Circular 65

STATE OF ILLINOIS WILLIAM G. STRATTON, Governor DEPARTMENT OF REGISTRATION AND EDUCATION VERA M. BINKS, Director

BY

T. E. LARSON

ILLINOIS STATE WATER SURVEY WILLIAM C ACKERMANN, Chief URBANA 1957

(89920—12-58) Printed by authority of the State of Iuinois REPRINTED FROM AND COPYRIGHTED AS A PART OF JOURNAL AMERICAN WATER WORKS ASSOCIATION Vol. 49, No. 12, December, 1957 Printed in U. S A

Evaluation of the Use of Polyphosphates in the Water Industry Thurston E. Larson A contribution to the Journal by Thurston E. Larson, Head, Chem• istry Section, Illinois State Water Survey, Urbana, Iu.

HE statements in this attempt to favorable data acquired in one appli• T separate fact from fancy in the cation are not necessarily applicable to sale and use of polyphosphates in water another. The first need in any applica• treatment derive from findings which tion is a statement of the end to be have been described in reliable scien• accomplished. tific literature. Several comments are contrary to the implications of sales Softening of Household Water representatives who lack either ade• Data are available to show that about quate training or the capacity to resist 5-10 ppm polyphosphate are required the temptation to oversell. Such defi• to counteract 1 ppm hardness. Thus, ciencies have resulted in misapplica• for example, water with a hardness tions which have not benefited the repu• 300 ppm will require an application of, tation of polphosphates in water 1,500-3,000 ppm of polyphosphate to treatment. soften it, or 6-12 tons per million gal• In a broad general sense, all poly• lons. On the other hand, considering phosphates show the same characteris• smaller quantities, 1 oz of polyphos• tics—with some notable exceptions phate would be required to soften 10 which are of minor consequence in this gal of water of 300 ppm hardness. discussion. These polyphosphates in• Thus, a household job such as dish• clude the crystalline pyrophosphate, washing represents a practical applica• metaphosphate, and tripolyphosphate, tion. Such softening is accomplished as well as a number of glassy polyphos• by a process called sequestration, or by phate blends. Polyphosphates behave the formation of soluble calcium poly• chemically in a manner quite different phosphate and magnesium polyphos• from the well known orthophosphates. phate complexes. In many ways the polyphosphates are This should not be confused with remarkable as a class of chemical what is normally considered as thresh• compounds. old treatment—a stabilizing or inhibit• Many claims have been made; some ing treatment which is designed for are adequately supported by data, but another purpose when treating mu• others are often misinterpretations or nicipal or industrial water supplies. inferences which cannot be substanti• Threshold treatment refers to the ap• ated by suitable records. In each case plication of polyphosphate in a concen• the circumstances of the application tration of about 2-4 ppm. This con• must be considered. Inferences from centration theoretically reduces the 1581 1582 THURSTON E. LARSON Jour. AW WA

hardness of the water by about 0.5 the point where this nondestructive ppm—for example, from 300 ppm to type of treatment can no longer be 299.5 ppm. applied with success. Polyphosphates are normally an inte• gral ingredient in dishwashing com• Red Water pounds for restaurants and household Rusty water may occur as the result use, as well as in many synthetic deter• of the natural presence of iron in the gents. A typical dishwashing ma• water pumped from ground water wells chine compound contains about 40 per or it may result from iron dissolved in cent polyphosphate, 40 per cent sodium the water as a product of corrosion. silicate, and 20 per cent soda ash. Iron in well water is normally present Thus, the use of polyphosphate for in the soluble (ferrous) state. The softening water for domestic purposes water is clear until, upon exposure to is practicable, but exceeds the bounds air or to chlorination, it is converted to of ridiculousness when it is implied the insoluble (ferric) state which re• that the same applies to municipal acts with water to form ferric hydrox• softening of hard-water supplies. ide, or rust. Polyphosphates have the property of Cleaning of Wells being able to combine with or seques• As cleaning and wetting agents, poly• ter soluble iron. They do not combine phosphates have achieved frequent no• with or sequester insoluble ferric hy• table success when used for cleaning droxide. In view of this property, the screened sand and gravel wells. When more effective use of polyphosphate confronted with a well problem, it is requires the application of this treat• important to be sure, first, that the ment to the water before it is exposed problem results neither from pump to air or before chlorination. Subse• inefficiency, lack of water in the aqui• quent oxidation by air or chlorine will fer, nor failures in the column pipe. convert the iron to the ferric form, but Where it has been determined that lack it remains dispersed until the polyphos• of water is due to clogged screen or a phate has lost its dispersing property by clogged water-bearing formation in the reversion to orthophosphate. immediate area of the screen, treatment The amount of polyphosphate re• with about 30 lb of polyphosphate per quired for this purpose is about 2-4 100 gal of water in the well bore has ppm per 1 ppm of iron in concentra• been found successful. Such treatment tions up to about 2-4 ppm. At higher is often accompanied by the use of 1-2 concentrations of iron, more than 9 lb of chlorine or chlorinated lime and, ppm polyphosphate is not always on a few occasions, has been in con• beneficial. junction with a wetting agent of the As it has been recognized that poly• non-ionic type. The practice of surg• phosphate is an excellent source of ing the well periodically during the phosphorous for the growth of bacteria, several hours following the application it is normally recommended that the of the treatment is recognized as highly concentrated solution of polyphosphate beneficial. For continued success, it is for injection into the water be treated desirable to repeat the application at with about 50 ppm chlorine each time scheduled intervals dictated by experi• the solution is prepared. This prevents ence so that incrustation does not reach the development of a breeding ground Dec. 1957 POLYPHOSPHATES 1583 for possible coliform organisms in the carbonate. Although either calcium polyphosphate treatment solution. or magnesium may be responsible for In certain areas where sufficient con• the scale, polyphosphate is effective centrations of ammonia, methane, or against calcium carbonate deposits both may be present in the well water, only; it has no effect on redissolving some iron removal water treatment or preventing the magnesium hydrox• plants are rendered ineffective by bac• ide deposits so frequently encountered terial growths in the filter bed. Such where lime softening plants are not growths deplete the oxygen which was operating properly. dissolved in the water by aeration for The first of four points where cal• the purpose of oxidizing the soluble iron cium carbonate deposits can be con• to the insoluble ferric hydroxide prior trolled is the sand filter. Lime soften• to filtration. In fact, even though chlo- ing plants without sufficient recarbo- rination is subsequently applied to the nation to prevent calcium carbonate water effluent, the water mains them• deposition on sand frequently acquire selves may be so infected with such an accumulation of calcium carbonate growths that chlorine is destroyed by on the sand grains. Continuous treat• the growths faster than the reverse can ment with polyphosphate in threshold occur. Subsequently, such insoluble concentrations (2 ppm) has frequently iron as may be deposited in the mains been reported to prevent such deposi• is redissolved as ferrous iron because tions. Magnesium hydroxide deposits of the anaerobic conditions created by which may occur at this point are no the bacterial growths. problem because they do not accumu• Chlorination before filtration has late on the sand grain and are easily proved to be successful in many cases. removed during backwashing. In persistent situations it may be de• Distribution systems constitute a sirable to use copper sulfate in con• second point of scale accumulation be• junction with polyphosphate for the cause lime softening plants with insuf• purpose of inhibiting the growth of ficient recarbonation may also deposit bacteria while the polyphosphate keeps a rather heavy calcium carbonate scale in soluble form such soluble iron as in the main as the water leaves the may be picked up from the mains. plant. Again, threshold treatment with The economics of such treatment is not polyphosphate will prevent this accu• prohibitive for small water plants. mulation at this point. However, such Normally, a treatment with 1 ppm cop• plants frequently also produce a water per (2.5 ppm anhydrous copper sul• at or near saturation with magnesium fate) is adequate when applied with hydroxide. Such lime-softening efflu• more than 2-4 ppm and less than 9 ents are usually high in magnesium ppm polyphosphate. and pH and low in calcium and alka• linity or both. Deposits of magnesium Scale Prevention hydroxide in the water main cannot be Scale is defined here as a deposit removed or prevented by the use of from hard water and does not refer to polyphosphates. There is no record either rust deposits or corrosion prod• in the literature to indicate the success• ucts. Threshold dosage of polyphos• ful use of polyphosphate for this phate in water has the unique power purpose. of preventing crystal growth of calcium 1584 THURSTON E. LARSON Jour. A WW A

Magnesium hydroxide or magnesium Another property of polyphosphate silicate, singly or mixed with calcium that should be considered lies in its carbonate or residual alum form a behavior when exposed to high tem• wavy, rippled deposit that has a very perature and to high pH. High tem• serious effect upon carrying capacity. perature and, to a lesser extent, high Records are available where the Hazen- pH in hot-water tanks and at the heating Williams C value has been reduced point, cause a significant reversion of from 120 to 90 or 80 with as little as polyphosphate to orthophosphate within a in. deposit of this rippled nature. a few hours. A number of overdosage Because polyphosphates will, not pre• instances are on record where the or• vent such a deposit and no chemical is thophosphate has reacted with calcium known for postcorrective treatment, the to form a calcium phosphate or calcium cure is one of prevention and not of hydroxyphosphate scale which actually treatment. did more harm than good. The rates A third point of scale accumulation for reversion for various polyphos• is in apartments, hotels, hospitals, and phates vary, but there is no record of similar places, where three kinds of any polyphosphate that possesses de• hardness can be considered : sirable useful properties which is not 1. Completely softened water which subject to partial or complete reversion will not deposit calcium or magnesium to orthophosphate within a matter of scales hours under high-temperature condi• 2. Unsoftened hard water which usu• tions. ally does not scale in cold water lines Magnesium hydroxide scale with or (At drinking fountains or faucets, without adsorbed silica in hot-water however, scale deposits may build up. tanks can easily be removed by flushing Polyphosphates should correct this with a hose. When such flushing of deposition of calcium carbonate for this highly heat-insulating deposit be• such hard waters. Scales in hot-water comes necessary on a weekly or monthly appliances and hot-water lines are usu• basis, the water user is justified in ally calcium carbonate. Treatment of complaining. this condition with polyphosphate in Scale is also frequently encountered threshold concentration should also be in cooling towers where polyphosphate of benefit if the retention time at the is frequently a part of the chemical heating point is not sufficient to revert treatment. Continuous treatment by the polyphosphate to orthophosphate.) application of polyphosphate alone is 3. Lime-softened water, which rarely usually not completely effective. Such produces heavy scales in cold water water use and accompanying treatment lines (In fact, some calcium carbo• is complicated by the rate of blowdown, nate scale is desired for corrosion pre• the pH of the water, the general water vention. Scales in hot-water lines are quality, the reversion rate of the poly• common, and if the scale is calcium phosphate, and the temperature of op• carbonate, polyphosphate can be of eration. The problem is frequently one benefit, particularly if applied in re• of treatment to prevent corrosion, scale, circulated systems. If, however, the and slime. Polyphosphate treatment scale is magnesium hydroxide, which is usually an adjunct to other treat• is highly insoluble in hot water, poly• ment or is supplemented by pH adjust• phosphate treatment is worthless). ment, by the addition of chromates, or Dec. 1957 POLYPHOSPHATES 1585 zinc salts (or both), by ferrocyanide polyphosphates, nor can prevention be or organics for corrosion control, and claimed where red water is avoided by even by chlorine or chlorinated phe• sequestration of this iron corrosion nols for slime control. product. Improvement but not pre• vention can be claimed in situations Corrosion where pitting-type corrosion is altered Corrosion is defined as the loss of to generalized corrosion. metal and may be evidenced by failure The effectiveness of polyphosphates of a structure, conduit, or container. is progressively greater at increasing In the case of ferrous metals, it may be turbulent velocities and at increasing detected by the appearance of iron in concentrations. Under essentially no the water from the solution of ferrous conditions are polyphosphates effective metal. It may also be detected by the in stagnant or nearly stagnant water, appearance of rust deposits on the such as in dead ends or service lines. water side of the metal. These deposits The only data available on its effective• may be uniform, as in general corro• ness at low velocities appear to indicate sion, or knobby growths, as in pitting a slight increase in corrosion at 0.5 fps. and tuberculation. It should be recognized that conditions The appearance of iron (red water) of 2-5 fps or more of turbulent flow as a result of corrosion may not be in velocities are not experienced continu• significant quantities at high velocities ally in all parts of almost any distri• because of dilution, but can be consid• bution system, and certainly not in erable at low velocities because of the communities where the grid has been longer contact time. On the other designed with 5-10 miles of 4- and 6-in. hand, the corrosion rate necessary to pipe for a 30,000 gpd consumption. induce structure failure may be greater Considerable embarrassment could be at high velocities but insignificant at avoided if vendors as well as super• low velocities. intendents would calculate the velocity of flow in 4-in. dead ends at normal The presence of iron in the water is domestic consumption rates by a lim• not necessarily an indication of corro• ited number of families. sion, because it may have been present in the water before contact with the In hot-water systems, the corrosion metal vessel. The appearance of a rates are usually greater than in cold. rusty scale is not always evidence of Recirculation improves the effective• corrosion because it may be a deposi• ness of polyphosphates or of any in• tion of hardness together with iron hibitor in hot-water systems, but, be• which was originally present in the cause of the low velocity, not in hot water. water tanks. It can, however, improve Polyphosphates have been reported the ability of the treatment to maintain to be effective in reducing corrosion by such iron as is dissolved by corrosion domestic waters. The reliable data in solution. show this effectiveness only under There is room for definition of water certain conditions and with certain quality as related to its corrosive prop• types of waters. Corrosion prevention erties. All waters are not equally cor• cannot be claimed, even though the red rosive and it should be recognized that water effects of natural iron in the wa• the quantity of polyphosphate required ter have been lessened by the use of may vary depending on the water 1586 THURSTON E. LARSON J our. AW W A quality as well as the specific conditions Nonmechanical removal of corro• of use. Any large-scale application and sion products and deposits other than most small-scale applications for cor• calcium carbonate has been proposed rosion control, warrant well planned and attempted, often with embarrassing and complete corrosion testing to es• results. It would appear that the pri• tablish data on effectiveness at the mary problem is that of defining the specific applications and velocities in• deposit to be removed and the quan• volved. tity. These unknowns are further complicated by a butcher's guess on the Cleaning of Water Mains quantity and method of application of polyphosphate. When it is recognized Reported experiences vary with re• that even the manner of action of poly• gard to the effectiveness and even the phosphate on the various unknown de• desirability of using polyphosphates posits or corrosion products is not for water main cleaning. The usual clearly established, nonmechanical wa• procedure consists of mechanical re• ter main cleaning with polyphosphates moval of bacterial slime, debris, mineral must be considered as nothing more deposits, or corrosion products. On nor less than a calculated risk. There occasions, little discussed in scientific is no established and proved plan or journals, polyphosphates or other chem• specification for nonmechanical removal icals have been used as an adjunct to of noncalcareous deposits. mechanical cleaning. No clear, reliable assessment of the value of this function is available, nor is it reasonable to ex• Conclusion pect a valid evaluation without more Polyphosphates are a truly remark• exact data than have been published able class of chemicals. Their effective• to date. ness for many purposes has been dem• The use of polyphosphates as an ad• onstrated in the laboratory and by junct to mechanical cleaning must be practice. Their ineffectiveness for followed by effective water treatment other purposes has also been established if the cleaning is to have a lasting ef• in the laboratory as well as by experi• fect. If the subsequent treatment is ence. It is unfortunate that no specific done with polyphosphates for corro• data are available for the cure-all mis• sion control, the limitations inherent in applications which have resulted in dis• low-velocity mains must be recognized. astrous failures and chaotic difficulties.