L I B RARY OF THE UNIVERSITY Of ILLINOIS 628 I£65c no ■ (-9 tl\IGlNt£RIN< The person charging this material is re- sponsible for its return on or before the Latest Date stamped below. Theft, mutilation, and underlining of books are reasons for disciplinary action and may result in dismissal from the University. University of Illinois Library SEP 6 1MB oc r r *° ^ h L161— O-1096 Digitized by the Internet Archive in 2013 http://archive.org/details/effectofphosphat07morg CIVIL ENGINEERING STUDIES SANITARY ENGINEERING SERIES NO. 7 A\ ENG!M:E*ING LHRARV UHIV ITy OF ILLINOIS URBANA, ILLINOIS EFFECT OF PHOSPHATES ON WATER TREATMENT Phase III. Effects of Phosphates on Coagulation and Sedimentation of Turbid Waters $ vW By JAMES J. MORGAN and R. S. ENGELBRECHT FINAL REPORT FOR THE PERIOD JUNE 15, 1955 TO SEPTEMBER 15, 1957 for the ASSOCIATION OF AMERICAN SOAP AND GLYCERINE PRODUCERS UNIVERSITY OF ILLINOIS URBANA, ILLINOIS EFFECT OF PHOSPHATES ON WATER TREATHENT Phase III, Effects of Phosphates on Coagulation and Sedimentation of Turbid Waters by ames Jo Morgan and S. Engelbrech' A Final Report for the Period e 15, 1955 to September 15 , 1957 ASSOC 3 AT DON OF AMERICAN SOAP AND GLYCERINE PRODUCERS Sanitary Engineering Laboratory University of Illinois August 1958 C2$ no .7 E h f i h CONTENTS ii NTRODUCTSON . Related Work of Other Investigators . . Purpose of the Investigation ...... Scope of the Investigation ....... General Methods of Investigation . . . . Significant Results of the Investigation EXPERIMENT At METHODS Characteristics of Waters Studied .... Coagulants Studied ............ Phosphate Compounds Studied ....... Analytical Methods ............ General Description of Pilot Plant Studies Pilot Plant Apparatus. .......... Pilot Plant Experiment Procedures .... General Description of Laboratory Jar Tests Laboratory Jar Test Procedures ...... EXPERIMENTAL RESULTS ............ iratory Jar Tests ....... Results of Laboratory Jar Tests Pilot Plant Experiments. . . . . . Results of Pilot Plant Experiments IV. DISCUSSION AND CONCLUSIONS . . . . . Discussion of Jar Test Results . . Discussion of Pilot Plant Results General Conclusions ....... V. RECOMMENDATIONS FOR FUTURE RESEARCH REFERENCES ............. Page 1 1 v 7 8 9 11 11 11 12 12 13 13 17 18 19 20 20 20 27 27 39 39 kO k] kh k6 iii ACKNOWLEDGMENT This report covers the final phase of an investigation conducted in the Engineering Experiment Station of the University of S Illinois, Department of Civil Engineering, and sponsored by the Association of American Soap and Glycerine Producers. The period of the investigation was from June 15, 1955 through September 15 s 1957. Acknowledgment is hereby made of the contributions to the planning and execution of this phase of the research by the following persons: Dr. Jess C Dietz 5 formerly Professor of Sanitary Engineering and initially Director of the project; Mr. Robert H. Harmeson, Research Associate and Instructor In Civil Engineering; J. C. Gulllou, Associate Professor of Hydraulic Engineering, consultant on the design of the treatment units; Mr. John R. Towers, Research Assistant; Mr. H. Khalifa, Research Assist- ant;, and Moss LuBelle Boice, Research Assistant. Particular acknowledgment is made to Mr. Moles Norton 2 Senior Laboratory Mechanic, for his contribu- tion to the construction and maintenance of the research equipment. iv ABSTRACT Interferences with turbidity removal by phosphate compounds in the coagulation and sedimentation processes of water treatment were studied in laboratory jar tests and pilot scale treatment plant tests. The major compounds investigated were STP and TSPPj representative of the phosphate builder compounds in synthetic detergents, Alum, ferric sulfate 3 and ferric chloride were employed as coagulants. The waters studied had a total hardness of 250 to 300 ppm s and a calcium hardness of 70 to 125 ppm= Phosphate levels of from 0„5 ppm to 6.7 ppm P ? 0,. showed interferences with coagulation and sedimentation. Magnitudes of the interference were observed to depend upon coagulant dosages, flocculation characteristics of the test apparatus, and sedimentation time. Increases in coagulant dosage were successful in over- coming interferences. Increased sedimentation time also decreased the degree of phosphate interference. For STP and TSPP phosphate levels of 0.5 to 1.0 ppm P^O,-, which were found to be the upper limits of complex phosphates in Illinois surface waters, interferences were overcome with coagulant dosages of 30-50 ppm and theo- retical pilot plant settling times of 3 to 6 hours. The increased coagulant demand per unit of phosphate concentration was observed to be In the order of 2 to 5 9 for adequate flocculation and settling of hard waters. TSPP and STP appeared to exert essentially the same interfering effects. The. orthophosphate compounds tested exerted considerably less interference than did these complex phosphates. LIST OF TABLES Number Description Page 1. Summary of Results of Previous Jar Test Investigations on the Effects of Phosphates on Coagulation-Sedimentation 5 2. Turbidity Removals in Jar Tests for Effects of Phosphates in Coagula- tion and Sedimentation, Series J-l 21 3o Turbidity Removals in Jar Tests for Effects of Phosphates in Coagula- tion and Sedimentation s Series J-l! 26 4o Turbidity Removals in Pilot Plant Tests for Effects of Phosphates in Coagulation and Sedimentation, Series P-8 28 5. Turbidity Removals in Pilot Plant Tests for Effects of Phosphates in Coagulation and Sedimentation, Series P-l 8 8 29 6. Turbidity Removals in Pilot Plant Tests for Effects of Phosphates in Coagulation and Sedimentation, Series P-8I8 30 7. Turbidity Removals in Pilot Plant Tests for Effects of Phosphates in Coagulation and Sedimentation, Series P-JV 31 8. Turbidity Removals in Pilot Plant Tests for Effects of Phosphates in Coagulation and Sedimentation, Series P-V 32 9. Turbidity Removals in Pilot Plant Tests for Effects of Phosphates in Coagulation and Sedimentation, Series P-VS 33 VI LIST OF FIGURES Number Descri ption Page 1. Pilot plant flpw diagram 15 2„ Turbidity remaining versus STP concentration for various ferric sulfate dosages employed in jar tests, Series J- ! I 23 3. Efficiency ratios for jar tests using ferric sulfate at various levels of STP, Series J- ! I 2k k. Observed ferric sulfate dosages required in jar tests to accomplish 90 per cent reduction in turbidity at stated STP concentration, Series J- I I 2$ 5. Efficiency ratios for alum coagulation in pilot plant experiments at various phosphate levels 36 6. Turbidity remaining versus alum dosage in pilot plant experiments at various phosphate levels 37 L INTRODUCTION Increased consumption of synthetic detergents In recent years has prompted research Into the effects of these substances on conventional water treatment processes- Operational difficulties have been attributed to the presence of synthetic detergent compounds In water treatment plant influents, and consid- erable research has been undertaken to establish what effects synthetic deter- gents, or their components, have upon treatment processes, chiefly coagulation- sedimentation and filtration,, A previous report (1) has reviewed the significant literature dealing with operational problems caused by synthetic detergents- The various related investigations concerning the effects of packaged synthetic detergent products and individual components, Le. s the surface active agents and the builder compounds, were also summarized- In order to provide a background for the present study on the effects of phosphate compounds on the coagulation-sedi- mentation process, a brief summary of the pertinent related research by other investigators will be presented- Related Work of Other investigators Cross (2) studied the effects of three domestic and three industrial synthetic detergents on the coagulation of Lake Michigan water- This water had a total hardness of approximately 130 ppm, and an alkalinity of approxi- mately 12.0 ppm, "he turbidities of the water studied ranged from 10 to 13 ppm. The test procedure consisted of flocculating 1000 ml. of sample in a 1500 ml. beaker for 15 minutes. A paddle speed of kh rpm was used. Detergent concen- trations of 0.1, 1.0, 3.0, and 5-0 ppm were used, and a control was run with each test. The coagulants investigated were alum, alum and silicate, and alum and lime; ferric sulfate and ferric sulfate with lime; and chlorinated copperas, with and without lime. 2 Interference with good coagulation was not usually experienced below detergent concentrations of 5-0 ppm. The use of lime appeared to reduce the magnitude of the interference in all cases. For alum coagulation with 5 ppm of detergent, denoting the three domestic detergents studied as D. , D„ , and D„ , Cross obtained the following results: Detergent Turbidity after 30 min. Per cent Removal D l 5 62 2 8 39 D 3 3 77 Control 1.5 88 These data indicate that different detergents exert different magnitudes of interferences on coagulation. Howell s and Sawyer (3), on the basis of their studies of synthetic detergent effects upon color removal by coagulation, con- cluded that the different interferences of synthetic detergents were caused by differences in their phosphate content. Studies by Smith, Cohen, and Walton (k) have indicated that the complex phosphates, sodium tripol yphosphate and tetrasodium pyrophosphate, which are the most Important builder compounds of commercial synthetic detergents, cause interference with alum coagulation of turbid waters at phosphate concentrations of about 1 ppm. Their studies consisted of jar tests on synthetic turbid waters of varying hardness. They investigated four dodecyl benzene sulfonates, tripoly- phosphate and pyrophosphate, and tri sodium phosphate. The four dodecyl benzene sulfonates investigated were found not to interfere with alum coagulation until levels of from 8 to 20 ppm were reached. 3 In soft water coagulation it was observed that complex phosphate interference could be overcome by a substantial increase in alum dosage* In the hard waters tested the alum floe which formed in the presence of complex phosphates was quite fine in size, and was observed to settle slowly. While the degree of interference was less than that for the soft waters, doubt was expressed about whether this floe could be adequately removed in an actual treatment plant. Orthophosphate concentrations which caused interference were much higher than those of complex phosphate causing the same degree of interference. Howells and Sawyer (3) used jar test coagulation of a colored natural water, believed free of detergent, to study the effects of packaged household synthetic detergents, surface active agents, and builder compounds upon water treatment. Alum was employed as the coagulant. Although no mineral analysis was presented, the water used is believed to have been a soft water. Tests on nine surface active agents indicated that none of the compounds pro- duced interference with coagulation in concentrations of less than 10 ppm. Tests on builder compounds indicated that sodium sulfate did not interfere with coagulation in concentrations less than 10 ppm, that sodium carboxymethyl- cellulose produced moderate interference, and that interference by sodium tripoly- phosphate and tetrasodium pyrophosphate was severe The effect of the interfering substances was to increase the amount of coagulant needed to accomplish the de- sired color removal. The effect was one of increasing the alum dosage at which adequate flocculation occured. Floe characteristics were apparently not impaired by the complex phosphates. Langelier (5) investigated the effects of certain surface active agents and of sodium metaphosphate upon flocculation of turbid waters, employing jar tests. No interference by any of the three anionic or three nonionic surface active k agents was found at concentrations of less than 17 ppm„ Alum was the coagulant,, In the case of sodium metaphosphate, a sequestering agent, the alum dosage required for successful coagulation was found to increase linearly with the con- centration of metaphosphate present- Furthermore, the presence of the metaphos- phate was reported to adversely affect floe quality, in spite of increased dosages of alum. The investigations referred to have all reported a quantitative definition of the effects of various phosphate compounds upon jar test flocculation of turbid or colored waters. Howells and Sawyer, and Langelier presented linear relation- ships between phosphate content and required coagulant dosage. These relationships are of the form A = KP + B, where P is ppm of phosphate, as P o o , A is ppm of alum, and B is the ppm of alum required for coagulation in the absence of phosphate. While Smith, Cohen, and Walton presented no equational relationship as such, they did present a plot from which an equation could be obtained. Table 1 summarizes the quantitative results of the above-mentioned investigations. The coefficient K in these relationships indicates the increased coagulant dosage required per unit of increased phosphate concentration. An inspection of the visual observations and the coefficients reported for jar test investigations reveals that; a) Hard waters pppear to be less affected by the presence of complex phosphates than soft waters, in jar tests. b) Size of floe appears to be adversely affected by the presence of complex phosphates in hard waters. o z h- o z =» = H- < CO H z Z o 1 1 1 £ H «° < a ■J ., j = to H- i to 2 III O > ™- z — » < _j CO 3 — ^ O o < n i =-= o -1 5» ••-> EG UJ < OS z h- °- o IU_ co o j; | H CO < *- X -J a. 3 i/l CO o iii 3^ CL a. u. U. © O >• cO a: H < ej £ tu £ Li. 3 L_ CO 11 i C9 O o o O o o o o o LA O o o O o o wo. X 4J a) c < (A I/) Oj C XJ s_ ro x XJ C o in o m o in o in I X a. c o XJ 3 I- c O fD !_ o z o o CN 3 XJ X L. 3 XJ C 3 D. E O u O u> E (D x o 3 X Q. X .— Q. E CD 3 4-» 4-> >« o O — ro E CD E i— in X XJ X 3 E 3 o fO Q. O Q. •— 03 •— Q. l_ o in in "D X XJ •— 4-> 1_ — O o a) o u a) >» l- X LO X CO 1- c o 4-1 ro 3 H a. c o 4-> CO 3 T Walton P CD en c CO «*!> c C CD X o c_> X 4-1 °E CO c CD X o o E CO **!> c - CO TJ c fO CD 5 o X o o o ?r\ CN o in o o o ra =— cf\ K CNJ CNJ a> CO + + + + + + + Q. a. a. a. a. o_ a. — t*\ _ o CNi m o 00 CO o^ r*> ^ CNi 4-» 4J ro (0 X X u Q. a. 4-> in in E ID O o 3 X X X • — CL Q. a. ■a c^ >- >> o O E ^ E i— (/) JC 5 o 3 o to c O 4J CO h- H a. fD CO T3 C fO o X E a. o. ai (D (/) O 13 E 3 cvj E Q. a. X "D tn C JB "D VO (1) •— ' <0 C 3 O CO O — -o CD C Q. — E x o s- >- Q) X 4-1 «B "D 5 0) -a h= c o x cn CO C 3 ro o L. CD 4J U >« 3 4J ~D — o U u O Q. cu cu > x c ro ^- en xi c ro ro \~ i- ro a, c E — o ro o +j s- a) ^-^ a* x r^ o " — o £ 4-> C o X <1> — +J S_ 4-> .— ro ro 01 3 •— w o ^- — u ro 3 o «4- tf> i — CD "4- "D C- — XJ 3 <+- O o — o X en sn •> 0) u H E O •_3 •? H- m U- »- 3 0)- — x m ro CD > XJ "D 3 - o >-x u m c CD in .— +j «+» 3 CD E O c — o +-> .- ro 4-1 c— ro 3 o o 3 XJ o c o a.._ 4-1 cu ro x — 4J 3 O O O +J O — <4- ro 3 *+- cr o CD XJ in O s_ O Cl cvj c en a) C SL ._ CD 4-t t- ro <+- 3 — — XJ ro > i- (D O »+- 1_ O J— «4- C3 in ro XJ CD *-i in CD en en 3 O) c CD CD X CD > ro x o o o o o J_ 14- l«_ CD O O 4-< CD in o> E CD CD ro 3 3 L. r— i— ro ro ro Q. > > E CL a. in O o 3 XJ CD 3f ^k to C 3 O 3 XJ CD c) Tet rased i urn pyrophosphate and sodium tripol yphosphate have approximately the same effects on coagulation of soft waters. The results of two inves- tigators indicate a slightly greater interference caused by pyrophosphate. d) Trisodium phosphate (orthophosphate) has a considerably smaller effect upon coagulation than do the complex phosphates. The reasons for the interference to coagulation of waters containing these phosphates are not precisely known. However, the following statements, appearing in recent literature, suggest a general theoretical basis for such interference: "Probably the most significant role of builders in synthetic detergent formu- lation is that of increasing detersive action by enchancing surface active properties and by acting as suspending and dispersing agents." (8) "Such phosphates may increase the surface activity of wetting agents; form complexes with ions of earth alkalies and heavy metal salts; and also peptize, disperse, suspend and emulsify pigments, oils, etc." (8) "Sodium tri pol yphosphate increases the foaming and dispersing tendencies of detergents and, in addition, possess sequestering power. Tetrasodium pyro- phosphate is another phosphate builder that: exhibits dispersing and seques- tering properties. (9) Purpose of the Investigation The purpose of this investigation was to study the effects of phosphate compounds, particularly sodium tri poll yphosphate and tetrasodium pyrophosphate, on the coagulation-sedimentation process of water treatment. Phase 1 of this final report presented the Jesuits of a survey of phosphate concentrations in Illinois surface waters (10). This survey established that complex phosphates in surface waters used as sources of water supply generally are below 0.5 ppm P^O^ 7 Therefore, the present, investigation concentrated upon an investigation of the effects of phosphate levels of 0.5 to 1 . ppm P 9 0r upon the coagulation-sedimenta- tion process. However, higher levels were also investigated, in order to describe more completely the effect of a range of phosphate levels upon water treatment. Scope of the Investigation This Investigation was limited essentially to a study of the effects of phosphate compounds upon clarification of turbid water, employing both laboratory jar tests and a pilot scale water treatment unit. No attempt was made to thoroughly study the effects of these compounds upon the filtration process. Thus, efficiency of turbidity removal in flocculation and sedimentation was chosen as the index of phosphate interference. The waters studied were hard waters. Total hardness varied from 250 ppm to 300 ppm, and calcium hardness varied from 70 ppm to 125 ppm. There were two reasons for selecting waters of these hardness characteristics. First, these characteristics corresponded generally to those found in the survey of Illinois surface waters. Secondly, although it has been reported that hard waters are less severely influenced by phosphate interference than soft waters, it has also been reported 'that phosphate interferences In hard waters are of a nature to affect floe settling characteristics. Such interferences could prove quite Important In plant, operation, as contrasted to laboratory jar test conditions. The phosphate compounds studied were those which chiefly compose synthetic detergent builders. Sodium tripolyphosphate, tetrasodium pyrophosphate, and three orthophosphate compounds were used. The coagulants employed in this study were filter alum, ferric sulfate, and ferric chloride. Adjustment of pH was used In some of the tests in which alum was employed. However, coagulant or flocculation aids were not investigated. 8 General Methods of investigation The investigation employed two different methods for evaluating the effects of particular levels of phosphate compounds upon the coagul at.ion-sediment.at ion process; pilot plant experiments and laboratory jar tests. Pilot Plant Experiments ; This method involved the comparison of turbidity removals in test waters, containing known amounts of phosphates, and phosphate-free control waters, by means of two identical water treatment units of pilot plant scale. Since test waters and control waters were identical, with the exception of their phosphate content, and since each water in a given test run received Identical treatment, any difference in turbidity removal efficiency was attributed to phosphate interference. Plant runs were of the order of eight hours in duration. Fine clays were used as sources of turbidity. Demlneral i zed or phosphate-free local waters were prepared for each test run. Modifications in plant design were made during the course of the investigation in order to observe the effects of mixing speeds, flocculation charac- teristics, and settling times upon plant performance in relation to phosphate inter- ference. Varying dosages of each coagulant and varying levels of phosphates were used In order to observe their effects on coagulation, as measured by turbidity removal eff i ciency. Laboratory Jar Tests; Jar tests provide a rapid method of determining the relative effectiveness of different, coagulant dosages In water treatment plant control. They are frequently used to measure the effects of coagulant aids and interfering sub- stances upon coagulation, but they have the disadvantage of being relatively un- standardi zed, so that results of different tests may not be directly comparable. Jar tests were usdd in this investigation to obtain a preliminary indication of 9 phosphate effects, and also to attempt a comparison of results with those of other investigators and wi th the pilot plant results of this investigation. Jar tests were performed using alum and ferric sulfate as coagulants. An earlier report (1) presented the results of preliminary jar test studies on alum coagulation of phosphate-bearing waters. The tests discussed in the present report were conducted using a different mixing speed and a different sample volume than those in the preliminary studies. Significant Results of the investigation Pilot Plant Results : interferences with coagulation and sedimentation efficiency were exerted by STP and TSPP at levels of 0.5 ppm P ? 0,- and greater. For a given coagulant dose, the intefference increased with increasing complex phosphate. For a given phosphate concentration, interference with efficiency of turbidity removal was decreased by an increase in coagulant dosage. Visual observations and efficiency results at different f loccul ation-sedimentation times suggested that the interference was exerted through an effect upon floe size and settleabi 1 S ty. An increase in f locculation-sedimentation time was successful in reducing the magnitude of the observed interferences. Improved rapid mixing and flocculating characteristics also reduced the degree of interference. If the effects of the presence of phosphates in a water are represented by C = KP + B, in which C Ss the coagulant required for 5 per cent residual turbidity, P is phosphate, expressed as ppm P o c , and B is the coagulant required in absence of phosphate, then K appears to be of the order of 2 to 5 for adequate flocculation characteristics and sufficient settling time, and of the order of 20 for poorer flocculation characteristics and shorter settling times. Jar Test. Results : Phosphate levels of STP from 0.5 to 5.0 ppm P-0- produced inte ference with efficiency of turbidity removal using ferric sulfate and alum. The r- 10 magnitude of the interference was decreased by increasing the coagulant dosage . Coagulant dosages of from 30 to 50 ppm were required to produce a settled effluent of 10 ppm turbidity or less, depending upon the concentration of STP. While the relationship of required coagulant dosage to phosphate content did not appear linear, a straight- 1 ioe relationship was fitted, in order to compare the results with those of other investigators. The following equation was obtained: F = 2.0 P + 31 where F is ppm of ferric sulfate required for a 10% residual turbidity, P is ppm of P 2°5- This relationship may be compared with the results of other investigators summarized in Table I. It should be noted that the value of GT for this series of jar tests was 5600. This lower flocculation characteristic, in addition to an alkalinity of 200 ppm, probably accounts for the high value of the intercept term, 31 ppm. A preliminary series of jar tests, previously reported (1), used alum, and studied the effects of STP, TSPP, monosodium orthophosphate, and disodium ortho- phosphate. The value of GT for that series was ^0,000, The observed approximate value q,f K, in A = KP + B, for the STP and ISPP series was slightly less than 1.0, indicating that benefits of improved flocculation characteristics. STP and TSPP produced more severe interferences than the orthophosphate compounds. !l. EXPERIMENTAL METHODS Characteristics of Waters Studied PS lot Plant : The waters used during the investigation can be characterized as hard waters. Two different sources of phosphate free water were used during the pilot plant studies. Demi neral i zed water, prepared using an ion exchange process, was used for a number of pilot plant runs. The addition of the appropriate chem- icals to this demi neral S zed water produced a water of the following mineral char- acteristics: alkalinity, 200 ppm; calcium hardness, 70 ppm; total hardness, 300 ppm, Calciuma chloride, magnesium sulfate and sodium bicarbonate were the chemicals added. For the remainder of the pilot plant runs, a phosphate- f ree local ground water was used. This water had the following mineral characteristics: alkalinity, 300 ppm; calcium hardness, 125 ppm; total hardness, 250 ppm. Turbidity was added to these waters In the form of powdered clays, and they were mixed for several hours, in order to produce a raw water turbidity of from 50 to 120 ppm. A grain size distribution of the clays used indicated that par- ticle size ranged from greater than 10 micron to less than 1 micron. Jar Tests : Distilled water, to which the approximate chemicals were added, was used In all jar tests. The mineral characteristics of the waters used were as follows: alkalinity, 200 ppm; calcium hardness, 70 ppm; total hardness. Calcium chloride, magnesium sulfate, and sodium bicarbonate were the chemicals added. Turbidity, in the form of powdered clay, was added to the waters used for jar tests. Raw water turbidities were approximately 10C Coagulants Studied The coagulants studied in the different series of laboratory jar tests were reagent grade aluminum sulfate and reagent grade ferric sulfate. The coagulants studied in the pilot plant series were filter alum (technical grade aluminum 12 sulfate), reagent grade ferric sulfate, and reagent grade ferric chloride.. Equiv- alent dosages of filter alum are approximately 11 per cent greater by weight than those of reagent grade aluminum sulfate. Phosphate Compounds Studied The compounds tested were those known to constitute a major fraction of the builder compounds in synthetic detergents. Sodium tripol yphosphate (STP) , Na r P„,0 _, 5 3 10 and tetrasodium pyrophosphate (TSPP) , Na.P„0 , are reported to be of greatest im- portance among the complex phosphate builders. In addition to the complex phos- phates, three orthophosphate compounds were investigated. Because complex phos- phates are known to degrade to orthophosphate (11), it is reasonable to expect that a considerable proportion of the phosphate material in a water supply influent would be simple orthophosphate. The three orthophosphates used were monosodium ortho- phosphate, disodium orthophosphate, and monopotass ium orthophosphate. Method of Expressing Phosphate Concentrations : All phosphate concentrations have been expressed as parts per million of P_0 . 1 ppm P o c = 1.73 ppm STP = 1.87 ppm TSPP = 1.34 ppm P0^ = 0.43 ppm P. Analytic al Hethods Determination of Phosphate Concentration : The method used for condensed phosphate and orthophosphate determination was based on the color iinetri c determination of orthophosphate. Acid hydrolysis was used to convert condensed hydrolyzable phos- phate to orthophosphate. A detailed discussion of the method used is included in Phase 1 of this report (10) . Turbidity Determinations : A Model 14 Coleman Universal Spectrophotometer was for all turbidity determinations in connection with jar tests. The spectro- photometer was also used for several pilot plant test runs. The majority of pilot plant turbidity determinations were made with a Hellige turbidimeter. 13 A Jackson Candle Turbidimeter was used for calibration purposes,, Hardness ; The Hach modification and reagents for the compleximetric hardness test using ETDA was used for analysis of total and calcium hardness. Results are ex- pressed as CaC.0~. Total Alkalinity : The t i trimetric method was used in all determinations; methyl purple indicator was used, and 0.Q2N standard HLS0» was the titrant. Results are ressed as CaCOL . 3 pH Determinations : pH determinations were made using a Beckman Model G electro- metric pH meter. Genera l Description of Pilot Plant Studies The effects of a given phosphate level upon coagulation and sedimentation were determined by periodic sampling of influent and effluent waters during a pilot plant test run of approximately eight hours duration. Each sample was an™ alyzed for turbidity, and the average efficiency of turbidity removal for both the test phosphate water and the phosphate free control water was determined. Compar- ison of these eff Iciencies, accompanied by a statistical evaluation of the signifi- cance of their difference;, indicated the effect of the phosphate compound upon the process. Pilot Plant Apparatus The apparatus used to measure the effects of various phosphate compounds con- sisted of two identical treatment units. Each unit consisted of a raw water basin, constant head supply tank, chemical feeder, rapid-mix tank with rapid mixing equip- ment, f 1 occu 1 at 5 on-sed I mentation tank, and rapid sand filter. A flow diagram of the treatment units is shown in Figure 1„ The walls of the floccul at 5 on- sedimenta- tion tank were of plexi-glass construction, to facilitate observations of floe size and settling characteristics. 14 Each raw water basin had, a storage capacity of 680 gallons. A value connected the two basins, and a circulating pump was provided for the purpose of completely mixing the contents of the two basins for a number of hours prior to each test run. Thus, it was possible to thoroughly mix the contents of the basins, and then to isolate each basin from the other prior to the addition of a phosphate compound to the water of one basin. The capacity of the basins was selected to provide for test runs of up to approximately 10 hours duration, at a design flow of 1 gallon per minute. Two centrifugal pumps were used to lift water from the raw water basins into the constant head supply tanks. Rate of flow through the treatment units was con- trolled by means of a valvie on the discharge line leading from each constant head tank. Each chemical feeder was a constant-rate, solution feeder, consisting of a con" stant-head siphon with calibrated glass top. The chemical feeders added coagulant directly to the rapid-mix tanks. The rapid-mix tanks provided a variable mixing period. The mixing apparatus was generally operated at a speed of from 400 to The flocculat ion-sedimentation tanks were rectangular basins, 12 feet long, J foot wide, and designed for a 2 foot flow depth. The first h feet of basin lengtl contained the flocculat ion mechanism, which consisted of a variable speed walkings beam f locculators. The f locculators were basket-type dasher plates, which -moved vertically through the flocculation zone. The area of the dasher plates normal to the direction of movement was 38^ square inches. The SR^ed of the f locculators was varied from 2.0 to 4,7 fpm during the several series of test runs. The theoretical retention time of the flocculation zone was 1 hour at 1 gallon per minute flow, and 2 hours at 1/2 gallon per minute flow. 15 or CD < O 1 1 A 1^ h- a a (0 C0 < -J Ql O * 2 < K 2 O H < H 2 UJ 5 Q ^ Ul CO 2 1 < 1- or LU 2 o X X £ 5 2 -I 3 0) Q Q O or Q_ < Q. < 3 UJ b or or U. u. - < UJ UJ 2 CO 1 ^ 2 < or Q tZ CO < co or < UJ X UJ Q UJ UJ u_ CL < or UJ »- _i 1 2 CO < o 5 £ 2 UJ < o X 0T o o cj to 16 The final 8 feet of bastn length constituted the sedimentation zone. A hang- ing baffle separated the flocculation and sedimentation zones. Theoretical reten- tion time was 2 hours at 1 gallon per minute flow and k hours at 1/2 gallon per minute. flow. Velocities of flow were 0.067 feet per minute and 0.033 feet per minute, respectively. The settling basins had a surface loading rate of 180 gallons per day per square foot of surface area, at a flow of 1 gallon per minute. A sludge removal device, consisting of a continuously moving scraper system, was designed for the removal of deposited sludge to collection hoppers. This system was found to be un- necessary, and was not used after an initial series of runs. The outlet device of each f locculat Son-sedimentation basin consisted of an 11 inch sharp=crested weir, set to provide a flow depth in the basin of 2 feet. The effluent from each basin was passed through a rapid sand filter. The filters consisted of 0.50 to 0.80 mm sand to a depth of approximately 30 Inches. Filter area was provided to allow filtration at a hydraulic loading of 2 gallons per minute per square foot. Filters were equipped with piezometer tubes to measure head losses. Short-circuitin g Analysis: It is known that flocculation and sedimentation basins are subject to short-circuiting, so that actual median times of flow are less than theoretical detention periods. A short-circuiting analysis conducted in the units at a flow of 1 gallon per minute indicated a median flow-through time of 130 minutes, as compared to the theoretical time of 180 minutes through the f locculat Son-sedimenta- tion tanks. This Indicates a hydraulic basin efficiency of approximately 60 per cent. Pilot Plant Flocculat ion Characteristics: Both the speed of the walking-beam floccu- lators and the theoretical detention time of flocculation were varied during the course of the pilot plant studies. The overall flocculation characteristics of a system may be represented by the product of G, the temporal mean velocity gradient, and T, the period of flocculation (6). G is a function of the power input for flocculation and 17 the viscosity of the water; G = (P/(j. ) 2 . The di mens 5 on less product, GT, for each flocculation condition, is given in the following table. Paddle Ve FPM loci ty Theoreti cal minutes Time, Ca Iculated G, Calculated GT 2.0 60 0.89 3200 2.9 60 1.6 5650 kj 60 3.2 1 1 , 300 kj 120 3.2 22,600 Each pilot plant run was of 8 to 10 hour duration. The water to be used in a given run was accumulated in the raw water storage basins on the date pre- ceding the run. Turbidity, in the form of powdered clay, was added to the water in each basin s and the contents of the basins were thoroughly mixed for 12 to 15 hours prior to the start of the run. At least one hour prior to the start of a run, the valve connecting the two raw water storage basins was closed, and a known amount of the phosphate compound to be tested was added to one of the tanks. The tank contents were then mixed thoroughly. A run was commenced by filling the coagulation-sedimentation tanks, while coagulant was added to the water at a constant rate of liquid feed!. Raw water turbidity samples were collected and analyzed periodically at the influents to both test water and control water units. The time required to fill the coagula- tion-sedimentation tanks was 3 hours, at a flow of 1 gallon per minute. For runs conducted at a rate of 1/2 gallon per minute, the units were filled at a rate of 1-2 gallons per minute, and chemical flow rate was adjusted accordingly. 18 From 8 to 15 determinations of influent and effluent turbidity were made on both the control and test unit during each run. Turbidity determinations were usually made within 15 minutes of time of collection. Observations of floe size and settling characteristics were noted. During the majority of the runs, deter- minations of filtered effluent turbidity were also made. In several runs, filter head loss readings were recorded. Concentrations of phosphate were determined by analysis, as a check on the calculated amounts added. Temperature measurements were recorded during the course of each run. II n order to verify the identical performance of each of the dual units, a number of runs were made in which all conditions in each tank were either exactly the same, or in which conditions were directly the reverse of those of some pre" vious run. For example, both units were used to treat a phosphate-free water under identical operating conditions; or one unit was used to test a given level of 'phos- phate, and then the other unit was subsequently used for the same level. These test runs demonstrated that both units performed identically under given conditions. Following each test run, all units were thoroughly cleaned with detergent-free water in preparation for the next test. Settled sludge was removed, rapid sand! filters were back-washed, and raw water storage tanks were washed and drained. The chemical feeding devices were calibrated prior to each run. General Descript ion of Labora tory Jar Tests A standard laboratory multiple stirrer apparatus was used to flocculate the samples tested, which contained known phosphate levels and coagulant dosages. Jnitla' and final turbidity determinations were made for each sample, and efficiencies of tur° bidity removal computed. A control sample was included for each sample containing phosphate. Comparison of turbidity removals at a given coagulant dose Indicated the effect of the phosphate compound upon the process. 19 Lab oratory Jar Test Procedures Two different jar test procedures were employed during the invest! gat. son. One of these procedures was employed in the preliminary jar test: series, and was dis- cussed in an earlier report (1). The procedure described in the present report differs from the earlier procedure with respect to volume of sample used and speed of f locculatson. in their other aspects the procedures were essentially the same. A standard six-place laboratory stirrer was used in the jar tests. Stirrer speed was variable from 15 rpm to 100 rpm„ The stirrer blades were 3 inches wide by 1 inch in height. The volume of sample used for each test was 2000 milliliters. Stock solutions of chemicals were added to each of the six 2000 milliliter test-samples, and the water was thoroughly mixed. A fine clay suspension was added to each sample to provide an initial turbidity of approximately 100 ppm, and the act- ual turbidity of each sample was then determined. A definite amount of the phosphate compound to be tested was added to each of three test samples. The remaining three samples served as controls. All samples were again mixed prior to the addition of coagulant. A calculated volume of coagulant stock solution was added to each of the six samples in succession, during a 30 second period of rapid mixing. Follow- ing this rapid mixing period, each sample was flocculated for 15 minutes at a paddle speed of 15 rpm. This corresponded to a paddle tip velocity of 12 fpm. At the end of the flocculation period the paddles were withdrawn, and the samples were settled for 30 minutes. A small portion of each sample was then carefully withdrawn, and the final turbidity determined. 20 lit. EXPERIMENTAL RESULTS Laboratory Jar Tests Two series of laboratory jar tests were conducted. The tests of series J- I employed reagent grade ferric sulfate as a coagulant, and studied the effects upon coagulation of concentrations of STP of 0.5, K0, 2.5> and 5«0 ppm P«0 » The pH of the waters was 8.2. A range of coagulant dosages of from 5 to kO ppm was used. The tests of series J- 111 employed reagent grade aluminum sulfate as a coagulant. Concentrations of STP of 0.5 and 5»0 ppm were studied. The range of coagulant dosages was from 10 to kO ppm aluminum sulfate. The pH range of the waters was from 6.6 to 7»0. The flocculation conditions of these jar test series were chosen to agree more closely with pilot plant flocculation conditions than did those of the preliminary jar tests, previously reported. Whereas the product GT for the preliminary tests was 40,000, the product for these two series was 5600. The GT value for pilot plant tests ranged from 3200 to 22,600. Results of Lab oratory Jar Tests Ferric Sulfate; The results of jar tests In which ferric sulfate was used as the coagulant are presented in Table 2. Each average efficiency removal has been computed from the results of three separate tests. The effect of the presence of varying amounts of sodium tripolyphosphate is indicated by comparing the average efficiency of turbidity removal for each phosphate concentration with that for zero phosphate. This comparison has been made by means of an efficiency ratio, which is the ratio of the efficiency of removal for a water of given phosphate content to that for the zero phosphate water, at a given coagulant dosage. Values of the final turbidity are also reported. These, afford an approximate basis of comparison of 21 TABLE 2 TURBIDITY REMOVALS IN JAR TESTS FOR EFFECTS OF PHOSPHATES IN COAGULATION AND SEDIMENTATION Series J- I Test Conditions Coagulant - Ferric Sulfate Phosphate - Sodium Tri polyphosphate (STP) Hardness - Calcium, 70 ppm; Total, 300 ppm Alkalinity - 200 ppm pH -8.2 Flocculatlon Time - 15 Minutes Settling Time - 30 Minutes Rapid Mix Speed - 45 rpm Flocculation Speed - 15 rpm Temperature - 30°C Final Run Coagul ant, Phosphate Avg. Efficiency Efficiency Turbidi ty No. ppm ppm P £ 5 Removal Per Cent Ratio ppm 1 5 0.0 57.2 K00 46.5 2 5 0.5 39.6 .69 58.2 3 5 1.0 30.2 .53 66.7 4 5 2.5 - - - 5 5 5.0 17.1 .30 75A 6 10 0.0 76.2 1.00 26.5 7 10 0.5 73.5 .97 25.9 8 10 1.0 68.7 .90 30.2 9 10 2.5 67.2 .88 34.0 10 10 5.0 60.2 .79 37.1 11 15 0.0 85.6 1.00 14.8 12 15 0.5 83.0 .97 16.6 13 15 1.0 81.9 .96 17.7 14 15 2.5 76.8 .90 24.0 15 15 5.0 71.2 .83 26.9 16 25 0.0 89.3 1.00 10.9 17 25 0.5 87.2 .98 12.5 18 25 1,0 87.1 .98 12.5 19 25 2.5 85.9 .96 15.2 20 25 5.0 79.8 .90 19.1 21 4o 0.0 94.3 1.00 5.8 22 4o 0.5 91.5 .97 8.2 23 ko 1.0 91.7 .97 8.2 24 ko 2.5 91.0 .97 9.2 25 ko 5.0 88.6 .94 10.9 interference effects, too, since Initial turbidities were approximately eqda I . However, only a per cent removal basis affords a true quantitative comparison. A graphical representation of turbidity removal data Is given in Figure 2. Per cent turbidity remaining has been plotted against concentration of sodium tripolyphosphate, as P«Qg- » for each coagulant dosage,, The effect of increased coagulant dosage upon turbidity removal at a given complex phosphate concentration is to be noted. Efficiency ratios have been plotted against coagulant dosage, for various levels of STP, in Figure 3. Low ratios are brought about by low coagulant dos- ages and high phosphate levels. Increased coagulant dosages tend to bring about a decrease in the magnitude of interference, as evidenced by the increased effi- ciency ratios at higher coagulant dosages. Figure h Is a plot of the. ferric sulfate dosage needed to bring about a reduction In turbidity against concentration of sodium tripolyphosphate, as P«0r Two curves have been fitted to the data. One is an attempt to represent the ob- served relationship, as shown by the solid curve. The second 5s a straight line fit to the data, Indicating a linear relationship between coagulant demand and complex phosphate. The equation of this line Is F * 2.0 P + 31 , where F Is the ppm of ferric sulfate needed for 90% turbidity removal, and P Is the concentration of STP, expressed as ppm P«©r • Thus, an Increase In coagu- lant demand of 2 ppm Is indicated for each ppm Increase STP, expressed as P^Qj- . Aluminum Sulfate; Results of aluminum sulfate coagulation of waters containing 0.5 and 5.0 ppm of STP, expressed as fM* , are presented In Table 3. The mag- nitude of Interference by STP concentrations is indicated by the value of the 23 12 3 4 5 6 Sodium Tripolyphosphate, As PPM P2O5 FIGURE 2. Turbidity remaining versus STP concentration for various ferric sulfate dosages employed In jar tests, Series J-l. 2k 0.90 0.80 o c 0) Q> C u c 0) UJ 0.70 0.60 0.50 0.40 0.30 0.20 20 30 Ferric Sulfate, PPM FIGURE 3. Efficiency ratios for Jar tests using ferric sulfate at various levels of STP, Series J- I . 25 0- Q. fl> O «^ 3 ifi o 12 3 4 5 6 Sodium Tripoly phosphate, As PPM P2O5 FIGURE 4. Observed ferric sulfate dosages required In jar tests to accomplish 90 per cent reduction In turbidity at stated STP concentration^ Series J- I. 26 TABLE 3 TURBIDITY REMOVALS IN JAR TESTS FOR EFFECTS OF PHOSPHATES IN COAGULATION AND SEDIMENTATION Series J- I I Test Conditions All conditions identical to those of Series J- 1 , except that alum is used as the coagulant, and pH was S.6 to 7.0, Run No. Co; agulant ppm Phospha ppm P 2 °5 Avg. Efficiency Removal Per Cent Eff iciency Ratio 1 2 15 15 0.0 0,5 83.3 83.1 1.00 .99 3 4 25 25 0.0 0.5 9%h 90.8 1.00 ;97 5 6 40 4o 0.0 0.5 94.9 94.2 1.00 .99 7 8 4o 40 0.0 5.0 94.9 81.5 1.00 .86 27 efficiency ratio at a given coagulant dosage. It 3s noted that an STP level of 0.5 ppm P o 0c caused a slight interference with removal of turbidity, while ^ 5 5.0 ppm P«0 caused a considerable interference. These tests were conducted at pH 6.6 - 7.0, which was found to be the optumum range for alum coagulation. A separate series of jar tests established this range, by investigating re- movals in alum coagulation over a pH range from 6.0 to 8.25» Pilot Plant Experiments Six different series of pilot plant tests were conducted. Concentration of phosphate and concentration of coagulant were the primary variables under consideration in each test series. However, coagulation-sedimentation is a very complex process, involving the attainment of a delicate physical and chemical balance, so that the conditions of operation exercise a considerable effect on the results obtained. Among the conditions which exert an influence upon success of turbidity removal are intensity of rapid chemical mixing, speed and duration of f locculation, settling time, pH of the water, temperature, the chemical and physical nature of the turbidity being flocculated, and the hard™ ness and alkalinity of the water. Several of these conditions were varied dur- ing the pilot plant investigation. The chemical and physical nature of the tt r r= bidity material remained essentially the same for all tests. For each of the six test series, as defined by its operating conditions, the test results taken together describe an observed relationship between concentration of coagulant, concentration of phosphate, and efficiency of turbidity removal. Results of Pilot Plant Experiments The results for each of the six pilot plant test series, designated as P=D to P"V! S are presented in Tables k to 9. Each table includes a summary of the tfcs conditions; values of temperature, coagulant dosage, phosphate concentration, efficiencies of turbidity removal; an efficiency ratio; and values of final 28 TABLE 4 TURBIDITY REMOVALS IN PILOT PLANT TESTS FOR EFFECTS OF PHOSPHATES IN COAGULATION AND SEDIMENTATION Series P- 8 Test Conditions Coagulant - Filter Alum Phosphate - Sodium Tri polyphosphate (STP) Hardness - Calcium, 70 ppm; Total, 300 ppm Alkal inity - 200 ppm No Rapid Mixing Flocculation Speed - 2.0 fpm Flocculation-Sedimentation Time - 3 Hours Phos- Coag- phate Avg, Eff i ciency Final Turb idi ty Temp. °C ulant ppm ppm - p 2 o 5 Per cent Eff. Ratio ppm Run Control Test Control Test 1 20 32 1.1 84.6 78.9 .93 16.6 25.4 2 21 63 2.2 87.7 82. 7 .94 14.2 22.8 3 19 63 4.5 86.4 79,9 .92 40.1 22.4 4 19 63 6.7 81.4 70.1 .86 15.6 22.4 NOTE: All runs showed a significant difference between control and test waters at a probability of 0.05, by statistical analysis. 29 TABLE 5 TURBIDITY REMOVALS IN PILOT PLANT TESTS FOR EFFECTS OF PHOSPHATES IN COAGULATION AND SEDIMENTATION Series P- I I Test Cond itions Coagulant - Filter Alum Phosphate - Tetrasodium Pyrophosphate (TSPP) Hardness - Calcium, 70 ppm; Total, 300 ppm Alkalinity - 200 ppm No Rapid Mixing Flocculation Speed - 2.9 fpm F 1 occu 1 at i on- Sed indentation Time - 3 Hours Phos- Coag- phate Avg. Eff iciency Final Turb i d i ty Tempo °C ulant ppm ppm P 2 G 5 Per Control cent Test Eff. Ratio ppm Run Control Test 1 22 5 0.5 57.0 48.6 .85 35.7 40.5 2 21 15 0.5 68.2 59.8 .88 28.9 34.1 3 24 25 0.5 75.5 70. 1 .93 20.1 27.6 4 21 4o 0.5 77.2 72.4 .94 23.6 26.2 5 23 5 1.0 55.6 44.6 .80 55.2 44.6 6 23 15 K0 68.2 56.8 .83 31.9 34.6 7 22 25 K0 75.5 68.4 .91 24.3 32.2 8 23 4o K0 77.2 69.1 .90 22.1 27.9 9 23 50 K0 84.2 78.7 .94 14.5 22.7 10 22 70 K0 73.9 70.6 .96 27.3 30.4 11 23 70 K0 78.5 74.9 .95 24.3 28.9 12 22 70 2.0 83.8 77.2 .92 14.5 18.6 13 21 50 5.0 80.4 72.1 .90 19.7 29.7 NOTE: All runs showed a significant difference between control and test waters at a probability of 0.05, by statistical analysis. 30 3 26 4 24 5 26 6 23 7 25 8 24 TABLE 6 TURBIDITY REMOVALS IN PILOT PLANT TESTS FOR EFFECTS OF PHOSPHATES IN COAGULATION AND S ED 1 MEMTAT 1 ON Series P-IH Test Conditions Coagulant - Filter Alum Phosphate - Tetrasodium Pyrophosphate (TSPP) except No. 6, STP Hardness - Calcium, 70 ppm; Total, 300 ppm Alkal ini ty - 150 ppm Rapid Mixing - 500 rpm Flocculation Speed - 2.9 fpm Fdocculation-Sedimentation Time - 3 Hours Phos- Coag- phate Avg„ Efficiency Final Turbidity Temp. ulant ppm Per cent Eff. ppm Run °C ppm ^o^c Control Test Ratio Control Test 1 25 25 0.5 84.5 81. 4 .96 15-2 18,5 2 26 25 0.5 86.9 84.0 .96 12.6 14.1 50 0.5 90.0 87.5 97 10.8 12.7 50 0.5 89.4 85.5 .96 11.0 14.4 50 0.5 89.9 86.6 .96 9.6 12.7 50 0.5$fP) 90.8 87.8 .97 5-9 8.0 50 1.0 90.3 85.8 .95 9.7 14.3 50 1.0 89.2 83.9 .94 12.0 16.3 NOTE: All runs, with exception of No. 1, showed a significant difference between control and test waters at a probability of 0.05, by sta- tistical analysis. Run No. 1 could only be viewed as significant at a probability of 0.10. .--'I TABLE 7 TURBIDITY REMOVALS IN PILOT PLANT TESTS FOR EFFECTS OF PHOSPHATES IN COAGULATION AND SEDIMENTATION Series P-IV Test Conditions Coagulant - Ferric Chloridie Phosphate - Tetrasodium Pyrophosphate (TSPP) Hardness - Calcium, 70 ppm: Total, 300 ppm Alkalinity - 200 ppm No Rapid Mixing Flocculation Speed - 2.9 fpm Flocculatlon-Sedimentation Time - 3 Hours Phos- Coag- -phate Avg. Efficiency Final Turbidity Temp. ulant ppm Per cent Eff. ppm Run °C ppm ^o^n Control Test Ratio Control Test 1 23 25 0.5 75.8 68.7 =91 25.6 25.5 2 23 50 0.5 83.2 80.7 .97 16.0 17.6 NOTE: Both runs showed a significant difference between control and Test waters at a probability of 0.05, by statistical analysis, 32 TABLE 8 TURBIDITY REMOVALS IN PILOT PLANT TESTS FOR EFFECTS OF PHOSPHATES IN COAGULATION AND SEDIMENTATION Series P=V Test Conditions Coagulant - Filter Alum Phosphate - Sodium TrS polyphosphate (STP) , except as note* Hardness - Calcium, 125 ppm; Total, 250 pH Adjustment Used Alkalinity =■ 100 - 300 ppm Rapid Mixing - 500 rpm Flocculation Speed - 4.7 fpm Phos- Coag- phate Avg„ Efficiency Final Turb idi ty Temp. °C Time Hours ulant PPm PPm P 2°5 Per cent Effo Ratio ppm Run Control Test Control Test 1 23 3 100 0.5 90.2 87.8 .97 10.6 13.3 2 27 6 50 1.0 9K6 91.1 .994 5.2 5.5 3 26 6 50 KG* 9K3 92.6 „„»« 5.8 5.0 4 28 6 50 ( )** 91.6 91.2 .996 4.1 4.3 5 24 6 25 2.0 88.0 85.6 .97 7.7 9.2 * Monosodium phosphate, as P ? ** (1 ppm TSPP, 1 ppm Orthophosphate) , As P 2"5 NOTE: Runs No. 1 and 5 showed a significant difference between control and test waters at a probability of 0.05s, by statistical analysis, 33 TABLE 9 TURBIDITY REMOVALS IN PILOT PLANT TESTS FOR EFFECTS OF PHOSPHATES IN FLOCCULATION AND SEDIMENTATION Series P-VI Test Conditions Coagulant - Ferric Sulfate Phosphate - STP and TSPP, as noted Hardness - Calcium, 125 PP«n; Total, 250 ppm Alkalinity - 300 ppm Rapid Mixing - 500 rpm Flocculation Speed - 4.7 fpm Phos- Coag- phate Avg. Eff i ciency Final Turb id! ty Temp. °C Time Hours ulant ppm ppm Per cent Eff. Ratio ppm Run P^Oj. Control Test Control Test 1 27 3 50 0.5(STP) 86.9 86 . 8 .999 8.1 8.2 2 26 6 50 0.5 (TSPP) 93.1 93.0 .999 4,2 4.2 3 26 3 50 1.0 (STP) 88.8 83.5 .94 7.0 10.0 4 30 6 50 1.0 (STP) 94.4 94.2 .998 3.5 3.6 5 24 6 50 1.2 (TSPP) 90.4 89.1 .986 5.7 6.4 6 27 6 50 2.0(TSPP) 92.7 89.5 .97 4.5 6.4 NOTE . Runs No. 3 and 6 showed a significant difference between control an« test waters at a probability of ©.©5, by statistical analysis. turbidity. The efficiency ratio is defined as the ratio of the efficiencies of turbidity removal in the test and control waters, respectively. The data of each test run consisted of a number of turbidity analyses on influent and effluent samples. From these data, which were subject to the fluctuations associated with sampling, the average efficiencies presented in the tables were computed. The significance of the difference between the average control and average test efficiency for each run was analyzed using the Table of "t 11 , Table IV of Fisher (12). The number of runs in each series which showed a significant difference between test and control efficiencies are noted in each table. Of the 38 test runs reported, 30 showed a significant difference in efficiencies. Of the 8 remaining, 7 were cases of interference having been essentially overcome, while 1 run could be considered to show a significant difference at a probability of 0.10, or ninety times out of a hun- dred. Results for alum coagulation in the presence of STP and TSPP are presented in Tables k, 5, 6, and 8; Series P-|, P-ll, P-lll, and P-V. The introduction of rapid mixing, in series P-Itl, was observed to cause an improvement In the overall process, and with it a reduction in the magnitude of phosphate inter- ference. Series P- IV indicates the benefits of Increased floccul at 5 on=sed 5 men- tation time, with relation to phosphate, interferences. In comparison with the results of series P-lll ft is noted that efficiency ratios of unity are approaches and that quite satisfactory final turbidities are obtained for both test and con" trol. Tables k, 5, 6, and 8 also indicate that TSPP and STP exercise essentially the same effect upon the process., Series P-JV, Table 7 9 consisted of two runs in which ferric chloride was used as coagulant. The significant effect of increasing coagulant dosage upon water containing a given level of phosphate is to be noted. 35 The results of coagulation employing ferric sulfate are presented in Table 9. Several important observations can be made. First, phosphate in- terference was practically eliminated at a dosage of 50 ppm and a phosphate level of 0.5 ppm P^Q,- , when adequate chemical mixing and flocculation were provided. Second, phosphate interferences at levels of 1.0 ppm P^Oj. were virtually elimin- ated at a coagulant dosage of 50 ppm by increasing f loccu II at ion-sedimentation time. Third, effects of STP and TSPP appeared to be essentially the same. Fourth, phosphate levels of 2.0 ppm PJ^q caused some interference, about 3 per cent, even at increased f locculation-sedimentation time. Results for alum coagulation are presented graphically in Figures 5 and 6. Figure 5 is a plot of computed efficiency ratio against alum dosage, for various phosphate levels, and under different conditions of operation. For a given phos- phate concentration, the efficiency ratio tends to increase with alum dosage. Lower phosphate levels are associated with higher efficiency ratios, at a given coagulant dosage. The effect of improved operation is seen in the generally higher level of efficiency ratio for series P-III and P-V. Figure 6 is a plot of per cent turbidity remaining against alum dosage, for different phosphate levels, and under different operating conditions. It is noted that adequate turbidity removals were obtained by increasing coagulant dosage and by increasing f locculation-sedimentation time. Filter Effluent Results : The objectives of the research were limited to the effects of phosphate compounds on the coagulation-sedimentation process. There- fore, the efficiency of turbidity removal, as determined in the settled effluent, was the index of phosphate interference. Since any factor adversely affecting coagulation and sedimentation would tend to Increase the load on the filtration units, efficiency of turbidity removal prior to filtration Is a significant in- dex. However, in order to obtain some general Indication of the way in which 36 1.00 0.95 u c .2 "5 UJ >> U c •*- LaJ <0 c o o II o o £T >» u c © UJ 0.90 0.85 0.80 0.75 7^ / / ^ -s* / S / o *£*- 90" 10 20 30 40 / SERIES p-n p-m P-V 0.5 P2O5 k L 1.0 P2O5 ■ D s 2.0 P2O5 5.0 P2O5 1 50 Alum Dosage, PPM. FIGURE 5. Efficiency ratios for alum coagulation in pilot plant experiments at various phosphate levels. 60 37 c Q) O w O Q. o» O E Q> Alum Dosage , PPM FIGURE 6. Turbidity remaining versus alum dosage in pilot plant experiments at various phosphate levels. 38 phosphates might affect filtration, rapid sand filter effluent turbidities were determined for several test runs. Filter effluent turbidities for both test and control waters ranged from 0.10 to 0.60 ppm in all test runs. Overall efficiencies of turbidity removal varied from 99.^ to 99.9 per cent. Efficiency differences between test and con- trol waters were very small or zero. The differences were not significant. The period of filter operation during each run was not of sufficient length to obtain significant data on rate of head loss increase in the filters. Therefore, the important matter of filter run length remains to be studied. 39 IV. DISCUSSION AND CONCLUSIONS Discussion of Jar Test Results Results of a preliminary series of jar tests previously reported (1), indi- cated that STP and TSPP exerted approximately equal interferences upon the coagu- lation-sedimentation process, and that both of these compounds exerted a much greater interference than did the orthophosphate compounds investigated. These results for alum coagulation in hard waters agreed essentially with the results of Smith, Cohen and Walton (k) for their study of phosphate effects upon alum coagu- lation in hard waters. The jar tests described in this report, using ferric sulfate and alum as coagulants, showed that STP exerted an interference effect upon coagulation of hard waters. The interference was reduced by increased dosages of coagulant. Figure 2 indicates that 40 ppm of ferric sulfate was capable of producing adequate turbidity removals at STP concentrations of up to 2.5 ppm PJ$e« At phosphate levels of 0.5 to 1.0 ppm P«QV » ferric sulfate dosages of 30 to 35 ppm produced adequate removals. As shown by Tables 2 and 3, alum and ferric sulfate produced approximately the same effects, although ferric sulfate appeared to have been somewhat more effective at higher phosphate levels. As reported En the results for ferric sulfate, the effect of phosphate con= cent, rat ion upon the coagulant dosage required for 90 per cent turbidity removal was approximated ""by the equation, F = 2.0 P + 31 . Comparison of this equation with theresults of Smith, Cohen, and Walton, gib/ en in Table 1 of this report, indicates agreement to a fair degree,. Their results for alum coagulation of a water containing 200 ppm calcium hardness and varying amounts of STP could be represented by A - 1.2 P + 2.0 . The apparently better flocculatlon characteristics of their experiments (GT = 57$000, as compared to GT = 5600) , and the higher calcium hardness (200, as against 70) suggest an explanation for the difference in values of the slope and intercept terms. Discussion of Pilot Plant Results The experimental investigation of pi lot scale coagulation and sedimenta- tion demonstrated that condensed phosphate compounds , as represented by STP and TSPP, have the ability to interfere with the removal of turbidity from hard waters. This interference is believed, on the basis of visual observations and quantitative results, to be a matter of opposition to coagulation of floccula- tion and a resultant decrease in floe size and settling characteristics. This is in agreement with the observations of Smith, Cohen, and Walton (4) for hard water f locculation, but differs from those of Howe! Is and Sawyer (3), who studied soft waters. The various series of pilot plant experiments Indicated that the degree of Interference at a given phosphate level was decreased by an Increase In coagulant dosage. This Is seen in the data of Tables 5, 6 S and 7» and Is also demonstrated by Figures 5 and 6. These results are in agreement with the findings of other investigators, for the range of phosphate concentrations studied. As would be expected, any factor tending to lead to poor flocculatlon in general, tended to intensify the extent of phosphate interference. It was ob- served that absence of rapid mixing produced greater phosphate interferences than when rapid mixing was employed. A comparison of Series P-5I and P-I!l, Tables 5 and 6, indicates that the use of rapid mixing speeds of 500 rpm decreased the interference attributed to condensed phosphate. The degree of interference attributed to SIP and TSPP was reduced by an increase of the f locculation-sedimentation time, suggesting that floe settling characteristics or floe size are affected by phosphate compounds. Tables 8 and 9 show the effects of increasing the total theoretical f loccul ation-sedimenta- tion time from 3 hours to 6 hours upon the degree of Interference. Virtual elimination of difference between control and test waters was achieved at coagulant dosages of 50 ppm. On the basis of the tests conducted there does not appear to be a signifi- cant difference between the effect of STP and that of TSPP. Runs 5 and 6 of Series P- 1 1 II , Table 6, showed efficiency ratios of 0.96 and 0.97 at 50 ppm of alum, for 0.5 ppm P^O,- of TSPP and STP, respectively. Runs 2 and k of Series P-1V, Table 8, showed almost identical results for 1.0 ppm P-0,. of STP and TSPP. While the majority of the experiments in the investigation studied coagula- tion with alum, a number of runs were made using iron salts. There was no out- standing difference in the degree of interference observed using iron salts, as compared to alum„ General Conclusions Based upon the studies of coagulation and sedimentation of hard waters having turbidities in the order of 60 to 120 ppm, using jar tests and pilot plant experiments, the following conclusions can be stated: !„ Condensed phosphate compounds, as represented by STP and TSPP, pro- duce interferences with turbidity removal in the coagulation-sedi- mentation process. 42 2. The magnitude of this Interference is reduced by an increase in coagulant dosage. Both iron salts and alum produce approximately the same effects, if the optimum pH range is employed,, 3. The magnitude of the interference is also reduced by increased periods of f locculat ion-sedf mentation. Practical elimination of interference at phosphate levels of 1.0 ppm P^CL or less appears possible, using this method. k. The degree of phosphate interference is affected by adequacy of chemical mixing and f locculat Son. Poor mixing characteristics weaken flocculation opportuni ty, and therefore enhance interference. 5. For condensed phosphate concentrations representative of those pre- sently found in Illinois surface water supplies, 0.5 ppm P«0j. or less, significant interferences were overcome using coagulant dosages of from 30 to 50 ppm. However, the required coagulant dosage for any given installation is dependent upon many factors. A more significant indication is the increased coagulant requirement per unit of phos- phate concentration. This Ss Indicated to be in the order of from 2 to 5 S depending upon the adequacy of flocculation and sedimenta- tion characteristics. Poor characteristics might lead to a factor of 10 to 20. 6. Filtration was capable of producing essentially the same overall efficiencies of turbidity removal for phosphate and zero-phosphate waters. However, rate of head loss development in filters was not studied. It would be desirable to relate, the coagulant dosages required for success- ful treatment in this investigation to those currently in use at water treatment plants. However, coagulation dosages vary considerably with each plant, and many plants in the State of Illinois frequently use dosages equal to or greater than hi the dosages studied in this investigation. Furthermore, the variability in plant flocculation facilities and sedimentation capacities, and increased use of floc- culation aids makes it difficult to directly compare dosages. This investiga- tion has confirmed that mixing facilities, settling time, coagulant dosage, and other variables affect turbidity removals in hard waters. The conclusions must necessarily be limited to a statement of the coagulant dosages required for the conditions of these experiments and of the increase In coagulant demand per unit of phosphate concentration. The latter Is perhaps more significant, as It has been shown to agree in order Of magnitude with results of other investigators. It also provides a more general basis for comparison than does an absolute value of coagulant dosage. in view of the results of this Investigation, It hardly seems justified to indict synthetic detergents as the cause of extreme Interferences with coagula- tion and sedimentation In conventional water treatment. Sat I sfactory process re- sults should be attainable In the presence of current representative levels of condensed phosphates, if reasonable Increases in coagulant dosage are made, and if adequate flocculation and sedimentation opportunity exists. hh V. RECOMMENDATIONS FOR FUTURE RESEARCH Degradation of Condensed Phosphates In order to have a complete understanding of condensed phosphate degradation, a thorough study of the biological factors associated with degradation should be made. Present knowledge indicates that both quantitative and qualatative informa- tion is needed regarding microorganisms and condensed phosphate degradation. Such factors as the kind and source of microorganism, the size of inoculum, the rate and extent of microorganism growth, and the metabolism of condensed phosphates should be studied. The significance of various environmental factors, as related to biological degradation, should also be investigated. These include: organic and inorganic chemical composition of the suspending medium, pH, temperature, dissolved oxygen, and condensed phosphate concentration. Effects of Phosphates on Water Trea tment Study of Soft Water Coagulation : This Investigation has been limited to a study of the effects of phosphates upon hard water coagulation. The work of other in- vestigators has indicated that phosphate interferences may be more severe in soft waters than hard waters. Therefore, an investigation of coagulation interferences over a range of hardnesses, preferably in pilot scale operation, or larger, would be desirable. The reported studies of soft water coagulation to date have all been 1 imi ted to jar tests. Use of Coagulant Aids: New polyelectrolytes appear to be valuable in reducing the concentration of coagulant required to attain adequate treatment. Recent research by Cohen and others (13) has ind electrolyte was successful in overcoming phate, thus reducing the alum dosage requ cated that a synthetic cation) c poly- nterferences by sodium tripolyphos- red. Depending upon their charge, some of the polyelectrolytes can function as true coagulants; while others are h5 capable only of aiding flocculation by increasing floe size and settling rate. In view of the fact that certain polyelectrolytes appear to be capable of over- coming dispersing tendencies, thus reducing the coagulant requirements for good coagulation and sedimentation, an investigation of these compounds should prove valuable for effecting the most economic solution to interference problems asso- ciated with phosphate compounds. The applications of conventional coagulant aids, such as sodium silicate, should also be investigated. 46 REFERENCES 1. Phosphate Compounds in Illinois Surface Waters and Their Effects on Water Treatment Processes. Research Report to A.A.S.&G.P. for the Period June 15s 1955 to June 15$ 1956; University of Illinois. 2. Cross, J. T. Effects of Synthetic Detergent Pollution. Jour. AWWA , 42: 17 (Jan. 1950). 3. Howells, D. H. and Sawyer, C. N. Effects of Synthetic Detergents on Chem- ical Coagulation of Water. Water and Sewage Works , 103:71 (Feb. 1956). 4. Smith, R. S., Cohen, J. M., and Walton, G. Effects of Synthetic Detergents on Water Coagulation. Jour. AWWA , 48:55 (Jan. 1956). 5. Langelier, W. F., Ludwig, H. F., and Ludwig, R„ G. Flocculation Phenomena in Turbid Water Clarification. Proc. ASCE , 78, Separate No. 118 (Feb. 6. Camp, T. R. and Stein, P. C. Velocity Gradients and Internal Work in Fluid Motion. J. Boston Soc. of Civil Engrs ., 30, 219 0943). 7. Fair, G. M. and Geyer, J. C. Water Supply and Waste_Water Disposal. John Wiley and Sons, New York, 1954. 8. Smith, R. S. 5 Walton, G., and Cohen, J. M. Synthetic Detergents and Their Effects on Sewage Treatment and Water Pollution, A Review of the Literature, U.S.P.H.S., June 1954. 9. Task Group Report. Characteristics and Effects of Synthetic Detergents. Jour. AWWA, 46:751 (Aug. 1954). 10. Morgan, J. J. and Engelbrecht, R. S., Effect of Phosphates on Water Treatment. Phase 1. Phosphates in Illinois Surface Waters. Unpublished research re- port. Sanitary Engineering Series No. 5» University of Illinois (March 1958). hi 11. Engelbrecht, R. S., Boice, L. B., and Morgan, J„ J., Effect of Phosphates on Water Treatment. Phase II. Studies on the Degradation of Condensed Phosphates. Unpublished research report. Sanitary Engineering Series No. 6, University of Illinois (July 1958). 12. Fisher, R. A., Statistical Methods for Research Workers. Oliver and Boyd, Edinburgh, 195^. 13. Cohen, J. M. , Rourke, G. A., and Woodward, R. L. Natural and Synthetic Polyelectrolytes as Coagulant Aids. Jour. AWWA , j>0:463 (April 1958). ft/