to ^//CIVII ENGINEERS i STLWIB ,/9Ze-S505 e.> SANITARY ENGINEERING SERIES NO. 11 C'-'-IOAN COPY MEASUREMENT OF REDOX POTENTIAL AND DETERMINATION OF FERROUS IRON IN GROUND WATERS ^ J 5s*. *<£ \3 *i • X 3 T 1 * ^4 fc ^ k J v> x^ ^> * \Nf L/^ \ /x X J *:W < Babcock (7)j> Fosnot (17), and Nordell (3*0 = The most widely used method for iron removal, which is also regarded as the conventional method, consists of aeration, sedimentation and filtration. This method is more economical than other methods. The principles on which the conventional method is based require that the iron which exists in ground water, mainly in the form of soluble ferrous iron, be converted by oxidation to insoluble forms of ferric iron which can be removed by sedi- mentation and rapid sand filtration. Hence the success of the treatment will depend very largely on the degree of oxidation attained. However, the mechanisms by which the iron is removed by sand filters are not yet under- stood. It is assumed that there are three possible mechanisms which may contribute to the overal efficiency of iron, removal (13): (a) catalysis by previously deposited material on the filter media; (b) a very long passage and contact time with thin films, which could promote the oxidation reaction; (c) adsorption and reaction phenomena "based on the adsorption of floes to the sand grains and subsequently oxidation. Many studies had "been performed with and without the use of pilot plants in an attempt to delineate the factors and mechanisms which control the iron removal process . However, many attempts have not succeeded owing to the lack of a reliable method of determining ferrous iron. The deter- mination of ferrous and ferric iron must "be reliable so that the state of oxidation of the iron present at any point in the treatment process can be determined with confidence. The conversion of ferrous iron to ferric iron is an oxidation- reduction reaction. Therefore, the measurement of oxidation-reduction potentials (redox) may help to reveal the reaction phenomena and may also be useful as a possible means of controlling the treatment process, since the relative concentration of oxidants and reductants involved can be measured by determining the prevailing redox potential. One of the great advantages of redox potential measurements, if it can be used as a controlling parameter during treatment, is that the measurements can be accomplished by simple techniques. The measurement depends solely on the use of instru- mentation which does not require a highly skilled analyst and can be ob- tained and installed within a relatively short period of time. It could pave the way for automatic control of water treatment processes involving iron removal, in a way similar to its successful application in many waste treatment processes. B. Purpose of the Study The present study can be divided into two phases: 1. Determination of the reliable method for the analysis of ferrous iron. Up to the present time, several methods have been proposed for the determination of iron, and many methods pertaining to the deter- mination of iron have been quite thoroughly summarized in a report presented by Oborn ( 36) However ^ almost all of the methods are concerned mainly with the determination of total iron^ and they cannot be used to distinguish be- tween the ferrous and ferric forms . It is particularly important to the succeeding work in this study that an accurate and reliable method for ferrous iron determination be selected. The ortho-phenanthroline method given in the 11th Edition of Standard Methods (2) was first selected for the study since it is claimed that for natural and treated waters ^ the ortho- phenanthroline method has attained the greatest acceptance for reliability (2, ^3)<> The method also can be performed with a relatively simple pro- cedure and can be adapted with ease to field determination. The reliability of the ortho-phenanthroline method was then compared with the 4,7 diphenyl - 1, 10 phenanthroline reagent method or bathophenanthroline method. The bathophenanthroline method was considered because it is a recently developed method which has been found to give a higher degree of accuracy and sensitivity than the ortho-phenanthroline method (27* 31) ° However,? this method involves a more complicated procedure which cannot be well adapted to field determinations. Therefore, the scope of the study in this phase has involved the determination of the accuracy and reliability of the ortho-phenanthroline method as compared with the bathophenanthroline method. Some modifications of the ortho-phenanthroline method were investigated in order to adapt it to field application without sacrificing accuracy and reliability. 2. The second phase of the study is devoted to evaluation of redox potential measurement as a parameter determining the efficiency in the iron removal of ground water. Another objective of the survey was to determine whether there is any relationship between the characteristics of raw water as indicated "by the redox potential and other chemical analysis to its treatability with respect to the iron removal. Thirty-one plants in Illinois were selected for this survey study and their locations are shown in Fie* 1, \ JOW± mo" Walnut* •La Moil le •Wyanet Forrest El Paso* Morton* •Hudson Tremont* *Danvers Chatsworth ^oe\\° • .<£• •OaWood n ; Deland «TftinSn Fa i rmoun t Arqentao*Clsco #To i§?? ad ', ands \No9 Warrensbu'rg* Bethany* Findlay* *Atwood •Areola J5I ►Windsor $5 Roxana •Edwardsville Marine J FIGURE MAP SHOWING TOWNS ANo\ VILLAGES IN ILLINOIS ) COVERED BY WATER \ / TREATMENT PLANT SURVEY "V A ( 43& J J II. HISTORICAL REVIEW A. Determination of Iron and Ferrous Iron The methods for determining iron which are in general use can be classified into two types, i.e., volumetric titration and colimetric procedures . The volumetric titration was widely used prior to the develop- ment of colorimetric methods about a decade ago„ The principles of the titration method are usually based on the chemical reduction of all iron which is present in the solution to the ferrous iron form» This is followed by titration with some oxidizing agents of know concentrations „ Numerous +2 - -2 +k oxidizing agents have been used., including MnO^ , I0*> Cr. „, and Ce However,? owing to their low sensitivity, these methods are limited to conditions where there are relatively large quantities of ferrous iron and where other reducing agents are absent « Hence they are not applicable to the determination of ferrous iron at the low concentration generally found in the natural waters ( 31) ° Because of their high sensitivity and the ease in performing the analyses, colorimetric procedures have become the most widely used methods for the determination o A variety of color- developing reagents have been used,, ranging from thiocyanate, to pyridyls, nitrosophenols and several derivatives of phenanthroline . Some of these common colorimetric reagents with their characteristics have been presented in a table by Sandell ( k-3) ° However, many of them, are limited in their use due to various interferences and the factors which also must be taken into consideration, as Sandell has pointed out, "There was no lack of colorimetric reagents for iron, but comparatively few are well, suited for the determination of traces of iron. Some reagents react with ferric iron, others with ferrous ■ The former generally have the disadvantage that they cannot be used in the presence of appreciable concentrations of such ions as fluroide and pyrophosphate which form stable complexes with trivalent iron in acid medium, in which the color reaction is carried out. In such cases better results can frequently he obtained by resorting to a reagent reacting with ferrous iron, which does not form such complexes or at least less stable ones . " Among all of the available colorimetric reagents, ortho-phenanthroline has been highly recommended, since it has proved to give very stable color in the determination of total iron (2, 36, k-3) » The other recommended derivatives of phenanthroline which also are employed include;, 2, 2° - bipyridyl and 2, 2", 2" - ter - pyridyl (2, 4-1) « Recently a new method employing the derivative ^,7 - diphenyl-1, 10 - phenanthroline also known as bathophenanthroline has been developedo This new reagent is quite specific and very sensitive to very small amounts of iron.. It has been successfully modified to detect the total iron concentration with accuracy up to one part per billion in high purity water (2?)° However, it can be noted that most of the methods of determination are primarily concerned with the total iron and do not distinguish between the ferrous or ferric iron forms o Since the ferrous iron is very unstable, it can be oxidized easily in the presence of oxygen as follows: 2Fe ++ + k HCOl 1 + H o + ^ C> * 2P (OH)., + k C0 o 3 2 2 2 e 3 2 If the sample is subjected to a strong reducing environment the reaction is reversed, and the solution may contain large quantities of ferrous iron (4l)„ In the determination of total iron by either ortho- phenanthroline or its derivatives, all iron present in the solution is generally reduced to the ferrous iron form by the use of hydroxylamine reagent (NH OH » HCl) „ If it is desired to determine the amount of ferrous iron in the solution, the reducing agent is then omitted* However, this 8 method is not very reliable. As pointed out by Lee and Stumm (3l)> these color imetric reagents are capable of disturbing the ferrous - ferric equilibrium. o There has been difficulties reported concerning the use of these reagents for specific ferrous iron determination, therefore. Another procedure whicn also is used for the analytical differen- tiation of ferrous iron from ferric iron is polarography, but this method is again limited to relatively large concentrations of ferrous iron and cannot be very well applied to the concentrations which are normally found in waters o Cooke and his associates (10) have developed a coulometric titration procedure which is very sensitive to low concentrations of ferrous iron. Titration have been carried out at concentrations as low as one part per billiono However, it requires of special instrumentation which is not normally available „ The 10th Edition Standard Methods (3) has recommended the use of ortho-phenanthroline with sodium acetate buffer as reagents for determining ferrous iron in waters, and only slight modification is made in the 11th Edition (2) by substituting ammonium acetate in place of the sodium acetate o It is claimed that the color developed is stable for several days- Lee and Stumm (31) noted that when the sample contained less than approximately 1 mg/l ferrous iron in the presence of ferric iron, the color formed was unstable and increased with time u They found HO to 60 per cent- increase in absorbance for various ferrous -ferric mixtures in a period of 30 minutes . From their investigations, they concluded that neither the ortho-phenanthroline procedure nor any of the previously reported procedures are adequate for the reliable determination of ferrous iron in solutions with concentrations less than 1 mg/l ferrous iron,, They attributed the failure of the methods to a lack of sensitivity or instability of the developed color o 9 It was concluded, from, the investigation that the ortho-phenanthroiine method be employed under certain conditions where \ l) 1 rag/l or more of ferrous iron is present in the solution being analyzed; 2) the molar ratio of ortho- phenanthroline to total iron (ferrous plus ferric) exceeds 30 j 3) the absorbance measurement is made within 10-15 min<. after color formation and k) the solution is free of interfering materials and protected from direct sunlight o Bathophenanthroline has been recommended as a substitute reagent to overcome these obstacles » The proposed bathophenanthroline procedure primarily follows the method which was developed by Smith and his associates (46) for the determination of total iron» The main modification is the omission of hydroxylamine hydrochloride when the ferrous iron is being deter- mined,, From the results of their investigation, they reported that the colorimetric procedure using bathophenanthroline as the color developing reagent had been found to give reliable determinations of ferrous iron in the presence of ferric iron in quantities measured in micrograms. However s they pointed out some of the disadvantages associated with the use of batho- phenanthroline ^ i«e.j the relatively high cost of the reagents and elaborate analytical procedure » Bo History and__Previous Work on Redox Potentials The significance of oxidation-reduction as a fundamental ecological factor has long been appreciated by investigators in the realms of physics, chemistry, biochemistry, bacteriology and soil microbiology ° According to Glasstone (l8), the studies of the oxidation-reduction systems were made by Wo D„ Bancroft as early as in 1.89'2» Though the practical, application of redox potentials have been widely accepted in various fields at the present time, it still has not been fully appreciated due to a number of difficulties which remain unsolved* 10 Theoretically each and every process Involving oxidation-reduction can be monitored by means of continuous redox potential measurement* In the presence of strong oxidants and reductants, the potential measurement can be obtained quite accurately as can be seen from its application to the potentio- metric titration, in which the types of the oxidant and the reductant present are known* In complex systems where a variety of redox components,* including some unknown species, are present, the redox potentials measured will repre- sent the gross effects of the redox components in the system* These may only indicate the relative effects of oxidants to reductants* Although the exact interpretation of these types of systems can not be established, the values obtained have been widely applied empirically* In cyanide plating waste treatment (^4-5, ^9)? £°*° example, redox potential has been used as a control parameter, since it has been found that the cyanide is oxidized to cyanates by chlorine at an approximate redox potential of +150 millivolts* The cyanates are further oxidized to nitrogen gas and carbon dioxide at a potential of approximately +700 millivolts* In the control of bleach manu- facture (15, 39) j? chlorine is added to a caustic solution (NaOH) at a rate controlled by the oxidation-reduction potential which is usually between +500 to +600 millivolts to form bleach* The treatment of toxic hexavalent chromium by reduction to the non-toxic trivalent chromium salts are also controlled by the use of redox potentials (25)* The applicability of redox potential as a control parameter in domestic sewage treatment was also claimed by many investigators e*g*, Nussberger (35), Hood and Rohlick (22) in the control of the activated sludge process; and was suggested by Grune and Lotze (19) in the control of sludge digestion* 11 The applications of redox potential as a ecological parameter in the natural environment have also "been studied "by many investigators. However, it is often pointed out that the systems in the natural environment usually are poorly poisedo Most of those who have made measurements therefore, have come to the conclusion that since they could not ohtain reproducible redox potential measurements, their measurements were unreliable. As a result, they often abandoned their investigations (6)„ Bass Becking and his associates (6) have accumulated data of redox potential and pH of waters from various environmental sources. They concluded that, at high pH» low redox potentials are more common than at low pH; whereas at low pH very high potentials are obtainable « They also felt that it may be possible to use the' relative redox potential and pH as a parameter indicating environmental limits of algae and few groups of bacteria. Pear sail and Mortimer ( 37) had studied the variation in redox potential with depth in natural waters and muds in relation to the seasonal change, they found that the redox potential of surface water was normally about 0.50 to 0.52 volts, although, during fall overturn, a slightly lower value was obtained. Allgeir and his associates (l) in their study of redox potential and pH of lake waters, recorded values of redox potential ranging from 0.580 to 0.505 volts. Pierce (38) in his investigations found that the redox potential, and the dissolved oxygen of ground water appeared to be important parameters in relation to the rate of tree growth. Deficiency of dissolved oxygen and low values of the redox potential were found to correlate with a slow rate of tree growth; such conditions of ground water were unfavorable, to the natural reproduction of most upland trees but not swamp species. In the application of redox potential to measure the corrosiveness of soil by Starkey and Wight (12, ^T), it was found that the bacterial 12 corrosion was most severe when the redox potentials ranged between and 100 m.Vo The sites at which moderate corrosion occurred had redox potentials ranging "between 100 and 200 mv. Well-aerated soils with redox potentials greater than 400 mv„ were classed as non-corrosive with respect to bacterial corrosion. The idea of employing the redox potential as a parameter in the water treatment plant is rather new. In a study of Weart and Margrave (50) in an attempt to explain the variatio n in performance of iron removal plants in Illinois, they found that a low redox potential in raw well water did not necessarily indicate unsatisfactory plant performance » The response of the water to aeration, as measured by redox potential, was an important consider- ation o When aeration alone raised the redox potential of a water to a level approaching 2h-k-mv,j the plant would function satisfactorily provided reasonable care was given to operation „ It was noted that in most plants using bicarbonate waters, containing ammonium but no sulfate,, and with a low initial redox reading, the potentials of these waters could not be raised to satisfactory levels with aeration alone „ Supplementary treatment with an oxidant such as chlorine had proved to be effective in maintaining oxidizing environment and increasing the efficiency of the process <> Weart and Margrave also found that in those plants with satisfactory iron removal, there was no -significant shift in the redox potential of water during filtration, even though the dissolved oxygen content was reduced. However, for those plants which showed unsatisfactory performance, the redox potential of the filter effluent reverted towards that of raw well water and the dissolved oxygen originally present in the influent was depleted. Under these conditions, completely ana.erobic -conditions existed in the filter, and some ferric iron was reduced to the soluble ferrous state with a resultant Mets Eeference Booat University of Illinois B10 6 NC1L 203 I. Soiaine Street TTfT-iOT-in T114, ■_ ,- -i ,-, ,■»-» 13 low iron removal efficiency « The investigators set a limit of 100 mv„ for raw water., above which iron removal is satisfactory, and below which iron removal is not satisfactory with conventional treatment,, Ik III. THEORETICAL CONSIDERATIONS Ao Princip les of Iron Determina tion by Ortho -p henanthroline . It lias been shown that the chemical reaction in the colorimetric determination of iron by ortho-phenanthroline is based on the chelating action between ortho-phenanthroline and ferrous iron which produces an orange-red color complex o This reaction can be represented by the chemical equation as follows ;. ^> L + "e II I F e ll X •" SJ (ortho-phenanthroline) (ferrous -or thophenanthroline complex) From the above chemical equation, it is evident that the two nitrogen atoms in ortho-phenanthroline each have an unshared pair of electrons that can be shared with the ferrous iron. Three such molecules of the organic compound attach themselves to the metallic, iron to form a orange-red complex ion. Ferric iron also will form a complex with three, molecules of ortho- phenanthroline in a similar manner and give a light blue color (ll)« The intensity of color formed varies with the concentration of iron present and follows Beer's Law y provided that the ortho-phenanthroline is in excess. The spectrophotometry determination then can be employed. The ferrous - ortho-phenanthroline complex which is orange-red in color may change to a blue complex due to the oxidation of ferrous ion to ferric ion [Fh 7 F ] 2+ — -> [Ph-J? } J+ f e" 3 e 3 e Orange - r sd Li girt ■ ■ blue 15 "Ph" denotes a molecule of ortho-phenanthroline . The complex salt of ferrous and ferric iron also are referred to as "ferrion" and "ferri-in". Upon prolonged standing the "blue complex of ferric iron may further change to a yellow- complex which is also produced by complexing ferric iron and ortho- phenanthroline ( 21) . In the colorimetric determination, there are important factors which also much be considered. Some of these factors are stability of the color developed, interferences by other substances, pH of the solution and the suitable wavelength to be employed in the colorimetric measurement. According to the kinetic studies on formation and dissociation of ferrous and ferric complexes with ortho-phenanthroline by Kolthoff and his associates (29), the quantitative conversion of ferrous iron to ferrous complex in acid solution is dependent on the ratio of excess ortho-phenan- throline acid. In order that the reaction be 99 percent complete, the ratio of excess (with reference to ferrous iron) ortho-phenanthroline to hydrogen ions must be 0.035 ox greater in the equilibrium mixture. Lee and Stumm (3l) also pointed out that: "Since phenanthroline is a base, it combines with a proton to form phenanthroline ion; therefore, the amount of excess phenan- tholine necessary is dependent on pH. At a pH of less than approxi- mately 5, the protons in solution compete with the iron for the available positions on the nitrogen atoms of phenanthroline. There- fore, the amount of excess phenanthroline is dependent on the acidity of color formation, particularly at pH values less than 4-5". As the results of studies by many investigators (16, 20, 33).? it has been found that in general the color developed with ortho-phenanthroline is independent of pH in the range between 3 and 8. According to Standard Methods (2), however, a pH between 2.9 and 3-5 insures rapid color development. Ortho-phenanthroline, as well as various derivatives, also react with a number of metals, particularly the divalent to form colored (e.g., Wi, Co, Cu) or colorless (e.g., Zn) complexes. The presence of any of these metals, although not interfering directly with the color measurement, may 16 consume reagent which in turn may prevent development of the full color in- tensity of iron. According to Sandell (k) who studied various interfering ions and their effect on the iron determination by ortho-phenanthroline: "Silver and bismuth give precipitates. Certain divalent metals such as cadmium, mercury, and zinc form slightly soluble complexes with the reagent and, moreover, reduce the intensity of the iron color, but the interference can be diminished by adding a larger excess of reagent. The maximum permissible concentration of these ions for 2 p. p.m. of iron is approximately 50 p. p.m. of Cd, 1 p. p.m. of Hg, and 10 p. p.m. of Zn. Hg may be present up to 10 p.p.m. (pH 3-9)° Beryllium (50 p. p.m.) does not interfere if the pH lies between 3.0 and 5.5; below the former pH a stable complex is formed with the reagent and above the latter pH the hydroxide is precipitated. Mo+° does not interfere above pH 5° 5.? but in more acid solutions it gives a turbidity. Tung- state decreases in color intensity but 5 p.p«m. of W does not harm. Copper may not exceed 10 p. p.m. and in its presence the pH must lie between 2.5 ancL ^-O. Nickel interferes by producing a change in hue and raising the transmittance below ^ko mju, so that no more than 2 p. p.m. may be present. Cobalt gives a yellow color and shoult not exceed 10 p. p.m. at . pH 3~5. Sn should not exceed 20 p. p.m. for pH 2-3,° and Sn should not be above 50 p. p.m. (pH 2.5). If the pH is kept above 6 and 3* respectively, oxalate and tartrate in concentrations as high as 500 p. p.m. do not inter- fere. In the presence of pyrophosphate the pH must be above 6; then 50 p. p.m. of P20y + causes an error of only 1 percent with 2 p.p.m. of iron. Phosphate may be present up to 20 p,p..m. of ^2^5 (P^2-9) » Fluoride ( 500 p.p.m.) does not interfere if the pH is kept above k.0. Chloride and sulfate are without effect, at least in moderate amounts. If perchlorate is present in more than small amounts, a precipitate of the slightly soluble 1, 10-phenanthroline perchlorate may be produced." Fortunately, these interfering substances usually are present in such low concentrations in ground waters that their effects generally are not of any significant concern in the normal determination. The ortho-phenanthroline color complex with iron has been shown to be very stable when employed in the determination of total iron and in the presence of excess ortho-phenanthroline and hydroxylamine-hydrochlorlde. Under normal laboratory conditions, the stability of color lasts up to six 17 months or more within the pH range of 2 to 9 (^3)« In an accelerated fading test using ultraviolet irradiation the color was found to be stable and unchanged after 100 hours (l6). However, in the presence of both ferrous and ferric complexes,, the color has been found to change with time due to the unstable equilibrium condition of ferrous and ferric iron in the solution. Harvey and his associates (2l) performed an experiment using ortho-phenan- throline in two solutions; one contained 10 mg/l of complexed ferric iron and the other 5 mg/l of complexed ferrous iron plus 5 mg/l of complexed ferric iron. They found that there was no appreciable change within 30 minutes. After that time, the change in readings was marked. The increase in absorbance (at 512 mu-) was greater for the mixture than for the ferric iron complex alone. After 15 days, the two solutions had approximately the same absorbance and became stable, from which it was concluded that an equilibrium between the two oxidation states had been reached. Lee and Stumm (3l) who performed the ferrous iron determination in the presence of the ferric iron by ortho-phenanthroline also found the color to increase with time. They has/e explained that the increase in color with time was the resul of the reduction of ferric iron to ferrous iron complex, since ferrous -ortho-phenan- throline complex has a larger stability constant as compared to the ferric- ortho- pkenanthroline complex . It has been found that the orange-red- ferrous- complex color has maximum absorbance at 512 mu. (2l) whereas the yellow color complex of ferric iron has its maximum at 3^0 mu. (20). From this, Harvey and his associates (2l) have suggested a rapid method of determining ferrous and ferric iron in a solution by measuring the absorbance at 513 niu. for ferrous iron and 396 mo. for total iron since he found that both ferrous and ferric-ortho- phenanthroine complexes have equal absorbance at 39°" m M-« However, the accuracy of this procedure is still questionable. The wavelength of 510 mu with a 18 light path of 1 to 10 cm has "been recommended in Standard Methods (2) for the determination of both ferrous and total iron by the ortho-phenanthroline method. The lower limit of sensitivity for iron using ortho-phenanthroline and a photometric measurement is found to be about 0.5 nig/l (^-6) > with a precision and accuracy under optimum conditions, according to Standard Methods (2), of 1 per cent (of absorbance) or 0.001 mg, whichever is the greater. However, the precision and accuracy also will depend on the method of sample collection and storage, and the iron concentration. The variability and instability in sampling will limit the precision and accuracy of this determination more than will the errors inherent in the analysis itself. The new reagent 4:7 - diphenyl - 1:10 - phenanthroline (batho- phenanthroline) has proved to be more sensitive than ortho-phenanthroline in many respects. Not only is the molar extinction coefficient of the ferrous bathophenanthroline ion (22, kQO) about twice that of ferrous - ortho-phenan- throline ( 11,100) but the new reagent can also be extracted from aqueous solutions with certain immiscible solvents, such as isoamyl alcohol and n-hexyl alcohol. Many important advantages are gained from this: (l) the iron in large samples can be easily concentrated into a small volume for measurement; (2) it is easy to free the necessary reagents from iron, which eliminates the blank correction; (3) the ferrous - bathophenanthroline complex is readily extracted by immiscible solvents which would tend to stabilize the developed colors; (k) the extraction procedure renders the method less subject to interferences from ferric iron and other interfering substances. B. Redox Potentials 1. Principles of Redox Potential Measurement Chemical reactions are classified into various types as "combination", "direct union", "decomposition" and "oxidation-reduction". As the name 19 implies, the oxidation-reduction (redox) potential measurement involves an oxidation-reduction reaction. Oxidation may be defined as a process in which (22): a. the proportion of oxygen in acid forming elements or radicals is increased; Id. hydrogen is removed; or c. a loss of electrons occurs with a resulting increase in positive valence; inversely.? reduction would involve (a) a loss of oxygen; (b) a gain of hydrogen; or (c) a gain in electrons , reflected by a decrease of positive valence. For every oxidation reaction, there must be a simultaneous re- duction reaction since free electrons cannot accumulate. A substance may be oxidized or reduced. However, another substance must be present to be reduced or oxidized; to either take or give electrons to the first substance. Although oxidation and reduction always occur together, the process may be written separately as half-reaction equations. For example the oxidation of iodide to iodine by hypochlorous acid (lk), H0C1 + H + + 2I~ = I + Cl" + H n may be separated into two half- equations, one 21" = I + 2"... ............. (a) representing the oxidation of iodide, and the other H0C1 + H + + 2e~ = H + Cl" ................ (b) representing the reduction of the H0C1. Addition of these half equations, neither of which occurs by itself, gives the complete oxidation-reduction equation. 20 When a metallic electrode Is placed in water, it will be dissolved to some extent . In this reaction, metallic ions are considered to pass into solution, and the electrode becomes negatively charged. The extent of the negative charge that accumulates on a metallic electrode is a function of the solution pressure of the metal, hence a potential is developed. This corresponds to the well-known electromotive force series of the metals. When the potentials are measured in terms of a normal hydrogen reference electrode and in a solution containing 1 gm. ionic weight of the metallic ion per 1,000 gm of water, the potentials obtained are called "normal electrode potentials", and they are usually represented by the symbol "Eo" ( hk) . The chart of normal electrode potentials for various metals have already been established and generally can be found in most texts which consider electrochemistry. Owing to different normal electrode potentials, when a pair of different metallic electrodes both immersed in a salt solution is connected through a conductor, a galvanic cell will result. The electrons will flow from, the electrode of higher potential to the electrode of lower potential. This reaction also is classified as an oxidation-reduction reaction. For example, if a pair such as zinc and iron is employed, the reaction of each electrode can be written as follows (26): Zinc electrode: Zn = Zn ' + 2e (Oxidation) Iron electrode: Fe + 2e = Fe (Reduction) The overall oxidation-reduction reaction is then _ o _ ++ _, ++ _ o Zn + Fe ss Zn + Fe There is another type of oxidation-reduction that involves only a change in the charge of the constituent Ions. The electrode material needs not participate in the reactno n; it may be an inert metal that serves as a source or sink for electrons and as a conductor. It is necessary, 21 however, that equilibrium conditions for the half-reaction of interest be established at the electrode surface . Potential can thus be determined, at least in theory, for any reaction yielding or utilizing electrons, i.e., for any oxidation-reduction reaction. For example, an inert platinum wire dipping into a solution of ferrous and ferric ions (half reaction .1... |. _L_L_L. _ Fe = Fe ' ' + e ), and also through a voltmeter to another system of the standard platinum hydrogen electrode, corresponding to the equilibrium of the half -reaction [H (gas) = 2H + 2e ], would result in the oxidation of the ferrous ions (Fe ) to ferric ions (Fe ) and in the reduction of hydrogen ion (H ) to unionized hydrogen molecule (H ). The reaction would be written as follows: Fe +2 + H + = Fe +3 + l/2H° The potential of the standard hydrogen electrode with the hydrogen ion at unit activity and the hydrogen gas at 1 atmospheric pressure is universally taken as standard and is assigned an electrode potential of zero at each temperature (ll). The magnitude of the oxidation potential of a system, referred to the hydrogen electrode, is expressed by the symbol Eh and is given by the equation derived from the Nernst equation (19, 50) . ■an- isi RT n I0x] Eh = Eo + -= In k= J -, nF [Red] E^ = E + £ log |§4t h o n to [Red] in which E, = The measured electrode potential in volts, in reference to the standard hydrogen electrode; E = A standard potential of the system at 30 Co when the activities of all oxidants and reductants are unity. 22 , - Rt 2,303 F R = 80 315 volt - joules coulombs (the gas constant) T = The absolute temperatures in degrees Kelvin F = The f araday constant = 96, 500 coulombs n = The number of electrons that participate in the potential system [Ox] and [Red] are molal concentrations of oxidant and reductant after correction for activity. Taking into account the difference between common and Napierian logarithms .. In = 2.3026 log x; the value of the constant k = ^ 5F = 0.06011 at 30% 0.0592 at 25°C J_J_ _L.Jm.-L -_. For example,, in the ferrous -ferric system (Fe = Fe ' ' + e + 0.771v) at 25 C the equation becomes : [Fe +5 ] E, = 0.771 + 0.0592 log L * n [Fe + ^] From, the general equation of E, it should be noted that the greater the proportion of reduced substance present., the lower will be the E, value; conversely , the greater the portion of oxidized substance the higher will be the E, value. In other words E, is not really the measure of concentration of the oxidized or the reduced substance^ but it is the measure of the ratio of the concentration. For example^, the E, of a system having the concentrations of [Ox] and [Red] of 100 and 10 molals respectively will be theoretically equal, to that of a system having [Ox] and [Redl concentrations of 10 and 1 respectively. When the molal concentrations of oxidant and reductant are equals the term E, becomes equal to E } regardless of concentration (26). 23 Although standard potentials are normally referred to the hydrogen electrode,, some other electrode is usually used as the reference since a potential is experimentally measured as the hydrogen electrode is somewhat inconvenient to use. Electrodes whose potential with reference to the hydrogen electrode are accurately known are normally chosen for their ease of pre- paration and reproducibility,. They usually consist of a system made of a metal, insoluble salt of metal., and a solution containing an ion in common with the salt. The most commonly used reference electrode is the calomel electrode. This electrode is essentially a mercury-mercurous ion electrode, in which the half-cell reaction can be represented as follows (ll) : Hg n Cl_ + 2e~ = 2H +2 Cl" °2 2 g Another reference electrode also commonly used is the silver- silver chloride reference half-cell consisting of a silver electrode (Ag) in contact with silver chloride (Ag Cl) . The silver and silver ions (Ag, Ag ) are the reduced and oxidized forms of this reference half cell. The silver-silver chloride electrode has an advantage over the calomel electrode in that it has greater stability at higher temperature ( 2*0 . All O.R.P. measurements which employ either a calomel or silver- silver chloride -as a reference can be converted to the potential with reference to hydrogen electrode by applying a correction factor. This can be expressed by the following equation (2*0. E, = E + voltage of reference electrode where K = The potential with reference to the hydrogen electrode. E = Observed voltage with calomel or silver-silver chloride. Potential of standard hydrogen electrode = 0.00 volts Potential of silver-silver chloride reference = 0.199 volts Potential of standard hydrogen electrode = 0,2*)-*+ volts 2h In practical applications., the conversion factor of the standard calomel is sometime taken as = 0.250 volts . 2» Factors Affecting Redox Potential. Measurements There are a number of conditions affecting the potential as well as the rate of change of potential. Generally., these influential factors are: 1, Oxidants and reductants in the system. All oxidants and reductants present and the tendency of the system to take or give off electrons. For example,? it has "been theoretically calculated that in a system where the redox potential is solely determine by the oxygen tension, a reduction in oxygen concentration from 10 gm/l to 0.1 mg/l will lower the potential "by 30 millivolts (48). 2. The pH of the system. The pH of the solution affects the activity of the ions involved. In addition,, through formation of hydroxide., it affects the precipitation of solid hydroxides. Bass Becking and his associates (6) has presented the relationship of redox potentials (E ) to pH for the equilibrium condition ass E , = E - 59( a / n )pH millivolts in which E = Standard oxidation potential. a = The number of protons, and n = The number of electrons taking part in the reaction. It can be seen that the gradient of the equilibrium line on an E, - pH diagram, is entirely determined by a/n. They also pointed out the fallacy in assumption which has been widely accepted that the slope is always -59 mv per pH unit by using the following illustrations as examples (6) Reactions Fe + 2H_S ^=±. Fe S + kiL + 2e 2 ^ 2 from which a/n = 2 and E, = -1^0 - 11.8 pH 25 Reaction: 3Fe +2 + k HgO ^=^ Fe 2 °4 + 8li+ + 2e ' From which, a/n = k and E h = +990 - 237 pH 3. The temperature of the system. The temperature of the solution influences the potential hut only slightly. The necessary correction will usually he less than 1 millivolt per degree centigrade. This value varies with the electrode "being used. However., temperature may exert the indirect effect significantly as it affects the oxidation-reduction reaction rate since^, in most processes,, reaction rates accelerate with increasing temperature . In connection with the factors discussed above,, the following factors have also been found to not only affect the potential measurement but also limits is applicability and accuracy as well. 1. The surface of the electrode can become desensitized by the effect so called "poisoning of electrode". A surface film of platinum oxides may be formed on the platinum electrodes when the electrodes are immersed in aqueous media which make the voltametric behavior differ dramatically from that of an unoxidize electrode (k? 48) . 2. The platinum surface can exhibit a memory effect by exposure to strong oxidant or reductants with subsequent residual influence (24). 3. The length of time allowed for stablized potential measure- ment. It has been often found that the equilibrium redox potential reading varies with time. It has been suggested that the time allowed for the readings to reach equilibrium, should be consistent. For example^ assuming that 10 minutes required to reach equilibrium in the initial sample, then all subsequent samples should he given the same time for equilibration. 26 Only the final stabilized reading (10 minute reading in this case) should be recorded, any and all intermediate figures serving only to show by their increments how dynarr.dc or static the system is.. 27 IV. EXPERIMENTAL EQUIPMENT AND PROCEDURES A. Analysis of Ferrous Iron 1. Preparation and Collection of Samples Samples used in determining ferrous iron were either made with distilled water or obtained from natural raw water. For distilled water samples, ferrous ammonium sulfate., Fe (NH. ) p S0. . 6 HO, was used to yield ferrous iron. Since the ferrous iron existing in this compound is very stable, the exact amount of ferrous iron then can be determined. The ferrous ammonium sulfate was weighed out to the closest 0.1 mg by the use of a Sartor ious balance. For example, if a solution of 0.10 mg per ml. of ferrous iron was required, 0.7022 gm. of ferrous ammonium sulfate was weighed out. The weighed amount of ferrous ammonium sulfate then was dissolved in 20 ml. *1 of cone. H p S0 r , and 50 ml» of distilled water. This solution was then diluted with distilled water to 1^000 ml. and mixed well. This solution may be called a "stock solution," and, in general, was prepared such that it would contain 0.1 mg of ferrous iron per 1.00 ml. The more dilute solutions were then prepared from this solution by pipetting varying volumes of the stock solution into a 1-liter volumetric flask and diluting to the mark with distilled, water. For standard ferric iron solution, ferric ammonium sulfate [Fe (lj, (SO.) 12H 0] was used and the preparation of the solution was similar to that of ferrous ammonium sulfate. For example, if 0.2 mg, of ferric iron per 1.00 ml. was required, I.7268 gm. of ferric ammonium sulfate would be' dissolved in 20 ml. cone. EL.S0,, and 50 ml. distilled water then 2 4 diluted to a liter. *1 All distilled water referred to here after is understood as "Iron-free distilled water." 28 For the distilled water iron solutions containing a mixture of ferrous-ferric iron, the ferrous ammonium, sulfate and ferric ammonium sulfate solutions which had been prepared as above were used by pipetting pro- portional amounts of the standard solutions into 1-liter volumetric flask accordingly and diluting with distilled water to the mark. For example,, if solution of 0.002 mg. ferrous - 0.002 mg. ferric iron per 1.00 ml., of sample was required, the solution was prepared by pipetting 20 ml. of the 0.1 mg- ferrous per 1.00 ml. standard ferrous solution and 10 ml. of the 0.1 mg-ferric per 1.00 ml. standard ferric solution into 1-liter volumetric flask. The mixture then was diluted to mark with distilled water and mixed. This solution then contained 0.002 mg. ferrous and 0.002 mg. ferric iron per 1.00 ml. Natural raw ground waters from various wells in Illinois were used as the natural sources of ferrous iron. Water samples were collected in the field,, using precautions not to aerate the water,? and transported to the laboratory in 18 gallon glass bottles. Since no constant temperature bath was available in transporting the water samples to the laboratory, a few degree rise in water temperature sometimes was observed. If the experiment could not be performed on. the same day on which the samples were collected,, the samples were stored over night in a refrigerator at about 10 C. 2. Method of Analysis Two methods for analysis of ferrous iron were used in this study: (l) the bathophenanthroline method which employed 4,7-diplenyl - 1,, 1,0 / \ *2 phenanthroline (C_k ELg N ). The procedure used was essentially that described by Lee and Stumm (3l) and (2) the ortho-phenanthroine method using *2 1, 10 phenanthroline monohydrate (C- Hn N . HO), *2 Both ort ho -phenanthroline and bathophenanthroline compounds were the products of the G. Frederich Smith Chemical Co.,, Columbus, Ohio. 29 a. Reagents for bathophenanthroline method; 1. Bathophenanthroline, 0.001 M. solutions. It was prepared by dissolving 0.0332 gm of bathophenanthroline in 50 ml of 95 percent ethyl alcohol and diluting to 100 ml. with distilled water. 2. Hydroxy lamine hydrochloride, 10 percent solution. The solution was prepared by dissolving 10 gm of hydroxylamine hydrochloride in 100 ml. of distilled water and stored in a glass stopper reagent bottle. 3. Sodium acetate, 10 percent solution. The solution was pre- pared by dissolving 10 gm. of sodium acetate in 100 ml of distilled water, ko Extracting alcohol. Reagent grade isoamyl alcohol was used. 5. Ethyl alcohol, 95 percent, reagent grade, 6. Standard iron solution. The standard iron solutions for ferrous and ferric iron were prepared from ferrous ammonium sulfate and ferric ammonium sulfate as previously described. The standard iron wire solution was also used as a reference. The standard iron wire solution was prepared by weighing out 0.2000 gm. of the well-polished "iron wire for standard! zin,g" and dissolved in 20 ml. 6n . H p S0, and diluted to the mark with distilled water. This stock solution then contained 0.20 mg iron per 1.00 ml. The more dilute solutions were then prepared from this stock solution by diluting with distilled water accordingly, b. Procedure for determining iron by bathophenanthroline; The sample (water being tested or standard iron solution) was pipetted into a 125 ml. separatory funnel. The usual volume used was 10 mlj however j, for very high concentration of iron, only 5 ml was used. A reagent blank consisting of 10 ml, of distilled water was also prepared. To each sample k ml. of the 10 percent sodium acetate solution was added. The pH of the solutions at this point were tested and found to be about k. An additional k ml of sodium acetate solution was used where the original, sample 30 of water had been acidified. Two ml. of hydroxylamine hydrochloride were added to each sample if the sample was "being tested for total iron. The hydroxylamine was omitted when the ferrous iron concentration was sought. 15 ml of 0.001 M. bathophenanthroline solution was then added, mixed, and followed "by 10 ml. of isoamyl alcohol. This mixture was shaken thoroughly. The liquids then were allowed to separate for at least 5 minutes after shaking . After the liquids had cleanly separated into two layers, the lower aqueous layer was drawn off and discarded. This isoamyl alcohol layer was drained into a 50 ml volumetric flask and the separatory funnel was then washed with about 5 nil of ethyl alcohol added from a pipette. This was done in such a manner that the upper stopper of the funnel and the walls of the funnel were uniformly washed at least twice by a film of alcohol as it drained from the top to the bottom into the volumetric flask. The solution was then diluted to the 50 ml. mark with ethyl alcohol and mixed by shaking. The intensity of the color was then measured and expressed as percent trans - *3 mittance by using a spectrophotometer at a wavelength of 533 mo- with a slit opening of 0.002 mm. and a light path of 1.0 em. From the values of percent transmittance obtained, the amount of iron present then can be calculated from the standard curve. The standard calibration curve was established by the use of the standard iron wire solution and the standard ferrous ammonium sulfate solution using about 10 different concentrations. The iron-bathophenanthroline color conforms to Beer's law, so that the plot of percent transmittance and concentration of iron on semi-logarithmic paper yielded a straight line. The ortho-phenanthrlline method employed was essentially that as given in the 10th and 11th Ediction of Standard Methods (2, 3), with some modifications . *3 Model DU, made by Beckman Instruments, Fullerton, California. 31 a. Reagents for orthophenanthroiine method: 1. Hydrochloric acid, cone. 2. Hydroxy lamine reagent. The solution was prepared by dissolving 10 gm, NH OH . HC1 in 100 ml. of distilled water. 3. Ammonium acetate buffer solution, 250 gnu of ammonium acetate (CH^COONH. ) were dissolved in 150 ml. distilled water, 700 ml. glacial acetic acid were added and the solution then was diluted to 1 liter with distilled water . 4. Sodium acetate buffer solution. The solution was prepared by dissolving 350 gm of sodium acetate (CH,C00Na) in 500 ml. distilled water and dilute to approximately 1 liter. The sodium acetate solution then was adjusted with distilled water such that 10,0 ml. of the solution when added to 2.0 ml. HC1 and 100 ml. distilled water, would yield the pH to a value between 3»2 and 3.3. 5. Ortho-phenanthroline solution. 1 gm. of 1, 10-phenanthroline monohydrate (C. Hg N . H 0) was dissolved in 1,000 ml. of approximately 80 C distilled water. The solution was kept away from the light in the dark brown bottle. 6. Standard iron solution,, For the calibration of a standard ferrous iron curve, the solutions were prepared from the freshly made standard ferrous ammonium sulfate stock solution. Both ferrous ammonium, sulfate and "iron wire for standardizing" stock solutions were used in the calibration of the standard total iron curve. b. Procedure of determining ferrous iron by ortho-phenanthir oline 1 Many modified methods for determining ferrous iron by the use of ortho-phenanthroline were investigated and included; 1, "Ortho-phenanthroline only": 10 ml. of the ortho-phenan- throline solution was placed in 100 ml. volumetric flask. The sample (water 32 'being tested or standard Iron solution) was pipetted into the flask and the solution diluted to the mark a The fixed sample then was filtered through *4 fine (Whatman No, 40) filter paper or 0.45|i membrane filter . 2, "Ortho-phenanthroline plus sodium acetate and HC1" : 2 ml. of cone. HC1 were placed in a 100 ml. volumetric flask,, and the sample being tested for iron was pipetted into the flask. 10 ml. of sodium acetate buffer solution and 10 ml. of ortho-phenanthroline solution were then added and the solution was diluted to the mark with distilled water, Jo "Ortho-phenanthroline plus sodium acetate": 10 ml. of sodium acetate buffer and 10 ml. of ortho-phenanthroline were placed in a 100 ml. volumetric flasks the sample being tested were pipetted into the flask and diluted to the mark with distilled water. 4. "Ortho-phenanthroline plus ammonium acetate" : 10 ml. of ammonium acetate buffer and 10 ml. volumetric flask, the sample being tested was then • pipetted into the flask and the mixture was diluted to the mark with distilled water, 5» "Acidified samples 111 ; Where the samples being tested had been acidified with hydrochloric or sulfuric acid to pH 1.0, the method as described under 4. (ortho-phenanthroline plus ammonium acetate) was used. The eolorimetric measurements for the methods described above usually were made against a pure distilled water blank. However, in some cases, the color and turbidity due to iron floes were quite apparent, each reading then was made against a "adjusted blank" in the case of both color and turbidity interferences, and against an "acidified adjusted blank" in the case of only the color interference. The "adjusted blank" was made up by using proportional volume of a sample added to distilled water. When *4 Cellulose nitrate membrane filter used in this study was manufactured by Millipore Filter Go, 35 the turbidity changed with samples ^ many "adjusted blanks" were prepared correspondingly for each sample. The "acidified adjusted blank." was prepared similarly to the "adjusted blank" but with a few drops of hydrochloric acid added. c„ Procedure for determining total iron by ortho-phenanthroline:: The procedure for determining total iron by ortho-phenanthroline was essentially identical to that described in the 11th Edition of Standard Methods (2). The sample was mixed thoroughly and 50 ml. (or smaller size if the iron concentration was high) were measured into a 125 ml- erienmeyer flask, 2 ml. for cone. HC1 and 1 ml for hydroxy lamine reagent were added with a few glass beads and the mixture then was heated to boiling. The sample was allowed to boil for about 5 minutes and was cooled to room temperatures. The solution was transferred to a 100 ml. volumetric flask, 10 ml. ammonium, acetate buffer solution and 10 ml. ortho-phenanthroline solution were added and the mixture was diluted to the 100 ml. mark with distilled water. The colorimetric measurement then was made against a pure distilled water blank. The intensity of color of both ferrous and total iron samples was measured as percent transmittance. The standard calibration curve for ferrous iron was established by the use of standard ferrous ammonium sulfate solution., while both standard iron wire and ferrous ammonium sulfate solutions were used, for the total iron standard curve. B. Redox Potential Measurement 1. Apparatus The redox potential of the water was measured with the use of a flow cell as shown in Fig. 2. The cylindrical cell was made of "Lucite", and had a diameter of approximately h inches,-, and was 5 inches in height. The flow cell was closed at the bottom with a large rubber plug. The top of 3^ Duplicate Platinum Electrodes — (Thimble Type) 4 M xl/4 M "Lucite , 'Tube 30" Leads to Beckman Model G pH Meter A Calomel Reference Electrode I/4" "Lucite" Cover ■ 3 /8 x, /|6 ,, Lucite" Tube Rubber Plug 4"0.D. Water Outlet (Free Discharge) SECTION A-A FIGURE 2 FLOW CELL DEVICE 35 the flow cell was covered with, a Lucite plate cemented to the cylindrical body at an angle of about ^5 as to facilitate the release of gas huhhl.es. The four openings at the top of the flow cell were stoppered with No. 5 rubber plugs, through which the thermometer and three electrodes were inserted into the cell. The thermometer was inserted through the hole at the lowest level while the calomel reference electrode was located at the highest level. The two mid-level holes were used for two duplicate platinum electrodes. The water was allowed to enter the flow cell through the inlet which was close to the bottom of the cell and was freely discharged through the outlet close to the top at a rate of about 0.5 to 1 gallon per minutes. The electrodes used in redox potential measurement consisted of two thimble type platinum electrodes and one fiber type calomel electrode. . The meter used for measuring redox potentials was a Model G pH meter manufactured by Beckman Instruments, Inc. The pH of water was measured separately in a beaker by a Beckman Instruments, Inc. Model N pH meter. 2. Procedure for Measuring Redox Potentials, After the flow cell had been arranged as shown in Fig. 2, the electrodes were connected to the meter and the water was allowed to enter the cell. The first reading was made at approximately 5 minutes after con- nection, when the needle of the meter became stable. Another reading was taken at about 5 minutes after the first reading. The first platinum electrode then was disconnected, and was replaced with the second platinum, electrode, and one reading again was made at about 5 minutes and another at 10 minutes after connection. All of the four readings taken with these two platinum electrodes should agree within 20 mv. It was found that the readings which were taken at the end of 10 minutes did agree within 20 mv. *5 Platinum electrode (thimble type) catalog part Wo. 39271,? Beckman Instrument, Inc. *6 Calomel electrode (fiber type) catalog part No, 39970, Beckman Instrument, Inc. 36 The reading was then recorded to the nearest 1.0 mv. If the readings did not give satisfactory agreement, both metal electrodes then were removed from the flow cell, and the platinum surfaces soaked in cone, sulfuric acid for about 30 seconds j, rinsed with distilled water , and thoroughly wiped with a soft tissue paper. Both electrodes then were inserted back into the flow cell and the whole process of measurement was repeated. After cleaning, the readings were usually improved and agreed quite well. The acidified redox potential measurements were made by lowering the pH of the water samples to pH 1.0 with cone, sulfuric acid, and the redox potential readings were obtained quite rapidly within 1-2 minutes since equilibrium, was reached in a much shorter time than that of the natural raw water. The redox potential measurement and other chemical analysis were made at various stages of the unit treatment process, which included raw, after aeration and "after filtration" or finished stages. At each stage two ferrous iron samples (one at the beginning and one at the end of redox potential measurement), one total iron sample and one dissolved oxygen sample were taken. Usually 25 or 50 ml. of samples were used for ferrous iron determination and was added directly to a 100 ml. volumetric flask which contained 10 ml. of ortho-phenanthroline solution only. The mixture was then diluted to the mark and filtered through fine (Whatman No. ko) filter paper. The dissolved oxygen determination was done in the field by the use of the Azide Modification of the Winkler Method as described in the 11th Edition of Standard Methods (2). Nitrate determinations were made in the field with the use of a test developed by Bray (8) known as "Nitrate Tests for Soils and Plant Tissues." The reagent consists of a mixture of powdered zine and dry citric acid, which are to reduce the nitrate to nitrite, and a standard indicator which produces pink color with nitrite. The color indicated can give the rough estimation of the nitrate existed. 37 The ranges of nitrate are divided into four classes according to the color prevailed, "high" for which the pink color appearance was very "bright, "medium" for moderate pink, "slight" for slight pink and nil for those which gave no pink color at all. The degree of increase in nitrate was obtained from the comparative values as described above. If the increase was from nil to medium or from slight to high, the increase in nitrate content would he re- ported as "large increase". The water temperature was also recorded. About one gallon of water at each stage was brought back to the laboratory for additional analysis which included total solids, total and calcium hardness, chloride and chemical oxygen demand. The methods of determinations in the laboratory follow those set forth in the 11th Edition of Standard Methods (2). Total Solids Residue on Evaporation Total Hardness EDTA Titration Method (Univer II ' as indicator) Calcium Hardness EDTA Titration Method (Calver 11 ! as indicator) Chloride , Mercuric Nitrate Method Chemical Oxygen Demand Dichromate Reflux Method (0.025 N Standard potassium dichromate was used, silver sulfate was added at 1 hr. after refluxing for complete chloride correction) *7 Both Univer-II and Calver — 11 compounds were the products of the Hach Chemical Company, Ames, Iowa. -f> V. EXPERIMENTAL RESULTS A. Analysis of Ferrous Iron Table 1 shows the results of determining iron "by modified methods of ortho-phenanthroline and bathophenanthroline using distilled water samples containing ferrous and ferric iron. Unless otherwise specified, the ortho- phenanthroline samples were compared with a distilled water blank, while the bathophenanthroline was read against a distilled water-reagent blank. The methods used are indicated at the end of the Table. Samples with varying ratios of ferrous to ferric iron were set up in order to determine and compare the accuracy of various methods of determining ferrous iron in the presence of ferric iron. The calculated amounts of ferrous and ferric iron were obtained by computation from the equivalent weight of ferrous and ferric ammonium sulfate standard solutions, respectively. Total iron is the sum of both the ferrous and ferric iron in the samples. The results in Table 1 are expressed in milligrams (mg), in terms of the amount present in the volume used for color development. The values of ferrous iron obtained using ortho-phenanthroline with the buffers, sodium acetate and ammonium acetate, or with the buffer plus hydrochloric acid are consistently higher than the calculated values, especially where samples have a low ratio of ferrous to ferric iron. The method which employed ortho-phenanthroline only was found to give results close to the calculated values and to those obtained by the bathophenanthroline method. However, when only ferrous iron was present, the results of all three methods were in generaly lower than the calculated values, but com- parable to one another. 39 H ■3 E-i a a 1 o in H CO o 1m u -CO OJ -$ d d t— vo ONVO i H tO d d 0\Q LIACOVO [-- t— -=h -* -3 CO N H tr\ 0AVQ OOOOHHHtO dddodddd OOOOOOOO -3 -? CO CO OJVO oo OOOOHHOJ^ OOOOOOOO NA + cd o •H -P OJ CO + Ph - CM on -=t" CM-* O H CM odd OlAO UA CM UA O H CM odd o o o H H H a o H o in SH CD Eh >i w ra ra d 3 ^ ^ Si H H H O ft ft ft CD si si •H -H H H O O -P -P Pi Si CD CD Xi 41 ft ft I I O 41 P H O H H O O 3 P O p 41 p o 3 si si p a a si Si si 03 Si xi 41 CD ft ft 41 1 ft o O O 41 41 41 p P P u U cd o o 43 H CM KA-=r pq si o u H H o3 P O EH 5 P CD S -d Sh 4 P CO -d 41 p H H Si O Si ?h O •H Sh •H H g3 H P o3 O P P O P O (D $1 CD •H Si H -H CD O H 3 U O o3 P !> si p g3 si -d si CD Si P 41 Cj ft 41 H i ft ?Q O O 41 41 H P P 03 U Ctf O O 43 gg kl The results of determining ferrous iron in natural waters by the three modified ortho-phenanthroline methods and the bathophenanthroline method are shown in Table 2. Five different water samples were collected at the Northern Illinois Water Corporation plant which serves Champaign and Urbana, Illinois. Samples were obtained from the inlet main (which is referred to as raw water) , during aeration, from the a reaction well, from the top of a filter after the water had been chlorinated, and from the filter effluent (finished water) . From, Table 2, it appears that the results using the "ortho- phenanthroline only" method of determining ferrous iron are in good agreement with those obtained by the bathophenanthroline method which is regarded as the reference method- The results for total iron as determined by both the ortho-phenanthroline (llth Ed. Standard Method ) and the bathophenanthroline show good agreement. Raw water from the Champaign-Urbana treatment plant (well Number 3) was brought to the laboratory and aerated. Ferrous iron determinations were made on samples taken at various time intervals following aeration. The results of these analyses are given In Table 3 along with a brief explanation of the various analytical methods at the end of the table. It can be seen that the results of determining iron by the acidified sample method give the highest values. The "ortho-phenanthroline plus ammonium acetate" method shows slightly lower values while the results obtained by the "ortho- phenanthroline only" method give the lowest values and show good agreement when compared to the bathophenanthroline method. Table k shows the results of ferrous iron determinations using aerated water treated in a similar manner but obtained from an abandoned well at Cisco, Illinois. In this case the concentration of ferrous iron in raw Metz Reference Rook University of Illinoisr B106 NC 208 I, Romine Stradt TlT*bnna Tll-in^-t^ c-i arte k2 OJ Q) H ■3 EH H a o d O H -P O EH pq H a d o H CO ^ OJ O u u CD H 03 -P CO O VD O H H O d OJ ON O H H O CO CO O o o o H CO O OJ OJ d o o H H •H El CO H d CO o o H d o o o o o CO o o o o H •H d CO O o OJ -4- _=*■ -4- CO t- OJ H O o OJ o H d H -P O 0) el •H rs ^--> -P 'd 3 ^ a3 d -p d O M CO d O d ,d -p O EH ^ rP> H ■3 EH a o u H H H -P O Eh bO a pq H bO ^-=1- d 03 d o 5-1 OJ 0) H d o -H -P cd !h 0) <; ^ d p a erf S °H EH O S3 3} LP CO -=J- OJ O OJ OJ H d o •H -P d) pq H OJ OJ H OJ OJ o H OJ ft OJ ft OJ o 0\ -=*- LP oo o ON O J- OJ | o o o OJ NA OJ H H CO O o o d £ vo o -3- LP o OJ vo co H o LP H H H CO o H o CO o OJ LP d d OJ CO lp R vo OJ VO 3 0\ H H H O O O LP LP o OJ o LP o VO OJ KA -=t" LP VO O- co d o u H 03 d o u u !> o o H H w w Pi P O O -P -P ■d tJ a tQ H d d H ti O P W d •H H O MM -p +5 OJ OJ d d 43 P P 1 1 -p <$ CtH cd CD U (D O a •H °H •H d H s H •H O o H 3 o H 5 o -P P a 3 d d d d p d •H d •H d d 43 d a d CD 4-" CD d 43 •H ,d o ft H 4=1 i 1 ft O W o d o 43 -p ■a p a .d p u d fH p d o H o •H 43 c P — £ H OJ pq d o H H d P O EH -d w p H H d o •H H d P O P O -d •d d d -p OJ OJ CQ bO d 3 U"\ ON CO lAt-rl LTNVO 0J H R *d a) d ^ 0) d o 1 •p co -d o3 co •\ -d OJ d H ^ co H f-lAKNrH H K> ON H t- OJ 1-3 w OJ H ""•v^ O O O O O o o -d d U -p y- S bO H H H O O ON CO CO t~- C— d CO OJ d u CO a <-\ r-\ r-\ r-\ H cq d to ^1 u 0) > •d »d +3 OJ o d 0) d •d OJ O VO CO O O K\-4" 0) d OJ g d s H HMD CO t- ONVO OJ > OJ ^1 3 0) d "^N^ O « O o o © ft 9 O U OJ •H • ■3 tj bO H O CO i t>-VO ir\-=J- -d- to ■d -p d -d ^t OJ a H H d 0) -d d c3 H •H o 0) X •H -p OJ 1 d H* m _ s 3 W ^1 OJ d •p bO '>> ~^ ">s CQ bO > CQ S 0) H H H d d bO d d d H •H CO o d OJ CO OJ t- CC ON O CO CO o o o P< cu bO -d Hf d I HIAW l>"CO t— t— O £"- o p d W d ,d H OJ OJ OJ ■a 05 =H a> o d d d d !h •d .d «1 •H •H •H cd °H 0) d -p 0) n=R- H o H H H H p OJ H Sh o o fd o o ,Q O qn 5h H ft 1 H LTN O LTN LTN LTNQ0 O M 0) X M a h ^ d ■P -w a H ^•KMAO t— VO ON H ON H •p •H -p -p ^ o O CQ Q H Q d d d SH no •p R o« o o t o * o 8 o3 i fi> CJ 1 d i 1 +3 H ON U"N O LTN LT\ lAtAO O irsco VO OJ H H H OJ OJ o H IO-3" H LTNVO O d """•Vw o o o o o o O o o hO OCOCOOOM3VO ITN_H^ _3- _=j- u d a H H d O o H H O O O CO o tO OJ ON Q CO VO t^N KN u d H MD CO CO H CO H u -p b3 , ***^^ p o o o o o a o o o a> o bO O CO t~~ t-~ -4" -=t- tO KN OJ OJ fe EH H a H 51 METHODS I Orthophenanthroline only read vs adjusted blanks o o — |0 min. reading © © — 30 min. reading • *■ — 6 hr. reading x * — 24 hr. reading 2 Bathophenanfhroline a a — I min. reading 30 60 90 120 TIME AFTER AERATI ON - minutes FIGURE 3 RELATIONSHIP OF FERROUS IRON REMAINING TO TIME IN AERATED DELANO GROUND WATER BY "ORTHOPHENANTHROLINE ONLY- READ VS ADJUSTED BLANKS 52 E z z < or o a: en Z> o THODS nthroline only-filtered reading at time indicated mln. reading 8 hr. reading 24 hr. reading anthroline min. reading 30 60 90 120 TIME AFTER AERATION-minutes FIGURE 4 RELATIONSHIP OF FERROUS IRON REMAINING WITH TIME N AERATED DELANO GROUND WATER BY tt ORTHOPHENANTHRO LINE ONLY -WITH FILTRATION 53 membrane-filtered just prior to the colorimetric measurement in Fig. k. Both Figs,, 3 and k show very similar results. The values of ferrous iron obtained at 10 minutes of both Figs. 3 and h are quite comparable to those obtained using the batho-phenanthroline method. It can be seen from both Figs. 3 and k that the values for ferrous iron which were obtained increased with time. The increase in the apparent amount of ferrous iron is greater as the amount of ferrous iron decreases. The results presented in Table 9 were also obtained from the same experiment as those shown in Figs. 3 and 4. However , the ferrous iron determinations were made by the "ortho-phenanthroline only" method. The fixed samples were filtered through membrane filters immediately after all reagents had been added. The indicated reading time refers to the period of time after filtration at which colorimetric measurements were made. The results show only very slight increases in the amounts of apparent ferrous iron as the time interval for color development increase. The values obtained at each time interval also show good agreement with those obtained by the "bathophenanthroline" method. The effect of time on bathophenanthroline color development with aerated Deland ground water is shown in Table 10. In contrast with the ortho-phenanthroline color development,, there is a slight decrease in ferrous iron as the time for color development increases. The accuracy of the "ortho-phenanthroline only" method., with filtration and a 10 minute color development time was compared with the "bathophenanthroline" method. A ferrous iron sample was prepared by adding a known concentration of ferrous ammonium sulfate stock solution to un- aerated ground water obtained from both Deland and Champaign, Illinois. Natural water was added to 10 ml. or less of the stock solution to make a 54 Table 9 EFFECTS OF TIME ON ORTHO-PKENANTHROLINE COLOR DEVELOPMENT WHEN FILTERED IMMEDIATELY Sample Ferrous Iron Present - mg/l Total Iron No. Bath. "Orthophenanthroline only" mg/l 10 30 6 2k min. rain. hrs. hrs, 1 3«36 3.16 3.22 3.24 3.22 4.80 2 3.36 3.^ 3.48 3.56 3.56 ^. ^5 3 2.87 2.96 2.92 2.96 2.96 4 2.30 2.46 2.44 2.44 2,46 4.80 5 L96 2.00 I.98 2.00 2.02 6 1.50 1.60 1.46 1.52 1.55 7 1.10 1.08 1.06 1.10 1.10 4.45 8 0.90 O.87 O.85 O.85 O.85 9 O.62 0.64 0.64 0.66 0.66 10 0.50 0.47 0.47 O.kQ 0.49 4.45 Bath: "bathophenanthroline Total Iron: ortho-phenanthroline (llth Ed. Standard- Method ) . 55 Table 10 EFFECTS OF TIME ON BA.THOPHENANTHROLINE COLOR DEVELOPMENT Sample Ferrous Iron (mg/l) Total Iron Reading Time m.g/1 10 18 kQ min. hrs. hrs* 1 3.32 3.17 3.15 kM 2 2.92 2.73 2.68 k.k5 3 1.96 1.7k 1.7k - k 1.3^ 1.16 1.18 h.Qo 5 0.94 0.85 O.80 - 6 0.66 0.60 0.60 4.8o 7 O.kQ o.kd O.kQ - 8 Oo36 Oo33 0.30 - 9 0o26 0.25 0.24 k.k H ■3 EH CVJ -d CO S-I d -d ■d § > -P H C0\ + bD o + S H rH ^-' o3 fs pel -d I O ti -P d -d +3 5-1 d cd (D d s 3 •H aJ !h -p — < o 01 ,d < O CVJ CM CVJ o CO UA O CO O co LT\ CO CVJ O CO OJ 1 OJ co OJ CO H CO KN CO LTN ltn v8 o vo o vo CVJ CO O H O ON ON CO ON CO On o o O O o 1 Q H OJ OJ O o o VO ON CO o\ ON I— ON co ON o- ON o D o O o H H H o o o o O VO £ LPs VO CO LTN O LPv -=J- LfN o o O o 1 o NA !<"\ tOv H H H VO O CO LTN VD vo VO O o O O KN KN KN H H H H OJ bO U 0) f> < d •H si -p •H > 03 u 5 H Pi O the "adjusted" values would be 170 mv. plus 2k mv. or l$k mv. It would be reported as 190 mv. to KN CD H ■3 EH O" o 03 CO M W O O O o o EH is; H TO H HH - KN CJ H H • ^^ d bO •H B P<4 Q o > • co -d" o -d > H O CO W •H O H • i d b.0 >> X • n w H d 8 d • O £h m <; H co P £ O co EH K H <> \ d bfj-H d o H fc CD co > d s -h I P>4 CD -P ■> 3 CD cm > CC co O « H * O EH d •H ^1 O H ,d o o OH W^^iONW ONCO -=t ITNVO H UN O CO VO CO t—co vo OJ H CO VO° UNVO ONHt^OWoJocOiAQcQt-O^HCOO H HH HHrlHWW-JiflAHlAOIH^ O H O0 K\60 H ONCO t~-V0 H CO UN H -3" -* ofuxiVO ONCO -=t" H ON KN D— KNcO H CO H O CM ON ON unoD O VO OJ f OV OH OJoJHHWr-IH^ai^Jj-WaiJ-IOCvlKN H O -=r O -^ -4-(M4-cO^)OVD(MJ-COOODO CO CM CO VO ONVD CO OJ -* H VO OJ KNVD G\ t— OJ VO CO O O O OJ t—-5F OrloiHoJHoJairlr)tAwi-4-4'(MH UAVO rO-d- CO H COGOOOJ-d--=|-H--*LJAHOHONOVOVOOJO N KNLrNKNKN OOHHOHH0JOHHO1 LTWO lt\ OJ On t> (A KN kn co OOOOOOOOOOOOOOOOOCOOOH VO UN LfNOOILTNOCOGOVOcOLrN. O-^O UAVO O LTNCO H KN CO H VO LTNVO OHr-CNcOVQcO-^O^-O-OJ LTN-d/ « . I „ . o » o . e . o „ « ° o „ * „ o o j OO HOOOOOOOOHHHHOOcOOO CO CO CO OJ OCMAOOCVlOlAOCOCOOCO O UN VO N-LfNCQ OWO-*H0J-^VD-4-H"N0J0N0\aiC0LrNC0 d.H0JH0JHWWHHKNW4-J^WoJtA0J-iv0 H o ooooooooot-000000000000 KACOOJCQLTNON. LCNCOH ° O OJ 0- OJ t— ONVD J-tf) Q rl 1A h-J IAIA4- (M-J-iAJ/CJfOWHCvlWHWHVO-? KN OJ O O KN OJ N~N UN 5 5$ VO J- KN-=± OOOOOOOOOOOOO ONO OJLrNLrNKNHHCOKNHAONON ^ KN CJ OJ (M Cvl KMA-J OJ H ^ CM oooooooo ON. ONVD VD VO O ON. ON H H H H H 0J H OJ * OOOOOOOOOOOOQ KN O OJ W W H CO L^CO VO IA W NN r-\ ^-\ r-\ H r-{ r-i K"N HHH ,->,0- OJ <$ rd * ,d Qj <; -P CO v_x O h> O > o S • > CD P P W en <: . to ^—^ -d pq d cd o i -p CD rd U FQ pq pq OJ * -p pq d^- o * OJ oj * . OJ * >> o a3 •-— W CD fd ^ ^ g <3j EH H O d d c3 H a o o o d o H O OJ • o <; -->. o o p o O CD W ^ d ^ ^ a d . , . CO >3 -H OPQLH^KO?? -' o H co H d OJ co ft X co O H OJ rA O r-\ <-\ <-\ CO O -=*■ CM OJ l^- O H ON H H f^O UN ON H OJVO -=± ONOO OJ CO OM>0 H VO CO O O d d d d vo co vo o H OJ H -5 d o o o -4-OVO00 00 OJ H t— HKNiAH -3- o o H O OVO OJ KNVO H CM CO CO -=i" LT\ O O O H dodo O O CO o ON LfN OJ NN d d d d ONCO co t— . H OJ OJ o OJ o o o o -=J- ON UN LPs LTN LTN -=t t— o o o o O- LTN O OJ Q O O O -3- ON OJ ON H I H • pq <; -P O to CO CD cO U Pm u O H Ph pq 6l oa CO 0) •H d o o H CU H & c3 EH t— H • \ Si S P4 i cu a u o O Pi •H P-4 W cu ft<: d K> CU . -P 5-i vd roa> 5-i <; -p •H ^ J, a3 O Eh CD a ■H !h O H O o £3 lAOWiAOOOOJOOOO^OOiAOWcOOiAOl K\ IX\ LT\ KA -3" LT\Lr\KALT\-4- irN-^-VO-H/VQOOVO-^l-VQ LT\ K> t— HHHHHHHHHHHHHHHHHHHHHH O LTNOOOOODOCOOOOOOJOOOOO H H HHHHHrHr-IHHr-IHH.-HHrHH.-IH 00(MOOOOWO"MAIA(MOCMCOOIA(M^OO K^_d-_d/ K\_3- n~\ ir\ OJ -=f KA rO\ na Lr\-=h ro* navq qj ir\ KA OJ MD HHrlr-{<-\r-\<-\<-ir-{r-it-{t-ir-lr-\r-{r-{r-\r\<-{r-\r-l<-l OiAOOOOOiAOOiAiAiAOlAOOOOOOm CO LTNJ 1 l^OJVOCO-^CO^-O LT\-=t- LT\H K\ CO VD LA H CO L^OJ o o CO LTN QUMAIAIAOOOO^OOOIAQOOIA -3 1 NA UA t>~-=," C"— \£> -3 1 LrNUAOJN^CTNUA-? HOO-* t— tv- C~— C — C— t — t- — t — t^- t — .C — t — t — t — t — t — t — t — t>- tr— OOOi^OOOLr\ir\OOOOOOOOLJAOOOO i^Lr\H^O\Lr\wai-d-Lr\c7\HOJOl>-^LAoocOLrN^ : o fvl t-1 fl t^lvo L^ L^ t^- t^- C^-VO* t^- C-^ l>-VO t^- C^-VD VD h^-!^- cOCOK!CQS2a!SWS2tLlSSSWK!SSScOCQ CO S iSgSSSSBStQCQSratQStQCOatfl t MS^giSSSSKiSSrassrasSfflSSSra -p o h> CO < CD -P M en 0) CO o en >->w O a5 -P pq d d -p >-P d 3 2 ratio d crt 41 d -p cu g a; o -P H a5 bfl cu d OJ OJ * OJ * -— -. >5 O *-f o o a3 d •H P>4 U Q co !H CU d cd Q OJ * O "c3 H H cu P» X c3 O o "-^ 5-1 O +3 O O CU CO ^ ri-d O ^ ^ t3 N^ H O L^- • • • • O O H rH O I O O O O O OOIAO G\VO LT\_^- r-\ <-\ r-\ <-\ OJ O OJ OJ VO -=? -3- tA r-\ M r-i r-1 OJ O O O Lf\-H7 -=|- K\ H r-\ r-\ r-t Lf\0 LT\0 t^- 1— ro>vo e « « t — t— t" — t— Q O O O 3- C— H O D— C— t— CO o o o o H KA O H t— t^-CO C— W 2 M 2 Mrawra W en en ^ pq <; 05 -p 6 5-. -P CQ CU u u o fr w 62 -* o •d CD O «5 cq •H \ id « Q • CD o en h • ""■"•-« d UO-H a p>4 i a w X • Q M t^--=J- OJ LPv ON VD KA KNVQ* CO KMM H (^ W OJ VO ON O H VO OJ CM f- CO H -3" LPv t^\ ON Ox O r°»co CO H N"N fOVO LPv o o « o e t— ONCO ON LPv o-^aioo • o « « » o o o o o CD 3 •H -P d o o kn H CO H cs5 Eh KN OJ H d s d • O fH U CD H O d eh cr; >-v d a h d o M U O o CD d -* OJ -4- OJ -=i- • a a • • O O H H H CMCMOt-O K^ LPv LPv OJ -=r OJ KN o -=t" N"N O CO CO MD en VO O 1 ® hP\VO OJ LPv r-i _=j- rr\ CO QOCQ4' O -? b- LPv O OOOOH r<~N o t— O t— VO O H O H O H H CO OJ O CO O ON O LPv H VO H lfNH-4- KN H a -H CD -P • W U d CD <^ PM is ffi d O « o o o o o OJ ONCO H CO t--=f OJ C— VD o o o o o CO LPv H ON t— OJ VO VO VO VD O O Q O Q ON LPvVO O -? OJ H H H O O hO H ^ !h •— ..-— ^ CD ^-% CD PQ U H O P ">? •s-^ — -■ h — « w d a? -h cu a O H O d CD ffl O 2 'H ^ -do u U O ' d SH d d d EH 1 W O •H O 0) ,....+ rH . «*5 HO^CM H •H H a5 d a P 3 o to ^d p (33 •H £ CO [S Ch • CD tfl o o o o o O CQ p w « ■d CD U bo o d o Cd CD ?H *H o CD ,d . O O O OJ O CD ,d d t * <* o o ,d P p • -H VO VO KA UA_^ P H O fe H H H H H aJ p o CQ CD OJ W £ CD P ?h • WOOIAO a5 o •\ d U J>3 «.» o • « * O -H CQ CX> -P CD j-vo rov_^-^|- 3 rQ p a3 a\ a o3 d O O LTN C? O KA !_r\ KA -HF o d O bO CD CD 5 a u a-* ?h a cs3 9 O « d tf-i • p -p w«\ « t— t-- C-— £-- t>- _, CM O H o5 p CD -d • •^v. bO CD p ■d CD O O bO d Sh bO p d p a •H P •H •H O ?H H bD c3 P OJ-* d CD CO dp . . -d ,d •H P O O O p II d ^ o -d •H SraSS i q a a ^ CD •H O H N"\ O CM «H O ft CD 'i^ -H -H -H CD p O CD >> II •H o5 >d to co co ,Q d ^ ^ ^i CD CD CD CD ^ £ ,Q p 4^ P P CD a? £3 !25 CQ CQ S25 co c3 d ci f> P « CD «H O O O o3 b .E} fi .H -H -ri p »h d ^.d p _ H -H U P a .. — -n PQ o h o ,a a5 O u X d aj -h cd a cd a a^ p •H O H O d CD i EH U o d S d W O g -H in |5 'do <: o R H CO ■3 C5 o o ft CO o pq PL. s H d •H -P d 0) -P O ft x o 'd o; a •H 03 bO ,d o o u CD !< 0) -p d 5h -P •H o pq\ < a bO co d a "d o d • ft 0) ft d d o •H 5h M (D bCd d o> d n ,d -h o - CM UA-Hr ^ooooia^oiao^u>ooo KS UAVO HOUALTNO-^^OLrNCMlA UA UA ,=J- • o o ooooooooooooooo + + + + + + + + + + + + + + + + oooo +oo H-OOOO +ooo + -=T O CM -=|- -=f ON LT\VO O^OJ-^MD-^f- CMONO OO CO ON CO t— VO t^--=J" t-OOMDCOMD C— COM3 t^- + + + + + + + + + + + + + + + + + H CO -=f UAVO VO H f^ OACO 4-HCAWOO^ riOO^OHAH^^OHOJiAONHVOfA HOOOOOOHO + + + + + + + UAQOOUAOOQQ CM -5 0\ UA UA 0\ H kakS oooodddod + + + + + I + + + oo+oo+o+o VO H CM UA H CO 0A H CO O\0\0 ua cO t— CO t—l>- H + + + + + + + + + COCO 1A-* 0-* Q CO CO UAOrHOt^-O-^OO OOOOOOOOOOOOOOOOO OOOHOOQOO OH^H-^W-4-O IAONCO lACO-^ f°* K"\VO t>- t>- co MD -3- vo I^cocoVQ kSiaiaiaco ua ka + + + + + + + + + + + + + + + + + H D— t— -4- f- _=h H ua cr\ t—C^-^VO UA f- f-- -=}" NA + ,+ + + + + + + ^fOvOLTNC30-=l-CMCMCMCMCOCMO UAVO Q O UA-* OMMWHOOOIlAlANCOrlO^O I 1 I I I i I I I I I I I I i 1 1 HC0CM-3-CMCMCMLr\ON H CM C---3- O H O CM On KMMO CVIHH-* HO 8 I 1 i I CM I 1 I 3 OQOQQOOOOQQQQQOO C-— kScO ^ IAo C— t— UA-? -3 KA Ra KA CM rH O ^lAQlWWWHHHHHHHHrldt- + + + + + + + + + + + + + + + + + o o o o o Q o o _. CTnVO lAKMAOJ o ff\0 LTNLr\LfALf\-^f-J--^- H tr\ + + + + + + + 4-H- 11 ,->;d'-» s~^ ^— >, ^-^ ^v<; w . v pq ft5 W P ^d «1 v_x o3 -_-• d P> d pq ^ p d ,o a) o f-i . > d a; d o 'd d r-l 0) 0) -P CO d d) 5h !>5 d 0) 54 a •H O 5^ pq co g o O O <; js :s w pq m Ph EH E-f o o o 03 o "d d d X o o "bOo' 5h' O < 0) P co d 54 cj 0) pq H-v-- H P •H CO O CJ a 5h 5-. CO o f-^l fo 0) H H pq > ^— ' CO d •d O S4 ro d ■d 3 d ■d w pq CO d 0) 5h CD d « ? a i i i i i i i i i i i i LT\ t— O O OHf^OH I I I I I I o . o + o e o o -H HiA^WHOOH d OOOiAOOOOOiAOiAOQiAOiAlAOO^O O OQOIAO^IAQ CO H -* LTNCVICOCOCO LTA-d/ LTN LT\VO VO K"NVO OJ -* H CO LT\ LTN VO r— VO Lf\ KN t— D— H KN CO C^- t— t— - t— t- t— t— C— t— t— C— t— t— t— t— C— t— VO t— D— t— E— t— t— t— t— t— C— t— C— CO coa.cocogcowacoscoSSSScocoWScogS a a a i W a W co a Ot^CVlH K>CM00Q0^ ONVO LTN ON-* O C— - OJ VO LTNcO H CO H t— VO COMDHOHL^C— COCVIONOJ-H/D— OOLfAcOCOO N"N H rH H OJ r-{ r-{ r-\ H fCN Lf\ ,-) OJ -* OJ H CO-* lA KMA KM-- ia OJ OOO O PJIAOHCO OlJ lAOH CO VO O LTNVO OOLTNO^^OOOOOVO OOOO O O VO O O LT\VO cOCV]Of^O-*COC5M^OCVJJ-KA(M ONVO OWOOWJ-HrlOONHWcOAoo ON-* -*" ON K"\ ON C— COMDCOCOMD-*COLr\ VD(M(MVDCOCQOJVDOO-=HAO-* OCvlOOHWCVlt-H t\ltc\4- H CvlVO O fO^vfloO^ O oHodooowoo^doooriiooodH CO N"N O CO -* H UAUAOJ-* KN o -* HVO OJ t- o OOOOOOOOOOOOOOOr-iOOOOO OOOHOOOOO oooooooooooooooooooooo AcOWmtA^HOGOOOHOOM^- C^-VO LTN OJ OJ ONC— -* I--VO ia**-**-? fAFNK>KN(\) (MOJWCvlCJCVlHHH oooooogoo LTN OJ HCO C— ON -* ON CO [>- t— t— VO VO LfNLfN-* OJ & o < OJ -p --— • co -p £h 03 d" O a) O d o h) a r 3 ^ rS o o CD O p> p d W K <; CO s PQ ^^ O "o ? o CO O d ^ -p O O O d d •H oqfq Pl-P d -h o 55 --^^--. ■v--' CQ d d o u p> d o d o ^-. CD O H H •H d s 76 © o >- _j CI) c < 1- '»_ 3 z o a UJ z> _J u. u_ UJ o x: 1 ® C o © o c o UJ _j CD h- < h- < o E CO CD Z> § I 5 LU or -J 0_ E st d o c o l_ > CD o> b e o c o V— <>- u_ • X o CO OJ ■*— *vT a> .. cu - <7 CD ^ ^-^ 1- UJ jyed value 5S than 0. b c E *t o i_ CD > Towns ian 0. n 0.2 E *t < • 2 CD CD ■4— b CD a> o ponding ) less th ) betwee b 1 1 1 o CD LU o o w CD £ .E CO O CD CD > o CO CD > O O N JZ • 4 a O X < m O ^^, ^-«, ^^ cc >*o a CD Jj*. , *""* o 'c X D => < «*— > — : OJ fEh < vS> vS> ^— »» O ^— ^ CD □ •4 *" J" O O o o m **°Z% **ZZ? D "T J o >- a a: _l < 3 < a o o CO UJ < ® Q_ _i CD ^ O u. o < 1- CD o LU o o 2E z _I 00 ■ < < ^^ x: z UJ Z LU ^-^ — h- o CO £2 CD o c CD »gg O 0_ o Q LU O o CD 3 .?Z Q n-gz or £ < "D -J Q CD < z o OJ CO h- < _) 2 s X> *~ LU < LL. Q CD o z O O Q_ ,JtJ 9: CD X >- o n CO x z o or LATIO ICAL o lu s 00 R TO CHE Q 2 Q ° ° c if) st rO OJ — ** I/6LU - l'Q'0'0) PUDLU9Q U96AXQ |D0jUJ9g0 J9|DM MDy CD 77 correlation between C.O.D. and redox potentials can be drawn. However t it is apparent that the plotted points of those plants which normally contained less than 0.20 mg/l of iron in the effluent seem to fall in a specific region of the plot. Those plotted values for finished waters which contained more than O.to mg/l iron were found in a separate region. The "Zone of Highly Treatable Quality" and the "Zone of Poorly Treatable Quality" were drawn arbitrarily to indicate these respective areas. The "Zone of Highly Treatable Quality 55 constitutes the area of C.O.D. from to 20 mg/l and where the redox potentials are above 150 mv. This implies that the treatment of those waters which have raw water C.O.D. less than 20 mg/l and raw water redox potentials above 150 mv, show high efficiency in iron removal by the conventional units used. The reverse is true in the other zone. There is one point in the "Zone of Highly Treatable Quality" where the iron remaining is indicated as "C". However, the value as obtained by the State of Illinois Department of Public Health in the month prior to this survey show the value of 0.08 mg/l iron remaining in the finished water. This plant also showed an increase in ferrous iron in the filter at the time this survey was made. Hence these might indicate that the poor quality effluent was more due to the plant operation rather than the poor water quality at the time this survey was made. 78 VI. DISCUSSION OF RESULTS A ° Analysis of Ferrous Iron Since 10 ml. of ortho-phenanthroline were used in both ferrous and ferric iron determination instead of 2 ml. as recommended in the 11th Edition of Standard Methods , it can "be assumed that the amount of ortho-phenanthroline present was in excess and did not limit color development. From the results shown in Table 1, it is evident that the bathophenanthroline method gives the values closest to the calculated values and was not significantly- affected by the ratio of ferrous to ferric iron. The presence of ferric iron did not interfere with the determination by bathophenanthroline, since the ferrous color complex was separated from the ferric iron upon extraction with isoamyl alcohol. Therefore, the bathophenanthroline method was re- garded as a reference in this study. In contract, it can be seen that for the ortho-phenanthroline method, the experimental values of ferrous iron increased with a decreasing ratio of ferrous to ferric iron. This is likely due to the conversion of ferric to ferrous iron complex in the presence of a high percentage of ferric, iron. Under this condition, the solution is not at equilibrium and the ferrous iron complex has a large stability constant as compared to the ferric iron complex (31). However, the mechanism of this reduction has not yet been fully explained. As the ratio of ferrous to ferric iron in the samples increased or when only ferrous iron existed, the experimental values by all methods of ortho-phenanthroline were quite comparable to the calculated values. It is understood that the conversion from ferric to ferrous iron might have been reduced. It is also evident that when samples were acidified and buffered or only buffered, the values of ferrous iron were consistently higher than both the calculated and bathophenanthroline values. It should be noted that lowering the pH of solutions by either 79 acidifying or buffering may help to increase the conversion of ferric to ferrous iron.,, especially where a high ratio of ferric to ferrous iron exists since the method of determination by "ortho-phenantbroline' only" show lower values and agree quite well with both the calculated and bathophenanthroline method. In the determination of total iron, by the ortho~phenan.throli.ne method (11th Edition » Standard Methods ) ., the values obtained as shown, in Table 1 are very comparable to the calculated values. Since all the iron was reduced to the form, of ferrous Iron by hydroxylamine hycrochloride and maintained in this stage by an excess amount of this reducing agent., the interference by ferric iron was eliminated. The values of total Iron., as obtained by ortho-phenanthroline and bathophenanthroline as shown in Table 2 also showed good agreement between the two methods. It may be concluded that the method cf determining total iron by ortho-phenanthroline as recommended in. the 11th Edition of Standard Methods is reliable. From, the results of the ferrous iron, determinations in natural water by the bathophenanthroline and various modifications of the ortho- phenanthroline methods as presented in Table 2 and Table 3.? It is obvious that the values obtained by the ortho-phenanthroline methods with either acidifying or buffering were higher than both the bathophenanthroline and the "ortho-phenanthroline only" methods. The differences seem to be greater at the lower ranges of ferrous iron. In. both cases,? the total, iron remained almost constant (with the exception, being the filter effluent shown in Table 2)., whereas the ferrous iron decreased gradually with time; therefore^ the ratios of ferric to ferrous iron, in the solutions increased with time. It then, may be concluded that the Interferences due to ferric iron increased at the higher ratio of ferric to ' ferrous iron. 80 Table k indicates results similar to those shown in Table 3, However, in this case natural water with a much higher iron concentration was used in the experiment . Results in Table k confirm the earlier statement that if the samples were acidified or buffered, the values of ferrous iron would be higher than both the "ortho-phenanthroline only" and the batho- phenanthroline methods. It is suspected that the methods of acidifying samples may disturb the states of iron in the solution and result in erroneously high values of ferrous iron. It is generally believe that the preservation of ferrous iron in samples can be done by acidifying the samples on the assumption that the presence of the acid tends to stabilize the original valence state of the metals(^-l) . From the results shown from Table 1 to Table k $ it is evident that the acidified methods give values consistently higher than the bathophenanthroline and the calculated values when both ferrous and ferric iron are present. Acidification of samples therefore may lead to errors in the apparent values of ferrous iron for the same reasons which have been pointed out. The ortho-phenanthroline method with an adjusted blank gave the lowest values of all the ortho-phenanthroline methods employed and closest to the bathophenanthroline. Though the volume of the original sample taken for each determination varied from 10 ml. to 25 ml.,, the effects -of turbidity due to the iron floe was quite prominent. By comparing • the results read against an acidified adjusted blank in which the iron floe was dissolved and the adjusted blanks it can be seen that the former showed consistently higher values. The differences were well over Img/l in most cases and can be attributed to the turbidity in the adjusted blanks. Hence, in the determination of ferrous iron in the presence of a high amount of precipitated iron, special care must be taken to compensate for the error which may be caused by the turbidity as well a.; the color- 81 Since it has "been noted that the time of color development was also a critical factor in the determination, a period of only about 10 minutes was allowed between the time the samples were fixed and the time color measurement was made so as to eliminate any error due to instability of color developed. It can be seen that the agreement between the method of "ortho- phenanthroline only" and the bathophenanthroline was greatly improved by using the shorter time for color development. Though the readings were made within 10 minutes to prevent any possible changes of color with time the "ortho-phenanthroline plus ammonium acetate" method still gave higher values than both the bathophenanthroline and the "ortho-phenanthroline only" method. In an attempt to improve the "ortho-phenanthroline only" method filtration was employed to eliminate the turbidity due to the iron floe so that distilled water could be used as a colorimetric blank instead of an adjusted blank which must be made up for each corresponding sample. From the results shown in Table 6. it is evident that the fil- tration method is very comparable to the adjusted blank method and the bathophenanthroline method. From the results shown in Table 7 it can be seen that filtration reduces the intensity of the color slightly and the results were lower. However s the reduction was not very significant as indicated by the percent transmittance. It appears that the reduction of the color due to filtration was well within the limit of experimental error and the benefit gained from filtration was well justified. From the results of the effect of time on ortho-phenanthroline color development shown in Table 8, it is evident that there was an increase in the intensity of the color development with time. It can be seen from the percent changes that the increase in color intensity became greater as the ferrous iron concentration in the samples decreased and the ferric 82 iron concentration increased. This result also may have been due to the conversion of ferric iron to ferrous iron which forms a colored complex with ortho-phenanthroline. However,, this increase in color was probably not •due to the increase in ferric ortho-phenanthroline complex., since at the wavelength employed ( 512mu.) the ferric iron, complex has very low absorbancy (21). It also was observed that the ortho-phenanthroline plus ammonium acetate method which has been recommended in the 11th Edition of Standard Methods (2) did not give a stable color as it was claimed. The results shown in Fig. 3 and Fig. k indicate that the ortho-phenanthroline color developed in the ferrous iron determination is not stable but changes with time. Figure k- clearly indicates that the changes in apparent amount of ferrous iron was comparatively small at the beginning where the ferric iron was relatively low and increased gradually as the ferric iron in the samples increased. The slight changes in. ortho-phenanthroline color,? developed within 30 minutes > can be noted in Fig. 3° However,, all values of ferrous iron which were obtained by reading at 10 and 30 minutes were very comparable to the bathophenanthroline values. From the results of Table 9j it is interesting to note that when the sample is filtered immediately after being fixed with ortho-phenanthroline. the developed color becomes quite stable. The contrast of results between Table 9 and Fig. k where filtration was delayed till prior to the colori- metric. readings > can. be clearly observed. In another similar set of sampleSj, the colored filtrates were permitted to stand for a period of 3 days. It was found that the increase in the indicated amount of ferrous iron was very slight. It appears that the ferric iron floe was removed by filtration; thereby eliminating the interference and resulting in a stable ortho-phenanthroline color. 33 The effect of time on bathophenanthroline color is shown in Table 10. In contrast to the ortho-phenanthroline color.? the "batho- phenanthroline color shows slight decreases in the apparent ferrous iron with time. In this case, since the ferric iron was eliminated upon extraction, the colored solutions which consisted largely of ferrous iron may have been oxidized to ferric iron in order to shift toward equilibrium condition which resulted in a fading of the color with time. However , the changes in color of bathophenanthroline were relatively small in comparison to those occurring with ortho-phenanthroline. The results shown in Table 11 indicate that the "ortho-phenan- throline only" method, followed by membrane filtration within 10 minutes after being fixed, is very comparable to the bathophenanthroline method. In addition, this experiment indicates that some organic and other inter- fering substances which might be present in raw water do not seem to show any interference with the ferrous iron determination, since the iron content of the mixture obtained experimentally agreed quite well with the corres- ponding calculated value which is the sum of the iron in the raw water and the standard solution. However, there is a question which still has to be resolved, as to whether the apparent ferrous iron as obtained in the raw water represents all the ferrous iron present. It is also possible that the ferrous iron which was revealed in the raw water might represent only that portion which, was not complexed with organic matter and other interfering substances. Lee and Stumm (31) recommended boiling the original samples with hydrochloric acid when determining ferrous iron in natural water by the bathophenanthroline method, in order to free the chelated or complexed iron. However, it is possible that boiling the sample with acid will change the original ratio of the ferrous -ferric iron in the sample. A further modification of the "ortho-phenanthroline only" method was employed by filtering samples through the Whatman No. ^40 filter paper and comparing the results with the membrane filter and the bathophenanthroline method. Filter paper was employed in order to see whether it could be used as a substitute for the membrane filter ^ since the use of paper would not require the use of a vacuum pump which is inconvenient to operate in the field, A period of about 30 minutes was allowed between the time for fixing the samples and the time at which readings were made in the case of ortho- phenanthroline in order to see whether there was any significant change in the stability of the samples when filtered through paper. The results can be compared with those which were obtained by filtering immediately through a membrane filter^ since it was previously observed that the color of the solution was stable after being filtered through a membrane filter. It is interesting to note that the method in which the samples were filtered prior to fixing consistently showed the lowest values among the three methods using Whatman filter paper; 'whereas the method in which the samples were filtered after 30 minutes showed the highest values; and the method in which the samples were filtered immediately after being fixed gave intermediate results. The method in which the samples were filtered prior to fixing also gave values very comparable to both the bathophenanthroline and the method using the membrane filter. It is possible that the method where filtering took place prior to fixing gave the best results because the ferric iron interference was eliminated before the reagents were added. However ^ in using Whatman filter paper a longer period of time is required for filtration and the possibility of ferrous iron being oxidized is in- creased. In another experiment^ a set of samples filtered through Whatman filter paper were kept for a period of two days under normal laboratory conditions and the colorimetric measurements were again determined. 85 It was found, that the color of the samples increased considerably,, Therefore, it can. he concluded that Whatman. Filter paper is not as effective as the membrane filter in. the removal of precipitated iron,, Bo Redox Potential Measurements It should be noted that in a survey study of this type, there are numerous factors in plant operation which vary from plant to plant that cannot be controlled. For example with regards to function efficiency, some plants have a raw water which is highly treatable but due to the failure in operation produces a poor quality water which results in poor overall efficiency. The organic and. mineral content of the water also varies from plant to plant. Until now, there are still no simple procedures available for the determination of the various constituents in natural well waters, hence a definite interpretation of the results of redox potential measurements which includes all the various factors is not possible. In this study, the chemical oxygen demand (C,Q,D,) of the various waters varied from 1 to 50 mg/l and, this was not reduced during treatment. It is possible that these CO, Da exerting substances may represent the amount of organic constituents which cause chelation and form complexes with iron, postulated by many in- vestigators. From the plot presented in Fig, 5.? it appears that raw waters having a high redox potential in general were low in G,0,D, and vice versa. Another factor which also affected, the results of this study was the lack of control at the sampling points in the plant. For example, in some plants the raw water could be collected directly from the well head; but in some cases raw water at this point was unobtainable and the raw water from storage was used instead. This water might not represent the actual characteristics of the raw water because of the possible addition of atmospheric oxygen and other substances to the water during storage, thereby changing the actual characteristics of the raw water. The sampling 86 points after aeration also varied from, plant to plant. In some plants the aerated water was available right after the aeration process, and in some plants the aerated water could not "be sampled until at the top of the filters In the latter case, the water usually had "been retained in a reaction tank for a period of about one hour., which would allow a greater degree of oxidation than if the water had been sampled immediately after aeration. In the experiments performed in the laboratory with natural and synthetic waters, it was found that both oxidized iron and redox potential increased gradually with time following aeration; therefore, sampling time after aeration must also be taken into consideration in order to compare the characteristics of the raw waters in their response to aeration. The redox potential measurements made In this study were quite reliable within a reasonable range of precision since the measurements were made by the use of two duplicate platinum electrodes and the values agreed within 20 mv. The electrodes were also standardized using a quinhydrone standard solution. The results are given in Appendix A, The maximum difference between the extreme values obtained by the two electrodes were within 20 mv and the maximum, differences from the standard values were within + 10 mv. In some cases when water contained a high gas concentration, the meter reading became sluggish and a reliable reading was difficult to obtain. Rinsing the electrodes in concentrated sulfuric acid greatly improved the readings in all cases. A correction factor for the raw water of 80 mv per pH unit as obtained by calculation from, the results of this survey was used instead of 60 mv per pH unit which is the widely accepted standard correction. Regardless of whatever factor is employed, the differences of the correction- was not greater than 10 mv in most cases. It is interesting to note that the average computed correction factors of both the aerated and the finished water was 60 mv per pH unit which is equal to the standard correction factor. University of Illinois B106 NCEL 208 N. Romine S1 rest p Urbana, Illinois I i I °7 However » these were averages from values ranging from ho to 86 mv in the case of aerated water, and from. 34 mv to 83 mv in the case of finished water. No doubt some values were the results of some experimental errros in the measurement. As Baas and others (6) indicate,, the correction of redox potential per unit charge in pH depends very largely on the species of iron present and ranging from 59 mv "to 237 mv according to their illustrations. The direct effect of dissolved oxygen alone on the redox potential reading was not significant as observed in both laboratory and field experiments. The treatment plant at Edwardsville^ for example,, the measurement of redox potential of aerated water was made almost immediately after aeration. The increase in redox potential was found to be only 30 mv whereas the increase in dissolved oxygen was 7-8 mg/l. Baas and his associates (6) had noted from the results of their experiments that the variation in the molecular oxygen content of water appeared to have no direct influence on the electrode potential^ except at very low oxygen tensions. The presence of small amounts of oxidizable matter^ both organic and inorganic wouid_, however 3 be reacted on by the molecular oxygen; this could poise the system at a lower E, , The effect of dissolved oxvgen on natural raw water hence can be explained similarly. It was not the purpose of this study to develop a definite quantitative relationship between the amount of" ferrous or ferric iron remaining and the redox potential. Theoretically this is not possible,* since redox potential only indicates the relative ratio of the amount of oxidants to reductants,, which in this case would include all reduced and oxidized organic and mineral substances besides iron. However.? it may be possible that redox potentials can be used empirically to indicate the ranges and oxidation stages of water in a manner similar to the application of redox potentials, in other fields. The changes in redox 88 potential may help to indicate the relative ranges of oxidation attained. From the results shown in Table 1^, the changes in redox potentials resulting from aeration varied from +370 to +70 mv. The changes in redox potentials from raw to aerated water, however, increased gradually with time following aeration, Since the sampling points varied from one plant to another plant, the elapsed time after aeration was not equal. Hence, the redox potentials of various plants with respect to their response to aeration should not "be compared due to the fact that the redox potentials change with time. An attempt was made to correlate the changes in redox potentials to the corresponding changes in ferrous and ferric iron at various stages of treatment. From the results shown in Table 1*4-, a general trend is observed since when the change in redox potential is high, the corresponding increase in the percent of oxidized iron or the percent decrease in ferrous iron is also high. It appears that a change in redox potential of +150 mv was a limit above which the increase in the oxidized iron (or the decrease in ferrous iron) generally would be above 60 percent., and below which the increase in the oxidized iron would be less than 6o percent. Weart and Margrave(50) suggested an increase of +100 mv as a limit for raw water in their response to aeration, above which iron removal would be satis- factory, and below which iron removal would not be satisfactory with con- ventional treatment. In this survey,, there was only one plant where the response to aeration as indicated by redox potential was less than +100 mv and the effluent was not satisfactory. There were many paints which showed increases in redox potential above +100 mv, and where the treatment still was not satisfactory. It would seem, however, that the response of water to aeration is not adequate to be used as a determining factor in the degree of success in treatment since the filter also plays a very important 89 role in the iron oxidation and removal processes. There were significant changes in redox potentials during filtration which must also be evaluated in determining the efficiency of the treatment., as can be seen from Table l6. Weart and Margrave (50) also noted that when aeration alone raised the redox potential of a water to a level approaching +2kk- mv (or mv by calomel electrode) , the plant would function satisfactorily if reasonable care was given to its operation. From the results of this survey as shown in Table ±3 } there were six plants with no chlorine which had an aerated water redox potential less than, or equal to 250 mv (or mv by calomel electrode) . Results at these plants were not satisfactory. Hence,, the results of this survey confirm the previous findings. For the group of waters., which were chlorinated,, there was a general, trend in that the increase in redox potential at every stage related closely to the increase in the residual chlorine irregardless of the changes in ferrous or ferric iron. This can be related to the fact that chlorine is a strong oxidizing agent. The changes in redox potentials from, raw to finished waters seemed to be a very good measure of the overall iron removal efficiency. It can be observed from Table 15 that a redox potential increase of +200 mv is a limit ; below which the treatment was not satisfactory. For thirteen plants where an increase in redox potential of less than 200 mv was observed,, there was only one case (Broadlands) in. which the total iron remaining in the effluent was less than 0.2 mg/lj and two cases (Walnut and Argenta) where the total iron remaining was lower than 0,k mg/l) whereas the other ten plants had greater than O.k mg/l of total iron remaining in the finished water. Hence 9 the measure of increase in. redox potential from the raw to the finished water was a better measure of the effectiveness of treatment than the change in redox potential from the raw to the aerated water. 90 However,, for those nine plants which no chlorine and where the increase in redox potential was greater than +200 mv, there were two (Roxana and Oakwood) whose effluent proved to he unsatisfactory. In both these plants the ratio of total iron to ferrous iron remaining was higher than in other plants within the same group. Since the redox potential is the measure of the ratio of the oxidants to the reductants,, the presence of a high ratio of total iron to ferrous iron might effect a high redox potential reading. However,, it is liekly that there was an improper function- ing of the filters involved since it is very unusual for a normal filter to have a concentration of total iron in the effluent greater than five times that of the ferrous iron concentration. In the case of Roxana, there appeared to he a "break through in the filter since the concentration of iron in raw water was unusually high (over 15 mg/l) . Because Oakwood was also having particular problems with its filters, these two plants should he regarded as exceptions. From the results shown in Table l6, no correlation can be drawn between the changes in redox potential across the filters and changes in other characteristics. However, it was noted that for all plants (in- cluding both those with and without chlorine) with a decrease in potential over -30 mv during filtration, the degree of success in the treatment was very poor. All plants which showed an increase in ferrous iron during filtration, are included in this group. It is interesting to note that in the case of La Moille, there was a large decrease in redox potential (-530 mv) across the filter which corresponded to the increase in the ferrous iron concentration and a depletion in residual chlorine. An increase of redox potentials resulting from filtration however, did not always accompany successful treatment. In almost all plants, with and without chlorine, there was a depletion of oxygen, during filtration 91 and a corresponding increase in the nitrate concentration. Weart and Margrave (50) have pointed out that such a depletion in dissolved oxygen during the passage of water through a filter was a sure indication of the existence of an extremely active biological flora and fauna. In what way and to what extent , this biological growth has effect on the filter is not yet understood. The reduction in oxygen across the filters for chlorinated water was in general less than that for unchlorinated water. This may have been due' to the fact that chlorine is a disinfectant which may have diminished the biological growth in the filters. Another question which is often raised concerns whether the dissolved oxygen introduced during treatment is a factor which limits the success of the treatment. Theoretically, only l.k mg/l of oxygen will be required to oxidize 10 mg/l of the ferrous iron. From this plant survey,, it can be seen that the dissolved oxygen in the aerated water in almost all plants exceeded 7 mg/l, whereas the ferrous iron was less than 5 mg/l in most cases and over 50 percent of this already had been converted to the ferric iron form during aeration. Hence,, it can be con- cluded that in the plants studied the dissolved oxygen was sufficient and was not a limiting factor. It is also obvious that the depletion of oxygen in the filter was not entirely due to the oxidation of iron. From the results shown in Table 17j it appears that" there is a general correlation between the redox potentials and the amount of ferrous and total iron remaining in water with no chlorine; at high redox potentials both the ferrous iron and total iron remaining were generally low and vice versa. It can be seen that for all finished waters which had a redox potential less than 270 mv treatment was not satisfactory since the iron remaining exceeded 0.4 mg/l; whereas for those finished waters with redox potentials higher than 270 mv which included thirteen plant s } nine exhibited a satisfactory finished water quality while four plants did not. Among 92 these four plants,, the percent of ferrous iron vas less than 30 percent. This limit of 270 mv can be very well applied empirically to most plants not using chlorine under normal operating conditions where the ferric iron is being effectively removed on the filters. For those plants using chlorine, there was no correlation between the redox potentials and either ferrous or ferric iron content. The redox potential however, does show a good correlation with the residual chlorine. From the plot of the redox potential against the C.O.D. of each raw water, it is apparent that the waters tested can be separated quite distinctly, into two groups with respect to their treatability for iron removal. The group of waters which contain less than 0.2 mg/l iron in the plant effluent fall in the "Zone of Highly Treatable Quality". This corres- ponds to the region of C.O.D. ranging from to 20 mg/l where the redox potential is above +150 mv. Those plants in which the effluent contained iron in quantities greater than O.^mg/l fall in the "Zone of Poorly Treatable Quality" 'which corresponds to the area of C.O.D. greater than 20 mg/l with the redox potential below 120 mg. It is possible that the redox potentials alone could be used to determine a water's respone to treatment in a manner similar to the application of redox potentials in determining the cofrosiveness of soils as suggested by Starkey and Wight (h r j). In the case of water, it can be concluded from the results of this study that a raw water which shows redox potentials higher than 150 mv will respond very well to treatment by the conventional methods, and those which show redox potentials less than 120 mv will not readily respond. If it is possible to predict the response of the raw water to treatment by this simple method, this measurement may be very useful in the treatment plant design. Proper adjustments can be made later within a plant to make it more efficient. 93 Since many aspects of water' treatment are similar to waste treatment,, it may "be possible that redox potential measurement can be applied to the treatment of water in the manner similar to that being used in the waste treatment. The similarity between the water treatment and the waste treatment is shown in Fig. 6, From the results obtained in this study , it can be; seen that the initial, redox potential of raw water was relatively low just as a raw waste. Upon aeration or oxidation by chlorine, the redox potential was increased. When water passed through a filter 3 the redox potential would either increase or decrease (as the results show in Table l6) . If there was a large drop in redox potential across the filter over a certain limit (-80 mv as indicated from the results of this investigation), it indicated an abnormal condition of the filter , perhaps indicating that the 'filter should be backwashed. If the redox potential of the effluent was lower than 270 mv, as observed from this study ^ it indicated a poor quality effluent. This is very similar to waste treatment where^ when the redox potential of the effluent decreases to about -100 mv. It is an indication of the start of unsatisfactory plant functioning. The appli- cation of redox potentials as an automatic control has been successfully incorporated into both industrial, and domestic waste treatment (25, h^ f ^-9) « It is possible that a similar technique can be applied to water treatment. The same limits may not be applicable in all plants, however, but by comparing the results of redox potential, measurement to the chemical analysis at a particular plant over a period of time^. an accurate and reliable limiting range of redox potentials can be established for that plant. One of the great advantages of water over a waste is that the characteristics of the well, water is normally more consistent and usually does not change over a period of years, whereas a waste is normally highly variable in its character- istics. CO CO UJ o o QC a. LJ < UJ or H- QC UJ H < CO CO UJ 1 ( o o cr OJ Q. y> 6 Report , ( January, 1961) . Ik. Fair and Geyer, " Water Supply and Waste Water Treatment , " John Wiley and Sons, Inc., New York (19 56) . 15. Fisher, H. S. and Carlson, R. E., "Automatic Continuous Manufacture of Hypochlorite Solutions", Paper Trade Jour., (May, 1956) . 16. Fortune, W. B. and Mellon, M. G., "Determination of Iron With 0- Phenanthroline", Ind. Eng. Chem. Anal. Ed. , V. 10, p, 60 (1938). 17. Fosnot, H. R., "Seven Methods for Iron Removal," Public Works, V. 86, No. 11, p. 81, (November, 1955). 101 BIBLIOGRAPHY (Continued) l8» Glas stone, So, Textbook of Physical Chemistry , D. Van Nostrand, Co., 2nd Edition Ind. Princeton, New Jersey, 19» Grune, W. N. and Lotze, To Ho, "Redox Potentials in Slude Digestion," Water and Sevage Works Jour. , V. 105.? No. 1, (January, 1958) • 20. Harvey, A. E., Jr., and Manning, D. L., "Spectrophotometry Studies of Empirical", Jour. Am. Chem. Soc , V. "Jk, p. h-^kk, (1952). 21. Harvey, A. E., Smart, J. A. a&d Amis, E. S. "Simultaneous Spectro- photometric Determination of Iron (II) and Total Iron with 1, 10 - Phenanthroline", Anal. Chem. .. V. 27> p. 26 (1955). 22. Hood, J. Wo and Rochlich, G. A., "Oxidation-Reduction Potentials in Sewage Treatment", Water and Sewage Works , V. 96, No. 352, (19^9). 23» Hutchinson, G. E., A Tretise on Limnology , V. 1, John Wiley and Sons, New York, (1957). 2^o Kehoe, T. J. and Jones, R. H., "Theory and Application of ORP Measurement in Waste Treatment Processes", Water and Sewage Works , Jour, ; , V. 107, No. 8, (August, i960). 25. Kelch, J. L., and Graham, A. K., "Electrometric System for Continuous Control of Reduction of Hexdvalent Chromium in Plant Waste/" Plating , V. 36, P. 1028 (19.49). " 26. Kennedy, R, H., "ORP Theory and Application", Industrial Wastes , V. 5, No. 2, (April, i960) . 27. Knapp, Wo Go, "Modified k, 7 - Diphenyl 1-10 - Phenanthroline Method Sensitive to 1 ppb of Iron in High Purity Water", Anal. Chem., V. 31, p. 141+5, (1959). 28. Kolthoff, I. M. and Belcher, R., Volumetric Analysis , Interscience Publisher, New York, (1957) « " 29. Kolthoff, I. M., Lee, T. S., and Leus sing, D. L,, "Equilibrium and Kinetic Studies on the Formation and Dissociation of Ferrion and Ferrin", Anal. Chem. , V. 20, p. 985/' (19^8) . 30. Latimer, W. H., Oxidation Potentials , Prentice-Hall, Englewood Cliffs, New Jersey, (2nd Ed., 1952). 31. Lee, G. F. and Stumm, W., "Determination of Ferrous Iron in the Presence of Ferric Iron with Bathophenanthroline", Jour. Am. W.W4A. , V. 52, p. 1567, (I960). 32. Lee, T. S., Kolthoff, I. M,, and Leussing, D. L, "Reaction of Ferrous and Ferric Iron - with 1, 10 - Phenanthroline. I. Dissociation Constant of Ferrous and Ferric Phenanthroline, Jour. Am. Chem. Soc , V. 70, p. 23^8, (19^8); II. Kinetics of Formation and Dissociation of Ferrous Phenanthroline s, Jour. Am. Chem. Soc , V. 70, p. 3596, (19^8). 102 BIBLIOGRAPHY (Continued) 33» Moss,, Mo Lo and Mellon, M. Go, "Color Reaction of 1,, 10 - Phenanthroline Derivatives", Ind, Eng. Chem., Analo Ed. ., V. Ik, p. 931, (19^2). 3h„ Nordell, E., "Iron and Manganese Removal", Water and Sewage Works , V. 100, No. 5, p. 181, (May, 1953). 35° Nussberger, F. E., "Application of Oxidation-Reduction Potentials to the Control of Sewage - Treatment Processes," Sewage Indo Waste , V. 25 <> No, 9, (Sept. 1953)o ~ """ "" """'" 36. Oborn, E. T., A Survey of Pertinent Biochemical Literature , Geological Survey Water Supply Pap er 1^59 - F. U. S. Government Printing Office, (i960). 37 • Pearsall, W. H., and Mortimer, C. H., "Oxidation Reduction Potentials in Waterlogged Soils, Natural Waters and Muds", Jour. Ecol. , V. 27, p. kB3, (193*0. """ "" 38. Pierce, R. S., "Oxidation-Redaction Potential and Specific Conductance of Ground Water: Their Influence on Natural Forrest Distribution", Proc. Soil Sci. Am. , V. 17, p. 6l, (1953) - 39» Pye, D. J., "The Control of Bleach Manufacture "by Oxidation Potentials", Jour, of Electro-Chemical Society . k0* Quispel, H., "Measurement of the Oxidation Reduction Potentials of Normal and Inundated Soils", Soil Sci. , V. 63, p. 265, (19V?). kl e Rainwater, F. Ho and Thatcher, L» L., Methods of Collection and Analysis of Water Samples , U. S. Geol. Survey Water Supply Papers No. 1^5^ (i960). k-2.- Reitz, L. K., O'Brien, A. So, and Davis, T. L., "Evaluation of Three Iron Methods Using a Factorial Experiment, Anal. Chem. V. 22, P. 1*4-70, (I950)o k-3° Sandell, E. B., Colorimetric Determination of Traces of Metals . Inter science Publishers, New York (3rd Ed., 1959). kk. Sawyer, Co N., Chemistry for Sanitary Engineering, McGraw-Hill Book Co., New York, (1960T7 45. Serta, I., "32 Cyanide Waste Treatment Hypochlorite Metal Finishing", Science for Electroplaters , (Jan}„ 1958). k-6. Smith, G. Eo, McCurdy, Wo Ho, and Diehl, H., "The Colorimetric Determination of Iron in Raw and Treated Muncipal Water Supplies by Use of ^ 7 - Diphenyl - 1, 10 Phenanthroline, Analyst , V. 77, p. klQ, (1952). 4-7. Starkey, R. L., and Wight, K. M., Anaerobic Corrosion of Iron in Soil , American. Gas Association, New York" (l9*+5) • 103 BIBLIOGRAPHY' (Continued) 48. Stumm, Werner, "Discussion of Application of ORP", Jour . Am . W . Wo A . , V. 53, No. 2, (Feb., 1961) . ^9° Walker, C. A., Eichenlaub, P. W., "Disposal of Electroplating Wastes by Oneida, Ltd.," 1 Sevage and Industrial Waste, (July, 195*0 • 50. Weart, J, G., and Margranve, G. E., "Oxidation-Reduction Potential Measurements Applied to Iron Removal", Jour. Am. W. W. A. , V. 4-9, No. 9j> (Sept. 1957). 104 APPENDIX A STANDARDIZATION OF REDOX POTENTIAL ELECTRODES WITH STANDARD QUINHYDRONE*! - BUFFER SOLUTIONS *2 Temperature range Extreme values of Standard Observed Potentials - mv. Potentials - mv, 25°C to 20°C 20°C 25°C pH 7»0 Buffer Solution * 5 53 to 38 47 41 pH 4, Buffer Solution 240 to 219 223 218 1 Quinhydrone - Fisher Reagent Chemical (Cat No. Q5), Fisher Scientific Co.^, Fair Lawn^ New Jersey 2 According to Kehoe (24), these readings are the ones most pH and millivolt meters indicated. 3 Reference Buffer solution produced by Leeds and Northrup Co. 4 Reference Buffer solution produced by Beckman Instruments, Inc. The above result shows the comparison of the potentials observed from the electrodes employed to the standard potentials in the quinhydrone- standard buffer solutions. The observed values shown are the extreme values which constitute the range of the observed values as obtained by various combinations of the platinum and calomel electrodes employed at various times throughout this study. The standard values are obtained from. a presentation by Kehoe (24),, in which he claimed that these values are the ones most pH and millivolt meters indicated and adopted as standard values. It can be seen that the electrodes employed gave quite good agreement with the standard values, even the extreme values lie within the range of allowable erros in the redox potential reading. 105 APPENDIX B COMPUTATION OF REDOX POTENTIAL CORRECTION FACTORS IN RELATION TO pH Raw Water Town Natural pH ORP mv. Acidified PH ORP mv. A pH A ORP mv. A ORP/A pH (mv./pH unit) Warrensberg 7.00 ito 1.00 690 6,0 550 92 Argenta 7.50 90 1.00 658 6.5 568 87 Cisco 7.00 120 1.00 650 6.0 530 89 Broadlands 7.50 160 1.00 680 6.5 520 80 Fai mount 7.35 130 1,00 660 6.35 530 814 Oakwood 7.50 110 1.00 650 6.5 5to 83 St. Joseph 7.50 150 1.00 650 6.5 500 77 Tolono 7.50 310 LOO 720 6.5 14-10 63 Areola 7.50 210 loOO 680 6.5 1470 73 Atwood 7.10 150 1.00 680 6.1 530 87 Bethany 7.20 70 1.00 630 6.2 565 91 Findlay 7.20 100 1.00 650 6.2 550 89 Winds on 7.00 130 1.00 720 6.0 590 93 Deland 6„70 130 1.00 610 6.7 I480 72 Wapella 6.85 90 1.00 570 6.85 I480 70 Heyworth 7-30 170 1.00 610 6.3 I4I4O 70 Roxana 6. 80 80 1.00 550 6.8 I470 69 Edwardsville 7oio 130 1.00 680 6.1 550 90 Marine 7. to 70 1.00 580 6.k 510 80 Mahomet 6.90 170 1.00 630 6,9 I460 67 Forrest 8.00 -100 1.00 650 7.0 750 108 Chatsworth 7.^5 _ 1.00 670 6.I45 - - Emden 7.10 110 1.00 620 6.1 510 8I4 Tremont 6.90 110 1.00 6kQ 5o9 530 90 Morton 7.30 70 1.00 630 6.3 560 89 Wyanet 7. to 90 1.00 6to 6.4 550 86 Walnut 7.25 170 1.00 650 6.25 I480 77 La Moille 7-35 130 1.00 610 6.35 1480 76 Danvers 7. to 50 1.00 550 6.1+0 500 78 Hudson 7o30 70 1.00 530 6.30 I460 73 El Paso 7.10 180 1.00 650 6.10 I470 77 Z 2kkk 30 Correction Factor = 2M4I4 30 = 80 mv, ./unit pH APPENDIX B (Continued) 106 4 derated Water Town Natural Acidi.fied A pH A ORP A 0RP/A pH pH ORP mv. pH ORP mv. mv. (mv./pH unit) Warrensberg 7.50 6kO 1.00 990 6. 50 350 5+ Argenta 7. 60 260'' LOO 690 6.60 1*30 65 Cisco 7.20 220* 1,00 700 6.20 1+S0 78 Broadlands 7.55 310* 1.06 .730 6.55 +20 6h Fairmount 7. ho - 1.00 690 6.1+0 - - Oakwood 7.80 3^ 1.00 720 6.80 380 56 St. Joseph ■ 7.50 490 1.00 750 6.50 260 1+0 Tolono 7.55 1+50 1.00 800 6.55 350 5+ Areola 7.60 610 ■ 1.00 9^0 6.60 330 50 Atwood „ - - - - - m Bethany 7.75 1+20' 1.00 760 6.75 31*0 50 Findlay 7.55 220 " ' 1.00 750 6.55 530 81 Windsor _ - - - - - _ Deland 7.30 290 1.00 690 6.30 1+00 61+ Wapella 7.^0 210 1.00 710 6/1*0 500 78 Heyvorth 7.80 280 1.00 660 6.80 380 56 Roxana 7.10 ^020 1.00 670 6.10 1+50 7+ Edwardsville 7.1+0 G150V ' 1.00 +700 6.1*0 550 86 Marine 7.65 +65o\ 1.00 +1050 6.65 1+00 60 Mahomet 7.35 +420 ■■ 1.00 +730. 6.35 310 ^9 Forrest 7.10 +390 1.00 +770 6.10 380 62 Chatsvorth 7.70 +350 1.00 +730 6.70 380 57 Emden 7.50 +220 1.00 +710 6.50 490 75 Tremont 7.1+0 +190 1.00 +710 6.1*0 520 81 Morton 7.70 +6l0 1.00 +900 6.70 290 +3 Wyanet 7.1+5 +260 " 1.00 +710 6. +5 1+50 70 Walnut 7- +5 j 1.00 +710 6.1+5 - - La Moille 7.90 '+^60 1.00 +970 6.90 1+10 60 Danvers 7.90 +290 1.00 +700 6.90 1+10 60 Hudson 7.60 +2^0 / 1.00 +670 6.60 1+30 65 El Paso 8,00 +660/ 1.00 +1090 7.00 1+30 2 n = 62 1691+ 28 ff\ s- * Correction factoi • = 1691+ 28 = 60 mv./pH unit 107 APPENDIX B (Continued) Finished Water Town Natural Acidified A pH A ORP A 0RP/A pH pH ORP mv. pH ORP mv. mv. (mv./pH unit) Warrens"berg 7.50 650 1.00 1020 6.50 370 57 Argenta 7.60 230 1.00 690 6.60 1*60 70 Cisco 7.10 210 1.00 650 6.10 1*1*0 72 Broadlan&s 7.60 250 1.00 630 6.60 380 58 Fairmount 7.50 350 1.00 770, 6.50 1*20 65 Oakwood 7.80 350 1.00 700| 6.80 350 52 St. Joseph 7.55 450 1.00 750 6.55 300 k6 Tolono 7.60 200 1.00 730 6.60 530 80 Areola 7.15 1*30 1.00 810 6.15 330 5^ Atwood 7. ho 500 1.00 730 6.to 230 36 Bethany 7.8o k05 1.00 750 6.80 31*5 51 Findlay 7.55 11*0 1.00 710 6.55 570 82 Windsor 7.25 220 1.00 720 6.25 500 80 Deland 7.35 250 1.00 630 6.35 380 60 Wapella 7.50 110 1.00 650 6.50 51*0 83 Heyworth 7.80 ^50 1.00 680 6.80 230 31* Roxana 7.10 280 1.00 670 6.10 390 6k Edwardsville 7.75 5**0 1.00 900 6.75 360 5^ Marine 7.6o 670 1.00 1050 6. 60 380 58 Mahom,et 7.20 430 1.00 730 6.20 300 1*9 Forrest 7.35 650 1.00 870 6.35 220 35 Chats worth 7.80 360 1.00 730 6.80 380 56 Emden 7.1*5 190 7.00 700 6.1*5 510 79 Tremont 7.55 270 1.00 690 6.55 1*20 61* Morton 7.70 550 1.00 800 6.70 250 37 Wyanet 7.50 290 1.00 720 6.50 1*30 66 Walnut 7.1*5 350 1.00 670 6,1*5 320 50 La Moille 8.30 200 1.00 680 7.30 1*30 66 Danvers 7.80 140 1.00 630 6.80 490 72 Hudson 7.70 680 1.00 1090 6.70 1*10 61 El Paso 7.50 710 1.00 1050 6.60 3k0 Z 52 18 1+3 n = 31 Correction factor = 181*3 7n 31 S 60 mv./pH unit APPENDIX C PAW WATER CHARACTERISTICS 108 0RP (E. u) Town Temp pH D.0. Ferrous Total v Corrected s-\ Iron Iron Natu- ral Acidi- fied to pH1.0 Corrected to pH 7.0 C.O.D. °C mg/l mg/l mg/l mg/l mv. mv. mv. mg/l Areola 16.5 7.50 0.4 5.02 5.08 +210 +680 +250 22.6 Argenta 13.5 7o50 0.1 1.20 I.36 +90 +658 +130 10.6 Atwood 14.2 7.10 2.58a- 2.60 +150 +680 +160 13.8 Bethany 15.0 7.20 2.09 2.12 +70p +630 +90 11.9 Broadlands I3o0 7.50 0.1 1.20 1.44 +160 +680 +200 18.1 Chatsworth 14,0 7.45 1.10 1.28 - +670 - 11.7 Cisco 14.0 7.00 0.3 4.10 4.24 +120 +65O +120 25.1 Danvers 13.8 7. to 2.28 2.80 +50 +550 +80 50.4 Deland 13.2 6.70 4,58 4.68 +130 +610 +110 48.8 Edwardsville 15.2 7.10 0.6 I.89 2.14 +130 +680 +l4o 2.8 El Paso 13.0 7.10 1.02 I.60 +180 +650 +190 10.4 Emden 13.5 7-10 2.60 2.96 +110 +620 +120 9.8 Fairmount 13.0 7.35 1.82 1.84 +103 +660 +l60 7.3 Findlay 15.2 7.20 - 4.45 4.72 +100 +650 +120 29.5 Forrest i4.o 8,00 21.4 26.0 -100 +650 -20 97+ Heyvorth 13.0 7-30 0.68 0.60 +170 +610 +190 1.0 Hudson 12.8 7.30 1.98 2.60 +70 +530 +90 36.2 La Moille 12.3 7.35 0.2 1.50 1.68 +130 +610 +160 12.9 Mahomet 14.0 6.90 0.1 2.00 2.20 +070 +630 +160 3.8 Marine 13.5 7.40 4.18 4.36 +70 +580 +100 37.0 Morton 14.0 7.30 0.8 2.78 3.00 +70 +630 +90 12.0 Oakwood 13.5 7.50 0.6 2.80 3.28 +110 +650 +150 22,9 Roxana 15.2 6.80 15.2 16.0 +80 +550 +60 43.6 St. Joseph i4.o 7.50 1.78 1.94 +150 +650 +190 9-1 Tremond 13.5 6,90 3.42 3.60 +110 +640 +100 12.1 Tolono 16.0 7.50 1,78 1.94 +150 +650 +190 9.1 Walnut 12.2 7.25 0.4 2.45 2.64 +170 +650 +190 8.8 Wapella 12.5 6.85 4.98 5.00 +90 +570 +80 12,5 Warrensberg 14,0 7.00 3.60 3.68 +140 +690 +140 28,1 Windsor 16.0 7.00 6,85 8.48 +130 +720 +130 49= 4 Wyanet 12.0 7.40 0.20 4.50 4,72 +90 +64o +120 17.8 APPENDIX C (Continued) PAW WATER CHARACTERISTICS 109 * *• -* * # Town Residue Hardne ss Alka- linity Chlo- ri de Sul- fate Nitr- ate Am- monia Treat- Ca Mg Tot, ability as CaCo-, mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l Areola k±3 203 I61 364 366 19 oO 33 nil 9^ C Argenta _ 183 122 305 418- 80 1.2 nil 6 A Atwood 470 207 130 358 468 2.0 nil -1.2 A Bethany _ 183 15 4 337 i+o6- 25 nil 5-2 B Broadlands 368 135 70 205 3^> 4.0 3.9 nil 1.7 A Chatsworth 730 237 201 m 346 250 slight 9.1 A Cisco 658 192 168 360 628 ,- 4.0 kA nil 11.4 C Dan vers 656 160 130 290 548 54.0 1.2 nil 11.2 r Deland 655 29^ 24-1 535 680 10 0.4 slight 36.0 C Edwardsville - 150 72 222 152 2.0 82 ' high A El Paso 44o 205 14-7 "352 ~ 422 «, nil 4.9 A Emden 379 2 40 1 c;q 340 386 3 slight 1.9 B Fairmount 521 230 160 390 292 25 86 nil 0.3 A Findlay 700 2V7 153 400 \ - 25 nil 1,8.0 C Forrest 480 180 176 356 u-68 6,0 175 slight- B Heyworth 351 _ - 336 28a 7 _ slight 5»2 A Hudson 644 230 195 425 624 0.4 nil 8.2 B La Moille 291 122 93 215 <264 1.0 1.0 slight C Mahomet - 255 205 46(3 3&7 15 96a slight 6.1 A Marine 571 163 111 274 500 2.0 0.6 slight 4,8 C Morton 6I+9 172 158 3.30 4l,4 15.0 2 slight 3.3 A Oakwood H3 178 127 305 354 13 _ nil 1.7 A Roxana 4-10' 600 300 900 200 19 106 nil 0.1 C St. Joseph 388 I96 129 325 372 4.0 1.0 nil 1.8 A Tremont 526 1.82 228 4l0 502 2.0 nil 1.0 A Tolono 388 196 129 325 372 4.0 1.0 nil 1.08 C Walnut 302 178 127 305 310 nil 1.2 A W ape 11a 4io 206 146 352 410 4.0 slight 2.0 C Warrensberg 527 170 l4d 310 422 8.0 nil 9.0 C Windsor - 217 180 397 630 7.0 11 slight 13 4 C Wyanet ^95 222 158 380 440/. 6.0 0- nil 2.8 c * Data shown in these columns was obtained, from "Mineral Examination of Water" data of Sanitary Engineering Laboratory., Department of Public Health, Springfield; Illinois. V*