Document Ne615opkM74DLLmRrzEx7NK4R

PLAINTIFF'S EXHIBIT A/C Pipe Producers Association Public Affairs Committee International Affairs Committee .4f~. ------- . frcC/J* Welch, Vice President Internal Correspondence DATE February 10,1982 subject U.S. Environmental Protection Agency (EPA) - Research on Corrosion of A/C Water Pipe ACTION REQUIRED: Review for information Enclosed is an article, "The Behavior of Asbestos Cement Pipe Under Various Water Quality Conditions: Part 2, Theoretical Considerations" by Schock and Buelow. This report summarizes the most recent results of EPA's research on the corrosion of A/C pipe. Beyond the theoretical considerations of complexation and solubility equilibria, the most important finding is that the Aggressiveness Index (AI) as well as other indices of water corrosivity are ... not theoretically sound from a chemical standpoint to predict fiber release and degradation of the interior surface in systems that are undersaturated with respect to calcite. Fortunately, it (AI) has generally proved to predict falsely pipe deterioration more often than to predict falsely pipe stability in field situations. The report goes on to state that "natural inhibitory factors" such as the presence of iron, manganese, silica and organic matter may bind asbestos fibers even though the pipe surface is softened from corrosive water attack. The finding with the most potential negative impact follows: The results of DWRD experiments conducted thus far and field observations indicate that addition of the corrosion-control compounds tested thus far cannot be relied upon to provide correction of A/C pipe deterioration. Advocates of the regulation of A/C pipe now may say that its use should be restricted because once deterioration ("surficial softening from aggressive water") begins, it is irreversible and consumer exposure to asbestos in drinking water will be continuous. In many respects, however, this research supports the conservatism of the Aggressive Index or Langelier Saturation Index as suitable criteria for predicting the long term structural integrity of A/C water pipe. Moreover, it bolsters the industry's long held opinion that tho^nly accurate means to determine whether an A/C water system is releasing asbestos fibers is to ensure that manufacturing or construction asbestos residues are thoroughly flushed and accurate asbestos in water analyses taken. Corrosion indices along will not suffice. If you have any questions, please do not hesitate to call. JFW/ajb . i\ i CAPCO JEN 0004-804. Enclosure cc: A. Kahn, Esq. N. Rahn, Esq. Asbestos International Association Asbestos Institute Special Counsel (Cal., Ariz., Texas, Conn.) Brian Com mins, Ph.D. Kirkland & Ellis copies to: Public Affairs Committee International Affairs Committee H. Olson B. Cook R. LeGault W. Perrell J. Woods J. Baker T. Dougherty D. Stinson W. McCallie B. Collier R. Dorner E. van der Rest B. Giboin N. Officer R. Hobbs A. Saoulis R. Jalan V. Pattabhi II. Hudson C. Barton S. Al-Tarkait H01PI20902 Chrono i CAPCO JEN 0004805  (JRESEARCH AND TECHNOLOGY a %At-*** , The behavior of asbestos-cement pipe under various water quality conditions: Part 2, theoretical considerations Michael R. Schock and Ralph W. Buelow A water chemical model Involving the complexation and solubility equilibria of calcium, magnesium, zinc. Iron, manganese, silica, chloride, orthophosphate, and carbonate has been developed and applied to laboratory pipe-loop-coupon and field data on asbestos-cement pipe. Calculation of the saturation states of the relevant solids enables graphic guides to be prepared for the prediction of dosage levels for in-situ pipe coating by zinc compounds as well as protection afforded by coatings formed by natural' Inhibitory constituents of the water. Because of concern about the possible problem of asbestos fibers being released from the walls of asbestos-cement (A-C) pipe, the AWWA Research Foundation reviewed the problem and in September 1974' suggested several research needs in this area. As a response ,td.this call for research, several projects were designed by the US Environmental Protection Agency, Drinking Water Research Divi sion (DWRD). to determine whether A-C pipe would be. attacked and asbestos fibers would be released from the pipe under various conditions of water quality. In general, this research has been divided into four phases: first, a field evaluation of 10 public water supply systems that used A-C pipe: second, an A-C pipe-loop system operuted under controlled condi tions; third, several tests of A-C "pipe coupons"; and fourth, two field research projects that are attempting to rehabili tate deteriorated pipe in place. Part 1 of the project report* contained the experi- mental results of the first three of these four phases, Additional research by DWRD into various methods of analy tical determination, primarily by micros copy, has been reported.* . This paper relates to the A-C pipe cou pon tests reported in part 1, particularly the studies of the protection of the A-C pipe against attack from aggressive waters. Table 1 shows the water quality of most of the DWRD recirculation experiments (as of July 1980) concerned with A-C pipe protection. Table 2 sum marizes the qualitative results of the 10 DWRD laboratory experiments referred later in this paper. This paper examines the solution interactions that were seen to govern the effectiveness of the zinc compounds as corrosion inhibitors. The model devel oped is then extended to show how sim ilar considerations of the aqueous equilibria of other metals and nonmetals may be used to predict more effectively. the true aggressiveness of a water towaTd ; A-C pipe in actual distribution systems. Critique of the "Aggressiveness Index" ; Classically, the tendency of a water to deteriorate the structure of asbestos-cement pipe has been described by the "Aggressiveness Index" (AI).4-5 given as . . i ' A1 = pH + log (AH) I where pH = --log aH' (--log of 'the , hydrogen ion activity); A = total alka- - linity in mg/L as CaC03;and H=calcium concentration in mg/L as CaC03. - Waters possessing an Al equal to or ' exceeding 12.0 are considered to be "non- aggressive"; those with an Al S 10.0 are 4 said to be "highly aggressive"; and those with an AI between 10 and 12 are "moderately^ aggressive." The AI is 4 derived5 from a simplified form of the Langelier Index of calcite saturation5-7 with factors introduced to compensate for the temperature dependency of the solubility product constant (K,[>) of cal cite and for the ionic strength5-5 of the solution. -The factors for temperature and ionic strength compensation utilized in the derivation of the 12.0 value of5 the Al correspond to a temperature of approxi- -mately 14C and an ionic strength of 636 RESEARCH AND TECHNOLOGY 0003-150X/81/120636-16S02.00 >1991 American Water Workt Association JOURNAL AWWA i CAPCO JEN 0004-806 r *\ .TABLE 1 WoIor qualities oj DWHD recirculation experiments Experiment Number Total Alkalinity mjt'L P os CoCOj Ca mg/t. 2n nitf/L PO* K Mjf Si02 ci TDS SPC m#/l. tnyfl* nyj/L mg/l mp'L mull my/L (ix Cl. m#tf. pS'tm Al} Treatment Chemical 1 2 9 4 5 `6 7 8 8 lOtt n 12 13 14 id 19 8.2 20*27 .. 5-13 6.2 20*25 4*8 0.3*0.$ 0.4*1 7.0 21*25 10-22 7JO 20*29 9*17 o.3*o.r 0.4*1 8.2 19*22 5-8 > 0.3-0.7 7.5 116-130 133*155 0.03 60** 7.9 120*131 129*160 0.04 60** 9.0 96*42 21*30 <0.02 6.2 42*50 9*16 0.02 0.2 7.0 25*27 13*16 0.1 8.2 18*19 7.5 18*21 0.6*4 9-14 0.3-0.4 0.3-0.4 0.03 0.03 10 <0.1 8*9 0.2 8.2 19*21 10*15 0.1-0.2 0.03 9*10 0.2 6.2 2*4 32-43 0.3-0.4 0.06 8.2 JC*19 7-10 04*0.5 <0.2 13 7.2 18*21 . 9*25 0.1 0.1 10 0.3 0.1*2 1-4 94** 1*4 94** 0.1*2 15*16 30 1.5 14-17 0.1 3 <10 0.6 2-3 <10 0.6 0.2 <10 1.3 1*6 27 0.8 <0.4 <10 1.0 <0.2 13 0.1-0.4 O.1-0.4 0.1-0.4 0.1-0.4 0.1-0.4 0*0.6 0-0.7 0-0.4 0.1-0.5 0-0.5 0.2-0.6 0.2-0.S 0.2-0.5 04-0.0 0.4-0.7 0.1-0.3 30-60 32-61 45*7) 49*72 36-41 340*360 330-360 80-IOt 146 108 24 41 44 63 47-63 43 231 78 63 104 100 78 10.2-10,6 10.1-10.5 9.3-0.7 94-9.7 10.2-10.4 11.7*11.6 12.1*12.2 11.9*12.1 10.6*11.1 0.5-9.6 9.2*10.1 9.7-10.0 10.5-10.7 10.0-10 4 10.2*10.5 9.4-9.9 none Zinc orthophosphate} none zinc orthophosphate} zinc chloride none CaCJj/NxHCCb CaCfe/N'aHCO* sodium tneUsillcate sodium mffasllicat* zinc chloride zinc chloride zinc chloride zinc chloride zinc sulfate zinc chloride and iron (HI) chloride Analyse* repmenl concentrations ol or ne.tr the end offrfch experiment unless n ran#* is given. All sulfate values were less than the 15*mg-'l detection limit. Based on the water blending comlitinns. the probable range was S-10mg'L. iron was less than the 0.1 mg/L calibration limit, except fur experiment 19 in which approximately 0.1 mg.'L of iron was added. tFree chlorine residual measured one week nfter each dose JAl Aggressiveness Index range; see text for definition JVirchein 931. a product of Virginia Chemical Corp. Portsmouth. V. 'Sodium and chloride concentrations are estimated from reagnnt additions. There was a large analytical scatter on the calcium and total alkalinity results. J'The same coupon was previously run for almost one month at S1O2 concentration of 10*30 mg'b and zinc approximately equal to 0.2 mg'L at a similar pH. total alkalinity, and calcium level. VJ -iroximately 0.01. The solubility prost constant for calcite was apparently ,;.kcn from the work of Larson and Buswell.7 The intent of the AI, when origi nally conceived, was to indicate environ ments that would tend to damage the structural integrity of the pipe. There are several major avenues by which the Al is open to criticism as an indicator or corrosivity and fiber release, the manner in which it has most fre quently been used. The foremost.problem is that the derivation of the Al isbased on the Langelier Index of calcium-carbonate saturation with some approximate ad justments for temperature and ionic strength.* Insofar as the A-C pipe would indeed be protected by a coating ofCaC03 if active CaC03 precipitation were occur ring, the AI would then givea fairly accu rate prediction of nonaggressiveness. However, in the case of .calcium-car bonate undersnturation, the aggressive ness of the water toward the pipe will be dictated by other chemical equilibria. Calcium carbonate is, at maximum, a trivial component of the matrix of A-C pipe;10 therefore, there is no a priori rea son to expect the AI as traditionally con ceived to give an accurate prediction of the pipe's interior surface condition in waters undersaturated with respect to calcite. A second criticism relates to the solu bility product constant of calcite used in the derivation of the index. Work by Langmuir," Jacobson and Langmuir,12 and Christ et al15 not only corrects for ion-pairing effects not considered pre viously7 during solubility experiments but also indicates that the quantitative temperature dependency of Kf0 found by Larson and Buswell7 is erroneous, particularly in the range of 0 to 15C, a common range for natural and drinking waters. The higher value of K.n given by Larson and Buswell7 probably results primarily from ignoring the ion pairs; to the extent that few environmental workers have applied such corrections to field and laboratory data, the error in Kk0 has not been obvious. Also, a metastable CaC03 phase of considerably higher solu bility than calcite is probably the initially precipitated form of CaC03 in soil root zones" and in water treatment.11 Precipi tation of aragonite, which is more soluble than calcite, in lake marls and marine environments has been commonly ob served." The occurrence of mixed pre cipitates of calcite and aragonite, as well as the stabilization of aragonite relative to calcite by such ions as Mg2*, Sr2*, trace levels of Zn2*, and some organic mate rials, has been documented.16*17 A similar phenomenon probably takes place in drinking water. Therefore, those study- ing calcium-carbonate precipitation may tend to regard the Larson and Buswell constant,7 which predicts higher solubil ity for calcite, as being closer to observa tion and therefore more correct, rather than noting its thermodynamic questionability and seek further explanation. The simple expedient of making temperature and ionic-strength adjust ments by averaging a range of values of temperature and total dissolved solids may result in more correct approxima tions in some cases, but it is an inherent source of error and uncertainty in other cases. When attempts are made to predict the amount of pipe softening or calcium leaching from the pipe interior by using the Al of the water, additional error is introduced by implying a correlation be tween the magnitude of calcite undersaturation and the rate of dissolution of the predominantly calcium-silicate pipe. A correlation between faster calcium leaching and greater calcite undersatura tion is, more probably, a coincidence resulling from some kinetic effect of cal cium concentrations in the waters (and probably the pH) with regard to calciumsilicate or calcium-hydroxide saturation. Later in this paper the Al will be shown to be further deficient in reliably predict ing the A-C pipe condition, because it DECEMBER 1981 M.R. SCHOCK & R.W. BUEIOW 637 CAPCO JEN 0004-807 TABLE 2 Summary of qualitative result.': of DIVHD A/C pipe treatment experiments Experiment Number 1 2 2 4 S 8 7 a 9 10 11 12 13 14 16 19 Treatment pti Process < 8.2 8.2 7.0 7J0 8.2 7.$ 7J> 9.0 8.2 7.0 8.2 7.5 . 8.2^ 8.2 8.2 7J2 none line (orthophusphale) 0.3-0.S tng'L none sine (orthophosphate) 0.3-0.$ tng/L sine (chloride) 0.3-0.7 mg/L . none CaCO? saturation CaCOs saturation sodium metasilicate sodium metnsilicale tine (chloride) 0.4 mg'L: low calcium sine (chloride) 0.4 mg'L line (chloridej 0.1-0.2 mg/L tine (chloride) 0.4 mg'L; low alkalinity zinc {sulfate) 0.$ mg/L iron (HI) (chloride) 0.1 mg/L: zinc (chloride) 0.1 mg'L Inner Wall of Coupon Softened? yes no yes slightly no yes very slightly - slightly no no no slightly yes very slightly no slightly TABLE 3 Aqueotis species considered in saturation index and solubility diogrom preparation H* OH* HjCOj* HCO3" C03*- H3PO4* H2PO4* HPO-4 P04*' Ca* CallCOj* CnCXV CaHPOa* CaP04* CaHjP04* CaOH' Mg-* MgHCOj' MrOH* MgCOj* MgHFOa* NV N11CO3* ZnCOj* ZnCl* Species ZnCk* ZnCb* ZnCU:* ZnCiOH* ZnHP04* ZnH2P04* ZnOH* Zn(OH)2# ZnlOHb* Zn(OH)4* Zn20H'* Fe* FeOH1* Fe(0H)2* Fe|OH|j* Fe(0H)4* FegfOHJz*' Fe3(OH)4FeSOs* FeCls* FeClj* FeCb* FeHP04* TABLE 4 Solid species considered in Saturation Index and solubility diagram preparation Solid (Chemical Formula) CaCOj Ca(OH): CaHPO-i -2H20 Ca8H2(P04)-5H20 Caj(P04)0H * MgCOa StglOHU Mg3(P04)2 ZnCOj Zn5|0H)6lC03l2 Zn(OH)2 Zn3(P04l2*4H2O Fc(OH}3 FePOa * II2O Common Name calcite portlandite brushlte octacalcium phosphate hydroxyapatite magnesitr brucite magnesium orthophosphate smithsonite hydrozincite zinc hydroxide o-hopeite amorphous iron (It!) hydroxidr strengite The corrosion of A-C pipe is governed almost completely by solubility consider ations. Thus, the dissolution and coating processes of the pipe may be described and predicted by bulk-solution chemical parameters and straightforward solid solubility reactions. The effects of pipe dissolution on the distribution system water quality, in addition to fiber release, are several. The "free-lime" component of the pipe (essentially equivalent to the solid portlandite) can dissolve as represented by the reaction Ca(OH)2(s) -=Ca'*+20H' (1) for which K,o is lO-520. This would increase the pH, titration alkalinity, and the calcium content of the water (and, therefore, the Al) during its passage through the pipe. The pH increase would enhance the ability of the water to absorb and hydrate any C02 gas present. Subse quently, the reaction of any absorbed CO* with the hydroxyl ions present would lead to the formation of bicarbonate ions, a step that would favor increased disso lution of.Ca(OH)2(s) by the mass-action effect. In the present commercial auto claved type II A-C pipe, the free-lime phase is <1.0 percent by weight.'1 Three of the predominant crystalline phases of the matrix of A-C pipe are tricalcium silicate (nominally Ca3SiOj), dicalcium silicate (nominally Ca2Si04), and tricalcium aluminate (CajAI206).10-21 Possible dissolution reactions may be represented by CajSiOs(s) + SHjO = 3Ca5- + H1SiO< + GOH' (2) Ca-SiOds) + 4H20 = 2Ca=- + H^SiO*0 + 40H" (3) Ca,Al;06(s) + 6H20 = 3Ca:* + 2A1>- + 120H" (4) fails to take into account protective chem matrix of A-C pipe is a very complicated ical reactions in drinking water aside combination of compounds and phases, from calcite deposition. some of which are poorly identified or are Mechanism of A-C pipe dissolution of indefinite composition. More than 100 compounds and phases important to the Attack on A-C pipe by drinking water chemistry of portland and related ce manifests itself as deterioration of the ments have been described and identi interfacial fraction of the interior pipe fied,21 and because of solid solution surface, which can lead'to the release of^possibilities, probably many more exist. asbestos fibers, about which there is The state of knowledge of the thermody considerable concern for possible human namic solubilities in water of the individ health effects."-20 Field and laboratory ual predominant compounds of the observations by DWRD have shown that cement lags far behind that of minerals sometimes adherent coatings form on the and related man-made compounds impor pipe that will prevent fiber release, but tant to drinking water chemistry. Until xvhich may or may not totally prevent more research is done, only some qualita some surficial softening.The cement tive generalizations can be made. 638 RESEARCH AND TECHNOLOGY 1 ; Estimates of the solubility constants for reactions (2) and (3) were obtained by using an internally consistent set of Gibbs free energy of formation data22 at 25Cand, at 103 kPa(3 atm) pressure, arc 10-* and 10-'-*. respectively. Free-energy data have thus far not been located for Ca3Al206. Since there is s'till considerable uncertainty as to the exact crystal struc ture of this solid,21 lack of thermody namic characterization would not be surprising. Because solids are often highly substi tuted in portland cement21, their true ther modynamic activities may be somewhat different from unity.22-21 Therfore, either the free energies tabulated may not exactly correspond to conditions in cement, or the activity of the solid may need to be included in the solubility con stant expression if it is used to determine JOURNAL AWVVA CAPCO JEN 0004-808 ......in pure drink inconclusive as to whether or not active ing water having a pH less than approxi precipitation was actually taking place ...c sunns tricalcium silicate mately 9 to 10. Importantly, the alkalinity (evidenced by no clear decrease or level ml dicalcium silicate are highly soluble increases observed in field studies* do not ing off of calcium or total carbonate con mler general drinking water conditions, necessarily indicate input of dissolved centrations). Increased stabilization of inactions of other solid phases can also carbonate from pipe materials. Carefully certain calcium-silicate phases of the asily mobilize sodium and potassium. obtained, analyzed, and preserved sam pipe by the lower pH and higher calcium Dissolution by reactions of the type of ples for potentiometric total alkalinity, present in experiment 7 (as opposed to (3), and (4),even ifthe estimated solu- pH, temperature, and major constituents experiment 8) is an alternative possibil lility constants are not very accurate, can be used to calculate changes in carbo ity-. The condition, of the .pipe coupons vould produce three major effects. First, nate concentration.*2-*1'-*7 does support addition of calcium andcar- fie levels'of calcium, aluminum, and silcon species (as well as any substitu- Protection by CaC03 saturation bonaie toljftain'calciui^^ ratjorTat VeliTTvety fow jHrvaluis~aTa inna! elements) will increase with time Three experiments reported in part yiaBSSEoIItiSon^rrnS^ lpon. standing, or with distance of pas- 1* (experiments 0, 7, and 8) evaluated Though protection by calcium-carbon uge through a pipeline, unless threshold the effectiveness of maintaining solu ate saturation j^qwell-esTiblTsIie -onditions are met for saturation and pre- tion composition near to slightly above corrosionri'nTu5inorWecRmquerthe~e>> ipitation of another less soluble phase, CaCO. (calcite) saturation by adjustment penseoTtEe^realmyircjimicalrTpr r until the available supply of leachable of calcium, total carbonate concentration waterMipplTesonovThardness and~afRa^ pecies is depleted for a given depth of enctration. (C|), total alkalinity (Aik!), and pH simul taneously. As would.be expected, the *U n--ity.ii,"asi- w*e* lTUTlifimi'des\--ir_a>B--i*l*iJivif*T+*o^ irthen Second, the pH of the system would pipe integrity was maintained with only end to increase because of several disso- a little softening when it was at the cal ution reactions and the two dissocia- cite saturation level26 and with virtually ions of silicic acid.*1 The overall solution no deterioration when it was slightly iH may also be increased slightly by dis above. Protection by metal precipitation olution of the asbestos fibers them- The small amount of pipe softening Just as calcium carbonate may precipi elves. For example, the overall process observed in experiment 8 is not particu tate and adhere to a pipe surface to form a f chrysotile-fiber dissolution as given larly surprising and may be explained by protective coating, so too could any other y the equation several factors. At least two of the major reasonably dense solid. Unfortunately Mg3Sii05(0H)4(s) + SHjO 3Mg'- + 2Si(OH)4 + 60H" (5) phases of portland cement, the primary matrix material of the pipe, are thermo dynamically highly unstable in most (from this standpoint), few combinations of cations and anions- are present in drinking water in sufficient quantity to drinking waters, as was shown earlier. provide enough mass for precipitation >uld tend to increase the pH by virtue of Another factor is the role that pH plays in and effective pipe coating. Four elements input of hydroxyl ion and soluble sil- the attainment of CaCOj saturation. jhat can form useful coalings areTron, inio the water, because chrysotile has A series of calculations, or the use of a zmcTn^ganes^a^d"snicgn5~^ . Kj, of--51.8,*6 but the total mass of graphic method such as a Caldwell-Lavv- THTIjrltpurpose^oTThis section is to . involved is quite small. rence diagram,*7*2* will show that a use computer calculations of chemical me hydrolysis of At** ion introduced greater mass of CaCOj can be deposited equilibria in hypothetical model systems ng dissolution of tricalcium alumi- at a low pH with high calcium and carbo to develop a predictive model of aggres may take place, but field observa- nate concentrations than at a high pH sive and nonaggressive water conditions. in distribution systems indicate with low calcium and carbonate concen A second purpose is to show by correla the pH-increasing reactions are dom- trations. tion of the computer model with field t.* Also, large pH gradients between As was mentioned previously, the data and laboratory experiments that the ripe surface and the bulk solution first-formed CaCOj precipitate (even if it true aggressiveness of drinking waters d not normally be expected. The is metastable) most probably is described with respect t o A-C pipe can be predicted r in a distribution main would have by a more soluble than calcite. It is. more accurately than by using the AI nal periods of suspended flow, such desirable to take this phenomenon into alone. Situations that are in conflict with a household system during over account, as well as complexation and ion the AnnosTprobably are not anomalous standing, except at the dead ends, pairing, when a conditioning scheme pheriomenaTiqtT^snlnToriTsKortcomTngs^ the buffer capacities of most, even based on calcium-carbonate saturation is. uTtHe^cHenriMlTjasis ot tKeTintj^Jlfim' isive, water should be sufficient to devised. TSn3eF|R^5n^m|nreTive^^eiiiJs3oQr ntiaily moderate a pH rise, espe- A more rigorous compulation of the FmaflyTthis "section will show how the Arhen such capacities are combined' " saturation conditions previously report model may be used to devise protective or instant water flow. Additionally, a ed* for experiments 6.7, and 8, was per corrective treatment conditions. :rge local rise in pH would often formed by means of the aqueous equili Calculation methodology. The model sys hiCOj deposition and subsequent brium calculation computer program tems considered were chosen mainly to ion, a situation that (ends to con- WATSPEC2,** an enhanced version of approximate laboratory experiments 'th the observation of continual WATSPEC.*0 All systems were calcu (completed or planned) or to illustrate ching in field and laboratory tests lated to be undersaturated with the esti some possible field situations. Obvious :d in part 1* and later in this mated freshly precipitated phase" but ly, not all combinations of important var were at or abovrSaturation with calcite. iables could be covered; however, the the alkalinity of the system may The calculated SI values suggest that general solution charcteristics covered w an increase. Considering (hat experiment G should not have provided by the example systems should enable y is a charge-neutralizing capac- protection by calcium-carbonate deposi them to be applied directly to most cases only newly-formed directly- tion, and the coupon was seen to be con not specifically shown. ting entities would be OH-, siderably softened.* Experiments 7 and 8 The models were calculated by utiliz and SiO-(OH)ri'. However, a were near the thermodynamic precipita ing the aqueous equilibrium modeling ase favors the production of tion threshold for aragonite and a fresh program KEDEQL.EPAK*'-** written in d COy-. normally the dominant precipitate. The analytical data are the FORTRAN* IV language and run on an 1 R 1981 M.R. SCHOCK & R.W. BUELOW 639 1 CAPCO JEN r TABLE 5 Aqueous reactions considered in the construction of the Saturation Index and precipitation diagrams Reaction lugKt Source Hydrogen H* + OH* * HsO 2H* + COj" = HjCOj' H' + CO:-" = HCOj> *11* + COj" COj(r) + H'Q ir + pos'* 11PO42H* + PO." = H:PO.' 14.00 16.7 10.3 16.14 12.4 19.6 3H* + PO4" * HjP04* 21.7 Calcium Cn- + H2O * CaOH* + H' -12.6 Ca-- + CO3- = CsCOa* 3.2 Ca- + COa- + H* CaHC03* 11.3 Ca- + H* + PO4- = CaHPOr Ca- + 2H* + PO4- CaH'P04* Ca=- + PO." *= C.PO.' 15.1 20.3 6.5 Magnesium Mg- + H2O MgOH* + H* -11.8 Mg- + COj- = MgCOj* Mr- + COa- + H* = MgHCOi* Mg- + II* + PO4- * MgHP04 3.0 11.4 15.3 Sodium NV + CO3- * NaCOa' 1.2 Zinc 2Zn- 4 H2O ZngOH- + H* Zn- x H.*0 ZnOH* + H* -9.0 -9.0 7.n- + 2H2O Zn(OH)2* + 2H* -16.9 Zn:* + 311:0 = Zn|OH)j' + 311' -28.4 . Zn- + 4H2O Zn(0H)4- + 4H* -41.2 Zn-- + IF - PO." = ZnHPOj* 15.C Zn1- + 21!' + PO-i1' s ZnH'POj' 21.2 Zn- + CO3- - ZnCOj* 4.8* Zn- + HjO x Ci' *= ZnClOH* + H -7.5 Zn- + C!` = Znd* . .. Zn- + 2C1** ZnCfc* 0.4 0.6 7m- 4 3C1* *= ZnClj* 0.5 Zn- + 4CP ZnCIs- 0.2 Iron Fe- 4 H2O TeOH- 4 H* Ft'- + 2H:0 = FcI0H)2' + 2H' -2.2 -5.7 Fe- 4 3HjO Fe(0H)3* 4 3H* -13.C Fc>- + 4H:0 = Fe(0Hl4* + 4H' -21.6 2Fe- 4 2)1*0 Fe2(OH)2- 4 2H* - 3.0 3FE'* + 4H:0 = Fc:(0H|4" + 4H' - 6.3 Fe- 4 H' x PO4- FeHPOs* 17.6 Fe'-Ci* FeO- 1.4 Ft - 4 2d* FeClg* 2.1 Fe- 4 3C1`= Feda* 1.1 1 2 3 2 4 4 4 5 5 S 5 S 5 5 5 5 5 5 1 1 1 1 1 4 4 6 7 4 4 4 4 1 1 1 1 1 1 3 5 5 5 1AH data are fur 25C. $Thc published data of BUinski et al" have been reinterpreted by Schock et nl` in terms of both ZnCOj* and Zn(COjJa-. w ith lop K=5.2 for ZnCOa* and with lop K - 7.5 for the reaction: Zn-- x 2CO3-- -- Zn(COa)i**. These constants were incorporated into the model after the calculations reported in this paper were performed. Sources: 1. DAKS, C.F. |R. i MESSIER. R.E._tlic ilydrolyUt of Cation*. Wiley-lnttrscience, New Ynrk (1976). 2. HARKED. H.S. A DAVIES. R. JR. /our. ACS. 65:2030 (1943). 3. HARKED. H.S. A SCHOLES. S.R. )R. /our. ACS. 63:1706 (1941). 4. SMITH. R.M. A MARTELL. A.E. Critical Stability Constants: Vof. 4. Inorganic Complexes. Plenum Press. New Ynrk (1976). 5. PLUMMER. I.N. ET At- WATEQF--A FORTRAN IV Version of WATEQ. A Computer Propram for CalrulatinpCheinic.it Equilibrium of Natural Water. U5GSWRI 76*13 (Dec. 1976/. r.. B1UNSK!. H. IT AL. Anof. Chim. Ado. 64:157 (1976). 7. WACJMAN. D.D.ETAL.NDSTech. Note 270-3 (Jan. 1966). Figure 1. Saturation Index diagram lor model Figure 2. Saturation Index diagram tor model system Ml system M8 TABLE 6 Solid reociions considered in Saturation Index diagram preparation* Reaction log K Source Calcium Ca- 4 COa- = CaCOa(s) Ca- 4 2H2O Ca(OH)2(s) + 2H* Ca- 4 H* x PO4- 4 2H*0 CaHPO* 2HjO(s) 8Cft- 4 6PO4- x 2H* 4 3H2O = Ca4H(P04)3 `SHaO 50- 4 3P04- 4 H2O = Os(P04)30H(5) x H* Magnesium Ms" + CO:'- = MsCO:(s) Mg- 4 2H2O = Mg(OH)2(s) 4 2H* 3Mg- 4 2P04` x 8H* Mg3(P04)2 *6H20(s) Zinc Zn- 4 CO3- = ZnCOj(s) 5Zn- 4 2CO3- x 6H2O - Zns|0H)6(C03)2(s) 4 6H* Zn- 4 2H'0 *= Zn(OH)2ls) 4 211* 37.n" + 2POi>- + 4H:0 = innll'Oj):-4H201S) Iron 111 Fe- 4 3H20 Ff(OHj3(s) 4 3ll* Fe** 4 PO4- 4 2HjO * FeP04 *2H20(s) 6.5 -22.8 16.9 46.9 40.6 7.S -10.6 .25.2 10.6 -9.7 -12.5 35.3 -4.9 26.4 1 2 3 3 4 3 7 2 6 6 7 $ 5 2 All data are foi; 25C. Sources: 1. JACOBSON*. R.L 6 LANGMUIR. D. CVochim. Cosmoch/m. Actc. 38:301 (1974). 2. SMITH. R.M. A MARTEM.. A.E. Critical Stability Con.stonts: Vol. 4. tnorconie Complexes. Plenum Press (1976). 3. LINDSAY. W.L. 6 MORENO, E C. Soil SW. Soc. Amer. fW.. 24:177 (1960). 4. MORENO. E C. ET At., /our. Res. NH5.72A:6:773 (Nov.-Dee. 1968). ' 5. PLUMMER. L. N. ET AL. USCS WRI 76*13. (Dee. 1978).- C. SCHINDLER. P. F.T AL. Ilelv. Chim. Aclo. 52:2327 (iy69). 7. BAES. C.F. JR. & MESMF.R, R.E. 1 he Hydrolysis of Colions Wiley-lnlerseience. New York (1976). 8. KRIAGU. J.O. Gcochim. Cosmorhim. Acla. 37:2357 (1973). .* TABLE 7 Model systems for zinc. iron (IJIJ. end calcium coating of A-C pipe ot 25C* Calcium Tola) Carbonate mg.'L mg/L System os Co os CaC03 Ml M2f M3 M4 M5t M6 M7t M8 MSP M9 M10 M1QP Mil M12P M13 M13P MI4 M15 0.4 0.4 0.4 0.4 0.4 0.4 0.4 1.0 1.0 4.0 4X> 4.0 40 4 JO 4.0 4J0 4.0 8JO 4Si 1JD 1.0 1.0 1.0 1.0 1.0 1.0 2.5 2.5 10.0 10.0 10.0 10.0 10.0 10.0 10 0 10.0 20.0 10.0 mg/L as COj 0.6 0.6 1.S 1.5 6.0 12 60 24 24 24 * 12 12 12 24 24 60 CO 2.4 2.4 mg/L MS Na os CaCOj mg.*!. mg-'/. 1.0 1.0 2.5 2.5 10.0 20.0 100.0 40.0 40.0 40.0 20.0 20.0 2041 40.0 40.0 100.0 100.0 4.0 4.0 0.1 5 0.1 5 0.1 10 0 10 0 10 1.0 12 0.1 CO 0.24 20 0.24 20 1.0 20 1.0 12 1.0 12 1.0 12 LO 20 1.0 20 1.0 30 1.0 30 2.0 20 1.0 20 d mg'L 4*9 4*9 11-17 11-17 8*17 9-23 42-92 11-33 11*33 18*40 15-29 15-29 15*29 5-26 5*29 4-56 4-56 45*52 3H-42 Zn Fe P04 mg/L mg/I. irg/L Figurt(s) 0.5 0.1 0.5 0 0.1 0.5 0.1 0.5 0.5 0.15 0.5 0.5 03 0.5 0.5 0.5 0.5 . 05 0.5 0 0 0 0.1 0 0 0 0 0 0 0 0 0.1 0 0 0 0 0 0 01 0 0 0 0 0 0 0 2.3 05 0 4.5 0 6.7 0.5 0 8.9 0 10.11 0.5 12.13 0 14.15 0.5 16.17 0 18 0 19.20 The pH ran#* of 5 to 10 was considered. tNo saturation of any solid in the pH range of S tn 10 { i | i ; { 1 * : \ ; } \ ' t ; : * *. ; * ; : V_______________________________________ 640 RESEARCH AND TECHNOLOGY . ,.----------------------------------- :: JOURNAL AVVWA CAPCO JEN 0004810 IBM 370/168 computer. Tables 3 and 4 level, providing that the thermodynamic summarize the aqueous and solid species equilibrium constants are not in error, considered, and the equilibrium con and that the system can be characterized stants employed in the calculations are by equilibrium thermodynamics. Figures given in Tables 5 and 0. The temperature 1-20 present SI and associated precipita was assumed to be 25C, and the compu tion diagrams for eleven of the model sys ter program calculated the solution ionic tems summarized in Table 7.* The model strength based on the equilibrium con systems are correlated with the DWRD centration of charged species. Activity experimental studies (Table 8). coefficients of uncharged molecules and Obviously, a degree of uncertainly the activities of water and all solids were must be attached to each calculation of SI assumed to be unity. or equilibrium-soluble concentration of The Saturation Index (SI) diagrams zinc, calcium, or orthophosphate shown were constructed by constraining in the diagrams. Since the computer mod REDEQL.EPAK to the condition that no els depend on the results from a combina solids be permitted to precipitate, even tion of many equilibrium constants though a state of saturation or supersatu interacting as a group, only semiquanti- ration might exist. The thermodynamic tative conclusions should be drawn Stale of saturation was'quantlfiecl byTfii unless correlation with detailed field or P Sl.n.KUi jefmeiT*as"the cdiwfiigrn5ga~- laboratory data suggests that more confi Figure 3. Precipitation diagram lor zinc in rithm of IheraTjopriKe ionacGvjfy'prp- dence can be placed in thepredictions. An model system M8 duct (IAFrio~tKe tfiermodynarmc~soIu- analysis by Langmuir" of the sizes of binTjTpfoduct "cqnsTanTT7'KTnre3<ifnplel errors generated in the.calculation of sat 'foTTfycIroxyapaTite^TCaslPOiljOH) the uration indexes of calcite and dolomite mass action equation for which the solu (CaMg(C03)3) from carefully collected bility product constant is defined is and analyzed geochemical field data indi- Ca5(P04!j0H(s] = 5Ca3- + 3POp- + OH' cates a probableuncertainty of 0.1 SI unit. TmTEitpeTience of~th/sautKorsTms' (6) tieen (Rat tfie~n5nlitY oCtile^atafrorh most environmental svstenis ii~erjatTy The SI expression is, therefore, infenorTahd a7TjimcertainIviiif0j^Sl unit is conservative. The uncertainty is SI (!o3Ca3-o3KPj0o43-oOH'' t?) eveiTgrealeTwhcn redox reactions are involved or when the solubility constants are poorly known. Major sources of poor When the solid and the solution are data include pH measurements, lack of ex^drh^irt^quTnBnuTJCT7?P^T?^ran3^r differentiation of dissolved versus sus ^O^/KenTHe^oIutiorns^upersaUifnleil pended forms for metals (particularly for with respecrtFtiTTsoTldrffi^ such elements as iron, manganese, and therefore.jn^oTTlonverselvrwhen a lead), poor means of estimation d the PH Figure 4. Saturation Index diagram for mode) system M9 st^ofunhefsaturariorrcxlstsr^i~:5~or' The, iqility^nhFSlrlTagrhmCcohsIst- ing_ofjLpIot ojLSl versus nHTTsTlianhe reRipiTorp^dyeSrdelinMa^neeorDll redox potential (Eh, pE, and so on) of the water, the frequent absence of equili brium in the case of redox-sensitive spe cies, and poor conversion of alkalinity TriwIucK^cnveTleDOsnTonTiracoati titration data to total carbonate concen tKermohvhamlcallvfaSSS. Laboratory tration (or differentiation among the var amTTieltl data can then be used to corrob ious carbonate species). orate the theoretical predictions. Once Protection by zinc precipitation. Aspects of corroboration is accomplished, the SI dia the geochemistry of zinc have been 1.0 r grams may be applied to new localities widely investigated for fresh natural and situations to develop treatment pro waters and sea water.35-"-37 Nriagu35 has grams without extensive (and costly) studied the solubility of o-hopeite experimentation. In common with obser (Zn3(P04)3 4H30) and some of the equili vations made previously concerning the bria of zinc-orthophosphate coraplexa- difference between the numerical satura tion. Solubility constants and other tion state of calcium carbonate and the thermodynamic data for relevant zinc oi - effectiveness of the coating produced, the solids and complexes have heen deter same limitation can Apply to any other mined or critically reviewed by many precipitated solid with any component(s) investigators.3-3*-35"55 indireclb^affected by pH. From the context of metal-removal pre Another type of diagram was prepared cipitation. Patterson et nl" have pres by having the computer program "allow" ented some experimental data interpret precipitation of whatever solids became ed in the framework of the solids Zn[OH)3 001 I---- 1...........-1 $ 7 -1..- . 1 . \ supersaturated at a given pH and then iind ZnCOj. plus four zinc-hydroxide 10 calculate the predicted chemical specia- complexes. Under several of the experi pH tion. By examining the final total concen mental conditions, however, the solid igure 5* Precipitation diagram for zinc in trations of dissolved forms of various basic zinc carbonate (also known as zinc model system M9 elements and the mass of precipitated hydroxycarbonntc. hydrozincitc, and solid, minimum treatment dosages may^ j be estimated by selecting concentrations \ for nuuM systrm* luvri in T-aMi* 7 bat rot UUiMratr>l in jMprr can In* obtainr1 Irtuti I he* that exceed the equilibrium solubility author. HCEMBER 1981 M.R. SCHOCK & R.W. BUELOW 641 _. i CAPCO JEN 0004-81 1 a ! '* \* iV * < r -V r : ' > !? I---  2nyCO) *>0H m ZnCOj A C2f*tCOOHj--C} *cit*. OC*OM> r r * 2ni coi ? O'- ZnCO) 4 * A Cat P9t >Oh O CaCOj Ca'c*t A tn OH } * OCt OK; Figure 6. Saturation Index diagram tor model system M10 Figure 9. Precipitation diagram for zinc and iron (III) in model system Mil Figure 12. Saturation Index diagram lor model system M12P r^ I 01 - i 2njCOj }OH II ZCO) I o C*CO) Ca>etr A ZnOH; O C OH-} V? V \. \\ < Jt kV'' 2<nc D Ca-c^tm a Wi5>ht*PO* a _J j., ID Figure 7. Precipitation diagram tor zinc in model system M10 Figure 10. Saturation Index diagram for model system M12 Figure 13. Precipitation diagram for zinc, calcium, and orthophosphate in model system' M12P ' 2"* CO> OH i * 2rCO) _ O CaCOj Cak'ta A 2n OH } O Cl OH i Figure 8. Saturation Index diagram for model Figure 11. Precipitation diagram for zinc in Figure 14. Saturation Index diagram tor model system Mil model system M12 system M13 Zns(OH)6[COs)2) and zinc-carbonate complexation were probably significant controlling factors also.**-5*-40 Larson44 has suggested that hydrozin cite may be an effective corrosion inhibi tor for galvanized steel piping in the pH range of 7.5 to 8,5 at alkalinities of 50 to 100 mg/L as CaCOj. Mah and Boatman44 642 RESEARCH AND TECHNOLOGY ditf*a" limited corrosion-inhibitor study with A-C pipe using a combination of lime and zinc orthophosphate. The zinc level given was 0.3 mg/L, the "lime + orthophosphate" concentration was "5.0 mg/L" {form unspecified], and no other solution parameters (pH or carbonate concentration] were definitively report ed. They observed some deposition of zinc and iron, and less apparent leaching of calcium and silicon, but no deposition of phosphorous. The fiber count data were inconclusive. The results of DVVRD experiments 1-5. 11-15, and 19 (Tables 1 and 2] may be compared with the SI and precipitation JOURNAL AVVVVA CAPCO JEN 0004812 Figure 15. Precipitation diagram (or zinc in model system M13 Figure 16. Saturation Index diagram for model system M13P Figure 17. Precipitation diagram for zinc, calcium, and orthophosphate in model system M13P diagrams calculated for the mode) sys tems [see Tables 7 and 8). Monitoring the rate of calcium leaching often gives a good prediction of surficial pipe condition.1 particularly when used in conjunction with a complete water analysis. A higher rate of calcium leaching was Figure 18. Saturation Index diagram for model system M14 Figure 19. Saturation Index diagram for model system M15 Figure 20. Precipitation diagram for zinc in model system M15 observed at pH 7.0 than at pH 8.2 in con trol experiments 1 and 3. Figure 21 shows the effect of 0.3 to 0.C mg/L zinc added ns zinc orthophosphate compared with the untreated equivalent, both at pH 8.2. The solution rate of the calcium-containing cement materials was greatly restricted. Figure 22 indicates, however, that at pH 7.0, after an induction period of about'80 days, the leaching rate of the coupon in the zinc-containing system (experiment 4) was essentially the same as that of the coupon in the control system (experiment 3}. Upon inspection, the coupon from experiment 3 was noticeably softened. Experiment 5 (Figure 23) consisted of the substitution of zinc chloride for zinc orthophosphate at pH 8.2; following a very brief induction period, the calcium solution rate was seen to be extremely low. A very hard surficial coating on the coupon was observed when it was removed and inspected. Experiment 12 consisted of the same quality of water as in experiment 5 at a pH of 7.5 and resulted in a coupon that was noticeably softened. Experiment 16 used zinc sulfate as the' zinc source in water similar to that of experiments 2 and 5. The coupon was protected by a hard bluish-grey-colored coating like that from experiments 2 and 5. The sulfate salt of zinc is interchangea ble with the chloride and orthophosphate ' salts because the formation constants of the sulfate complexes of calcium and most other cations present in a drinking water are small enough that zinc solubil ity is only significantly affected at unrea listically high sulfate levels. The experimental observations corre late extremely well with Figures 10-13, which predict an onset of hydrozincite precipitation at a pH of approximately 7.8. Reduction in the zinc levels in the dosed tanks also showed a correlative slowing in rate around 0.3 mg/Lin experi ments 2 and 5. Virtually no reduction in zinc was noted in experiment 4. Model system M8 was developed to anticipate the effect of lowering the cal cium level in the water, while maintain ing the same pH, carbonate level, and zinc dosage. Zinc chloride was chosen for all but one of the remaining experiments because it gave performance equivalent to the orthophosphate salt, and it did not generate the discoloration (microbiologi cal growth) evident at the surface of the O-ring seal around the tank lid (seepart 1 for experimental apparatus1]. The pipe coupon from experiment 11 was covered by a hard grey coating similar to that of coupons from experiments 2 and 5. Figures2and3predict deposition ofahydrozincite coating, an indication of good protection. The main difference in protec tion afforded by model systems M8 and M8P versus M12 and M12P is that less protection is theoretically available at extremely high pH in the absence of orthophosphate, ns a result of the low calcium-carbonate precipitation poten tial. The zinc equilibria are clearly domi nant (Figures 2, 3, C-13). The absence of calcium docs not detract from the zinc coating process, but its presence is prob ably desirable to help stabilize the calciumsilicate phases of the porlland cement. DECEMBER 1981 M.R. SCHOCK & R.W. BUELOW 643 CAPCO JEN 0004.813 644 RESEARCH AND TECHNOLOGY Experiment 14 involved n reduction in one involving a minimal dosage of added - the cnrbonate level in the water to chemicals. Second, there is the possibil- ' approximately 2 mg/L ns CaCOj, but an ity of zinc toxicity in fish, which could ! increase in the calcium content to an result if wastewater treatment provided initial value of 32 mg/L as CaCOj. Sub inadequate zinc removal. This possibility stantial calcium leaching was observed is briefly discussed in a later section of for approximately the first 50 days, with ' this paper. There is no evidence of any a lesser rate following (Figure 24). The danger to humans, however. Third, even . decrease in rate is interpreted as the if a high dosage is employed by a utility i result of the onset of hydrozincite forma as a conditioning step, the equilibria may tion after a period of initial leaching of be reversible (such as with calcium car- easily available calcium and the increase bonate) to some extent, and continued in carbonate content. The carbonate dosing at the threshold of hydrozincite increase is attributed to C02 absorption saturation would probably be necessary. from the air. There is a strong correlation A test ofa one-week dose of zinc all mg/L among the time of the zinc decrease, the followed by a lowering of the zinc concen- I increase of carbonate, and the slowing of tration to 0.3 mg/L (conditions otherwise the calcium leaching (curves for zinc and similar to experiment 2) is presently ; total carbonate are not shown). Of the under way to test the effectiveness of J coupons from tests using zinc addition at such a procedure. Perhaps only intermit- j pH 8.2, this coupon appears less pro tent supplementation of zinc would be ; tected than all others, except possibly possible if the reaction reversibility were j that from experiment 13. The low level of slow. i carbonate puts the system depicted by The role ot orthophosphate.The role of the. ; model M14 and demonstrated by experi orthophosphate in the coating process of j ment 14 on the threshold of hydrozincite the DWRD experiments is apparently ; precipitation (pH ~ 8.3), thereby making minor or nearly irrelevant. The published i the protection by zinc minimal. The solubility constants and phase diagrams 1 protection by calcium-carbonate deposi for zinc orthophosphate3*-47 indicate that ! tion would also be impossible, even up to these compounds are too soluble to reach j pH 10. . a stale of supersaluration under any of j The last laboratory zinc-protection the conditions of the DWRD laboratory experiment to be discussed here is experi experiments. Additionally, supersatura ment 13, where the zinc concentration tion is not indicated at any of the combi- was reduced to 0.1 to 0.2 mg/L with cal nations of pH. zinc, and orthophosphate cium and alkalinity levels of 10 to 15 concentrations of the computer-modeled mg/L as CaCOj and 19 to 21 mg/L as systems. Evidence is provided by the ; CaCOj, respectively. The coupon was energy-dispersive x-ray spectrometry > noticeably softened as would be pre data of Mah and Boatman43 and this dicted by model system MS) (Figures 4 laboratory4* which show that phosphor- ! and 5). The low level of zinc employed ous is not present on the surface of the 1 was insufficient to provide hydrozincite pipe samples, but that zinc is. Also, no precipitation protection below a pH or phosphate decrease in solution was approximately 8.5. observed within the precision of the ana- i Figure 5 shows clearly that even at a lytical method used. : higher pH, very little zinc would be de Considerable uncertainty exists in the posited compared with many of the other selection of the proper value for the solu- ; systems. bility constant for hydroxyapatite. The The relationship of zinc solubility to precipitation process involved and the total dissolved carbonate concentration exact crystal chemical nature of the solid ; and pH is displayed in a slightly different formed are both highly variable and very ) manner by Figure 25. This solubility dia sensitive to the experimental conditions : gram was constructed considering the employed.4*-33 Lundager Madsen34 has \ aqueous zinc species and zinc solids suggested that although hydroxyapatite ; given in Tables 5 and 6 and the ionic possesses only one true thermodynamic ; strength. The calculation procedure was solubility constant, it is useful operation essentially the same as that utilized in a ally to employ two different solubility previous lead solubility study.4* One of constants. One would be used when pre- : the significant predictions obvious in this cipitation is being considered and the diagram is that littie advantage is derived other when dissolution processes are from carbonate concentrations above contemplated. The value selected in ; .approximately 50 mg/L as CaCOj in the Table 6 reflects a relatively high estimate ! pH range of 8 to 9. A pH of 8.5 affords of hydroxyapatite solubility. ; considerably more protection than pH The difficulty in quantitatively de 8.0, but pH 9.0 shows only a small scribing the calcium-orthophosphate improvement over pH 8.5. equilibria arises in large part from the Experiments of long duration at zinc tendency for octacnlcium phosphate or concentrations greater than 0.5 mg/L brushite to be the initially precipitated have not been performed, primarily for phase.3''.33-33-" with the onset of precipita- : three reasons. The first is that the most tion occurring when their solubility pro desirable treatment procedure would be ducts are exceeded. Subsequently, these ' ! JOURNAL AWWA CAPCO JEN 0004-814 tu iil UAJ`3* level had risen to 0.3 mg/L, where it patite, though possibly after a very long remained throughout the final threeand a time.50." A more correct approach may be half months of the test. Duplicate sam the use of one of the alternative salts to ples were filtered by using an all-plastic predict the onset of precipitation. There support unit, and 0.45-pm polycarbonate is also significant evidence of the inhibi membranes showed virtually identical tion of precipitation by low concentra zinc concentrations, indicating that the tions of carbonate, magnesium, and zinc was entirely in dissolved form.jrjj strontium ions. similar concentrations of zincaltainejlby In conclusion, concerning the role of Ensiolulmniwcperirnen^ iTiTwni^elylhnryinc- "tl orTcmcimmoHRophosphatesolidscould provItleTirotefLlioalaj, A-LfpipeTexcept pbssibljjQitv'ery hjgiTNpH~Teyeji""ar nt reversible'and'quarmfafiver c^remelylngKleveU ofcajcIunC7.inc.~0r ThemajcJmfliculTy^wnh using this o7riioplT^j^gT^1~her^vas~esi9TitiMv mechanism by itself is the hardness5*.** of rio~~<mTerenceinl!ie. fardnesses ofthe ` the surfaces developed on the zinc- coatingSDrovIHed bythe zinc^chloride protected coupons. In mineral form sysX|mst~T5B^?lnc-orlKopEo5pii^e^\^ hydrozincite is considered to havea hard temsr^nHmF~zTnc^suTfate systemTTTlie. ness of 2 to 2.5 on the Moh's hardness scale,** which roughly corresponds to a Knoop value range57.*5 of 32 to 80. The observed hardness, which is somewhat difficult to assess because of the thinness pnale complexes are of minor importance of the coating and because of occasional in~ watersystemsat the~HosnoeTeveIs~ unevenness and porosity of the substrate, suggesHE is more on the order of 4 on a dried Substitution of other phosphate com coupon. pounds for the chloride, sulfate, or Several zinc-silicateminerals have tab orthophosphate-zinc salts is not justified ulated hardness values comparable to the on-the basis of the experiments reported observed hardness and color of the cou here. The molecular structure and aque pon coatings. Table 9 lists hardness ous chemical behavior of "condensed," values for several zinc solids. The viabil "glassy," or "poly" phosphates are quite ity of a conversion of hydrozincite to a different from that of the orthophos zinc silicate coating may be.qualitatively phates.45-50 The ability of many of the tested with some of the existing thermo condensed, glassy, or polyphosphates to dynamic data"34 that is available for wil- f-rm strong aqueous chelates or com- lemite (Zn2SiOj) and ZnSiOj. Hemimor- <es with Cal* and Al3', which are phite (ZnSi:Oj(OHJ2 H20), for which jor pipe components, as well as with thermodynamic data are scarce, is a very ther potentially protective ions (such as reasonable possibility, based on its Zn7*, Fe'\ or Mn1*) may enhance the solu occurrence as an alteration product in bility of the pipe under many conditions. oxidized zones of zinc ore bodies57 and in Further theoretical and laboratory eva galvanized pipe corrosion.80-Leggett61 has luation of the behavior of the alternate described experiments where zinc-silica phosphate compounds is necessary solids of undetermined form were created before substitution is carried out. by titrating dilute, acidic solutions of Mechanism ot A-C pipe protection by zinc. monomeric silicic acid and zinc to various Positive identification of the exact mech- pH values. The results are somewhat dif mism involved in the zinc coating pro- ficult to quantitatively interpret with cer ess has not been made. However, a tainty, because no effort was made to wo-step mechanism consistent with the exclude C02 gas from the systems, but tboratory and field data gathered thus the formation of hydrozincite, or smith- ir can be postulated. The first step is the sonite, or both, is veryprobable under the ttainment of a state of saturation with experimental conditions reported. The ispect to zinc hydroxycarbonate hydro- given solubility curves are also quite ncite: comparable to the DWRD laboratory data 5Zn7* + 60H* + 2COjZnj(OH)e(COj)j(s} and model systems forthesilica-free sys (8) tems. The concentrations of dissolved sil ica employed by Leggett (0.001 M in Si(OH).i) were much higher than the con The zinc levels in the experiments centrations in the DWRD bulk solutions, iere a coating is formed tend to but they may be comparable to levels crease to values in close agreement attained ftfferstitially upon pipe dissolu th equilibrium total-aqueous-zinc conitrations predicted by the compulerd model. An experiment was set up h no zinc addition, but a section of ised galvanized pipe as a zinc source, h water quality otherwise the same as .xperiment 12. After 40 days the zinc tion. Gilmour and Kittrick*7 have discussed several zinc-solubility experiments con cerning behavior in soil systems, and they concluded that in several cases Zn-SiQj(s) could be a viable control on zinc activity when hydrogen sulfide lev- MEMBER 1981 TADLE 8 Correspondence of model ond experimental systems DWRD Experiment Applicable SI Diagram Number Model System |F*Sur| Precipitation Diagram (Figure) 1 M12 10* 2 M12P 12 13 3 M12 10* 4 M12P 12 13 S Mil 10 , tl It M6 2 9 12 MS 2 3 19 M9 4 6 14 Ml* IS 19 M12 10 11 16 M12 10 11 *lgnort fields of xine-containtng solid* in the diagrams; the ether fields will be changed little. TABLE 9 Moh's hordness volues for several zinc . minerals54 Common Name Hvdrozincite Smithsonitc WiUemlte Hemimorphitr Zincite Formula Zns(OH)b|CO>)2 2nCOj Zn2SiO< ZntlSisO?) (OH); *20 ZnO Hardness 2*2.6 6 5*5.6 4.5*5 4*4.5 TABLE 10 uiuw jigc cucifry uj fuiimmvii vujmc* ui 25C employed in the construction of the E-pH diogroms in figures 26 and 27 Species yCf*--col'mol Source H:0 OH* HiCOj* Hcoy co*- Fes* FefOH)2* F.IOHJj- Fe* FeOH*** Fc(OHJr FefOHJj* r.[OH)j- Ft(OH)3ts). fresh ppl. Mn* Mn** MnOH* Mn|OHjj" MnHCQs* MnOa* MnO*>* MnOitsJ, pyrolousite MnjOjis). bixbyite MnjOals). hausminite Mn(OH)2i*}. pvroebrosite Mn(OH)jls) MnOOH (s). manpanlte MnCOafs). rhodochrosite a2(o,i HCIO* OCT -56 690. -37604. -148950. -140280. -126 390. -18 650. -104 200. -146630. -1060. -54 770. -106700. -152600. -198400. -164 500. --64 490. -19700. -96800. -177900. -197100. -106900. -119700. -111400. -2)0300. -306 900. -147300. -181100. -133400. -105100. 41650. -19100. -8800. 1 1 2 2 2 4 4 4 4 4 4* 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 3 3 3 Sources: ]. PARKER. V.B. CT AL. Selected Thtrmpchemic.il Data Compatible with the CODATA Recommendation*. NBS NBSIR 75*965 (1976). 2. Calculated from equilibrium constant* plven by Harnrd and Davis in four. ACS. 66:2030 (I943J am! b> Horned and Srholes in faur. ACS. 63:1706 (1941) for 25PC. with the value for ACf* )l*0 from this table. 9. WACMAW D.D. ET AU Selected Value* of Chemical Thermodvnamie Properties. NflS Tech, Note 2*0*3 (1968). 4. Calculated from equilibrium constant* and equations tabulated by Plummer ct al in W ATKQF--A FORTRAN IV' Version of VVATKQ. A Computer Program for Calculating Chemical Equilibrium of Natural Waters. USCS VR176*13 11976). with other \atur> given in lhi> table. J M.R. SCHOCK & R.VV. BUELOW 645 CAPCO JEN OOOA815 r TABLE 11 Soturalion Jnc/ex calculations for several potential!y protective solids* Wattr Supply System Tulut pH Alkalinity Ca Ms Yc Mn .sio* Al Calcitr Saturation Index** CaCOjppt0 r|OHb m fitted Quartz DtlecULV FiUrs* Multiple samplin R* I) 74 g * 0.4 C*1 8.0 JJhJ 6.5 12 6.6 13 C.C 14 6.4 15 6.6 Kk 74 Single samplings Piant-L 7.3 7.4 Planl-M 7.6 8.1 Plant-N 74 8.3 Plant>0 6.5 Sy*tem0 8.4 Plant-P SysttinP 7.1 7.0 220 100 50 17.6 19 4.4 4 2.3 7 2.0 23 04 10 7.2 16 74 76 3A 155 163 180 165 174 26.6 210 216 130 141 14 14 4.0 C.S 5.3 0.6 2.3 2.6 23 25 NA 0.3* NA NA 0.7 0.06 0.5 0.6 0.6 0.11 2.3 0.19 1.7 0.11 24 0.01 6.1 0.06 2.3 042 2.5 046 1.6 0.12 0.0 <0.10 2.4 1.4 2.0 0.05 0.5 <0.1 0.5 <0.1 6.4 0.24 9.0 3.8 NA NA 0 0.03 0.05 0.02 0.01 0 O.CH 0.03 0.03 0.06 0.03 0.05 0.02 <0.03 <0.03 0.42 0.4G NA NA 6 6.7 7.4 14 12 12 0.4 12.5 12.7 10.3 7.0 8.1 0.3 8.7 8.0 114 51 11.0 49 11.1 35 10.9 36 11.6 16 0.6 11 11.1 15 11.6 10 11.6 60 11.0 60 10.0 0.6 0.8 -1.5 -3.0 -3.6 -2.5 -3.1 -2.7 -0.0 -0.6 -0.7 -0.9 -04 -2.1 -04 -04 -04 , -0.6 --0.6 0.1 0.3 -2.0 -4.4 -4.1 -3.0 -3.6 -34 -1.4 -14 -14 -1.4 -0.8 -2.6 -1.1 -0.7 .-0.7 -1.3 -1.3 34 t 2.6 24 -2.1 04 1.4 -1.5 0.3 1.6 -14 0.5 14 -3.0 0.5 0.4 0.5 1.5 -0.6 -0.9 2.5 2.4 0.0 2.7 2.8 0.9 2.5 4.0 0.6 6.6 0.6 3.3 24 * 2.1 64- * 0.3 0.4 0.2 2.4 2.7 1.0 3.5 2.4 w nunc none . occasion.^ lew NAl BDl NSS * 1.0 24 BDL NSS 8.4 26.6 1.5 NSS Temperature is assumed to he I5eC for systems D-K; 25*C for l.-P. K was estimated at +0.C2 V. Analysis values ate mg'l. and mjj'L as CaCOa for alkalinity. Notes: NA--not analyzed; a--Aggressiveness Index pH + log (AH) (see text): b--Saturation Index = Log (lAP/K*) (see text); c--fresh phase described by Suarez (1977); d-- solubility constant used for p> rolusite: e--millions of fibers per litre fur single sampling* (NSS-- nut statistically significant; BOL--below detection limit): f--raw water contained jj mg/LFe and unspecified Mn; No Mo %'dluegiven nfteraerallon./iltratiun.rmdchiorinatiomg--sedimentation and filtration reportedly incomplete after ferricsulfatecoagulation: h*>w.iteramil>ses>were provided by the utility:--a water blended from the five systems shown is the actual system input, following chlorination; j--inspectiunof the pipeshnwed smooth lining on inside. * V els are low enough to preclude the forma tion of ZnS(s). Additionally, in a paper on corrosion inhibition on metal pipes, Lehrman and Shuldener3 reported the proba ble formation of a zinc-silicate phase on galvanized iron pipe carrying hot water following a dosage of 8 to 12 mg/L Si02. To calculate the tendency of hydrozin cite to convert to willemite or ZnSi03, equilibrium constants for the following proposed reactions were calculated directly from Gibbs free energy of forma tion tables by standard procedures,51`14 using a value of --755 kcal/mol for zinc hydroxycarbonate derived from the equilibrium constant and equation tabu lated in Table 6. 2Zn5(0H)r,(C03)2(s) + 5Si[OHJ^ . (9) SZn3Si04(s) + 4HC03` + 4H*`+ 12H20 ' Zns{0H)6fC03)2(s) + 5Si[OH)i 5ZnSi03(s] + 2HC03` + 2H* + 11H20 The equilibrium constants correspond ing to reactions 9 and 10 at 25C are 10",6J and 10-154, respectively. In DWRD experi ments 2 and 4 the bicarbonate con centration was approximately 10-'4 M. Therefore, for reaction 9 the equilibrium silica concentration would be approxi mately 0,2 mg/I. Si02 at pH 7.5 and 0.05 mg/L Si02 at pH 8.2. For reaction 10 the concentrations would be 0.002 mg/LSi02 at pll 7.5 and 0.001 mg/L Si02 at pH 8.2. These levels of dissolved silica are exceeded by any of the test waters 64G RESEARCH AND TECHNOLOGY employed, and they should be exceeded to a greater extent by silica concentra tions in the interstices and at the coupon surfaces caused by cement dissolution. While the bulk-solution zinc concentra tions are obviously not controlled in the experimental time frame by these zinc silicates, conversion of aJhydrozinite surficial_coaling_to a zmc~silicateis clearly in^icatecTasTdistlnct possTBilitvi ffTacHlition to thecolorandnarHnessof the pipe-coupon coatings, further evi dence that some surface reaction does take place is seen by the presence of some grey zinc-compound coaling, even on the pH 7.5 coupons, and on those coupons from systems with insufficient zinc for hydrozincite precipitation at pH 8.2. The presence of a large amount of zjneat the pjneJsuHac^aD~pears~to~T3r7rnecesRarv preregtimTeTorelTectivejleyeloproenl of a hard coating; ' FuturSTesearch will include the use of some surface analysis techniques, such as polarized light microscopy or Raman spectroscopy, to attempt to define more exactly the nature of the coating. Prptection by sodium-melasilicate "solution Two experiments (9 and 10) were per formed to determine the efficacy of sodium-melasilicate (Nn2Si03 9H.O) addition as a protective procedure for A-C pipe. The pipe coupons, representing a silica concentration of 14 to 18mg/Lnt a pH of 7.0 and 8.0 under similar calcium and carbonate chemical conditions (see 1 Table 1), were found upon inspection to be clean and not deteriorated. This result is somewhat in contrast to that reported in a Seattle study,*4 but the silica dosages used in the DWRD tests were somewhat higher than those used in the Seattle experiments (approximately 10 mg/L as Si02). The protection of the coupons in the DWRD tests showed a good correlation with the relatively small increase in cal- cium observed for both experiments. Scanning-electron-microscope photo graphs of both coupons in several loca tions on their surfaces showed similar appearance, with few fibers obvious on the surface. The coupon from the pH 8.2 experiment had a slightly more uniform appearance, but the difference was small. The protective mechanism provided by tfrjPchssolved silica~is iKiPknown. The conceliTrationlsTKnTugircnougfu favor spontaneous precipitation of amorphous silica, but it is high enough for quartz formation. There is a possibility that quartz for mation may be enhanced by the struc tures of many of the silicate solids in the Portland cement itself.11 Some cement phases tnay tend to induce epitaxial growth, providing a mechanism for rela tively rapid coating, because the unsup ported crystallization of quartz is very slow.11 X-ray diffraction analyses'5 of surface samples from two different dis tribution systems having dissolved silica levels between C and 14 mg/L Si02 indi cate the presence of quartz to possibly a JOURNAL AWWA CAPCO JEN 0004816 greater degree than woutd be expected from the cement matrix itself. Two of the attractive aspects of the satijum-silicafiPtrealment are the possihility~^irso projecting severnTTypesrof m5aT~pi6esw^'"~~wilhout The adverse side effects of ortliophTisnhnle^ahTrThe ppssjEHify' ot TrcjrtmentatnTower^T. Some^ITrfcuIt^encoimteretDiovv-eyeri ht^cermin'lnduslBaCnrocesses^^vv^hen -appreciable dissolved silica iFpreseptjn. `the waterTpartictilarlv whenlngwateris heated. Treatment of deteriorated pipe Several additional experiments involv ing coupons taken from distribution sys tems suffering some surficial softening from aggressive water have also been performed or are in progress. Coupons softened in previous experiments'and from use in a distribution sys tem were placed in a test system similar in chemical composition to experiment 2. The coupons did not show appreciable calcium leaching, but the coating formed was not as hard or smooth as on the new pipe coupons. Presently, a test is in pro gress in which softened samples are placed in a water slightly over-saturated with calcite at pH 8.65. The results of.DWKD experiments conducted~jEyst{arjn3TTieiyqhsecyations ind icatelhait^cidJUonoT^liecorrosionl' controTcompouSisjesteynTusTarcannor bc'reli'edupdnto^Dro\jde7correcjionof A-C pipe deterioreljpiu If possible, all nftened'matenal sEouId be removed to - vide a hard, clean surface; this surface t then be maintained by proper water litioning. The feasibility of cement lining of deteriorated A-C pipe is cur rently being studied. The success of application of corrosion-control chemi cals in water distribution systems will depend on the quality of the pipe surface at the time corrosion-control treatment is initiated. The control of microbiological growths through proper disinfection practice is extremely important, and corrosion-control treatment must be operational before pipes are cleaned. Results of DWRD in-situ pipe rehabilita tion studies will be reported at a later time. Natural inhibitory factors This section discusses several constit uents commonly present in drinking water that may serve to inhibit deteriora tion of the A-C pipe without any special water-treatment procedure. Several metals (such as lead, cadmium, and copper) not discussed in this section could also form protective coalings, but they are comparatively rare and are gen erally present at very low levels. There fore, only the most common metals are covered in detail. Iron and manganese. Two metals fre quently present in drinking waters that could provide protection to A-C pipe by virtue of their low solubilities are iron and manganese. In part Id samples from water systems l`ah'd~G~were reported to jiave obvious brown coatings'largely' compos~ed 61 'iron,, and ~tKe jpuryjyed waters were shown to have minima'Tor jionquantifiiible fi'Eer counts. PimTsections from the large-scale DWRD experi mental pipe loop were alsoseen toreceive a protective iron coating, even at pH 5.5 and total alkalinity of 1 to 3 mg/L CaCOj. Systems D and E discussed in part 1 have water conditions that could favor the for mation of iron coatings. Pipe inspections and detailed chemical analyses were not performed, presumably because virtually no fibers were detected in the drinking water samples, and the AI'value was above 12 in each case. At the completion ofexperiment 19 (see Tables 1 and 2), duplicate 1-L water sam ples were taken for fiber counts.'After months of operation resulting in a calcium-level increase from 9 to 25 mg/L as CaCOj, one of the samples was deter mined to have 0.45 million fibers perliter, and the fiber level of the other sample was not statistically significant. The presence of 0.1 mg/L of total iron was observed to impart a fightTSafjngjbTlie pipe coupon, which served to helpbind fibers even though thtTpipB^urface'was softened from water attack (as evidenced, also by~the calriuin leaching). Several water supplies tested by USEPA, but not discussed in detail in part 1, have also shown low or nondetectable fiber levels in spite of having AI values below 10. In at least one case, a manganese-containing solid was found, in addition to that pound in system I. ' The general subject of the chemistry of iron in water has been reviewed by Stumm and Morgan." and Hem.0-07 Many possible reactions of iron in groundwa ters have been discussed in some detail by Hem,67 Hem and Cropper,00 Lang muir," and Langmuir and Whittimore.TM The influence of particle size, surface area, and oxidizing conditions on the solid phases formed and the solubility constants involved have also been inten sively studied for both ground and sur face waters.70-71 All of these studies can be used to provide a general descriptive model of conditions that would promote the formation of potentially protective coatings on A'-C pipe. Figure 26 is a potential-pH diagram for 0.1- and 1.0-mg/L iron activities in water contfttfiing 40 mg/L as CaCOj dissolved carbonate species activities, constructed using a WATFIV computer program writ ten by Froning et nl.72 The data used to construct this and subsequent E-pH dia grams are given in Table 10. The forma tion of a ferric-oxyhydroxide precipitate (written as Fe(OH)3) is profoundly influ enced by the redox potential of the wtiter involved, as well as hydrogen ion, metal- V lie species, and carbonate species activi ties. The oxygenation and precipitation processes are also affected to some extent by sulfate ions70-72 and the carbonate buffer capacity," but these influences will not be included in this qualitative discussion because of their complexity and relatively small impact'. Few studies have been done on the exact redox potentials developed in drinking water, but typical treatment practices such as aeration and disinfec tion are likely to generate potentials that could near the theoretical stability limit for water,21-" depending on the partic ular oxidant. Ozone has been observed to generate potentials in the range of approximately 850 to 1000 mV at concen trations of 0.5 to 2.8 mg/L." This would be well into the ferric-oxyhydroxide field at pH values anywhere from 5 to 10.8. Experiments successfully oxidizing iron . (II) to iron (III) by ozonation have been conducted." Similar results are provided by chlorination, because the species Cl3 (aq), HOCl and OCl` are all strong enough to oxidize iron easily. Such a mechanism has been suggested by Stumm and Morgan.21 The iron oxidation has also been linked to dechlorination in the distribution system, giving further corroboration.70 The combinations of iron concentrationTpfiTcafBoiiate ^oncentratigp^nd, ^ioSsfBly,^dfganTrm^terTfiaTproyi5ejbe_ mosTlsfiectivir'A-C pipe poatmgs'Kave* not beenstudieej. LaBorafory'fesls~per^ FormecTTly dWRD, energy-dispersive spectrometric and electron microscopic examinations of pipe samples,`and a few field observations show clearly that the iron coatings often serve to cover and bind fibers to the pipe surface. However, calcium leaching arid pipe softening is not prevented in many cases, indicating con tinuing deterioration of the cement matrix. Increased hardening with time, resulting from the recrystallization of the iron precipiate,05-70 may be found and would be a logical consequence, provided that the corrosiveness of the water to the Portland cement is not too great or the leaching too rapid. The.J.QW -solubility of several manga nese solids under oxidizing conditions is anotEeTtfieoretTcaltypreHicTable protect. tive~mechansisrnTorX^plpeTThe aqueous^ieimsti^'^orTnanganese and its oxides is extremely complicated. In addi tion to the numerous complexation and hydrolysis reactions it can undergo, man ganese also has five valence states attain able in natural and drinking waters.77 Manganese solids may also participate in redox reactions with other metals;77-7* like iron hydroxides, they can addition ally function as strong adsorbers of other trace metals" that may also ultimately interact with the pipe surface. The var ious aspects of the solubility and redox chemistry of manganese have been stu- DECEMBER 1981 M.R. SCHOCK & R.W. BUELOW 647 CAPCO JEN 0004817 pursued in detail here. f. tiwutuiiidilibdlt: as an iron and manganese deposition A poton'.inl-pH diagram for dissolved preventor in communities with either tnnrfgnnese species activities of 0.05 mg/1, iron problems or iron pipe, or both. In 'and dissolved carbonate species activities their study, where most of the waters had of 40 ing'/L ns CaC03 is shown in Figure high silica levels (10 to 25 mg/L) to begin 27. Selecting the proper solids to use in "diagram construction is complicated be cause of the several polymorphs of a with, the silicate showed signs of acting sometimes as a coprecipitator, sometimes as a former-of a thin corrosion-reducing V . i V* _ given stoichiometry.53"TM and the possible dominance of metastable solid species. Oxidation of manganese to form MnO/ is reportedly possible by chlorination,33 ns would be predicted by Figure 27, and by ozonation. The formation of MnO* by air oxidation is also common." .Feldspalhic sand (approximately 45 film on the pipe surface, and sometimes as a sequestering agent for ferric iron in solution. Though ihe. eagast nature of the--silica reactions, inyolving?v-ti pipe hasnot. ~bcen extensfyelyTlucld^Tnidrthefjeid^d M8 6 L |E u laGpratonTjevigence^ 10 - clearly~lndicates'TKaT^njc^arTpIayjn^ Esprtfrtnt U O It' percent quartz. 21 percent orthoclase fmpwfanTrolFip~^rCl^rpe^rotecHonr^^ feldspar, and 34 percent plagioclase ''Analysis ol tieid^ataTTnereliance-that feldspar) has been reported to be capable has historically been placed on the AI has 50 . WO ISO MO of oxidizing Mn3*to form solid MnCHnear pH 8." Whether-or not the various sil icates present in portland cement could serve in a similar capacity has not been studied. unfortunately served to focus analytical attention almost solely on pH, alkalinity, and calcium. Thus, very little additional water chemical data are available when correlations among minor constituent Figure 24. Comparison of calcium teaching at pH 8.2 between a system with high calcium but iow carbonate (experiment 14) and a system with low calcium and carbonate at the same level of zinc-chloride addition (experiment 11) Because of the general objectionability equilibria, pipe conditions, and fiber lev orTnanpanese^in^cirlnirin^'wateri^ncr els in the water are attempted. Some `PWKD experiments haveTieencon^ examples of how such additional data ducted, or are planned, with this metal. can be utilized to predict more accurately At least two lncaljties tested bvUSEPA the true aggressiveness of a water toward do have-iderUiTtablelnanRanese coalings, A-C pipe follow, utilizing the SI concept. onTheirA^C pipe, ancfTTremaiti sa_\dnule" Table 11 summarizes analytical data pos^ibHiryToFtTiHeFlocanons as yetnot from eleven systems, the first six of closely examined! which were described in part 1.* Systems Silica, 'l'he role silica can play in A-C L.M.N.O, and P arc located in the south pipe protection is quite varied. As was ern and southwestern United States and discussed in a previous section on the DWRD laboratory experiments, quartz overgrowths or another form of cement solubility stabilization may take place, represent only one sampling event at each plant before chlorination and one location in the distribution system. The SI for five potentially protective solids ld* COj-n9/L it C*CO) Figure 25. Solubility diagram for zinc versustotal inorganic carbonate concentration at fixed pH an occurrence progressively favored as was calculated using the computer pro the dissolved silica concentration rises grams WATSPEC2" and WATEQF,36 /= 0.005; 25 C above G mg/L. Silica may also combine which derive the total inorganic carbo with other metals directly to some extent nate concentrations from titration alka to form a coating. A study done by Kato" linity analyses correcting for other V shows significant coprccipitation of 9.3 reactive species. Corrections are also mg/L silica with hydrated oxides of man made for complex and ion-pair forma ganese and iron in the pH range of 5.0 to tion, temperature, ionic strength, and morphs. However, System D was also 8.0. The coprecipitation was observed to redox potential. In order to calculate the very oversaturated with respect to ferric increase with pH, temperature, and oxy speciation of the electrochemically active oxyhydroxide, and there is indication gen content, with iron being the much elements, it was necessary either to know that a similar state exists in system E.3 No more effective precipitator. Though the from measurement or to estimate the data were available for silica or manga concentrations of iron and ipanganese redox potential. Because of the absence of nese. Because the pipe was not inspected were much higher in Kato's study" than such potential measurements, an esti in either case, one cannot be certain to would be found in drinking water, the mate of + 0.62 V was used, in view of the what extent the various factors are indi mechanism remains plausible for the fact that all waters in the distribution vidually responsible for the pipe protec development of a thin surficial coating on systems were chlorinated. Some bias is tion. the A-C pipe. Presumably, similar behav introduced, therefore, for samples ob The actual nonaggressiveness of the ior might be observed in the presence of tained prior to such treatment. In refer other water supplies shown correlates aluminum hydroxide precipitates. ence to the previous discussion concern well with both ferric-oxyhydroxide and The relationship between dissolved sil ing disinfectants, the estimate of+0.G2 V quartz oversaturation and very little ica and iron is somewhat complicated by is probably low and will therefore pro- with the AI. Pipe samples from systems the ability of the silica to act both as a vWe a conservative estimate of the pre G and I have been examined by energy- sequestering agent and as a precipitator, cipitation tendency of iron and manga dispersive x-ray spectrometry*'-" and depending on. concentrations and solu nese. show the presence or manganese ns well tion conditions. Several studies have The two systems that are nonaggres- as iron. Manganese may be present as the attempted to determine the stoichiometry sive on the basis of their AI value (>12) result of adsorption, or as a solid, because and formation constants of ferric-silicate do indeed show an absence of fibers in the it would take only a small increase in the complexes with mixed success,"-"3 large water samples. The oversaturation with estimated redox potential of the water to ly because or difficulties with polymeri calcium carbonate is also shown by the SI reach, for example, pyrolousile (MnOa) zation in the solutions. Dart and Foley"3 for various calcium-carbonate poly- saturation (see Figure 22) in system I. A 648 RESEARCH AND TECHNOLOGY journaLawwa i - CAPCO JEN OOOA818 undersaturated with respect to'calcium have been tabulated in this area,'*-*'* but carbonate. research has only occasionally focused The ability of ferric oxyhydroxide to on the effect of the zinc speciation. The become more insoluble with age"-71 could aqueous forms present have often been be an important factor in the reduction of shown to be of critical importance in calcium leaching after many years of ser numerous copper toxicity studies.'1 vice. The existence of a role for organics in the protection observed in these sys Conclusions tems has not been studied. The experimental field and theoretical Obviously, more detailed field data data presented combine to suggest sev must be gathered to test further the accu eral important conclusions with regard to racy of the predictions of the more com A-C pipe stability and treatment method prehensive chemical model of pipe ologies, in addition to those previously protection presented in this paper. Partic stated.2 ular attention should be paid to changes 1. The Aggressiveness Index as tradi- 0 1 >9 < S (7 I I II II II U II in the various indices during travel in tionally"conceived Is "not theoretically PH Figure 26. Potentia!-pH diagram of iron in carbonate-containing water at 25C systems as an indicator of active deposi sound from a chemical standpoint to pn tion or cement leaching. Correlation with ycTTfSeFrelease and degradation ortiie fiber counts and pipe inspections (cer mtenorpipesurface irTsyslemsThat are Stability fields are shown for dissolved iron species activities of 0.1 mg/L (----- ) and 1.0 mg/L (------). Dissolved carbonate species `activities are 40 mg/L as CaCOj. Data are from Table 9. tainly as important forprotected systems as for deteriorated ones) is necessary to define more closely the individual or combinations of chemical factors in volved in giving the best prediction of the 'undersafurated with respect to^calcTieT* Fortunately, it has generally ptovetflq, 'predlcTTalsely pipe de.tenoTatio~n ~more oltenTKaTjo^redictjnirely'pipesiainTjTy" ^field^siruafions.'Hbvveverran'accurate' physical condition of the pipe. The index "of A^Cpipe condition must take detailed field data and correlations with into account the presence of protective fiber counts will also serve to define bet constituents such as silica, iron, manga ter the water chemical conditions favor nese, and zinc as well as pH, carbonate, ing fiber retention over release. The rates calcium, disinfectant concentration, and and patterns of flow in a distribution temperature. The role of organic mate sytem may also be of some importance in rials in the protective process still needs this regard. The vital need for properly investigation, and future field studies representative samples, carefully pre need to gather full chemical data to more served and accurately analyzed, cannot precisely define the conditions of optimal be overstressed. The data available are of A-C pipe protection. Observations^ gf 0 1 2 3 4 S 6 7 1 10 II U IS II limited usefulness in rigorous chemical A-C pipe sections from jrmnvlocalities*' OH `ngure 27. Potential-pH diagram of manganese in carbonate-containing water at 25 C studies, largely because they were not obtained with the aforementioned type of chemical data analysis in mind. ^nHusivelydcrnpnstrate^haOmnjTaf inhibitory factors, sued aiTlKose'^mem' tionedTare cornmonlylhe pipe protection" ibility fields are shown for dissolved tganese species activities of 0.05 mg/L and toi dissolved carbonate species activities of 40 mg/L as CaCOj. Data are from Table 9. Environmental impact of zinc treatment The use of zinc orthophosphate for cor %iSClljrffismOiQrc^iiesaturolTpm~" " ^ZrSomecoating mechanisms serve to bind the asbestos fibers against release but not to seal the cement pipe matrix J rosion control in a water distribution sys tem2 (system G) resulted in the fouling of from dissolution. Thus, cnlciurqand nH increases duringflow through aoistribu^ an industrial filter when zinc concentra fum system mustbe^infer^FelecfwiTjrcaut tions greater than 0.5 mg/L were used. Ron", as tK^HojiorpecessaHFngcncate similar situation of near-MnOj satura Subsequently, the dosage was reduced to fiber releaseTtnlierclieimHlrMctionsm tion exists for system K. which addition 0.3 mg/L. Because it is necessary to pro tHewatermay also affect the pH. ally may derive benefit from the slight vide a sufficiently high zinc level to cause 3-Tbechlgride, sulfat.e, a_nd ojJhpphos- ferric-oxyhydroxide oversaturation and zinc-hydroxycarbonate supersaturation ' phafgs^tSfifzmcTiayeTiaSSjaAtedLofld the comparatively high level of calcium in order to obtain an effective pipe coat tound^ provjd^dbstamTalprolection that may tend to slow calcium-silicate ing, special provisions may have to be pTjnLrC"pipewuen^tfey^areliBderLtO lhe dissolution. made for or by some water users. In the water at jhe proper concentration and pH Several-interesting trends are apparent cited case, some zinc was found to adhere rangesan5IiynCTtlipsercIicmlcaT~conqrr in systems L through P. In system O, in to sections of the pipe at the lower dosage, tions are maintained ^throughout the dis_- spite of the highest A1 of the group, some but the coating was not as well developed UjEugonsSSp^Zinc is the active agent. pipe deterioration may have taken place, as in laboratory studies at the higher zinc It acts to coat the pipeandprotectstlie indicated by the fiber, calcium, and pH level. Recent investigation has shown pipe against fiber release and water increases. The higher true nggressivity of that zinc levels were substantially attack initially by forming a zinc- this water is suggested by the low iron reduced during passage through the sys hydroxy-carbonate precipitate at most and manganese values, the marginal SI of tem dtfcfthat the precipitation potential of carbonate concentrations. Evidence has quartz, and the low calcium level. In sys zinc was minimized or even lost entirely. been presented that the zinc solid then tem P, fibers are actually reduced, possi Zinc levels throughout a distribution sys reacts with the pipe surface.itself, possi bly resulting from iron precipitation or tem must be monitored to assure that bly converting some or all of the coating from a combination of iron and silica. zinc-hydroxycarbonate saturation is to a harder zinc-silicalc solid phase. The Systems L and P show n large amount of maintained to the furthest extremities.- chemical dosages required for a given iron and a lesser amount of quartz super- Concern must he generated regarding general water quality are closely estima saturation through the system, as well as the ultimate fate of the zinc, in view of its ble by diagrams constructed from com two of the higher calcium levels of those potential toxicity to fish and other nqfla- puter-assisted calculation of simultan systems tabulated. Goth are significantly tic organisms. Numerous investigations eous aqueous equilibria. OECEMBER 19B1 M.R. SCHOCK & R.VV. BUELOW 649 CAPCO JEN 0004819 4. When systems involve both metal and. A-C.pipe, the zincrchlnride tnd sul fate alts provide little or no protection to the metal pipe. However, DWRD tests at pH 8.2 with lead tmd A-C pipe in the They are also appreciative of the patience of Pat Pierson with an extremely difficult manuscript and are indebted to Joseph C. Jackson, A-C Pipe Producers Assn., for comments on the history and intent of the York (1967). *8' Kitano. Y. et ai.. Adsorption of Zinc and Copper Ions on Calcite and Aragonite. and its Influence on theTransformation of Aragonite to Calcite. Ccochem. Jour.. ` 10:175 (1976). same recirculation system show that the Aggressiveness Index and to Bryant 19. McCabe, L.J. & Miliette. J.R. Health Orthophosphate salt provides corrosion Mather, US Army Corps of Engineers Effects and Prevalence of Asbestos Fib inhibition for both types of piping mate Waterways' Experimental Station, for ers in Drinking Water. Proc. Ann. Conf. rial because of the action of the ortho comments on the behavior of cement and AWWA, San Francisco, Calif. ()un. phosphate ion. The lead solubility in such systems appears to be governed by the formation of lead-orthophosphate {rather than zinc-orthophosphate) compounds, while the A-C pipe is protected, as has been discussed in this paper. Cement lin concrete in water. References 1. A Study of the Problem of Asbestos in Water, Part 2. Jour. A IVIV'A, 66:9:1 (Sep. 1979). 20. Cooper. R.C. & Cooper, C.W. Public Health Aspects of Asbestos Fibers in Drinking Water. Jour. AWWA. 70:6:338 ()un. 1978). 21. Guide to Compounds of Interest in Cement and Concrete Research. Special ings should behave essentially the same 1974). Rept. 127. Highway Res. Bd., NRC, or be even more amenable to zinc coating 2. Buf.low, R.W. et al. The Behavior of NAS. NAE (1972). than the A-C pipe matrix. 5. Substantially softened A-C pipe showmg~~Toosk Haers. a rough .surface caiisea bv considerablecalHumanosilica~leacKmgToiiHmtfr. niav~not be ableLo Be rehabnitatepEyoniy in-situ zincjreatment, unlessa"clean and hard surfaceTree Asbestos-Cement Pipe Under Various Water Quality Conditions: A Progress Report, Part 1--Experimental Results. Jour. AWWA, 72:2:92 (Feb. 1980). 3. School M.R. etal. Evaluation and Con trol of Asbestos-Cement Pipe Corro sion. EPA 600/D-81-067. NACE Cor rosion/81, Toronto. Can. (Apr. G-10 ' 22. Wacman, D.D. etal Selected V'alues of Chemical Thermodynamic Properties. NBS Tcchn. Note 270-3. ()an. 1988). 23'. Stumm. W. fc Morcan. (.). Aquatic Chemistry. Wiley-Interscience, New York (1970). 24. Garrels. R.M. & Christ, C.L. Solutions, Minerals, and Equilibria. Freeman. onooseTTBerecatuHe^stfirel " 1980). Cooper & Co.. San Francisco (1985). '"C^everaTDWRDTestsnave indicated 4. AWWA Standard for Asbestos-Cement 25. Bars. C.F. )r.& Mesmer, R.E.The Hydrol corrosion of new galvanized piping can Pressure Pipe. 4 in. Through 24 in., for ysis of Cations. Wiiey-Interscierice, probably be inhibited by the formation of a passivating coating of zinc hydroxycarbonate, which is generally more insoluble than zinc orthophosphate. Therefore, the result of the zinc solubility modeling presented in this paper is applicable to Water and Other Liquids. AWWA C40077, Rev. of C400-75, AWWA, Denver, Colo. (1977). 5. Definition* of Aggressive Waters. Cer tain-Teed Products Corp., Valley Forge, Pa. (ca. 1978). 6. Lanceuer, W.F. The Analytical Control New York (1976). 26. Plummer. L.N. et al. WATEQF: A FOR TRAN IV Version of WATEQ. A Com puter Program forCnlculating Chemical Equilibrium of Natural Waters. USGS Water-Resources Investigations 76-13 (Dec. 1976). that situation as well. of Anti-Corrosion Water Treatment. 27. Loewenthal. R.E. & Marais, G.R. Carbo The levels of. orthophosphate neces Jour. AWWA, 28:10:1500 (Oct. 1936). nate Chemistry of Aquatic Systems: sary to protect other types of piping 7. Larson. T.E. & Busweu, A.M. Calcium Theory and Application. Ann Arbor material, such as ductile iron, have not been studied by DWRD. Detailed studies of lead and galvanized pipe solubility will be reported elsewhere. In final summary, the.results, of field, pilot-plant, anH^theoreticaT^UidiessKow Carbonate Saturation Index and Alka linity Interpretation. Jour. AWWA, 34:11:1667 (Nov. 1942). 8. Fair, G.M. etal. Water and Wastewater Engineering. Vol. 2. John Wiley & Sons. . Inc.. New York (1968). 9. Hem. ].D. Study and Interpretation of the Science. Ann Arbor, Mich. (1976). 28. MERRtLL, D.T. & Sanks, R.L. Corrosion Control by Deposition of CaCO) Films. AWWA. Denver. Colo. (1978). 29. School M.R. Unpublished WATSPEC program modifications. DWRD. USEPA Cincinnati, Ohio (1979). tharash^os-cenfgntpipgcanbedurable Chemical Characteristics of Natural 30. Wicley. T.M. WATSPEC: A Computer amT'ehsure sa[eseryiceTlpfovTcle3~TK/it Water. USGS Water-Supply, 1473 (2nd Program for Determining the Equili cerfaTn~consTrgTn{s~areplaced ojt Jthe ed.. 1970). brium Speciation of Aqueous Solutions. water~quplitjeslnJaeJjaDSBflCiM- These constraints are a combination of compli cated chemical factors that can be pre dicted by considering the total water chemistry of the system. The term "aggressive water," when it is applierLlP 10. Troxell. G.E. et al. Composition and Properties of Concrete. McGraw-Hill Book Co., New York (2nd ed.. 1968). 11. Lancmuir. D.The Geochemistry ofSome Carbonate Ground Waters in Central Pennsylvania. Gcochim. Cosmochim. Acto. 35:1023 (1971). '- British Gcomorphological Res. Group Techn. Bull. 20. Geo Abstracts Ltd.. Univ. of East Anglia. Norwich. England (1977). 31. Ingle, S.E. et al. A User's Guide for REDEQL.EPA. A Computer Program for Chemical Equilibria in Aqueous Sys ^^mpe7nee3sTBBereclei[inedlnJhtxif 12. Jacobson, R.L. & Langmuir. D. Dissocia tems. EPA-600/3-78-024 (Feb. 1978). fnesejgtTuJ^srjterSTielH^tydiesMcom- tions Constants of Calcite and CaHCOj" 32. Ingle, S.E. et ai. REDEQL.EPAK Aque passing pipe inspections-and complete from 0 to 50C. Gcochim. Cosmochim. ous Chemical Equilibrium Computer water analyses are needed to improve the definition of truly aggressive waters. Acto. 38:301 (1974). 13. Christ, C.L. et al. Stabilities of Calcite and Aragonite. Jour. Hes. USGS, 2:2:175 Program. Marine and Freshwater Ecol ogy Branch. Corvatlis Envir. Res. Lab.. Corvallis. Qre. (1979). Acknowledgments The nuthors wish to thank William Mueller, Carl Shadix, Lou Trombly, Jim (Mar.-Apr. 1974). 14. Suarez, D.L. Ion Activity Products of . Calcium Carbonate in Waters Below the Root Zone. Soil ScL Soc. Am. Jour., 33. Wigley, T.M.L. ion Pairing and Water Quality Measurements. Con. Jour. Earth Sci.. 8:468 (1971). 34. Hem. J.D. Chemistry and Occurrence of Caldwell, Vern Maxon, and William 41:310 (1977). Cadmium and Zinc in Surface Waterand Cushingberry for theiranalytical chemis 15. WiecherS. H.N.S. El' al. I.ime Treatment Groundwater. Water. Res.. 8:3:681 ()un. try support, plus Robert Canter and Ken v of Wastewater:Development and Appli neth Thomas for assistance with the A-C cation of a Simple Graphical Technique coupon studies. They also wish to thank Gary Logsdon and James Symons of USEPA DWRD, Carol H. Tate of JMM Consulting Engineers, and Sara E. Ingle, for Predicting the Chemical Composi tion of Lime-Treated Secondary Efflu ent. Prog. Water Tcchnol.. 12:347 (I960). 10. Deer, W.A. etai- An Introduction to (lie Rock Forming Minerals. Longman formerly of the Department of Chemistry, Group Limited. London (1966). 1972). 35. ZiRlxo. A. & Healy, M.L. Inorganic Zinc Complexes in Seawater. Limmd. Occonog.. 15:6:956 (Nov. 1970). 36. ZiRiNO. A. & Yamamoto, S. A pHDependent Model for the Chemical Spe ciation of Copper. Zinc, Cadmium, and Lead in Seawater. Limnol. Oceanog.. University of Minnesota at Duluth, for 17. Kr.wskopk, K.B. Introduction to Geo ' 17:5:661 (Sep. 1972).. thoughtful review of this manuscript. chemistry. McGraw-Ilill Book Co., New 37. Mann, AAV. & Deutschek. R.L. Solution 650 RESEARCH AND TECHNOLOGY JOURNAL AWWA CAPCO JEN 0004820 Geochemistry of Lead and Zinc in Water Containing Carbonate. Sulphate and eratures. Acto Chem. Scond.. A26:477 (1974). tryofWnterSupply.TreatmentandDis- * tribution (Alan J. Rubin, editor). Ann Chloride Ions. Chem. GeoI., 29:2!)3 58. Cohsaro, G. & Sutherland, S. The For Arbon Science, Ann Arbor, Mifch (1980). . mation and Behaviour of Hydroxylapat- (1973). 38. Nkiauii, J.O. Solubility Equilibrium Con ite. Principles and Applications of 76. Williams, D.B. Dechlorination Linked to . stant of n-Hopeite. Ccochim. Cosino- Water Chemistry. John Wiley and Suns, Corrosion in Water Distribution Sys chiin. Actu, 37:2357, (Apr. 1973). New York. (1967). tems. Water and Sewage Works, (Mar. 39. Schindler. P. ET At- Loslichkeitspro- 57. Sinkankas.J. Mineralogy. Van Nostrand 1958). dukte und Freie Bildungsenthalpien von Reinhold Co., New York (1964). 77. Morgan. J.J. Chemical Equilibria and Zinkoxid, amorphem Zinkhydroxid,/]t-, 58. Hurlbut, C.S. Jr. Dana's Manual of Min Kinetic Properties of Manganese in Nat Pz-> o>- $- und <-Zinkydroxid. Helv. eralogy. John Wiley and Sons. New York ural Waters. Principles and Applicn- Chim. Actn. 47:4:982 (1964). (1971). . lions of Water Chemistry (S.U. Faust 40. SCHINDLER, P. ET AL. Loslichkeltspro- 59. Weast, R.C. CRC Handbook of Chemis and J.V.' Hunter, editors). John Wiley dukte und Freie Bildungsenthalpien von try and Physics. CRC Press, Cleveland, and Sons. New York (1967). ZnCOj und ZN5(0H)B(C03)2 bei 25. Ohio (1975). 78. Hem, J.D. Increased Oxidation Rate of Helv. Chim. Acto, 52:8:2327 (1969). 60. Lane. R.W. et Al-The Effect of pH on the Manganese Ions in Contact With Felds 41. Bilikski, H. ET AL. Determination of the Silicate Treatment of Hot Water in Gal par Grains. USGS Prof. Paper 475-C Stability Constants of Some Hydroxo vanized Piping. Jour. AWWA. 69:8:457 (1963). and Carbonate Complexes of Pb (II). Cu (Aug. 1977). 79. Hem, J.D. Redox Processes at Surfaces of (II). Cd (II). and Zn (II) in Dilute Solu 61. Legcett, G.E. Interaction of Silicic Acid Manganese Oxide and Their Effects oh tions by Anodic Stripping Voltammetry with Copper and Zinc and Chemical Aqueous Metal Ions. Chem. Geo and Differential Pulse Polarography. Changes of the Precipitates with Aging. 21:199 (1978). Anol. Chim. Acto. 84:157 (1976). Soil Sci. Soc. Am. Jour.. 42:263 (1978). 80. Kato, K. Behavior of Dissolved Silica in 42. Smith. K.M. & Martei.l, A.E. Critical 62. Cilmour. |.T. & Kittrick. J.A. Solubility Connection with Oxidation-Reduction Stability Constants: V. 4, Inorganic and Equilibria of Zinc in a Flooded Soil. Cycle in Lake Water. Geochem. Jour., Complexes. Plenum Press. New York Soil Sci. Soc. Am. Jour.. 43:899 (1979). 3:07 (1969). (1976). 63. Lehrman, L. & SciiULDENKR. H.L.The Role 81. Porter, R.A.& Weber, W.J.jR..The Inter 43. Patterson, J.R. etal. Carbonate Precipi of Sodium Silicate in Inhibiting Corro action of Silicic Acid with Iron (III) and tation for Heavy Metals Pollutants. sion by Film Formation on Water Piping. Uranyl Ions in Dilute Aqueous Solution. Jour. WPCF, 49:2397 (1971). Jour. AWWA. 43:3:175 (Mar. 1951). Jour. Inorg. Nucl. Chem..33:2443 (1971). 44. Larson, T.E. Corrosion by Domestic 64. Seattle Internal Corrosion Study. Phase 82. Olson, L.L. The Interaction of Fe (111) Waters. Bull. 59. Illinois State Water II Rept. Kennedy Engineers. Inc., with Si(OH),. Jour. Inorg. Nucl. Chem., Survey, Urbana, 111. (1977). Tacoma, Wash. (1977). , 35:1977 (1973). 45. Mah. M. & Boatman, E.S. Scanning and 65. X-ray analyses, performed courtesy of 83. Dart. J.J. & Folly, P.D. Silicate as Fe. Mn Transmission Electron Microscopy of the Johns.-Manville Corp., Denver, Colo. Deposition Preventative in Distribution New and Used Asbestos-Cement Pipe (1974,1979). Systems. Jour. AWWA. 64:4:244 (Apr. Utilized in the Distribution of Water 66. Courchene. ].E. & Kirmever. G.]. Seattle 1972). Supplies. Scanning Electron Micros Internal Corrosion Control Plan-- 84. Hoover, T.B. Inorganic Species in Water: copy. 1:85 (1978). ' Summary Rept. Proc.. Ann. Conf. Ecological Significance and Analytical 46. Schock. M.R. Response of Lead Solubil AWWA. Atlantic City. N.J. (1978). Needs (A Literature Review). EPA- ity to Dissolved Carbonate in Drinking 67. Hem, J.D. Equilibrium Chemistry of Iron 690/3-78-064 (Jul. 1978). Water. Jour. AWWA. 72:12:704 (1980). in Ground Water. Principles and Appli 85. Quality Criteria for Water. USEPA, Erratum, p. 36 (Mar. 1981). cations of Water Chemistry. John Wiley Washington. D.C. (Jul. 1976). 47. Goloshchapov, M.V. S Filatova, T.N. and Sons. New York (1976). 86. Weatherly, A.H. et al. Zinc Pollution The P2Os-ZnO-H20 System. Russion 68. Hem. J.D. & Cropff.r, W.H. Survey of and the Ecology of the Freshwater Jour, of Inorg. Chem.. 14:3:424 (1969). Ferrous-Ferric Chemical Equilibria and Environment: Zinc in the Environment. :. Clark. P. et AL. Unpublished Data. Redox Potentials. USGS Water Supply Part I. Ecological Cycling (J.O. Nriagu. DWRD. USEPA. Cincinnati. Ohio (1979). Paper 1459-A (1959). editor). Wiley-lnterscience. New York 49. Van Wazkr, J.R. Phosphorus and Its 69. Lancmuir. D. Geochemistry of Iron in a (1980). Compounds, lntcrscience Publishers. Coastal-Plain Ground Water of the 87. Brown. V.M. ET AL. Aspects of Water Inc., New York (1958). Camden, New Jersey. Area. USGS Prof. Quality and the Toxicity of Copper to 50. Brown, W.E. Solubilities of Phosphates Paper 650-C. (1969). Rainbow Trout Water. Reseorch, 8:797 and Other Sparingly, Soluble Com 70. Langmuir. D. & Whittemore. D.O. Varia (1974). pounds. Environmental Phosphorus tions in the Stability of Precipitated Fer 88. Pagenkopf, G.K. Zinc Speciation and Handbook. Wiley-lnterscience. New ric Hydroxides. Nonequilibrium Sys Toxicity to Fishes, Toxicity to Biota of York (1971). tems in Natural Water Chemistry (J.D. Metal Forms in Natural Water. Proc. 51. McDowell, H. et al. Solubility of Cas(PO.)jOH in the System GA(OH)2 Hem. editor). ACS. Washington. D.C. Workshop. Duluth. Minn. (Oct. 1975). (1971). 89. Smith. M.J. & Heath. A.G. Acute Toxic -HaP04-H20 at 5. 15. 25. and 37C. 71. Whittemore, D.O. & Langmuir. D. The ity of Copper, Chromate. Zinc, and Cya Jour. Hes. NBS. 81:2&3:273. (Mar.-Jun. 1977). Solubility of Ferric Oxyhydroxides in Natural Waters. Ground Woter. 13:4 nide to Freshwater Fish: Effect of Different Temperatures. Bull. Envir. 52. Moreno. E.C. et al. Preparation and Sol (Jul.-Aug. 1975). Contom. Toxicol.. 22:113 (1979). ubility of Hydroxyapatite. Jour. Res. 72. Fkoninc. M.H. et al. An Improved 90. Chapman. G.A. Effects of Continuous NBS. 72A:6:773 (Nov.-Dee. 1968). Method for Calculation of Potential-pH Zinc Exposure on Sockeye Salmon Dur 53. Nancollas, C.H. et al. Calcium Phos- Diagrams of Melal-Ion-Water Systems ing Adult-to-Smolt Freshwater Resi phates-Specintion. Solubility, and Kinetic Considerations. Ch. 23 in Chem by Computer. Corrosion Sci.. 16:371 (1976). dency. Trons. Am. Fish Sci.. 107:6:828 (Jun. 1978). ical Modeling in Aqueous Systems. ACS 73. Dodsma. J. etal. The Influence of Sulfate 91. Macnusox. V.R. etal. Relationships of Symposium Scries 93. ACS. Washing "Ions on the Formation of Iron (III) Activities of Metal-Ligand Species to ton. D.C. (1979). Oxides. Jour. Inorg. Nucl. Chem.. Aquatic Toxicity. Chemical Modeling in 54. Lundackr Madsen, H.E. Ionic Concen 41:1565 (1979). trations in Calcium Phosphate Solu 74. Ghosh, M.M. Oxygenation of Ferrous Aqueous Systems (Everett A. Jenne. editor). ACS Symposium Series 93. tions. II: The Solubility of Hydroxyln- Iron (II) in Highly Buffered Waters. ACS. Washington. D.C. (1970). patile in Water or Salt Solutions at Aqueous-Environmental Chemistry of 37C. Acto Chem. Scond., A29:745 Metals (Alan ). Rubin, editor). Ann (1975). Arbon Science. Ann Arbor, Mich. Michael R. Schock is a research chemist with ',5. Jervoe, P. & I.undaokrMauskn.H.E.Cal (1974). . and Ralph W. Buelow is retired from the Drink cium Phosphates with Apatite. Struc 75. Kat/.kneiaon, E. it At- Inactivation of ing Water Research Division. US Environmen ture I: Precipitation at Different Temp Viruses and Bacteria by Ozone. ChAnis- tal protection Agency. Cincinnati. OH 45263. DECEMBER 1981 M.R. SCHOCK & R.W. BUELOW 651 CAPCO JEN 0004.821