01 µg/liter. Thus, both of these early field studies of water in the distribution system indicated an increase in asbestos fiber content.

In 1975, the EPA conducted a study in the Seattle area to determine if water quality changes were occurring as a result of exposure of aggressive water to A/C pipe (Dangel, 1975). Prior to entrance into the distribution system, the pH of the raw water was low (5.4), as was its alkalinity (approximately 4 mg of calcium carbonate per liter) and calcium content (2 mg/liter), and its Aggressiveness Index was 6.7. After exposure to approximately 3.2 km of A/C pipe mains, the water exhibited changes in pH, alkalinity, calcium content, and conductivity, all of which increased with longer exposure to the pipe. Dangel concluded that as the cement binders dissolved, asbestos fibers may have been leaching from the pipe walls. This study was conducted by the EPA to determine if asbestos pipe was degrading. Later studies found that the concentration of asbestos in the raw water was approximately the same as in some parts of the distribution system that were not exposed to A/C pipe. Therefore, subsequent studies were conducted on removal of asbestos from the raw water itself (Kirmeyer et al., 1979).

Another aspect of the A/C pipe deterioration—potential economic losses—was addressed by Hudson and Gilcreas (1976). In a water utility described by the authors, alkalinity of water leaving the treatment works increased by 6 mg of calcium carbonate per liter by the end of the distribution system, which consisted of reinforced concrete, asbestos-cement, and cement-lined cast-iron pipe. The plant was not able to provide sufficient lime-feeding capacity to produce an effluent stable in calcium carbonate content. Based on alkalinity increases in the distribution system, the authors concluded that the water system was losing approximately 455 metric tons of transmission and distribution piping annually as calcium carbonate. No comment was made about the potential introduction of asbestos fibers from the A/C pipe.

McFarren et al., (1977) reported an EPA survey of six public water supply systems that used A/C pipe for distribution. These systems had various combinations of pH, alkalinity, and calcium hardness with Aggressiveness Indexes ranging from 5.34 to 12.85. The results of this yearlong survey are summarized in Table III-10. Asbestos fiber counts, measured by transmission electron microscope, were consistently quantified only in the two systems with very aggressive waters. The authors stated that the pH, alkalinity, and calcium hardness increased as the water passed through the pipe, presumably as a result of the water's reaction with the cement in the pipe, which caused the water to become less aggressive to pipe downstream. In a sample of pipe taken from the King County water district system in Seattle, pipe exposed to the water was shown by scanning electron micrographs to have been changed substantially as compared to new pipe.

TABLE III-10. EPA Survey of Systems Using A/C Pipe.

TABLE III-10

EPA Survey of Systems Using A/C Pipe.

Craun and Millette (1977) studied the use of A/C pipe in Connecticut public water supplies and the incidence of gastrointestinal cancer. A total of 149 public water supplies in 82 towns were evaluated. Preliminary electron microscopy measurements indicated that 19 water samples exposed to A/C pipe contained chrysotile fiber ranging from below detectable limits (10,000 fibers/liter) to 700,000 fibers/liter. Some amphibole fibers were detected, but concentrations were less than 50,000 fibers/liter. The majority of towns had an Aggressiveness Index under 9.8. The observation of relatively low fiber counts leads to the conclusion that the Aggressiveness Index alone is not a sufficient predictor of the release of fibers from A/C pipe.

Information on exposure of A/C pipe to other types of water quality has been presented by Webster (1974). Photographs of A/C pipe that had been carrying brine (water with a high Saturation Index) for several years showed a well-marked deposition of calcium salts in the interior of the pipe, which should protect the pipe from corrosion. A/C pipe subject to acid conditions while carrying sewage developed rough interior surfaces, which presumably could release fibers through erosion. The sewage probably contained hydrogen sulfide, which can attack A/C pipe (McCabe and Millette, 1979). High levels of hydrogen sulfide in a Florida well water source were apparently responsible for attacking A/C pipe. The Aggressiveness Index calculated for this water would not lead to such a conclusion, but it does not take into account the corrosive effects of water quality characteristics other than pH, calcium, and alkalinity. Removal of hydrogen sulfide from the Florida well water supply is being studied.

Hallenbeck et al., (1978) reported the results of a study of water samples from 15 public water supply systems in Illinois before and after exposure to A/C pipe of various ages, length, and diameter. Five were groundwater systems, and 10 were surface water systems from Lake Michigan. Aggressiveness Indexes ranged from 11.2 to 12.8; Langelier Saturation Indexes were calculated at -0.8 to 0.5. Pipe ages ranged from 0.5 to 50 years. Measurements by transmission electron microscope indicated that there was no statistically significant difference in the fiber content of water samples collected before and after exposure to the A/C pipe. Thus, there was no statistically significant release of chrysotile fibers from the A/C pipe into these moderately aggressive to nonaggressive water supplies.

In an investigation of domestic water supplies in the San Francisco Bay Area, Cooper et al., (1978) reported on asbestos concentrations within distribution systems that contained an A/C pipe. The authors concluded that there was no substantial increase in fiber counts, measured by transmission electron microscopy, after the water passed through A/C pipe. Some increases were observed, but they were not significant compared to the error of measurement.

Millette et al., (1979) compiled a summary of more than 1,500 analyses of asbestos in the water supplies of 43 states, Puerto Rico, and the District of Columbia to assess the overall exposure of the U.S. population to asbestos in drinking water. Some of the highest concentrations of asbestos fibers were attributed to A/C pipe (Table III-11). This report also contains tabulations of the miles of A/C pipe, concentrations of fibers, and other information for the 1,500 water samples, which had been collected by the EPA.

TABLE III-11. Selected Locations in Which Asbestos Concentrations Have Been Attributed to A/C Pipe.

TABLE III-11

Selected Locations in Which Asbestos Concentrations Have Been Attributed to A/C Pipe.

Millette et al., concluded as follows:

The majority of persons receiving water from asbestos cement pipe distribution systems are not exposed to significant numbers of fibers from the pipe. Many residents using asbestos cement pipe may be exposed to intermittent amounts of asbestos fibers in their water if pipe tapping work is done improperly. In areas of very aggressive water (estimated to be 16 percent of the U.S. water utilities) consumers using asbestos cement mains may be exposed to high concentrations of fibers, over 10 million fibers/liter.

Tarter (1979) has conducted a statistical study of the size of asbestos fibers in the San Francisco Bay Area water systems. A slight but not necessarily statistically significant shift to longer fibers was observed in water after it flowed through A/C pipe in the East Bay Municipal Utility District distribution system. Examination of samples obtained before and after passing through A/C pipe in the East Bay system and in the San Francisco Water Department Hetch-Hetchy system led to the conclusion that water exposed to A/C pipe contained a significantly larger portion of long fibers than did raw water. Craun and Millette (1977) also reported that samples obtained after passing through A/C pipe in a Connecticut system had a higher percentage of fibers exceeding 1 µm in length than a sample with a natural source of fibers obtained from the San Francisco raw water reservoir. The respective median values for the asbestos fiber length were 2.0 µm from the A/C pipe distribution system and 0.7 µm from the natural fiber source.

McCabe and Millette (1979) have summarized a number of EPA-sponsored studies pertaining to deterioration of A/C pipe under varying water quality conditions. In an update on the Connecticut study, they indicated that source waters in 45 Connecticut A/C pipe systems were thought to be very aggressive because the Aggressiveness Indexes were less than 10. However, an epidemiological evaluation of the exposures revealed that there were no high concentrations of asbestos in the distributed water sampled after passing through A/C pipe in any of the systems. Furthermore, none of the pipe that had been dug up over the years had been reported to be significantly deteriorated. With one exception, all samples collected from the Connecticut A/C pipe system contained asbestos concentrations below 1 million fibers per liter, measured by electron microscopy. Distribution system maintenance work and tapping were believed to be the sources of asbestos in one Connecticut sample, which contained 10 million fibers per liter on one occasion and less than I million fibers per liter on another. The higher concentration was believed to be the result of pipe tapping. If devices are not used to flush the cutting debris from the pipe or if samples are collected from dead-end areas or from fire hydrants that have not been completely flushed, samples may contain high concentrations of asbestos fiber that are not an accurate representation of the system as a whole. Thus, what may be attributed to pipe deterioration in some cases may actually be a result of accumulation of sediments from previous pipe tapping.

In a summary of EPA field work, Buelow et al., (1980) discussed evaluation studies at nine water utilities using A/C pipe with various water qualities (Table III-12). Asbestos fiber counts, made by transmission electron microscopy, were consistent in four of the five systems where the Aggressiveness Index was less than 10.0. Inspection of one system with an Aggressiveness Index below 10 (third line in the table) revealed that the pipe had a protective coating resembling iron rust. Fibers were also found in one system with an Aggressive Index greater than 10.0. In the systems with an Aggressiveness Index exceeding 11.56, there were either no asbestos fibers or their occurrence was very inconsistent, regardless of the combinations of pH. alkalinity, and calcium hardness. Unfortunately, water quality data from field studies were insufficient to calculate Saturation Indexes.

TABLE III-12. Summary of EPA Field Studies.

TABLE III-12

Summary of EPA Field Studies.

Results from this field study also indicated that the longer the water was exposed to the A/C pipe. the greater were the increases in pH and calcium. This should cause the water to become less aggressive. Therefore, the investigators concluded that major pipe deterioration will usually occur in the sections of the pipe located just after the water enters the A/C pipe distribution system. However, a section of pipe farther from the source will not always be attacked less than the pipe closer to the source due to different flow patterns through different distribution lines.

Buelow et al., (1980) reported the major conclusions from this study as follows:

1.

Calculation of the Aggressiveness Index alone is not always sufficient to predict actual behavior of A/C pipe.

2.

Collecting a single sample for an asbestos fiber count is often insufficient to judge the actual behavior of A/C pipe in a given situation.

3.

Wet drilling and tapping of A/C pipe. if not performed with a flushing device on a tapping machine, can cause a major release of fibers.

4.

Metals such as zinc, iron, or manganese in the water can change A/C pipe behavior.

5.

Water is not expected to be attacking A/C pipe when the initial Aggressiveness Index is above about 11. the pH or the concentration of calcium do not change significantly as the water flows through the pipe. and no asbestos fibers are found consistently in representative water samples....

6.

Water is likely to be attacking A/C pipe when the initial Aggressiveness Index is below 11. and the pH and the concentration of calcium as the water flows through the pipe increase significantly, and the water does not contain iron. manganese, or similar metals....

In summary. this study has demonstrated that asbestos cement pipe behaves much like other piping materials, except PVC, that are commonly used for distribution of drinking water. If aggressive water conditions exist, the pipe will corrode and deteriorate: if aggressive water conditions do not exist, the pipe will not corrode and deteriorate.

Control of A/C Pipe Deterioration

Various methods have been suggested for curtailing or preventing the corrosion of A/C pipe within water distribution systems. These methods fall into the following basic categories:

  • adjustment of water quality to control Langelier Index:
  • adjustment of water quality to control Aggressiveness Index;
  • addition of materials expected to form protective films or coatings such as zinc, iron, and manganese;
  • elimination of hydrogen sulfide;
  • rehabilitation of distribution system;
  • institution of proper maintenance procedures;
  • limitation of the use of A/C pipe; and
  • judicious selection of new locations for A/C pipe use.

Adjustment of water quality by controlling Langelier Index is also used to prevent corrosion of metal pipe. Therefore, it is a method with which many utilities are already familiar. The EPA listed adjustment of water quality to a positive Langelier Index as a method for corrosion control and mentioned it specifically for A/C pipe (U.S. Environmental Protection Agency, 1979b). McFarren et al. (1977) suggested that A/C pipe corrosion might be controlled by adjusting the pH upward and adding calcium (in effect adjusting either the Langelier Index or the Aggressiveness Index upward). He added, however, that some people might object to an increase in hardness of their water. Buelow et al. (1980) agreed that calcium carbonate saturation can be used to prevent attack on A/C pipe. However, it is acknowledged that the greatest weakness in the use of the Saturation Index occurs with water of relatively low alkalinity and calcium. Nevertheless, corrective lime treatment is an option.

Adjusting the Aggressiveness Index is another alternative for controlling the corrosion of A/C pipe. Methods for accomplishing this would include adjustment of pH upward or the addition of calcium or alkalinity to the water. Adding lime, caustic, or soda ash might be considered, but pilot tests should be conducted for an exact selection of chemicals.

In certain cases, the addition of small amounts of metal salts such as those of iron, zinc, and manganese may contribute to the formation of a protective film. The EPA is in the process of evaluating natural inhibitory factors in a Massachusetts drinking water supply to determine why water that would otherwise be classified as very aggressive is not extremely corrosive to A/C pipe and other materials (U.S. Environmental Protection Agency, 1979c). The potential protective effect of these metal compounds is being investigated. Since coatings resembling iron rust have been observed in laboratory experiments on existing systems, the addition of iron may be able to protect A/C pipe. Zinc and manganese have also been shown to have protective effects. However, applicable drinking water standards and potential effects on industrial users would have to be considered before control measures using either compound could be implemented. In addition, the anion of choice needs to be carefully considered. For example. a study in Seattle showed that although zinc phosphate was effective in reducing corrosion, the phosphate demonstrated a potential to stimulate algal growth in open distribution reservoirs (Courchene and Kirmeyer. 1978). This finding suggests the need for study of other forms of zinc, such as zinc chloride. The costs for adding zinc orthophosphate and zinc chloride have been estimated by the EPA (1979b). Coating of pipe with zinc orthophosphate and pH adjustment are being evaluated as control measures to prevent deterioration of A/C pipe in Greenwood. South Carolina (McFarren et al., 1977).

In a few cases, deterioration of A/C pipe may be curtailed by removing hydrogen sulfide, which may attack the pipe under low pH conditions. A Florida town is currently studying the hydrogen sulfide problem and is planning to implement treatment to eliminate it from water prior to exposure to A/C pipe (McCabe and Millette, 1979). In an EPA-sponsored research project, a utility is adding zinc to the water to determine if it can reduce the amount of asbestos fibers released from A/C pipe that is slightly attacked.

The institution of proper maintenance practices is another approach to controlling occasional additions of asbestos from the distribution system itself. This may be applicable in situations where the water itself is not attacking the pipe, but where occasional high asbestos fiber concentrations have been attributed to residues remaining from improper maintenance procedures. Some tapping devices on the market today force debris from cutting operations to be flushed from the pipe, preventing contamination of drinking water with those fibers. The A-C Pipe Producers Association (1980) recommends that equipment with a positive purge should be used to tap A/C pipe under pressure. This equipment is reported to eject 99% of the asbestos cement chips. Another alternative is the use of heavy walled tapped couplings supplied directly from the factory. According to Buelow et al., (1980) many regular pressure tapping machines can be modified by adding a flushing valve from a commercially available kit. New tapping machines can be purchased with a flushing valve.

The American Water Works Association (1978a,b) has published detailed procedures for the use of A/C pipe in water utilities. For ''dry'' tapping of pipe not under pressure, it recommends that dust and cuttings should be removed from the pipe's interior by flushing with water, wet mopping, or vacuuming prior to placing the pipe in service. This should minimize the fouling of valves, regulators, meters, and other equipment with chips and minimize the unnecessary addition of asbestos to drinking water. For "wet" tapping of pipe in service, it recommends that provision should be made for downstream flushing or use of tapping equipment with positive purge or "blow-off" features.

Another approach to preventing the occurrence of asbestos fibers in drinking water supplies has been to restrict the use of A/C pipe by legal means. Some cities have either banned the future use of A/C pipe or are in the process of considering such bans.

The more careful selection of locations where A/C pipe should be used is a preventive measure with merit. During the past 10 years, pipe manufacturers and trade organizations have recommended caution in the use of A/C pipe with aggressive waters, but this advice has not always been heeded. By using predictors such as the Saturation Index and the laboratory coupon tests developed by the EPA. a utility should be able to determine judiciously if A/C pipe is an appropriate material for a given water quality.

Summary and Conclusions

Under certain water quality conditions, A/C pipe can deteriorate. Evidence from laboratory and field tests has shown that asbestos fibers can be released from deteriorated pipe. Other water quality changes observed after exposure to A/C pipe include increased pH, calcium, and alkalinity. Consistent with these observations, the cement portion of the pipe, consisting principally of calcium silicates plus a small amount of free lime, can dissolve if water is not saturated with a number of calcium compounds. Release of cement constituents would be followed by increased pH, calcium, and alkalinity. After cement is dissolved or otherwise weakened, asbestos fibers in the matrix are released. Since asbestos in A/C pipe is principally the chrysotile form, the majority of fibers released to the pipe are chrysotile.

A number of indexes have been developed to predict pipe performance under given water quality. The Langelier Index estimates whether water will be oversaturated, undersaturated, or at equilibrium with respect to calcium carbonate. If water is oversaturated, this index is positive and a protective calcium carbonate film can deposit on A/C or other types of pipe. The Aggressiveness Index, a simplified version of the Langelier Index, has been used extensively in the water supply field.

Since both the Langelier Index and the Aggressiveness Index are based on calcium carbonate solubility, there are shortcomings in their ability to predict the performance of A/C pipe. When the indexes indicate oversaturation, experimental and field observations have frequently (but not always) shown that pipe is not deteriorated. When exposed to water predicted to be aggressive by the Aggressiveness Index, some pipes deteriorate and others do not. In the latter instance, protective films of zinc, iron, or other material compounds have been found. Thus, it is clear that water quality factors other. than pH. calcium, and alkalinity should be considered before water's aggressiveness to A/C pipe can be determined.

Schock and Buelow (1980) have proposed a chemical equilibrium model that accounts for the potential protective action of silica, iron, manganese, and zinc as well as that of pH, carbonate, calcium, temperature, and disinfectant residual in determining a Saturation Index for several compounds. This appears to be a promising method for predicting performance of A/C pipe, but it will require field tests for confirmation.

Plastic Pipe

In recent years, plastic pipe materials of various compositions have been used in transmission mains of water distribution systems and in individual home services. Plastics were brought into use to provide low-cost pipe materials that were not subject to the action of aggressive waters. Within the past few years, however, there has been some concern over possible leaching of organic compounds and other constituents of plastic pipe into the water. The limited number of tests that have been conducted on this subject have dealt primarily with polyvinyl chloride (PVC) pipe. This section contains discussions of the types of plastic pipe commonly used in the United States, their composition, their use in water distribution systems and home plumbing, the effects of water quality as measured by laboratory and field studies, control measures, and conclusions.

Use of Plastic Pipe

Thermoplastic pipe was first introduced as a commercial product in the United States about 1941, but initially its use in potable water supplies was limited. By 1968, Farish (1969) reported that approximately 234 million kg of pipe and fittings were produced from three materials used in potable water applications: polyethylene (PE); polyvinyl chloride (PVC), including chlorinated polyvinyl chloride (CPVC); and acrylonitrilebutadiene-styrene (ABS).

The most common plastic piping material. PVC, has been in use in the United States since approximately 1960 (Dressman and McFarren, 1978; McFarren et al., 1977). In 1975, Rawls reported that approximately 180 million kg of PVC resin was used in the production of water pipe annually. The use of plastics in distribution systems and home services within the United States has been summarized by compiling the results of a questionnaire distributed by the American Water Works Association (1979). Based on responses received from 514 water suppliers, the approximate percentages of all types of plastic pipe used for distribution system mains were as follows: 80% PVC, 13% PE, 5% polybutylene (PB), and 2% fiberglass-reinforced plastic pipe (FRP). For home services, the percentages of various plastic materials were as follows: 43% PE. 20% PB. and 36% PVC. The relative percentage of PVC used was shown to be increasing.

In a survey of water distribution materials present in European households. Haring (1978) reported that PVC and PE (as Polythene) comprise between 1% and 12% of the total materials used in Belgium. France, the Federal Republic of Germany, Ireland, Italy, the Netherlands, and the United Kingdom.

There are several advantages and disadvantages in the use of plastic pipe for distribution systems and services. The American Water Works Association (1979) has reported the following advantages: ease of installation, low cost, external and internal noncorrodibility, better flow characteristics, ease of repair, and greater strength relative to A/C pipe. Disadvantages noted by utilities using plastic pipe include problems in cold weather, difficulty in locating nonmetallic pipe, need for careful bedding and backfilling, capping problems, problems with solvent weld joints, deterioration from storage in direct sunlight, costly and often unavailable transition fittings, nonstandardization of fittings and adaptors, weakness relative to cast or ductile iron, and problems with splitting. cracking, and scoring of the pipe. A few users were concerned about possible migration of vinyl chloride monomer from PVC pipe into potable water.

Composition of Plastic Pipe

Plastic is defined by the National Sanitation Foundation as "a material that contains as an essential ingredient an organic substance of high molecular weight, is solid in its finished state, and at some stage in its manufacture or its processing into finished articles can be shaped by flow." Thermoplastics are plastics that are "capable of being repeatedly softened by an increase of temperature and hardened by a decrease of temperature" (Farish. 1969).

The polymerized product is the principal constituent of plastic pipe. In some cases, there is a small residual of unpolymerized monomer in the finished pipe. Some monomers, such as unpolymerized vinyl chloride in PVC pipe, have toxic properties and are a potential source of concern if they leach out into potable water. Vinyl chloride is a known carcinogen in humans and animals (National Academy of Sciences, 1977).

As a result of potential health effects of unpolymerized monomers, steps have been taken in industry to curtail the use of certain materials that could be leached from the plastics. Efforts to reduce the amount of vinyl chloride monomer in PVC have been made since 1973, when the Food and Drug Administration found that vinyl chloride leached into liquor contained in PVC bottles, which were subsequently banned for liquor sales by the Department of the Treasury. New methods for manufacturing PVC are aimed principally at reducing the residual level of vinyl chloride monomer in the plastic. Rawls (1975) reported that technology for reducing the amount of vinyl chloride in vinyl chloride polymers is improving rapidly and that a goal of less than 1 ppm residual vinyl chloride in PVC is reasonable.

In addition to the organic polymer, which is the main component of plastic pipe, other materials are included in the pipe for various purposes. Therefore, any impacts on water quality related to the use of plastic pipe may derive from leaching of these components. For example, unplasticized PVC (uPVC) pipe contains pigments, lubricants, and stabilizers in addition to the polymer PVC (Packham, 1971a). Pigments, added to make the pipe opaque, commonly include carbon black or titanium dioxide. Lubricants, which reduce the adherence of pipe material to extrusion tools used in pipe manufacture, include a range of materials such as stearic acid, calcium stearate, glycerol monostearate, polyethylene wax, and montan wax. Stabilizers, incorporated to reduce the rate of decomposition of PVC at elevated temperatures used for extrusion and injection molding, include compounds of lead, cadmium, barium, tin, calcium, and zinc. In general, pigments and lubricants are physically and chemically inert, but many of the stabilizers are toxic.

Products used to join plastic pipe sections may also, under some circumstances, leach into water in contact with the pipe. Some jointing methods involve the use of solvent cement, push-on joints, mechanical joints, and heat-fused joints. Primers and cements used in solvent-welded joints contain such constituents as methyl ethyl ketone, cyclohexanone, tetrahydrofuran, and N, N-dimethylformamide.

Plastic Pipe and Water Quality

Although plastic pipe does not appear to be subject to major degradation by contact with the water, there has been some concern that small amounts of materials in the plastic pipe could leach into the water. Similar concerns have been expressed about materials that come into contact with food. There has been a limited amount of testing on plastic pipe in contact with potable water. Effects on water quality observed in laboratory studies and field tests are summarized below. Most of the information pertains to PVC.

The quality of water appears to be of minor importance in the release of substances from plastic pipe materials (McFarren et al., 1977). This is in contrast to other types of pipe from which the release of constituents is definitely affected by water quality.

Laboratory Tests

In the United States, some of the first studies of plastic pipes were conducted by the National Sanitation Foundation in conjunction with the University of Michigan School of Public Health. The potential toxic effects of plastic pipe were studied during a 3-year research program beginning in 1952 (Farish, 1969). In these studies plastic pipes (composition not specified) were exposed to aggressive water at pH 5.0 for 72 hours at 37.78°C. Taste and odor were evaluated, and chlorine residuals were measured. Wistar white rats exposed to plastics in their drinking water were studied for 18 months. Autopsies showed no evidence of damaging effects attributable to contaminants in water resulting from continuous exposure to various plastics.

Materials other than organic chemicals can be extracted from plastic pipe. The National Sanitation Foundation (Farish, 1969) found that lead, cadmium, strontium, lithium, antimony, and other toxic elements used as stabilizers in plastic pipe could be extracted by typical potable waters in concentrations exceeding the standards established in 1962 by the U.S. Public Health Service. In laboratory tests (Farish, 1969) of lead-stabilized PVC plastic pipe produced in the United States and abroad, lead concentrations in water ranged from 0.3 to 2.0 mg/liter. The concentrations were lower at high pH's, but even at pH 9.0 they exceeded the U.S. Public Health Service drinking water standard of 0.05 mg/liter. Whether these concentrations are typical of those found in water distribution systems was not reported.

Subsequent to these laboratory tests, the National Sanitation Foundation developed a standardized testing procedure for plastic pipes. Both physical and toxicological properties of pipes, fittings, and joining compounds used to transport potable water have been tested in the prescribed manner.

A number of laboratory test procedures have been used in Europe to determine the concentration of lead extracted from unplasticized polyvinyl chloride (uPVC) pipe. Packham (1971a,c) argued for the necessity of a standardized test. With his procedures, he found lead could be extracted from uPVC at concentrations typically ranging from 0.01 to 2 mg/liter, depending on flushing time, extraction procedure and sequence, and pipe manufacturer. The amount of lead extracted depends on the concentration of lead in the pipe material, the form in which it is present, and the process used in the manufacture of the pipe. Results from other researchers supported a conclusion that extractable lead was present in the form of a thin film on the inner pipe surface.

Eklund et al., (1978) conducted laboratory scale tests to evaluate the influence of different pipe materials on water quality. Sections of PVC and nonplastic water pipes were thoroughly rinsed with tap water for several hours, then filled with drinking water and stored for 24 hours. Volatile organics were analyzed with closed loop stripping and gas chromatography/mass spectrometry. With homogeneous PVC pipe, there were indications of three contaminants added to the tap water; however, these were present in concentrations too low (< 1 ng/liter) for structure elucidation. Nonvolatile organic substances and inorganic compounds were not determined.

The most extensive testing of PVC pipe to date has been conducted under the auspices of the California Department of Health Services. The goal of one study was to determine the health risks, if any, posed by the use of plastic pipe systems in homes. James M. Montgomery, Consulting Engineers, Inc. (1980), conducted laboratory tests to evaluate water quality in PVC and CPVC pipe for a number of constituents: solvent cements used for joints, volatile halogenated organics, volatile aromatic organics, base/neutral extractable organics, and metals.

Both static and usage simulation tests were made with piping and joints typical of those found in a two-bedroom house. The static system was designed to simulate new plumbing, containing stagnant water prior to occupancy so that leachable organics would be similar to those encountered during the first several times the system was used by the consumer. A 2-week period of stagnation was followed by flushing, refilling, and sampling intervals. The second system, postoccupancy usage simulation, was set up to simulate normal household use for a 30-day period, alternating 12 hours of flowing water with 12 hours of stagnant water. The kinetics of the static samples were also tested. Other variables in the tests included water temperature, water source, pipe material, type of solvent cement, and cementing procedure.

Solvents used for PVC cement and primers typically include methyl ethyl ketone, tetrahydrofuran, N,N-dimethylformamide, and cyclohexanone. The diffusion of solvent in 2-week static CPVC samples described above ranged from a low of 0.11 mg/liter for dimethylformamide in conventional systems filled with cold water to 375 mg/liter for tetrahydrofuran in the cold water in a poorly constructed system in which excess cement and primer were used. The most important variable affecting solvent diffusion appeared to be the quality of workmanship in the system. Bonded surfaces with excess cement and primer leached appreciably more of each solvent than the conventional systems. In subsequent refills of the CPVC pipes, the level of each solvent dramatically decreased although their residence time in the pipe was shorter. Most of the solvents were gone after the first 2 weeks of use.

In the usage simulation described above, solvent concentrations were measured in samples taken when 360, 1,800, 3,600, and 10,800 gallons (1,368, 6,840, 13,680, and 41,040 liters) of water had flowed through the system. Solvent levels decreased rapidly between samples analyzed at the 360- and 1,800-gallon mark. The level of methyl ethyl ketone, tetrahydrofuran, and cyclohexanone measured at the 1,800-gallon mark were nearly constant throughout the remainder of the experiment. Detectable levels of methyl ethyl ketone and tetrahydrofuran (12 and 33 mg/liter, respectively) continued to be leached after 34,280 gallons (130,264 liters) had passed through the system during 98 days of operation. N,N-Dimethylformamide was not detected in any sample; cyclohexanone was present at 0.05 mg/liter in the sample taken at 360-gallons but was not detectable at 10,800 gallons.

In these tests, 32 volatile halogenated or volatile aromatic organic compounds identified as priority pollutants by the EPA were analyzed by gas chromatography/mass spectrometry (GC/MS). Several volatile halogenated organic compounds were detected at levels higher than those found in the raw water, suggesting that they had leached from the PVC pipe. Concentrations of chloroform in the 2-week CPVC samples were 5 to 11 times greater than those in the raw water. Levels of carbon tetrachloride were 52 to 125 times higher than the concentration in the raw water. The maximum concentrations (146 µg/liter for chloroform and 50 µg/liter for carbon tetrachloride) were found in the excess-cement/cold-water system. These two compounds may have been formed when residual chlorine in the raw water reacted with methyl ethyl ketone and tetrahydrofuran acting as precursors or they may have diffused from the pipe material. The total trihalomethane concentration in water residing for 2 weeks in the excesssolvent/cold-water system (152 µg/liter) exceeded the 100 µg/liter standard set by the EPA. No other system in this study exceeded the limit.

Vinyl chloride monomer was not detected in any of the samples. The detection limit was 0.1 µg/liter. In the PVC pipe tests. 30 nonvolatile organic compounds were analyzed by base neutral and acid extraction followed by GC/MS. All that were found were attributable to contamination. Heavy metals were also analyzed in the PVC pipe test since they are sometimes used as heat stabilizers. After 2 weeks in the static PVC system. concentrations of metals in the water from the pipe were below EPA maximum contaminant levels and generally at or below levels in the influent water (Table III-13).

TABLE III-13. Metal Concentrations in Static Polyvinyl Chloride Systems.

TABLE III-13

Metal Concentrations in Static Polyvinyl Chloride Systems.

Field Tests

The few field tests conducted on water collected from the distribution system or home services also concentrated on PVC pipe. Packham (1971b) summarized findings of studies in European systems in which uPVC with lead stabilizers was used. In a study conducted in the Netherlands, 7 of 15 distribution systems contained no detectable lead, several were at or below 0.05 mg/liter, and the highest was 0.10 mg/liter. Pipes ranged in diameter from 15 to 100 mm and in age from 2 months to 5.5 years. In Italy, water from 32 uPVC distribution pipes containing lead in the pipe varying from 0.031% to 1.619%, had less than 0.07 mg/liter lead in 31 water samples.

Packham sampled 53 points in English distribution systems with uPVC pipe, under both normal flow and night flow conditions. The pipe had been in service from 0 to 74 months and ranged in length from 9 to 4,300 meters. Most lead concentrations were less than 0.01 mg/liter; the highest was 0.05 mg/liter.

In 1975, Dressman and McFarren (1978) studied five U.S. water distribution systems using PVC pipes to determine vinyl chloride monomer concentrations in samples collected before and after exposure to PVC pipe. Very low concentrations of vinyl chloride (0.03 to 1.3 µg/liter) were detected in four of the five water samples studied. The lowest concentration was found in a 9-year-old system and the highest was found in the newest system. In the fifth system, no vinyl chloride was detected at a detection limit of 0.03 µg/liter. These findings are summarized in Table III-14.

TABLE III-14. Vinyl Chloride in PVC Distribution System Piping.

TABLE III-14

Vinyl Chloride in PVC Distribution System Piping.

Vinyl chloride concentrations in the samples prior to exposure to the PVC pipe were less than the detection limit (i.e., < 0.03 µg/liter). The newest and longest system appeared to have the highest vinyl chloride concentration and the next longest had the next highest concentration. Traces of vinyl chloride at the nanogram per liter level were still present in two California systems, Pioneer and Roseburg, approximately 9 years after they were installed.

Control Measures

Since plastic pipes have only recently been tested for deterioration and possible leaching of contaminants into the water of distribution systems, control measures have not been fully developed. However, some control measures have been mentioned in the literature. These include requiring tests on pipe materials to demonstrate that the materials are acceptable for use in water distribution systems, specifying desired properties of plastic pipe materials, and limiting the use of plastic pipe in individual distribution systems.

The concept of requiring tests prior to using new pipe materials for potable water supplies is not a new one. Farish (1969) described the testing procedures used by the National Sanitation Foundation to evaluate physical properties of the pipe materials and toxicological and organoleptical evaluations. In some states, plastic pipe manufacturers are being required to demonstrate the leaching characteristics of their material. In California, for example, these tests have been used to determine if the state should restrict the use of PVC pipe materials (James M. Mongomery, 1980).

Another approach is to require conformance to standards. Both the American Water Works Association (1975a, 1978b) and the American Society for Testing and Materials (1974, 1978) have developed standards for PVC and PE pipe. In the Federal Republic of Germany, recommended quality criteria for PVC pipe encompass the following factors: raw materials, catalysts and additives, residues of decomposition products from catalysts, residues from emulsifiers, protective colloids, residues of precipitating materials, and stabilizers (Anonymous, 1977). Adherence to these pipe material standards could control the amount of unpolymerized monomer, stabilizers, or other constituents that may have toxic properties. Requiring that only certain evaluated components be permitted in cements and primers is another quality control method (James M. Montgomery, 1980).

Other recommendations involve the alteration of construction procedures so that both construction workers and homeowners are protected from solvent cements and primers that could leach into water. The California Department of Health Services (James M. Montgomery, 1980) has recommended the use of warning labels on all solvent-cemented pipe outlets in new construction and a specific protocol to be used by contractors for flushing newly installed pipe systems.

A final method of control includes regulatory actions such as banning the use of specific plastic materials for certain uses. For example, the U.S. Department of the Treasury has banned the sales of liquor in PVC bottles, and the FDA at one time proposed regulations to ban the use of vinyl chloride polymers in materials coming into contact with food to preclude the migration of vinyl chloride gas into the food (Rawls, 1975). However, these proposed regulations excluded certain uses such as water pipes.

Based on the evidence collected to date, a combination of materials testing and standards specification has been the typical approach to control.

Summary and Conclusions

There has been only a limited amount of work concerning the effects of plastic pipe on water quality, and this research has concentrated on PVC. Plastic pipe was developed to provide a material that would be inert to water, but some materials can leach out in trace quantities. In general, the initial water quality is not a major factor in the leaching of materials from these pipes in contrast to the definite effect of water quality on corrosion of metal and A/C pipe.

Plastic pipe is composed principally of polymerized organic compounds, but there can be residual unpolymerized monomer present in low concentrations. This is especially important in PVC pipe, since the vinyl chloride monomer is known to be carcinogenic in humans and animals (National Academy of Sciences, 1977). Plastic pipe can also contain other components, such as pigments, lubricants, stabilizers, and plasticizers. Pipe sections may be joined with cement and primers, which contain volatile solvents. When predicting the effect of piping on water quality, all of the pipe's constituents should be considered.

Laboratory tests have shown that effects of plastic pipe on water quality are a function of the type of pipe, the manufacturer, the flow pattern and duration, the contact time, the jointing material, the residual monomer concentration, the stabilizer type and concentration, and the testing protocol. A variety of laboratory tests of lead-stabilized PVC have shown that the lead is extracted by water and can be found in concentrations ranging from less than 0.01 to 2 mg/liter, depending on test conditions and pipe composition. There is a possibility that unpolymerized vinyl chloride monomer can be leached into water, but recent tests have shown no detectable monomer. Solvents from cement and primers as well as heavy metals have been found in laboratory tests of PVC and CPVC.

Samples of water exposed to PVC pipe in distribution systems have contained low concentrations of leachates from PVC pipe. presumably because the pipe had aged and the contact times were short. Samples from uPVC pipe in Europe contained lead concentrations typically between 0.01 and 0.05 mg/liter. In the United States, an EPA survey of five PVC distribution systems found vinyl chloride monomer concentrations ranging from below detection limits (0.03 µg/liter) to 1.4 µg/liter.

Possible control measures include standards for pipe composition, laboratory tests to ensure leachate concentrations are below accepted limits, or curtailment of pipe use.

Lining Materials

Nonmetallic linings are commonly used in potable water distribution systems to prevent corrosion of underlying metal components. Strictly speaking, linings are materials on the interior surfaces of pipes, tanks, or other facilities, thereby coming into direct contact with water. Coatings, on the other hand, are applied to the external surfaces of pipes or tanks and, thus, are not in contact with the water. Sometimes the terms ''linings'' and "coatings" are used interchangeably, but "linings" in its strict definition will be used herein for the sake of accuracy and to avoid confusion.

Frequently, steel or ductile iron pipes are lined with cement mortar, asphalt, coal tar, or compounds containing coal tar, vinyls, or epoxies. Steel water-storage tanks are lined with materials similar to those used in metallic pipe.

Although lining materials have been used successfully to help prevent deterioration of water quality by controlling corrosion and its by-products, the possibility that the materials might themselves affect water quality has recently been raised. Laboratory and field studies have shown that some of these materials, especially coal tar compounds, can leach contaminants into the water with which they come into contact. There is also some speculation that small particles of intact lining material can also be released from deteriorating linings.

The concern about linings and their potential impact on water quality is based primarily on the fact that most of these materials are organic in composition and many of them contain carcinogens. Examples include the polynuclear aromatic hydrocarbons (PAH's) found in coal tar compounds.

The amount of information on concentrations of materials released into the water from linings and coatings is limited. There have been a few laboratory studies of lined panels and several studies of water quality before and after exposure to distribution system materials; however, the extent to which the general population is exposed to materials released from linings is as yet unknown.

To obtain more in-depth information regarding the effects of linings on water quality, the EPA has formed an Additives Evaluation Branch within its Office of Drinking Water. This group is conducting further research to answer some of the questions about linings.

Many types of materials are used to line pipes, water storage tanks, and appurtenant facilities in the distribution system. For pipelines, the most commonly used materials in the United States are cast iron, ductile iron, asbestos-cement, and steel. Cast-iron and ductile-iron pipes are typically lined with cement mortar or cement mortar plus a bitumastic (asphalt combined with a filler) sealing coat. Steel pipes are usually lined with cement mortar. A/C pipe is typically unlined, although vinyl linings are used occasionally with aggressive water. Other linings for pipes include coal tar epoxy, other epoxy, coal tar emulsion, and polyamide cured epoxy.

To protect the approximately 0.5 million steel water-storage tanks in the United States from corrosion, a number of linings are used on the interior surfaces, including coal-tar-based materials, vinyl, epoxies, zinc rich, metallic rich, chlorinated rubber, and phenolic based materials. There are no statistical data on the number of tanks lined with specific materials (Goldfarb et al., 1979). We know, however, that the predominant paints are vinyl, epoxy, and coal tar. Estimates of the percentages of these various materials used vary widely.

The earliest U.S. water tanks were made of riveted wrought iron. At that time, little consideration was given to protective paint. Iron plates were usually primed with iron oxide paint, and the interior surfaces were painted with a single coat of asphalt paint or red lead. Most of the protection from corrosion resulted from the inherent resistance of the wrought iron itself. In the early 1900's, wrought iron was replaced with steel structures, which were painted the same way as the iron. However, since the steel had a lower corrosion resistance than the iron, paint failures were frequent. As a result, new linings were developed (Brotsky, 1977).

The protective mechanism of lining materials has been described by Wallington (1971). Paints protect steel principally by their high ionic resistance and low permeability to ions such as chlorides and sulfates, which accelerate corrosion. Paints do not have sufficiently low permeability to water and oxygen to stop corrosion simply by excluding them from the metal surface, although very thick linings probably retard this process.

The formulations of the different tank lining materials have been described by Goldfarb et al., (1979), Wallington (1971), and others. A practical guide to the selection of materials has been published by Banov and Schmidt (1978). Characteristics of typical lining materials are described below.

Coal Tar

Coal tar linings have been used in the United States since 1912 to protect the interior of steel pipelines carrying potable water (Goldfarb et al., 1979). They have also been used to line steel water-storage tanks.

Coal tar linings are made by combining coal tar pitch with other material to provide the desired properties. According to definitions developed by the American Society for Testing and Materials, coal tar is "a dark brown to black cementitious material produced by the destructive distillation of bituminous coal" and pitches are "black or dark-brown solid cementitious material which gradually liquifies when heated and which is obtained as residue in the partial evaporation or fractional distillation of tar." Coal tar pitch is ''composed almost entirely of polynuclear aromatic compounds and constitutes 48-65 percent of the usual grades of coal tar" (American Petroleum Institute, 1971).

In the United States, coal tar is usually produced as a by-product of the manufacture of metallurgical coke, and it is obtained by the destructive distillation of bituminous coal. The tar is recovered from coke oven gases by partial condensation. The material that escapes the initial partial condensation is "light oil" and gas. Further processing of the tar by fractional distillation produces tar acids, tar bases, naphthalene, and creosote oil. The residue that remains after distillation is commonly referred to as pitch. The pitch fraction resists penetration by water and deterioration by water action. For this reason, it is used for waterproofing and roofing and as protective coatings on buried or submerged iron and steel structures and pipelines (Goldfarb et al., 1979).

The coal tar material most commonly used in the United States for lining is coal tar enamel, as described in Standard C-203 of the American Water Works Association (1978a). The enamel is manufactured by dispersing coal in a mixture of coal tar pitch or refined coal and coal tar base oil with a high boiling point. It is strengthened by the incorporation of talc at approximately 30% by weight.

The resulting coal tar enamel has a long service life (in excess of 50 years), good resistance to erosion, a smooth surface, and seems to resist the buildup of algae and other growths (Goldfarb et al., 1979).

It is well established that coal tar contains polynuclear aromatic hydrocarbons (PAH's). Some of the PAH's found in coal tar and water are shown in Table III-15. The constituents of tar, summarized by Lowry (1945), include PAH's such as naphthalene, 1-methylnaphthalene, 2-methylnaphthalene, acenaphthene, fluorene, anthracene, phenanthrene, chrysene, pyrene, and fluoranthene. Other PAH's, heterocyclic nitrogen compounds, phenols, and a number of other constituents have also been found. Similarly, Wallcave (1971) has reported a number of PAH's in coal tar pitches.

TABLE III-15. Polynuclear Aromatic Hydrocarbons Found in Water.

TABLE III-15

Polynuclear Aromatic Hydrocarbons Found in Water.

Other coal tar derivatives are described below.

Asphalt

Asphalt is a "dark-brown to black cementitious material in which the predominating constituents are bitumens which occur in nature or are obtained in petroleum processing" (American Petroleum Institute, 1971). In contrast to coal tar, which is derived from coal, asphalt is derived from petroleum.

Ductile-iron pipes used in water distribution systems are typically lined with a sealing coat of petroleum asphalt, which is in direct contact with the water. This lining serves primarily to control the curing rate of an inner layer of Portland cement, which is used to retard corrosion of the pipe by aggressive waters (Miller et al., 1980). The asphalt lining aids the cement curing process by retaining moisture, and it protects the cement mortar from decalcification in soft water (Goldfarb et al., 1979). The asphalt lining for ductile pipe is more completely described by the American National Standard Institute (1980). According to the Ductile Iron Pipe Research Association, all U.S. manufacturers of cast- and ductile-iron pipe use an asphalt sealing material derived from the distillaton of petroleum products; this material does not contain any coal tar pitch (Stroud, 1980).

Concentrations of PAH's are much lower in asphalt linings than they are in coal tar linings. Miller et al., (1980) analyzed asphalt paint for several PAH's (fluoranthene, benzo(b)- and benzo(k)fluoranthene. benzo(a)pyrene, perylene. ideno( ,2,3.-cd)pyrene, benzo(ghi)perylene. anthracene, and chrysene). They found these compounds in concentrations ranging from 0.1 to 10 µg/g. Concentrations of the alkyl derivatives of these PAH's were 2 to 10 times greater. Wallcave et al., (1971) analyzed eight asphalts. They found 17 PAH's, typically ranging from 1 to 10 µg/g. Again, alkyl derivatives were frequently found in concentrations up to an order of magnitude higher.

Other Materials

Other materials used to line water distribution piping or tanks include coal tar epoxy, other epoxy, cement mortar, coal tar emulsion, epoxy phenolic formulations. and polyamide cured epoxy formulations (Goldfarb et al., 1979). When water is known to be aggressive to A/C pipe, vinyl linings have been used. Some of the other materials used to line water storage tanks include chlorinated rubber, metalized zinc, wax (applied cold or hot), phenolic, and zinc rich. As actually applied, the linings include coal tar enamel (hot), three-coat vinyl, metalized zinc, four-coat vinyl, chlorinated rubber, coal tar paint (applied cold), coal tar epoxy, two-component epoxy, asphalt, wax, one-component epoxy, phenolic compounds, and zinc (Goldfarb et al., 1979).

Release of Lining Materials into Distribution System

Research on the leaching of lining components into water has been limited. Most of it has been concentrated on PAH's from coal tar and asphalt material. Due to experimental conditions, laboratory investigators have typically found higher concentrations of these materials in the test water than are actually found in water distribution systems.

The magnitude of contamination in distribution systems is determined by a number of factors. These include the age of the coating, contact time of the water, ratios of surface area to water volume, the type of coating used, manner of application, and deposits of materials such as carbonate (Sorrell et al., 1980).

Many PAH's have been found in water in distribution systems (Sorrell et al., 1979). Structures of some of these compounds are shown in Table III-15. PAH contamination of raw waters has also been found.

There are no drinking water regulations for PAH's or other materials released from linings of distribution systems in the United States. The International Standards for Drinking Water (World Health Organization, 1971) contains limits for PAH's since several of them are known to be carcinogenic. These standards indicate that "the concentration of six representative PAH compounds (fluoranthene, 3.4-benzofluoranthene, 11.12-benzfluoranthene. 3,4-benzpyrene. 1.12-benzperylene. and indeno[1,2,3-cd]pyrene should ... not, in general. exceed 0.2 µg/liter." An update on PAH's and their significance to health is discussed in Chapter VII of this volume.

Laboratory Tests

The few laboratory tests conducted on the effect of linings on water quality have evaluated the leachings from coal tar, coal tar enamel, asphalt, and other materials. There is no standard procedure for conducting these tests. Thus, a variety of experimental conditions have prevailed. Furthermore, they have not been very representative of actual exposure conditions in the distribution systems. Therefore, the data from these tests can be used to show whether compounds do or do not leach from the coatings, but they are not valuable in predicting the concentrations of materials that would be found under actual conditions in the distribution system. A summary of the findings of those laboratory tests is contained in Table III-16.

TABLE III-16. Summary of Laboratory Tests on Linings.

TABLE III-16

Summary of Laboratory Tests on Linings.

Alben (1980a) conducted tests on steel panels coated with coal tar. After 1 week of static testing, several PAH's were found in the leachate samples. The highest concentration was phenanthrene (and possibly anthracene) at 125 µg/liter. According to Alben, phenanthrene and its isomer anthracene were not resolved by the GC/MS technique used in this study. but anthracene was considered to be the minor constituent based on the relative solubilities of the two compounds and analyses of coal tar. Concentrations of naphthalene, three methylated naphthalenes, fluorene, fluoranthene, and pyrene ranged from 13 to 56 µg/liter.

The U.S. Army Corps of Engineers (Lampo. personal communication, 1980) coated steel panels with coal tar pitch and immersed them in water. Compounds found in the leachate by liquid chromatography and GC/MS were not the typical PAH's. and no benzo(α)pyrene was found. Aza-arenes were detected; acridene was found in concentrations of 5 to 10 µg/liter. Although concentrations of the aza-arenes in the coal tar pitch are lower than those of the PAH's, the heterocyclic nitrogen compounds are more soluble in water, a factor that Lampo credits for the concentrations of aza-arenes in the leachate being higher than those of the PAH's. In a similar test with a 5-year-old panel, no compounds of these types were detected in the leachate; detection limits were not specified.

In an EPA laboratory, coal-tar-based coatings were tested on glass plates with flowing tap water (Sorrell et al., 1980). Concentrations of PAH's in the water after 25 and 165 days were phenanthrene at 230.000 to 290,000 ng/liter; fluoranthene. pyrene. and anthracene at 14.000 to 46.000 ng/liter: chrysene and benzo(a)anthracene at approximately 1.000 ng/liter; and others at approximately 100 ng/liter or less.

Petroleum asphalt coatings on ductile-iron pipe have been investigated by Miller et al., (1980). Samples were analyzed by high-pressure liquid chromatography with fluorescence detection and GC/MS. Using a recirculation test system and sampling times ranging from 10 minutes to 293 hours. Miller found that the sum of the concentrations of the six PAH's [fluoranthene. benzo(b)fluoranthene. benzo(k)fluoranthene. benzo(a)pyrene, benzo(ghi)perylene. indeno(l,2,3,-cd)pyrene] listed in the standards of the World Health Organization (1971) was less than 10 ng/liter. Concentrations typically were less than 1 ng/liter. Fluoranthene was measured at 7 ng/liter. the highest concentration for any single PAH covered by the World Health Organization standards. In one sample, anthracene was detected at a maximum concentration of 41 ng/liter. This test suggests that lower concentrations of PAH's are leached from asphalt coatings than from coal tar coatings, presumably because the asphalt initially contains lower concentrations of these PAH's.

Eklund et al., (1978) evaluated cast-iron water pipe with cement lining and with asphalt and polyurethane linings. Analyses were conducted by closed-loop stripping enrichment and GC/MS. Leachate from polyurethane-coated pipe was contaminated with chlorobenzene at about 1 µg/liter and by aromatic substances at approximately 5 µg/liter. Naphthalene, methylnaphthalene. biphenyl. methylbiphenyl. and dibenzofuran were also detected. For the cast-iron pipes with cement and asphalt linings, no influence on water quality was observed; however, the investigators did not analyze for nonvolatile organic substances and inorganic compounds.

An epoxide-alkyl enamel and a vinyl chloride/vinyl acetate copolymer paint proposed for painting submersible well pumps were tested as coatings on iron metals (Kupyrov et al., 1977). Neither the enamel nor the paint had any significant effect on the organoleptic or chemical properties of the water.

Sorrell et al., (1977) tested cement- and asphalt-lined cast-iron pipe. Using a number of analytical techniques, they identified phenanthrene (65 ng/liter), fluoranthene (4 to 6 ng/liter), pyrene (6 ng/liter), and 1-methylpyrene (2 ng/liter) in the water that had been in contact with the asphalt lining.

Field Tests

Most of the limited number of field studies conducted on this subject have focused on surveys of organic compounds, particularly PAH's. in the distribution system. This emphasis is apparently due to the fact that many pipes and reservoirs are lined with organic materials. However, there is no systematic information on leaching rates, changes in leachate composition over time, or decomposition products, nor is there complete information on constituents and their concentrations to which the public is generally exposed. Therefore, it is difficult to estimate the health effects that may result from these exposures.

Field studies often consider an inadequate number of distribution system materials that might come into contact with the water. Initial water quality data (other than PAH's) are generally not included in reports of distribution system sampling, but it is assumed that water quality is not an important factor in the leaching of these types of materials.

After surveying the data on the leaching of PAH's in U.S. water distribution systems. Sorrell et al., (1980) reported that only limited information is available. These data, including the presence of coal tar or asphalt lining, are summarized in Table III-17. Several interesting observations can be made from this accumulation of data. First, the concentrations of the majority of PAH's analyzed were less than I ng/liter. Second. phenanthrene appeared in higher concentrations than those of the other compounds, which is in agreement with the laboratory studies of leaching described above. Third, the Portland. Oregon, system, which had a coal tar lining, contained substantially higher concentrations of some of the PAH's than did the three systems with asphalt linings; however, the PAH concentrations in the water of the other two systems with coal tar linings were similar to those in systems with asphalt linings. It would be necessary to have more information on the size and age of the distribution system. the materials used in them, and the amount of water flow to make more substantive comments on the findings.

TABLE III-17. Polynuclear Aromatic Hydrocarbons in Finished and Distributed Waters.

TABLE III-17

Polynuclear Aromatic Hydrocarbons in Finished and Distributed Waters.

Another survey of PAH's in distribution systems has been conducted by Saxena et al., (1978). Treated water was collected from the treatment site as well as from various locations in the distribution system. Information from the analyses of the six PAH's included in the drinking water standards of the World Health Organization (1971) is summarized in Table III-18. Except for the samples from Wheeling, samples contained low concentrations of PAH's at ppt (ng/liter) levels. In many cases, water sampled from the distribution system contained higher concentrations of individual PAH's as well as of the total of the six PAH's in comparison to -concentrations in undistributed treated water at the water treatment plant. At two of the sites, the major PAH's introduced by the distribution system lines were benzo(ghi)perylene and indeno(1,2,3,-cd)pyrene. At two of the sites (Elkart and Fairborn). trace quantities of some PAH's were found in the undistributed treated water, but after the water was exposed to the distribution system piping, all six PAH's were detectable. In some cases, notably in Wheeling. levels of PAH decreased after exposure to the distribution system. This suggests that PAH may have been removed by adsorption onto the surface of certain kinds of pipes. The authors could not link PAH levels to distribution system linings because definitive information on lining materials at these locations could not be obtained.

TABLE III-18. Effect of the Water Distribution Process on the Levels of Polynuclear Aromatic Hydrocarbons.

TABLE III-18

Effect of the Water Distribution Process on the Levels of Polynuclear Aromatic Hydrocarbons.

An EPA-sponsored study (Saxena, 1979) is currently in progress. Water samples from several water supplies throughout the United States are being analyzed for PAH's and mutagenic activity. Preliminary results from four cities indicate that the six PAH's contained in the World Health Organization standards range in concentration from below detection limits to 6 ng/liter. These results are only preliminary, since the work is expected to continue for some time.

Zoeteman and Haring (1976) have studied PAH concentrations in European distribution systems. They found detectable levels of PAH's in 88% of 25 tap waters tested in the United Kingdom and in the Netherlands. The mean concentration of six PAH's in ng/liter was as follows: fluoranthene, 17 ng/liter; 1,12-benzoperylene, 5 ng/liter; 11.12-benzofluoranthene, 3 ng/liter; indeno(1,2,3,-cd)pyrene. 3 ng/liter: 3.4-benzofluoranthene, 3 ng/liter; and benzo(a)pyrene, 3 ng/liter. The authors also analyzed for PAH's at pumping stations and taps in 10 cities. Concentrations ranged from less than 3 ng/liter to 30 ng/liter for individual compounds at the tap. In most cases, there was no difference between the concentrations before and after exposure to the distribution system. In two cases, the concentration of total PAH's was less at the tap. The authors concluded that there was no significant PAH contamination in the tap waters resulting from exposure of the water to asphalt linings of cast-iron pipes. However, they also note that such linings are not used frequently in the Netherlands and more studies should be conducted in countries such as the United Kingdom where these materials are more commonly used.

Another approach to analyzing water in the distribution system is to measure the mutagenic activity of concentrates collected from various distribution systems. Schwartz et al., (1979) analyzed concentrated samples of water from three medium-sized water supplies using the Ames Salmonella test for mutagenicity. Thirty-liter raw water samples and 60-liter finished water samples were concentrated using polyurethane foam, solvent extracted, and evaporated to 1 ml for mutagenic testing.

Materials in the distribution systems included cast iron in one midwestern city, cement-lined ductile iron in another midwestern city, and a mixture of copper, cast iron, cement-lined ductile iron, and asbestos-cement -in a southeastern city. No further details on linings were described. Assays of the samples from the first two sites showed that mutagens were being contributed by the distribution process. Samples from the third site were not mutagenic. Only nonvolatile mutagens could be measured by this technique. Volatile compounds such as trihalomethanes or vinyl chloride would be lost during the sample preparation procedures. Results suggested that the distribution systems probably introduce two different classes of mutagens: low-molecular-weight polar compounds and highmolecular-weight. nonpolar compounds.

Schwartz et al., (1979) suggested that several mechanisms may be responsible for increased mutagenic activity in the water after distribution. First. mutagenic compounds could be leached from the interior surface of tanks and pipelines, which are frequently lined with coal tar—a mixture containing proven mutagens and animal carcinogens. Second, mutagens could be synthesized in water during its travel from the treatment plant to the sampling site as a result of chemical reactions (e.g., reaction of residual chlorine with organics). oxidation, or microbial action, which converts inactive chemicals into mutagens.

In addition to the general surveys of PAH's and the mutagenic properties of distribution system samples, there have been some case studies relating changes in water quality to linings at individual sites. The most notable case studies have been conducted in Pascagoula, Mississippi, and Portland. Oregon.

In Pascagoula (McClanahan, 1978), two water storage tanks were cleaned and lined with coal tar pitch. The larger Bayou Cassotte tank held 750.000 gallons, the Beach Tank was a smaller, elevated storage tank. After the tanks were filled, there were immediate taste and odor changes in the smaller tank. The tanks were drained, refilled, and again put into service, after which the appearance of bacteria were reported. Following another draining and filling cycle, complaints were again received, and a citizen inquired about the advisability of using the coal tar pitch as a lining material. Subsequent to two more draining, inspection, and filling cycles, the EPA collected samples from the Bayou Cassotte tank. Several PAH's were detected in concentrations ranging from I to 10 ug/liter. A number of solvents and other organics were also found in the water. The compounds in the highest concentrations were phenanthrene/anthracene (apparently not resolved by the analytical techniques) at approximately 9 µg/liter and naphthalene at about 6 µg/liter. Approximately 4 months later, the tank was drained and filled again. Samples taken shortly thereafter again showed a number of PAH's at µg/liter concentrations: phenanthrene/anthracene was approximately 5 µg/liter and fluorene was about 2 µg/liter. Again, several solvents and other organics were detected. From a sample collected about 1 month later, there were again several PAH's ranging from 1 to 10 µg/liter, and phenanthrene/anthracene was found in a concentration of 14 µg/liter in the bottom of the tank. Results are shown in Table III-19. The samples collected at the top and the bottom of the tank were affected differently by flow rate and solar heating. The EPA Region IV used the results of these analyses to estimate the concentration of phenanthrene/anthracene in water that had been exposed to a storage tank lined with coal tar pitch for approximately 1.5 days, an average exposure time under regular operating conditions. They estimated that a phenanthrene/anthracene concentration of approximately 5 µg/liter would be produced.

TABLE III-19. Constituents Leached from Bayou Cassotte Tank, Pascagoula, Mississippi.

TABLE III-19

Constituents Leached from Bayou Cassotte Tank, Pascagoula, Mississippi.

Subsequent to the sampling program in Pascagoula, the EPA recommended that the tanks be placed back in service on an interim basis to meet the water needs of that city (Kimm, 1978), but that monitoring should continue for 1 year. According to McClanahan (1978), the phenanthrene/anthracene concentrations decreased over time until they were less than 1 µg/liter in water samples taken from the outlet of the Bayou Cassotte Reservoir.

Leaching of material from coal-tar-based lining materials was also studied by the EPA in Portland, Oregon (Robeck, 1978). Water samples were collected from the source and at the terminal point of a pipe lined with coal tar of an unspecified type. The pipe was 24 in. (60.96 cm) in diameter and 2.43 mi (3.89 km) in length. Samples were analyzed for a number of PAH's (Table III-20). The PAH's were undetectable or at concentrations of approximately 1 µg/liter in the water obtained at the source, while detectable quantities of five PAH's were found after exposure to the coal-tarlined pipe. Phenanthrene was again the compound found at the highest concentration (3,225 ng/liter, while fluoranthene and pyrene were found at approximately 600-700 ng/liter). Other compounds were present at lower concentrations.

TABLE III-20. Compounds Detected in the Water Samples Obtained in Portland, Oregon.

TABLE III-20

Compounds Detected in the Water Samples Obtained in Portland, Oregon.

Alben (1980a) collected samples at the inlet and outlet of a 12,000-gallon (45,600-liter) storage tank, which had a 5-year old commercial coal-tar lining. Concentrations of PAH's in the effluent from the tank were 5 to 30 times higher than those in the influent. Influent and effluent concentrations in µg/liter were: naphthalene (0.004 and 0.025), fluorene (0.001 and 0.021), phenanthrene/anthracene (0.019 and 0.210), fluoranthene (0.003 and 0.081), and pyrene (0.002 to 0.071). The sum of the PAH's increased from 0.029 to 0.410 µg/liter.

In Champaign, Illinois, Lampo (personal communication, 1980) collected water samples from a tank relined a year earlier with a coal tar pitch/epoxy coating. Using liquid chromatography and GC/MS analyses, he did not detect materials in the water after exposure to the tank.

One feature regarding water quality deterioration, evident from several studies, is that some linings can introduce taste and odors into the water if the linings are improperly cured. For example, taste and odors noticed shortly after a tank was relined with coal tar pitch/epoxy in Champaign were believed to be caused by a solvent, xylene (Lampo. personal communication. 1980). In the Pascagoula incident, taste and odors resulting from the new coal tar lining were responsible for the initial investigation of organic contaminant concentrations. The compound(s) responsible for the taste and odor were not identified by McClanahan (1978). In addition to the solvents used to formulate or apply the coatings, another possible source of taste and odors is naphthalene, which is reported to have a threshold odor concentration in water of approximately 1 µg/liter (Zoeteman and Haring. 1976).

Another type of lining material, vinyl, has recently been found to leach organic compounds into water in the distribution system. In April 1980, the EPA recommended that installation of vinyl-lined A/C water pipe be suspended in New England because concentrations of tetrachloroethylene as high as 665 µg/liter were found in drinking water carried by the pipe. Tetrachloroethylene has been used to apply the lining to the A/C pipe. EPA is currently gathering information on the problem, including concentrations of the chemical in the pipe of different ages and in pipes that receive varying degrees of use (Anonymous. 1980).

The literature contains a few general comments on the deterioration of cement linings. For example, the AWWA Committee on Control of Water Quality in Transmission and Distribution Systems (Anonymous. 1971) has noted that there can be an increase in pH in mortar-lined pipe. even with a fairly continuous flow, if there is sufficient contact time. A pH increase of 0.4-0.6 was noted in an 28.8-km mortar-lined pipe in Grand Junction. Colorado. In Seattle (Dangel. 1975). increases in pH. alkalinity, and calcium have been observed in water from freshly relined mains. Water quality changes were reported to cease after a few weeks of flow. presumably after the uncombined calcium oxide had been removed from the cement.

Summary and Conclusions

In the United States and abroad, a number of lining materials are used in drinking water distribution systems to protect pipes and tanks. The most common of these materials are coal tar. petroleum asphalt. vinyl, epoxy. or some combination thereof.

Generally, concern about the use of these lining materials is rooted in their formulations, which are based on organic compounds. Some materials, such as coal tar, are comprised primarily of polynuclear aromatic hydrocarbons (PAH's). several of which are known to be carcinogenic in animals and humans. Constituents of the linings could reach the drinking water at the tap through leaching processes or possibly through physical deterioration and release of small particles of the lining material.

There have been a few laboratory studies on leaching from linings, principally from coal tar or asphalt linings. A number of PAH's and some heterocyclic compounds have been found to be leached from these types of coatings. Higher concentrations appear to leach from coal tar than from asphalt. However, these laboratory studies have not been conducted under standardized conditions, so it is difficult to compare one study to another and to predict from these studies what the actual concentrations of constituents would be in a water distribution system.

A number of field studies have been conducted to measure organic constituents after exposure to linings in the distribution system. Some of these studies have found PAH's at nanogram to microgram per liter levels. The compound most commonly found in these studies has been phenanthrene and/or anthracene. In some cases, these compounds are not resolvable by mass spectrometric techniques, but the isomer is believed to be phenanthrene. In addition to releasing PAH's, some linings have been found to release solvent constituents and compounds that produce tastes and odors if the lining is not properly cured. Finally, a limited amount of work has been done on mutagenic activity of water in the distribution system. In two of three samples studied, water after exposure to lining materials had higher mutagenic activity than water before exposure to these materials.

There is a need for more research on the effects of linings on water quality. The recently formed Additives Evaluation Branch of the EPA should begin to answer some of these questions. In particular, there are needs for specific and systematic data collection procedures, including information on the type of material, time in service, and water flow rates. Some standard exposure factor, such as square meters of lining area per liter of flowing water, is needed to establish a basis from which results of various research can be compared. Most studies have been conducted either on coal tar or on asphalt linings. More work needs to be done on other types of lining materials to gain a complete picture of potential water quality changes upon exposure to distribution system materials.

Deposits and Their Effects

Following water treatment, chemicals can deposit in water distribution systems, thereby affecting the quality of water distributed to the consumer. The most widely studied elements exhibiting this behavior are iron and manganese, but these studies have been examined primarily for their aesthetic and economic effects instead of their effects on health. Nevertheless, one can reasonably estimate the exposures of humans to these and other elements associated with their deposition and resuspension, as well as exposures to chemicals added for their control.

Aside from the aesthetic concerns—the visual impacts of iron and manganese suspensions including their discoloration of materials—these two elements are frequently associated with the growth of microorganisms in distribution systems (O'Conner, 1971). Their deposits are often resuspended, due to changes in flow, water temperature and quality, and water hammer.

High concentrations of iron and manganese affecting water quality in distribution systems are found not only in groundwaters but also in surface water supplies from deep lakes and eutrophic lakes and in surface waters exposed to mine drainage or acid industrial wastes (O'Connor, 1971). Calcium carbonate deposition and its effect on water quality have been discussed earlier in this chapter in the sections on chemical water quality indexes and corrosion. This discussion will briefly consider the behavior and control of chemical deposits, primarily to indicate possible exposures of humans. However, such exposures will necessarily be speculative because of the paucity of available information.

Approaches to Iron and Manganese Control

The control of manganese and iron in water supply systems has been widely practiced by one of three principal methods: oxidation (followed by precipitation and filtration), ion-exchange, and stabilization (O'Conner, 1971). The latter involves the use of dispersing agents to prevent their deposition in distribution systems. Despite these well-established techniques, however, water treatment plants are frequently unsuccessful in controlling these elements. In Nebraska, a survey of 29 water treatment plants practicing iron and manganese removal indicated only 60% were effective based on their ability to meet EPA secondary criteria of 0.3 mg/liter for iron and 0.05 mg/liter for manganese (Anderson et al., 1973). Even when the concentrations of iron and manganese in the raw water were quite low in one plant, the deposits accumulated and were flushed in the distribution system following fluctuations in the water's flow pattern. Similarly. O'Conner (1971) cited a 1960 survey of Illinois plants that indicated that iron in a reduced oxidation state was often found in the effluent from the filter at the water treatment plant.

Stabilizing Agents

Both naturally occurring and purposefully added organic or inorganic complexing agents can prevent oxidation and settling of iron or manganese in distribution systems. O'Conner (1971) discussed organic substances that behave in this fashion, including tannic. gallic, and ascorbic acids. Two common stabilizing agents applied at water treatment plants are phosphates and silicates.

One of the early uses of silicates for the prevention of iron deposition in a distribution system was reported by Henry (1950). He found that a silica sol (activated sodium silicate) produced by the partial neutralization of sodium silicate was effective in preventing or delaying the settling out of iron and produced no turbidity. In these pilot studies, it was also observed that large amounts of polyphosphates provided some similar protection, but produced objectional amounts of turbidity.

More recent studies of addition of silica to prevent iron and manganese deposition have been reported by Dart and Foley (1970, 1972) for well water supplies in Ontario, Canada. They noted that for waters with a natural silica content of 30 to 40 mg/liter, little decrease in iron concentrations resulted either from aeration or from chlorination followed by filtration. Although the finished waters contained concentrations of iron as high as 0.3 to 1.0 mg/liter. they appeared to be acceptable to the consumers. In contrast to the study of Henry (1950), Dart and Foley (1970, 1972) found that it was not necessary to activate (partially neutralize) the silicate, which in their studies was an effective additive retaining the iron in solution when used simultaneously with chlorination.

Dart and Foley (1970. 1972) followed the efficacy of such silicate treatment for approximately 2 years. They found it to be successful with initial iron concentrations up to 1.3 mg/liter and silicate additions up to 6.2 mg/liter. They noted that the iron-staining properties of the distributed waters diminished. There had been some concern that domestic hot water tanks might accumulate iron by thermal break-up of the silicate complex. but this apparently did not happen. They judged that the likely mechanism of iron control by the addition of silicate depended on the chelation of freshly produced ferric ions by the orthosilicate anion. Considering the adsorption by anion exchange resins, they concluded that the complex species formed by the ferric and silicate ions was negatively charged, although Weber and Stumm (1965) have demonstrated the formation of a positively charged iron silicate [FeSiO(OH)32+] complex. Dart and Foley (1972) also found that silicate sequestered (complexed) manganese in a well water when chlorination and elevated pH resulted in the desired oxidation of the manganese.

Although a variety of inorganic phosphorus compounds can be used in water treatment. Aulenbach (1971) noted that polyphosphates are the most effective in controlling iron, although the dispersion of the iron may not be permanent, particularly as the polyphosphate breaks down. hydrolyzing to orthophosphate. To be most effective, the polyphosphate should be added before oxidation, e.g., by aeration or chlorination, and applied at concentrations ranging from 1 to 5 mg/liter as phosphate at a suggested weight ratio of two parts phosphate to one part iron.

Anderson et al., (1973) reported the frequent use of sodium hexametaphosphate, a polyphosphate, in controlling iron and manganese in water supply systems, but described one such groundwater supply in which this additive was not successful. Of special interest is their finding that copper was also precipitated into the distribution system with the iron and manganese deposits. They noted that copper ions can sorb onto ferric hydroxide colloids at pH values greater than 5.

Tuovinen et al., (1980) have also shown that heavy metals are associated with tubercles in distribution systems. Such metals as iron, manganese. lead, and copper can, when they are freed by hydraulic stresses and increased flow velocities, increase in concentration in the water delivered to the consumer. Table III-21 compares heavy metal concentrations in ''red waters'' with annual average concentrations in finished water in the Columbus, Ohio, distribution system.

TABLE III-21. Heavy Metals in Analysis of Red Water and in Finished Water in the Columbus Distribution System.

TABLE III-21

Heavy Metals in Analysis of Red Water and in Finished Water in the Columbus Distribution System.

Summary and Conclusions

Several metals can be deposited and resuspended in water supply distribution systems. There may be fluctuations in the concentrations of other organic or inorganic chemicals that coprecipitate or sorb with the metals and may influence the bacterial concentrations.

Of special interest is the possible potential adverse health effects of silicates and polyphosphates, which are added as stabilizing agents. particularly when complexed with manganese. Other heavy metals are also maintained in solution by the addition of these dispersing agents. Whether this can cause a substantial increase in the exposures of humans to these metals is not known.

Reactions Involving Water Treatment Chemicals in the Distribution System

Reactions initiated in a water treatment plant may not achieve chemical equilibrium at the treatment site. Thus, entirely new reactions may be initiated between treatment chemicals and extraneous substances in the distribution system. The potential for systemic chemical change through homogeneous and heterogeneous chemical reaction is too large to support the sanguine notion that water quality leaving a treatment plant is identical to that flowing from individual consumer outlets (Larson, 1966).

The reactions of chlorine with organic carbon in the treatment plant (Rook, 1974; Thomason et al., 1978) and with model organic compounds (Larson and Rockwell, 1979; Norwood et al., 1980; Rook, 1974) have received increased attention in recent years. Studies have been focused on reactions in the plant itself or on reactions under simulated treatment plant conditions. Emphasis has been placed on the apparently ubiquitous trihalomethane reaction product. Unfortunately, there are only limited data on the continued reactions of chlorine in actual distribution systems. There are more limited experimental data on the effects of ozone (Sievers et al., 1977) and chlorine dioxide (Gordon et al., 1972; Stevens and Symons, 1976) on selected organic structures and treated river water. Virtually no scientific data exist on the reactions of other disinfectants, e.g., ozone, chloramines, and chlorine dioxide, in actual distribution systems.

Therefore, this section is necessarily limited to the few reports of extended chlorine-trihalomethane reactions in distribution systems and the reactions of chlorine residuals with organic substrates that may exist in distribution systems.

The rapid hydrolysis of chlorine (Reaction 1) is well documented, and the equilibrium constant (3.96 × 104 at 25°C) requires that very little molecular chlorine be present at pH values above 3 and total chlorine concentrations below 1,000 µg/liter (Morris, 1967):

Image img00016.jpg

Similarly, the initial reaction of chlorine with ammonia in aqueous solutions is rapid (Weil and Morris, 1949):

Image img00017.jpg

When concentrations of ammonia are low, this reaction is followed by further and slower chloramine reactions, which eventually oxidize ammonia nitrogen to N2 with subsequent loss of oxidant (Wei and Morris, 1974). The latter reactions can occur in the distribution system along with chlorine exchange reactions with organic nitrogen (Wajon and Morris, 1980).

Most studies of chlorine interactions with nitrogen compounds have focused on reactions with ammonia. In most high-quality raw water, only a small part of the total nitrogen is present as free ammonia. Reactions with organic nitrogen compounds involve both cleavage of compounds such as protein and heterocyclic nitrogen-containing compounds and the formation of N-chloro organic species (Morris, 1967), which may be analytically mistaken as free chloramines. There are insufficient data to permit further characterization of these reactions or their effects on water quality in distribution systems. It is clear from even the earliest data on hypochlorite ion and aquatic humic reactions that the ultimate concentration of total trihalomethanes (TTHM) is a function of reaction time, temperature, and pH, given an initial total aqueous carbon value and the presence of chlorine as hypochlorite ion. The analytical methodology for TTHM recognizes these variables and distinguishes between trihalomethane values measured at any point in time (instantaneous THM) and those values for samples held in bottles for longer periods (5-7 days) (THM formation potential) (Stevens and Symons, 1976). Recent studies (Brett and Calverley, 1979; DeMarco, personal communication, 1980) have verified that THM values actually increase with residence time in distribution systems as long as both chlorine and organic precursors are available.

It cannot be ascertained whether this phenomenon is due to simple homogeneous reaction kinetics of the hurric materials and hypochlorite ion or to more complex heterogeneous reactions controlled by the physical size and shape of the humic macromolecules. It is also possible that complex homogeneous reactions occur with rate-controlling steps involving the production of chloroform from several sites in the humic macromolecules, the reactivities of which are dependent on partial oxidation by hypochlorite ion. It is attractive to assume that the THM increase is not due solely to additional reaction of hypochlorite ion with extraneous organic precursors in distribution systems, since good correlations have been observed for municipal systems (Brett and Calverley, 1979) between treatment plant effluent samples aged in the laboratory and samples withdrawn from the distribution system after equivalent periods. In these cases, supported by data on real systems, samples at the consumer tap (3 days system residence) may be approximately twice the THM values leaving the plant (Brett and Calverley, 1979).

Reaction of hypochlorite ion with extraneous organic material in a distribution system is probable, although the dominant reaction products may not be THM's. Organic nitrogen compounds have already been mentioned and additional humic input from soil contact should not be disregarded.

Humic/hypochlorite ion reactions form a variety of other chlorinated and unchlorinated reaction products in laboratory experiments (Table III-22). Since the rates of these processes have not been investigated, it is not possible to state whether their concentrations might be expected to increase in distribution systems. Indeed, investigators have not even searched for them in real water distribution systems.

TABLE III-22. Nonvolatile Reaction Products of Humic Materials and Hypochlorite Ion,.

TABLE III-22

Nonvolatile Reaction Products of Humic Materials and Hypochlorite Ion,.

As discussed above, the ubiquity of PAH's in water distribution systems is well known (Blumer, 1976). They may enter drinking water via atmospheric deposition in open reservoirs or through leaching from lining materials in distribution systems.

The presence of hypochlorite ion in distribution systems may affect the qualitative distribution of the PAH's in drinking water. Alben (1980b) reported that abundant oxygenated and halogenated PAH's were found in chlorinated coal tar leachate samples, whereas parent PAH's, alkyl- and nitrogen-substituted PAH's, were predominant in unchlorinated samples. At chlorination levels of 50 mg/liter, the dominant PAH in leachate samples was fluorene, whereas phenanthrene dominated unchlorinated samples. Carlson et al., (1975, 1978) have shown that exposure of PAH's to aqueous chlorine reduces their concentration and produces material more lipophilic than the parent hydrocarbon (Table III-23). The relevance of the reactions and reaction products listed in Table III-23 to real distribution systems has not been established.

TABLE III-23. Chlorination Products of Selected Polynuclear Aromatic Hydrocarbons.

TABLE III-23

Chlorination Products of Selected Polynuclear Aromatic Hydrocarbons.

The effect of increased lipophilicity on bioaccumulation factors is unknown as is the nature of the effect of chlorine substitution on carcinogenicity of the compounds. However, it is known that the carcinogenicity of chemical compounds is enhanced by the halogen content.

The growth of algae in the reservoirs of distribution systems may result in the release of significant quantities of metabolic products into the water. The excretion of a wide variety of relatively complex organic structures is apparently common to almost all species of algae and is not confined to stressed cells (Barnes. 1978). Excretion of glycolate by Chlorella and Chlamydomonas is well documented, and green algae tend to reduce glycolate excretion in favor of higher molecular weight compounds as the cultures age. Many other types of compounds have been identified in cultures of various species (Table III-24).

TABLE III-24. Some Extracellular Products of Algae.

TABLE III-24

Some Extracellular Products of Algae.

Decomposition of algal biomass is another source of reactive organic material. Approximately one-half of the biomass may be converted to soluble. short-chain fatty acids in the presence of oxygen and bacteria. The remainder may be converted to refractory, humic-like substances. The reactivity of these materials with hypochlorite ion or with chloramines is virtually unexplored.

Summary And Conclusions

Although one can describe possible reactions between various organic substrates and different oxidants in distribution systems. hard scientific data on real distribution systems are extremely limited. Data suggest that THM concentrations continue to increase in the distribution system as long as both organic precursors and chlorine are present. It is probable that the nonvolatile reaction products of humic material and chlorine also increase in distribution systems, although there are no data for real systems.

It is attractive to assume that chlorine will react with trace amounts of other organic substrates in various distribution systems. e.g., PAH's from pipe linings or excretion products from algae in open reservoirs. Unfortunately, no existing experimental evidence would permit testing of these assumptions.

References

  • A-C Pipe Producers Association. 1980. A/C Pipe and Drinking Water. A-C Pipe Producers Association. Arlington. Va. 20 pp.
  • Ackerman, J. 1980. Bellotti weighs suit over water pipe hazard. The Boston Globe. June 16, 1980. pp.17-24.
  • Alben, K. 1980. a. Coal tar coatings of storage tanks. A source of contamination of the potable water supply. Environ. Sci. Technol. 14:468-470. [PubMed: 22288638]
  • Alben, K. 1980. b. Gas chromatographic mass spectrometric analysis of chlorination effects on commercial coal tar lechate. Anal. Chem. 52:1825-1828.
  • American National Standard Institute. 1980. American National Standard for Cement Mortar Lining for Ductile-Iron and Gray-Iron Pipe and Fittings for Water. Standard A21.4-80. American National Standard Institute. New York.
  • American Petroleum Institute. 1971. Introduction. Pp. 1-2 in Petroleum Asphalt and Health. Medical Research Report No. EA 7103. American Petroleum Institute. Washington. D.C.
  • American Society for Testing and Materials. 1974. ASTM Standard Specification for Polyethylene (PE) Plastic Pipe (SDR-PR). ASTM D2239-74. Pp. 132-140 in Annual Book of ASTM Standards. American Society for Testing and Materials. Philadelphia. Pa.
  • American Society for Testing and Materials. 1976. Standard Methods of Testing Asbestos-Cement Pipe. ASTM C500. American Society for Testing and Materials. Philadelphia. Pa.
  • American Society for Testing and Materials. 1978. ASTM Standard Specification for Polyvinyl Chloride (PVC) Plastic Pipe (SDR-PR). ASTM D2241-78. Pp. 141-147 in Annual Books of ASTM Standards. American Society for Testing and Materials. Philadelphia. Pa.
  • American Water Works Association. 1971. Control of Water Quality in Transmission and Distribution Systems-Committee Report. Quality Control in Distribution Systems. AWWA Annual Conference June 17. 1971.
  • American Water Works Association. 1975. a. Specification 4 in. Through 12 in. Polyvinyl Chloride (PVC) Pipe. AWWA C900-75. American Water Works Association. Denver. Colo.
  • American Water Works Association. 1975. b. Standards Committee on Asbestos-Cement Pressure Pipe. AWWA standard for asbestos-cement transmission pipe. for water and other liquids. AWWA C402-75. J. Am. Water Works Assoc. 67(8):462-467.
  • American Water Works Association. 1978. a. AWWA Standard for Coal-Tar Epoxy Coating System for the Interior and Exterior of Steel Water Pipe. ANSI/AWWA C210-78. American Water Works Association. Denver. Colo. 9 pp.
  • American Water Works Association. 1978. b. AWWA Standard for Polybutylene (PB) Pressure Pipe. Tubing. and Fittings. 1/2 in. Through 3 in.. for Water. AWWA C902-78. First Edition American Water Works Association. Denver. Colo. 11 pp.
  • American Water Works Association. 1979. The use of plastics in distribution systems. Committee on the Use of Plastics in Distribution Systems. J. Am. Water Works Assoc. 71:373-375.
  • American Water Works Association. 1980. AWWA Standard for Asbestos-Cement Distribution Pipe. 4 in. Through 16 in. (100 mm Through 16 in.) (100 mm Through 400 mm) NPS. for Water and Other Liquids. AWWA C400-80. American Water Works Association. Denver. Colo.
  • American Water Works Association Research Foundation. 1974. A study of the problem of asbestos in water. J. Am. Water Works Assoc. 66:1-22.
  • Andelman, J.B. 1974. The effect of water treatment and distribution of trace element concentrations. Pp. 423-440 in A.J. Rubin, editor. . ed. Chemistry of Water Supply. Treatment. and Distribution. Ann Arbor Science Publishers. Inc.. Ann Arbor. Mich.
  • Anderson, D.R.. D.D. Row. and G.E. Sindelar. 1973. Iron and manganese studies of Nebraska water supplies. J. Am. Water Works Assoc. 65:635-641.
  • Anonymous. 1971. Control of water quality in transmission and distribution systems. Committee report. Quality control in distribution. J. Am. Water Works Assoc. 63:741-742.
  • Anonymous. 1977. Gesundheitliche Beurteilung von Kunststoffen und anderen nichtmetallischen Werkstoffen im Rahmen des Lebensmittel und Bedarfsgegenständegestzes für den Trinkwasserbereich. Bundesgesundheitblatt 7:10-13.
  • Anonymous. 1980. TCE leaching prompts EPA to seek halt to installation of vinyl lined pipes. Chem. Regulation Reporter 4(3):63-64.
  • Aulenbach, D.B. 1971. Determining phosphate additive for iron control in water. J. Am. Water Works Assoc. 63:197-198.
  • Banov, A. and H.J. Schmidt, Jr. 1978. Techniques and materials for preventing corrosion. In R.L. Sanks, editor. . ed. Water Treatment Plant Design for the Practicing Engineer. Ann Arbor Science Publishers. Inc.. Ann Arbor. Mich.
  • Barnes, D.B. 1978. Trihalomethane-Forming Potential of Algal Extracellular Products and Biomass. Master's Thesis. Virginia Polytechnic Institute and State University. Blacksburg. Va.
  • Bean, E.L. 1970. Water quality goals and corrosion by soft waters. Pp. 1-16 in Control of Corrosion by Soft Waters. Education Committee. American Water Works Association. and U.S. Public Health Service. Environmental Control Association. Washington. D.C.
  • Blumer, M. 1976. Polycyclic aromatic compounds in nature. Sci. Am. 234(3):34-45. [PubMed: 1251182]
  • Booth, G.H., and A.K. Tiller. 1960. Polarization studies of mild steel in cultures of sulfatereducing bacteria. Trans. Faraday Soc. 56:1689.
  • Brett, R.W., and R.A. Calverley. 1979. A one-year survey of trihalomethane concentration changes within a distribution system. J. Am. Water Works Assoc. 71:515-520.
  • Brotsky, B. 1977. Interior maintenance of elevated storage tanks. J. Am. Water Works Assoc. 69:506-510.
  • Buelow, R.W., J.R. Millette, E.F. McFarren, and J.M. Symons. 1980. The behavior of asbestos-cement pipe under various water quality conditions: A progress report. J. Am. Water Works Assoc. 72:91-102.
  • Bull, R.J., and G.F. Craun. 1977. Health effects associated with manganese in drinking water. J. Am. Water Works Assoc. 69:662-663.
  • Camp, T.R., editor. , ed. 1963. Water and Its Impurities. Reinhold Publishing Corp. New York. 355 pp.
  • Carlson, R.M., R.E. Carlson, H.L. Kopperman, and R. Caple. 1975. Facile incorporation of chlorine into aromatic systems during aqueous chlorination processes. Environ. Sci. Technol. 9:674-675.
  • Carlson, R.M., R. Cople, A. R. Oyler, K.J. Welch, D.L. Bodenner, and R. Liukkonen. 1978. Aqueous chlorination products of polynuclear aromatic hydrocarbons. Pp. 59-65 in R.L. Jolley, editor; , H. Gorchev, editor; , and D.H. Hamilton, Jr., editor. eds. Water Chlorination: Environmental Impact and Health Effects. Vol. 2. Ann Arbor Science Publishers. Inc. Ann Arbor. Mich.
  • Christman, R.F., J.D. Johnson, F.K. Pfaender, D.L. Norwood, M.R. Webb, J.R. Haas, and M.J. Bobenrieth. 1980. Chemical identification of aquatic humic chlorination products. Pp. 75-83 in R.L. Jolly, editor; , W.A. Brungs, editor; , R.B. Cumming, editor; , and V.A. Jacobs, editor. , eds. Water Chlorination: Environmental Impact and Health Effects. Vol. 3. Ann Arbor Science Publishers. Inc., Ann Arbor. Mich.
  • Cook, P.M., G.E. Glass, and J.H. Tucker. 1974. Asbestiform amphibole minerals: Detection and measurement of high concentrations in municipal water supplies. Science 185:853-855. [PubMed: 4843791]
  • Cooper, R.C., M. Kanarek, J. Murchio, P. Conforti, L. Jackson, R. Collard, and D. Lysmer. 1978. Asbestos in Domestic Water Supplies in Five California Counties. Progress Report for Period April 25. 1977 to June 30. 1978. Contract No. R804366-02. U.S. Environmental Protection Agency, EHS Publ. No. 78-2. School of Public Health. Environmental Health Sciences. University of California. Berkeley.
  • Courchene, J.E., and G.J. Kirmeyer. 1978. Seattle Internal Corrosion Control Plan. Phase 1. Summary. Seattle Water Department, Seattle. Wash.
  • Craun, G.F., and J.R, Millette. 1977. Exposure to asbestos fibers in water distribution systems. Pp. 1-13 in Proceedings of the American Water Works Association 97th Annual Conference. Anaheim. Calif., May 8-13, 1977. American Water Works Association. Denver, Colo.
  • Cunningham, H.M., and R. Pontefract. 1971. Asbestos fibres in beverages and drinking water. Nature 232:332-333. [PubMed: 5094836]
  • Dangel, R.A. 1975. Study of Corrosion Products in the Seattle Water Department Tolt >Distribution System. EPA-670/2-75-036. U.S. Environmental Protection Agency. Cincinnati. Ohio. 22 pp.
  • Dan, F.J., and P.D. Foley. 1970. Preventing iron deposition with sodium silicate. J. Am. Water Works Assoc. 62:663-668.
  • Dart, F.J., and P.D. Foley. 1972. Silicate as Fe. Mn deposition preventative in distribution systems. J. Am. Water Works Assoc. 64:244-249.
  • Davis, J.B. 1967. Petroleum Microbiology. Elsevier Publishing Company. New York.
  • Dressman, R.C., and E.F. McFarren. 1978. Determination of vinyl chloride migration from polyvinyl chloride pipe into water. J. Am. Water Works Assoc. 70:29-30.
  • Eklund, G., B. Josefsson, and C. Roos. 1978. The leaching of volatile organic compounds from different types of water pipes. Vatten 34:207-208.
  • Farish, C.A. 1969. Plastic pipe and water quality. J. Am. Water Works Assoc. 61:480-482.
  • Goldfarb, A.S., J. Konz, and P. Walker. 1979. Interior Coatings in Potable Water Tanks and Pipelines. Coal Tar Based Materials and Their Alternatives. MTR 7803. Prepared for EPA Criteria and Standards Division. Office of Drinking Water. EPA Contract 570/9-79-001. Mitre Corp. McLean. Va. 140 pp.
  • Gordon, G., R.G. Kieffer, and D.H. Rosenblatt. 1972. The chemistry of chlorine dioxide. Pp. 201-286 in S.J. Lippard, editor. . ed. Progress in Inorganic Chemistry. Vol. 15. WileyInterscience. New York.
  • Hadley, R.F. 1948. Corrosion by microorganisms in aqueous and soil environments. Pp. 466-481 in H.H. Uhlig, editor. . ed. The Corrosion Handbook. Sponsored by the Electrochemical Society. Inc., New York. John Wiley & Sons. Inc.. New York.
  • Hallenbeck, W.H., E.H. Chen, C.S. Hesse, K. Patel-Mandlik, and A.R. Wolff. 1978. Is chrysotile asbestos released from asbestos-cement pipe into drinking water? J. Am. Water Works Assoc. 70:97-102.
  • Haring, B.J.A. 1978. Human exposure to metals released from water distribution systems. with particular reference to water consumption patterns, Trib. CEBEDEAU No. 419:349-355.
  • Henry, C.R. 1950. Prevention of the settlement of iron. J. Am. Water Works Assoc. 42:887-896.
  • Hoyt, B.P., G.J. Kirmeyer, and J.E. Courchene. 1979. Evaluating home plumbing corrosion problems. J. Am. Water Works Assoc. 71:720-725.
  • Hudson, H.E., Jr. and F.W. Gilcreas. 1976. Health and economic aspects of water hardness and corrosiveness. J. Am. Water Works Assoc. 68:201-204.
  • Iverson, W.P. 1974. Microbial Corrosion of Iron in Microbial Iron Metabolism. Academic Press. Inc.. San Francisco. Calif.
  • James M. Montgomery. Consulting Engineers. Inc. 1980. Solvent Leaching From Potable Water Plastic Pipes. Final Report. Prepared for the Hazard Alert System. California Department of Health Services/Department of Industrial Relations. James M. Montgomery. Consulting Engineers. Inc.. Pasadena. Calif. 67 pp.
  • Karalekas, P.C., Jr. G.F. Craun, A.F. Hammonds, C.R. Ryan, and D. J. Worth. 1976. Lead and other trace metals in drinking water in the Boston Metropolitan area. J. N. Engl. Water Works Assoc. 90:150-172.
  • Karalekas, P.C., Jr., C.R. Ryan, C.D. Larson, and F.B. Taylor. 1978. Alternative methods for controlling the corrosion of lead pipe. J. N. Engl. Water Works Assoc. 92:159-178.
  • Kay, G.H. 1974. Asbestos in drinking water. J. Am. Water Works Assoc. 66:513-514. Kennedy Engineers. 1978. Internal Corrosion Study. Prepared for City of Seattle Water Department. Seattle. Wash.
  • Kimm, V.J. 1978. Memorandum on Coal Tar Pitch Coating. Pascagoula. Miss.. to Gary Hutchinson. Water Supply Representative. Region IV. Atlanta. Ga.
  • Kirmeyer, G.J., G.S. Logsdon, J.E. Courchene, and R.R. Jones. 1979. Removal of naturally occurring asbestos fibers from Seattle's Cascade Mountain water source. Pp. 903-927 in Proceedings of the American Water Works Association Annual Conference. Part II. June 24-29. 1979. San Francisco. Calif.
  • Kupyrov, V.N., V.I. Vorobets, R.K. Gakal, and S.N. Starchenko. 1977. Hygienic characteristics of some polymer coatings suggested for use in drinking water supply. Gig. Sanit. No. 3:96-97. [PubMed: 852727]
  • Lane, R.W., and C.H. Neff. 1969. Materials selection for piping in chemically treated water systems. Materials Protection 8:27-30.
  • Langelier, W.F. 1936. The analytical control of anti-corrosion water treatment. J. Am. Water Works Assoc. 28:1500-1521.
  • Larson, R.A., and A.L. Rockwell. 1979. Chloroform and chlorophenol production by decarboxylation of natural acids during aqueous chlorination. Environ. Sci. Technol. 13:325-329.
  • Larson, T.E. 1966. Deterioration of water quality in distribution systems. J. Am. Water Works Assoc. 58:1307-1316.
  • Larson, T.E. 1971. Corrosion phenomena-Causes and cures. Pp. 295-312 in Water Quality and Treatment. A Handbook of Public Water Supplies. Third Edition. Prepared by the American Water Works Association. McGraw-Hill Book Co., Inc., New York.
  • Lee, S.H., and J.T. O'Connor. 1975. Biologically mediated deterioration of water quality in distribution systems. American Water Works Association 95th Annual Conference Proceedings. Paper 22-6. American Water Works Association. Denver. Colo.
  • Levine, R.J. 1978. Asbestos-An Information Resource. DHEW Publ. No. (NIH) 79-1681. U.S. Department of Health. Education. and Welfare. Bethesda. Md.
  • Macrae, I.C., and J.F. Edwards. 1972. Adsorption of colloidal iron by bacteria. Appl. Microbiol. 24:819. [PMC free article: PMC380669] [PubMed: 4629705]
  • McCabe, L.J. 1970. Corrosion by soft water: Metal levels found in distribution samples. Presented at Seminar on Control of Corrosion by Soft Waters. Education Committee of American Water Works Association and U.S. Public Health Service. Washington. D.C. 6 pp.
  • McCabe, L.J., and J.R. Millette. 1979. Health effects and prevalence of asbestos fibers in drinking water. In Proceedings of the American Water Works Association Annual Conference. June 24-29. 1979, San Francisco. Calif.
  • McClanahan, M. 1978. Water Supply Branch. EPA Region IV. Memorandum on Coal Tar Pitch Coating, Pascagoula. Miss., to J.A. Cotruvo. Director. Criteria and Standards Division. Office of Drinking Water, U.S. Environmental Protection Agency. Washington. D.C.
  • McFaren, E.F., R.W. Buelow, R.C. Thurnau, M. Gardels, R.K. Sorrell, P. Snyder, and R.C. Dressman. 1977. Water quality deterioration in the distribution system. Pp. 1-10 in Proceedings of the 5th Annual AWWA Water Quality Technology Conference. Kansas City. Mo.. Dec. 4-7. 1977. American Water Works Association. Denver. Colo.
  • Michaels, L., editor; , and S.S. Chissick, editor. . eds. 1979. Asbestos. Properties. Applications. and Hazards. Vol. 1. John Wiley & Sons. Inc. 568 pp.
  • Miller, H.C., W.J. Barrett, and R.H. James. 1980. Investigation of potential water contamination by petroleum asphalt coatings in ductile-iron pipe. Pp. 561-578 in 1980 Annual Conference Proceedings: Water for the World Challenge of the 80's. Atlanta. Ga.. June 15-20. 1980. American Water Works Association. Denver. Colo.
  • Miller, J.D.A., and A.K. Tiller. 1970. Microbial corrosion of buried and immersed metal. In J.D.A. Miller, editor. , ed. Microbial Aspects of Metallurgy. American Elsevier Publishing Co., Inc. New York.
  • Millette, J.R., P.J. Clark, and M.F. Pansing. 1979. Exposure to Asbestos from Drinking Water in the United States. EPA-600/1-79-028. Health Effects Research Laboratory. U.S. Environmental Protection Agency. Cincinnati. Ohio. 87 pp.
  • Millette, J.R., A.F. Hammonds, M.F. Pansing, E.C. Hansen, and P.J. Clark. 1980. Aggressive water: Assessing the extent of the problem. J. Am. Water Works Assoc. 72:262-266.
  • Morris, J.C. 1967. Kinetics of reactions between aqueous chlorine and nitrogen compounds. Pp. 23-53 in S.D. Faust, editor; and J.V. Hunter, editor. , eds. Principles and Applications of Water Chemistry. Proceedings of the Fourth Rudolfs Research Conference. Rutgers-The State University. New Brunswick. N.J. John Wiley & Sons. New York.
  • Mullen, E.D., and J.A. Ritter. 1980. Monitoring and controlling corrosion by potable water. J. Am. Water Works Assoc. 72:286-291.
  • National Academy of Sciences. 1977. Drinking Water and Health. A Report of the Safe Drinking Water Committee. Advisory Center on Toxicology. Assembly of Life Sciences. National Research Council. National Academy of Sciences. Washington. D.C. 939 pp.
  • Norwood, D.L., J.D. Johnson, R.F. Christman, J.R. Hass, and M.J. Bobenrieth. 1980. Reactions of chlorine with selected aromatic models of aquatic humic material. Environ. Sci. Technol. 14:187-190.
  • O'Connor. J.T. 1971. Iron and manganese. Pp. 378-396 in Water Quality and Treatment. A Handbook of Public Water Supplies. Third Edition. Prepared by the American Water Works Association. McGraw-Hill Book Co., Inc. New York.
  • Olsen, H.L. 1974. Asbestos in potable-water supplies. J. Am. Water Works Assoc. 66:515-518.
  • Olsen, E., and W. Sybalski. 1949. Aerobic microbiological corrosion of water pipes. Parts I and II. Acta Chemica Scand. 3:1094.
  • Packham, R.F. 1971. a. The leaching of toxic stabilizers from unplasticized PVC water pipe. 1. A critical study of laboratory test procedures. Water Treat. Exam. 20(Pt.2):108-124.
  • Packham, R.F. 1971. b. The leaching of toxic stabilizers from unplasticized PVC water pipe. II. A survey of lead levels in uPVC distribution systems. Water Treat. Exam. 20(3):144-151.
  • Packham, R. 1971. c. The leaching of toxic stabilizers from unplasticized PVC water pipes. III. The measurement of extractable lead in uPVC pipes. Water Treat. Exam. 20(Pt. 3):152-161.
  • Radziui, J.V., S.J. Campbell, R.M. Brinkos, and E.J. Shervin. 1967. A liquid cementcalcite pipe-coating process. J. Am. Water Works Assoc. 59:1413-1426.
  • Rawls, R.L. 1975. PVC makers confident on food-contact uses. Chem. Eng. News 53(37):11-12.
  • Robeck, G.G. 1978. Health Effects of PAH's. Memorandum to J. Garner. Director. Health Effects Research Laboratory. U.S. Environmental Protection Agency. Washington. D.C.
  • Rook, J.J. 1974. Formation of haloforms during chlorination of natural waters. Soc. Water Treat. Exam. J. 23:234-243.
  • Rook, J.J. 1977. Chlorination reactions of fulvic acids in natural waters. Environ. Sci. Technol. 11:478-482.
  • Saxena, J. 1979. Technical Progress Report for Period February 10. 1979-October 10. 1979. Study of the Water Distribution System as a Potential Source of Mutagens/Carcinogens in Drinking Waters. EPA Grant No. R-806 413. U.S. Environmental Protection Agency. Washington. D.C.
  • Saxena, J., D.K. Basu, and D.J. Schwartz. 1978. Methods development and monitoring of polynuclear aromatic hydrocarbons in selected U.S. waters. Pp. 119-126 in J. Albarges, editor. . ed. Proceedings of the International Congress on Analytical Techniques in Environmental Chemistry. Barcelona. Spain. Nov. 27-30. 1978. Pergamon Press. New York.
  • Schock, M.R., and R.W. Buelow. 1980. The behavior of asbestos-cement pipe under various water quality conditions: A progress report: Part 2-Theoretical considerations. Unpublished.
  • Schwartz, D.J., J. Saxena, and F.C. Kopfler. 1979. Water distribution system. a new source of mutagens in drinking waters. Environ. Sci. Technol. 13:1138-1141.
  • Schweitzer, G.W. 1970. Zinc-glassy phosphate inhibitors in potable water, editor. . Pp. 1-4 in Control of Corrosion by Soft Waters. Presented by Education Committee. American Water Works Association. and U.S. Public Health Service. Washington. D.C.
  • Sievers, R.E., R.M. Barkly, G.A. Eiceman, R.H. Shapiro, H.F. Walton, K.J. Kolonko, and L.R. Field. 1977. Environmental trace analysis of organics in water by glass capillary column chromatography and ancillary techniques. Products of ozonolysis. J. Chromatogr. 142:745-754. [PubMed: 914947]
  • Sorrell, R.K., R.C. Dressman, and E.F. McFarren. 1977. High pressure liquid chromatography for the measurement of polynuclear aromatic hydrocarbons in water. Pp. 1-22 in Proceedings of the AWWA Water Quality Technology Conference. Kansas City. Mo.. Dec. 4-7. 1977. American Water Works Association. Denver. Colo.
  • Sorrell, R.K., R. Reding, and H.J. Brass. 1979. Analysis of polynuclear aromatic hydrocarbons in selected water supplies. Pp. 318-320 in Papers Presented at the 177th Meeting of the American Chemical Society. Division of Environmental Chemistry: Symposium on Analysis of Polar and High Molecular Weight Compounds in Water. held in Honolulu. Hawaii. April 1-6. 1979. American Chemical Society. Washington. D.C.
  • Sorrell, R.K., H.J. Brass, and R. Reding. 1980. A review of occurrences and treatment of polynuclear aromatic hydrocarbons. Environ. Int. 4:245-254.
  • Stevens, A.A., and J.M. Symons. 1976. Measurement of Trihalomethane and Precursor Concentration Changes Occurring During Water Treatment and Distribution. Appendix 4 to Interim Treatment Guide for the Control of Chloroform and Other Trihalomethanes. U.S. Environmental Protection Agency. Washington. D.C.
  • Stroud, T.F. 1980. President's Message. Ductile Iron Pipe News. Ductile Iron Pipe Research Association. Oak Brook. Ill. 20 pp.
  • Tarter, M.E. 1979. Data Analysis of Drinking Water Asbestos Fiber Size. Final Report. Environmental Protection Agency No. 600/1-79/020. National Technical Information Service. Springfield. Va. 71 pp.
  • Thomason, M., M. Shoults, W. Bertsch, and G. Holzer. 1978. Study of water treatment effects on organic volatiles in drinking water. J. Chromatogr. 158:437-447. [PubMed: 721940]
  • Tracy, E.L. 1950. Observations on water samples from cement-asbestos pipe systems in Vermont. J. N. Engl. Water Works Assoc. 64:164-178.
  • Tuovinen, O.H., K.S. Button, A. Vuorinen, L. Carlson, D.M. Mair, and L.A. Yut. 1980. Bacterial. chemical. and mineralogical characteristics of tubercles in distribution pipelines. J. Am. Water Works Assoc. 72:626-635.
  • Uhlig, H.H., editor. . ed. 1971. Corrosion and Corrosion Control: An Introduction to Corrosion Science and Engineering. Second Edition. John Wiley & Sons. New York. 419 pp.
  • U.S. Environmental Protection Agency. 1979. a. Interim Primary Drinking Water Regulations; Amendments. July 19. 1979. Fed. Reg. 44:42246-42259.
  • U.S. Environmental Protection Agency. 1979. b. National Secondary Drinking Water Regulations. July 19. 1979. Fed. Reg. 44:42195-42202.
  • U.S. Environmental Protection Agency. 1979. c. Quarterly Reports. Health Effects Research Laboratory. U.S. Environmental Protection Agency. Cincinnati. Ohio.
  • U.S. Environmental Protection Agency. 1979. d. Statements of Basis and Purpose for Proposed Amendments to the National Interim Primary Drinking Water Regulations and Proposed Special Monitoring Requirements. Office of Drinking Water. Criteria and Standards Division. Municipal Environmental Research Laboratory. U.S. Environmental Protection Agency. Cincinnati. Ohio.
  • U.S. Environmental Protection Agency. 1980. Interim primary drinking water regulations: Amendments. August 27. 1980. Fed. Reg. 45:57332-57357.
  • Wagner, J., Jr. 1970. Instrumentation for evaluating corrosion and corrosion inhibitors. Pp. 1-4 in Control of Corrosion by Soft Waters. Presented by Education Committee. American Water Works Association. and U.S. Public Health Service. Washington. D.C.
  • Wajon, J.E., and J.C. Morris. 1980. Bromamination chemistry: Rates of formation of NH2Br and some N-bromamino acids. Pp. 171-181 in R.L. Jolley, editor; , W.A. Brungs, editor; , R.B. Cumming, editor; , and V.A. Jacobs, editor. . eds. Water Chlorination: Environmental Impact and Health Effects. Vol. 3. Ann Arbor Science Publishers. Inc. Ann Arbor. Mich.
  • Wallcave, L., H. Garcia, R. Feldman, W. Lijinsky, and P. Shubik, 1971. Skin tumorigenesis in mice by petroleum asphalts and coal-tar pitches of known polynuclear aromatic hydrocarbon content. Toxicol. Appl. Pharmacol. 18:41-52. [PubMed: 5542832]
  • Wallington, R.H. 1971. Protective coatings for pipes and tanks. Water Treat. Exam. 20:209-219.
  • Webster, 1. 1974. The ingestion of asbestos fibers. Environ. Health Perspect. 9:199-201. [PMC free article: PMC1475405] [PubMed: 4470936]
  • Weber, W.J., Jr. and W. Stumm. 1965. Formation of silicate-iron (III) complex in aqueous solution. J. Inorg. Nuclear Chem. 27:237-239.
  • Weil, I.W., and J.C. Morris. 1974. Dynamics of breakpoint chlorination. Pp. 297-332 in A.J. Rubin, editor. . ed. Chemistry of Water Supply. Treatment. and Distribution. Ann Arbor Science Publishers. Inc. Ann Arbor. Mich.
  • Weil, I., and J.C. Morris. 1949. Equilibrium studies on N-chloro compounds. II. The base strength of N-chlorodialkylamines and of monochloramine. J. Am. Chem. Soc. 71:3123-3126.
  • World Health Organization. 1971. International Standards for Drinking Water. Third Edition. World Health Organization. Geneva. 70 pp.
  • Zoeteman, B.C., and B.J. Haring. 1976. The nature and impact of deterioration of the quality of drinking water after treatment and prior to consumption. Report prepared for the Commission of the European Communities. National Institute for Water Supply. Voorburg. The Netherlands. 123 pp.
  • Zoldak, J.J. 1978. Analysis of Drinking Water for Trace Level Quantities of Organic Pollutants. Thesis. Miami University. Oxford. Ohio.
Copyright © National Academy of Sciences.
Bookshelf ID: NBK216607

Views

Related information

  • PMC
    PubMed Central citations
  • PubMed
    Links to PubMed

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...