LI E> RARY OF THE UNIVERSITY Of ILLINOIS l£65c no •f-9 — ■—? :: The person charging this material is re- sponsible for its return on or before the Latest Date stamped below. Theft, mutilation, and underlining of books are reasons for disciplinary action and may result in dismissal from the University. University of Illinois Library SEP 6 1988 oc L161— O-1096 Digitized by the Internet Archive in 2013 http://archive.org/details/fundamentalfacto09enge \t+ (f one** uwM " CIVIL ENGINEERING STUDIES SANITARY ENGINEERING SERIES NO. 9 628 77o.9 ENGINEERING LIBRARY UNIVERSITY OF ILLINOIS UN URBAKA, ILLINOIS FUNDAMENTAL FACTORS IN TREATMENT OF IRON BEARING WATERS By R. S. ENGELBRECHT, G. E. MARGRAVE, J. M. LONGLEY, L R. ROBINSON, and J. G. WEART PROGRESS REPORT FOR THE PERIOD FEBRUARY 1, 1959 TO DECEMBER 15, 1960 for the NATIONAL INSTITUTES OF HEALTH U. S. PUBLIC HEALTH SERVICE RESEARCH PROJECT RG-6436 DEPARTMENT OF CIVIL ENGINEERING UNIVERSITY OF ILLINOIS URBANA, ILLINOIS and ILLINOIS DEPARTMENT OF PUBLIC HEALTH SPRINGFIELD, ILLINOIS First Progress Report on Project RG-6436 February 1959 to December I960 FUNDAMENTAL FACTORS IN TREATMENT OF IRON BEARING WATERS R. S. Engelbrecht, Director G. E. Margrave, Asst. Director Department of Civil Engineering University of Illinois Urbana, I 1 1 i noi s and Illinois Department of Public Health Spri ngf ield, I 1 1 inoi s January 10, 1961 ivision of General Medical Sciences National Institutes of Health b ' ENGINEERING LIBRARY vn o. 9 SUMMARY A pilot-scale treatment plant, consisting of a storage reservoir, diffused air aerator, react ion- sedimentation unit, and rapid sand filter, has been used to investigate the fundamental factors governing the treatment of iron-bearing waters. A satisfactory procedure for synthesizing waters containing known components has been developed, including a method of maintaining iron in the reduced state to simulate natural waters. Twenty-four pilot plant studies have been completed. Waters studied included 10 synthetic waters, 8 natural waters studied in the laboratory; and 6 natural waters studied in the field. The waters selected for investigation varied with respect to concentrations of ammonia, sulfate, hardness, alkalinity, and iron. The overall removal of iron by the pilot-plant, regardless of water composition, has been satisfactory with one exception. Bench-scale studies have indicated some difficulty in iron removal by usual processes when a dicarboxylic acid was present. In the pilot plant, the aerator unit, which consistently increased the DO to about 80 per cent of saturation, was found to remove approximately 25 per cent of the iron, the reaction-sedimentation unit about 3 per cent, while the filter removed from 90 to 100 per cent of the iron applied to it. A greater rate of oxidation of iron appeared to occur in the filter as opposed to that in the aerator and reaction-sedimentation unit. The detention time in the reaction-sedimentation unit was found to have no effect on overall iron removal. High sulfate concentra- tions apparently improved removal of iron by the reaction-sedimentation unit. Con- centrations of alkalinity when in excess of the hardness of a water enhanced ferrous iron oxidation. Little correlation to date has been found between COD, organic nitrogen and iron removal efficiency. It should be noted, however, that COD values as high as 52 mg/1 have been found in natural waters. 0RP measurements made to date have shown only that raw waters with a negative value change to a positive value as a result of treatment. Table of Contents Page I Introduction A. Nature and Importance of the Problem 1 B. Iron Removal Methods 2 I I Experimental Equipment and Procedures A. Pilot Plant Equipment 3 B. Analytical Methods 14 C. Water and Equipment Preparation 14 D. Operation of Pilot Plant 16 II Experimental Results A. Preparation of Synthetic Ground Water 17 B. Characterization of the Waters Tested 20 C. The Aeration Unit and Dissolved Oxygen Study 21 D. 1 ron Removal 25 E. Carbon Dioxide Reduction and pH Considerations 30 Fo Effect of Variations in Sedimentation Time 34 G. Miscellaneous Determinations 34 IV Discussion of Results A. Preparation of Synthetic Ground Water 39 Bo Aeration Unit and Dissolved Oxygen Study 39 C. I ron Removal 41 D. Carbon Dioxide Reduction and pH Considerations 44 E. Miscellaneous Determinations 45 V Conclusions A. Design and Construction of Equipment 47 B. Special Techniques for Study 47 C- Preparation of Synthetic Ground Water 47 Do Results of Testing Pilot Plant on Various Waters 48 VI Publications, Staff and Foreign Travel A. Publications 51 Bo Staff 51 Co Foreign Travel 5' Fundamental Factors in Treatment of Iron-Bearing Waters R. S. Engelbrecht, G. E. Margrave J. M. Longley, L. R. Robinson, Jr. and J. G. Weart I INTRODUCTION A. Nature and Importance of the Problem The undesirable presence of dissolved iron in potable or industrial process water supplies has been recognized for a very long time. The first treatment of a water supply for the removal of iron was made in Germany in 1868. Since that time, much has been learned regarding the nature of the existing forms of iron in raw water, and several satisfactory treatment techniques have been developed. As far as is known, however, the fundamental principles of iron removal and their relationships to the small concentrations of seldom evaluated, but important, constituents of a water are not understood to a degree which per- mits the design of a facility with any assurance of satisfactory iron removal. In fact, the variations in waters which seem to make the difference between a successful or an unsuccessful application of a treatment process are not usually included in a routine mineral analysis of a water. The result is that most iron removal plants are designed on the basis of an existing plant which is success- fully treating water of a similar mineral characteristic. Proposed treatment methods are only occasionally tried on a pilot plant scale. An extensive litera- ture survey made in association with the current research project has confirmed these observations. The problem of iron in water supplies is world wide, but varies con- siderably in magnitude. In the United States, and particularly in the central states, where deep wells are the source of water for many communities, the problem Is especially acute. For example, 70 per cent of the public water supplies in Illinois contain iron in excess of the U. S. Public Health Service limit of 0,3 mg/1 . Current data indicates that one-third of these plants having iron removal facilities are not reducing the iron content of the finished water to a satisfactory level. B.. Iron Removal Methods The removal of iron from water can be accomplished by the following methods: a. Conversion of soluble iron compounds to insoluble forms, followed by removal of the insoluble material, bo Removal by ion exchange. c. Conversion of soluble iron compounds to insoluble forms by oxidation utilizing an ion exchange resin especially treated to possess oxidized ions at the exchange sites. (This is the so-called Manganese Zeolite.) do Removal by the conventional lime-soda ash softening process. e. Combinations of the above. The method being investigated to date in this study is that of oxidizing the soluble iron forms in the water to insoluble forms by aeration and removing these insoluble forms from the water by sedimentation and rapid sand filtration. Oxidation of the ferrous forms present in natural waters is usually accomplished by aeration; however, in the case of waters containing organic matter which form interfering colloids with the iron, all the iron is not readily oxidizer by aeration. To date, the nature of this interference is not fully understood. 3 II EXPERIMENTAL EQUIPMENT AND PROCEDURES A. Pilot Plant Equipment The pilot plant equipment used in this study was designed as an inte- grated series of unit processes. This approach was taken to provide flexibility in operation and to facilitate modification of the treatment processes when applied to waters of widely varying mineral character. Rubber, glass and transparent rigid plastics ("Lucite" and "Plexiglas" trade names) were used throughout the apparatus to insure that results would not be influenced by the pick up of iron from the equipment. The water being studied does not contact any metal during the treatment process, and care was taken to avoid the use of ferrous accessories where rust or corrosion products could fall into any water during operation or sampling. The entire apparatus was made so that it could be easily disassembled and transported for field studies. Figure 1 is a photograph of the pilot plant set up in the laboratory. The units built for the initial study of removal of iron by aeration, sedimentation, and filtration were a) a feed tank and constant head flow regulator, b) aerator, c) reaction-sedimentation unit, and d) a rapid sand filter. Figure 2 is a line drawing of the assembled units. For field studies, an overflow stand pipe to provide a constant head was substituted for the feed tank and constant head device. Necessary stands, supports and accessories position the equipment for a gravity flow through the plant from the elevated feed tank or the overflow stand pipe. No pumping was necessary for laboratory studies; for field studies, the municipal utility's well pump was used to feed the overflow stand pipe. Figure 3 is a photograph of the unit in the field. The feed tank and constant head flow regulator are diagramed in Figure k. The feed tank also served as a mixing vessel when synthetic waters were used. A device with six sintered glass diffusers connected to a manifold and com- mon supply pipe served to introduce nitrogen for oxygen stripping, CO- to facili- tate solution of carbonates, and for mixing. The aerator is a multiple pass, diffused air unit with the water intro- duced under gravity pressure. The water leaves via an overflow weir and col- lection channel. The unit is diagramed in Figure 5° Air is introduced through a grade M sintered glass diffuser 30 mm in diameter. The combination of the velo- city of the water leaving the jet and the lighter density of the water and air bubble mixture causes a vertical current within the small central tube which FIGURE 1 The Pilot Plant as Set Up for Laboratory Studies ELEVATION INCHES 191 _ MAX 53.5 135 120.5 HEAD TANK TABLE a SUPPORTS CONSTANT HEAD DEVICE TOP OF SECOND FLOOR A RAW SAMPLE TAP B AERATED SAMPLE TAP C SETTLED SAMPLE TAP D FINISHED SAMPLE TAP AERATOR REACTION TANK FIGURE 2 LINE DRAWING OF PILOT PLANT FIGURE 3 The Pilot Plant as Set Up at Cisco, Illinois The arrangement differs from the laboratory set up only by the use of the overflow stand pipe in place of the constant heac supply unit. PLASTIC COVER CO OR N„ GAS 2 CYLINDER 1 a ■ WOOD /PLASTIC SHEET M FLOATING MEMBRANE 55 GALLON HEAD TANK PLASTIC LINED / PLASTIC TEE M r^ IX) PLASTIC RUBBER PLUGS GAS DIFFUSER DEVICE TABLE L J I CONSTANT HEAD REGULATOR ■TO PLANT SUPPORT APPROX. SCALE 1/8"= I INCH FIGURE 4 FEED TANK 8 CONSTANT HEAD FLOW REGULATOR TRIANGULAR SPACER k SHALL SPACERS 8 CEMENT BOTTOM IN LOOSE FIT H CEMENT SPACERS TO TUBES HERE- CENTRAL TUBE 2"x 1/8" WALL x 16" BARRIER TUBE 3"x 1/8" WALL X 18" SHELL TUBE a 6" x I A" WALL x 16" DRILL 12 HOLES 1/2" DIA. IN TWO ROWS EQUALLY SPACED AROUND DRILL 6 HOLES 1" DIA. ONE ROW EQUALLY SPACED AROUND LOOSE DISC TO LOCATE TUBES BOTTOM - 1/2" PLASTIC CEMENTED I TO TANK* -CEMENT OUTLET IN NOZZLE RUBBER SEAL DIFFUSER k MM GLASS RUBBER , PLUG \ GLASS TEE AIR IN— 1^ ~A RUBBER SEAL ■ Hi WATER IN — » £ DETAIL OF DIFFUSER 2 x SCALE FIGURE 5 THE AERATOR 9 houses the inlet nozzle and diffuser. The water spills over the top of the small central tube and passes down the annular space between the central tube and the barrier tube, since the barrier tube extends above the central tube. Holes at the bottom of both tubes permit a division of the flow; thus some water is recircu- lated up the central tube, and the balance moves into the space between the shell and the barrier tube. Water diverted outward eventually passes over the overflow weir into the collection channel and on to further treatment- Figure 6 shows the aerator in action; the profuse bubble pattern is easily seen, and the air-water "froth" at the top of the central tube can be seen. The walls of the central tube and barrier tube are barely discernable. The sizing and volume of the aerator were largely a matter of con- venience. With a water flow of h 16 cc/min (which loads the filter at 2 gallons per minute per square foot) the detention time in the aerator is about \k minutes; this is quite in agreement with the contact times suggested in the literature for aeration devices. The compressed air for aeration is supplied at 125 psig either by the central laboratory compressor, or by a portable compressor unit. The air is passed through a trap with glass wool to remove dust, oil, and moisture. A flow regulator is used to keep the flow at a chosen rate. A Rotameter is used to measure the airflow. A needle valve following the Rotameter provides additional control of the air flow rate, and a mercury manometer is connected to measure the pressure after the needle valve* This system provided a sensitive control of the air flow. The tank for the combined purpose of providing time for completing the chemical reaction and for sedimentation was designed with a length-to-width ratio of 2 to 1, so that the path any particle traversed from inlet to outlet could be doubled or halved, depending on the orientation of the tank with respect to inlet and outlet connections. The height was made 18 inches so that a 1„5 hour detention at k\6 cc/min flow could be studied, since 1.5 to 2 hours represents the maximum usually encountered in practice for the reaction-sedimentation time. The line connecting the aerator to the reaction tank was constructed so that when chlorine was introduced at the aerator end it would be thoroughly mixed by the time the water entered the reaction tank. This line has a diversion tap for sampling* To facilitate sampling for dissolved oxygen analysis, the diversion tube was constricted to make it flow full without aspirating air. The line between the aerator and reaction tank does not flow full, and the diversion device imposes a turbulence which improves mixing and aeration. 10 FIGURE 6 The Aerator in Operation Attention is directed to the concentric tubes, visible as faint shadings, and to the holes which permit the division of the flow. The outlet of the collection channel is seen in the upper r i ght . 11 Water leaves the reaction tank via the overflow weir and a plastic tube connected to a plastic needle valve immediately ahead of the sand filter inlet. A sampling tap was provided between the weir and the needle valve. The needle valve can be adjusted to maintain the desired water level in the reaction tank. Figure 7 is a diagram of the reaction-sedimentation tank and the inlet-outlet devi ce. The filter was constructed as a conventional rapid sand filter using a three inch diameter transparent plastic tube. The filter medium has been a carefully screened sharp grain sand. In the laboratory preparation, the random chosen sand was screened into several fractions. The results of the screening were plotted on log-probability paper. Sieve fractions were selected to give a filter sand having an effective grain size of 0.^5 mm and a uniformity coeffi- cient of 1.^. A large supply of this sand was prepared, so that the filter could be replenished, as required, with the original mixture. This was done to establish the filter media as nearly constant as possible. The sand bed was held as near as possible at 30 inches, after back-washing. The filter was pro- vided with a backwash connection at the bottom. Backwashing was usually done with laboratory tap water, and was manually controlled to make a bed expansion of about 15 inches (50 per cent). This amount of expansion provided a good scour action, visible through the filter walls. Backwash was continued until the effluent was clear. The filter column was fitted with a series of sample tubes positioned at various depths and spacing below the sand surface. The sample tubes were 8 mm glass tubes fitted with sintered glass cylinders. Each sintered glass cylinder was centered on the axis of the filter column. The filter column and a detail of the tap locations is shown in Figure 8. For field studies, an overflow stand pipe with a diversion tee in the bottom was built. The stand pipe is adjustable in height, and maintains a constant head on the supply line, with flows up to 7-0 GPM. Waste water is carried away by a length of fire hose. The diversion tee at the base was faced so that, the incoming flow was split; the desired flow for the treatment plant being taken off, and the balance going Into the stand pipe, up and over the top weir. It was felt that the water taken off for the treatment plant would be more representative of the true raw water than it would be if it were taken out of the stand pipe itself. Also, possible aeration-oxidation was avoided by this connec- tion. This stand pipe, with the overflow weir and waste water hose is shown in Figure 3« 12 ADJUSTABLE "fr BRASS BOLT s OUTLET DEVICE LUCITE TANK 8" x 16" x 18" HIGH INSIDE DIMS. 1/2" WALLS -SZ- 3" GLASS FUNNEL TYGON TUBE SECTION A-A RUBBER PLUG ' ■- ■■ ■■'- ■ .■/////,■////. . , /.//.■ s //////// , - /- . ■ ■ ,. . GLASS <-TEE SECTION B-B FULL-SCALE ALL PARTS "LUCITE" UNLESS NOTED APPROX. SCALE 1/8" = 1 INCH 1 x 1/8" WALL 4 \s I HOLES 1/8" DIA. REACTION -SEDIMENTATION TANK INLET DEVICE DETAIL FIGURE 7 REACTION - SEDIMENTATION TANK INLET AND OUTLET DEVICES 13 SUPPORT RODS (OVERFLOW) I (INLET) 2 3 4 6 SAND DATA EFFECTIVE DIA.-0.45 m U.C.-1.4 TUBE IS 3" x 1A" WALL LUCITE (OUTLET) DATU M RUBBER PLUG _SZ_ \ /-— SA £ 34 TOP OF SAND 1/8" 8 — 9 ANGULAR LOCATION OF HOLES HOLE SCHEDULE - FRONT X - REAR HOLE ELEV. ANGULAR DIA. USE NPt INCHES LOCATION INCHES — _ 1 49-1 A C 3/4 OVERFLOW 2 39 E 5/8 INLET 3 34-5/16 D *3/8 MANOMETER k 33-11/16 F 5/8 SAMPLE 5 33-3/8 B 5/8 SAMPLE 6 32-1/8 D 5/8 SAMPLE 7 27-3/8 F 5/8 SAMPLE 8 19-1/2 C 5/8 THERMO. 9 15-7/8 B 5/8 SAMPLE 10 2-1/2 D *3/8 MANOMETER 11 1-1/2 C *3/8 SAMPLE 12 1-1/2 A *l/2 OUTLET 13 1-1/2 E *3/8 BACKWASH *PLASTIC TUBE CEMENTED IN HOLE APPROX. SCALE: 1/8" - 1 INCH ALL LUCITE EXCEPT AS NOTED ELEV. 4-1/8" BOTTOM OF SAND GRAVEL UNDER DRAIN 1/8-1 A MESH SUPPORTED ON PERFORATED PLASTIC DISC PLASTIC RING & DISC TO RETAIN PLUG FIGURE 8 SAND FILTER DETAILS 14 B. Analytical Methods Table 1 is a summary of the analytical methods used during this study. C. Water and Equipment Preparation The synthetic waters were prepared by adding carefully weighed amounts of dry chemical to a measured volume of deoxygenated distilled watei — usually 220 liters or 5^ gallons, the capacity of the supply tank. Care was taken in filling the tank so as to avoid turbulence and consequent aeration as much as possi ble. The solubility of calcium and magnesium carbonate in distilled water is low, so carbon dioxide gas was bubbled through the water to make carbonic acid, which in turn reacted with the carbonate to make soluble bi carbonates. The gas was introduced through six glass diffusers connected to a manifold which was held submerged to the bottom of the tank by the rigid plastic supply pipe. The gas was bubbled in until all the solid chemical was dt ssolved--usual ly a period of five to eight hours. The bubbles created some mixing, but occasional mechanical agitation was beneficial: After the carbonate compounds were dis- solved, the water was stripped of dissolved oxygen by bubbling compressed high purity dry nitrogen through the same diffuser device. When dissolved oxygen tests showed less than 0.5 mg/1 dissolved oxygen present, the iron compound, usually ferrous chloride, was added. Natural well waters were handled as follows. On the afternoon of the day before a scheduled experiment, 60 gallons of the water to be used were collected in glass carboys. At the time of collection, the collecting equip- ment was thoroughly flushed until it was felt that representative water was being pumped. The carboys were completely filled with a minimum of turbulence and aeration, and capped with aluminum foil. The water was analyzed in the field for dissolved oxygen; temperature, pH, and ORP measurements made; and samples for iron analysis were taken and fixed with acid for later determination in the laboratory. Sufficient sample was also collected to make a complete mineral analysis of the raw water. The filled carboys were transported to the third floor of the labora- tory and, on the day of the experiment, carefully syphoned into the feed tank. The surface of the water was protected from aeration by a floating plastic membrane. No gas of any kind was Introduced. For field studies, the equipment was disassembled and transported to a field site. The water was supplied through the overflow stand pipe device; no special handling was required. 15 -* TJ 1- CD O 4J 2 „ l_ TJ CD c !_ TJ +J •M 1_ CD ^— s <---~ CD •— ro — +J ™— — ■M 4-> i_ CD rO — — ca CD > . — r ~~ !_ i_ CD TJ C Q- O 3 ro l_ CD +j 05 O CD O o > > >> TJ +j <+- •— 1 — u- J- l_ TJ ro C CD O O CD J_ E i_ TJ Z3 O CO +J ro O CD '»-»' s_ - j*. +J TJ E 4- on c (V"\ ro c O 3 TJ CD T5 CD CD x i_ c X ro TJ CD tj TJ 1/1 in X c CD CD •> 00 CD 3 o CD 3 3 , • — 4-1 CD 4-> on TJ X m CD CD O C TJ CD +J +J 3 l_ s_ > TJ E s_ CD TJ c . * O c CD O CD •— 4-» E CD ro CD O - ■M +-» . * 4-> i_ •— c 3 •M O !_ E r— CD CD !_ O ■M Z O 1_ O i_ c Q. E (D r— O O CD CD .— ._. 4-> CD CD ro O +J fO •— •— +J s_ • — 4-1 CD +-> I/) *— +J O 4- o OJ l_ TJ TJ CD CD ■— ■_ ro c co CD .— O «- a> C C E 4-» O C C O _ ~ CD E 4- ■M c .— ._ O CD O O O .— c O ._ C CD -t-j 2 E 4- 3 4-> C CD O C CD •— ■— O CD -d-E .— CD c O CDTJ (0 CD x CD 4-> < TJ ™ x • — CO +J o 4-> •— »—> O- CD fD O ♦— Q. 00 E ro • — (/) O O !_ 1 — JMcrt SI c O O s_ >* i_ L. l_ <4- ~— *T* ro ro E i_ CD E .— 3 ~— CD CD +J — CD E 3 4-1 1— O M- O ro E E c ■M X CD' TJ _^ • — , O +J TJ — i_ 4-> E E CD in 4J E .— O TJ CD +J 1 — TJ a) ro O O Q. s_ • — •— (0 N CD Q. • — CD O 2: O O CO — O 2 co < CO CO CO _l CO S XI ro co o o 1- LU < o >• < ■z. < ti- ro >- < CO in i/)I 4-> COl 3 — [ o o 1 TJ CD E 3 in c o o O O o o ro — cv! ro ro CD a£Elo 00 oou. CD in TJ o -C ■M CD (D or TJ O CD SZ 4-1 4-j ro CD E E O 00 O v. xi CD JZ O — O — LU 2: Q i/> in 1/1 CD CD CD >- >- > CD c c c _c o o +-> — ._ C 4-1 4-1 CO ro ro c s_ l_ CD 4-> +J x: — ._ Q. I- I- • O < < x S — 1 — -*-* Q Q 1- LHUJO CD CD CD >^ >- > ZOO. TJ CD »- CD i+- TJ ,- O C TJ 1- O O 4-> E O 4J CD to — - — — X !- UJ — ro CD .— -o — in — _^ id CD c ro S?5 CO 4-> 1/1 CD >- t/1 CO CO CD CD CD en TJ en .- o — -3" O CO OO O ro CM m co O O ZC i_ CD s_ l_ CD c CD CD 4J — 4-) 4-> CD — CD CD E O E E O L- O — CD 4-> X O E 1- 3 O 4-> — 1_ 4-> — X C — 01- cd — co a. co CD — 4-> x h- O c E i_ E O C 4-1 X — CD — >» •— CO U Q. CO E TJ L. l_ — CD I O ._ ._ 3 +j >- ao > X O O X CO X I ro i_ s_ a) +-> 4-1 1-3 CD — CD 1- 4-J ca h~ 2: ixi 2; o a. ui (/1 i/ii/i CD CD CD CD O >- >- >- >- Z 1 T) l_ (0 X E Zl O ^— ro I/) in TJ CD C C ro TJ !_ ^~ CD fD X O en c on CD to CD 3 TJ c 4-> CD CD s_ X CD 4-> TJ •— CD • — TJ O in LU c c CD O X 1_ 4-> a> O >4- in f— 14- in .— CD ~ c ID CD 4-) in ro on o CM O O ro TJ O O CD ro c < CD ro O X Q. to !_ o E O ro o in CD 1 >- i TJ c i <0 CD CT CO 2 CD CO 4-1 O m c\ E X LU CD -CI .- CD C TJ — 1- to O J* — Q — x o < 00 on on on cd on c CD TJ c u TJ to l- x ro c O s_ E — 3 to o fO 4-> — 4-1 O ro o h- o K- c CD ai o c c CD CD D1 CO o >- 1- X 4-1 O ■Z TJ CD O > 'c o ro on CD on i_ — x o o a. CD CD 3 4-> TJ CO — M- m 1 — CD 3 CD TJ CD 3 CO 3 4-> C 10 >- CD 1_ 4-> CO CD — O a-o l E — TJ CD O >- I— < X C o 4-> o 3 TJ C CD — O a: to i_ 1 — - c +j O c .- CD 4-J +J ro O 3 o I- T> O. X o CD TJ *>< O c O X 1_ CO o on on CD c TJ s_ ro X E TJ 3 1_ — fO in TJ CD C c ro CO 4-> to CO 16 D. Operation of the Pilot Plant Each experiment was started by filling, in turn, the aerator, the reaction-sedimentation tank, and the sand filter. Samples for water analysis were taken after about three pore volumes of test water had passed through the sand filter; this insured that the distilled water (kept in the filter between runs) was completely displaced. Air was applied at a rate of about 1 600 cc/min during this initial period, as well as during the course of an experiment. After the plant was filled and operating at equilibrium, the hourly sampling to evaluate the operation was started. The sampling procedure varied with various experiments, but generally the following basic samples were taken: 1. Raw water entering aerator: a. Alkal i ni ty b. Dissolved oxygen c. Iron—always total, and occasionally ferrous d. pH e. Temperature 2. Aerated water at a point just beyond the aerator collection channel \ a. Dissolved oxygen b. Iron — always total; occasionally ferrous c. pH d. Temperature, from top of aerator 3. Settled water, at a point between the weir of the reaction tank and the filter inlet regulating valve: a. iron—always total; occasionally ferrous b. pH Co Temperature, at the weir k. Fi 1 tered water: a. Iron— always total, occasionally ferrous b. Alkalinity c. Dissolved oxygen, occasionally d. pH e. Temperature 5. Temperature at the mid point of the filter. As soon as possible after an experiment, the equipment was washed up with hydrochloric acid (except the sand filter) to remove iron and CaC0~ deposits. The filter was backwashed vigorously, then left filled with distilled water. 17 I I 1 EXPERIMENTAL RESULTS A. Preparation of Synthetic Ground Water Nine experimental synthetic ground waters were successfully prepared and used with the experimental pilot plant in the laboratory. Various combi- nations of nitrogen stripping, carbonation, and chemical additions were tried in order to find the most economical combination (in terms of gas consumption) and the most "satisfactory" iron bearing water in terms of stability of the ferrous iron. This stability was evaluated by the length of time between the addition of a ferrous compound and the observation of a yellow or red-brown color typical of ferric iron in the water. The choice of chemicals to provide hardness and alkalinity in desired ratios was based on the selection of the most soluble compound available, which would not introduce an undesfred anion. The iron compound was selected to pro- vide ferrous ions and a desirable anion; usually ferrous chloride was used. Table 2 is a summary of the procedure and sequence of stripping, carbonating, chemical addition, final character, and the observed time until the ferric ion was visually noted . Since natural well waters are more or less stable in their ferrous- ferric ratio; e.g., the yellow color is either present immediately and remains quite uniform, or it does not appear for quite some time, the evaluation of the synthetic waters was made in terms of time of appearance of yellow color. The sequence and duration of applying the C0„ gas was determined by the need for carbonation to convert insoluble carbonates to soluble bicarbo- nates. The N~ was used for stripping dissolved oxygen and for mixing chemicals, and was applied as needed for these purposes. As Table 2 shows, C0„ also turned out to be effective as a stripping gas, so N„ stripping prior to adding carbo- nates and sulfates was discontinued. Table 2 also indicates that waters with a hardness approximately equal to, or greater than, the alkalinity maintain a ferrous stability longer than those with excess alkalinity. The dissolved oxygen at the time of addition of the ferrous compound was obviously important, since as little as 0.5 mg/1 DO could oxidize all of the 3 mg/1 of ferrous iron added. So as to prevent the immediate oxidation of fer- rous iron, it was found that the 00 concentration should not exceed 0.10 to 0.15 mg/1 following stripping with either N or CO,.. However, there are other com- pounds which apparently can also influence this oxidation rate. The presence of free C0_ in the water when the iron compound was added seemed to show a direct 18 1-o o c a LU 00] «M [2 I LA o (A I CM O CM O LA O O LA O o o o z CM CM CTV O — QJ Z o CO o a. o o oi uj < 3 cm z: :d o — a: -Q o H !„;.« X h- z >• CO o < CO o z o v£> E CD co "o, -C T3 o < CD CD CD CD (D o s o 3: o s o o z CA fA CA fA fA ?A .J" fA CA ooooooooo OOOOOOCOOO O) dj CD ft) CD CD CDCD CD =3" CA o o CD o «M "O o c o| [LU 00 CM I Z LA O LA CM o CU Q H > ■*-> z Li. CM o - £ 19 LA VD CM cm r- — CM CM O — o LA 00 LA O LA -4- LA LA O — \ 00 00 CM t^. i — vO CA CO -3" o aj cr CM *— O E 00 -3" CM 00 1^. ca LA 00 00 X • Q. LA vO r-^ v£> vO vO t^. vO O i/) {/) CD — C \ LA O o O O O o LA LA T3 cr 00 LA vO LA O O CA O CA L- E CM CM CM CM CA CA CM CA CM 03 DC >- +-» c — — \ LA O LA LA O CA vO LA O •— cr CM ~— J" -d- CD CD CD a) o T3 c C j* -* -* .* .* c OJ • » — • • — • c C C c c 3 3 fA Csi (/) u a c- a ro 03 03 03 03 l_ C r— O s_ 3 X O 3 X •— O x o a) z z o J- Q. 1- D_ L_ Q. l_ a. 1_ Q. i- a L. CL CD •— z i_ o 3 X 3 X O 3 X 3 X 3 X 3 X 3 X O CD X z Q CD O CD Z Q O O co 3 <+- «+- -o O o •— c o »— ■a -a UJ o o o (D c CD o CM z u x: ■o CM CM o o z fA o CM (A o 03 +J C 03 E 03 CO J3 J3 3 jQ -o CD O CA E o U < z z z z z Z z z z Z O c CD -a 4-< c CO 03 s u. u. o o o OH < 00 C3 a) 2 i_ > c CO 0) \- 03 03 03 3 CO CO o CM CA J- LA vO 1^ CO CA O > — z — ~r> < 20 correlation with the stability of the ferrous iron. It appeared that at least 30 mg/1 free C0_ was necessary for a one hour stability, while 100 mg/1 or more C0« extended the stability to several hours. The pH of the water was observed to be probably the most precise indi- cator of stability. The experimental waters showed a linear correlation between pH and stability, Figure 9. Waters for which the pH at the time of iron addition was not accurately known have been eliminated. Data to extend the curve of Figure 9 below pH 5.8 are not available; however, it is expected that the stability would become asymptotic to the time axis at some lower pH. Ferrous iron standard solutions, at low pH, keep indefinitely. The observed pH-stability is in agreement with the C0„-stabi 1 i ty obser- vations; the high C0« concentrations create the lower pH values, and these lower pH waters were found to be the most stable with respect to ferrous iron. The entire study of synthetic waters was based on a system containing no organic matter and, thus, no organic complex of the iron forms. This is not truly representative of the surface waters which might be encountered, but it is reasonable for many ground waters. As has been pointed out before, the organic complexes hold the iron in solution for an indefinite time, and are quite indepen- dent of pH and C0_ conditions. B. Characterization of the Waters Tested The 9 synthetic water and 14 natural water experiments performed embrace nine "characteristic" waters or groups of waters with very similar mineral con- stituents. Although the experiments were performed on the 9 synthetic waters first, followed by 8 laboratory and 6 field studies of natural waters, it seems more desirable to discuss the results of the experiments arranged according to mineral characteristics. The following outline is used in correlating waters in terms of mineral characteristics. Water Characteristics Water Designation I Nitrogen and Sulfate Variable: Iron 3~6 mg/1 A. High Sulfate 50-75 mg/1 1 . Ni trogen mg/1 A 2. Nitrogen 4-6 mg/1 B B. Low Sulfate 0-2 mg/1 1 . Ni trogen mg/1 C 2. Nitrogen 4-6 mg/1 D 3. Nitrogen 14-16 mg/1 E 21 II Non-carbonate Hardiness Present: Iron 3~6 mg/1 F III Reducing Agent Present, Sulfite added: Iron 3 - 6 mg/1; G Nitrogen, mg/1; Sulfate, 50-75 mg/1; chloride, 0-15 mg/1. IV Chloride High: Iron 3~6 mg/1; Nitrogen, mg/1; H Sulfate, 0-2 mg/1; Chloride, 50-60 mg/1. V Iron High: iron, 12-17 mg/1; Nitrogen, 14-16 mg/1; J Sulfate 0-2 mg/1; Chloride, 0-15 mg/1. Hardness and alkalinity varied, but the variation is not significant in the results. The various waters, their source, and their characteristics are sum- marized in Table 3- The letter and subscript nomenclature will be kept uniform throughout the presentation of results and the discussions which follow* C. The Aeration Unit and Dissolved Oxygen Study The primary function of the multiple pass aerator was to subject the water to vigorous aeration; secondarily, to remove any particles (of floe, etc.) large enough to settle against the upflow velocity. The measured volume of the aerator was 5-9 liters. At a flow rate of 416 cc/min, this provided a theoretical detention time of 14.2 minutes. From a study of the detention time by the salt tracer technique, it was found that the flow- through was almost ideal. The action of the aerator as an upflow clarifier might be expected to remove some iron if floes large enough to settle out were formed. This floe formation would depend upon the rate of the oxidation reaction and the coagula- tion phenomenon. As subsequent discussion will show, the aerator appeared responsible for some iron removal. During each experiment, a complete record of dissolved oxygen of the water before and after aeration was made. Air flow was maintained constant and the same for all experiments, so as to eliminate one variable throughout the entire study. Excluding the waters with a high sulfate concentration, the aerator produced a water with a final mean DO of 6.9 * 1°0 mg/1. In general, this represented a DO increase of 6.36 - 0.8 mg/1. Sn terms of per cent of satu- ration, the uniformity of the performance on all types of water was even more noticeable. The final mean DO was 79.5 per cent of saturation, and represents a mean increase of 7^.0 per cent of saturation. This performance included water o o with temperatures from 13 C to 27 C. Table k summarizes the performance of the 22 u o .—J o U ^ > 9* o W a U *• t Pi X .2 a> S o fH u to o h o u u fll fl> o (LI VH a 0) a CQ < 0) o •4-> s .r-1 (0 H M S T3 O fl ffi a X a. sO nO nO vO sO Hd 'uoi;bj;u3DU09 uoj usSo-ipA^ 23 m> -Q fO UJ h- o < < O < 1x3 >■ to LU o a, O o o o o o o o 4-> +J 4-J 4-J +j — ■— . — . — +-> 4-> 4-> 4-J . — . — . — . — 0) (1) Q) 0) 0) ro - to -C 4-J C to Synth Synth Synth Natur 4-» (0 2: 4-J CO z 4-> (0 z Synth Synth Synth Synth !_ 3 4-J z s_ 3 4-J CD Z s_ 3 4-1 CO Z L. T) T) T3 "O T) T3 3 O CO 4-J CD CL) Q) (U a) a> ZLlU-U.Ll.Ix Ll a) ko — -of J — lo ltf\ E O I JCILA — O I ! D1 [O E c l— 0)1 J Ol— E L.I 4-J Zvfl I ' —Li- e tol +JI o —I 1 H- CM I I O E 4 LT\ v^l f^J-vD CM CM c<\ .00 I I -5— ) -y ~i —) ~) ~i ~» 24 Table 4 SUMMARY OF AERATOR PERFORMANCE Percent DO After Saturation Gai n in 1 ncrease Experiment Aeration After DO of Percent No. mq/1 Aeration mq/1 Saturation Synthetic, in Laborato ry 3 6.41 79.0 6.02 74.2 4 7.33 87.2 6.60 78.6 5 7.03 85.6 6.93 84.4 6 6.65 83.9 6.53 82.4 7 7.36 88.0 6.07 71.4 8 7.15 85.4 6.38 76.1 9 6.85 85.3 6.81 84.8 10 4.81 67-3 4.81 67.5 Mean 6.7 82.7 6.3 77.4 Nat :ural , i n Laborator y 11 7-39 84.6 6.57 75.1 12 6.87 82.0 6.51 77.6 13 7.40 84.4 6.59 77.5 14 6.77 86.1 5.54 72.1 15 7-36 84.2 6.19 70.8 16 7.33 84.5 6.05 67-7 17 6.39 77.5 6.39 77.5 18 6.66 80.4 6.25 77.4 Mean 7.1 83.0 Natural , i n Field 6.3 74.5 19 6.87 71.1 6.87 71.1 20 7.07 70.4 7.07 70.4 21 7.83 77.3 6.36 71.9 22 6.10 60.4 6.10 60.4 23 6.88 65-7 6.88 65.7 24 6.12 60.5 6.12 60.5 Mean 6.8 67.6 6.5 66.7 25 aerator; the grouping into synthetic water experiments, laboratory natural water experiments, and field natural water experiments shows small but significant differences in performance. The study of dissolved oxygen concentrations associated with treatment unit processes following aeration was confined to the J series natural water which contained large iron concentrations and was the most difficult to treat for complete iron removal. Table 5 summarizes the concentrations of DO found after aeration, after reaction-sedimentation, and after filtration. The theo- retical consumption of oxygen for oxidation of ferrous Iron Is Q.]k mg. for every 1.0 mgo of iron removed but the average DO decrease observed was only 0.07 mg. which Implies that 50 per cent of the total Iron removed by the reaction-sedimen- tation plus filtration had not been oxidized by the time the water left the aerator, but was oxidized (and later removed) by the reaction-sedimentation unit followed by filtration. D . ! ron Remova 1 The basic objective of the entire experimental work was to remove Iron from waters with varying characteristics. The method of aeration, reaction-sedimentation and filtration success- fully removed the iron from all experimental waters except one, and, in subse- quent experiments, this water was successfully treated. Table 6 shows the overall per cent iron removed during each experiment. The waters used are grouper according to the scheme presented earlier in this section. Table 7 Is a statis- tical summary of the total iron analysis data accumulated. Mean, mode, maximum and minimum values for concentrations of iron at each sampling point of each experiment are shown. By examination, it is evident that the mean values are representative of the concentrations of each sampling point, despite an occa- sional wide divergence. Table 8 Is a tabulation showing the iron removal effected by the aera- tion process alone. The aerator removed iron by sedimentation of floe as pre- viously discussed, and by the formation of Insoluble oxide films which adhered to the Interior surfaces of the aerator. The variation in iron removal was great among the groups; the conclusion drawn is that iron removal during aeration was greatest and most consistent for the natural waters brought to the laboratory for study. The variation among waters in the same group indicated that the iron re- moval during aeration was Influenced by some characteristic other than those evaluated In these early studies. 26 Table 5 OXYGEN CHANGES DURING IRON REMOVAL BY SEDIMENTATION & FILTRATION FIELD STUDIES, CISCO, ILLINOIS Parameter Measured ^ Experiment Number or Calculated 19 20 21 22 23 24 DO after aeration mg/1 6.78 7-07 7-83 6.10 6.88 6.40 DO after sedimentation, mg/1 — — 7.85 6.31 Not used 6.45 Change, sedimentation +0. 10* +0.10* +0.02 +0.21 +0.05 1 ron removed by sedimentation, mg/1 0.0 0.04 0.0 0.0 — - 0.0 DO after sedimentation, mg/1 6.88* 7.17* 7 = 85 6.31 Not used 6.45 DO after f i Iter mg/1 5.96 6.46 6.36 5.45 5.97 6.34 Change, f i 1 trat ion 0.92 0.61 1.49 0.86 — 0.11 1 ron removed by f i 1 trat ion, mg/1 10.68 10.40 11.32 10.90 11.05 11.09 Ratio Fe/ DO 13.0 17.1 7.6 12.6 12.1 10.08 0.47 10.44 10.80 10.56 11.05 10.80 0.08 0.06 0. 13 0.06 0.08 0.01 Total DO change from after aeration to effluent mg/1 -0.82 -0.61 -1.47 "0.65 -0.91 -0.06 Total iron change for above mg/1 Ratio DO/ i ron Expected DO change: 0.14 mg/1 for 1 mg/1 iron change Found DO change 0.07, or 50 per cent *Best estimate from limited data No. 21 & 22. 27 Table 6 REMOVAL OF IRON BY AERATION PLUS REACTION-SEDIMENTATION PLUS FILTRATION Water Group Type Per Cent of Applied Iron Removed A B C 1 C 2 D 1 D 2 D 3 D k F 1 F 2 G 1 G 2 H 1 H 2 I 2 3 k 5 6 7 8 Syntheti c Synthet I c Syntheti c Syntheti c Synthetic Natural Natural Natural Natural Syntheti c Syntheti c Synthet i c Syntheti c Natural Natural Natural Natural Natural - Natural - Natural - Natural - Natural - Natural - Field Field Field Field Field Field 100.0 99.5 99.8 99.0 100.0 99.5 99 = 5 100.0 99.0 99.0 99-6 100.0 100„0 99. 97 . 28 99 98 97 97.8 97.7 97.9 99.7 Average including J 1 Average wi thout J 1 96.2 per cent 99.5 per cent x l_ CD 4-> ■M C ■ _] < z < z O cm O z LU 2: < LU < o h- co £ co 1- +J a) c j-> a) CO i/1 S CD !_ T3 a. a> — c +j o +J L. .— 3d) 2: !_ — ■O CL\ CD CD (U +j c E "O fD o o CD < X ifl .— ■M CD 2: CD 1- — 2 0- *>> en CD I 5- c e -a CM 00 LA o co 00 LA o LA 00 is CM LA CM LTV CM o vO CM co co a\ CO CM IS IS CM CM cr» is CM CM — is cr> is CM CM J- vO CO CO is CM isv£> CM CM 00 CM 00 00 CM CM O OO CO CM co co co co is .3- 00 vO CM CO CM CM 00 vO — ~ CM CO CO CO 00 vO CM \iD is^o — o CM CO CO CO 00 is cm i . o . I CO CO CO I 00 CM vO I CM CO CM S camo — 1 CM CO CO I vO CO 00 CO 00 00 o rs CO CA CO vO O co co — M3 CM CM CM CM CO CO CM CM CM 00 CM CM Co — CM CO CM CM la 00 is rs CM CM CM CM 00 CM O LA isoo is is CM CM CO CM CO CO O Is \0 N CM CM CM CM vO CO CM CM CO O — LA O 00 vO vO IS 00 — CM CO— — — CM — — — 00 is — oooooj-isOOvO — — — 0000000 0000 vO CSOvD J - - CDO . . . . . . . • c ° — — CM — — O — — O — CM — 00 00 o\ a\ is is 00 00 — CTkvO 00 OO LA 00 CO LA CA |s vO LA— CM — CMOOO — — — — J- — 00 CM 00 CO LA CO CO O — — — ! — 00 CM CO O CM 00 « ■ • a o OC^OONffl o o la is j- — CA 0° — — CO — OO — — I — 00 CM vO co CXV CO 00 o CM CO — CM LA CO o CO co CO vO CM CO 00 co co co LA — -3" 00 LA CO •J- — LA O IS CAvO O I CM CO CO I vO O 00 CO CO LA LA-3- 00 CO LA O CO CO — — co co co co J" - LAN- co^j .4- _3- — coco cm CM LA J" CT\ J- CO CO J" J- CO vO o LA -A CM CM LA O CM CO CO 00 CM CM co a~\ CM CM 00 00 CM CM o a\ ^00 rs, is IS O LA CM LA — J- rs j- la is o co 00 ! I o — — CM — OO — O — CO LAOOLAvOrs IsCTv COCO CM — — — — I — — CM CM — — — I — — CO — — — CM J" CM CM CO v£> O co vO O O _jl co — — — 1 — — CTva) coco isoocois cdo) • c •• .....ICC CMO CMCM vDvD - — — I OO Z IZZ CMCT\ OOvO — is CMCO-4"Cr\OLACr\\£> coco ' — cm cors CAcocMcors' — cmco CM CM CO LA CM CM ISMD 0000 oocaooca oo cors 0000 coca enm-oocnoo — CM — CM CO J- — CM — CM — CM — CMcOJ-LAvOlsOO OO OOOa LU U-U. CD CD II -i—)—i-)-i—)-i—3 a. ° x o LU Z 5* 00 o ■l-J CD c CO IS o +J CD c CO J- LA MD CO J- vO o o u •M CD CD CD CD ■M +J 0) CD x: ji x:, l. s- l. +j +j -m 3 a d C C C 4-1 +J A-l >» >> >- CD CD CO CO CO CO z z z CM CM CO o o CD CD JZ JZ +J +-> c c CO CO (T\ O — LA o o 4-1 +J CD CD _c x: +-> +J c c >- >- CO CO CD CD !_ U =3 ZJ +J +J CD CD IsOOCAO — CM M4" — — — CM CM (N CM CM cd m l. L.-o-a-o-a-a*o +J+->CDCDCDCDCDCD 03 CD — — .- — — — ZZ1lU.IlLU.Ll 29 Table 8 REMOVAL OF IRON BY THE AERATOR Water Per Cent of Applied Iron Removed 44 12 35 Avg. 19 3.2 8.1 20.5 Avg. 20 31.4 I ncomplete Natural 24 Group Type A Syntheti c B Synthet i c C 1 Syntheti c C 2 Syntheti c D 1 Syntheti c D 2 Natural D 3 Natural D 4 Natural F 1 F 2 G 1 G 2 H 1 H 2 J 1 J 2 J 3 J 4 J 5 J 6 J 7 J 8 Syntheti c Synthetic Syntheti c Syntheti c Natural Natural Natural Natural Natural - Field Natural - Field Natural - Field Natural - Field Natural - Field Natural - Field 5.6 1.7 Avg. 3.6 10.7 45 Avg. 27.8 22.5 35-4 Avg. 29 28.7 30.4 Avg. 29.6 6.1 5.9 5.6 3-0 0.1 5.0 Avg. 4.3 30 Table 9 is a tabulation showing the removal of iron by the reaction- sedimentation process. The results of removal of iron by the reaction-sedimen- tation process were more consistant than those for aeration, and showed that zero to about 12 per cent removal can be accomplished by sedimentation after aeration. The synthetic waters showed the most reduction of iron by sedimenta- tion, while the natural waters in the laboratory studies and field studies showed very little removal. Table 10 summarizes the removal of iron based on that amount of iron applied to the sand filter. With the exception of J I, the sand filter averages 98.9 per cent removal of the iron applied. The lower per cent removals are for natural waters, and occur with both field studies and laboratory studies. Table 11 gives the percentage of the total iron in the raw water re- moved by the sand filter,, when the removal is accomplished by aeration, reaction plus sedimentation, and sand filtration. It is readily seen that the removal varies from 50 to 98 per cent, and shows two trends with respect to various waters. First, there are situations where the removal of the iron is almost entirely accomplished by the sand filter. Experiments F 1 and 2 (waters with the hardness and alkalinity approximately equal) demonstrated this. The waters used in the field studies also showed that most of the iron removal occurred during the sand filtration. The second trend is that some waters, notably those with high sulfate concentrations, settled more successfully than the others. Waters with alkalinity greater than hardness oxidized very easily, but removal by the filter appeared unaffected. E. Carbon Dioxide Reduction and pH Considerations The carbon dioxide (C0_) study was confined to an evaluation of the CO reduction and accompanying pH changes resulting from aeration, and to the relationship between the changes, the water group characteristics, and iron removal. The mean value of C0« removed was 77 per cent, the maximum was 88 per cent, and the minimum 5& per cent. The change in pH is an Important and quite obvious result of changing the CO- content of the water. There was a significant increase in pH of the water following aeration. The reaction-sedimentation pro- cess and the sand filtration generally increased the pH slightly, but occasionally a water showed a drop in pH. The final pH of the waters investigated was above ~J .k in all but one experiment. There was no relationship between final pH achieved and the degree of iron removal. 31 Table 9 REMOVAL OF IRON BY THE REACTION-SEDIMENTATION TANK Water Group Type Per Cent of Applied Iron Removed A B C 1 C 2 D 1 D 2 D 3 D k F 1 F 2 G 1 G 2 H 1 H 2 J 1 J 2 J 3 J k J 5 J 6 J 7 J 8 Syntheti c Syntheti c Syntheti c Syntheti c Syntheti c Natural Natural Natural Natural Syntheti c Syntheti c Syntheti c Syntheti c Natural Natural Natural Natural Natural Natural Natural Natural Natural - Field - Field - Field - Field - Field Natural - Field 12.5 9.1 0.0 8.6 7.5 0.0 0.0 Incomplete data 0.0 Incomplete data 0.7 k.S hj 0.0 0.0 2.2 3.6 0.0 0,4 0.0 0.0 0.0 32 Table 10 REMOVAL OF IRON BY THE SAND FILTER AS A UNIT PROCESS Water Group Type Per Cent of Applied Iron Removed A B C 1 C 2 1 D 2 D 3 D 4 Synthetic Syntheti c Synthet i c Syntheti c Syntheti c Natural Natural Natural 100.0 99.0 99.0 99.0 100.0 99.5 99.7 100.0 Natural 99.0 F 1 F 2 G 1 G 2 H 1 H 2 J 2 Syntheti c Syntheti c Syntheti c Syntheti c Natural Natural Natural Natural 99.0 99.5 100.0 100,0 99.4 95.5 0.0 99.8 J 3 J 4 J 5 J 6 J 7 J 8 Natural - Field Natural Natural Natural Natural Natural Field Field Field Field Field 99.0 97*3 98.0 97 = 8 97.9 99.4 Avg. Except J 1 98.9 33 Table 11 PERCENTAGE OF ALL IRON REMOVED BY THE SAND FILTER, FOR VARIOUS WATERS Water Group Type A Syntheti c B Synthet i c C 1 Syntheti c C 2 Synthetic D 1 Synthetic D 2 Natural D 3 Natural D 4 Natural 1 ron i n Per Cent of Total 1 ron Raw i n Raw Water Removed Water, mq/1 by Fi Iter Only 5.1 50 3-3 72 4.5 65 3.1 88 3-2 85 4.5 80 4.5 69 3.9 78 Natural 5.1 75 F 1 Syntheti c 2.3 F 2 Synthetic 2.9 G 1 Synthetic 3.2 G 2 Synthetic 5.3 H 1 Natural 2.3 H 2 Natural 2.8 J 1 Natural 17.9 J 2 Natural 16,4 J 3 Natural - F ie Id 11.2 J 4 Natural - F ie Id 11.4 J 5 Natural - F ie Id 11.7 J 6 Natural - F ie Id 11.2 J 7 Natural - Fie Id 11.3 J 8 Natural - F ie Id 11.4 94 97 85 52 78 63 67 95 92 96 98 98 98 34 Attempts to relate either percentage of iron removed, or iron re- maining in the effluent, to either final pH or the change in pH showed no corre- lation. F. Effect of Variations in Sedimentation Time The 6 field experiments, the J series, were designed to study the effect of variations in sedimentation time. The following times were used: Experiment No. 19, 75 minutes; No. 20, 60 minutes; No. 21, 90 minutes; No. 22, 30 minutes; No. 23, minutes; No. 24, 30 minutes. For the detention times studied, overall efficiency and the efficiency of the various treatment units was apparently not affected. Iron removal efficiency remained high whether the sedimentation time was 90 minutes or zero. During Experiment No. 23, where no sedimentation time was provided, a rapid build-up in headloss in the filter was noted. After the second hour of operation, the filter developed a negative head. From the analyses of water taken from the sample taps in the filter, it was noted that the iron steadily penetrated deeper into the filter. However, the removal of iron by the filter was 98 per cent which was about the same as that observed for the other experiments. After the completion of this experiment, a re-calibration of the flow rate revealed that it was 555 cc/min or 133 per cent of the design rate. This excess flow rate could possibly account for the rapid build-up in headloss. G. Miscellaneous Determinations The routine analysis of the waters provided data for some evaluations of the following parameters: 1 . Al kal i ni ty 2. Hardness 3. Total solids (residue) 4. Sulfates 5o Chlorides 6. Nitrogen and COD 1. Alkalinity Table 12 summarizes iron removal in relation to the alkalinity changes for each experiment. Twelve of the 23 experiments showed an overall reduction of alkalinity, while the other 11 showed a small net gain or no change in alkalinity. For experiments 22 through 24 the increase in alkalinity after treatment was pos- sibly the result of a delay in analysis of from 24 to 48 hours after the samples Table 12 RATIO OF ALKALINITY REDUCTION TO IRON REMOVAL 35 Water Group Type 1 ron Removed mq/1 Aikal i ni ty Change mq/1 Ratio Aikal i ni ty 1 ron 5.1 + 1 Gai n 3.42 -2 0.59 4.44 3.10 +3 -2 Gai n 0.65 3.21 4.51 4.47 3.92 -4 -3 -4 -3 1.25 0.67 0.89 0.77 D 4 F 1 F 2 G 1 G 2 H 1 H 2 J 1 J 2 Syntheti c Syntheti c Syntheti c Syntheti c Syntheti c Natural Natural Natural Natural Synthetic Syntheti c Syntheti c Syntheti c Natural Natural Natural Natural Natural - Natural - Natural - Natural - Natural - Natural - Field Field Field Field Field Field 5.02 2.30 2.88 3.18 5.26 2.30 2.69 5.18 16.35 1.16 1.11 1.45 0.90 1.06 1.32 -9 Not avai lable -6 44 +4 -4 -20 -14 -4 -12 + 11 +5 44 1.80 1.88 Gai n Gal n 0.78 1.22 1.25 0.36 1.05 Gai n Gai n Ga i n Mean Value Expected Value, Reduction of Alkalinity for 1 mg/1 Fe 1.01 1.78 36 were taken. The reduction of alkalinity per mg/1 of iron removed showed a spread from 0.59 to 1.88 mg/1, as CaC0~, with a mean value of 1.0 mg/K Stoi chiometr i - cally, the reaction of oxidizing the iron bicarbonate predicts an alkalinity reduction of 1.78 mg/1 for every mg/1 of Iron oxidized and removed. The experi- mental data thus indicated that only about 59 per cent of the expected alkalinity reduction occurred. The 59 per cent reduction is for the overall treatment, but occasional analysis of intermediate alkalinity values showed that almost the entire change in alkalinity took place after the aeration. On this basis, the percentage of expected alkalinity reduction was about the same as the percentage of the expected dissolved oxygen reduction-~59 per cent and 50 per cent. This unexpected close agreement may be indicative of some removal mechanism which does not utilize dissolved oxygen and alkalinity. 2. Hardness The hardness of the raw water which in this study varied from 253 mg/1 to 427 mg/1 has shown no effect on iron removal. 3. Total Sol Ids Data on the reductions of total solids were available only for the experiments conducted on the water having a high iron concentration. From 8 experiments on this water, it was found the mean total solids change was 22 mg/1, with a range from 2 to 45 mg/1. The mean iron reduction as iron was 11 mg/1; converted to iron hydroxide, and assuming the waters of hydration would be driven off, this reduction is equivalent to 19 mg/1 of residue. The agreement between the 22 mg/1 observed difference and the 19 mg/1 calculated difference seems significant, especially since the raw water carried some very fine silt, and the amount of silt seemed to vary from experiment to experiment. k. Sulfate Sulfate analyses showed no appreciable change in sulfate concentration between raw and finished water, either natural or synthetic. One exception was the experiments where sulfite was introduced as an oxygen scavenger to maintain a low raw water 00. This water showed a sulfate increase almost exactly as pre- dicted from the sulfite added. Both high and low sulfate waters were tested; both were equally satis- factory In iron removal. 37 5. Chloride Chloride analyses of raw and finished waters showed occasional gains in chloride during the early stage of an experiment. This was due to carry over of HC1 from the cleaning operation. There was no significant change of chloride with respect to iron removed, either with natural or synthetic waters. Both high and low chloride content waters were tested. 6. Nitrogen and COD Table 13 gives a summary of the ammonia nitrogen (NhL-N) , organic ni- trogen (Org.-N), and chemical oxygen demand (COD) analysis made on natural raw waters and some effluents. It was found that the raw water total nitrogen (Total -N) concentration in no way affected the amount of iron removed, or the ease with which iron was removed from the experimental waters. Determinations of nitrogen for both raw and finished water in the J series showed a 2.6 per cent reduction of the total nitrogen as a result of the iron removal treatment. A relatively high COD was found in those waters which had a high Total-N concentration. The significant COD and nitrogen levels indicated that organic matter was present in some waters; and, for Experiments 12 plus 17 through 2k, contamination by an Organic-N waste is suggested. The difference in the ratio of nitrogen to COD also suggests that the organic matter present differs from source to source, or that the COD was influenced by a variety of other oxygen consuming consti tuents. Only the J series was analyzed for effluent COD; the change in COD was about 7.0 mg/1 , or 13=2 per cent, representing an appreciable oxidation by the aeration unit. 38 J CM (T\\£) CA • a LA — LA — 4-> O to . ° . . o o z o o o o i — M I CA ■ — CA_J" POM I I CM PA CM CM o o o o o o pa -Q CO r- < O o o o «a5 LU C3 O a: I- o >■ < CO Ol ,1 Cl-N. 1—8 en E 2\ fDI CD as e O l_ h- a) +j ai 3 Z -a . CD en -G u 1/1 O E C CD °i ■Ml « PA 4-» o Ul en fD IO i fid Z 1 4 A Q mooovo — o cm pa 0000 tQ • • • • z o o o o LA (0 CO CO «D -Cf O Z Z Z Z — CA CA O vO (D CM _^ O vD O z pa cm — — (1) (D (0 (0 <0 (D Z, Z. Z. Z Z. Z. ro ro ro ro fo ro — CM OO CO CM -4* CM r>.-3~ LAfAvO la o o o o la LA LA CM CM vO CM OOOOOO 00 cm oinN4 — cm 1 1 cm — — o 001 i o o o o . .i| . . . o 00 0000 r-~ r^. CA LA vD PA vD vD CO (D r d" LA PA O — O J- 00 CM NO CO (0 CM CA O LA LALAZZLA-3"LAv£> CM CA 00 — PA I ! LA -j- 1 LA LA LA LA CM CM LA O CO CA |^ — 00 CO CA crjco°°°>°* ZZOO — OOO cm cm o mm ro ro .4- -j- roLA^r_d- MNj^NOON \DvO LA -4 LALA^f LA O O LA CO PA\D LAvO m~OvDONCO N ——0—000 ^0 NvD\DvD M .4- \£> -4" CM CM CM — LA -V0 4" Lf\CMvO LA LA -4 -3" -4 -4- 4" LA — CM PA 4" LA \£> r^.COCAO — NCAJ — — .— — — r— ^-^-.— (Njpgijvlisvigvj — cm oa cm 4- xuidaxo — CM PA -4 LA vD r-^OO -^—$-i-i—)~>-)-5 cd ro c O 4 s CM C cd E -.s_ CD a x UJ i_ o c a) o a) si c ro -C o CD +j a) 13 at — c O ro J3 o < J- LAvO CM — CM CM PA CM O — -3" O o o r-^ -O a) jC !_ in a) 4- r^. — 00 •— . 4-J » c ro -4 LA LA ._ 3 — 4" o« u. a) !_ 3 O (/) CD E ro it) u 2 cu vO CO LA 00 E ro ■u O cc ro -4 LA CM v_ 3 — LA \ en en CT) E E ■ — !_ E \ a) z z en 4-» Z F a> E PA Cn +j O ro X L- O O s_ Z O l~ O in D_ c c c c c ro CO CO ro CD a) a) cu ro •m z 39 IV DISCUSSION OF RESULTS A. Preparation of Synthetic Ground Water There are very few references available on how to prepare synthetic ground waters, or on how to accomplish economical carbonation and oxygen strip- ping which is important in preparing waters for studying iron removal. The most advantageous sequence of gas application seems to be carbonation, concurrent with filling the reservoir and with the compounds for alkalinity and hardness present. This may be followed by N» stripping, if necessary, to lower the DO to a concentration of about 0.10 mg/1 prior to the addition of the ferrous com- pound. Since N„ also strips out CO., and thus raises the pH, it is desirable to re-carbonate slightly after the stripping. Mixing during an experiment was accomplished by N~ agitation; this resulted in gradual pH shift during the ex- periment. Water characterized by alkalinity greater than hardness appears less stable; i.e., iron is more readily oxidized and removed. This observation agrees with the formulation of the reaction from ferrous bicarbonate to hydrated ferric oxides, in which bicarbonate is required to complete the reaction to a floe which will precipitate. The observation that free CO- reduces the pH was in accord with the expected equilibria conditions; the waters prepared for laboratory experiments were generally buffered by the presence of other salts. The pH values for the more stable waters was in agreement with the pH of natural waters. The synthetic waters used in this study substantiate the concept that high C0_ retards oxida- tion because the high CO- experimental waters were more stable and remained clear (unoxidized) for longer periods. This is true only for waters without organic complexing. B. Aeration Unit and Dissolved Oxygen Study The aeration device provided vigorous aeration, with long exposure of the water to the bubbles. The salt tracer technique was found to be a convenient means of measuring the detention time. Calculations based on the tracer study showed an effective recirculation of 1.5 times through the multiple-pass aerator and a detention time of 14 minutes. The aerator may have served as an upflow clarlfier, as was shown. The ability of the experimental aerator to consistantly produce a water of about 80 per cent saturation, without regard to initial DO (as long as It is low) or mineral characteristics is an advantage not possessed by conventional aerators as reported in the literature. It should be remembered, kO however, that the air rate used in the experimental studies is far in excess of any rate used in practice. Further study of the aeration ability at different air rates are to be made. The overall decrease in DO through the reaction-sedimentation plus sand filtration processes was found to be sufficient to account for the oxidation of only 50 per cent of the total iron present after aeration, but removed from the finished water. Since the ferrous iron determinations positively showed ferrous iron present after aeration, the most plausible reason for the decrease in DO is that the DO was, in fact, used in the oxidation of the ferrous iron,, This decrease in DO occurred mostly in the filter. There are three possible mechanisms which might contribute to the overall efficiency of iron removal: a) catalysis by previously deposited material In the filter; b) a very long passage and contact time with thin films, etc., which could promote the oxidation reaction; c) adsorption and reaction phenomena based on the adsorption of floes to the sand grains and subsequently oxidation. The catalytic effect of "ripened" filters is widely discussed in the literature; the other possibilities are not. The same sand was used throughout all experiments, so it is possible that the bed became "ripened". The time of passage in the filter was actually quite short, since about two liters of water would fill all the voids. Two liters would represent a five minute flow-through time during which accelerated oxidation by tortous paths and close contact might be expected (postulate b) . The oxidation and re- moval rate would then have to have been greatly accelerated, particularly since oxidation and removal in the reaction-sedimentation process was very slight for times of 30 to 90 minutes. Vigorous backwashing after each experiment removed as much floe as possible from the filter, so that any floe adhering to the sand would have to be placed as an experiment progressed. The removal by adsorption would thus be expected to increase, as more floes were deposited on the sand, or, therefore, as an experiment progressed. The effluent usually showed an iron concentration greater during the first few hours of an experiment than during the remaining time, up to a definite break-through. This early higher concentration lends support to postulate c, that removal did proceed via an adsorption mechanism, and that the mechanism was more effective during the middle time period of an experiment. Ifl It is also possible that the filter was not operating at peak iron removal until the first large floes had been deposited on the top of the sand and thus created a micro strainer which was effective in trapping all subse- quent floes. A filling of pores in the upper sand layers was observed, but no real build up of floe deposit was noted. Co S ron Removal The consistantly high removal of iron from a variety of waters showed that none of the routinely determined mineral characteristics of the waters studied had any appreciable effect on the overall efficiency of iron removal by aeration s reaction-sedimentation, and sand filtration. The one experiment which gave only 28 per cent removal of iron should not be completely discounted. The probable explanation of poor removal may be related to the fact that the water was taken from a well which had not been used for some time; the iron content as a result of standing was about 17 mg/l~-much higher than for the other experiments. Subsequent experiments with this water gave satisfactory results. The conclusion drawn, then, is that the extremely high iron content, plus the possibility that the iron may have been of a dif- ferent character either because the water was in contact with the well casing for some time, or because iron was actually picked up from the casing, accounts for the unsatisfactory removal performance. Other experiments with water from the same source showed satisfactory removal, but. the raw water iron content was less--about 12 mg/1 , and the well was flushed before the water was collected. The use of a flocculator in Experi- ment 19 did not appreciably alter the removal by the reaction-sedimentation unit, but it may have improved the floe so that the iron was removed more efficiently by the f i Iter. In the aerator, removal is by two mechanisms, by adhesion of the ferric oxide to the surfaces of the aerator (especially noticed at the foam region of the center tube), and by sedimentation in the up-flow portion of the aerator. The diffuser in the aerator gradually clogged during an experiment, making it necessary to gradually increase the air pressure to maintain a steady flow. From this experience, it is believed that proper maintenance of diffusers would be very important to the successful operation of a plant using diffused air. Iron removal by the reaction-sedimentation tank was consistantly small. Synthetic waters appeared to be more likely to form settleable floe. The waters with high sulfate concentrations show consistantly better removal by sedimentation, while the low sulfate waters occasionally showed some removal by sedimentation. kl The natural samples brought to the laboratory showed very low iron removal by the reaction-sedimentation process. This may be correlated with the aeration device removal since it was shown that the aerator removed more iron from natural samples brought to the laboratory than from others, and that the removal was most probably by sedimentation in the aerator. The field studies showed no iron removal in the reaction-sedimentation tank. Possibly this resulted from insufficient time to complete the oxidation, plus insufficient mixing and time for sedimentation to be effective. Synthetic samples (which showed some sedimentation removal of iron) did not contain organic matter, so no complexing of the iron, or interference with coagulation would be expected for this reason. Jar tests showed that the typical fluffy, gelatinous floes were formed with the synthetic waters; the typical floes were not formed with most natural waters. A separate experiment showed that a very long sedimentation time in- creased the percentage of iron removed. After one experiment was completed, the sedimentation tank supernatant was sampled without disturbing the small precipitate on the bottom. After 3 days of quiet sedimentation, the super- natant liquid was again tested for iron. The iron concentration of the super- natant decreased from about 11 mg/1 to 0.1 mg/1. In this experimental work, the sand filter accomplished most of the iron removal --50 to 98 per cent of the total amount removed for the waters tested, The balance of the removal was done by the aerator and the reaction-sedimentation unit. In experiment J 1 all iron removal took place prior to the filter. In all remaining experiments, the filter removed from 95 to 100 per cent of the applied iron, and averaged 98.9 per cent. There was no trend for separation of per- formance by water groups. It is noted in Table 10, however, that natural waters less frequently showed a 100 per cent removal of applied iron by the filter. This might be the result of organic complexing, since only the natural waters had organic matter present. Even at that, only the J series had much organic matter, as evidenced by a COD of about 50 mg/1, and this water showed a fair iron removal by the filter. In this study, the iron concentration in the effluent followed a definite time oriented pattern; a typical plot is shown in Figure 10. The relatively high iron in the first hour or two indicated that iron passes all the way through the filter until some "adjustment" occurred. This "adjustment" took an hour or two, and might represent the time required for the collection of floes on the surface of the filter, or time for individual grains of sand in the M IT) i— i O 1 2 o c o — ID — •M TO •w c S (0 » " V. . 0) J3 UJUI 1- « C ec — <£ C p l. 3»