A COMPARISON OF ALUM'AND FERRIC CHLORIDE AS COAGULANTSI 1N WATER PURIFICATION Thesis for the Degree of B. S. - J. H. Pomeroy M. E. Noecker 1935 . - v \ aaaaa \wlgmwoza’a: {3,— uuvuvv .cg. ' "In"; A Comparison.of Alum and Ferric Chloride as Coagulants in Water Purification A Thesis Submitted to The Faculty of MICHIGAN STATE COLLEGE of AGRICULTURE AND APPLIED SCIENCE By » ( . ti fl" (“jet en ‘4; J. ‘ "3" t‘ J. H. Pomeroy ‘M. E. Noecker A Candidates for the Degree of Bachelor of Science June 1935 HA1; , 015.41... 11‘ ACKNOWLEDGMENT We wish to thank the Sanitary Engineering Depart- ment for the use of the laboratory and equipment, and the Bacteriology Department for use of the bacteriology laboratory. We owe Special thanks to Hr. E. F. Eldridge for his assistance during the experimental work on this thesis. PURPOSE OF PROBLEM The purpose of this thesis is to make a comparison of alum and ferric chloride as coagulants in the purification of water by coagulation, sedimentation, and filtration. INTRODUCTION Water supply and more especially, water purification has been problems of mankind since man began living in communities. Perhaps the earliest steps made in water purification took place in China and India. For centuries it has been known that the chinese put alum in tubs of water to purify it. (1). The Egyptians did the same thing. EurOpe was far behind the East in water treatment until comparatively recent years. As late as the middle of the 19th century, European and American cities were using water highly contaminated with no means of purification. About the middle of the 19th century, however, high death rates and epidemics of water borne diseases compelled recognition of a serious problem. Purification at first consisted of slow sand filtration. Filtration of the water for the London Metr0politan district was made compulsory in 1885. (l) (2). Shortly afterwards, cities in America began a movement to improve water supply. In 1866, James B. Kirkwood was sent to Europe to study water purification systems in behalf of the City of St. Louis. (2) (3). While his suggestions were not adopted at St. Louis, other American cities followed his plans in their systems. (2). In 1884 Alpheus Hyatt secured a patent for water purification involving use of a coagulant. The sane year, the first plant designed on that principal was constructed at Sumerville, New Jersey. (5). Coagulation was not an essential feature of water purification until the rapid sand filter was studied and designed by Charles Kermany and G. W. Fuller in 1895, at St. Louis, (2). Hyatts patent expired in 1901 but the rapid sand filter did not become popular fer a few years afterward. (During this time alum remained the common coagulant for both.types of filters. It was known that other coagulants existed. Ferrous sulphate was used in a few installations but in general was not very satisfactory. Dry feed was considered impossible at this time and a coagulant was required whidh would dissolve quite readily in concentrated solutions. (9). In 1910 it was discovered that ferric chloride could be prepared commercially from gaseous chlorine and scrap iron. (6). In spite of work with other coagulants alum remained the favorite. Ferric chloride, because of its expense, was not used extensively in water treatment until recent years. Enslows work with ferric chloride beginning in 1927 has started a new trend. (6). The ever increasing demand for water free from color, odor, and turbidity has led to the abandonment of a1um.in some locations where it ‘ has proven incapable of meeting special requirements or fluctuations in influent. While it would be impossible for any comparisons to be made upon coagulants used in purification of different water supplies, comparisons can be made where different coagulants have been used upon the same water. The number of cases are rare where this is true. The expense of ferric chloride naturally indicates that alum will be used where performance in plant Operation permits. Results of eXperience in water purification show that alum can be used very successfully as a coagulant where there are no wide variations which call for a sharp change in dosage. Cases have been known where, in low turbidities with soft, highly colored water, alum was not satisfactory. In some situations of that kind satisfactory results have been obtained activated carbon in addition to alum. (7). One water supply, however, having a high fluctuation in turbidity was handled by ferric chloride. (4). This incident occurred at the Westchester Water Works No. 1. The source of supply was the Mamaronick river. This plant was designed with the intent of using alum as the coagulant. It was soon discovered that the rapid fluctuations in influent properties compelled constant attention and change of dosage rate to prevent over and under dosage, since either resulted in a pin point floc. It was found that an excellent floc was formed upon use of either lime or soda ash. Their use caused a setting of color so that it could not be removed. Color removal is accomplished at a pH of 5.0 or 5.5. (4) (5) (6). Alum does not give satisfactory floc at that pH. Trouble increased.when summer brought low flows and algae trouble with so intensified a taste as to be a nuisance resulted. At this time, trials were made with ferric chloride. Satisfactory results were obtained and the use of ferric chloride was adapted. Comparisons at this plant indicated superiority of ferric chloride in low temperatures since alum required lime for floc formation at these temperatures. Because of a wider pH range, ferric chloride is better than alum where sudden changes in pH values may take place. Over dosage will not be detrimental except from the expense standpoint in the use of ferric chloride. Under dosage will fail to bring down all the color but will not pass the filter as will alum in like circumstances. Ferric chloride required a shorter time for sedimentation. Although three hours was sufficient sedimentation time with the use of alum under flood conditions, the filters were over loaded. No increase of load on the filters was noticed after the change to ferric chloride. Before changing to ferric chloride at this plant, trouble was encountered with Short filter runs and air binding due to heavy filter mats. As would be expected.from the above statement, alum left a troublesome gelatinous coating on the sand grains. Use of ferric chloride remedied this trouble. From the above paragraph one might get the impression that ferric chloride is considered superior to alum as a coagulant. Such is not always the case. There are conditions which make ferric chloride desirable but most plants can use alum more economically. In general both coagulants operate on the same principle. Both form insoluble hydroxides which settle out of solution in form of a floc. The chemical reactions which.result in that floc are: A12 ($04)5 . 3 Ca (H005)2 - 2 Al (0H)3 + 3 oaso4 + 6 008 and PeCl + 3 Ca (H005)2 - Fe (0H)3 + 3 CaClz + 3 CO 3 2 The Ca (3005):; is natural alkalinity already in water. Where it is not present soda ash or lime must be added. The alum requires less alkalinity than does ferric chloride for the sane dosage rate. (11). The floc formed by the Nabove reactions is the insoluble Al, (OH)5 and Fe (0H)5. Since aluminum hydroxide is amphoteric, the pH value:must be such that the hydroxide form results from.the reaction. Too wide a variation in pH will result in no floc at all or a fine colloidal floc not suitable for coagulation. Ferric chloride, as stated before, is not so sensitive to pH variations. Any fine comparison made upon coagulants must be made under the same conditions. Any particular supply would require a study to determine the most efficient coagulant. An attenpt has been made in this experimental work, to make certain general comparisons which.might apply to any source of water. The source was the Red Cedar river. No doubt the experiments as outlined in the following pages would apply closely to any like raw water supply. 7 DESCRIPTION OF PLANT The plant consists of five units, as follows: a flow control box which can be used to proportion river and tap water, mixing chamber, coagulation basin, primary settling, and sedimentation tank, and finally two rapid sand filter units with one rate control trap. Since we did not attempt to soften the water, the primary settling tank was cut out of the system. The original supply of raw water consisted of river water piped under a head from.the M.S.C. power plant and East Lansing tap water. A gravity head moves the flow throughout the system, and in all of the tanks except the sedimentation the influent enters at the bottom and passes out at the surface of each unit so as to obtain as nearly uniform flows as possible. The chemicals and added turbidity were supplied from five gallon bottles placed over the flow control box. An adjustable overflow pipe in the flow control box establishes a constant head and flow of water to the mixing chamber. The power fur the mechanical mixing and coagulation is supplied by a small electric motor. Theoretically, the retention periods in the two basins is fifteen and sixty minutes respectively. The mixture of water and floc next bypasses the primary settling tank to the top of the sedimentation basin. It was not possible to follow the general design of the plant and enter the water I at the base of the tank because of sludge formation. So in.this tank "around-the-end" baffles were installed to give a uniform flow. This tank retention period is approximately four hours. Tracing the flow on from this tank, next in consideration are the twin filter units. Each rapid sand filter consists of an upright six inch glass pipe containing fifteen inches of gravel and about twenty inches of sand. They are underdrained by a one- half inch orifice which leads to the rate control box. This box has an adjustable float Which restricts the flow through the filter and bypasses the excess from the sedimentation tank immediately to the waste drain. Although the theoretical retention of the entire treatment is about six hours, we have reason to believe by tracing the floc deposition across the effluent pipe in sedimentation, that the actual time is considerably less. Kk<< 4 1 l: o. . u d! ,‘; .' a fi ' ‘ ‘ ’ ," . (;0 ~ ‘ .V, | .' '1. . IAIR‘: 7‘. ' ‘ | ' ‘ l . ,' L ‘ A “ ‘ \ r cr“ '