Division of Agricultural Sciences UNIVERSITY OF CALIFORNIA Chemistry of Lime in Ammoniated Waters for SEALING CONCRETE PIPELINES L. D. DONEEN K. K. TANJI CALIFORNIA AGRICULTURAL EXPERIMENT STATION BULLETIN 841 Repairing cracks in concrete pipelines used for irrigation water is expensive. The pipe must be excavated and patched from the outside or sealed on the inside if entry can be gained. A less expensive method involves the possibility of treating the flowing irrigation water with ammonia and other amendments so that lime will precipitate, adhere to the concrete, and seal the cracks. This bulletin is intended for irrigationists, agriculturists, engineers, concrete-pipe manu- facturers, and any others concerned with large-scale irrigation operations and the repair of pipelines. It reports results of laboratory studies on California waters and discusses the chemistry of lime formation and the nature of the precipitates. In the Appendix you will find • A technical discussion of the theory involved in precipitation and crystal- lization • A discussion of the possibility of predicting lime precipitation qualitatively and quantitatively • A table containing analyses of 83 representative California waters, against which a given water may be compared with respect to its possible capacity for precipitating lime JUNE, 1969 THE AUTHORS: L. D. Doneen is Professor of Water Science, and Irrigationist in the Experi- ment Station, Davis. K. K. Tanji is Associate Specialist, Department of Water Science and Engin- eering, Davis. In brief here are some of the results of the experiments reported : • Ammoniation of waters containing similar concentrations of Ca ++ and HCOl, but varying in Mg ++ content, produced different levels of lime precipitation. Magnesium has an inhibiting effect on lime precipitation, depending on Ca ++ content of the water. • Coprecipitation of Mg ++ with CaC0 3 occurred in measurable amounts in some waters, but this was not associated with Mg ++ content, Mg:Ca ratio, or other parameters. When Mg ++ precipitated in significant amounts, how- ever, it was related to calculated Mg(OH) 2 precipitation. The precipitation of Mg(OH) 2 is a separate factor from the inhibiting effect of Mg. • Common-ion Ca ++ had a strong effect on lime precipitation in waters con- taining similar concentrations of HCOl and Mg ++ but varying in Ca ++ content. Lime precipitation was greatly increased by addition of common- ion Ca ++ to ammoniated waters containing adequate amounts of HCOl. The inhibiting effect of Mg was partly overcome by adding Ca amendments. • Waters deficient in Ca ++ and HCOl were reconstituted with Ca ++ and HCOl amendments to increase precipitation. Results show considerable promise for field application. • Under magnification, a variety of crystalline forms was found in the precipi- tates, with calcite, vaterite, and aragonite predominant. All the crystals adhered strongly to glass flasks in the laboratory. • For most waters, 100 ppm NH 3 are sufficient to promote precipitation. • Application of amendments at the same point at which NH 3 is added is not recommended. Under field conditions, it will be necessary to introduce them separately, and with proper spacing, so that they are completely dissolved and mixed with the flowing water before it reaches the next injection site. ABSTRACT The sealing of leaky concrete pipelines involves the precip- itation of crystalline lime (CaC0 3 ) from flowing ammoniated waters. Many waters are not suitable for lime precipitation, but can be corrected by the addition of amendments. This study reports on the forms of lime, rate of crystallization, buffer capacity, common-ion effects, coprecipitation, and inhibition. A laboratory method was developed to test waters for lime precipitation, to determine use of amendments, and to predict the quantity of precipitates and sealing action. The results have practical application to field conditions. CHEMISTRY OF LIME IN AMMOMATED WATERS FOR SEALING CONCRETE PIPELINES 1 2 INTRODUCTION Sealing leaks in monolithic concrete pipelines commonly used for distribu- tion of water is expensive. Leaks are generally repaired by excavating to the leak and either patching with a sealant or, if entry into the pipe is possible, sealing from the inside. A more eco- nomical method would be to treat flow- ing irrigation waters so that they would precipitate lime (CaC0 3 ) in pipelines. The Chowchilla Water District in California sealed 159 out of 171 leaks in a three-mile length of 30- and 36- inch monolithic concrete pipelines by injecting anhydrous NH 3 into certain well waters flowing through the pipe (Eastman, 1966). Lime precipitating from these ammoniated waters adhered strongly to concrete surfaces, particu- larly within cracks and splits in the line. Preliminary tests conducted in the Chowchilla Water District, the Solano Irrigation District, and the Depart- ment of Water Science and Engineering at Davis indicated that many waters failed to precipitate lime selectively, and that their chemical composition and ammonia concentration are critical factors in such precipitation. Well-documented measures are avail- able for preventing lime precipitation and crustations in water lines of boilers and heaters, in flotation, in stabiliza- tion of sols, and in other industrial processes. The fertilizing of crops by injection of ammonia into water dis- tributed through pipelines has been recognized as a hazardous procedure (Reitemeir and Buehrer, 1940) because lime crustations reduce flow and clog sprinkler jets. Many unknown factors are involved in the formation of crystalline lime with adherence properties. The present study was conducted in an attempt to deter- mine the conditions favorable for such precipitation. The first phase of the re- search involved a systematic examina- tion of the effect of various dissolved salts, ammonia concentrations, and ambient conditions on lime precipita- tion. In order to separate the limiting 1 Submitted for publication November 9, 1967. 2 This study was supported by the Department of the Interior, Bureau of Reclamation, and the University of California Water Resources Center. [4] variables, pure salt solutions were used in preference to natural waters. The second phase was concerned with the chemical characteristics of surface and ground water that are suitable for pre- cipitation of lime. Waters with unsuit- able chemical characteristics were re- constituted to the desired composition by addition of amendments, and then treated with ammonia. An Appendix, beginning on page 33, discusses chemical equilibria and crys- tallization, and is intended to serve as a reference for interpretation and evalu- ation of the results obtained in this study. LIME PRECIPITATION IN SALT SOLUTIONS The first phase of the investigation was concerned with precipitation of lime in ammoniated salt solutions. Single and mixed salt solutions were examined to determine the effects of a particular solute or combination of solutes and NH 3 concentrations on lime formation. Prior to these studies, how- ever, certain laboratory techniques and procedures were developed. The influence of temperature, buffer capac- ity, common ions, inhibiting effect of Mg, rate of crystallization, and crys- talline species was investigated. EXPERIMENTAL TECHNIQUES AND CONDITIONS Analytical Methods The study of precipitation from pure salt solutions was conducted under laboratory conditions and with glass containers. Initially, solutions prepared from chemically pure or reagent-grade salts were studied in preference to natural waters because the latter con- tain dissolved materials varying in com- position and concentration. By using known salt solutions, we could control the number of variables and study their effects comprehensively. Laboratory glassware allowed direct microscopic observations of the precipi- tants. From the standpoints of adhe- sion and presence of colloidal impurities which may act as seeds for crystalliza- tion, the relatively rough surface of concrete pipe should be more conducive to precipitation than was the smooth glassware surface. t Standard analytical methods (Rich- ards, 1954; Orland, 1965) were em- ployed. Calcium and magnesium were determined by EDTA (ethylenediamine tetraacetate) titration or by flame pho- tometry. The flame photometric meth- od was also used for sodium. The carbonates and hydroxide were titrated with standard acid, and pH was meas- ured with an expanded-scale pH meter. Sulfate was determined turbidimetric- ally; chloride, by titration with silver nitrate. The amount of calcium precipitated from solutions was determined by dif- ferences between initial and final Ca ++ values. In general the calcium precipi- tated in ammoniated waters is in the CaC0 3 form, and only trace amounts may precipitate as other Ca salts in natural waters. The final analysis for Ca" 1-1 " was made on supernatant solu- 5] 0.8 Temperature (° C) Fig. 1. Effect of temperature on the precipi- tation of lime in 1.56 m.e. per liter Ca(HC0 3 ) 2 solution treated with 100 ppm NH 3 . tions obtained by filtering the ammoni- ated solution through No. 41 Whatman filter paper. Stock solutions of aqueous NH 3 (NH 4 OH) were approximately 1 N in strength, to initiate crystallization and to simulate the injections of anhy- drous NH 3 in water. Stock Ca(HC0 3 ) 2 salt solutions, =b 5 m.e. per liter, were prepared by dissolving CaC0 3 under 0.2 atmosphere of C0 2 . Effect of Temperature Laboratory room temperatures were about 22 =b 2°C. The temperature effect on lime precipitation in an ammoniated Ca(HC0 3 ) 2 salt solution over a six-hour period is shown in figure 1. Increases in temperature lower the solubility of NH 3 and CO2 in water. Conversely, increases in temperature cause a greater dissoci- ation of aqueous NH 3 into OH - and NHt and a greater ionization of HCO^ into CO! and H + . The net result of rising temperature was an increase in precipitation. Buffer Capacity and Spontaneous Precipitation A series of Ca(HC0 3 )2 salt solutions, ranging in concentration from 0.5 to 5.3 m.e. per liter, were titrated with NH 3 solution in a closed system, with continuous stirring. The pH titration curves for these buffered HCOl solu- tions are presented in figure 2. Small additions of NH 3 to the 0.5 and 1.3 m.e. per liter salt solutions caused rapid changes in pH up to an NH 3 concen- tration of about 100 ppm. Thereafter, further titration with this weak base resulted in only small changes in pH. As the HCOl concentration increases, its buffer capacity increases, and more NH 3 is required to obtain a unit pH change than is required in less concen- trated solutions. Spontaneous precipi- tation was detected when the pH of the solutions was raised to 9 or slightly higher. The ammoniated solutions at this pH range are at a state of super- saturation (Appendix fig. 1) with re- spect to CaC0 3 . Nucleation and crystal growth can probably occur at lower pH, but not spontaneously. Rate of Crystallization Ammonia was applied at rates of 50, 75, 100, 150, and 200 ppm to a salt solu- tion containing 1.56 m.e. per liter of Ca(HC0 3 ) 2 . Following the addition of NH 3 , the solutions were shaken and allowed to stand for 2, 8, 24, 48, and 96 hours of precipitation time. The am- moniated solutions were then analyzed for Ca ++ and pH. Spontaneous precipi- tation did not occur and crystal forma- tion was relatively slow. Microscope observations (100 X magnification) of the two-hour series indicated the pre- sence of particles ranging in distribution from sparse (50 ppm NH 3 ) to numerous (200 ppm NH 3 ) , which could not be posi- tively identified. Apparently these minute particles were nuclei or crystal seeds. Similar particles were found with [6] 160 NH 3 (ppm) 320 480 640 0.956 N NhL OH (ml) Fig. 2. Buffer capacity of Ca(HC0 3 ) 2 salt solutions, and the NH 3 concentration and pH at which spontaneous precipitation takes place. the eight-hour series, except that the solution treated with 200 ppm NH 3 contained identifiable calcite and arago- nite crystals (Appendix fig. 3). With increasing precipitation time, the size of the crystals was larger for a given level of NH 3 . With increasing NH 3 con- centrations, the number of crystals increased for a given time period. Lime precipitated in the bottom and adhered strongly to the flasks. The amount of CaC0 3 precipitated for a given NH 3 concentration at dif- ferent quiescent time periods is given in figure 3. Lime precipitation was in- creased either by lengthening precipi- tation time or by increasing NH 3 con- centration. The pH of the solutions is shown in figure 4. The 48-hour pH values fell between the 24- and 96-hour series. Because of insufficient C0 3 very little CaC0 3 was formed below pH 8.5. Cyclic Precipitation and Renewal In the field, a more dynamic condition exists than that in the laboratory. Anhydrous NH 3 is introduced into water as it flows past the point of injec- tion. Because of differences in velocity distribution, the water is continuously mixed or agitated in the pipeline. Depo- sition of nuclei and crystals on previous deposition sites or on new areas occurs as ammoniated water flows past each cross-section. For example, if a J^-mile section of the distribution line is to be treated with NH 3 , the amount of water required to fill that pipeline is in pro- portion to its diameter. The time re- quired to fill or displace the water in the line at a flow rate of 1 cfs (cubic [7] 0.0 50 100 150 NHa (ppm) 20 Fig. 3. Relations between CaCOs precipitation and NH 3 concentrations for different quiescent precipitation periods. The initial concentration of Ca(HC0 3 ) 2 was 1.56 m.e. per liter. 10 - x 200 NH, (ppm) Fig. 4. The pH of ammoniated Ca(HC0 3 ) 2 solutions for different quiescent precipitation times. Initial pH was 6.80. [8] feet per second) is as follows : PIPE CAPACITY TIME TO FILL DIAMETER WHEN FULL AT 1 CFS inches CU. ft. hrs. 24 8,290 2.30 30 12,931 3.59 36 18,652 5.18 42 25,366 7.05 48 33,158 9.20 The ammoniated water may be suc- cessively displaced and released down- stream until leaks are sealed to the operator's satisfaction. To duplicate such a condition in the laboratory would require a model, and then only limited measurements and observations could be obtained without elaborate instrumentation. Instead, the following laboratory technique was used. In suc- cessive six-hour periods, lime was pre- cipitated from ammoniated salt solu- tions in a flask. At the end of each period, the solution was replaced with a fresh one. In addition, the reaction flasks were placed on a horizontal shak- ing machine set at 60 oscillations (1- inch strokes) per minute to simulate movement of water in pipelines. This experiment was continued through eight cycles of six-hour precipitations with NH 3 rates of 25, 50, 75, 100, and 200 ppm. Results are presented in figures 5 and 6. For a given NH 3 concentration (fig. 5) lime precipitation increased with number of cycles, except for the 25- ppm NH 3 treatment. The deposit of nuclei and crystals from the previous cycles tended to intensify crystalliza- tion from succeeding solutions. The critical NH 3 concentration was about 75-ppm NH 3 . The range of pH obtained from these treatments is shown in fig- ure 6, for three selected cycles. The amount of lime precipitated and the pH for the eighth cycle of this study are 100 2 4 6 No. of 6-hr precipitation-renewal cycles 10 Fig. 5. Relations between CaC0 3 precipitation and NH 3 concentration for successive cycles of 6-hour precipitation and renewal of ammoniated Ca(HC0 3 ) 2 solutions. Initial Ca(HC0 3 ) 2 concen- tration was 1.58 m.e. per liter. [9] II 10 1 « O I st cycle D 5th cycle A 8 th cycle 50 100 150 NH 3 (ppm) 200 Fig. 6. The pH of Ca(HC0 3 )2 solutions ammoniated to varying degrees and undergoing cyclic 6-hour precipitation and renewal. Initial pH of the Ca(HCOJ 2 solution was 6.89. similar to those for the 48-hour quies- cent treatment in the study on crystal- lization rate. Minute specks of barely discernible calcite were observed in the 25-ppm NH 3 flasks after eight cycles of precipi- tation. Large calcite crystals with fewer smaller ones were found in the bottom of the 50-ppm NH 3 flasks. A profuse growth of calcite and aragonite, with many smaller, single crystals, was ob- served in the 75-ppm NH 3 flasks, with some CaC0 3 adhering to the walls of the vessel. With the 100-ppm NH 3 treatment, a heavy deposit of calcite and aragonite crystals covered the entire flask up to the water level. By the eighth cycle, precipitation in 200- ppm NH 3 solutions was so dense that individual crystals were not discern- ible. The flask was opaque, and glazed over to such a degree that very little light was transmitted for microscopic observations. Procedure Adopted A rapid technique for determining data related to field conditions is desirable for facilitating large-scale studies on the effect of dissolved salts on lime pre- cipitation. The cyclic precipitation and renewal experiment approached field conditions; however, it is a time- consuming method and permits only a few salt solutions to be systematically examined within a reasonable time. In [10] view of the close approximation of data obtained from the 48-hour precipita- tion (fig. 3) with that of eight cycles of precipitation and renewal (fig. 5), the former procedure was selected. The initial pH and concentration of the HCOl ion influence the amount of NH 3 required for lime formation (fig. 2). A solution containing about 1.58 m.e. per liter of Ca(HC0 3 ) 2 was chosen as the base solution for the addition of various salts. The concentration of salt admixture to the Ca(HC0 3 )2 solution was in ratios of 0.5, 1, 2, and 4. The pH of the solution mixtures was then ad- justed to about 6.85, before ammoni- ation, by adding minute amounts of a strong base (NaOH) or acid (HC1). A 100-ml subsample of the salt solu- tion was measured into 125-ml Erlen- meyer flasks, aqueous NH 3 (NH 4 OH) from a 0.934 N stock solution was added, the solution was mixed, stop- pered, and allowed to stand under room conditions. Concentrations of ammonia were 50, 75, 100, 150, and 200 ppm. After 48 hours, the ammoniated solu- tion was filtered, and Ca ++ and pH of the filtrate were measured. The amount of lime precipitated was the difference between the initial and final Ca ++ con- tent. The precipitants were examined by microscope at 100 X magnification, for the presence of CaC0 3 and other crystalline or amorphous particles. All treatments were replicated threefold. SALT-SOLUTION STUDIES Influence of Neutral Salts on Precipitation The influence of seven different salts on lime precipitation from ammoniated Ca(HC0 3 ) 2 solutions was investigated. The neutral salts studied were NaCl, Na 2 S0 4 , and MgCl 2 , representing 1-1, 1-2, and 2-1 electrolytes, respectively. The amounts of lime precipitating for different levels of added NH 3 are shown in table 1. For a given ratio of salt added to Ca(HC0 3 ) 2 , symmetry con- centration (SC), higher NH 3 rates in- creased the amount of lime formed. Varying the concentration of neutral Na salt (NaCl, Na 2 S0 4 ) produced little or no effect on lime precipitation for comparable NH 3 additions. However, admixtures of MgCl 2 resulted in smaller amounts of lime precipitation (fig. 7), and with increasing symmetry concentration this effect became more pronounced. [ The ionic activity coefficient for di- valent ions (Ca ++ and CO! before pre- cipitation) in solutions containing 200 ppm NH 3 and at symmetry concentra- tion 4 was calculated to be 0.94 for the NaCl and 0.93 for the Na 2 S0 4 and MgCl 2 systems. Since the CI - in the NaCl system had no significant effect on lime precipitation, Mg ++ of the MgCl 2 system must be the ion that interferes with or inhibits lime forma- tion. Influence of Common Ion HCOl on Precipitation The admixture of the common ion HCOl, more correctly the CO! formed upon ammoniation (equation [11], Appendix A), resulted in greater lime precipitation than did the neutral salt admixtures (table 2). The common-ion effect may be described by referring to equation [9] (Appendix A) . When CO! 11] Table 1 EFFECT OF NEUTRAL SALTS AND VARIOUS CONCENTRATIONS OF NH 3 ON LIME PRECIPITATION AFTER 48 HOURS NH 3 added Added salt Symmetry concen- tration* NaCl Na 2 S0 4 MgCl 2 CaCOs P H CaCOs pH CaCOs pH ppm m.e./l m.e./l m.e./l 0.5 1.56t 6.84f 1.56f 6.85t 1.58t 6.84f 50 1.02 9.49 0.98 9.56 0.70 9.47 75 1.17 9.70 1.15 9.72 1.04 9.64 100 1.28 9 85 1.24 9.92 1.14 9.78 150 1.35 10.00 1.33 10.03 1.24 9.96 200 1.40 10.09 1.36 10.12 1.24 10.08 1.0 1.56f 6.85f 1.56f 6.86f 1.58f 6.85f 50 0.94 9.53 0.98 9.68 0.59 9.40 75 1.22 9.71 1.23 9.79 0.98 9.64 100 1.23 9.84 1.25 9.85 1.12 9.82 150 1.31 10.02 1.33 10.02 1.20 9.98 200 1.32 10.12 1.37 10.16 1.21 10.10 2.0 1.56f 6.86f 1.56f 6.86f 1.58f 6.87f 50 1.00 9.58 1.00 9.39 0.48 9.44 75 1.13 9.78 1.12 9.62 0.87 9.58 100 1.25 9.95 1.25 9.83 1.02 9.69 150 1.29 10.07 1.32 10.00 1.09 9.88 200 1.36 10.10 1.35 10.13 1.16 9.98 4.0 1.56f 6.87f 1.56f 6.85f 1.58t 6.88f 50 0.98 9.63 0.83 9.60 0.27 9.42 75 1.04 9.77 0.94 9.77 0.64 9.59 100 1.16 9.89 1.13 9.87 0.86 9.72 150 1.25 10.04 1.26 10.03 1.04 9.87 200 1.30 10.13 1.29 10.17 1.10 10.01 * Concentration ratio of added salt to Ca(HCOs)2. t Initial concentration and pH of Ca(HCC>3)2 solution. (or Ca ++ ) is added to a saturated lime solution, the concentration (activity) of Ca ++ (or COl) is depressed since ^caco a is a fixed value. The decrease in Ca ++ (or COl) concentration leads to a lower solubility of lime and hence greater precipitation. With increasing additions of common ion (either COl or Ca ++ ) to this solution, further de- creases in solubility would occur. In the NaHCOs series, increases in the concentration of added HCOl pro- duced no significant effect on lime precipitation for given rates of NH 3 . Contrary to the lime solution cited above, the salt solution mixture of Ca(HC0 3 )2 and NaHC0 3 was not initi- ally a saturated solution of lime. In order for lime to precipitate, NH 3 was added to convert HCOl to COl. The COl formed then reacted with Ca ++ to form lime. By increasing the concentra- tion ratio of NaHC0 3 to Ca(HC0 3 ) 2 the buffer capacity of the solution was also increased, which fortuitously counter- acted the common-ion effects in this system. The relationship between HCOl concentration and buffer capacity of salt solutions is illustrated in figure 2, where additional NH 3 was required to attain a given pH level as the HCOl concentration increased. The pH values [12] 250 NH 2 (ppm) Fig. 7. Average percentage precipitation of Ca ++ as CaCO :! for several NH :! concentrations in MgCI 2 admixtures of varying symmetry concentrations (SC). for the NaHCOs system (table 2), as compared with the NaCl system (table 1), were found to be lower, substanti- ating the buffer action of added HCOl. Further increases in NaHC0 3 would probably result in a decrease in lime precipitation and pH as the buffer action overcame the common-ion effect. Although the common-ion effect was subdued to a certain extent in the NaHC0 3 series, precipitation of lime was greater when compared with the NaCl and Na 2 S0 4 systems (fig. 8) . In contrast, lime precipitation was reduced as the concentration ratio of Mg(HC0 3 ) 2 was raised (table 2). The buffer capacity of Ca(HC0 3 )2 solutions was increased with additions of HCOl as was the case for the NaHC0 3 sys- tem. With the exception of the 0.5 ratio [13] Table 2 EFFECT OF THE COMMON ION HCO^ AND VARIOUS CONCENTRATIONS OF NH 3 ON LIME PRECIPITATION AFTER 48 HOURS NH 3 added Added salt Symmetry concentration* NaHCOs Mg(HC0 3 ) 2 CaC0 3 pH CaCOa pH 0.5 ppm 50 75 100 150 200 m.e./l 1.57f 1.15 1.26 1.39 1.42 1.43 6.86f 9.38 9.57 9.73 9.87 10.01 m.e./l 1.59f 0.98 1.13 1.20 1.27 1.30 6.84f 9.39 9.68 9.83 10.01 10.10 1.0 50 75 100 150 200 1.57f 1.02 1.19 1.32 1.39 1.41 6.86f 9.32 9.51 9.67 9.86 9.98 1.59f 0.99 1.07 1.17 1.28 1.29 6.85f 9.20 9.49 9.67 9.86 10.00 2.0 50 75 100 150 200 1.57! 0.92 1.25 1.36 1.42 1.43 6.86t 9.20 9.46 9.66 9.83 9.94 1.59f 0.88 1.02 1.14 1.18 1.29 6.84f 9.02 9.25 9.44 9.67 9.85 4.0 50 75 100 150 200 1.57f 0.95 1.30 1.39 1.42 1.43 6.86f 9.12 9.39 9.52 9.76 9.86 1.59f 0.74 0.91 1.02 1.14 1.19 6.86f 8.67 8.92 9.13 9.38 9.60 * Concentration ratio of added salt to Ca(HC0 3 )2. t Initial concentration and pH of Ca(HC0 3 )2 solution. solutions, pH values were lower in Mg(HC0 3 ) 2 than in MgCl 2 systems for identical rates of NH 3 . In addition to the buffer effect, Mg apparently inhib- ited lime formation, as was the case with the MgCl 2 system. The combined effects of buffer action and inhibition by Mg are shown in figure 9. With the lowest concentration of Mg(HC0 3 )2, percentage of precipitation was less than with NaHC0 3 , and with increasing Mg(HC0 3 ) 2 concentrations precipita- tion was reduced. Apparently the bene- ficial effects of common ion were nearly nullified by the buffer action and Mg interference, as shown by the curves [ for MgCl 2 and Mg(HC0 3 ) 2 . The exact mechanism (s) by which Mg inhibited lime formation is not known. The degree of coprecipitation of Mg ++ with CaC0 3 was relatively insignificant (0.1 m.e. or less per liter), with Mg ++ concentrations in the Ca(HC0 3 ) 2 solutions ranging from about 0.8 to 6.4 m.e. per liter. In a pre- vious study (Doneen, 1964), Mg ++ coprecipitated in significant amounts in Ca(HC0 3 ) 2 solutions only when the Mg ++ concentrations exceeded 20 m.e. per liter. Several views on the inhibit- ing mechanisms of Mg are found in the literature (Brooks, Clark, and Thurs- 14] c U i_ 4. $ Water + CaCh. [25] ratio is unaffected and only the HC0 3 : Ca ratio is increased. The added HCOl contributes to the buffer capacity of the water. Moreover, comparison should be made between Ca ++ and CO!, de- rived primarily through ammoniation, and not with HCOl. Thus the magni- tude of the common-ion HCOl effect is generally less than that of the common ion Ca++. Reconstituted Waters Waters lacking sufficient Ca ++ for maximum precipitation (table 8) re- ceived either gypsum or CaCl 2 as an amendment. The dual effect of common ion Ca ++ and reduction in Mg:Ca ratio increased precipitation substantially. Certain waters that precipitated appre- ciable amounts of lime (for example waters 37, 50, and 62) were also treated with calcium additives, and still larger amounts of lime were produced. Waters deficient in HCOl were re- constituted with NaHC0 3 salt solu- tions. The impact of common ion HCOl was less than that of the Ca ++ ion (table 9). Admixtures of NaHC0 3 and CaS0 4 amendments were applied to waters containing limiting concentra- tions of both HCOl and Ca++. The addition of only NaHC0 3 to water 5 resulted in a small amount of precipi- tate, but when both NaHC0 3 and CaS0 4 were added (water 54), the amount of lime was greatly increased (table 9) . Similar results were obtained with water 16 and its reconstituted waters, 19 and 67. If Ca ++ is not limit- ing (water 17), the addition of HCOl provides substantial gains in the amount of precipitation. Table 9 PRECIPITATION OF LIME IN ORIGINAL AND RECONSTITUTED (WITH BICARBONATE AND BICARBONATE-CALCIUM AMENDMENTS) AMMONIATED WATERS Water no. Water composition* CaC03 precipitation in water treated with NH3 at: Ca++ HCO; Mg:Ca ratio 50 ppm 100 ppm 150 ppm 5 m.e./l 0.4 0.4 m.e./l 0.6 2.1 0.4 0.5 Ib/acre-ft 7 33 Ib/acre-ft 7 38 Ib/acre-ft 7 8f 44 5 0.4 2.3 0.6 2.5 0.4 0.1 7 268 7 279 7 54 1 284 7 0.4 2.5 0.6 2.8 2.3 0.5 4 271 8 284 8 58| 295 16 0.8 0.9 1.1 4.1 0.5 0.5 7 98 20 103 27 19f 105 16 0.8 3.3 1.1 3.7 0.5 <0.1 7 412 20 426 27 67$... 428 31 1.3 2.1 2.3 3.2 0.4 0.3 87 231 97 248 114 49J 252 17 2.2 2.3 0.7 3.3 0.1 <0.1 109 286 131 292 152 55f 296 * Complete analyses given in Appendix C. t Water + NaHCOa. | Water + NaHCOs + CaS04. [26] Nature of Precipitates The precipitates formed in the test waters were examined under 100 X magnification. A variety of crystalline species of lime was produced. Vaterite, calcite, and aragonite (Appendix fig. 3) were the most common forms. Infre- quently, CaCCV6H 2 and CaC0 3 -H 2 were observed in ammoniated waters. For many waters yielding relatively small amounts of lime because of high Mg ++ content, the crystalline species observed were ill-defined and often appeared as globular masses of a col- loidal or gel-like nature. The pre- dominant crystalline species produced by solutes or amendments were : Amendments NaHC0 3 NaHC0 3 + CaCl 2 NaHC0 3 + CaS0 4 CaS0 4 (low Mg++) CaS0 4 (moderate Mg ++ ) CaS0 4 (high Mg++) CaCl 2 CaCl 2 (Mg++) Predominant Crystalline Species aragonite aragonite, calcite aragonite, calcite calcite vaterite, calcite aragonite, vaterite calcite, aragonite vaterite, calcite Although exceptions were noted, these observations are in general agreement with those of other investigators (Johnston, Merwin, and Williamson, 1916; Brooks, Clark, and Thurston, 1950; Buehrer and Reitemeir, 1940). The rate of nucleation and crystal growth in the test waters was generally slow, and all of the readily identifiable crystalline forms of lime adhered ten- aciously to the surface of the glassware. But where precipitation was instantan- eous, upon ammoniation, the precipi- tants were of colloidal sizes and ap- peared as minute dots at 100X mag- nification. Under these spontaneous precipitation conditions the precipi- tates tended to be softer and loose, especially with some of the highest concentrations of gypsum amendments (waters 79, 81, 82, 83, Appendix C). (Suspended and adhered precipitates are discussed in more detail on p. 31.) Table 10 COPRECIPITATION OF Mg++ IN AMMONIATED WATERS, CALCULATED Mg(OH) 2 PRECIPITATION, AND OTHER PARAMETERS Mg""" precipitation in water treated with NH3 at: Water treated with 100 ppm NH3 Water no. 50 ppm 100 ppm 150 ppm Calc. Mg(OH) 2 Initial Mg^ Initial Mg:Ca ratio Final Mg:Ca ratio Super- saturation ratio Mg:Ca precipita- tion ratio 83 m.e./l 0.70 0.57 0.71 0.63 0.35 0.37 0.40 0.26 0.30 0.29 0.25 0.20 0.13 0.14 0.21 0.20 0.18 m.e./l 0.74 0.75 0.66 0.60 0.52 0.52 0.44 0.39 0.33 0.33 0.32 0.27 0.25 0.19 0.19 0.18 0.17 m.e./l 0.74 0.74 0.68 0.60 0.50 0.37 0.52 0.34 0.32 0.31 0.34 0.25 0.27 0.20 0.11 0.17 0.22 m.e./l 0.51 0.74 0.77 0.50 0.50 0.74 0.75 0.72 0.29 0.33 0.44 0.49 0.50 0.30 0.31 0.43 0.50 m.e./l 2.9 5.6 4.8 2.7 2.8 5.5 5.4 5.5 0.6 0.7 1.3 2.4 2.6 0.6 0.7 1.3 2.3 0.4 1.5 1.5 0.5 1.0 1.6 2.8 2.0 0.2 0.7 0.5 0.6 1.2 0.1 0.3 0.7 0.8 0.4 7.6 14.9 0.9 3.5 6.1 6.1 7.2 0.8 0.8 2.6 1.6 3.0 0.5 1.6 2.3 5.8 11.7 18.2 13.4 10.2 13.0 12.9 13.8 17.7 86 5.4 12.7 8.9 11.2 10.0 8.8 10.4 9.5 0.22 78 68 0.24 23 81 0.19 64 0.23 69 20 37 40 65 0.19 61 0.14 71 46 59 15 75 11 50 18 79 05 49 10 36 14 63 07 [27] Coprecipitation of Magnesium In general, only trace amounts of Mg++ were precipitated from ammoniated waters, with the exception of the waters listed in table 10. The mechan- ism (s) by which Mg ++ coprecipitates in these waters is not fully known. The degree of coprecipitation was not re- lated to initial Mg ++ concentration or initial Mg:Ca ratio. Only trace amounts precipitated from other waters having similar or higher Mg ++ contents or Mg:Ca ratios. Magnesium precipita- tion is also not related to the final Mg:Ca ratio — i.e., after lime precipita- tion — nor is it related to the initial or final pH of the waters. The supersatur- ation ratio (defined by equation [19], Appendix A) is related to rate of nucle- ation and crystal growth — the higher its value the greater the chance for Mg ++ being occluded or entrapped in the precipitating lime. Apparently, co- precipitation is not significantly cor- related to the supersaturation ratio (table 10) or the Mg:Ca precipitation ratio. ACTION OF AMENDMENTS UNDER SIMULATED FIELD CONDITIONS In the procedures adopted to test natural and reconstituted waters, CaCl 2 and NaHC0 3 were added as liquids and CaS0 4 was added as a solid, but dis- solved before ammoniation, to simu- late introduction of amendments up- stream in relation to the NH 3 injection site. The tests were carried out under a 48-hour quiescent precipitation period. The effect of applying CaS0 4 and/or NaHC03 in the dry state, with gypsum applicators and other similar devices, at the site of NH 3 injection or just up- stream was explored. These and other related studies were carried out under simulated field conditions, e.g., with stirring or agitation. Dissolution Rate of Gypsum The solubility of gypsum is affected by temperature, but the concentration re- quired to provide sufficient Ca ++ to precipitate lime in waters is about one- tenth its solubility. The rate of gypsum dissolution in waters is more important. Powder-form gypsum at concentrations of 3 and 5 m.e. per liter was added to C0 2 -free distilled water, and the reac- tion flask was placed on a shaker. The horizontal shaking machine was set at 60 oscillations (1-inch strokes) per min- ute to simulate agitation not too dif- ferent from flow of water through pipe- lines. The dissolution rate of gypsum was determined for 5 to 60 minutes of elapsed time by analysis for super- natant Ca ++ . The ratio of supernatant Ca ++ to added CaS0 4 provided a meas- ure of dissolution (fig. 12). Within 20 minutes, more than 95 per cent of the added gypsum was dissolved. The dif- ference in rate of dissolution between 3 and 5 m.e. per liter of added gypsum is relatively small. Gypsum in granu- lated form has a lower dissolution rate. Putah South Canal water contains sufficient HCOl and 0.4 m.e. per liter of COl but insufficient Ca ++ for sub- stantial lime precipitation. Gypsum was added to this nonammoniated water at rates of 4 and 6 m.e. per liter. The flasks were placed on a shaker and the waters were analyzed at periods from 15 minutes to 6 hours. Within 15 minutes most of the gypsum had dis- solved (table 11), and apparently solu- bilization increased only slightly for the remaining period. Calcium in the solid [28] 1.00 — I — t » i r i — ..A.. -,. A 0.95 p y^ - u 0.90 . 6 - Mt>ft— i 1 1 + 1 Saturation index Fig. 6A. Saturation index of waters, before ammoniation, and potential lime precipitation values calculated from 100-ppm NH 3 waters. 3 5 7 9 M 13 Stability index Fig. 7A. Stability index of waters, before ammoniation, and potential lime precipitation values calculated from 100-ppm NH 3 waters. saturation index of 0.0 ± 0.5 are related to large differences in potential lime precipitation, the index appeared to be unsuitable for our study. The stability index for waters, proposed by Ryznar (1944), is an empirical modification of the saturation index. According to Langelier (1936), the saturation index provides a means of ascertaining the directional tendency of the water to deposit lime, and is in no way a measure of its capacity. Ryznar (1944) points out that in some cases two waters may have an identical saturation index but widely different scale-forming tendencies. To overcome this anomaly, Ryznar suggests placing emphasis on pH s by, Stability index = 2pH s — pH, [30] A plot of potential lime precipitation values calculated from data obtained from waters treated with 100 ppm NH 3 and the stability index is presented in Appendix figure 7. A marked improvement over the saturation index is noted in curvilinear relations. Lime appears to form when the stability index is less than 9. In a more recent study by Bower et al. (1965), pH s of synthetic waters was highly correlated with the solubility of lime. Substantial correlation between pH s of waters before ammoniation, and potential lime precipitation values is evident (Appendix fig. 8). Momentary Excess, Quantitative Prediction A more quantitative measure of a water's degree of saturation with respect to [44] i 1 r i e 3 4 s Calc. lime (m.e./l) Fig. 9A. Above: Prediction of experimentally determined CaC0 3 precipitation in 100-ppm NHy waters from equation [32] and Mg inhibit- ing factor (MIF). Fig. 8A. Left: Relation between pH, of waters, before ammoniation, and potential lime pre- cipitation values obtained from 100-ppm NH 3 waters. lime was suggested by Dye (1958). The momentary excess or initial precipitation of lime is calculated by the following expression (Ca ++ X) (COT - X) = K' s X 10 1 [31] where X is the momentary excess, and the parentheses denote concentration in terms of ppm CaC0 3 . In a similar manner, Tanji and Doneen (1966a) computed lime precipitation by (Ca++ - X) Tea (COT - X) 7co 3 = K.. [32] where parentheses denote molar concentration and y is the ionic activity coef- ficient. Since equation [32] is more precise than equation [31], it was applied to the waters in this study. Equation [32] was solved for X after CO! was obtained from equation [14]. The predicted lime precipitation, X, was found to be much higher than the experimental values. This discrepancy was reduced when X was multiplied by the respective Mg inhibiting factor of the water. Substantial overall agreement between experimental and calculated values, r = 0.915, is indicated in Appendix figure 9. Discussion on Predictions Several methods for prediction of lime precipitation in ammoniated waters [45] have given varied results. These computations indicate that precise predictions are not readily obtainable. With some exceptions, the precipitation of lime from waters was carried out at room temperatures of 22 d= 2° C. The constants appearing in the equations were values applicable to 25° C. Moreover, the K sp values employed were for calcite. In the experimental waters of widely differing chemical composition, crystalline species other than calcite precipitated with solubilities greater than calcite. This may be the reason that the regression line in Appendix figure 9 is less than a 1 :1 correspondence. Furthermore, some of these calculations assumed chemical equilibrium, which may or may not have been attained over the 48-hour precipitation period. The inhibiting effect of Mg on lime precipitation is not accounted for in any of the published formulas. Adjustments were made by calculating the potential precipitation of lime from these waters by taking the product of experimental lime data and its reciprocal Mg inhibiting factor. The various qualitative indices were then evaluated against potential lime values. For the quantitative method (equation [32]), the first approximation on lime precipitation, X, was modified by the Mg factor to obtain predicted values. The Mg factor was empirically derived from limited data, but its application to experimental waters appears to be significant. The probability of overcorrection or undercorrection for Mg ++ effects in certain waters, however, should not be discounted. For more quantitative consideration, other factors should be accounted for, such as: the formation of Mg(OH) 2 ; ion pairs of CaS0 4 (Tanji and Doneen, 19666); MgC0 3 and CaC0 3 (Garrels, Thompson, and Siever, 1961); and com- plexes of Ca(HC0 3 )+ Mg(HC0 3 )+, and (MgOH)+ (Garrels et al, 1961). These computations can, however, become quite unmanageable or lengthy. An examination of widely divergent points, in Appendix figures 4 through 9, indicates that some waters recur frequently, but many do not. The reasons for their deviation from the trend are not very apparent. Because of the complexity and dynamic nature of lime chemistry in ammoni- ated waters, the qualitative methods of predicting lime precipitation are less than satisfactory. The slopes obtained from the relations between potential lime precipitation and various qualitative indices are so steep that small changes of the index result in large changes in predictions. The quantitative method pre- sented above, however, not only yields a more accurate prediction but also predicts in terms of actual lime instead of potential lime precipitation. [46] APPENDIX C SI ^ 2 O c3 I- Q g ^ as i CO "A .2 • CO a .2 s > ° Cl, l-H 1 ce ^ ^ -d 73 +■= Pi «« o3 O *-.E 08 ^Q :3 o .§ _£ a a w o H ttf # g H H c r/3 c8 £ co O 8 B a -•§ CD !7h P 0) ^ _d o 73 CO QJ o h-j ^ S ^ ± go S3 g £ "43 12 o8 a o « 08 ° CD fl 02 O « 2 ^ 08 3 v. ° Is 3 •a s 5 ^ -Q ^ 03 °^ ,d o ? cd 5 » hh .2 Ph o M S > "a 1 I I s J o8 o d * ^ d a o 73 0) 03 nj +3 Jh CD 4-> a o - rt > 'd +j + d or! 3 D s 00 CO 00 CO 01 © © © lO CO ^> nr ( ?HC» s © © © © ©©©©©©©©© © © © © © © © 0©©©©©©©© © © © ^H r-l ^ "S O ft A bH w S a tDOOlfltMOOtOCOtNlO 00 O0 -* © t^ CNI 00 -t ^ ft T3 ft -•'•/- S3 U 6.& ^ ss©§ 101<®OOMHtO(0'* ©©©(MO©©©© N NIOIO OitOO © © ^H © © t^ C3 IOhNhCStPCONN h>»whMOhh« 00 rt CO © CO t^ lr^ ^h © t-- i g © © © ©©©©©©©©© 0) O ft A 10 © *tfl © ©t~-crt©>o»o©>co 00 N * -r* © »o t^ 02 OC COrtlOrttNNNcOlC H CD W3 lO O CO NClO 10lr}tNtO©NMt^t» HNCOQCONNQC N tO t- 0C CO ft t^ 00 00 00 t>»C»00CX1000OOOO0 C»N N N CO CON XCOMNCOOOCONOe 00 !>• t~ t^ O0 4 ^ MCOrtTjHH^cOdTf CM •>* CM © T © IT. COCOtOtOrHOCOCOff TtlMCOrHlOffitOlOK; N CO r-i Oi S i_l «3 © ^H ,-H ^H rtnOOtMMCOrt^OO o s © © © O ©©©©©©©©© © © © l-H © © © ~ © CO CO « ©©©^-H©t^OCHtr o s © © © © 0©©©©©©©© CM CO © © © © © co©©©co©©©c dlOCOO G(^_ '1 ft o3 h— lOlOWOOOOTtirtrt 10 co O O a COCSNinOCKrHCOTt <*NCO0 © ^H CO S o ^ © a> 00 ex HHHO!0>0©00 oocococo t N ^f 02 CN CO CO CO £ s © © © OOOtNrHrtOtNtN tH N OOO COC (MrtOrt^rtC40« OCONOOT 1 03 T ^ + \ N lO S CM CO N CT ■*ooOTt<>mflN^ fjrt OOCOH «j ONNCC rH^COtNlOCOMOOtN >rt t- CO CO © ■>* CN lr^©CNl^t^oo-v©t^ s © © © © ©©©©©©^H© — OOOH NOr- OOHCOOONMC © © ^H © C H * N O) CO * >C oioioaoooocoN^ (N H lO CO"5 , CM CO © © © V «HHr. IfiinCOiHHHHOOK © t~ © © CO ^h V s © ,-H r-< ^- OOONtMNNMO OOHHN^t H«NNNCNMMC rH (N rt rH N ' t 03 O •* lO (O lO CO N tC COOtN'J'NCOCOOU © Tt< © © CO tN tN CN « Tti-rti-^i-tti-^tO»O>C N O N CO CO CO CO OOhOOOIOOICT 5 © © © C ©©©©©©©C>© © © © © © © <= rt^-HT-(©©O^H©C «• V, >r £' J •^1 IO © t^ CO © © -HCNCO'O'lOtDNCOff © 1-1 CM CO Tti 1 '■'• NNO COhOOOCO'OO'OOOOHh (OlOtjllOUJlOfMOeOHiOM OHO 1/5 to o o co o h h O CO OS h fiOOOO) W O) 00 00 ^ l« CO CO CO CO O CO O UO CO h CO CO h CO CO O tD O tD N rf •* O O U5KJ10HHOCOO OS CD H CO CO NOO OOOO) o o © ooooooooooooo O O O O O oooo o o o o o o O O OS O OS OOOOOOSOOS OS OS O O OS r~ co h ^Hiona>t>>- i o o co co oo h O 00 "0 CO CD O t^^o^Mn o lO to cow N HOON lOlOTf^HtDNO M'O'HMtO'3'HCQ OOt-HMN m >n co "5 co O H H OhhhOOOOOhOhh CO CO r-c CO CO r-H H CO CO i-l CO CO CO CO CO CO CO CO CO CO NHHMNtOWCO CO ^t* CO ^ CO NHOO NTtlOMlflNOOlflOON^OlO lOCOCOCOCOCOCOHOOOCOCO OS H O! (O N OOO N OOJ HOOOIO CO CO CO t~ o oo co oo h o CO O O O GO H t-h t~- H CO o H OS -^ H t* COOlOOCOOcO-ti COCOCOCOOOOi— ICO m oo h — i -* »o td t* »o t- 1 ra o o OOOOOOOOOOOOO 010)0) O O OS OS O O O O OS o O OS OS O OS OOOOOSOSOOS O O O OS OS CO CO rH HcOHc©OOOOOCOcOf~WCOCO COOUOCOOOCOOOOOCOOOCO H(-«lO» O O t- o h h§§2 CO H CO CO OO o CO h OS tt< CO h CO CO O CO o H H OS CO t* HHHl-^otOHCO MiOCON^^Oh CD CD 00 l« CO •*lOH^n O H H OOhhOOOOOhOhh CO h h CO CO h h CO CO H CO H CO CO CO CO CO CO CO CO NhhWNMMM ■"""■" HNOO 3 OO h OHcDHcOCOOCOCOOOOt— CO COOOOOHOOOOt-t^OO OH MOOt* O OlOlQIO ^ H .«! H 00 H Tf ■* H H OS 00 ■**! H 00f» !ON-* 00 CO O CO 00 h »o lO O t— CO HNQNNtOlOtO OOOHOOKSntOM CO O O t* 00 H t* CO O CO 09)0) OOOOOOOOOOOOO OS ~. O o o OS O OS OS OS OS O OS OS OS OS OS O 00 o o o co h oo OOOCOOt^OCOHHOCOOCO MOONHNCOOOOfNNM lO CD h r— o OOMMOH i-H iC OS o 00 OO O h «5 CO CO CO OS OO ■*! H UO O H OO O CO O H OO O O CO CO CO tMSHlOOOHCO t^COHOHT-it^OS CO CO h OO CO O t* OS o o o »-« o OOHrHOOOOOHOr-H,-! V O O h CO r-l CO H CO CO r-t CO CO CO CO CO HHHCONCCMIN CO CO CO ^ CO t- O O NrtM COCOHCOOOhcOOOCOOOOCO COTfiCOOOOOOCOt^COCOOOCOO T-H ift O t~ t* (OHOtOlO CO CO CO CO lO CO CO OS m co m co h co «M l<3 N rt H f OSMH OCO«5COCOHHIO (DNtDtOHH WO CO t— O CO CO O CO h H H OO 00 t- OOWOONNNNl'OOOONOOOO O OO F— O O » 00 OON t- OO t— OO 00 00 O) C) t- M t» NOOWOOOOOOOOt* N OO M 00 00 o Tti co HTtiOffllDNNNf^'l'O'* 0)HHO>HrtNNH10t-«M onooo h CO CO h H 00 O CO CO CO O O CO o h h h oo o d rH N N M ■* 00 O CO CO CO co co r- co co OOOOOOHtON HO*N10tOrtt> CO -H/i 00 Ui 00 HMIOOU5 J* i-l O oomonmioooohnmh o o co o o CO CO H CO CO CO »h tH CO CO o o o «o h OCOIOOhOCOO CO H H CO »-t sss lOSONNNNOJOlNTfCJ^O COHTHt^C^t^HHOHHCOH OO O O 00 00 o h co o o CO CO O H CO CO CO t- CO OS CO CO O 00 CO W5 CO CO -* 1 o h co oo m -& tti «5 o co o tOHHTftOCOHffl HN00 1«COCOOOOO i-l H O CO i-< 00 OS ^ O CD <3 O h ,-iC0O.-h,-ihOO»-iOOcDO CO CO CO CO CO O O CO o O t* 00 OO O CO CO CO H CO H COCOCOCOOSCOCOCO N NNOIO N 1(5 M c© h h Ot-i-lcOCOCOCOCOC©CDCO«5CO HNOOOOJOOOOOtDlONM h o «o h o H CO CO H CO NHNtO CO CO O lO o to O «5 o o CO CO CO CO CO CO h o CO CO • HiOUJO • ONNIOOOMIOM COCOCOCOOlOOlO CO Oho OS ■ CO O CO O CO CO OOOOOOOOhOOOOh H H H CO CO O O CO CO O OO t-- 00 t-H h CO CO H CO • OOOHt^cOCOCO t~ -NOON IO H H .-H CO "tft NOht)ih^(OONWhNOO ONWOMHHOrtNMffllO CO CO H CO o CO CO O CO CO h t- h o NONH lO O 00 H CO OO CO CO H CO 00 h OOO^OtO COCN OOOO coiOh-OKStomio MtOHCOCONCOlO h CO h CO CO tOMMOO *_(< rH HO OOOhCOCOHcOOhOOOCO O O CO o o H WH« O O CO CO CO lO o o •* >o co ONIOONHMW O CO CO O CO OhO> TjlSNHrti-irtrt«HO0CTH to N H ION 1-1 H I— H o co o o h o NNU5N tO h lO O CO OS CO -* t» oo»-ihh.-ii-ihi-icoi-icoh »-l CO CO CO CO i-H H CO ■* -h CO CO CO h rf NMH^IO ^-iH-HCOi-irfi'*it~ CO *0 H CO i-H t* CO CO %C? t- o» avrtooMNtooHiaHd ococooooohcohhho O 00 CO OO O h CO CO CO CO H h t- CO t* CO h CO OS t-» OS OS *h o N O) O) t* O) OO (0^00*0 h CO — I h CO OHONtOlflOOtO NtOlOMM^tON o h m co h CO O CD h CO hJ cococococococococococococo CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO HUSONOO '\ : : IOCS OOOlOrtNeO^WlDt-OOOlO HNW*>0 IQIOUJ lOlO CO t- OO OS "OiOiOiO o h co co •>* tti CO CO CO CO CO CO CO t-- CO OS o CO CO CO CD t- HNCCflOtONOO O O H CO CO N 00 00 00 OO LITERATURE CITED Akin, G. W., and J. V. Lagerwerff 1965a. Calcium carbonate equilibria in aqueous solutions open to the air. I. The solubility of calcite in relation to ionic strength. Geochim. Cosmochim. Acta 29:343-52. 19656. Calcium carbonate equilibria in aqueous solutions open to the air. II. Enhanced solubility of CaCOs in the presence of Mg 2+ and S0 2 1. Geochim. Cosmochim. Acta 29:353-60. Berner, R. A. 1966. Diagenesis of carbonate sediments: Interaction of magnesium in sea water with mineral grains. Science 153:188-91. Bower, C. A., L. V. Wilcox, G. W. Akin, and M. G. Keyes 1965. An index of the tendency of CaC03 to precipitate from irrigation waters. Soil Sci. Soc. Amer. Proc. 29:91-94. Brooks, R., L. M. Clark, and E. F. Thurston 1950. Calcium carbonate and its hydrates. Trans. Roy. Soc. London (A) 243:145-67. Buckley, H. E. 1951. Crystal growth. New York: John Wiley and Sons, Inc., 571 pp. BUEHRER, T. F., AND R. F. REITEMEIR 1940. The inhibiting action of minute amounts of sodium hexametaphosphate on the precipitation of calcium carbonate from ammoniacal solutions II. Jour. Phys. Chem. 44:552-74. California State Department of Water Resources 1964a. Quality of ground waters in California, 1960. Sacramento: Calif. State Dept. of Water Resources Bui. 66-60, Parts I and II. 19646. Quality of surface waters in California, 1960-61. Sacramento: Calif. State Dept. of Water Resources Bui. 65-61, Vol. I, Part II, and Vol. II. DONEEN, L. D. 1964. Notes on water quality in agriculture. Davis: Dept. of Water Science and Engineering Paper 4001, 48 pp. Dye, J. F. 1958. Correlation of the two principal methods of calculating the three kinds of alkalinity. Jour. Amer. Water Works Assoc. 50:800-20. Eastman, H. V. 1966. The miracle leak sealer. Reclamation Era 52(1): 11. Frear, G. L., and J. Johnston 1929. The solubility of CaC0 3 (calcite) in certain aqueous solutions at 25°. Jour. Amer. Chem. Soc. 51(2): 2082-93. Garrels, R. M. 1960. Mineral equilibria. New York: Harper and Brothers, pp. 43-60. Garrels, R. M., M. E. Thompson, and R. Siever 1961. Control of carbonate solubility by carbonate complexes. Amer. Jour. Sci. 259:24-45. [50] Harned, H. S., and R. Davis 1943. The ionization constant of carbonic acid in water and solubility of carbon dioxide in water and aqueous solutions from 0° to 50°. Jour. Amer. Chem. Soc. 65:2030-37. Harned, H. S. and S. R. Scholes, Jr. 1941. The ionization constant of HCO~^ from 0° to 50°. Jour. Amer. Chem. Soc. 63:1706-09. Hem, J. D. 1961. Calculation and use of ion activity. U. S. Geol. Survey, Water-Supply Paper 1535-C. Hodgman, C. D. (ed.) 1955-56. Handbook of chemistry and physics. Ohio Chemical Rubber Pub- lishing Co. (37 ed.), 3513 pp. HURLBUTT, C. S. 1955. Dana's manual of mineralogy. New York: John Wiley and Sons, Inc. (16th ed.), 530 pp. Johnston, J. 1915. The solubility-product constant of calcium and magnesium carbonates. Jour. Amer. Chem. Soc. 37:2001-20. Johnston, J., H. E. Merwin, and E. D. Williamson 1916. The several forms of calcium carbonate. Amer. Jour. Sci. (4th ser.) 41(246): 473-512. KlELLAND, J. 1937. Individual activity coefficients of ions in aqueous solutions. Jour. Amer. Chem. Soc. 59(9): 1675-78. Kline, W. D. 1929. The solubility of magnesium carbonate (nesquehonite) in water at 25° and pressures of carbon dioxide up to one atmosphere. Jour. Amer. Chem. Soc. 51(2): 2093-97. Langelier, W. F. 1936. The analytical control of anticorrosion water treatment. Jour. Amer. Water Works Assoc. 28(10): 1500-21. Larson, T. E., and A. M. Buswell 1942. Calcium carbonate saturation index and alkalinity interpretations. Jour. Amer. Water Works Assoc. 34:1667-84. McCauley, R. F. 1960. Use of polyphosphate for developing protective calcite coatings. Jour. Amer. Water Works Assoc. 52:721-34. Moore, E. W. 1939. Graphic determinations of carbon dioxide and the three forms of alkalinity. Jour. Amer. Water Works Assoc. 31:51-66. Orland, H. 0. (ed.) 1965. Standard methods for the examination of water and wastewater (12th ed.), New York: Amer. Public Health Assoc, Inc. 769 pp. [51] Reitemeir, R. F., and T. F. Buehrer 1940. The inhibiting action of minute amounts of sodium hexametaphosphate on the precipitation of calcium carbonate from ammoniacal solutions I. Jour. Phys. Chem. 44:535-51. Richards, L. A. (ed.) 1954. Diagnosis and improvement of saline and alkali soils. Washington, D. C: U. S. Dept. Agr. Handbook 60. 160 pp. Ryznar, J. W. 1944. A new index for determining amount of calcium carbonate scale formed by a water. Jour. Amer. Water Works Assoc. 36:472-83. Salutsky, M. L. 1959. Precipitates: their formation, properties, and purity. In: Treatise on analytical chemistry. (I. M. Kolthoff and P. J. Elving, eds.). New York: Interscience Encyclopedia, Inc. Part I, Vol. 1, pp. 733-66. Sillen, L. G. 1964. Stability constants. Spec. Pub. No. 17. London: The Chemical Society. 754 pp. Tanji, K. K., and L. D. Doneen 1966a. A computer technique for prediction of CaC0 3 precipitation in HCO~jj salt solutions. Soil Sci. Soc. Amer. Proc. 30:53-56. 19666. Predictions on the solubility of gypsum in aqueous salt solutions. Water Resources Research 2:543-48. Tanji, K. K., L. D. Doneen, J. V. Kubes, and L. R. Simmons 1968. Field techniques for sealing leaky concrete pipelines with ammoniated irrigation waters. Denver, Colo. : Chem. Engr. Branch, Div. of Research, U. S. Dept. of Interior, Bureau of Reclamation Rept. No. ChE-79, 63 pp. Washburn, E. W. (ed.) 1928. International critical tables. New York: McGraw-Hill Book Co. Vol. III. Van Hook, A. 1961. Crystallization theory and practice. ACS Monograph No. 152, New York: Reinhold Publ. Corp. 325 pp. Weir, C. D. 1957. The ammonia-carbon dioxide-water equilibrium in boiler feed-water conditioning. Jour. Appl. Chem. 7:505-11. 7Jm-6,'69(J7218s)LL