Citation
Electrophoretic studies of turbidity removal by coagulation with ferric sulfate.

Material Information

Title:
Electrophoretic studies of turbidity removal by coagulation with ferric sulfate.
Creator:
Walters, James Vernon, 1933-
Publication Date:

Subjects

Subjects / Keywords:
Coagulants ( jstor )
Coagulation ( jstor )
Colloids ( jstor )
Dosage ( jstor )
Ions ( jstor )
Kaolinite ( jstor )
Montmorillonite ( jstor )
pH ( jstor )
Sulfates ( jstor )
Turbidity ( jstor )
City of Vernon ( local )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright James Vernon Walters. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
20127810 ( OCLC )
0025035127 ( ALEPH )

Downloads

This item has the following downloads:


Full Text












ELECTROPHORETIC STUDIES OF

TURBIDITY REMOVAL BY COAGULATION

WITH FERRIC SULFATE










By

JAMES VERNON WALTERS


A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY












UNIVERSITY OF FLORIDA
August, 1963














ACKNOWTED kINTS


The author wishes to express his gratitude to his cormzittee

chairman, Dr. A. P. Black, for guidance and encouragement he has given during the investigation and for his generosity in the giving of himself to his students. He deeply appreciates the friendship and helpfulness of Prof. J. E. Kiker, Jr., Prof. G. B. Morgan, Jr., and Prof. T. deS. Furman, members of his supervisory committee, who have helped him in many ways during his graduate study at the University of Florida. Dr. T. R. Waldo, of his supervisory committee, has generously given her time to guide the author in his dissertation preparation. He gratefully acknowledges his indebtedness to her.

Dr. Norihito Tanbo's sketch of the Brigs cell which is included herein, will remind the author of the hours of consultation Dr. Tanbo gave him. Mrs. A. L. Smith, Dr. R. F. Christman, and Dr. S. A. Hannah have also given the author the advantage of their experiences in water

coagulation research. The author thanks Mrs. J. G. Larson, Mr. W. T, Walters, and Mr. C. Chen for their help in the execution and reporting of his experiments.

The research was directly supported by Water Supply and Pollution Control Research Grant WP-139 from the Public Health Service, and was indirectly supported by Ford Foundation loans and Public Health Service Traineeships which financed the author's graduate study.








The author shall forevermore try to express his appreciation and gratefulness to his sons and his wife, Barbara, whose love and understanding have sustained him.














CONTETS


Page


ACKNOWLEDG27NTS . . . . . . . . * � � * � � � � * . . . * 0


LIST OF TABLES . . . . . . . . .


LIST OF FIGURES o � � . � � � � � � � � � � � � � � � CHAPTER
1, INTRODUCTION . . . * * *

1i, HISTORICAL REVIEW . o. . . . o o . . . . o


III. COAGULATION THEORY . .


a a * 0 9 a 0 09 * 0* 0 * 0*


V. EXERI=ZTAL MATERIALS AND PROCEDURES .


MatUerials a 9
clays . . . 0 0 9 � 9
Ferric Sulfate... Procedures # o *


*� . � 0 �
*@ � @ � O
* 9 0 �
* 0 0 �


Preparation of Clay Suspensions Preparation of Ferric Sulfate Coagulation Tests . . . . . . .
Sample Preparation . . . *.
Flocculation .. . . , o , .


*90 4 * * 0


* O 0 � � @ @ �0 � � � @* � @ � @4 � @. 0 � @ � 9 � �. �9
* . 9 �
* 9 �0


Initial and Residual Turbidity Measurements
Residual Iron Deterinations .... � � �
Measurement of pH . . �.. 0 0 Electrophoretic Mobility Determinations . . . .
Dosing and Xi g . e * e * * 9 � * * * *
Conductance Measurement . . . ..... �
MeasurementofpH .
Particle Mobility Measurement .


'VI. DISCUSSION OF RESULTS


0 9 * 0 * 0 * * * 9 v O 9


VII. CONCLUSIONS . . . . . . . . . . . . . . . .

APPENDIX . . & � * 9 0 9 * 9 * 9 - � 9 * � � 9 * 9 9 & 9 � * *


0 9 , #** 0 a S * 9 0 0 0 *


vi vii


111. PURPOSE AND SCOPE . . . . . . . . . . . . .. . . e







Page

BIBLIOGRAPHY .. � � 0 � 0 0 0 � S � � � � 0 � � 0 � 0 � 0 � 102 BIOGRAPICAL SKETCH . . . . . 1 . . . . . . . . . . . . . . .07














LIST OF TABLES


Table Page

1. Analysis of Ferric Sulfate Used for Coagulation . . . 27 2. Properties of the Clay Suspensions. . .-. . . * . . . 38

3. Coagulant Dosages ...................... 39

4. Effect of pH on Electrophoretic Mobility and Sedimentation of Montmorillonite Clay . . . .*. . . . 79

5. Coagulation of Montmorillonite Clay With 3.0 Milligrams Per Liter of Ferric Sulfate . . ..... 80

6. Coagulation of Montmorillonite Clay With 5.0 Milligrams Per Liter of Ferric Sulfate .. ....... 8

7. Coagulation of Montmorillonite Clay With 50
Milligrams Per Liter of Ferric Sulfate . . . . . . . 85

8. Effect of pH on Electrophoretic Mobility and Sedimentation of Fuller's Earth . . . . . . . . . . 87

9. Coagulation of Fuller's Earth With 3.0 Milligrams Per Liter of Ferric Sulfate ............ . . . . . 89

10. Coagulation of Fuller's Earth With 5.0 Milligrams
Per Liter of Ferric Sulfate ..... ............. 91

11. Coagulation of Fuller's Earth With 50 Milligrams
Per Liter of Ferric Sulfate ............... . 93

12. Effect of pH on Electrophoretic Mobility and
Sedimentation of Kaolinite Clay . . . . .......... 94

13. Coagulation of Kaolinite Clay With 3.0 Milligrams
Per Liter of Ferric Sulfate .............. ...95

14. Coagulation of Kaolinite Clay With 5.0 Milligrams
Per Liter of Ferric Sulfate ............... . 97

15. Coagulation of Kaolinite Clay With 50 Milligrams
Per Liter of Ferric Sulfate . . . . . . . . . . . . . 100













LIST OF FIGURES

Figure Page

1. The Helmholtz Layer Model .... .............. . . . 16

2. The Gouy-Chapman Diffuse Layer Model. . . . . . . . . . . 18

3. The Stern Layer Model .................... 20

4. The Briggs Cell .................... 35

5. The Effect of pH and Mobility upon Coagulation. Clay:
Montmorillonite Ferric Sulfate Dosage: 0.0 rag/1 . . . . 42

6. The Effect of pH and Mobility upon Coagulation. Clay:
Montmorillonite Ferric Sulfate Dosage: 3.0 mg/l . . . . 44

7; ThpT ffegt o. p11 and Mobili6y upon Coagu1 ati . Clay:
morktmorilioit'o Fe rYI o 3oulatr Dosage; 6-0 g19 1 45

8. The Effect of pH and Mobility upon Coagulation. Clay:
Montmorillonite Ferric Sulfate Dosage:, 50. mg/1 . 46

9. The Effect of pH and Mobility upon Coagulation.
Clay: Montmorillonite ....... . ................. 48

10. The Effect of pH1 and Mobility upon Coagulation. Clay: Fuller's Earth Ferric Sulfate Dosage: 0.0 mg/1 . . .. 49 11. The Effect of pH and Mobility upon Coagulation. Clay: Fuller's Earth Ferric Sulfate Dosage: 3.0 mg/1 . . . . 50 12. The Effect of pH and Mobility upon Coagulation. Clay: Fuller's Earth Ferric Sulfate Dosage: 5.0 mg/1 . . . . 52 13. The Effect of pH and Mobility upon Coagulation. Clay: Fuller's Earth Ferric Sulfate Dosage: 50. mg/1 . . . . 53 14. The Effect of pH and Mobility upon Coagulation. Clay: Fuller's Earth . . . ....... . . .................... 54

15. The Effect of pH and Mobility upon Coagulation. Clay: Kaolinite Ferric Sulfate Dosage: 0.0 mg/i . . . . . . . 55 16. The Effect of pH and Mobility upon Coagulation. Clay: Kaolinite Ferric Sulfate Dosage: 3.0 mg/1 . . . . . . . 56









Figure

17. The Effect of pH and Mobility upon Coagulation. Clay: Kaolinite Ferric Sulfate Dosage: 5.0 mg/i . . . .... 18. The Effect of pH and Mobility upon Coagulation. Clay: Kaolinite Ferric Sulfate Dosage: 50. mg/l. ..........

19. The Effect of pH and Mobility upon Coagulation.


Clay: Kaolinite .


20. The Effect of pH on Iron Residual.
Ferric Sulfate Dosage: 3.0 mg/l .

21. The Effect of pH on Iron Residual.
Ferric Sulfate Dosage: 5.0 mg/1 . .

22. The Effect of pH on Iron Residual.
Ferric Sulfate Dosage: 50. mg/1 . .

23. The Effect of pH on Iron Residual.
Ferric Sulfate Dosage: 3.0 mg/! . .

24. The Effect of pH on Iron Residual.
Ferric Sulfate Dosge: 5.0 m/1 . .

25. The Effect of pH on Iron Residual.
Ferric Sulfate Dosage: 50. mg/1

26. The Effect of pH on Iron Residual.
Ferric Sulfate Dosage: 3.0 mg/1 . .

27. The Effect of pH on Iron Residual.
Ferric Sulfate Dosage: 5.0 mg/1 .

28. The Effect of pH on Iron Residual.
Ferric Sulfate Dosage: 50. mg/1 . .

29. Quantities of Acid or Base Required
of Fuller's Earth . . . . . ....

30. Quantities of Acid or Base Required
of Fuller's Earth . . . . . ....

31. Quantities of Acid or Base Required
of Fuller's Earth . . . . . ...

32. Quantities of Acid or Base Required
of Fuller's Earth . . . . . ....

33. Quantities of Acid or Base Required
of Fuller's Earth . . . . . .


. . . . . . . . . 0


Clay: Clay: Clay: Clay: Clay: Clay:


Clay: Clay: Clay:


Page 57 58 59


Montmorillonite Montmorillonite Montmorillonite Fuller's Earth Fuller's Earth
� . 9 9 9 �


Fuller's Earth Kaolinite Kaolinite


Kaolinite


for pH Adjustment for pH Adjustment for pH Adjustment for pH Adjustment for pH Adjustment


Vi i


. � . � 0 * . . � .


� q













I. INTRODUCTION


Clays are the most common source of turbidity in surface waters used for municipal and industrial water supplies. Before surface water is satisfactory for domestic and industrial use, most of the clay and other particulate matter present in the water must be removed. Removal of the suspended matter is usually accomplished by alum or ferric sulfate coagulation, sedimentation, and rapid sand filtration. Water treatment plants which utilize this process produce finished waters which regularly exceed the minimum quality required for potable water by the U. S. Public Health Service.1'2'3 The specific Public Health Service recommendation concerning turbidity is that the turbidity of drinking water be less than five units.

Prediction of the optimum coagulation conditions for the production of such high quality water from a given raw water is very difficult. In the absence of records of previous treatment of water from the same source it is rationally impossible at present. The difficulty of coagulation prognosis is the result of complexly related effects of the numerous properties of the raw water and the chosen coagulation conditions which affect the efficacy of coagulation. Among these conditions and properties are: specific coagulant chosen; coagulant dosage; pH, alkalinity, ionic constituents, and base exchange capacity of the raw water; and size, shape, chemical nature, hydration, and charge of the colloidal particles in suspension. Most of the factors -hich affect coagulation also affect the electrophoretic mobility of the suspended


-1-






-2-


colloidal particles.

Electrophoresis is the movement of electrically charged particles, suspended in a conducting liquid medium, which results from the impression of an electric field. Electrophoretic mobility is the ratio of the speed of electrophoretic movement to the intensity of the electrical field which produced the motion. It is commonly expressed in microns per second per volt per centimeter.

Mineral content, coagulant, coagulant dose, alkalinity and

nature of clay particles, of the numerous parameters which affect electrophoretic mobility and coagulation, can be chosen and controlled for a selected synthetic clay suspension. In addition to these parameters which can be selected, pH, base exchange capacity, effectiveness of coagulation and particulate electrophoretic mobility for a given suspension can be directly measured.

The present research has been performed in order to study the empirical relationships among electrophoretic mobility, residual turbidity, coagulant dosage, and residual iron (the mensurable parameter) for the ferric sulfate coagulation of suspensions of three different clays over the pH range between three and ten. To my knowledge, such a study has not been attempted before. The relationships discovered are described herein, and the effectiveness of ferric sulfate coagulation of these clays is compared with that of their alum coagulation, reported earlier by Black and Hannah.4














II. HISTORICAL REVIEW


The earliest records of the coagulation process have been traced by Black who wrote:

Although various crude methods of atez puX-ication generally
characterized as coagulation, have been kno. and used .ince
ancient times, knowledge of the Thndamental factors involved in
the process has been acquired comparatively recently.

The earliest references of scientific interest in coagulation as a process for the treatment of water are references to the works of DtArcet6'7 and Jeunet.6,8 DtArcet at the beginn.g of the nineteenth century and jeunet in 1865 sought to establish the value of the process, but it rs not used for the treatment of a public water supply until 1881. After its initiation in Bolton, England,6 the process was soon adopted in Holland and in the United States.

The first coagalation patent was granted, in 1884, to Isaiah

Smith Hyatt.5 Following the suggestion of Col. L. H. Gardner, Superintendent of the New Orleans Water Company, Hyatt successfully txeat turbid water by combining the use of perchloride of iron as a coagulant vith his process of rapid filtration. His patent covered not only the use of perchloride of iron, but also of "any other suitable agent

which is capable of coagulating the impurities of the liquid and preventing their passage through the filter bed."

The 1884 Annual ecort of the State Geologist of New Jersey contained results of tests of various salts as coagulants. Austen and Wilber6,9 concluded that of the salts investigated, aluminum sulfate 10
was most effective. Fuller published a description of similar studies

-3-






- 4

in 1898. He found sulfates of iron and aluminum to be most effective. The chloride of these elements followed next in order, but their use as coagulants in water treatment has not developed on a practical scale.

In the year of Fuller's publication, W. B. Bull,5 at Quincy,

Illinois, began using a mixture of ferrous sulfate and lime for coagulation. Fourteen years later, E. V. Bulll reported the first use of chlorinated copperas. This chemical was not tried again until 1928, when Hedgepeth and Olsen12 used it for the successful treatment of a highly colored water. Ferric ion was produced by oxidation of the ferrous sulfate with chlorine.

As use of the process has become prevalent, the number of researchers and the scientific disciplines they represent have proliferated. Black13 and Packham6 in recent reviews of coagulation theory and related literature, thoroughly cover the many facets of historical and contemporary research, theory, and practice. The literature cited in the present study will consequently be limited to that directly

related to the research herein reported.

In 1923 the first of a series of studies by Theriault, Clark,

and Miller was reported. The paper by Theriault and Clark14 described their treatment of several buffered solutions with various amounts of alum to determine the effect of pH upon the rate of floc formation. They assumed that the optimum pH for coagulative treatment of water would be the pH at which the minimum time was required for the formation of the aluminum floc. They found that generally the best floc formation occurred in the pH range 4.95 to 5.40. Higher alum doses resulted in broader ranges of pH over which rapid coagulation was observed. These workers were unable to explain the difference between values of pH






-5
which accompanied rapid coagulation in their study and the higher values observed to be most effective in water treatment plants.

In the same year, Baylis,15 using alum to coagulate natural waters of varying alkalinity, noted that for a given pH, waters of higher alkalinity required larger quantities of coagulant to accomplish satisfactory treatment. For any chosen water he was able to reduce the alum dosage by adjusting the pH with a strong acid, thereby neutralizing a portion of the alkalinity originally present. The values of pH for optimum coagulation were in the range 5.5 to 7.0.

Late in 1923, Miller16 published the second of the Theriault,

Clark, and Miller papers. He described their constituent study of precipitates formed by the mixing of potassium alum solutions with sodium hydroxide solutions of various molarities. Although the anion concentration of the coagulant had no practical effect, the amount of alkali mixed with a given quantity of aluminum exerted considerable influence upon the composition of the precipitate. The ratio of aluminum to sulfate ion was determined for the floc over the useful pH range, but the exact chemical nature of the floc was not determined. His review of treatment plant records revealed that optimum coagulation occurred within the pH range 5.4 to 8.5.

In another series of tests (1925) Miller17 investigated the

effects of various anions upon the formation of floc in solutions of potassium alum. Generally, the anions of higher valence exerted greater influence upon coagulation, and-less alkali was necessary in their presence.

The composition of precipitates from ferrous and ferric salt solutions was the subject of another 1925 publication by Miller.18






- 6

There precipitates from the iron salts were found to be very similar to floe from the analogous aluminum salts, except that they were formed at pH values significantly lower than the minimum pH values observed for alum floc and that they remain insoluble above the maximum pH at which alum floe exists.

The results of Miller' s experimental work allowed him to make three statements concerning the optimum conditions for floe formation with both iron and aluminum coagulants. (1) There must be present in the water a certain minimum quantity of the metallic coagulant ion.

(2) There must be present an anion of strong coagulating power such as the sulfate ion. (3) The pH must be properly adjusted.

In 1928 Bartow and Peterson19 examined the effects of various salts upon the formation rate for alum floe over the useful pH range. Many of the salts studied increased slightly the rate of floe formation and extended," on the acid side, the pH range over which floe occurred.

In a paper published in 1928 and now regarded as a classic in its field, Mattson20 described the coagulation of clay suspensions and the electrophoretic mobility measurement of clay particles as mobility was altered by various dosages of aluminum salts. His work demonstrated the effect of the different anions and cations upon particle charge. He specifically predicted the utility of electrophoretic studies in water coagulation investigations and pointed out the predominant role played by the hydrated oxides formed during coagulation.

In three articles Black and others21,22,23 report investigations, made on a laboratory and semi-plant scale, of the effects of several anions upon the rate of floc formation for both aluminum and ferric salts. They proposed improved conditions for laboratory study of







-7

coagulation. Also they recommended the use of larger samples and containers, constant stirring, and isohydric indicators. With such improvements they could duplicate pilot plant results. Anions were found to be useful to increase the rate of coagulation and to broaden the pH ranges of rapid floc formation. The pH range for floe formation was wider for ferric sulfate than for alum, but no iron floe formed between pH values 6.5 and 8.5. The authors stressed the importance of adsorption as an important mechanism of turbidity removal and also attributed zones of no floc formation to a change in particle charge.

In the slightly more than a decade since the publication of

Black's 1934 article, the interest of water chemists seem to have been focused upon activated silica and other coagulant aids. The work of
Ir ~ 2492
Langelier and Ludwig 4 25 marked the return of many researchers to the field of basic coagulation theory. They initiated the use of synthetic clay suspensions and called attention to the importance of base exchange capacity and particle size distribution of raw water, the action of metallic hydrolysis products as "binders," and the occurrence of perikinetic coagulation.

In 1958, Pilipovich, Black, and others26 examined the relationships among pH, zeta potential, base exchange capacity, coagulant dosage, and turbidity removal for the coagulation of suspensions of five clays with alum. Clays of high base exchange capacity required considerably larger doses of coagulant to accomplish coagulation than did clays of lower base exchange capacity. The authors reaffirmed the superior effectiveness of hydrolysis products over that of trivalent aluminum ions as coagulating agents.

Matijevic and others27 (1961) coagulated lyophobic colloidal






-8

suspensions of known particle charge and concentration with aluminum nitrate to determine the nature of the species formed during alum floc formation. Results indicated that below pH 4 the simple trivalent hydrated aluminum ion prevails. Between pH 4 and pH 7 the probable prevailing complex is the tetravalent A18(OH)20. At higher pH values the tetravalent hydrate is transformed into a divalent form which in turn yields to a complex of zero charge.

In the same year Black and Hannah,4 using alum and various polyelectrolyte coagulants and coagulant aids, investigated the relationships among pH, electrophoretic mobility, base exchange capacity, coagulant dosage, and turbidity removal for the coagulation of three different clay suspensions. Several equivalents of coagulant per equivaIet of clay base exchange capacity were necessary to effect satisfac. tory coagulation. The pH range of best coagulation was 7.5 to 8.5 in which range the electrophoretic mobilities of the clay particles was slightly negative. In some cases fair perikinetic coagulation occurred at pH values less than 4.5.

In his study of the precipitation of aluminum hydroxide resulting from addition of alum to solutions containing bicarbonate, carbonate, or chloride of sodium, Packham28 (1960) evaluated the degree of precipitation by measuring soluble aluminum residuals and by photometrically recording turbidity of the system as a function of time. The maximum amount of precipitate was formed in the pH range 5.5 to 7.2. Maximum coagulation rate occurred between pH 7.2 and 7.6, and maximum immediate precipitation was observed at pH 7.1 or 7.2.

As his criterion of coagulation effectiveness, Packham chose the coagulant dosage required to reduce the initial turbidity by one-half






-9

for a given set of conditions. The pH range he observed for optimum coagulation conditions based upon the use of this criterion does not agree with the pH range for good coagulation which others have reported. The difference results from the selection of different coagulation efficiency criteria. Use of Packham's criterion to select conditions for coagulation in a treatment plant would result in water of unsatisfactory quality. If he had chosen a much greater turbidity removal efficiency as his criterion, his prediction of optimum coagulation pH range would have coincided with those commonly reported.

Packham expressed the opinion that the coagulation of clay is

principally accomplished by the enmeshment of the clay particles in the metallic hydroxide precipitate formed by reaction of the coagulants. Orthokinetic coagulation is the term now used to name the mechanism he described.

Mackrle29 (1962) has presented arguments favoring the acceptance of the physical theory of stability and coagulation of colloidal clay suspensions. He has attached greater significance to the value of the psi potential, the potential difference between the "Helmholtz layer" and the bulk of the suspending medium, than to the value of the zeta potential, the potential difference between the plane of shear and the

bulk of the suspending medium. He has suggested that the metallic coagulants react to form crystalline hydrous oxide sols which destabilize the clay suspension by mutual coagulation. His conclusions were based upon his work with ferric sulfate and clay suspensions and upon crystallographic studies of such precipitates reported by others. In his tests pH and coagulant dose were allowed to vary simultaneously, and no measurements of electrophoretic mobility were performed to






- 10

permit an examination of the independent relationships among pH, coagulant dosage, turbidity removal and zeta potential.

In 1962 Stumm and Iorgan30 cited the work of others and identified, for both iron and aluminum salts, the hydrolysis products of highest effective charge and assigned tentative formulae to them. They also called attention to the aging effect of iron and aluminum sols and emphasized the chemical nature of particle charge caused by ionization as contrasted with charge produced by physical adsorption. The majority of their laboratory tests were performed to determine the specific chemical interaction between various functional groups normally found in water and the metallic coagulants in common use. They concluded that metallic complexes other than hydrous oxides can be formed as a result of the interactions and that the Lateraction of the functional groups can appreciably affect the pH at which optimum conditions of coagulation occur. They also presented a method for carrying out laboratory jar tests at constant pH and convincingly justified the desirability of such a procedure.













11. COAGULATION THEORY


Any explanation of coagulation theory must be preceded by a description of the physical and chemical characteristics of the colloidal suspensions to be coagulated. The suspensions customarily encountered in the treatment of water for municipal and industrial use are dilute. Even here raw water is taken from streams which carry heavy sediment loads, the water to be coagulated is dilute because presedimentation is used to remove the readily settleable particles. The materials 'which are removed from the dilute suspensions by coagulation can be generally classified as turbidity or "color."

In their study of the nature of colored surface waters Black and Christman31 have ascertained that "the materials responsible for color in water exist primarily in colloidal suspension in the water." They found particle size and number of particles to vary with pH, but particle size was generally less than 10 rap. Although the present research is not directly concerned with coagulation for color removal, color compounds do merit mention, because their colloidal nature and susceptibility to coagulation are similar to those of colloidal turbidity particles.

The primary source of turbidity to which the present investigation pertains is colloidal clay. Clays are complex aluminum silicates of sedimentary origin. Their lattice structures consist of layers of silicon-oxygen sheets and aluminum-oxygen sheets. These sheets, common to all clays, occur in different orders in the different classes of clays.


- ll-






- 12

Kaolinte, montmorillonite, and miller's earth have been used in the present research.

An important physico-chemical characteristic of clays which is considerably affected by the placement order of the sheets is base exchange capacity. Kaolinite and montmorillonite, of the clays used in the present research, represent the low and high extremes of base exchange capacity which results from crystalline structure.

Kaolinite consists of an alumina sheet and a silica sheet -wich are combined to form one layer, -hereas montmorillonite has an alumina sheet included between two silica sheets to form one layer. These layers have a definite thickness, measurable by x-ray diffraction

methods, which is constant for a specific clay; but the lattices extend to the irregularly broken edges of the crystals in the other two directions. In water the parallel layers tend to split apart and become dispersed. Because of the small coulombic forces between similar adjacent silica sheets in montmorillonite, the surfaces of the layers become hydrated and therefore separate to a considerable extent. Kaolinite layers are held together tightly by hydrogen bonding between hydroxide ions on the bottom of one layer and oxygen ions on top of the next layer and thus resist separation. A more detailed discussion of clay structure is given by Hendricks.32

Marshall and Krinbil133 have stated one of the results of the

structural differences between kaolinite and montmorillonite. The base exchange capacity of kaolinite varies with particle size, whereas the base exchange capacity of montmorillonite is almost independent of particle size.

Particle size is the property upon which the definition of






- 13

colloids is based, but colloid behavior results from phenomena involving the tremendous surface area inherent in particles of colloidal size. An example of the importance of surface phenomena is that the classifications of colloids, lyophilic and lyophobic, result from such a phenomenon. Distinction between the two classes depends upon the wettability of the surface of the particle. The particles possessing poorly wettable surfaces are called lyophobic. Lyophobic colloids are unstable, because by coalescing they offer a smaller surface area at which a solid-toliquid interface must be formed. Since the smaller interface results in a lower surface energy state and a lower total energy state for the system, the coalescence is spontaneous.
Lyophilic colloids, on the other hand, are stable because of the attraction of the surface for the solvent .- the wettability of the surface. These lyophilic colloidal solids which exhibit surface interaction with the solvent are, therefore, the colloids of primary concern to water chemists.

Another surface phenomenon which greatly influences the properties of a suspensoid is the electrical charge possessed by the particles. The charge can be initiated by several different surface mechanisms; still other surface phenomena affect charge magnitude and sign.

Of the several mechanisms which can produce particle charge, we shall first consider the process that might be termed preferential ionization. As an example of this mechanism, we shall consider minute crystals of silver iodide in equilibrium with a saturated solution in which the ionic concentration product is roughly 10-16. The force with which the two ions are held in the lattice differ greatly. Despite its larger size, the more polarizable iodide ion is held much more






- 14

forcefully than is the silver ion. When the concentrations of the two ions in the solution are equal, more silver ions escape into the solution than do iodide ions; hence, the crystals are negatively charged. If the silver ion concentration is increased, the charge can be reduced to zero; a further increase of silver ion concentration can cause a

charge reversal.

Ionization of functional groups which are connected to the

particle by covalent bonds is a second and important source of particle charge. Examples of particles thus charged are proteins and ion exchange materials. Proteins are formed by long chains of amino acids joined by peptide linkages. Some of these amino acids carry an additional carboxyl or amino group -hich can remain free and exposed to the solvent. Thus they can form CO0" and NH 3+ ions which are covalently attached to the particle. The ion exchange materials are porous solids wich have built-in acids or basic groups. The natural zeolites and many clays are included in this general category.

Among the other mechanisms which can cause particle charge are chemisorption and adsorption of specific ions resulting from van der Waal's forces. For particles which are charged 1y the adsorption of specific ions, the surface charge density is greatly affected by concentration changes of those ions in the solution.

In all of the mechanisms mentioned above, particle charge was

achieved by the separation of unlike charges (the destruction of electroneutrality). Because of the large amount of energy involved in the separation of these charges, electroneutrality is disturbed only on the ultramicroscopic scale. When any one, or a combination of these mechanisms, produces an excess of charges of one sign in any locality,






- 15

there must exist an equal excess of charges of opposite sign in the immediate vicinity. Thus, we find surrounding and very near each charged particle an excess concentration of ions of counter sign. These ions are called counterions or gegenions. The counterion excess occurs only in the solution very closely surrounding each charged particle. Throughout the rest of the suspending medium electroneutrality prevails.

When we speak of electroneutrality and areas of excess charge, we implicitly assume that the conditions we describe are the averages, with respect to time and space, of the several dynamic equilibria which prevail. Brownian movement is one spatial equilibrium which must be averaged. Ion exchange is a physico-chemical equilibrium of considerable effect. Each of the equilibria, including these two, cause fluctuations with respect to time.

Based upon these averaged conditions, several models have been proposed to explain the possible arrangement of the charged surface and its surrounding array of counterions. These models have been used by chemists, physicists, and mathematicians as bases for their computation of theoretical values of parameters for many colloid systems. Probably none of the models proposed is correct, but each has been useful in the explanation of some observed phenomenon.

The simplest of these was proposed by Helmholtz.34 Fig. 1 is a schematic section which shows the charged surface and the surrounding medium. By ionization or some other mechanism, the surface of the particle is charged. This excess of negative charge in or on the surface attracts toward the surface the positive ions in the solution. Helmholtz proposed a model in which all of the counterions so attracted





- 16 -


CHARGED SURFACE -


PARTICLE


V-HELMHOLTZ LAYER OF COUNTERIONS
SURFACE OF SHEAR


0 ,0.BULK OF
K SOLUTION
' 1: 0============= :'� ::i::::�9 ..: .::::.. .., . .
9 I �::: :: :i i!i !::::~:: :iiii~ ::": : :: : . ..: iii:: � u �N:!:IIL : !iii


BOUND
SOL VE NT


DISTANCE


-THE HELMHOLTZ LAYER


FIG. 1


MODEL






- 17

are located in a single surface parallel to and very near the surface of the particle. His model is analogous to a parallel plate condenser. The graph of potential in Fig. 1 represents the potential difference between the bulk of the solution and the point within the solid-liquid interface that corresponds with the abscissa.

Fig. 2 depicts the Gouy-Chapman35 model which is named for the two scientists who first considered it in detail. They recognized the Helmholtz assumption to be an oversimplification of the spatial arrangement of the neutralizing counterions. Reasoning that thermal agitation would prevent such an uniform arrangement of gegenions, they proposed a diffuse layer of variable but finite thickness which contains an excess concentration of counterions. In formulating their model they also considered the shielding effect of the counterions located near the particle upon the attractive force existing between the particle and the more distant counterions. The proposed model is characterized by a potential versus distance graph, for which the potential decreases at a decreasing rate as distance from the surface of the colloid increases.

In addition to the gegenions which are attracted to the particle, a sheath of the solvent is tightly bound to lyophilic colloids. This sheath is pulled along with the colloid whenever some force such as that of gravity causes relative motion between the particle and the bulk of the solvent. Thus, there exists surrounding a colloid a surface of shear which contains the bound solvent that accompanies the particle as it migrates.

If an electric field is impressed upon a colloid system, the colloids migrate toward the pole of opposite charge at a speed










CHARGED SURFACEPARTICLE


- 18 -


GOUY-CHAPMAN DIFFUSE LAYER


010







G
-SURFACE OF BOUND- SOLVENT


BU LK OF SOLUTION




SHEAR


PSI POTENTIAL S-ZETA POTENTIAL


DISTANCE


FIG. 2 -THE GOUY-CHAPMAN
DIFFUSE LAYER MODEL






- 19

proportional to the potential which exists between the surface of shear and the bulk of the solvent. (The "bulk of the solvent" refers to the portion of the solvent out of the effective interaction range of the charged colloids, that is, where electroneutrality occurs on micro as well as macro scale.) This shear surface potential upon which the velocity of a charged colloid in an electric field depends is called the zeta potential,t . The movement of the charged colloid that is caused by impression of the electric field is referred to as electrophoresis; and the ratio of particle speed to the intensity of the electric field is termed electrophoretic mobility. In units commonly used for expression of mobility and potential, a lP/sec/v/cm mobility is equivalent to a potential of 13 mv.

Since the velocity of colloid movement caused by a measurable field intensity can be observed through an ultramicroscope or by moving-boundary frontal methods, the electrophoretic mobility of colloids is a directly measurable parameter. This parameter is closely related to particle charge, surface-charge-density diffuse-layer thickness, and particle surface potential, which are important parameters but are not directly measurable. (The surface potential, *, mentioned above is diagramatically presented in Fig. 2).

A model suggested by Stern36 represents further sophistication of the Gouy-Chapman diffuse layer theory. It includes the Stern layer of adsorbed counterions which are held in actual contact with the surface of the colloid. A section of the Stern model and a typical potential curve for it appear in Fig. 3. In addition to the zeta potential and the surface potential, the potential at the interface between the Stern layer and the Gouy-Chapman portion of the double diffuse










CHARGED SURFACE -PARTICLE


- 20 -


r-STERN


LAYER


SURFACE OF




GG BOUND SOLVENT


SHEAR


BULK OF SOLUTION


PSI POTENTIAL-0,







E d
=EA POTENTIAL


DISTANCE


FIG. 3 -THE STERN LAYER MODEL






- 21

layer is indicated and represented by the symbol'.

The models described above are useful as they provide a basis

for appreciating the phenomena which influence the stability or instability of a given colloidal suspension. The primary phenomenon favoring stability is, of course, the mutually repulsive force between similarly charged colloids.

A particle with its complete double layer is electrically

neutral, so that it exerts no net coulombic force upon another. This situation exists in the case of particles which are sufficiently distant from each other. As two particles approach, however, the double layers interpenetrate and interact. Should the two particle surfaces finally touch, there could be no more diffuse layers between them to screen them from the effect of the repulsive coulombic force. It is the work required to thus distort and finally destroy a part of the diffuse double layers which causes most of the repulsion.

Charge density upon the particle surface and concentration of

ions in the solution surrounding a particle are parameters which directly affect stability. Low ionic concentration in the solution causes greater thickness of the diffuse layer. High charge density results in large psi potential. Thus low ionic concentration and high charge density contribute to colloid stability. The relationship of these parameters to zeta potential can be seen in Fig. 1. The conditions which contribute to stability result in high values of zeta potential; therefore, zeta potential, or more important, the directly measurable electrophoretic mobility can be a useful index of stability. A less important phenomenon contributory to stability is the van der Waal adsorption of similiions. This occurs to a significant degree






- 22

only under very specialized conditions, and the effect is more important with similiions of large molecular weight.

The phenomena which are influential toward instability are much more numerous than those causing stability, because the former consist of all the phenomena which can impair the effectiveness of the latter. Of the instability mechanisms, the two most important are the mutual attraction (and coagulation) of oppositely charged colloids and the van der Waal attraction. The other mechanisms are of much less direct importance, or they effect instability through their influence upon or in conjunction with one of the two mechanisms mentioned.

Van der Waal forces are always attractive; but because they result from dipole interaction, they decrease approximately with the third power of the distance between the particles. Thus they are effective only when the particles are brought into extreme proximity. Coulombic forces between similarly charged colloids oppose their approach toward each other with force which decreases roughly with the second power of the distance between the particles, so that some other force is required to push them close enough together for van der Waal forces to prevail. Brownian motion caused by thermal agitation is one source of such force. Mechanical agitation of a suspension can similarly contribute to instability.

Some other factors effect instability by reducing the particle charge or potential of the colloids. The concentration of "potential determining ions" in the solution exerts a large influence upon this parameter. The hydrogen ion concentrations will determine the charge density and, hence, the potential of any particle which depends upon the ionization of carboxyl or amino groups as its source of charge.






- 23

The concentration of inert ions in the solvent can affect stability by increasing the probability of particle collisions as their number is increased and by causing compaction of the diffuse layer.

Ionic concentration is not the only ionic variable which

influences stability. The valence and size of the specific ions present exert effects upon particle charge and potential. The Schulze-Hardy rule is a general statement of the observed phenomena that divalent counterions are more potent in the production of instability than monovalent counterions and that trivalent ions are much more powerful coagulation agents than the divalent counterions are. Ion size and less important anomalous characteristics of specific ions influence their ability to cause colloid instability as is evidenced by the Hoffmeister series in which various anions and cations have been listed in order of flocculating power.

Because of the numerous mechanisms which are responsible for destabilization, the many variables which affect stability, and the

differing chemical and physical nature of the sundry colloidal particles which may be encountered, it is practically impossible to calculate the conditions under which satisfactory coagulation of a given colloidal suspension will occur. It is desirable, therefore, to examine the empirical relationships of the parameters that can be measured so that the work involved in cut-and-try coagulation can be reduced.














IV. PURPOSE AN D SCOPE


The preceding chapters have reviewed the most significant work

dealing with water coagulation and have summarized contemporary theories concerning the basic mechanisms involved. Because of the complexity of relationships among the numberous variables which affect coagulation, many researchers have sought to discover empirical relationships among the measurable parameters in order that they might more fully understand basic coagulation mechanisms.
4 26 3
Black and Hannah,4 Pilipovich and others, Black and Willems, and Black and Christman38 have pursued such a course in the electrophoretic studies of the coagulation of several colloidal nate:ials with various coagulants and coagulant aids Packham39 and Mackr"e9 have reported some feric sulfate coagulation research, but neither of them performed electrophoretic studies of coagulation in which the effects of coagulant dosage and pH have been separately determined.

The primary purpose of the present research is to add to the body of knowledge resulting from the work of Black, Hannah, Willems, Christmuan, and Pi li po-vch: it will seek to learn the relationships among electrophoretic mobility, residual turbidity, coagulant dosage, and residual iron content of three clay suspensions resulting from their ferric sulfate coagulation, The three clays were chosen so that suspensions containing equal concentrations of them would exhibit low, medium, and high base exchange capacities respectively. The literature contains no references to work of this nature.


- 24 -






- 25

Because certain materials and procedures were utilized, the

study, in addition to its primary purpose, will allow direct comparison of the results of ferric sulfate coagulation with the results of alum coagulation of these clays reported by Black and Hannah.4 Moreover, the results of these tests may yield experimental confirmation of coagulation theories previously presented by others.

The scope of the work was limited to study of electrophoretic mobility, residual turbidity, and residual iron content in the ferric sulfate coagulation of suspensions of a low, a medium, and a high base exchange capacity clay over the pH range 3 to 10. The ferric sulfate dosages were 0, 3, 5, and 50 mg./l respectively. The procedures involved were jar tests, residual turbidity and iron determinations, and electrophoretic mobility determinations.














V. EXPERIMENTAL MATERIALS AND PROCEDURES


The experimental portion of the research consisted of (1) jar tests for the ferric sulfate coagulation of three clay suspensions and

(2) the determination of mobilities for the flocculating colloids in' those suspensions. Coagulant dosages of 0.0, 3.0, 5.0, and 50 mg/l of ferric sulfate were used, and pH was varied over the range 3 to 10. Final dissolved iron content, final pH, and initial and final turbidity of the jar test suspensions were measured, and pH values of mobility

samples were determined.



Materials


Clays

The clays used to prepare turbid waters were Kaolinite 4 and

Montmorillonite 23, obtained from Ward's Natural Science Establishment, and fuller's earth, obtained from the Floridin Company. Ferr7ic Sulfate

The ferric sulfate used as coagulant was the commercial grade

manufactured by the Tennessee Corporation, Atlanta, Georgia. Ferri-Floc is the registered trademark by which the manufacturer designates this material. The analysis of the particular sample of ferric sulfate used appears in Table 1.


- 26 -







- 27 -


TABLE 1

ANALYSIS OF FERRIC SULFATE USED FOR COAGULATION



Constituent Per Cent by Weight Total Water Soluble Iron 21.50 Water Soluble Ferrous Iron (Fe++) 0.70 Water Soluble Ferric Iron (Fe+++) 20.80 Water Insoluble Matter 2.00 Free Acid (as HSO,) 2.55 Moisture 2.51







- 28 -


Procedures


Preparation of Clay Suspensions

Hannah40 has given the following description of the clays used and of the first steps in their preparation:

The kaolinite and montmorillonite consisted of large lumps of dry clay. These materials were crushed and ground with mortar
and pestle and were then ball-milled for 24 hours. The fuller's
earth, which was obtained as a dry powder, was ball-milled for
24 hours.

The preparatory efforts of others, described above, were performed early in 1959. At that time Hannah40 determined the base exchange capacity of each of the three clays in the manner prescribed in Official Methods of Analysis of the Association of Official Agricultural
41
Chemists. Since their preliminary preparation the dried clays had been stored in closed glass containers. With the clays in the condition described, the present experiment was began with the weighing out of four 3.75-g aliquants of each clay. Each aliquant, in turn, was mixed with 500 ml of demineralized water and dispersed by mixing in a Waring Blendor for five minutes. The four aliquants of each clay suspension were placed in a separate covered beaker and allowed to hydrate for 24 hours. At the end of the hydration period, each suspension was slowly passed through a sodium-cycle, ion-exchange column to replace the natural, exchangeable cations of the clay with sodium ions. A glass column of 25 mm diameter containing a 12-inch depth of Nalcite HCR cation-exchange resin was used for the ion exchange process, which was




The Waring Blender is manufactured by Waring Products Company, New York, New York.







- 29
42
developed by Lewis. Just before the treatment of the three clay suspensions the resin in each was regenerated with one liter of ten per cent NaCl solution.

Stock suspensions were prepared by diluting each of the three

suspensions to twenty liters with demineralized water. The three suspensions were stored in polyethylene carboys. Preparation of Ferric Sulfate

A 2-kg portion of commercial-grade ferric sulfate was taken from a 100-lb bag of the material which was supplied by the manufacturer. After being mixed thoroughly, the sample was quartered, and 500 g were reserved for analysis. Approximately 100 g of the remaining material was finely ground with mortar and pestle. Aliquants of 1.000 g were weighed out and stored in glass vials having tight-fitting polyethylene stoppers.

Because the hydrolysis reactions of ferric sulfate are dependent upon time and concentration, a fresh coagulant solution of the chosen concentration was prepared daily. One of the previously weighed aliquants was quantitatively transferred to a 200-ml volumetric flask, which was then filled to the mark with demineralized water. This particular concentration was chosen, because less concentrated solutions became cloudy and were found to contain considerable volumes of hydrolyzed material within four hours after the initial mixing. A minimum of twenty minutes was allowed for the mixing, which was accomplished with a magnetic stirrer.






- 30 -


Coagulation Tests


Sample preparation. After the preparation of the aforementioned clay suspensions, approximately six months passed before coagulation tests were begun. A high-speed mixer, equipped with a stainless steel shaft and propeller, was used to thoroughly mix the suspensions. The minimum time of mixing for any suspension was one hour. Mixing was continued while 200-ml aliquants were pipetted into 8-oz polyethylene bottles. Also, triplicate 25-ml samples of each suspension were pipetted into beakers, and residue upon evaporation was determined for each in accord with the procedure specified in Standard Methods for the Examination of Water and Wastewater,43 which hereafter will be referred to as Standard Methods. To provide an ionic concentration in the suspensions comparable to those of surface waters, 20 ml of 5.00-g/1 sodium bicarbonate solution was added to each of the 200-ml suspension aliquants. In the final clay suspension samples, diluted as described below, the sodium bicarbonate concentration was 50 mg/1.

In the morning of a typical day of coagulation tests, immediately

after preparation of the coagulant solution, six of the 8-oz bottles containing the desired clay were selected, and their contents were quantitatively transferred to separate two-liter volumetric flasks in which 500 ml of demineralized water had been placed. Ten ml of chlorine solution containing 0.400 g/l of Clp were pipetted into each flask.

(During chlorine demand tests, that specific quantity had proved sufficient to satisfy the demand of the suspensions and still to leave about 1 mg/i of free available chlorine to oxidize the ferrous iron present to







- 31

the ferric state. Such a prechlorination dosage is common for treatment of surface water.) Sufficient amounts of 0.1 N NaH or HC1 were added to the flasks to yield the desired pH of the final suspension, and the flasks were filled to the mark with demineralized water. A tefloncovered magnetic bar was placed in each flask, and the flask's contents were nixed for five minutes with a magnetic stirrer. One-liter graduated cylinders were used to transfer half of each flask's contents to separate 48-oz square jars which were placed on the multiple laboratory

stirrer (hereafter referred to as the jar test machine). The remaining contents of the flasks were retained for mobility determinations.

Flocculation. The flocculation process was begun by rapidly stirring the suspensions while adding the correct coagulant dosage to each of the six suspensions. After receiving the coagulant dosages, the suspensions were mixed rapidly for 2 minutes and then slowly for 28 minutes. Next the stirring paddles were removed, and the suspensions were allowed to settle for 10 minutes. The rate chosen for rapid mixing was 100 rpm. The speed initially selected for slow mixing was 40 rpm, but soon after laboratory tests were begun, 5 rpm was adopted for slow mixing since it appeared that higher rates might cause disintegration of the floccules.

Promptly at the end of the sedimentation period 250-ml samples were drawn from each jar. An apparatus similar to that described by Cohen was used to siphon each sample from approximately an inch below the surface of the supernant. The settled samples so obtained were used for residual turbidity and residual iron determinations.




The six-jar, variable-speed stirrer used is a product of Phipps and Bird, Inc., Richmond, Virginia.






- 32 -


Initial and residual turlbidity measurements. Turbidity of the samples -as measured in a Lumetron Model 450 Filter Photometer. The 45
procedure recommended by the maufacturer required the preparation of a calibration curve for each of the three clays used. Calibration for a single clay involved the preparation of eight to ten suspensions so that their turbidities uniformly covered the desired turbidity range. The optical density of each suspension -as measured in the Lumetron for 650-m, light over a 75-n-m oath. ALlso, the turbidity of each of the suspensions was detennined uzcLth the Jackson Candle Turbidimeter in the manner set forth in St-P-nrd I7ethods46

Because coagulation efficiency was to be judged upon the basis of turbidity removal, determination of initial turbidities -as necessary. An average value of initial turbidity for each of the three clay suspensions ras obtained by measaring the initial turbidity of the suspensions for the jar tests, in dhich coagulant dosage was zero.

sidul iron de-terminaticn. The supernatant samples Vrach were obtained in the ,anner described above were filtered prior to the deterrrination of their iron content0 A fine, smooth, quantitative filter paperW' was used. Filtration ias necessary, because some of the iron present was chemically or physically bound to floc particles of such small size that they did not settle out in the short settling period allowed. in a -ater treatment plant the iron bound to such small




T.e DL'rmetron Photometer is a product of Photovolt Corp., New York, New York.
VThe paper used was paper No. 13061 manufactured by Will Corp., New York 52, New York.







- 33

floccules would not be found in the finished water, because sedimentation of longer duration or passage through the rapid sand filter would remove them. Filtration of the iron samples thus makes the laboratory jar tests more nearly analogous to plant conditions. The iron content of several samples was determined without prior filtration. In a later chapter results of these determinations are identified and are compared with results for similar filtered samples. The variance of iron content between the two kinds of samples emphasizes the necessity for filtration.

The phenanthroline method appearing in Standard Methods47 was

followed for the iron determinations. A 530-mro filter and a 75-m light path were chosen for the Lumetron Model 450 Filter Photometer used for the colorimetric iron determinations.

Measurement of pH. The final pH of the coagulated clay suspensions was measured with a Beckman Model G pH meter.



Electrophoretic Mobility Determinations


Dosi and mixing. A period of four to eight hours usually elapsed after the final dilution of the clay suspensions before

mobility determinations were made. In order to attain thoroughly mixed suspensions after the quiescent period, each suspension was stirred for five minutes on a magnetic stirrer just prior to the mobility determinations. The individual suspensions were dosed with the appropriate amounts of coagulant at the beginning of the mixing.




The Model G is a product of Beckman Instruments, Inc., Fullerton,
Ca.Lifornia.







- 34
Conductance measurement. Specific conductance was determined by the procedure described in Standard M.ethods.48 A Model RCl6Bl Conductivity Bridge with a pipette-form conductivity cell having a cell constant of 1 cm1 was used for the measurements.

Meareet of oH. The pH of each suspension was measured immediately before the mobility of its particles was determined. A Beckman Model G pH Meter was used for these measurements.

Particle ncbility measurement. The equipment and procedures used for the microelectrophoretic mobility determinations are those described in the paper by Black and Smith.49 The detailed description of the physical dimensions and construction methods for the glass cell appear in an article by Briggs,50 who designed it and for whom it is named. Since both procedure and equipment are described in detail in the references cited, only a general discussion of them -will be included here. In addition to the general account, however, any deviation from the suggested methods of Black and Smith will be delineated specifically.

The Briggs cell is constructed of Pyrex4 glass. Its rectangular cross-sectional shape and area are practically constant over the central portion of its longitudinal axis. A filling funnel is located at one end, and removable electrodes and outlet stopcocks are connected to either end of the cell. it is mounted in a metal stand which holds it in proper position on the microscope stage and supports it when it is not in use. Fig. 4 is Tanbo's sketch51 of a Briggs cell.




The Conductivity Bridge is a product of Industrial Instruments, Inc., Cedar Grove, New Jersey.

Pyrex is a product of the Corning Glass Company, Midland, MIchigan.






- 35 -






- 36

Two Briggs cells were assembled, mounted, and calibrated before
*
jar tests and mobility determinations were begun. Plastic Wood was found to be a more satisfactory packing material than any previously tried. If removal from the holder is ever necessary, the cell can be freed easily by soaking the assembly in acetone or a similar solvent.

During the course of this study several improvements in technique were developed. (1) The cell holder was attached to the microscope

stage with a pair of small clamps. This procedure prevented accidental misalignment of the cell during determinations. (2) Another deviation from the suggested procedure was the use of a 4-mm glass bead in each of the electrodes to prevent rapid mixing of the mercuric nitrate and potassium nitrate solutions in the electrode. This procedure was proposed by Briggs.50 (3) It was found that the mercuric nitrate solution could be introduced into the electrode most easily with a hypodermic syringe fitted with a 2-in needle. (4) When sufficient mercuric nitrate solution was injected to bring the liquid level about 3 mm above the narrowest portion of the opening between the upper and lower chambers of the electrode, no trouble was experienced in placing the glass bead without entrapment of air bubbles beneath it.

Serendipitously it was discovered that for the suspensions

studied, electrodes prepared in the manner described above could be used, over periods as long as a week. Moreover, for these particular suspensions it was possible to rinse the cell by passing demineralized water through it while it remained clamped in place on the microscope stage. The application of a low-intensity vacuum to the outlet tube caused




Plastic Wood is a product of Boyle-Midway, Chamblee, Georgia.







- 37
water velocity of sufficient magnitude to remove the settled floccules; therefore, it was not necessary to remove the cell from the microscope stage more often than once a week. Once the system was aligned and the cell clamped securely in place, a check was necessary only at the beginning and ending of each day's determinations to be sure that the microscope was focused on the stationary layer inside the cell. The water aspirator which was used to pull cleaning water through the cell was also useful as an aid in filling the cell with a suspension without the entrapment of bubbles.

The particle velocities were determined by visual observation and were timed with a stopwatch. A previously calibrated occular micrometer located in the plane of the microscope image of the stationary layer served as the measured course over which the flight of the particles were timed. The polarity of the electrodes was alternated between individual timings. Thus, of the twenty particles timed for each suspension, ten were observed to move toward the left, and ten moved toward the right. An exception to this characteristic movement was noted at the isoelectric point. At the isoelectric point various particles were observed simultaneously to move very slowly toward either electrode. The mobility of the particles of a suspension was computed from the mean of the time observed for each particle to travel a fixed distance; therefore, the computed mobilities are time-averaged rather than velocity-averaged mobilities.

















Soiu physical charactrisics of the clays which u used and the procedures fo: pr-caratiom of the suspensions stdied have been described in preceding sections. Table 2 presents the turbidities, gr:av-etri c concentrations and base exchange capacities of these






T2= 2
?P::P::cI2S OF TzE x C ' SiSPSON~S



a I ..ch.aZ : Base Excha e Rssidae Upon Capac-ity of Capaci ty of a! Ea , a oaon Clays Susoensions C7 lay, �aricity ms/i ce/i beq/I


_c te 5862.0 1107


Jle'�c sarth 79 6. 26 i7.

2 "4 74.0 87 6.4 kicro0"ats wi_ll be abbrevi ated Ueo .


aC~ base -xha~ ~a acit 0S a a Lim p o a.,at r ta t- C o co3a zu oatn dcsaces, Also, thle effects of .. . . . ..obility and coalition in . o .stdy theill be Study, ...co......d with s _ia"-"r effects

:or al: ant dosge i-n the alum coagulation stud:y reoported by Hannah.

to..... ,-~ &ie ahot te coalant dosages used Ln bothi






- 39
investigations be expressed in terms comparable to the units of base exchange capacity employed herein. These expressions appear in Table 3TABE 3

COAGUTLkNT DOSAG.S



Ferric Sulite Alum
1g/ p~eq/1 mgfL P.ea


3.0 34 5.0 45 5.0 57 15.0 135

50.0 570 100.0 900




Although n "nal do ...s of D.0 ng/! were included in both

studies, the effective dosage of aluminum ion is considerably smaller than that of the ferric ion because water of hydration constitutes a larger portion of the al dosage. Prably the most meaningful comparisons beaten t two ivastig.ations can be made for dosa es of (1) 5.0 .g/l of al an d 3.0 mg/1 of ferric sulfate, (2) 5.0 .mg/i of al, and 50 mg/l of ferric sulfate, and (3) 10' mgii of alum and 50 mg/l of ferric sulfate. These combinations of dosages are not eQUal, but the two lower ferric sulfate concentrations bracket the lower alur dosage, and the highest dosages of the tio studies represent concentrations greatly in excess of those necessa-_-y for good coagulation.

The choice of a satisfactory criterion of good coagulation is pearticularly difficult. Adequate coagulation must result In an extrenely high degree of turbidity removal. The optical property, turbidity must be relied upon as an index of removal efficiency. Use of this







- 40

property is complicated by the effects of shape, size, nuxmber and refractive index of the particles responsible for the turbidity in a given sample.

An exarole of the manifestation of these effects has been presented in Table 2. The suspensions contain equal gravimetric concentrations of the three clays, but the turbidities they exhibit are widely divergent. For water treatment plant operation a practical criterion commonly used requires that the turbidity of coagulated and settled water be five units or less. Such a cr-terion is probably too stringent for use in this study because the short period of sedimentation was insufficient for the settlingz- of the sin.ler tart- c1 which contribute most to the turbidity of a suspension. moreover the selection of a specific residual turbidity as the cr-iterion would be to recuire a much larger removal efficiency for the kaolinite suspension ( initial turbidity: 294 units) than for the contmorillonite (initial turbidity: 58 units).

Packha2 has arbitrarily chosen as his coag!ation critcrion

that dosage which reduces initial tu-oidity by 50 pr cent, This choice is undesirable, because it differs so widely from acceptable conditions encountered in water treatment plant operation. We have abitr rily chosen 90 per cent removal of initial tu'iCbidity as the criterion of satisfactory coagulation and results of the two studies will be compared on this basis.

The residual turbidities a-d aoI a ilitis for each -cmut1a'aton of clay and coagulant dosage have been graphed as functions of pH. Both graphs for a single permutation appear in the same figure. (The residual turbidities are expressed as percentages of the initial turbidity of each suspension.) Figares 5, 6, 7, and 8 are based upon







- 41

the turbidity and mobility results for the montmorillonite suspensions which were treated vith 0, 3, 5, and 50 mg/1 of ferric sulfate respectively.

In Fig. 5, pH is observed to exert minor influence upon montmorillonite mobility in the absence of any coagulant. Final turbidity is little affected by pH. Some turbidity removal was accomplished by the agitation and sedimentation of the jar test procedure, but the clarification so achieved was not significant.

Hannahs 40 electrophoretic studies revealed that the mobility of the suspended clay alone was relatively independent of pH, but the magnitude of the mobilities he reported were approzdmately 30 per cent smaller

than those recorded during the present investigation. The latter agree much more closely with the , negative clay mioblitiLe3 reported by Mattson20 and by others than do Hannahss Such agreement c an not be interpreted to be a proof of accuracy, because Mattson, for e:ample, worked with a different clay and did not report pH values. However, as consideration of figures below will indicate, the graph of the nobility

versus pH for zero coagulant dosage does fairly well define the mamLn negative mobilities observable for the suspensions and coagulant inder consideration. Ln view of this relationship, comparison of mobilities for zero coagulant dosage with Eattson~s maminum negative clay mobilities appears to be wrth7vhile,

The lack of agreement with Hannah's values led to a reconsideration of the equipment and procedure for mobility measurement. Of the mechanical devices which are used, the ammeter is the unit most probably capable of introducing an error of such size and consistency. The precision of the meters is about one or two per cent of the full scale





- 42 -


3 4 5 6 7 8 9 1011


34


10 1


2 3 4 5 6 7 8 9 10 11
pH


rI I


FIG. 5. -THE EFFECT OF MOBILITY UPON COAGUL C LAY: MONTMORILLONITE


DOSAGE:0


pH AND ATION. FERRIC MG/ L


2


U

0

w C,)


+1
0


+1
0
-1

-2
-3
1
120 100
80 60 40 20
0


I
0





-J
D

w
x


2
120 100
80 60


40 20
0


I I I I I I I U



* a
SLOW MIXING:
" 40 RPM "
5 RPM -o--o* a


0.0


SULFATE








reading, and the accuracy is not likely to be as good as the precision. When conditions such as the electrolyte conductivity or the available

power supply voltage make it necessary to measure a current corresponding to only ten or fifteen per cent of the full scale reading, the error involved in the determination may be six' to ten times as large as that for a reading at full scale. Perhaps the use of some more sensitive method of field intensity measurement circuitry would allow closer agreement among the results of inde-endent investigators.

Fig. 6 shows the effect of pH upon mobility and turbidity removal of the montmorillonite clay suspension for the 3-mg/! coagulant dosage. Increasing ph was accompanied by a gradual increase in the magnitude of the negative mobility. The sane trend wna shown for Hannah. s 5-mg!/l dosage, and, as was previously mentioned the nagi.. of the mobiities were only about 75 per cent of those from the present study. WThereas the only good coagulation rulting fro- th,,t alum. dose was perikinetic coagulation at pH less than 4.5, 3-mg/I of ferric sulfate yielded good coagulation from ph 5.5 to -H 6.6.

With an increase of ferric sulfate to 5 mrg/l as shown in Fig. 7, the increase in mobility with increasing pH becomes greater, although the curve is not so smooth. The most noticeable effect of the increased coagulant dose is the broadening of the zone of good coagulation over the range pH 4.6 to 7.8.

The 50-mg/l ferric sulfate dosage resulted in change reversal for values of ph less than 5.3. in Fig. 8, the algebraic decrease in mobility with increasing ph was observed again as it had been i the previous igr hannah's comoarable alum dose of 100 mg/ resulted in a mobility cuave of similar shape above pH 4.7, but the isoelectric





-44-


3 4 5 6 7 8 9


2
120r


100.
80 60F


40 20
0


3-4 5


2 3 4 5 6
pl-


9 10


11 120 100


7 8 9 10 11


80 60 40 20
0


FIG. 6. -THE EFFECT OF pH AND MOBILITY UPON COAGULATION. CLAY: MONTMORILLONITE FERRIC SULFATE DOSAGE: 3.0 M G / L


10


+2



0 U-


I I I 5 I



*


m
I I I I - ~ ~I


11


+1
0
-1

-2
-3


I
0
5-.
0

H


D
H
-J
D

w


I* I I 0


I a I I I I "I


L





- 45 -


4 5 6 7 8 9


1011


+1
0

-1
-2
-3


2 3


4 5 6 7 8 9


10


11
-1120


40 RPM-.-- - 100 5 RPM ----o- -80

-60

-40

00 0 20 2 0
2 34 5 67 81 011
pH


FIG. 7 -THE EFFECT


OF pH AND


MOBILITY UPON COAGULATION. CLAY: MONTMORILLONITE FERRIC


2


0


U


0
U
LU (-I


_ p p ' A a A


0
10
I




D
F
LU


SULFATE


DOSAGE:


5.0


MG / L


+;,,,





- 46 -


2 3 4 5 6 7 8 9 1011 +2i v , , , i , +2


+1

0
-1
-2
-3


2
1201


100 80
60 40 201


3 .4 5 6 7 8 9


pH


10 11 , - 120


8 9 101


FIG. 8. -THE EFFECT OF MOBILITY UPON COAGUL CLAY: MONTMORILLONITE SULFATE DOSAGE: 50.


100 80 60
40 20
0


pH AND ATION. FERRIC MG/ L


2


0

U
Ldi (.1


I
0




D
-J

ii
LU


I I I I I I I


. SLOW MIXING: .
40 RPM
- 5 RPM -o--o-




0 1


I I IIII I I IIII M III I II I






- 47 -


point (pH 6.6) was considerably higher, and below pH 4.6 as the solubility of aluminum increased the charge reversal effect of the alum also decreased. The alum dose resulted in good perikinetic coagulation below pH 4.5 and both alum and ferric sulfate produced good orthokinetic coagulation from pH 5,5 to 8.8 and from 5.5 to 10.0 respectively.

Fig. 9 is a composite of the four preceding figures. it allows ready comparison of the broadening of the zone of good coagulation with increasing coagulant dosage. Also the zone of perikinetic coagulation

below pH 3.7 and the zone of poor or no coaalation from roughly pH 4 to pH 5 are easily identified. Probably the mcst important relationship graphically illustrated in the figure is that the isoelectric point

occuring at pH 5.3 for the one dosage causing charge re orsal marks the beginning of the pH zone of most economical coagulation of the particular clay and for the arbitrary good coagalation criterion chosen.

The next series of figures pertain to the coagulation of fuller's earth. Fig. 10 illustrates the effect of pH in the absence of coagulant. As was the case with montmorillonite, in the absence of coagulant, negative mobility increased with increasing pH value. Perikinetic coagulation occured below pH 4.2 although no such coagulation had been observed for montmorllonite.

Beginning with Fig. 11, the mobility curves become more complex as the varying effectiveness of the coagulants for charge reversal at

different values of pH becomes more evident. For the 3-rag/l dosage, the original particle charge is reversed over a pH range of slightly more than one unit. For this particular dose, comparison of mobility curve shapes for ferric and alum coagulation are virtually impossible because of lack of similarity. Good ferric coagulation was observed over the








J INO-1-!1RO NiNON " ,
NOV-Ir9VOD NOdf fn Z?, ONV CGNV Hd Jo . 3H2-'6 C_t


Hd


0L OL 6 2 9 t7


I -~ - Tm


6o--'- . -


. L19 01'


11 Ile 0 Ir


6 9 L, 9 g


-0










0F
H
C)
09




tO





i',K


, f , .t ' . _ ,, _ . . _ '0,





- 49 -


2


3 4 5 6 7 8 9 1011


2 3 -4 5 6 7 8 9 10 11


2 3 4 5 6 7 8 9 10 11
pH


+1

0
-1
-2
-3


120
100 80 60 40 20
0


FIG. 10.


-THE EFFECT OF pH AND


MOBILITY UPON COAGULATION. CLAY: FULLER'S EARTH FERRIC SULFATE DOSAGE: 0.0 MG / L


.!.2
+1

0


0

U O/


I 5 I I I I I I a


0



._.


W F-


12C boc
80 60 40 20
0


- ' I i' I " ' I I
0 00
0 0
0



40 RPM "
5 RPM -o-0II i II I I I I I 1 ..








3 4 5 6 7 8 9 1011
I I I I I i 1 . 2


8 9


6
pH


2


0"
0 U>-2
3~J r2


10 11 , "120
100 " "80


7 8 9 10


11


60 40 20
0


FIG. 11. -THE EFFECT OF p H AND MOBILITY UPON COAGULATION. CLAY: FULLER'S EARTH FERRIC SULFATE DOSAGE: 3.0 MG / L


- 50 -


+1
0
-1
-2
-3


4 5


I
0
0

H


H
-J
D

IL]


2
1201


100 80
60r


40 20
0


I 1 I I I I


1 -1 , p I_ A I, a i I I


2 3


4 5







- 51 -


pH range 5.6 to 7.3.

When the ferric sulfate dose was increased to 5 mg/1 (Fig. 22)

the zone of charge reversal was increased to 3 pH units, and good coagulation occurred above pH 5.5. A most interesting effect noted in this particular portion of the study is the greater efficiency of 5 rpm slow mixing for perikinetic coagulation and the better orthokinetic coagulation that accompanied the 40 rpm slow mixing.

The mobility curve for the 50-mg/l ferric sulfate dosage (Fig. 13) appears to be typical of excess ferric sulfate dosage for the three clays studied. Particles were positive below pH 6.3, were isoelectric at that point, and became increasingly negative as pH was raised to 10. The slope of the curve is steepest ina aediately above and below the isoHannah's comparable alum curve changes less abraptly at the isoelectric point, and the apparent effect of increasing solubility at low pH is evident. Ferric coagulation is good above pH 4.2.

The composite graph for fuller's earth (Fig. 14) clarifies the relationship between the isoelectric point of the overdosed suspensions and the beginning of the zone of efficient coagulation. As was the case for montmorillonite, it appears that the "overdosed isoelectric point"

does mark that beginning. Also, the broadening of the zone of good coagulation with increasing dosage is evidenced.

Individual mobility and turbidity graphs for the kaolinite clay

are shown in Figures 15, 16, 17, and 18, and Fig. 19 is the corresponding composite. The mobility curves are very similar to those for fuller's earth, and the comments above concerning the latter are qualitatively applicable. Regarding coagulation, the kaolinite is much more nearly





- 52 -


0

U
Li

oI


FIG. 12. - THE


EFFECT


MOBILITY UPON COAGU CLAY: FULLER'S EARTH SULFATE DOSAGE: 5.0


OF pH AND


LATION.
FERRIC MG/ L


2345678,9,10�11
2 +2 +1
0
-1

2 -2
3
2 3 4 5 6 7 8 9 1011
Sr, , , , 1 120 D- SLOW MIXING: - 100
40 RP M - 80
5 PPM -o--o0 0[ [-60

40
0! 0 0
0 0

2 3 4 5 6 7 8 9 10 11
pH


12 10


I
0
0
H
C

F
-j
D
C
Lii





- 53 -


2 3


3 4 5 6 7 8 9 1011


4 5


+1

0 .-1

-2
-3


6 7 8 9 10 11


2 3 4 5 6 7 8 9 10 11
pH


FIG. 13.


-THE EFFECT


OF pH AND


MOBILITY UPON CLAY: FULLER'S


SULFATE


COAGULATION. EARTH FERRIC
-: 50. MG/ L


2


U+1

0


I I 5 5 5 I I I 5 5 I


0

-.
Q

D
-.
_J C<
w3


DOSAGE





- 54 -


2

(.9
-+F

0
>o .


C)2r
>-3


2
120-


0

F



D

_J


Li CtO


3 4 5 6 7 8 9 1011

MG/L FE2(SO4)3 O" '5---500


"-3
3 4 5 6 7 8 9 0 11
120


80 60
40'
20


100


2 34 5 6 7 89 9 0 11
pH


FIG. 14. - THE EFFECT OF p H MOBILITY UPON COAG!_iLATl CLAY: FULLER'S EARTH


10


AND ON.





- 55 -


2 3 4 5 6 7 8 9 1011
+2 +2
+1 -+1
o 0


:2 2 45 7 91 1
0
-2- -2


00--% 2 3 -4 5 6 7 8 9 10 11 1-.- 12 0 ..... , -, ...., - ,, , ,12 0

>. 100 100
80- 80 03
60- 60 D6 SLOW MIXING: F 40 R PM -o--o40_ 540 40 < j 5 RPM---D 20 - 20
L , p 0
2 3 4 5 6 7 8 9 1011
pH
FIG. 15. -THE EFFECT OF p H AND MOBILITY UPON COAGULATION.
CLAY: KAOLINITE FERRIC SULFATE DOSAGE: 0.0 MG / L





- 56 -


3 4 5 6 7 8 9 1011
Ii+2


3 4 5 6 7 8 9


8 9


2




:0
0
W-2-


10 11
~~~ n~r


10


100 80 60
40 20
0


11


pH
FIG. 16. - THE EFFECT OF p H AND MOBILITY UPON COAGULATION. CLAY: KAOLINITE FERRIC SULFATE DOSAGE: 3.0 MG/ L


+1
0
-1
-2
-3


2
12010080 60
40
20-


0




D LU


1 1 I I I I I I I I~~__ J~~___ L ~ ~ I ~I


2 3 4 5


I -A I I





- 57 -


3 4 5 6 7 8 9 1011


2 +2



0
0 >-3"


2 3 4 5 6 7 8 9 10 11
120 -- I I 120


SLOW


MIXING:


40 RPM -0--5 RPM -o--.o- o
0 0
0 0
0
0
p A A I


3 4 5 6 7 8 9 10 11
pH


100 80
60 40
20
0


FIG. 17


-THE EFFECT OF p H AND


MOBILITY UPON CLAY: KAOLINITE SULFATE DOSAGE


COAGULATION.
FERRIC
5.0 MGI L


+ I

0
-1
-2
-3


I I p I


4



ft I I I I I a


1001


FH
0
0
F



-J


80 60
40 20


Ii





- 58 -


34 5 67 81 011


+1
0
-1
-2
-3


3 4 5 6 7 8 9


10 11


3 4 5 6 7 8 9 1011
pH


100
80 60 40 20
0


FIG. 18. -THE EFFECT MOBILITY UPON COA CLAY: KAOLINITE SULFATE DOSAGE: ,


OF pH AND REGULATION.
FERRIC 50. MG/ L


2 +2r


0

U


0
0

I

D
H

D Lii


2
120


10OF


80 60 40


20


2


I








I


!










K) ~ (I


I)


N 4









1,



K) L



\rL C') L


3 4



11))
lii
33 ~/
Iy

I I I I I I II I


I

NI

~! ,i' 33

'~~\J C1) I I I


'I


/ J-UOAQ/ DBSlUl )


/
/


(~\f &)


(NJ
3(9
(N'
V..-


,(" I,


()V

.3/





i(),1


3 .3.


C)


( )

C) ~ .1)
(\j


K
~K2>
1~


~' (~)


'1



.~ ) i.,)


I


K) (9iL i )5

w ( ). o



iii ( hi;


I

I
I
I
Ci.


(v\!-


C-L /n- ",i v n -s%.-






- 60

similar to the montmorillonite. Most important, however, is that the "overdosed isoelectric point" near pH 6.0 marks the beginning of the zone of good coagulation. Black and others52 have shown that the same relationship between the "overdosed isoelectric point" and the zone of good coagulation exist for the ferric sulfate coagulation of natural colored waters.

An additional observation of possible importance concerns base exchange capacities of the suspensions. There was no observation of charge reversal in the present investigation for dosages for which the ratio of coagulant dose to base exchange capacity (both expressed in ieq/1) was less than 1.0. Good coagulation occurred for all three clays, however, with a dosage of 3 mg/l. For this dose the ratio mentioned above for montmorillonite is 0.48, for fuller's earth is 2.0, and for kaolinite is 5.3. Work with ferric sulfate doses in a range which would yield ratios near unity should reveal more interesting information on charge reversal phenomena.
.1
The U. S. Public Health Service has set 0.3 mg/l of iron as the maximum allowable concentration of that constituent in drinOking water. In the present research, therefore, determination of the residual iron content in the supernatant was necessary in order that suitability of the various treatments could be evaluated. The results of the determination are graphically presented in Figures 20 through 28.

The most important information obtained from the iron determination was that for all the pH zone of good coagulation, the residual iron content is below the maximum allowable concentration. Another interesting observation which is generally applicable for the ferric sulfate coagulation of the three clays is that residual iron values are less











2 1.4-


1.0


0.4
0.2
0.01
2


A


| I .. . I " I .. . I 'I .........


1.2 1.0
0.8 0.6
0.4 Q2


6. 7


6 0


3456789 1011


pH


FIG.20.- THE EFFECT OF pH ON IRON RESIDUAL. CLAY: MONTMORILLONITE FERRIC SULFATE DOSAGE: 3.0 MG/L


-, 61 -


4 5


10


11
1.4


1.2F


0.8" 0.6-


0 - too





- 62 -


2 3 4 5 6 7 8 9 10 11
1.4 " 1 1 1 , I I I 1 , 1.4


z
0 0.8- -0.41
0.6- -0.
-J
< 0.4- -0, D* *
c 0.2- .0Lr 0. 0 1

2 3 4 5 6 -7 8 9 1011

pH


FIG. 21. -THE EFFECT OF pH ON IRON RESIDUAL. CLAY: MONTMORILLONITE FERRIC SULFATE DOSAGE: 5.0 MG/L





- 63 -


2 3 4 5 6 7 8 9 10


11


2 3 4 5 6 7 8 9 1011


pH


FIG. 22.-THE EFFECT OF pH ON IRON RESIDUAL. CLAY: MONTMORILLONITE FERRIC SULFATE DOSAGE: 50. MG/L











10 11


3 4 5 6 7 8 9


11.4 1.2
1.0 0.8 0.6 0.4
0.2


10 11


pH


FIG. 23.-THE EFFECT OF pH ON IRON RESIDUAL. CLAY: FULLER'S EARTH FERRIC SULFATE DOSAGE: 3.0 MG/L


- 64 -


)


4 5


I 4 1


.
2


0


D
[JJ:


12 1.4
1.2 1.0 0.8


0.4 0.2 0.01
2





- 65 -


2 3 4 5 6 7 8 9 10 1
1.4 .. . L , , , " " .


2 3 4 5 6 7 8 9 10 11


pH


EFFECT


FIG. 24. - THERESIDUAL. C
FERRIC SULF


SLAY: AT E


OF pH


FULLER'S
DOSAGE:


ON IRON
EARTH
5.0 MG/L


(9


0 LU





- 66 -


2 3 4 5 6 7 8 9 1011
1.4 1-- - -1 -- 1--- I' I i --1,,.... 1. 4


2 3 4 5 6 7 8 9 1011


pH


FIG. 25.- THE EFFECT


OF


RESIDUAL.


CLAY:


FULLER'S EARTH


FERRIC SULFATE


DOSAGE:


pH ON IRON


50. MG/L





- 67 -


2 34 5 6 7 8 9 10 11


2 3 4 5 6 7 8 9 1011


pH


FIG. 26.- T RESIDUAL.
E RR C SL


HE EF
CLAY: JLFAT E


ECT OF p KAOLINITE DOSAGE:


H ON IRON 3.0 MG/L


_J



z
0

D
IJ


=1





- 68 -


2 3 4 5 6 7 8 9 10 11
1.4 1 a " 1 . . -.1 .4


2 3 4 5 6 7 8 9 10 11


pH


F!G. 27 - THE RESIDUAL. (


EFFECT OF LAY: KAOLINI


5.0 MG/L


pH ON IRON TE


S ULFAT E


DOSAGE:





- 69 -


2 3'4 5 6 7 8 9 10 11 2345678910114


2 3 4 5 6 7 8 9 1011


pH


FIG. 28.- T RESIDUAL.


-E EF
CLAY:


ECT OF pH KAOLINITE


ON IRON


50. MG/L


.9
0

_J fZ
LU


F E R PIC


S ULFAT E


DOSAGE:






- 70

than the maximum allowable concentrations in the low pH range dow.m to pH 3.8. Furthermore, even at pH 3.0 no supernatant contained more than

1 mg/1 of iron in solution. These low residual iron values for clay coagulation are much lower than those reported for colored water coagulation.

Black and others52 have reported residual iron values for comparable coagulant dosages in colored water that are six times as large as the largest observed in the present study for the same pH of 3.0. At pH values up to 4 or 5, the ratio was even larger. The difference in behavior of the clay and color colloids in this respect point out the need for a more thorough understanding of basic coagulation mechanisms.

For the present research it was necessary to determine whether

the pH of the samples employed for mobility determinations were identical with the pH of the corresponding jar test suspensions. Figures 29, 30, 31, and 32 show the close agreement between the respective pH values. Fig. 33 is a composite of the four, which shows the effect of coagulant dosage upon the value of pH. It should be remembered that the suspensions which were used were synthetic preparations containing 50 mg/i of sodium bicarbonate. A more poorly buffered solution might not yield such agreement between the two. Even with the systems used it was difficult to dose the suspensions with the exact quantities of acid or base required to obtain evenly spaced mobility and turbidity curve points. Only the curves for fuller's earth are shown, since those for the other clays were similar.





- 71- -


2 3 4 5 6 7 8 9 101
1 A F I a
1" MOBILITY SAMPLES -01i TEST JARS -o-nZ Gel


O21-


1
1.2 1.0 0.8


-40.2


0~


GOS


0 L0.


SULFATE: MGIL
I A I


,3 4 5 6 7 8 9 10 11


pH


,G. 29.- QUANTITIES OF ACID OR RCQUWIF-ED FOR pH ADJUSTMENT FULLER'S EARTH.


BASE OF


0.0 0.2


S1.0


1.4 1.6


1 0.4 0.6


0.8


11.0


1.2 1.4 1.6


C





- 72 -


2 3 4 5 6 7 89 1011


2 3 4 5 6 7 8 9 11


pH


r 1. 30.-


QUANTITIES


OF ACID OR


BASE


F.U ED F ULL ER'S


FOR pH ADJUSTMENT OF EARTH.





- 73 -


2 3 4 5 6 7 8 9 1011


2 3 4 5 6 7 8 9 1011


FIG. 31.


- QUANTITIES


OF ACID OR BASE


QUz-, M--D FOR pH ADJUSTMENT OF FULLER'S EARTH.





- 74 -


2 3 4 5 6 7 8 9 10 11
121' 1.2
MOBILITY SAMPLES -c-TEST JARS
2c Q8, -0.8
1_08


~0.4
SC2 0.2 0.,1 0D _ 0.0

10.2
H ~0.4
_ C1 ~0.6 o - 10.8
- 1c 1
- i
FATE-1.2
50. MG/ L

2 3 4 5 6 7 8 19 O11
pH

iG. 32.- UANTITIES OF ACID OR BASE T.7 , D FOR pH ADJUSTMENT OF FULLER'S EARTH





- 75 -


3 4 5 6
F -E FlR S UL FAT-E
L 0.0 MG/L
Q3.0 MG/L06L 5.0 MG/L 1 50. MG/L


7 8 9 1011


2 3 4 5 6 7 8 9 1011


pH


FK'3. 33. - QUANTITIES OF ACID OR IrEQU ,77'ED FOR pH ADJUSTMENT FULLER'S EARTH.


BASE
OF














VII. CONCLUSIONS


In the present investigation, charge reversal at low pH was observed for all three clay suspensions when the ratio of coagulant dosage to base exchange capacity (both expressed in Peq/1) exceeded three. With increasing pH the mobilities of these suspensions decreased until the isoelectric point (in the pH range 5 - 7) occurred. This "overdosed isoelectric point" was found to mark the beginning of the zone of efficient orthokinetic coagulation of all three of the suspensions studied. Such behavior has been reported for the coagulation of colored water with ferric sulfate by Black and others.52 Furthermore, an analysis of the results of clay coagulation with alurn reported by
40
Hannah revealed that the "overdosed isoelectric point" was an indicator of effective coagulation conditions for those suspensions.

The base exchange capacities of the suspensions studied were not found to be directly or proportionally related to the coagulant dosages required to effect satisfactory coagulation, but they did significantly affect the coagulant dosages required to cause charge reversal at low

pH.

For all of the combinations of ferric sulfate dosages and clay

suspensions studied, it was observed that in the pH zones of good coagulation the residual iron values were less than the maximun allowed in
1
drinking water by the U. S. Public Health Service. Moreover, even at low pH the residual iron values were less (by a factor of six to ten) than those reported by Black and others for the coagulation of colored


- 70 -





- 77

water with comparable dosages of ferric sulfate. The difference between the behavior of the colored and turbid waters mentioned above may be caused by the ability of organic matter (including organic color) present in the water to form complexes with both ferrous and ferric ions.

The conclusion of greatest practical importance resulting from the present investigation is that the microelectrophoretic technique may be useful for the prediction of the pH zone of most effective coagulation of natura, turbid water. Further study is necessary to determine whether there is an "overdosed isoelectric point" for natural, turbid waters which marks the zone of good orthokinetic coagulation.


























APPFIDIX












EFFECT OF


TABLE 4

pH ON ELECTROPHORETIC MOBILITY AND SEDIMENTATION OF MONTMORILLONITE CLAY


pH of Final pH of Final Turbidity Jar. HCIl NaOH Mobility Mobility Coagulated Final* as Per Cent of No. meq/1 meq/1 Sample p/sec/v/cm Water Turbidity Initial Turbidity


3.72 6.51

6.79 7.22

7.44

9.20 5,85

6.46 4.40


-1.58

-1.69

-1.58

-1.60

-1.66
�i94

-1.98, -2.16




No coagulant


3.80 6.69

6.95 7.30 7.52

8.80 6.20 6.70


*Turbidity of initial suspension: 58 .


1 2

3 4 5 6

181A 182A 187


0.81 031

0.26 0.02


0.02 0.18


0.55
0.40 0.70


used3 for this series.












COAGULATION OF


TABLE 5

MOINTMORILLONITE CLAY WITH 3.0 MILLIGRAMS PER LITER OF FERRIC SULFATE


pH of Final pH of Final Turbidity Residual Jar HC1 NaOH Mobility Mobility Coagulated Final* as Per Cent of Iron No. maeq/1 meq/1 Ssrpie I/sec/v/cm Water Turbidity Initial Turbidity mg/l


0.00 0.00

0.07 0,20 0.30

0140

0.60

0.55

0.30
0.25


7.37 7.66 8,93

9.35 9.80 10.01


-1.55

-1.57

-1,67

-1.61

-1-93

-1.91


0.07 0020 0.30


7.36 7.66 8.96 9 , 4ro 9.75
10,06 6.22

6.99 7.18 8.20

9.44
9.87


0.14 0,21

0,29 0.,41

0.38 0,38

0.07



0008

0, 42 0.39

0.53










TABLE 5 - CONTINUED


pH of Final pH of Final Turbidity Residual Jar HC1 NaOH Mobility Mobility Coagulated Final* as Per Cent of Iron No. meq/1 neq/l Sample P/sec/v/cm Water Turbidity Initial Turbidity mg/l


109 110 ll

112 113 114 115

116 117 118 119

120 157


0.15


0.05 0.12

0.18 0.25

0.35


1.50
0.80 0.70 0.60

0.50 0.25 0.68


6.90

7.62 7-77 8.81

9.36 9.65

2.91 3.68 4.13 5.21 5.82 6.58

4.14


-1.29
-1.37

-1.36

-1.50

-1.76

-1.67

-1.37

-1.10
-1. 21

-1.39

-1.38

-1.51

-1.68


6.88 7.70 7.87 8.78 9.38 9077

2.92 3.73 4.20 5.31

5.94
6.60 4.31


0.26 0.33
0.45 0.54 0O. 60 0.73 0.14 0.04 0.25

0.24 0.16 0.35 0.23









TABLE 5 - CONTINUED


pH of Final pH of Final Turbidity Residual Jar HC1 NaOH Mobility Mobility Coagulated Final* as Per Cent of Iron
No. meq/1 meq/1 Sample p/sec/v/cm Water Turbidity Initial Turbidity pig/1

158 0.60 5.07 -1.54 5.25 7 12 0.01 159 0.35 6.61 -1.75 6.72 7 12 0.00 190 0.68 4.33 49 85 0.21 191 0.60 5,31 10 17 0.13


*Turbidity of initial suspension: 58.










TABLE 6

COAGULATION OF MONTMORILLONITE CLAY WITH 5.0 IILLIGRA14S PER LITER OF FERRIC SULFATE


pH of Final pH of Final Turbidity Residual Jar HC1 NaOH Mobility Mobility Coagulated Final* as Per Cent of Iron No. meq/l meq/1 Sample P/sec/v/cm Water Turbidity Initial Turbidity mg/i


1.70 0.60

0.53 0.00


25

26 27 28 29 30

55 56 57
58

59 60

153


0.00 0.14 O.60


0.68 0.59 0.25


0.20 0.30 0.40


0.90


2.92 4.87 5.62 7.15 7.22 10.00 3.89

4.19 6.46 8.90 9.49 9.79 3.52


-0.93
-1.27

-1.17

-1.78

-1.78

-2.20

-1.22

-1.35

-1.77

-1.75

-2.37

-1.80

-1.72


2.90 5.06

5.85 7.20

7.50 9.61 3.89 4.13 6.58 8.90 9.42 9.69 3.58


1.19 0.37

0.32 0.18

0.32 0.92 0.14 0.15 0.08

0.24 0.40 0.16 0.14










TABLE 6 - CONTINUED


pH of Final pH of Final Turbidity Residual Jar HC1 NaOH Mobility Mobility Coagulated Final* as Per Cent of Iron No. meq/1 meq/1 Sample 11/sec/v/cm Water Turbidity Initial Turbidity mg/1


154 155 156

176 177 178 192 193

194 195 208

209


0.70 0.55


0.00 0.62 0.6o 0.48 010

0.62 0.59


0.28 0.22 0.12 0.00


4.03 5.61

9.49


-1.80

-1.37

-2.26


4.11 5.68 9.47 9.17 8.46 7.65 4.69 5.10 6.06

7.12 4.87

5.30


*Turbidity of initial suspension: 58.


0.11 0.05

0.43 0.09 0.05

0.00 0.07 0.05 0.03 0.04 0.02 0,04












COAGULATION OF


TABLE 7

MONTMORILLONITE CLAY WITH 50 MILLIGRAMS PER LITER OF FERRIC SULFATE


pH of Final pH of Final Turbidity Residual Jar HC1 NaOH Mobility Mobility Coagulated Final* as Per Cent of Iron No. meq/1 meq/1 Sample ./seo/v/c, Water Turbidity Initial Turbidity mg/1


1.70 0.60 0.53 0.00


0.00 0.07 0.60


1.20 0.25


0.25
0.40 0.75 1.00


2.90 3.40 3.47 5.56 5.91 7.57 3-17

4.10 6.46 6.78 8.87 9.67


+1.03 +0.71 +0.58

-0.17

-0.34

-1.65
+0. 38 +1.07

-0.99

-0.85

-2.37

-2.43


2.90 3.40 3.45
5.90 6.20 7.63 3.15

4.03 6.44 6.82 8.88 9.72


9.8 3.1 3.7

0.88

0.56 018

1.02 0.57 0.06 0.07

0.18 0.24









TABLE 7 - CONTINUED


pH of Final pH of Final Turbidity Residual Jar HCl NaOH Mobility Mobility Coagulated Final* as Per Cent of Iron No. meq/1 meq/1 Sample P /sec/v/cm Water Turbidity Initial Turbidity nag/1

151 0.18 4.30 +0.67 4.41 18 31 0.03 152 0.08 4.95 +0.16 5.22 10 17 0.12 202 0.42 3.39 135 233 0.59 203 0.15 4.37 10 17 0.05 204 0.00 0.00 5.67 5 9 0.00 205 0.60 7.62 1 2 0.02 206 0.65 7.91 2 3 0.01


*Turbidity of initial suspension: 58.










TABLE 8

EFFECT OF pH ON ELECTROPHORETIC MOBILITY AND SEDIMENTATION OF FULLER'S EARTH


pH of Final pH of Final Turbidity Jar HCl NaOH Mobility Mobility Coagulated Final* as Per Cent of No. meq/1 meq/l Sample p/sec/v/cm Water Turbidity Initial Turbidity


1.62 0.62 0.51 0.05


7 8 9
10 11

12 183 184 188

189 214 215


0.05

0.35


0.80


0.12


0.72


o0o9


3.00 5.70 6.18 7.28 7.65

9070

4.00 9.29 4.49 9.18


-0.89

-1.60

-1.78

-1.72

-1.94

-2.16

-1.50,

-2.29

-1.53

-2.13


0.72 0.58


3.00

5.80 6.25 7.37 7.80 9.75
4.18 9.06


85
105 109 110 101 lo

109 33 80

67 81 62 81


4.66 5.90










TABLE 8 - CONTINUED


pH of Final pH of Final Turbidity Jar HCl NaOH Mobility Mobility Coagulated Final* as Per Cent of No. meq/1 meq/1 Sample p/sec/v/cm Water Turbidity Initial Turbidity


216 0.25 6.86 68 86 217 0.08 7.41 66 84


*Turbidity of initial suspension: 58. No coagulant used for this'series.












COAGULATION


TABLE 9

OF FULLER'S EARTH WITH 3.0 MILLIGRA 4S PER LITER OF


FERRIC SULFATE


pH of Final pH of Final Turbidity Residual Jar HCI NaOH Mobility Mobility Coagulated Final* as Per Cent of Iron No. meq/1 meq/1 Sample P/sec/v/cm Water Turbidity Initial Turbidity mg/i


121 122 123 124 125 126 127 128 129 130 131 132


1 . 25
0.90 0.75 0.70

0.68 0.65 0.60 0.50 0.40 0.20

0.05


0.05


3.07 3.60 4.07 4.45 4.70 5.00 5.58 6.05

6.38 6.82 7.34 7.87


-0.46

-0.26

+o.68 +0.55

+0.63

-0.17

-0.66

-0.98

-0.84

-1.54

-2.08

-2.62


3.12 3.54 4.07 4.43 4.67 4.99 5.56

6.03 6.38 6.92 7.36 7.90


0.10

0.07 0.07 0.08 0.05 0.05 0.00 0.06 0.02 0.02 0.08

0.14









TALE-9 - CONTINUED


pH of Final pH of Final Turbidity Residual Jar HC1 NaOH Mobility Mobility Coagulated Final* as Per Cent of Iron No. meq/l meq/l Sample P/sec/v/cm Water Turbidity Initial Turbidity mg/l 160 0.18 9.03 -2.38 9.21 59 75 0.05 161 0.10 7-32 -1,74 7.60 11 14 0.00 175 0.15 9.09 -2.10 8.80 56 71 0.01


*Turbidity of initial suspension: 79.










TABLE 10

COAGULATION OF FULLER'S EARTH WITH 5.0 MILLIGRAMS PER LITER OF FERRIC SULFATE


pH of Final pH of Final Turbidity Residual Jar HC1 NaOH Mobility Mobility Coagulated Final* as Per Cent of Iron No. meq/1 meq/1 Sample P/sec/v/cm Water Turbidity Initial Turbidity mg/l


1.52
o.60 0.46

0100


0.00 0.07 0.4o


1,00 0.75 0.25


0.15 0,20 0.30


3.10 5.4o

6.30 7.40 7.57 9.62

3.54 4.09 6.62 8.40 9.02 9.55


+ 0.36 +1.38

-0.43

-2.02

-2.41

-'3 00

+0.68 +1.45

-0.80

-2.36

-2.30

-2.41


3.15 5.12 6.00

7.25 7.52 9.73 3.54
4.24 6.99 8.38 8.82 9.41


1.60

0.86 0.34 0.18

0.16

0.26 0.16 0.12 0.08

o.o6 008 0.08










TABLE 10 - CONTINUED


pH of Final pH of Final Turbidity Residual Jar HCI NaO0H Mobility Mobility Coagulated Final* as Per Cent of Iron No. ieq/1 meq/1 Souple ./sec/v/cm Water Turbidity Initial Turbidity mg/l


0.62

1.52

0.82 0.62 0.55

0.25 0.4,8


162 196 197 198 199 201 200 207 210

211 218 219


0.08

0.15

0.20


0.00

0.72


6.42


-1.28


6.52 2.98 3.66 5.08 5.66

6.62 6.00


0,00 0.27 0.05

0,05 0.05 0,02 0.03 0.00 0.00

o.o6


7.77 8,82 9.16 7.43 4.22


*Turbidity of initial suspension: 79.




Full Text

PAGE 1

1 ELECTROPHORETIC STUDIES OF TURBIDITY REMOVAL BY COAGULATION J WITH FERRIC SULFATE By JAMES VERNON WALTERS J A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNrVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA August, 1963

PAGE 2

ACKWOl^IEDG^lENTS The author wishes to express his gratitude to his coinmittee chairman, Dr. A. P. Black, for giiidance and encouragement he has given during the investigation and for his generosity in the giving of himself to his students. He deeply appreciates the friendship and helpfulness of Prof. J. E, Kiker, Jr., Prof. G. B. Morgan, Jr., and Prof. T. deS. Furman, members of his supervisory committee, xvho have helped him in many ways during his graduate study at the University of Florida. Dr. T. R, Waldo, of his supervisory committee, has generously given her time to guide the author in his dissertation preparation. He gratefully ^ acknowledges his indebtedness to her. Dr. Horihito Tanbo's sketch of the Briggs cell which is included herein, will remind the author of the hours of consultat-ion Dr. Tanbo gave him. Mrs. A. L. Smith, Dr. R. F. Christman, and Dr. S, A. Hannah have also given the author the advantage of their experiences in water coagulation research. The author tharJcs Mrs. J, G. Larson, Mr. . T, Halters, and Mr. C. Chen for their help in the execution and reporting of his experiments. \ The research was directly supported by VJater Supply and Pollution Control Research Grant WP-139 from the Public Health Ser^/ice, and was indirectly supported by Ford Foundation loans and Public Health Service Traineeships which financed the author's graduate study. ii

PAGE 3

The author shall forevermore tr^,'to express his appreciation and gratefulness to his sons and his wiies Barbara, whose love and understanding have sustained him. ) X2.-J.

PAGE 4

CONTENTS Page ACKNOVJLEDGMENTS ,..... ii LIST OF TABLES vi LIST OF FIGURES vii CHAPTKR I. INTROUJCTION 1 II. HISTOHia^L REVIE/J 3 III. COAGULATION THEORY H r/. PURPOSE Aira SCOPS ......... • 2^ V. EX?ERIi4EI^TAL MATERIALS AND PROCEDURES ....... 26 26 26 26 Ma.t.eria.ls .........••••• ClSLys .....•...•••••••• Ferric Suli'ate .....•.*<••••• Procedures ..... ...*•*••* 2o Preparation of Clay Stispensions .,..... 28 Preparation of Ferric Sulfate ......... 29 Coagulation Tests ......•.•*• 3^ Saraple Preparation ......*<>• 3^ Flocculation .•.......•••* .5— 1 Initial and Residual Turbidity Measurements 32 Residual Iron Determinations • 32 Measurement ofpH 33 Electrophoretic Mobility Determinations 33 Dosing and Mixing ..... 33 I '' Conductance Measurement ....... 3"^ ^ Measurement oipH • 3^ J Particle I-bbility Measurement ........ 3^ VI. DISCUSSION OF RESULTS 3S VII. CONCLUSIONS • 76 APPENDIX ..,..........••••• 7 XV

PAGE 5

Page BIBLIOGRAPHY 102 BIOGRAPHICAL SKETCH 10? J )

PAGE 6

J J LIST OF TABLES Table Page 1. Analysis of Feirric Sulfate Used for Coagulation ... 27 2 Properties of the Clay Suspensions .......... 38 3* Coagulant Dosages 39 k. Effect of pH on Electrophoretic Mobility and Sedimentation of Montraorillonite Clay 79 5. Coag-alation of Montmorillonite Clay Hith 3.O Milligrams Per Liter of Ferric Sulfate 80 6. Coagulation of Montmorillonite Clay With 5.0 Milligrams Per Liter of Ferric Sulfate 83 7. Coagulation of Montmorillonite Clay With 50 Milligrams Per Liter of Ferric Sulfate 85 8. Effect of pH on Electrophoretic Mobility and Sedimentation of Faller's Earth ....".,..... 87 9. Coagulation of Fuller's Earth With 3.O Milligrams Per Liter of Ferric Sulfate ..... 89 10. Coagulation of Fuller's Earth V/ith 5.O Milligrams Per Liter of Ferric Sulfate 91 11. Coagulation of Fuller's Earth With 50 Milligrams Per Liter of Ferric Sulfate ... ..... 93 12. Effect of pH on Electrophoretic Mobility and Sedimentation of Kaolinite Clay ..... ^k \ 13. Coagulation of Kaolinite Clay With 3.O Milligrams ^ Per Liter of Ferric Sulfate ...... 95 l^'. Coagulation of Kaolinite Clay With 5.O Milligrams ; Per Liter of Ferric Sulfate 97 \ 15. Coagulation of Kaolinite Clay With 50 Milligrams Per Liter of Ferric Sulfate 100 vi

PAGE 7

) LIST OF FIGURES Figure Page 1. The Helmholtz Layer Model 16 2. The Gouy-Chapman Diffuse Layer Model. 18 3. The Stem Layer Model 20 -4-. The Briggs Cell 35 5. The Effect of pH and Mobility upon Coagulation. Clay: Montmorillonite Ferric Sulfate Dosage: 0,0 mg/l .... 42 6. The Effect of pH and Mobility upon Coagulation. Clay: Montmorillonite Ferric Sulfate Dosage: 30 mg/l . ,. ^ 7. The Effect of pH and Mobility upon Coagulation, Clay: Mofifensriilsfiltg Fefj*iss Sulfate DsaagiSi $Q pig/l k$ 8. The Effect of pH and Mobility upon Coagulation. Clay: Montmorillonite Ferric Sulfate Dosage: 50. mg/l .... k-6 ; 9. The Effect of pH and Mobility upon Coagu.lation. Clay: Montmorillonite k-8 10. The Effect of pH and Mobility upon Coagulation. Clay: Paller's Earth Ferric Sulfate Dosage: 0.0 mg/l .... k-9 11. The Effect of pH and Mobility upon Coagulation. Clay: Fuller's Earth Ferric Sulfate Dosage: 3.0 mg/l .... 50 12. The Effect of pH and Mobility upon Coagulation. Clay: Fuller's Earth Ferric Sulfate Dosage: 5-0 mg/l .... 52 13. The Effect of pH and Mobility upon Coagulation. Clay: J Fuller's Earth Ferric Sulfate Dosage: 50. mg/l .... 53 Ik. The Effect of pH and Mobility upon Coagulation. Clay: Fuller's Earth 5415. The Effect of pH and Mobility upon Coagulation. Clay: Kaolinite Ferric Sulfate Dosage: 0.0 mg/l 55 16.. The Effect of pH and Mobility upon Coagulation. Clay: Kaolinite Ferric Sulfate Dosage: 30 mg/l 5^ vii

PAGE 8

; Figure Page 17. The Effect of pH and Mobility upon Coagulation. Clay: Kaolinite Ferric Sulfate Dosage: 5.0 mg/l 57 18. The Effect of pH and Mobility upon Coagulation. Clay: Kaolinite Ferric Sulfate Dosage: 50. nig/l 58 19. The Effect of pH and Mobility upon Coagulation. Clay: Kaolinite 59 20. The Effect of pH on Iron Residual. Clay: Montmorillonite Ferric Sulfate Dosage: 3.O mg/l 51 21. The Effect of pH on Iron Residual. Clay: Montmorillonite Ferric Sulfate Dosage: 5.O mg/l 62 22. The Effect of pH on Iron Residual. Clay: Montmorillonite Ferric Sulfate Dosage: 50. mg/l ..... 63 23. The Effect of pH on Iron Residual. Clay: Fuller's Earth Ferric Sulfate Dosage: 3.O mg/l (^ 2^. The Effect of pH on Iron Residual. Clay: Fuller's Earth Firrio Sulfat Dosages 5.0 mg/l ..,..,., 65 J 25, The Effect of pH on Iron Residual. Clay: Fuller's Earth Ferric Sulfate Dosage: 50. mg/l ............. eS 26. The Effect of pH on Iron Residual. Clay: Kaolinite Ferric Sulfate Dosage: 3.O mg/l ..... 67 27. The Effect of pH on Iron Residual. Clay: Kaolinite Ferric Sulfate Dosage: 5.0 m.g/l 68 23. The Effect of pH on Iron Residual. Clay: Kaolinite Ferric Sulfate Dosage: 50. mg/l ..... 69 29. Quantities of Acid or Base Required for pH Adjustment of Fuller's Earth ............. 7I 30. Quantities of Acid or Base Required for pH Adjustment ) of Fuller's Earth ^2 o 31. Quantities of Acid or Base Required for pH Adjustmen of Fuller's Earth , 03 32. Qxiantities of Acid or Base Required for pH Adjustment of Fuller's Earth 74 33. Quantities of Acid or Base Required for pH Adjustment of Fuller's Earth 1 . 75 •'J2.-^:l

PAGE 9

I. INTRODUCTION Clays are the most common source of turbidity in surface -waters J used for municipal and industrial water supplies. Before surface water is satisfactory for domestic and industrial use, most of the clay and other particulate matter present in the water must be removed. Removal of the suspended matter is usually accomplished by alum or ferric sulfate coagulation, sedimentation, and rapid sand filtration. Water treatment plants -which utilize this process produce finished waters > xdiich regularly exceed the minimum quality required for potable water by th U. S. Public Health Service.*^' ^'-^ The specif io Public Health \ Service recommendation concerning turbidity is that the turbidity of drinking water be less than five units. Prediction of the optimum coagulation conditions for the production of such high quality water from a given raw water is very difficult. In the absence of records of previous treatment of x^ater from the same source it is rationally impossible at present. The difficulty of coagulation prognosis is the resiilt of complexly related effects of the numerous properties of the raw water and the chosen coagulation conditions which affect the efficacy of coagulation. Among these conditions and properties are: specific coagulant chosen; coagulant dosage; pH, alkalinity, ionic constituents, and base exchange capacity of the raw water; and size, shape, chemical nature, hydration, and charge of the colloidal particles in suspension. Most of the factors vMch affect coagulation also affect the electrophoretic mobility of the suspended 1 J/

PAGE 10

2 colloidal particles. Electrophoresis is the movement of electrically charged particles, suspended in a conducting liquid medium, -vAiich results from the impression of an electric field, ELectrophoretic mobility is the ratio of the speed of electrophoretic movement to the intensity of the elecJ trical field -Kfcich produced the motion. It is commonly expressed in microns per second per volt per centimeter. Mineral content, coagulant, coagulant dose, alkalinity and nature of clay particles, of the numerous parameters vjhich affect electrophoretic mobility and coagulation, can be chosen and controlled for a selected synthetic clay .suspension. In addition to these parameters v*ich can be selected, pH, base exchange capacity, effectiveness of coagulation and particulate electrophoretic mobility for a given suspen\ sion can be directly measured. The present research has been performed in order to study the : empirical relationships among electrophoretic mobility, residual turbidity, coagulant dosage, and residual iron (the mensurable parameter) I for the ferric sulfate coagulation of suspensions of three different I clays over the pH range between three and ten. To my loaowledge, such I a study has not been attempted before. The relationships discovered are described herein, and the effectiveness of ferric sulfate coagulation of these clays is compared -with that of their alum coagulation, reported earlier by Black and Hannah,

PAGE 11

) ) II. HIST0Ria4L KSVIE/J The earliest records of the coagulation process have been traced by Black i&o torote: Although various crude methods of \^tec purification, generally .' characterised as coagxilationj have been kno^'n and used oince ^ ancient tiixieSj knowledge of the fundamental factors invol.ved in the process has been acquired comparatively recently. The earliest references of scientific interest in coagulation as a process for the treatment of "trater are references to the trorks of D'Arcet '"^ and Jeunet. D'Arcet at the beginning of x.he nineteenth century and Jeunet in I865 sought to establish the v£.lue of the process, but it t-ras not used for the treatment of a public x-rater supply until 1881 After its initiation in Bolton, England* the process ^v'as soon adopted in Eolland and in the United States. The first coagulation patent \^s granted j in 188^ j to Isaiah Smith Hyatt. ^ Folloiving the suggestion of Col. L. H. Gardner^ Superintendent of the New Orleans Water Company* Hyatt successfally treated turbid x^rater by combining the use of perchloride of iron as a coagulant vn.th his process of rapid filtration. His patent covered not only the use of perchloride of iron, but also of "any other suitable agent v^ich is capable of coagulating the impurities of the liquid and preventing their passage through the filter bed," The 188^!Ar.nual Re-port of ^le State Geologist of New Jersey contained results of tests of various salts as coagulants. Austen and Wilber^'^ concluded that of the salts investigated, aluminum sulfate was most effective. Puller-^ published a description of similar studies 3

PAGE 12

'I ; ) > ^i. in I898, He found sulfates of iron and aluminum to be most effective. The chloride of these elements followed next in order, bat their use as coagulants in vjater treatment has not developed on a practical scale. In the year of Fuller's publication, W, B. Ball,^ at Quincy, Illinois, began using a mixture of ferrous sulfate and lime for coagulation. Fourteen years later, E. V. Bull reported the first use of chlorinated copperas. This chsnical was not tried again until 1928, 12 •i&ien. Hedgepeth and Olsen used it for the successful treatment of a highly colored water. Ferric ion was produced by oxidation of the ferrous sulfate with chlorine. As use of the process has become prevalent, the number of researchers and the scientific disciplines they represent have proliferated. Black -^ and Packham in recent reviews of coagulation theory and related literature, thoroughly cover the many facets of historical and contemporary research, theory, and practice. The literature cited in the present study will consequently be limited to that directly related to the research herein reported. In 1923 the first of a series of studies by Theriault, Clark, 1^ and Miller was reported. The paper by Theriault and Clark described their trea-bnent of several buffered solutions with various amounts of alum to determine the effect of pH upon the rate of floe formation. They assumed that the optimum pH for coagxiLative treatment of viater vrould be the pH at vjhich the minimum time x-ras reqtiired for the formation of the aluminum floe. They found that generally the best floe formation occurred in the pH range ^.95 to 5^ Higher alum doses resulted in broader ranges of pH over vdiich rapid coagulation was observed. These workers were unable to explain the difference between values of pH

PAGE 13

5x-Aiich accompanied rapid coagulation in their study and the higher values observed to be most effective in "water treatment plants. In the same year, Baylis, using alvm to coagulate natural I waters of varying alkalinity, noted that for a given pH, vjaters of higher alkalinity required larger quantities of coagulant to accomplish y satisfactory treatment. For any chosen tjater he isas able to reduce the alum dosage hy adjusting the pH -with a strong acid, thereby neutralizing a portion of the alkalinity originally present. The values of pH for optimum coagulation were in the range 55 to 7.0, -1 ^ Late in 1923, Miller published the second of the Theriault, Clark, and Miller papers. He described their constituent study of precipitates formed by the mixing of potassium alum solutions with sodiiom hydroxide solutions of various molarities. Although the anion concenJ tration of the coagulant had no practical effect, the amount of alkali ; mixed with a given quantity of al-uminum exerted considerable influence I upon the composition of the precipitate. The ratio of aluminum to sulfate ion was determined for the floe over the useful pH range, but the exact chemical nature of the floe i^xas not determined. His review of treatment plant records revealed that optimum coagulation occurred within the pH range 5.4 to 8.5. In another series of tests (1925) Miller investigated the effects of various anions upon the formation of floe in solutions of potassiiim alum. Generally, the anions of higher valence exerted greater influence upon coagulation, and~less alkali was necessary in their presences The composition of precipitates from ferrous and ferric salt -1 Q solutions was the subject of another 1925 publication by Miller.

PAGE 14

) 6 There precipitates from the iron salts were found to be very similar to floe from the analogous altrniintun salts, except that they were formed at pH values significantly lower than the minimum pH values observed for alxm floe and that they remain insoluble above the maximum pH at \lch. alum floe exists. The results of Miller's experimental vrork allowed him to make three statements concerning the optimum conditions for floe formation with both iron and alumin-um coagulants, (l) There must be present in the water a certain minimum quantity of the metallic coagulant ion. (2) There must be present an anion of strong coagulating power such as the sulfate ion, (3) The pH must be properly adjusted. In 1928 Bartow and Peterson examined the effects of various salts upon the formation rate for alum floe over the useful pH range, ) Many of the salts studied increased slightly the rate of floe formation • and extended, on the acid side, the pH range over which floe occurred. In a paper published in 1928 and now regarded as a classic in its 20 field, Mattson described the coagulation of clay suspensions and the electrophoretic mobility measurement of clay particles as mobility was altered oj various dosages of aluminum salts. His work demonstrated the effect of the different anions and cations upon particle charge. He specifically predicted the utility of electrophoretic studies in water ^ coagulation investigations and pointed out the predominant role played by the hydrated oxides formed during coagulation. In three articles Black and others^"'" ^^'^^ report investigations, made on a laboratory and semi-plant scale, of the effects of several anions upon the rate of floe formation for both aluminum and ferric salts. They proposed improved conditions for laboratory study of

PAGE 15

) 7 coagulation. Also they recommended the use of larger samples and containersg constant stirring, and isohydric indicators. With such improvements they could duplicate pilot plant results. Anions were found to be useful to increase the rate of coagulation and to broaden the pH ranges of rapid floe formation. The pH range for floe formation was wider for ferric sulfate than for alum, but no iron floe formed between pH values 6,5 and 8,5. The authors stressed the importance of adsorption as an important mechanism of turbidity removal and also attributed zones of no floe fo3:*mation to a change in particle charge. In the slightly more than a decade since the publication of Black's 193^ article, the interest of water chemists seem to have been focused upon activated silica and other coagulant aids. The work of oh. o e Langelier and Ludwig '^ marked the return of many researchers to the y field of basic coagulation theory. They initiated the use of synthetic clay suspensions and called attention to the importance of base exchange capacity and particle size distribution of raw water, the action of metallic hydrolysis products as "binders," and the occurrence of perikinetic coagulation. 96 In 1958, Pilipovich, Black, and others examined the relationships among pH, zeta potential, base exchange capacity, coagulant dosage, and turbidity removal for the coagulation of suspensions of five clays with alum. Clays of high base exchange capacity required considerably larger doses of coagulant to accomplish coagulation than did clays of lower base exchange capacity. The authors reaffirmed the superior effectiveness of hydrolysis products over that of trivalent aluminum ions as coagulating agents. Matijevic and others (I961) coagulated lyophobic colloidal

PAGE 16

8 suspensions of kno-wn particle charge and concentration vjith alvuninum nitrate to determine the nature of the species formed during alum floe formation. Results indicated that below pH 4 the simple trivalent hydrated aluminum ion prevails. Between pH 4 and pH 7 the probable j prevailing complex is the tetravalent AloCOH)^^. At higher pH values J the tetravalent hydrate is transformed into a divalent form -tiiich in turn yields to a complex of zero charge. In the same year Black and Hannah, using alum and various polyelectrolyte coagulants and coagulant aids, investigated the relationships among pH, electrophoretic mobility, base exchange capacity, coagulant dosage, and turbidity removal for the coagulation of three different clay suspensions. Several equivalents of coagulant per equivaImt of clay base exchange capacity werg necessary to effect satisfac; tory coagulation. The pH range of best coagulation was 7.5 to 8,5 in ifiiich range the electrophoretic mobilities of the clay particles was slightly negative. In some cases fair perikinetic coagulation occurred ]. at pH values less than 4,5. In his study of the precipitation of aluminum hydroxide resrulting from addition of alum to solutions containing bicarbonate, carbonate, or chloride of sodium, Packham (i960) evaluated the degree of precipitation by measuring soluble aluminum residuals and by photometrically 1 recording turbidity of the system as a function of time. The maximum amount of precipitate was formed in the pH range 5,5 to 7,2, Maximum coagulation rate occurred between pH 7,2 and 7*6, and maximxam immediate precipitation was observed at pH 7,1 or 7,2, As his criterion of coagulation effectiveness, Packham chose the coagulant dosage required to reduce the initial turbidity by one-half ^

PAGE 17

9 for a given set of conditions. The pH range he observed for optirram coagulation conditions based upon the use of this criterion does not agree mth the pH range for good coagxdation x-ihich others have reported. The difference results from the selection of different coagulation efficiency criteria. Use of Packh3ins criterion to select conditions for coagulation in a treatment plant x-rould result in -prater of unsatisfactory quality. If he had chosen a much greater turbidity removal efficiency as his criterion, his prediction of optimum coagulation pH range would have coincided -with those commonly reported. Packham expressed the opinion that the coagulation of clay is principally accomplished by the enmeshment of the clay particles in the metallic hydroxide precipitate formed by reaction of the coagulants. Orthokinetie eoagolation is th term now used to n^iie the mechanism he described, on Mackrle (I962) has presented arguments favoring tlae acceptance of the physical theory of stability and coagulation of colloidal clay suspensions. He has attached greater significance to the value of the psi potential, the potential difference between the "Helmholtz layer" and the bulk of the suspending mediura, than to the value of the zeta potential, the potential difference between the plane of shear and the bulk of the suspending medium. He has suggested that the metallic coagulants react to form crystalline hydrous oxide sols li^ch destabilize the clay suspension by mutual coagulation. His conclusions were based upon his xiork -with ferric sulfate and clay suspensions and upon crystallo graphic studies of such precipitates reported by others. In his tests pH and coagulant dose were allowed to vary simultaneously, and no measurements of electrophoretic mobility -were performed to

PAGE 18

10 permit an examination of the independent relationships among pH, coagulant dosage, turbidity removal and zeta potential. In 1962 Sttrnim and Ihvgaxv^ cited the -work of others and identified, for both iron and al-uminxim salts, the hydrolysis products of highest effective charge and assigned tentative formulae to them. They J also called attention to the aging effect of iron and aluminum sols and emphasized the chemical nature of particle charge caused by ionization as contrasted -with charge produced by physical adsorption. The majority of their laboratory tests were performed to determine the specific chemical interaction betvreen various functional groups normally found in ijater and the metallic .coagulants in common use. They concluded that metallic complexes other than hydrous o:3ddes can be formed as a result of the inttractions and that the interaction of the fimctional groups ^ can appreciably affect the pH at i-jhich optimum conditions of coagulation occur. They also presented a method for carrying out laboratory jar i I tests at constant pH and convincingly justified the desirability of such a procedure.

PAGE 19

> in. COAGULATION THEORY Any explanation of coagulation theory must be preceded by a description of the physical and chemical characteristics of the colloidal suspensions to be coagulated. The suspensions customarily encountered in the treatment of water for municipal and industrial use are dilute. Even -vhere raw water is taken from streams idaich carry heavy sediment loads, the water to be coagulated is dilute because presedimentation is used to remove the readily settleable particles. The materials vMch are removed from the dilute suspensions by coagulation can be generally classified as turbidity or "color." In their study of the nature of colored surface vjaters Black and Christman*^ have ascertained that "the materials responsible for color in water exist primarily in colloidal suspension in the water," They found particle size and number of particles to vary with pH, but particle size was generally less than 10 mn.. Although the present reseai*ch is not directly concerned with coagulation for color removal, color compounds do merit mention, because their colloidal nature and susceptibility to coagulation are similar to those of colloidal turbidity particles. The primary source of turbidity to which the present investigation pertains is colloidal clay. Clays are complex aluminum silicates of sedimentary origin. Their lattice structures consist of layers of silicon-oxygen sheets and aluminum-oxygen sheets. These sheets, common to all clays, occur in different orders in the different classes of clays. 11

PAGE 20

12 Kaolinte, montmorillonite, and fuller's earth have been used in the present research. An important physico-chemical characteristic of clays ishich is considerably affected by the placement order of the sheets is base exchange capacity, Kaolinite and montmorillonite, of the clays used in y the present research, represent the low and high extremes of base exchange capacity -tiiich results from crystalline structure. Kaolinite consists of an alumina sheet and a silica sheet i/Mch are combined to form one layer, idiereas montmorillonite has an alumina sheet included between two silica sheets to form one layer. These layers have a definite thickness, measurable by x-ray diffraction methods, which is constant for a specific clay; but the lattices extend to the irregularly broken edges of the crystals in th other tm directions. In viater the parallel layers tend to split apart and become dispersed. Because of the small coulombic forces between similar adjacent silica sheets in montmorillonite, the surfaces of the layers become hydrated and therefore separate to a considerable extent. Kaolinite layers are held together tightly by hydrogen bonding between hydroxide ions on the bottom of one layer and oxygen ions on top of the next layer and thus resist separation. A more detailed discussion of clay structure is given by Hendricks, -^^ Marshall and Krinbill"^^ have stated one of the results of the structural differences between kaolinite and montmorillonite. The base exchange capacity of kaolinite varies with particle size, vftiereas the base exchange capacity of montmorillonite is almost independent of particle size. Particle size is the property upon which the definition of ^*^**c% aiifcr-i iB^n a\^emt^m

PAGE 21

13 colloids is based, but colloid behavior results from phenomena involving the tremendous surface area inherent in particles of colloidal size. An example of the importance of surface phenomena is that the classifications of colloids, lyophilic and lyophobic, result from such a phenomenon. Distinction between the two classes depends upon the wettability of the surface of the particle. The particles possessing poorly wettable surfaces are called lyophobic. Lyophobic colloids are vmstable, because by coalescing they offer a smaller surface area at which a solid-to• liquid interface must be formed. Since the smaller interface results in a lower surface energy state and a lower total energy state for the system, the coalescence is spontaneous. Lyophilic colloids, on the other hand, are stable because of the attraction of the eurface for the solvent the wettability of the surface. These lyophilic colloidal solids which exhibit surface interaction with the solvent are, therefore, the colloids of primary concern to water chemists. Another surface phenomenon which greatly influences the properties of a suspensoid is the electrical charge possessed by the particles. The charge can be initiated by several different surface mechanisms; still other surface phenomena affect charge magnitude and sign. Of the several mechanisms which can produce particle charge, we shall first consider the process that might be termed preferential ionization. As an example of this mechanism, we shall consider minute crystals of silver iodide in equilibrium with a saturated solution in which the ionic concentration product is roughly 10" The force with which the two ions are held in the lattice differ greatly. Despite its larger size, the more polarizable iodide ion is held much more

PAGE 22

14 forcefully than is the silver ion. When the concentrations of the two I ions in the solution are equal, more silver ions escape into the solui tion than do iodide ions; hence, the crystals are negatively charged. If the silver ion concentration is increased, the charge can be reduced i to zero; a farther increase of silver ion concentration can cause a / charge reversal, I Ionization of functional groups -tMch are connected to the particle by covalent bonds is a second and important source of particle charge. Examples of particles thus charged are proteins and ion exchange materials. Proteins are formed by long chains of amino acids joined by peptide linkages. Some of these amino acids carry an adI ditional carboxyl or amino group [ihich can remain free and exposed to the solvent. Thus they can form COO and M^ ions x^aich are covalently I attached to the particle. The ion exchange materials are porous solids f ] -tdiich have built-in acids or basic groups. The natural zeolites and many clays are included in this general category. Among the other mechanisms -which can cause particle charge are chemisorption and adsorption of specific ions resxilting from van der VJaal's forces. For particles vjhich are charged by the adsorption of ;i specific ions, the surface charge density is greatly affected by concen[ t i tration changes of those ions in the solution. i : I In all of the mechanisms mentioned above, particle charge was achieved by the separation of \inlike charges (the destruction of electroneutrality ) Because of the large amount of energy involved in the separation of these charges, electroneutrality is disturbed only on the ultramicroscopic scale. When any one, or a combination of these mechanisms, produces an excess of charges of one sign in any locality,

PAGE 23

15 there must exist an equal excess of charges of opposite sign in the immediate vicinity. Thus, we find surrounding and veiy near each charged particle an excess concentration of ions of counter sign. These ions are called counterions or gegenions. The counterion excess occurs only in the solution veiy closely surrounding each charged ^ particle. Throughout the rest of the suspending medium electroj neutrality prevails. I When we speak of electroneutrality and areas of excess charge, I we implicitly assume that the conditions we describe are the averages, with respect to time and space, of the several dynamic equilibria which prevail. Brownian movement is one spatial equilibrium which must be averaged. Ion exchange is a physico-chemical equilibrium of consider. able sffeot. Eaoh of the equilibria, ineluding these two, caus fluctuations with respect to time. Based upon these averaged conditions, several models have been proposed to explain the possible arrangement of the charged surface and its surrounding array of counterions. These models have been used by chemists, physicists, and mathematicians as bases for their computation of theoretical values of parameters for many colloid systems. Probably none of the models proposed is correct, but each has been useful in the explanation of some observed phenomenon. The simplest of these was proposed by Helmholtz.^^ Fig. 1 is a schematic section which shows the charged surface and the surrounding medium. By ionization or some other mechanism, the surface of the particle is charged. This excess of negative charge in or on the surface attracts toward the surface the positive ions in the solution. Helmholtz proposed a model in which all of the counterions so attracted

PAGE 24

16 CHARGED SURFACEPARTICLE r-HELMHOLTZ LAYER OF COUNTERIONS M te y ;0;\va:v^:.0i -SURFACE OF SHEAR BOUND SOLVENT ,. BULK OF SOLUTION < hO a. DISTANCE FIG. 1 '-THE HELMHOLTZ LAYER MODEL

PAGE 25

17 are located in a single surface parallel to and very near the surface of the particle. His model is analogous to a parallel plate condenser. The graph of potential in Fig. 1 represents the potential difference between the bulk of the solution and the point within the solid-liquid interface that corresponds vfith the abscissa. Fig. 2 depicts the Gouy-Chapraan^^ model v/hich is named for the two scientists who first considered it in detail. They recognized the Helmholtz assumption to be an oversimplification of the spatial arrangement of the neutralizing counterions. Reasoning that thermal agitation would prevent such an unifom arrangement of gegenions, they proposed a diffuse layer of variable but finite thickness which contains an excess concentration of counterions. In formulating their model they also considered the shielding effect of the counterions located near the particle upon the attractive force existing between the particle and the more distant counterions. The proposed model is characterized by a potential versus distance graph, for which the potential decreases at a decreasing rate as distance from the surface of the colloid increases. In addition to the gegenions which are attracted to the particle, a sheath of the solvent is tightly bound to lyophilic colloids. This sheath is pulled along with the colloid whenever some force such as that of gravity causes relative motion between the particle and the bulk of the solvent. Thus, there exists surrounding a colloid a surface of shear which contains the bound solvent that accompanies the particle as it migrates. If an electric field is impressed upon a colloid system, the colloids migrate toward the pole of opposite charge at a speed

PAGE 26

18 CHARGED SURFACEPARTICLE GOUY-CHAPMAN DIFFUSE LAYER ^ — — f^ ^ I ^ @ r^ ^ r\ BULK OF ^ ^ SOLUTION --^SURFACE OF SHEAR BOUND SOLVENT PSI POTENTIAL j N^^EJA POTENTIAL DISTANCE Fia 2 -THE GOUY-CHAPMAN DIFFUSE LAYER MODEL

PAGE 27

19 proportional to the potential which exists between the surface of shear and the bulk of the solvent. (The "bulk of the solvent" refers to the portion of the solvent out of the effective interaction range of the charged colloids, that is, where electroneutrality occurs on micro as well as macro scale.) This shear surface potential upon which the velocity of a charged colloid in an electric field depends is called the zeta potential, 2 The movement of the charged colloid that is caused by impression of the electric field is referred to as electrophoresis; and the ratio of particle speed to the intensity of the electric field is termed electrophoretic mobility. In units commonly used for expression of mobility and potential, a Iji/sec/v/cm mobility is equivalent to a potential of I3 mv. Since th valecity of colloid movement caused by a msasurabl field intensity can be observed through an ultramicroscope or by moving-boundaiy frontal methods, the electrophoretic mobility of colloids is a directly measurable parameter. This parameter is closely related to particle charge, surface-charge-density diffuse-layer thickness, and particle surface potential, which are important parameters but are not directly measurable. (The surface potential,)^, mentioned above is diagramatically presented in Fig. 2),. A model suggested by Stem^^ represents further sophistication of the Gouy-Chapman diffuse layer theory. It includes the Stem layer of adsorbed counterions which are held in actual contact with the surface of the colloid. A section of the Stem model and a typical potential curve for it appear in Fig. 3. In addition to the zeta potential and the surface potential,-}^ the potential at the interface between the Stem layer and the Gouy-Chapman portion of the double diffuse

PAGE 28

20 CHARGED SURFACEr STERN LAYER PARTICLE -^ E) SURFACE OF SHEAR BOUND SOLVENT BULK OF SOLUTION t < 1— U O Q_ SI POTENTIAL-^^ ZETA POTENTIAL DISTANCE FIG. 3 -THE STERN LAYER MODEL

PAGE 29

21 layer is indicated and represented by the symbol'}^ \ The models described above are useful as they provide a basis for appreciating the phenomena which influence the stability or instability of a given colloidal suspension. The primary phenomenon favoring stability is, of course, the mutually repulsive force between similarly charged colloids. A particle with its complete double layer is electrically neutral, so that it exerts no net coulombic force upon another. This situation exists in the case of particles which are sufficiently distant from each other. As two particles approach, however, the double layers interpenetrate and interact. Should the two particle surfaces finally touch, there could be no more diffuse layers between them to icr^in them from th iff get of th repulsivt coulombic feroe. It ia the work required to thus distort and finally destroy a part of the diffuse double layers which causes most of the repulsion. Charge density upon the particle surface and concentration of ions in the solution surrounding a particle are parameters which directly affect stability. Low ionic concentration in the solution causes greater thickness of the diffuse layer. High charge density results in large psi potential. Thus low ionic concentration and high charge density contribute to colloid stability. The relationship of these parameters to zeta potential can be seen in Fig. 1. The conditions which contribute to stability result in high values of zeta potential; therefore, zeta potential, or more important, the directly measurable electrophoretic mobility can be a useful index of stability. A less important phenomenon contributoiy to stability is the van der Waal adsorption of similiions. This occurs to a significant degree

PAGE 30

22 only under very specialized conditions, and the effect is more important with similiions of large molecular weight. The phenomena which are influential toward instability are much more numerous than those causing stability, because the foimer consist of all the phenomena which can impair the effectiveness of the latter. Of the instability mechanisms, the two most important are the mutual attraction (and coagulation) of oppositely charged colloids and the van der Waal attraction. The other mechanisms are of much less direct importance, or they effect instability through their influence upon or in conjunction with one of the two mechanisms mentioned. Van der Waal forces' are alxjays attractive; but because they result from dipole interaction, they decrease approximately with the third power of the distance betv;een the particles. Thus they are effective only when the particles are brought into extreme proximity. Coulombic forces between similarly charged colloids oppose their approach toward each other with force which decreases roughly with the second power of the distance between the particles, so that some other force is required to push them close enough together for van der Waal forces to prevail. Brownian motion caused by thermal agitation is one source of such force. Mechanical agitation of a suspension can simi3^rly contribute to -instability. Some other factors effect instability by reducing the particle charge or potential of the colloids. The concentration of "potential determining ions" in the solution exerts a large influence upon this parameter. The hydrogen ion concentrations will determine the charge density and, hence, the potential of any particle which depends upon the ionization of carboxyl or amino groups as its source of charge.

PAGE 31

23 The concentration of inert ions in the solvent can affect stability by increasing the probability of particle collisions as their niimber is increased and by causing compaction of the diffuse layer. Ionic concentration is not the only ionic variable which influences stability. The valence and size of the specific ions present exert effects upon particle charge and potential. The Schulze-Hardy rule is a general statement of the observed phenomena that divalent counterions are more potent in the production of instability than monovalent counterions and that trivalent ions are much more powerful coagulation agents than the divalent counterions are. Ion size and less important anomalous characteristics of specific ions influence their ability to cause colloid instability as is evidenced by the Hoffmeister series in which various anions and cations have been listed in order of flocculating power. Because of the numerous mechanisms which are responsible for de stabilization, the many variables which affect stability, and the differing chemical and physical nature of the sundry colloidal particles which may be encountered, it is practically impossible to calculate the conditions under which satisfactory coagulation of a given colloidal suspension will occur. It is desirable, therefore, to examine the empirical relationships of the parameters that can be measured so that the work involved in cut-and-try coagulation can be reduced.

PAGE 32

IV. PURPOSE A1\TD SCOPEThe preceding chapters have re\aeued the raost significant work dealing with vjater coagulation and have siiianarized contemporary theoii.es concerning the basic mechanisras involved. Because of the complexity, of relationships among the n-aniberous variables X'jhich affect coagulation,' many researchers have sought to discover empirical relationships among the measurable parameters in order that they might more fully understand basic coagoilation mechanisms Black and Hannah i, Pilipovich and others.; Black and WillemSj '5 and Black and Christman have pursued such a course in the electrophoretic studies of the coagiilation of several colloidal materials vri.th 39 29 various coagailants and coagulant aids. Packham-^ and Kackrle have reported some ferric sulfate coagulation research, but neither of them performed electrcphoretic studies of coagulation in which the effects of coagulant dosage and pli have been separately determined. The primary puarpose of the present research is to add to the body of knowledge resulting from the work of Black, Hannah, Willems, Christman>j and Pilipovich: it vjill seek to learn the relationships among electrophoretic mobility, residual turbidity, coagulant dosage, and residual iron content of three clay suspensions resulting from their ferric sulfate coagulation. The three clays were chosen so that suspensions containing equal concentrations of them would exhibit low, medium, and high base exchange capacities respectively. The literature contains no references to vrork of this nature. 2i^

PAGE 33

25 Because certain materials and procedures i^rere utilized, the study, in addition to its primary purpose, will allow direct comparison of the results of ferric sulfate coagulation with the results of alum coagulation of these clays reported by Black and Hannah. Moreover, the results of these tests may yield experimental confirmation of coagulation theories previously presented by others. The scope of the work was limited to study of electrophoretic mobility, residual turbidity, and residual iron content in the ferric sulfate coagulation of suspensions of a low, a medium, and a high base exchange capacity clay over the pH range 3 to 10. The ferric sulfate dosages were 0, 3, 5, and 50 mg./l respectively. The procedures involved xjere jar tests, residual turbidity and iron determinations, and electrophoretic mobility determinations.

PAGE 34

V. EXPEKQ'IENTAL MATERIALS AND PROCEDURES The experimental portion of the research consisted of (1) jar tests for the ferric sulfate coagulation of three clay suspensions and (2) the determination of mobilities for the flocculating colloids in those suspensions. Coagulant dosages of 0.0, 3.0, 5.O, and 50 rag/l of ferric sulfate were used, and pH was varied over the range 3 to 10. Final dissolved iron content, final pH, and initial and final turbidity of the jar test suspensions were measured, and pH values of mobility samples were determined. Materials Clays The clays used to prepare turbid waters were Kaolinite 4 and Montmorillonite 23, obtained from Ward's Natural Science Establishment, and fuller's earth, obtained from the Floridin Company. Ferric Sulfate The ferric sulfate used as coagulant was the commercial grade manufactured by the Tennessee Corporation, Atlanta, Georgia. Ferri-Floc is the registered trademark by which the manufacturer designates this material. The analysis of the particular sample of ferric sulfate used appears in Table 1. 26

PAGE 35

27 TABIE 1 ANALYSIS OF FERRIC SULFATE USED FOR COAGULATION Constituent Per Cent by Weight Total ater Soluble Iron 21.50 ater Soluble Ferrous Iron (Fe-H-) O.7O Water Soluble Ferric Iron (Fe-H-f-) 20.80 Water Insoluble Matter 2.00 Free Acid (as HgSO^) 2.55 Moisture 2.5I

PAGE 36

28 Procedures Preparation of Clay Suspensions Hannah has given the following description of the clays used and of the first steps in their preparation: The kaolinite and raontmorillonite consisted of large lumps of diy clay. These materials were crushed and groujid with mortar and pestle and were then ball-milled for Zk hours. The fuller's earth, vrhich was obtained as a dry povrder, was ball-milled for 24 hours. The preparatory efforts of others, described above, were performed early in 1959At that time Hannah determned the base exchange [' capacity of each of the three clays in the manner prescribed in Official Methods of Analysis of the Association of Official Agricultural 4-1 Chemists Since their preliminary^ preparation the dried clays had been stored in closed glass containers. With the clays in the condition described, the present experiment was begun x^rith the weighing out of four 3'75-g aliquants of each clay. Each aliquant, in turn, was mixed x^fith 500 ml of demineralized xjater and dispersed by mixing in a Waring Blendor for five minutes. The four aliquants of each clay suspension were placed in a separate covered beaker and allowed to hydrate for 24 hours. At the end of the hydration period, each suspension was slowly passed through a sodium-cycle, ion-exchange column to replace the natural, exchangeable cations of the clay with sodium ions. A glass column of 25 mm diameter contairiing a 12 -inch depth of Nalcite HCR cation-exchange resin was used for the ion exchange process, which was The Waring Blender is manufactured by Waring Products Company, New York, Kex-r York.

PAGE 37

29 developed by Lewis. Just before the treatment of the three clay suspensions the resin in each was regenerated with one liter of ten per cent KaCl solution. Stock suspensions were prepared by diluting each of the three suspensions to twenty liters with demineralized water. The three suspensions were stored in polyethylene carboys. Preparation of Ferric Sulfate A 2-kg portion of commercial-grade ferric sulfate was taken from a 100-lb bag of the material which was supplied by the manufacturer. After being mixed thoroughly, the sample was quartered, and 500 g were reserved for analysis. Approximately 100 g of the remaining material was finely ground -with mortar and pestle. Aliquants of 1.000 g were weighed out and stored in glass vials having tight-fitting polyethylene stoppers Because the hydrolysis reactions of ferric sulfate are dependent upon time and concentration, a fresh coagulant solution of the chosen concentration was prepared daily. One of the previously weighed aliquants x^as quantitatively transferred to a 200-ml volum.etric flask, which was then filled to the mark with demineralized water. This particular concentration was chosen, because less concentrated solutions became cloudy and x-jere found to contain considerable volumes of hydrolyzed material within four hours after the initial mixing. A minimum of twenty minutes was allowed for the mixing, which was accomplished with a magnetic stirrer.

PAGE 38

-30 Coagulation Tests Sample preparation After the preparation of the aforementioned clay suspensions, approximately six months passed before coagulation tests were begun. A high-speed mixer, equipped with a stainless steel shaft and propeller, was used to thoroughly mix the suspensions. The minimum time of mixing for any suspension was one hour. Mixing was continued while 200-ml aliquants were pipetted into 8-oz polyethylene bottles. Also, triplicate 25-nil samples of each suspension were pipetted into beakers, and residue upon evaporation was determined for each in accord with the procedure specified in Standard Methods for the Examination of Water and Wastewater ^ which hereafter will be referred to as Standard Methods To provide an ionic concentration in the suspensions comparable to those of surface waters, 20 ml of 5.00-g/l sodium bicarbonate solution was added to each of the 200-ml suspension aliquants. In the final clay suspension samples, diluted as described below, the sodium bicarbonate concentration was 50 mg/l. In the morning of a typical day of coagulation tests, immediately after preparation of the coagulant solution, six of the 8-oz bottles containing the desired clay were selected, and their contents were quantitatively transferred to separate two-liter volumetric flasks in which 500 ml of demineralized water had been placed. Ten ml of chlorine solution containing O.^qo g/l of CI2 were pipetted into each flask. (During chlorine demand tests, that specific quantity had proved sxifficient to satisfy the demand of the suspensions and still to leave about 1 mg/l of free available chlorine to oxidize the ferrous iron present to

PAGE 39

31 the ferric state. Such a prechlorination dosage is common for treatment of surface water.) Sufficient amounts of 0.1 N NaOH or HCl were added to the flasks to yield the desired pH of the final suspension, and the flasks were filled to the mark with demineralized water. A tefloncovered magnetic bar was placed in each flask, and the flask's contents were mixed for five minutes with a magnetic stirrer. One-liter graduated cylinders were used to transfer half of each flask's contents to separate ii-S-oz square jars which were placed on the multiple laboratory stirrer (hereafter referred to as the jar test machine). The remaining contents of the flasks were retained for mobility determinations. Flocculation. The flocculation process was begun by rapidly stirring the suspensions while adding the correct coagulant dosage to each of the six suspensions. After receiving the coagulant dosages, the suspensions were mixed rapidly for 2 minutes and then slowly for 28 minutes. Next the stirring paddles were removed, and the suspensions were allowed to settle for 10 minutes. The rate chosen for rapid mixing was 100 rpra. The speed initially selected for slow mixing was 40 rpm, but soon after laboratory tests were begun, 5 rpm was adopted for slow mixing since it appeared that higher rates might cause disintegration of the floccules. Promptly at the end of the sedimentation period 250-ml sajtiples were drawn from each jar. An apparatus similar to that described by Cohen x^as used to siphon each sample from approximately an inch below the surface of the supernant. The settled samples so obtained were used for residual turbidity and residual iron determinations. _P p-../''? six. jar, variable-speed stirrer used is a product of Phipps ana bd.Td, Inc., Richmond, Virginia.

PAGE 40

In itial and residual turbidity Eeas-ax-ements Turbidity of the samples \^2.s aieas'ared in a Lumetron Model ^-50 Filter Photometer. The procedure recoimnended by the manufacturer required the preparation oi a calibration curare for each of the three clays used. Calibration for a single clay involved the preparation of eight to ten suspensions so that their turbidities uniformly covered the desired turbidity range. The optical density of each suspension >ras measured in the Luraetron.; for 650-i3|i light over a 75ma path;, Also, the turbidity of each of "che suspensions was detenrdned -i-rlth the Jackson Candle TurbidiKieter in the 46 manner set forth in Standard ^ Methods Because coagulation efficiency t-jss to be judged upon the basis of turbidity removal, determination of irdtial turbidities t^ras necessary. An average value of initial turbidity for each of the three clay s-aspensions X'las obtained hj measaring the initial turbidity of the suspensions for the jar tests^ in which coagulant dosage w-as zero. Re sidual iron detex-'mination ^ The supernatant samples which were obtained in the manner described above were filtered prior to the deterrflination of their iron content. A fine, smooth ^ quantitative filter paper^^ t-ias used. Filtration vsb.s necessary, because some of the iron present t-ras chemically or physically bound to floe particles of such small siae that they did not settle ou'c in the short settling period alloi^ied. In a vrater treatment plant the iron bound to such small *The L\imetron Photometer is a product of Photovolt Corp., New lorks KexT Tork. %.e paper used was paper No. 130^1 manufactured by VJill Corp., New Toi-k 52,, Ix^ew York.

PAGE 41

33 floccules would not be foup.d in the finished v:ater, because sedimentation of longer duration or passage through the rapid sand filter would renove thera. Filtration of the iron samples thus makes the laboratoryjar tests more nearly analogous to plant conditions. The iron content of several samples was determined x-iithout prior filtration. In a later chapter results of these determinations are identified and are compared with results for similar filtered samples. The variance of iron content between the two kinds of samples emphasizes the necessity for filtration. The phenanthroline method appearing in Standard Methods "^ was followed for the iron determinations. A 530-mM. filter and a 75-mi. light path were chosen for the Lumetron Model ii-50 Filter Photometer used for the colorimetric iron determinations. Measurement of pH The final pH of the coagulated clay suspensions was measured with a Beckman Model G pH meter. Electrophoretic Mobility Deterainations Dosing and mixing. A period of four to eight hours usually elapsed after the final dilution of the clay suspensions before mobility determinations were made. In order to attain thoroughly mixed suspensions after the quiescent period, each suspension was stirred for five minutes on a magnetic stirrer just prior to the mobility determinations. The individual suspensions were dosed vixth. the appropriate amounts of coagulant at the beginning of the mixing. The Model G is a product of Beckman Instruments, Inc., Fullerton, California

PAGE 42

3^Condnc-t^ mee j n8a^suremerit, Specific conductance was determined by the procedure described in Standard .. Methods A Model RCI6BI Conduc>! tivity Bridge vdth a pipette-form conductivity cell having a cell constant of 1 cm""" -was used for the measurements. Mgasy^j^eaent^^of _r)H The pH of each suspension was measured immediately before the mobility of its particles t-jas deten-nined. A Beckman Model G pH Meter was used for these measurements, ^g; r ."& J^lg---^QMliti:measureiGent Tae equipment and procedures used for the roicroelectrophoretic mobility determinations are those described in the paper by Black and Smith. ' The detailed description of the physical dimensions and construction methods for the glass cell appear m an article by Briggs,-" who designed it and for T-iom it is named. Since both procedure and equipment are described in detail in the references citedj, only a general discussion of them mil be included here. In addition to the general account, however any deviation from the suggested methods of Black and Smith xvlll be delineated specifically. The Briggs cell is constructed of Pyrex^ glass. Its rectangular crosssectional shape and area are practically constant over the central portion of its longitudinal axis. A filling funnel is located at one end, and removable electrodes and outlet stopcocks are connected to either end of the cell. It is mounted in a metal stand wiiich holds it in proper position on the microscope stage and supports it yhen it is not in use. Fig. 4 is Tanbo's sketch^"^ of a Briggs cell. The Conductivity Bridge is a product of Industrial Instruments, inc.. Cedar Grove j New Jersey. ""Pyrex is a product of the Coming Glass Company, Midland, I-'iichipan*

PAGE 43

35 _J -J LlI U 00 o L?J I Ll

PAGE 44

-36Two Briggs cells were assembled, mounted, and calibrated before jar tests and mobility determinations were begun. Plastic Wood was found to be a more satisfactory packing material than any previously tried. If removal from the holder is ever necessary, the cell can be freed easily by soaking the assembly in acetone or a similar solvent. During the course of this study several improvements in technique were developed. (1) The cell holder was attached to the microscope stage with a pair of small clamps. This procedure prevented accidental misalignment of the cell during determinations. (2) Another deviation from the suggested procedure was the use of a ^-rara glass bead in each of the electrodes to prevent rapid mixing of the mercuric nitrate and potassium nitrate solutions in the electrode. This procedure was proposed by Briggs. (3) It was found that the mercuric nitrate solution could be introduced into the electrode most easily with a hypodermic syringe fitted with a 2-in needle, (if.) When sufficient mercuric nitrate solution was injected to bring the liquid level about 3 mm above the narrowest portion of the opening between the upper and lower chambers of the electrode, no trouble was experienced in placing the glass bead without entrapment of air bubbles beneath it. Serendipitously it was discovered that for the suspensions studied, electrodes prepared in the manner described above could be used over periods as long as a week. Moreover, for these particular suspensions it was possible to rinse the cell by passing demineralized water through it while it remained clamped in place on the microscope stage. The application of a low-intensity vacuum to the outlet tube caused Plastic Wood is a product of Boyle-Midway, Chamblee, Georgia. /

PAGE 45

37 water velocity of sxifficient magnitude to remove the settled floccules; therefore, it was not necessary to remove the cell from the microscope stage more often than once a week. Once the system was aligned and the cell clamped securely in place, a check was necessary only at the beginning and ending of each day's determinations to be sure that the microscope was focused on the stationaiy layer inside the cell. The water aspirator which was used to pull cleaning water through the cell was also useful as an aid in filling the cell with a suspension without the entrapment of bubbles. The particle velocities were determined by visual observation and were timed with a stopwatch. A previously calibrated occular micrometer located in the plane of the microscope image of the stationaiy layer served as the measured course over which the flight of the particles were timed. The polarity of the electrodes was alternated between individual timings. Thus, of the twenty particles timed for each suspension, ten were observed to move toward the left, and ten moved toward the right. An exception to this characteristic movement was noted at the isoelectric point. At the isoelectric point various particles were observed simultaneously to move very slowly toward either electi-ode. The mobility of the particles of a suspension was computed from the mean of the time observed for each particle to travel a fixed distance; therefore, the computed mobilities are time-averaged rather than velocity-averaged mobilities.

PAGE 46

x'-L:J.OUiDii_'^J '^i Sorie physical charactoristias of tha clays i-jhich xjere •ased and tha procedures for preparation of the suspenaior-s studied have been d3seribed in preceding sections. Table 2 pressnts the turbidities, gravimetric concentrations and base exchange capacities of these 1 Li, jCy O o TArJUi 2 i-cj^ijxiiiib On mil; oi_-.I bUb.:Ld'iblOj.\!S iitial Base Exchange Base Exchange Residue Upon Capacity of Capacity of Lsaj Turbidity nig/l Evaporation Glays ieq/. Suspensions ^ieq/l Mcntjiorillcnite 58 Fuller's Earth 79 Kaolin! te 29i> 62.0 C-i-.O 7i:'0 1,150 265 87 71 3 6.^ Microequivalents X'Till be abbreviated P-eq, The base exchange capacities are important in relax.icn to ccagu -Lant dosages, Alsos the effects of coagulant dosage upon mobility and coagulation in the present study tjill be compared with sirtilar effects of coagulant dos-ge in the aluia coagulation study reported hj Hann _"c IS uesirabj-c;, "cnerefore.^ ohat the coagulant dosae"es used in both p. n ~ JO

PAGE 47

39 investigations be expressed in terras corn.parable to the units of base exchange capacity eraployed herein. These expressions appear in Table 3' TABIE 3 COAGULANT DOSAGES Ferric Su?!.fat e^. Altai Big/l 't^eojl mg/l M-eq/l 3.0 3^ 50 k5 5.0 57 15.0 135 50.0 570 100,0 900 Although norcdnal dosages of 5-0 ng/l -were included in both studies 5 the effective dosage of alusinura ion is considerably smaller than that of the ferric ion because water of hydration constitutes a larger portion of the arora dosage. Probably the most raeaningful comparisons between the txm investigations can be siade for dosages of (1) 5.0 mg/l of aluia ana 30 m.g/1 of ferric sulfate ^ (z) 5.0 rr.g/l of al-om and 5.0 ffig/l of ferric sulfate and (3) 100 lug/l of alum and 50 mg/l of ferric sulfate. These corabinaticns of dosages are not equals but the two lox^rer ferric sulfate concentrations bracket the lower alxira dosage j and the highest dosages of the ti:o studies represent concentrations greatly in excess of those necessary for good coagulation. The choice of a satisfactory criterion of good coagulation is particularly difficult. Adequate coag"alation must result in an extremely high degree of turbidity reraoval. The optical property 5 turbidity,^ isus'G be relied upon as an index of reraoval efficiency. Use of this

PAGE 48

^ Zi-O property is coiriplicated hj the effects of shape 5 size 5 number and. refractive index of the particles responsible for the turbidity in a given sample An example of the manifestation of these effects has been presented in Table 2 The suspensions contain equal gravimetric concentrations of the three clays, but the turbidities they exhibit are widely divergent. For water treatment plant operation a practical criterion comiTionly used requires that the turbidity of coagulated and settled water be five units or less. Such a cr-iterion is probably too stringent for use in this study because the short period of sedimentation xxTas insufficient for the settling of the srcaller particles v;hich contribute most to the tv.rbidity of a suspension „ Moreover., the selection of a specific residual turbidity as the criterion would be to recraire a much larger removal efficiency for the kaolinite suspension (initial turbidity: 29^ units) than for the montmorillonite (initial' turbidity: 58 units) Packham~ has arbitrarily chosen as his coagulation criterion that dosage which reduces initial turbidity by 50 V'^'^ cent„ This choice is undesirables because it differs so x-.\idely from acceptable conditions encountered in water treatment plant operation. Vie have arbitrarily chosen 90 per cent removal of initial turbidity as the criterion of satisfactory coagu.lation and results of the ti-;o studies Xv^ll be compared on this basis. The residual turbidities and particle mobilities for each permutation of clay and coagulant dosage have been gr-aphed as fxxnctions of pH. Both graphs for a single pertautation appear in the same fig-are. (The residual turbidities are expressed as percentages of the initial turbidity of each suspension.) Figures fs 6j 7s and 8 are based upon "'— l/*'--^V*VWai^l*fc ^ IM

PAGE 49

„ ill the turbidity and mobility results for the montmorillonite suspensions x^-hich were treated vjith 0, 3, 5, and 50 mg/l of ferric sulfate respectively. In Figo 5 J pH is observed to exert minor influence upon ir^ontiKorillonite mobility in the absence of any coagulant. Final turbidity is little affected by pH. Seine turbidity removal w-as accomplished by the agitation and sedimentation of the jar test procedure, bat the clarification so achieved i-jas not significant. Hannah's electrophoretic studies revealed that the mobility of the suspended clay alone vras relatively independent of pH, but the magnitude of the mobilities he reported were appro:ximately 30 per cent smaller than those recorded during the present investigation.. The latter agree Jiraeh more closely Xvlth the raaxlmm negative clay mobilities reported by 20 Matt son and by others than do Eannah-So Such agreement can not be interpreted to be a proof of accuracy, because i-iattson for example TOrked with a different clay and did not report pH values. However, as consideration of figures below \-AH indicates the graph of the mobility versus pH for zero coagulant dosage does fairly well define the masijdum negative mobilities observable for the suspensions and coagulant under, consideration. In view of this relationship ^ comparison of mobilities for zero coagulant dosage x-Jith Mattson^s maxiraum negative clay m-obilities appears to be worthi-jiiile. The lack of agreement with Hannah's values led to a reconsideration of the eq-uipment and procedure for mobility measurement. Of the mechanical devices ^jhich are used, the ammeter is the unit m^ost probably capable of introducing an error of such size and consistency. The precision of the meters is about one or two psr cent of the full scale

PAGE 50

-42r^K'K: ,W^'^c< "*' -2 3 4 5 6 ^120i 1 r — I r .o > 100 Q 80 cr 60 _j 40 < Q 20 if) yj 7 8 9 10 11 120 SLOW MIXING: 40 RPM -o-Q5 RPM -o-o100 80 60 40 20 2 3 4 5 6 7 8 10 11 FIG. 5. -THE EFFECT OF p H AND MOBILITY UPON COAGULATION CLAY: MONTMORILLONITE FERRIC SULFATE DOSAGE: 0.0 MG / L

PAGE 51

reading, and the accuracy is not likely to be as good as the precision. Vlhsn conditions such as the electrolyte conductivity or the available pov:er supply voltage make it necessar;,' to measure a current corresponding to only ten or fifteen per cent of the full scale readings the error involved in the determination may be si:-: to ten times as large as that for a reading at full scale. Perhaps the use of some more sensitive method of field intensity measurement circuitrj;would allow closer agreement among the results of independent investigators. Fig. 6 shows the effect of pH upon mobility and turbidity removal of the ffiontmorillonite clay suspension for the ^~To.g/l coagulant dosage. Increasing pH was accompanied by a gra.dual increase in the magnitude of the negatixi-e mobility. The sasie trend was shoxai for Hannah's ^-rngjl. dosage, and^ as was previously mentioned, the magni-iv-ides of the mobilities x-jere only about ?5 psr cent of those from the present study. Ifcereas the only good coagulation resulting from that alun dose was perikinetic coagulation at pH less than ij-Os 3-i^-g/l of ferric sulfate yielded good coagulation from pH 5 '5 to pH 6.6. With an increase of ferric sulfate to 5 ifig/l £s shown j-n Fig. 7? the Increase in mobility vjith increasing pH becomes greater jalthough the curve is not so smooth. The most noticeable effect of the increased coagulant dose is the broadening of the zone of good coagulation over the range pK k-.S to 7.8. The 50-iGg/l ferr-ic sulfate dosage resulted in change reversal for values of pH less than 50* '^'^Fig. 85 the algebraic decrease in mobility with increasing pH was observed again as it had been in the previous figures = Hannah^ s comparable alura dose of 100 mg/l resulted in a mobility curve of similar shape above pH k-^7i but the isoelectric

PAGE 52

^^ ;>yv il4 6 7 8 9 10 11 pH FIG. 6. -THE EFFECT OF pH AND MOBILITY UPON COAGULATION CLAY : MONTMORILLONITE FERRIC SULFATE DOSAGE: 3.0 MG/ L

PAGE 53

-4523456789 10 11 10 11 .120 SLOW MIXING: 40 RPM -a-a5 RPM -o-o100 H80 60 40 20 O 9 10 11 FIG. 7 -THE EFFECT OF p H AND MOBILITY UPON COAGULATION C L AY : MONTMORILLONITE FERRIC SULFATE DOSAGE: 5.0 MG / L

PAGE 54

i^6 > 1001n 80 ^ 60 40 9 10 11 120 SLOW MIXING: 40 RPM -o-D5 RPM -o-oQ hi O 5 6 7 8 9 10 11 2 FIG. 8. -THE EFFECT OF p H AND MOBILITY UPON COAGULATION C LAY : MONTMORILLONITE FERRIC SULFATE DOSAGE: 50. MG/ L 100 80 60 40 20

PAGE 55

if? point (pH 6.6) was considerably higher, and below pH k.6 as the solubility of aluiTxinuia increased the charge reversal effect of the alum also decreased. The alum dose resulted in good perikinetic coagulation below pH i!..5 and both alum and ferric sulfate produced good orthokinetic coagulation from pH S^S to 8.8 and from 5.5 to 10.0 respectively. Fig. 9 is a composite of the four preceding figures. It allows ready comparison of the broadening of the zone of good coagulation \n.th increasing coagulant dosage. Also the zone of perikinetic coagulation below pH 3,7 and the zone of poor or no eoag-alation from roughly pH -^ to pH 5 are easily identified. Probably the most important relationship graphically illustrated in, the fig-are is that the isoelectric point occuring at pK 5,3 for uhe one dosage causing charge re-.-arsal marks the beginning of the pH zone of most economical coagulation of the particular clay and for the arbitral^ good coagulation criterion chosen. The next series of figures pertain to the coagulation of fuller's earth. Fig. 10 illustrates the effect of pH in the absence of coa,gulant. As was -ohe case ivrith montmorillonite in the absence of coagulant, negative mobility increased vjith increasing pH value. Perikinetic coagulation occured below pH if .2 although no such coagulation had been observed for montmorillonite Beginning with Fig. 11 3 the mobility curves become more complex as the varying effectiveness of the coagulants for charge reversal at different values of pH becomes more evident. For the 3-mg/l dosage 5 the original particle charge is reversed over a pH range of slightly more than one xanit. For this particular dose, comparison of mobility curve shapes for ferx-ic and alum coagulation are \!lrtually impossible because of lack of similarity. Good ferric coagulation was observed over the

PAGE 56

v;-V ? is 9 10 1 1 ''" 1G/L FE2SQ4)3 I ^ mn io. 9. -THb EFFECT OF p H AND MOBILITY UPON COAGULATION CLAY-: MONTMORILLONUE

PAGE 57

49 3 4 T r 7 8 9 10 11 120 D D SLOW MIXING: 40 RPM -D-D5 RPM -K>-o3 4 5 6 7 8 9 10 11 pH FIG. 10 -THE EFFECT OF p H AND MOBILITY UPON COAGULATION CLAY: FULLER'S EARTH FERRIC SULFATE DOSAGE: 00 MG/ L

PAGE 58

-50ri 2 3 4 5 K i20r— T 6 7 8 9 10 11 120 7 8 9 10 11 FIG. 11. -THE EFFECT OF pH AND MOBILITY UPON COAGULATION CLAY: FULLER'S EARTH FERRIC SULFATE DOSAGE: 3.0 MG / L

PAGE 59

51 pH range 5.6 to 7,3. When the ferric sulfate dose was increased to 5 mg/l (Fig. 12) the zone of charge reversal was increased to 3 pH tmits, and good coagulation occurred above pH 5.5. A most interesting effect noted in this particular portion of the study is the greater efficiency of 5 rpra slow mixing for perikinetic coagulation and the better orthokinetic coagulation that accompanied the ^0 rpm slovr mixing. The mobility curve for the 50-mg/l ferric sulfate dosage (Fig. I3) appears to be typical of excess ferric sulfate dosage for the three clays studied. Particles were positive below pH 6.3 were isoelectric at that points and became increasingly negative as pH was raised to 10. The slope of the curve is steepest iniiiiediately above and belox-J the isoiigiStris psiat hi@h sltarly indiaatig a -mm q£ mMMosa stability, Hannah's comparable alum curve changes less abruptly at the isoelectric point, and the apparent effect of increasing solubility at low pH is evident. Ferric coagulation is good above pH 14-. 2. The composite graph for fuller's earth (Fig. l^i-) clarifies the relationship between the isoelectric point of the overdosed suspensions and the beginning of the zone of efficient coagulation. As was the case for montmorillonite, it appears that the "overdosed isoelectric point" does mark that beginning. Alsoj the broadening of the zone of good coagulation -with increasing dosage is evidenced. Individual mobility and turbidity graphs for the kaolinite clay are shown in Fig-ares I5, 16, 1?, and 18 ^ and Fig. 19 is the corresponding composite. The mobility curves are very similar to those for fuller's earth, and the comments above concerning the latter are qualitatively applicable. Regarding coagulation, the kaolinite is much more nearly

PAGE 60

-528 9 10 1 1 r 9 10 11 120 D D D SLOW MIXING: 40 RPM -o— D5 PPM 3 45 6 78 9 10 11 100 -80 -60 40 20 FIG. 12. -THE EFFECT OF pH AND MOBILITY UPON COAGULATION CLAY: FULLER'S EARTH FERRIC SULFATE DOSAGE: 5.0 MG/ L

PAGE 61

539 10 1 1 1 1 + 2 5 6 7 8 9 10 11 FIG. 13. -THE EFFECT OF pH AND MOBILITY UPON COAGULATION CLAY: FULLER'S EARTH FERRIC SULFATE DOSAGE: 50 MG/ L

PAGE 62

s^ 7 3 9 10 11 T 6 7 8 9 10 11 RG. 14. -THE EFFECT OF p H AND MOBILITY UPON COAGULATION CLAY: FULLER'S EARTH

PAGE 63

' 55 3 5 6 7 8 9 10 1 1 120 SLOW MIXING: 40 RPM 5 RPM 80 60 40 20 2 3 5 6 7 10 11 FIG. 15. -THE EFFECT OF p H AND MOBILITY UPON COAGULATION CLAY : KAOLINITE SULFATE DOSAGE FERRIC 0.0 MG / L

PAGE 64

56 1^.^20 7 8 9 10 11 120 6 7 8 9 10 11 pH FiG.16. -THE EFFECT OF p H AND MOBILITY UPON COAGULATION CLAY : KAOLINITE FERRIC SULFATE DOSAGE : 3.0 M G / L

PAGE 65

57 9 10 1 1 I— ^ 120 9 10 11 1 i120 10 11 FIG. 17 -THE EFFECT OF pH AND MOBILITY UPON COAGULATION CLAY: KAOUNITE -FERRIC SULFATE DOSAGE: 5.0 MG / L

PAGE 66

-588 9 10 1 1 1 1 s 1 ^ 2. 9 10 11 FIG. 18. -THE EFFECT OF p MOBILITY UPON COAGULATION CLAY : KAOLINITE FERRIC SULFATE DOSAGE: 50. MG / L

PAGE 67

59 J ^ — I ^ i i i i__ i i j i \ MOBILITY UPON CGA CLAY : KAOUNITE-

PAGE 68

60 similar to the montmorillonite Most important, however, is that the "overdosed isoelectric point" near pH 6.0 marks the beginning of the 52 zone of good coagulation. Black and others have shown that the same relationship between the "overdosed isoelectric point" and the zone of good coagulation exist for the ferric sulfate coagulation of natural colored waters. An additional observation of possible importance concerns base exchange capacities of the suspensions. There was no observation of charge reversal in the present investigation for dosages for which the ratio of coagulant dose to base exchange capacity (both expressed in l^eq/l) was less than 1.0. .Good coagulation occurred for all three clays, hovTever, with a dosage of 3 Kig/l. For this dose the ratio mentioned above for montmorillonite is O.i+Q, for fuller's earth is 2.0, and for kaolinite is 5<3> Work with ferric sulfate doses in a range which would yield ratios near unity should reveal more interesting information on charge reversal phenomena. The U. S. Public Health Service has set O.3 mg/l of iron as the maximum allowable concentration of that constituent in drirJcing water. In the present research, therefore, determination of the residual iron content in the supernatant was necessary in order that suitability of the various treatments could be evaluated. The results of the determination are graphically presented in Figures 20 through 28. The most important information obtained from the iron determination was that for all the pH zone of good coagulation, the residual iron content is below the maximum allowable concentration. Another interesting observation which is generally applicable for the ferric sulfate coagulation of the three clays is that residual iron values are less

PAGE 69

61 9 10 11 _j z O < Q (/) LlI 9 10 11 F1G.20.-THE EFFECT OF pH ON !RON RESIDUAL. CLAY: MONTMORILLONiTE FERRIC SULFATE DOSAGE: 3.0 MG/L

PAGE 70

62_J o z O < Q CO LjJ DC 1.4 1.2 1.0 0.8 0.6 04 020.0^ 3 T 4 5 6 7 8 9 10 11 ^ FILTERED UNFILTERED-^^-^^ ^ 1.4 1.2 1.0 0.8 0.6 04 02 OO 9 10 11 pH FIG. 21. "THE EFFECT OF pH ON IRON RESIDUAL. CLAY: MONTMORILLONITE FERRIC SULFATE DOSAGE: 5.0 MG/L

PAGE 71

63_J 5 o •mimJ < Q (/) tr 1.4 1.2 1.0 0.8 0.6 0.4 0.2 2 r 0.0^ 2 3 4 5 6 7 8 9 10 11 FILTERED -0 0UNFILTERED -)^— ^^ 1.4 -11.2 1.0 0.8 0.6 04 3 4 5 9 10 11 pH FIG. 22. -THE EFFECTOR pH ON IRON RESIDUAL. CLAY: MONTMORILLONITE FERRIC SULFATE DOSAGE: 50. MG/L

PAGE 72

64 2 3 4 5 O o QC < 00 UJ DC l.4r 1.2| i.o[ Q8| 0.6 04 7 8 9 10 11 i1.4 1.2 1.0 0.8 -p. 6 0.4 9 10 11 pH FIG. 23.THE EFFECT OF pH ON IRON RESIDUAL. CLAY: FULLER'S EARTH FERRIC SULFATE DOSAGE: '3.0 MG/L

PAGE 73

65_J o y Q CO LlI ^L *^ 5 6 7 1.4 1.2 .0 9 10 11 FILTERED -0 0UNFiLTERED -^— ^<~1.4 1.2 1.0 0.8 ^9 10 11 pH FIG. 24. -THE EF RESIDUAL. CLAY: FERRIC SULFATE ECT OF pH ON IRON FULLER'S EARTH DOSAGE: 5.0 MG/L

PAGE 74

^ 66 ^ 10 11 5 Q CO QC 7 8 9 10 11 pH FIG. 25.THE EF RESIDUAL. CLAY: FERRIC SULFATE ECT OF pH ON IRC )N FULLER'S EARTH DOSAGE : 50. MG/L

PAGE 75

67(D < Z) Q 2 3 4 5 6 9 10 11 FIG. 26THE EFFECT OF pH ON IRON RESIDUAL. CLAY: KAOLINITE FERRiC SULFATE DOSAGE: 3.0 MG/L

PAGE 76

68 O 2 o < Z) Lijfci 1.2 I. Or Q8 L W 0.2 0.0^ 2 3 4 5 6 7 9 10 11 T i il4 7K FILTERED -0 0UNFILTERED -^^— ^|1.2 no ^ ^ 0.4 ^ J 02 9 10 11 pH FIG. 27 -THE EFFECT OF RESIDUAL. CLAY: KAOLIN! FERRIC SULFATE DOSAGE pH ON IRON TE 5.0 MG/L

PAGE 77

69 8 9 10 11 O ,y' J 3 ftf 9 10 11 FIG. 28.THE EFFECT OF pH ON IRON RESIDUAL. CLAY: KAOLINITE FERRIC sulfate' DOSAGE: 50. MG/L

PAGE 78

70 than the inaximuia allox'^able concentrations in the lox^; pH range doi-m to pH 3.8. Furthermore, even at pH 3.O no supernatant contained more than 1 mg/1 of iron in solution. These low residual iron values for clay coagulation are much lower than those reported for colored v:ater coagulation Black and others-' have reported residual iron values for comparable coagulant dosages in colored water that are six times as large as the largest observed in the present study for the same pH of 3.0. At pH values up to ^ or 5, the ratio was even larger. The difference in behavior of the clay and color colloids in this respect point out the need for a more thorough understanding of basic coagulation mechanisms. For the present research it vjas necessary to determine whether the pH of the samples employed for mobility determinations were identical mth the pH of the corresponding jar test suspensions. Figures 29 30s 31, and 32 show the close agreement between the respective pH values. Fig. 33 is a composite of the four, which shows the effect of coagulant dosage upon the value of pH. It should be remembered that the suspensions which viere used were synthetic preparations containing 50 mg/l of sodium bicarbonate. A more poorly buffered solution might not yield such agreement between the two. Even i;-iith the systems used it \ias difficult to dose the suspensions vjith the exact quantities of acid or base required to obtain evenly spaced mobility and turbidity curve points. Only the curves for fuller's earth are shoxra, since those for the other clays were similar.

PAGE 79

71 n I MOBIL! '^•^tlEST J SAMPLES RS -n— DES -1.2 1.0 OB 0.6 0.4 0.2 0.0 02 H FIG. 29.QUANTITIES OF ACID OR BASE REQUIRED RDR pH ADJUSTMENT OF FULLER'S EART

PAGE 80

72 2 3 FIG. 30.QUANTITIE REQUIRED FOR pH FULLER'S EARTH S OF ACID OR BASE ADJUSTMENT OF

PAGE 81

70 1.21 (T T" 08h t 0.6 Xna Lij MOBiLl "EST h> 3 niirW FiG. 31. REQUIRED FOR FULLERS EART i I itio S OF ACIDOR BASE H ADJUSTMENT OF

PAGE 82

74 50. MG/ L J, ^o fiaiae O ^ RG.32. QUANTITIES OF ACID OR BASE REQUIRED FOR pH ADJUSTMENT OF FULLER'S .EARTH

PAGE 83

75 .2F A p prpPir^ Or 0.0 "Al LlI UJ £L t/) H UJ i 6 o .or -^-0 Q6h ^-^ FIG. 33. QUANTITIES OF ACID REQUIRED FOR pH ADJUSTM FULLER'S EART OR BASE ENT OF

PAGE 84

VII COKCLUSIONS In the present investigation, charge reversal at low pH was observed for all three clay suspensions when the ratio of coagulant dosage to base exchange capacity (both expressed in M-eq/l) exceeded three. With increasing pH the mobilities of these suspensions decreased until the isoelectric point (in the pH range 5-7) occurred. This "overdosed isoelectric point" was fo-ond to Eiark the beginning of the zone of efficient orthokinetic coagulation of all three of the suspensions studied. Such behavior has been reported for the coagulation of 52 colored water vjith ferric sulfate by Black and others. Furthermore, an analysis of the results of clay coagulation vrith aloiri i-eported by i{,0 Hannah revealed that the "overdosed isoelectric point" was an indicator of effective coagulation conditions for those suspensions. The base exchange capacities of the suspensions studied were not found to be directly or proportionally related to the coagulant dosages required to effect satisfactor)!coagulation j but they did significantly affect the coagulant dosages required to cause charge reversal at low pH. For all of the combinations of ferric sulfate dosages and clay suspensions studied, it was observed that in the pH zones of good coagulation the residual iron values were less than the maxiiriuia allowed in drinking water by the U. S. Public Health Service. Moreover, even at low pH the residual iron values were less (by a factor of six to ten) 52 than those reported by Black and others for the coagulation of colored 76 I iga i.iirifci>'-tri

PAGE 85

APPENDIX

PAGE 86

79 P ^H -rj •H O "O a .H •H -P Xi f-i i o -P' r-i:^ o H O CV rt O •-a S O CO ON NO C^ NO CO H I ON NO ON NO CO O CN! O NO NO -P ca tH .ri rH CNJ H ON oa O H q^ £>J^ £>Cvi •ri & 1 • A K /i OS C^ NO NO oP, O CO s H w — 5 o^ oJ K S H CD NO CM CM O CM C^ CM O O o CO CO C!N o CM ON CO H Vn NO O CM NO CO C3N : 00 o NO NO 9 CM NO VA NO o CO CO H H On o o CO c5 t o o o 00 s W I g H cti -P O -P

PAGE 87

80 H irci W H •H. PI O -^ •H -P ,o 7i EH -^ •ri -P •ri H K-P H -H EH O TJ O K -p -1 & CiJ H H On C\2 S ON VO ^ ON CA NO o NO NO NO On CO ^ CQ DCO o B o o e o I o ^ ^ ON S On O NO O CN> Cvi CN2 NO NO H H H !>!>!>ON On CA NO C>N NO CNi CNJ --'A NO o CNi ON r-\ CN> -^ s o o ON o NO NO £>• CO ON H CO ON -p o H > 2/ o O o 2: K> H -H-H O H a •d ^ W ,0 B Oi o en ct! O rt o *-3 s UA C^ t>H C^ (—; VPi ^J"^ NO NO ON ON w e H H H H H H 1 1 i 1 1 8 INCA o o o o CO NO NO Cno NO CO 0^ On CO o CNJ CO o 00 CO o CO ON ON o
PAGE 88

81 rH pi G H MD C^ u^ ^ o 0^ -:^ ^ >J-\ ^ MD iA C^ -, n o(NJ C^ -* VD oH o CM CM H C^ CM •rl f-1 txO • o • • a • a • e O O O O O o O o O O O O O l>j > -P -P Vi ^ -O -H •H -P rO rQ C ?^ ?-4 0) ?i ^ ^ ^ H Cv! H ^ ^, CJN ^ C?v NO CM CO 00 !h H H CO ?\ DCM 0^ o ci -:a.9 •^A y H CD CO DCO rt>ON oCM c^ CJN NO CA H rt Eo cd • • • • • • ^ VO £>£>00 ON ON CM c^ ^ ^ U~N NO ^ 1 G rt Is td^ s. vr\ W ^ -^g g' ;^^ CM ^ C^ o ^ ^^ s rf g^^ ^ u{ S •H'^-^ • e • • e • B o r-i H H H H H H H H H H H H O (D •ST" tn t i i I i s 1 1 S 1 l>5 -p H fi. o ra o H H fi rd o •-3 S o CM iNH NO >A H CiN NO £>00 C^N NO CJN NO CT\ O H O CM CO CO r-i C3N CM O C3N CM CA o s CA CA rH d u^ o CO • VA NO ^ o O o O O ^-A vr\ CO £>NO lA CM o H CM CA ^ VA NO cv CO C3N q C>H f-t H H (— H rH H CM ^A H H H iH 1— H rH H H H

PAGE 89

82 H 6-1 O o -P ^H .ri •H O 'd •d ,J3 pi ^ O E-i H ^ -P E-t

PAGE 90

-83^J^ •H (0 H a H O — .rj -P f-i a) E-t H 0) nJ Oh Id •H -P •H th-l O -ci Q w •P p. C3 u H o 3 a -p s 8 y^ -P o H l> ^ O O OJ CO OJ C\J -^ va 00 ^ § VO ^ c^ c^ H o ON H H O CM H H ^ H CJ H ?\ c\a H C^ ON H >^ -p H .H CS -Td c>CM O •S:^ c^ H H [in U H CO CM 5^ ca H C^ o C?N CM o Vi^ 00 u^ o CM O VO ON C!N CO c^ H -4NO ??. o CTN CO CM C3N VO CJN c^ cv C^ 00 CO o ca ^ {>u-\ £r O CM CJN Csl H £>c>CM CM CA t>!>c^ CX) D• • • 9 9 ft • o iH H r^ !-\ Cvj H H H H CM H H a I i i i 1 1 1 I 1 CTn CvJ CO CM VO ^A O CM o VO VO CM o o o o CM r-i O O e o O MO CJN CO C^ ON i — I VO w VO o ON 00 o CM S^ ON o c^ o ON CJN o CM C^ 9 o CJN CM O O CTN !>-OQC3NO"AVDrv-CQC3NQ O^

PAGE 91

84H O c^ ^ ON o o o o o o o o o o VA 9^ ON >A UA -P P4 5w H o -n 3 1^ H (D "^ a 4J ClJ c| 03 3: i^cS CA CA ^ o ^ lA CA CA £NDNO lA Gn O NO w .v_ o ^ H ^ NO nD H o He C3 CA • e • • • ON ON 00 D^ >A NO c^ ^ VPv !>i e -P o ;^ ">" o 00 ?^ NO •H• • ^ o H H 5 -P H .rl H CA H On O H eo NO ^ w !q • • ON p, o ra S H CO (M CNJ O mCNl CO CNi C^ ^ lA 03 >-3 S VA vr\ >A IN[>c^ On On On ON O H H H H H H r-i H rH H CM CTN C7N o CM 00 >A O H H •H •P O

PAGE 92

85-d 'd O" — • WHS -P o •^ ij -P •H G H PU CO •H H •H a CO ON CO CO 5i 0^ c^ CO H o o o CO o o A CM H c^ CVi c\J A iA C5 w ^ ^ o O g •a; K-P O ti C(} W H 'H H O H gn ft o w ciJ O H a o .^J_ C^ H^ o n^ ^ O H O CD 4^ C\i (3N C^ 00 o C\J o CV2 O C7\ H O c^ H O + o o o c^ 00 o + ^ d o H o I vO C^ "^ o o o o CvJ CvJ o Cvl U-\ \0 H w o I H "^ o VO H I O CX) o + ^ o\ <^ C^ o CO \o CO CO
PAGE 93

86 I •KO 'Ci rH G U >5 o -p **-*^ o w H .H H O H a ft O CO H o c 53; S o cr a o •-3 S o r-i O o o o H o H H H On oa C^ CO H o c-i H O H u^ Oi ON P\ u^ c^ ^ u^ CM H ON o + o H o + ON CO H CO O in H o o o o o NO NO H OJ CM r-i r-i Oi cn ^ ^A 000 CM CNJ C\i NO o CM CO o •H CO o CO •H •S O -P

PAGE 94

8? •Hi O 'O c •H •H H O 0) Cm CO "^ OS o o H H H ON c^ o £>H CJ H q o c^ CO v£) CO \o 00 H H VO 00 o o o CO VO 00 CO CO VO VO C^ f^ ^ o\ 00 CM VO "^ so ^ CM CO w^ VO 5^ o CO C3N 00 c-i M3 O CTs VO O OS o to Ov CO o i o VO • H I CO H I Cvl H 1 H a VO H • CM i O H i C3N CM CM i H I CO H • CM i •P O H o r^ o o CO CO lO o O o\ C3N CO & o 1>H CM VO INo CM ^ H s • • • • • • • • • CO CO v^ VO C^ CvCJv .* C^s •:* c;n CO H 5 cr o cr* S o CO CM H O 0>j VO C\J VO H VO O o CO CM CO a o 00 ON H CvJ CO ^ CO ON r-i H CO 00 00 00 H H iH H

PAGE 95

-88 H •P .So. •P -P •H C H •d •P •H H O (D H & CO H Oj CD o a* s a o •-3 is; •P Xi p; ^. VO .* CO 00 s VO VO VO VO .aCvi CO o VO d cd EO 0) CO H Cii O O O 00 o CO s Pi CQ H c •H o E-t

PAGE 96

-89H •H CO Pi C H U to rH S O H O o 00 o o o o o o CM o o CO o On EH i-1 CO o (in P>4 o E-t H I 3 O s E-< w s 03 O o H EH !=2 o -p O -o '8 -p !q •H S fH ^ O EH Eh f^ r-j H P^ -H nJ -P C m .H P^ M tH K ^ U to !^ H o Id o o n NO NO o I 00 >n CO ON UA o >n NO CO o I CO en I CNi 00 00 o eM t ^ NO NO CNi NO CnI I CNoo o CM O On H oo en CNi CNi CNl r-i r-i r-i O CNI H 00 NO vn CN! H vn NO NO CNI H O NO o CNl H O CO
PAGE 97

90 H CD ^ G H •H ^ M o a> •ri H I to s f-1 • tfl o •-3 S O O O H O H H CM On O o CO 00 CO en i 0^ o On CO H ^ o CvJ C^ o H ON o On H O NO H H 1^ NO c^ H H ON o CO & O

PAGE 98

-91 CO H oi pi. s:; H 73 O" •H U o M ;=?>. ^ -r> H .H •rj O 'O 'O .rl .H -P ^ !^ H •rt CO -H H o VO -* VO VO VO (N! 00 V3 CO CO MD CO C^ H H CSJ H H O O o o U\ 0^ vpk & lA -:3O H en lA Eh pi 5h -P O Xi CD P^ TO r-i Kl So nJ 1>5 S -p o H O H CM CvJ CA VA CA (r\ A CM o u^ cv CA ^ -* On CO OJ H H H o oj lA D"^ A MD CJN CA MO CO V3 CA 00 H CA O i o CM I 5^ CM i O o t o + lA H 4o CX) o I 00 On V3 CA CM I O CA CM CM I d O o H s o o -)-> o •ri H •H ^ p. Q 02 ci Q> H o cr rf O O H CA O >A CM ^A H C^ o NO CA o NO NO CA O o o o o •?. A OO NO ON O >A CA ON o vr^ NO CA CA O o ^ ^A O lA CM NO NO lA C^ H VA O CO ^A CM O CJN o CM O VA VA C3N O CA CM lA CA lA ^

PAGE 99

m E 8 H ^ ^^ WHS +^ liH -H 4 •H +J ^ U Pi ?-i cc! p_, ^ M H -H o r; Q m -if H O H ? -P cti So cci :rl O >^> H o S w 92 o o ^ o o o O C\2 O CA o o o o o o 00 H CA ON H CO CO C7\ CM ^ vO CA H ^ MO CA C^ VO ON H c^ ?^ CM CO vo CO NO Cn! O VA ON NO o NO NO o NO CNl c^ ir\ tr\ NO NO CO GO NO ON CV Cn! CN! CO CNJ H i >> i-^ Q 'H -H H OJ O H Oh -:3-ri S K A 3 NO a,o to •^ H IX! D 'a' rf o K S H cr o S C^J C^J CN3 CM U^ u~\ rH NO "> 00 NO >A CM O o e m o H o o o o 5^ ci O CVi ND I>00 ON H •-a e; NO ON CJN CTs CTN O H H H H H CNl CO o o o CNJ O O o £>o o o I-i CM CNJ c^ CO O CiN H ON O g •P O -p I E-I it

PAGE 100

93 ^ ri U O M oa -*, o JNu^ U>v OD VO ^ VO VO o o CV' H O O o O o o o o H o H H E-i CO o H Hi o EH to oi 1-5 >^ O =5 CD O H -So. .5^ 4 3 £-1 •n H O Gi HCl o S Jar CM c^ CO H CM CO o c^, c^ A C^ H ^ C^ c^ r^ CV! c^ CJ CM H ON -* -:^ C^J d •J^ CM CM !>CVl VO o CO CO i>VO £>.. CO s o\ CM CM i — I lA H ^ C^J CO H ^-^ CO ^ .— i ON CO iA u^ 00 lA i>O lA J>o H o o O O o O H H CM CM CM 4-i-Jh 4i! 1 1 ! i 4^ O CO 'h .hi h CM H C-N O CO o c^ c<-.. r-i VO ^A O H P^ o -* O VO -^ o lA C3\ !N CO O Ouiti 8 C o o o a 9 e 6 o c^ CA -^ ^ ^A VO VO VO JNc^ CTV C3v P^ O W S .^^ CO CM o CM CO CO O O VO VO O o o o o O O H o VO lA O CM ••A CO O C5N VO o C^ C!V O H Cs! C^ _^ !>£>CO CO 00 CO CO crv o •ri & g •H O E^

PAGE 101

9^ H O Eh Is o H f=5 Q o o H Eh O ^ >^ -P +J Vi .rj O X) •H H -P /:i fn (D pS ^ o e-i f-! H H Q> rt Cd CLi -ri -H W .H i-1 O ^ Q) H fi -P !>j o -p---~ •H > -H O O W vn CvJ Do Du^ c^ vO o H 00 co o\ CO £>00 c^ CO >'-3 Jt* oJ H 4 c^ c^ IJ^ \o i>o u-\ ^ u> d -H -* U-A VA vO -'-\ CJ ^ CJ C^ •Pi 2 CO ca OJ CO CJ ca CM CO ^ Sh On CJ CO o o MO CO u^ >-'^ ON VO CO lA H C^ C^ o o (t> H CJ CO CO 8 ! II CJ c^ ON NjO CO e — i' Hi) O I CO I CO I ON 1 o CO i CO o o^ CO ON 'A CA o o NO H CO H ^A NO u^ o o CA O o VA CO H CO -3ON VA H NO CO H CNlCO rj
PAGE 102

95 •r! U W H tiO f ON o c•-^ o o o o o o o o o o o o o o o o o o o o •r) O ^ •H -P ^ U CD Ji Eh H O ra •H W -H H •O ON C'-^ ,-s_ CV NO -^ i^ H (M H (M C^. C^ H 00 NO o NO 00 H : — i -p rt -ri •ri ^ t<-i o 1-7-4 "& cS C*_. r-i o iH ? _o 'S h.-) C3 f^ c NO CO c^ NO 0-J NO CTn NO ^"^ VTN, o' o o NO 0^ c<~\ (r\ CNJ NO o CO NO f—l o (.Ni O ON Cv] CO c^ 0^ C^ ^' u^^ U-\ NO NO CO >> ;^ -p O H" '-^-^ r-! —S ^ }^ o C) o r-i CN> (X! o H Cvl o NO Cs! ; NO o'^ C-N oi ON vo o o o & O H o o i' -i48 9^. 0^ CN Cn QN CO CV ^ o o O iH :s> CNi 1 i c o =ii -P o "in -H H O rH .O^ H Ri O NO C^ ON c~\ s D-' s o s S r-i l-'-l O H ..^^ c\> On O & o o K s> O 0"% ON (X\ xn CO o c'-\ :— i CN! ON C\ CO H O CNl !>00 CO xn NO NO NO c~>. -1^'~\ NO oCO ON O c^ T-'"\ r^ 0-\ C--N c^ C^ -* r-i H !-i H H r-i H H NO O O 00 1— ^-^ o o o o !;NO NO u-^ -^" c^: O C\] r-i >H

PAGE 103

. 96 iii S P, H -d o "~b3 i •H ?-! CC M S i 0) 1 Pri >^ 1 >-5 -i^ -P C,H •H •H O •Ti T3 •H. •H +^ ,Q r-d s:; ;^ ^ O f^ :i u E-i r-i 1 ?.• rH O Ci5 rt Ph •H C -iJ •H W 'ri P^ i-d i=5 i M -.-^' H .r^ ci ti ri •H -H ^• t. U ^ e-i OJ o ON CM o o 0~\ A H •H O -P rf r-: 1— : O S3 iU O c-^ CN o CO o C\ o J VO -P 0) ^ T-i rH O H p.. •rS S p4 Q 03 o J-1 Pn O D-' E3 S cS O O" CM o 0.5 O o Co O IH

PAGE 104

^ 97 •" n rH \o V-\ X! c---. r-i CO ri U -cQ o O u"j — ; ^ r-'. o G Pi I o c^ vO oo r^ -Ju -+ cv OJ CM CJ C\! f>a -p 1-1 01 o •n o ^ -c;! H •H -P Xi n^ e:; ;^ SH fM H r-I o Pi ra P^ -H .C -P •H, W .H fe. nJ G CM \o o H cr [>V3 C^! c-^ (r\ c^ E-i -~'-P r— 1 -H !:; -H -H. ^ ^ O c^ MD CO o CO ^Cii H H O t:! a) K -P ft cti ?^ H O rH d -P ITS' Go C3 '•J-^ c^\ -ir c-^ o -CjVD ON G\ C^ CO MD CO CM OO o a OX o H fc-J H ?s -' "'--^ CO rH ^'^. _-r vC•-0 -^" CO VT';--^ •^ CO j^ C^ C'-^X ^'^. CO 02 CO i>CO 'J-i .;>_ S o U o o o V s o o ;:; (9 :;' o o r— I o r^' C-J CM O r— ; o C^ CM CO J o ti -I-.I i i a _i. s I s il :~1 o o '^ 'H — i O I— I jft fin O DO CO CO CN MO ON CO '-O o OC' o -o ON CO CO H. H ?-^ MD CO ON q t3 CD o o rH o c^ rf o s S H H O O^ K o <^i o o iNco o O a~) o ~o o ON CO o o o o o o CO CO o o CO >JO CO •-o i-i CM 9 O O o -4o o CO

PAGE 105

BIBLIOGRAPHT

PAGE 106

LITERATURE CITATIONS lo Public H ealth s ervjce Driiiking Water Standards. 1962, Pub* 956 USPHSj VJashingtciis D Co (1^62 j a 2, USPHS Drinking Water Standards 1962, Federal PLegister (I^r. dj, 1962). p. 2152 3, U So Public Health Serviceo Drinking Water Standards, I96I, JG-o.r. BMA, 53J935 (^g 196l) 4, Blacks, A„ Po aiid Hannah, S. A, Electrophoretic Studies of Turbidity Removal by Coagulation With Aluirinum Sulfate* Jourc AWi-IA 53:438 (Apr. 1961), 5 Water Quali ty and Treat^iient ,, AWWA, New York N. Y (2nd ed 19jo)„ p. 13I0 6 Packham, R„ F^ The Theory of the Coagulation Process A Survey of the Literature^ 2 Coagialation as a 'iater Treataient Process. ProCa of the. See o for Water Treatment and Examination il:106 (19^2") 7 D'Arcetj F. Note Relative to the Clarification of the Water of the Nile and Water in General l-Jhich Holds Earthly Substances in Suspension. (Articles from French Journals Translated by J. Griscoia). J. Franklin. In st^, Ken Series, 22:258 (I838). 8 Jeunet, Ko Co j-ionit __,S cio g 7?age Works 108:P.192-9? (Oct. 1961), 103

PAGE 107

104 14a Theriault, E^J. ajad Clark, W. Mo i'-n Experimental Study ox the Relation of H' Concentration to the Formation of Floe in Alum Solutiono ^b. H eal,th Repts , 38:181 (1923) 15^ Baylis, Jo R<, Use pf Acids With Alum in Water Purification and the Irfjportance of h"^ Ion Concentration. Jouro AVjWA, 10:365 (1923). l6 Fiiller, L. B„ On the Composition of the Precipitate From Partially All^alini zed Alum Solutions. Faho Health Repts , 38:1995 (1923) l?o Fiiller s L 3. A Study of the Effects of Anions Upon the Properties of MuiTx Floe. Fub Health Repts., 40:351 (1925). l8o J'Uller, L. B, Some Properties of Iron Compo^ands and Their Relations to Water Clarification^ Fabo. Health Rspts 40:1413 (1925). 19. Eartow, E. and Peterson, B. F. Effect of Salts on the Rate of Coagulation and the Optimum Precipitation of Al-'on Floe, Indc, Eip: CheH.s 20:51 (1928). 20. Mattsons S J. Cata-ohoresis and the Electrical Neutralization of Colloidal Material, "j. Phys.. Chem. 32:1532 (1928). 21. Black, A P.. Rics; Owen, and Bartow, Edi-rard. Formation of Floe by ALlminum' Sulfate. Ind. Sigo Cham.. 25:811 (1933)22. Bar-cow, Sd^^rd^ Black, Ao P. and Ssnsbarvj W. E. Formation of Floe by Feriuc Coagulants. Indc Elig:^ Chem 25:898 (1933); Proc. ASCS, 59:1529 (1933) <> 23. Black, A. P. Coagulation With Iron Compounds. Jour. Al'Jl'JA, 26:1?13 (193^) 24. Langeliers W. F. and Lud-c-ig, K. F. Mechanitsiri of Flocculation in the Clarification of Turbid Waters. Jour a 3MA 41:i63 (Feb, 1949). 25. Langelier^ W. F. and Lud.x-jig, E.^ F. Flocculation Phenomena in Turbid Water Clarification. Proc.. ASCB 78slfo. 118 (1952). 26. Pilinc^Jdchj J. Bo, et al. Elecxrophoretic Studies of Water Coagulation. Jo ur., mik 50:146? (Kov. 1958). 2?. Matiievic, E. et al. Detection of Metal Ion Hydrolysis by GoagCulation. III." Alumin-om. J. Phys^..... Chem ^ 6yiSz6 (I96I). 28. Pacldiam, Re F. The Coagulation Process. III. The Effect of pH on the Precipitation of Aluminura Hydroxide. Tech. Paper No.. I7 British Water Research Assn., Redhlll, Surrey, England (I96O). 29. Flackrle, S. Mechanism of Coagulation in Water Treatment. Jour, of the Sanitary I^. Div. ASC3> 8S:No. SA3, p. 3117-1 (^-^^7. I962).

PAGE 108

105 30. Stxjirffii, W, and Morgan^ u\ J, Chemical Aspects of Coagulation. Jour, .mi A 5ii.:971 (Aug I962). 31. Black, A. P. and Christman, R, F. Characteristics of Colored Surface Waters. Jour, AWA 55:753 (J^ I963). 32. Hendricks, S B. Base Exchange of Crystalline Silicates. Ind. Ehgo Chem .. 37:625-30 (19^5). 33. Marshall, C. E. and I^rinbill, C. A. The Clays as Colloidal Electrolytes. J. Phys<, Chern ,, 46:1077-90 (19^2). 31^. Helmlioltz, L. F. Imn. Physik. 703? (l8?9) Has been translated. Bocquet P. Tto Monop-raphs Qn...Eae ctrokinetics._jbe rg Research Bull,. 5Io. 33 a University of Michigan, Ann Arbor (,1951 )• 35. Mysels, K J, In.trod^Q'^-O ^ Colloid Chemistry Interscience Publishers, New York (1959). p, 321. 36. Mysels, K. J. Introductio n to Co?-loid Cheaistry Interscience Publishers, Kexv York (1959). p. 330 3?. Black, A. P. and Willems, D. C-, Slectrophoretic Studies of Coagulation for Removal of Organic Color. Jour AM'JA 53:589 (May 1961), 38. Black, A. P. and Christman, R F. Slectrophoretic Studies of Sludge Particles Produced in Lirae-Soda Softening. Jour. AWA 53:737 (J-^1961). 39. PacMiam, R. F. The Coagulation Process A Reviet-r of Soae Recent Investigations. Proc,., o f the Soco for Water Treatment and Examination 12:15 (19^3) • 40. Hannah, S. A Effects of pH and Polyelectrolyte Coagulant Aids on Coagulation of Clay Suspensions, MS, Thesis, Dept. of Chemistry, Univ. of Flordda, Gainesville, Fla. (I96O). p. 24. 41. Replaceable Bases in Soils Devoid of Carbonates. Offlci^ljfethods of Analysis of .the Association of Offic ial Agr icultural Chemists. George Banta Publishing Co., Menasha, vas. (8th ed., 1955)* pp. 39-42. 42. Le^^s, D. R. Replacement of Cations of Clay hj Ion Exchange Resins. Ind. Eng. Chem ., 45:1782-3 (1953). 43. Standard Methods for the Examination of Water and Waste--?ater. APPIA, AVMA and VI? CF, Hew York (11 th ed. I96O), p. 213 44. Cohen, J. M. Improved Jar Test Procedure. Jour. AWVJA, 49:1425 (Kov. 1957).

PAGE 109

106 45. Lvimetron Photoelectric Colorimeter Model 450 for Nessler Tubes. Photovolt Corp., 95 Madison Ave, Nex^^ York l6, K. Y. 46. Standard Methods for the Examination of Water and Waste1^^ter APHAs AWA and VJPCF, New York (11th ed. 19^0)1 p. 261, 47. Standard Methods for the Exa-aination of Wa ter and Wastet-jater APHA, AWA and ^AipCF, New York (llth ed., I96O). p. 140. 48. Standard Methods... for the Examination of Wat^r and Wastex-ater APM, AVMA and WCF, Nexj York (llth ed., 190071 p. 114. 49. Black, A. P. and Smith, A. L. Detennination of the Mobility of Colloidal Particles by Microelectrophoresis. Jour. AWWA 54:926 (Aug. 1962). 50. Briggs, D. R. A Pyrex All-Glass tlicrc electrophoresis Cell. Anal. Chem .. 12:703 (1940), 51. Tanbo, Norihito. Private communication. 52. Black, A. P., Singley, J. E. VMttle, G. P., and Ifeulding, J. S, The Stoichiometry of the Coagulation of Organic Color trxVa Ferric Sulfate. Jouro AWA In press.

PAGE 110

BIOGRAPHICAL SKETCH James Vernon Walters was bom May I3, 1933s 3-t Dublin, Georgia. In June, 1951> he was graduated from I>ablin High School. He received the degree of Bachelor of Civil Engineering in June, 1955 from the Georgia Institute of Technology. In September of that year he enrolled in the Graduate School of the Georgia Institute of Technology. He was commissioned in the Public Health Service of the United States in November, 1956. His station of duty was the Public Health Service Regional Office in Atlanta, Geox'gia. That assignment continued through January, 1959. During June, 1958, he received the degree of Master of Science, Civil Engineering, from the Georgia Institute of Technology. Mr. Walters joined the faculty of the University of Alabama as. Assistant Professor of Civil Engineering in February, 1959He continued in that position until September, I96I5 when he x-;as granted educational leave to enroll in the Graduate School of the University of Florida. From that time until the present, he has pursued his work toward the degree of Doctor of Philosopher, James Vernon Walters is married to the former Barbara Ann Daniell. They have two sons. Mr. Walters is registered in the states of Georgia and Alabama as a Professional Engineer and Land Surveyor. He is a member of the American Chemical Society, the American Society of Civil Engineers, the American Society for Engineering Education, the American VJater Works Association, the Water Pollution Conurol Federation, the ComiTiissioned Officers Association of the U. S. Public Health 107

PAGE 111

108 Service, the Alabama Public Health Association, the Alabaraa Water and Sewage Association, Tau Beta Pi 5 Signia Xi, Chi Epsilon, Alpha Chi SigHia, Phi Kappa Phi, and Kappa Kappa Psi.

PAGE 112

This dissertation ^ra.s prepared iinder the direction of the chairman of the candidate's supervisory committee and has been approved by all members of that coHnaittee. It was subinitted to the Dean of the College of Engineering and to the Graduate Coiincil, and was approved as partial falfillment of the requirements for the degree of Doctor of Philosophy, August 10, 1963 Dean, College of Engineeringf Supervisory Committee: Cw^ Dean, Graduate School