Introduction and Purpose of Demineralization

In many process boiler systems, zeolite softening of water is not sufficient to meet feed water quality standards. In these cases (typically high pressure boilers >1 psig!, the water is demineralized. "his process remo#es most of the mineral content of the raw water. Demineralizers depend upon both cationic and anionic e$change processes. "hey can be arranged in many different combinations, based upon incoming water contaminants and desired final water quality. %y the end of this chapter you should ha#e a basic understanding of what demineralization does, how these systems wor&, and the basic steps necessary to troubleshoot a system.

"heory of 'peration
Demineralization on a commercial scale became possible after resin manufacturers de#eloped methods of mass producing a new form of ion e$change resin. (hen regenerated with caustic ()a'*!, this resin remo#ed silica and carbon dio$ide from the effluent of a zeolite resin operated in the hydrogen cycle. + standard cation resin of the type commonly found in zeolite softeners, which is regenerated with sulfuric acid as opposed to salt (hence, hydrogen cycle cation!, is called a strong acid cation resin. +n anion resin with a strong affinity for silica and carbon dio$ide is described as a strong base anion resin. +n offshoot of the strong base anion resin de#elopment was the production of a chemically different resin with a strong affinity for strong-acid anions, sulfate (.'/!, chloride (0I!, nitrate ()'1!, phosphate (P'/!, etc. "his was called a weak base resin. "he wea& base resin does not remo#e silica or carbon dio$ide, as does the strong base resin. "herefore, it cannot be used alone to produce the water purity demanded by high-pressure steam plant operations. 2ess caustic is required to regenerate the wea& base resin and it is often coupled with a strong base resin in the same #essel to produce the most economical anion unit operation. "he wea& base resin is always loaded into the #essel last when the unit is readied for downflow ser#ice. Putting the wea& base resin on top of the anion resin bed uses its higher capacity for strong acid remo#al to insure that the strong base resin can concentrate on remo#ing the wea& acids associated with the carbon dio$ide and silica (carbon and silicic acids!. "he cation and anion #essels of a demineralizer train are piped in series so that the effluent of the cation unit enters the top of the anion unit and the discharge from the anion unit goes to the treated water header or into a polishing unit. + discussion of what ta&es place as the raw water is processed through the #arious units in a demineralizer may help to e$plain the resins and their function.

Strong Acid Cation

+s mentioned pre#iously, the strong acid cation unit contains a zeolite resin that is regenerated with sulfuric acid (*0I can also be used, but it is more e$pensi#e than * 3.'/!. +s the untreated water passes through the strong acid cation unit, the hydrogen ions that occupy the e$change sites in the resin are replaced by 0a, 4g, 5e, etc., ions. "he companion ions of the cations remo#ed-0'1, .'/, 0l, ) 1, P'/, etc.-pass through the resin and lin& up with the re6ected hydrogen to form the strong acids *30'1, *3.'/, *0I, *)'1, *1P'/, etc. 'b#iously, the effluent from a strong acid cation unit is #ery acidic-, it often has a 3. -1. p*. 5or this reason, the cation #essel and the interconnecting piping and #al#es are lined with rubber. "he order of ionic preference for a strong acid resin is, 1. *ydrogen

3. 0alcium 1. 4agnesium /. Potassium 7. .odium +s one might e$pect, sodium, the least preferred cation, is also the most wea&ly bound. +s a strong acid cation unit approaches the limit of its capacity, the ions shown in the list abo#e begin to lea& through in re#erse order, i.e., sodium will lea& first, followed in order by potassium, magnesium and calcium. In fact, a strong acid cation resin8s affinity for sodium is so low that some sodium will always lea& through, e#en when the resin is freshly regenerated. In actual operation, if a cation unit is to be run to near its 9brea&9 point for economic reasons, a parameter &nown as free mineral acidity is monitored to determine when e$haustion is approaching. 5ree mineral acidity, or 54+, is present when the water p* is less than /.1 (the methyl purple or methyl orange end point in the total al&alinity test!. %y definition, 54+ : the sum of the sulfuric, nitrate, phosphoric and hydrochloric acid in the water sample. "he 54+ of the effluent from a strong acid cation unit, therefore, is proportional to the le#el of the total e$changeable cations in the raw water. 4onitoring the strength of the acid in the cation effluent by 54+ analysis is therefore a good indicator of the performance of the unit. +s the rate of cation e$change decreases due to the decrease in a#ailable e$change sites in the resin, the amount of acid in the effluent (54+! decreases. "herefore, decreasing 54+ heralds the end of the ser#ice cycle in a strong acid cation unit. 0ation units are not run to e$haustion because of the need to double regenerate them to get bac& the total capacity and because high-pressure systems using this water could not tolerate the hardness associated with the end of the run. In practice, most demineralizer trains are designed to produce the ma$imum amount of water per desired ser#ice cycle. In#ariably, the limiting resin #olume is that of the anion unit. .trong base resins, e#en when preceded by a wea& base resin, can process less water than a strong acid resin. 5requently, the anion #essel is designed to accommodate the resin #olume necessary to treat the desired amount of water and the strong acid cation unit dimensions are duplicated from the anion unit design. "his, in effect, insures that when a train brea&s, it will brea& on silica first. In a high-pressure system, silica intrusion is more easily handled than hardness intrusion. + boiler system can function properly or cease operation as a result of the quality of demineralized water that is used for ma&eup. ;#en the best internal treatment programs ha#e their limits. "his is why a thorough understanding of the owner8s demineralizer system is so important to the water treatment consultant. (ith e$cellent feed water, e#en a mediocre water treatment program can be made to wor& in a high pressure system. 0on#ersely, e#en the best chemicals and the most carefully thought-out treatment program can fail miserably when the demineralizer system cannot be counted on to deli#er a quality feed water.

Dealkalization
'ne of the troublesome constituents in water used as a source of ma&eup for boiler systems is al&alinity. "he so-called total al&alinity of a water is the sum of the 0' 1 and '*< found in that solution. "he 0'1 portion of the total al&alinity is especially troublesome. +s raw water is processed through a strong cation unit in a demineralizer train, the 0a or 4g normally associated with the 0'1 is e$changed for *< and the cation effluent contains *30'1. "his acid, called carbonic acid, is #ery unstable. It disassociates into carbon dio$ide (0'3! and water #ery rapidly. 5or this reason, many strong acid cation units are followed by a mechanical de#ice called a decarbonator. + decarbonator is nothing more than a #essel filled with pall or Raschig rings supported on a grid o#er a plenum. + fan blows atmospheric air up

through the fill and out a #ent at the top of the tan& or #essel. 0arbon dio$ide, which brea&s out of the strong acid cation effluent stream, e$ists at the top. "he effluent from the decarbonator normally contains 1 ppm 0'3. Decarbonization of the strong acid cation effluent can also be accomplished by passing the acidic solution through a strong base anion resin that has been regenerated with caustic. + more complete discussion of this process follows.

Weak Acid Cation

0ertain waters that contain a high percentage of hardness associated with al&alinity can be economically treated by passage through a wea& acid cation resin. %y definition, the wea& acid resin will remo#e 0a<<, 4g<<, and )a< which enters the bicarbonate (*0'1!- form. %ecause most industrial water sources contain some noncarbonate hardness (0a.' /, etc.!, it is necessary to follow the wea& acid cation unit with a strong acid cation unit to achie#e truly demineralized water. (ea& and strong acid cation resins can be placed in different #essels or they can be placed in two distinct layers in the same #essel. "he regeneration efficiency of a wea& acid resin is #ery high compared to that of a strong acid resin. "herefore, it is possible to utilize the regenerant acid stream from the strong acid unit to regenerate the wea& acid unit. (hen wea& and strong acid cation resins are loaded into the same #essel, the strong acid resin settles on the bottom of the unit after bac&wash because of the density difference between the two resins. %ecause the wea& acid resin contains some strong acid sites, after regeneration with sulfuric acid, a 1 = brine solution must be passed through the unit. "he brine solution e$hausts any strong acid sites in the wea& acid resin and regenerates the strong acid resin in the sodium form. If this is not done when raw water enters the wea& acid resin, noncarbonate hardness e$changes at the strong acid sites. 54+ e$its the wea& acid resin and pre#ents the e$change of residual noncarbonate hardness in the strong acid resin. )ormally, a wea& acid resin produces 54+ for / -> = of its ser#ice cycle. "his combination would not be suitable for higher pressure boiler applications because of the presence of e$cess sodium in the effluent from the sodium-form strong acid resin. ?egeneration of a wea& acid cation resin with sulfuric acid must be carefully monitored to insure that the acid concentration during the regeneration does not e$ceed .@=. *igher concentrations of sulfuric acid can react with the 0a<< in the e$change sites of the e$hausted resin and result in the precipitation of calcium sulfate (0a.' /!. 0alcium sulfate, or gypsum, is insoluble e#en in the concentrated form of many acids. 'ften, mechanical remo#al is the only satisfactory way to rid the resin of this contaminant. 5rom an operational standpoint, it is ob6ectionable because it produces a pressure drop across the unit.

Strong Base Anion

+ strong base anion resin remo#es the anions associated with the wea& acids in the effluent from a strong acid cation unit. .pecifically, strong base anions remo#e silica (from * 3.i'1! and carbon dio$ide from *30'1 (carbonic acid!. In addition, the strong base anion resin remo#es anions associated with the strong acids in the strong acid cation effluent. .ulfate (from *3.'/!, chloride (from *0I!, and other anions are remo#ed. "he e$changeable anion is '*-, which is pic&ed up by the strong base anion resin during regeneration with caustic ()a'*!. "he cation in the water passing through unmodified (hydrogen ion is the cation! combines with the '*- liberated during the e$change and lea#es the anion unit as molecules of water (*3 !. .ince pure water is nonconducti#e, the conducti#ity of the effluent from a strong base anion unit is #ery low, about 1.A micromhos. .odium slippage through the strong acid cation unit combines with the liberated '*- in the anion effluent as )a'*. %ecause )a'* is more conducti#e than *3 , the sodium slippage through the cation increases as does the conducti#ity of the anion effluent.

'ne helpful tip to remember is that if strong acid cation resin e#er gets into the anion unit (due to a bad strainer in the cation #essel bottom!, it will wind up on the bottom of the anion unit because of its higher density, and will pic& up sodium during the anion regeneration cycle. (hen the anion is put bac& in the ser#ice cycle, sodium will lea& from the cation resin into the anion effluent and gi#e a higher than normal anion effluent conducti#ity. "his problem can be #erified by ha#ing samples of cation effluent and anion effluent analyzed for sodium by atomic absorption. +ny significant difference between the sodium le#els in and out of the cation resin means that sodium has migrated into the anion #essel. "his situation can be #ery troublesome, because most high-pressure boiler water treatment programs are ad#ersely affected by sodium intrusion. .trong base anion resins are quoted as ha#ing two capacities, salt-splitting capacity and total capacity. "he so-called salt-splitting capacity of a strong base anion resin is a measure of its ability to remo#e silica and carbon dio$ide. "he total capacity is a combination of the saltsplitting capacity and the resin8s capacity to remo#e .'/ -, 0l-, )'1- etc. .trong base anion resins also absorb organics from the water going through them. 4ost source waters contain some organic compounds. .urface waters are notorious for their propensity to carry a combination of iron and organic compounds. "hese compounds pass through the cation unit because the iron is not in the form of free ions. (hen it reaches the anion, on the other hand, the anion resin absorbs the compound and, by analysis, appears to be fouled with iron. (hat is really there is the ironBorganic comple$. "his material is not remo#ed during regeneration it occupies e$change sites normally reser#ed for other anions. "he final result is a reduction in the capacity of the anion train. +nion resins also remo#e other forms of organics during regeneration. "he organics that are remo#ed not only occupy #aluable e$change sites but also react with the resin to alter its character. "his alteration manifests itself in the transformation of the salt-splitting capacity of the resin to strong acid capacity. +s such, the anion resin begins to lose its ability to remo#e silica and carbon dio$ide but shows an increase in its capacity to remo#e the strong acid anions. "he total capacity of the resin does not decrease, but its salt-splitting capacity does. "he end result in a system pro#iding water for a high-pressure boiler will be shorter and shorter run lengths, because the silica le#el in the effluent of the anion unit is the main control parameter. 4icrobiological fouling can also occur in the anion unit. 5ungi grow nicely in an anion unit. "reatment with a solution of formaldehyde is one method of sterilizing an anion train.

Weak Base Anion

+ wea& base anion resin does not ha#e the ability to remo#e wea& acid anions (silica and carbon dio$ide!. It does, howe#er, ha#e a high capacity (&gBft1! for the anions associated with the strong acids (.'/-, 0l-, )'1-, etc.!. "he wea& base anion resin would be the resin of choice where a decarbonator was in place at the cation effluent and where silica le#els in the treated water were not a ma6or consideration. + demineralizer train that does not incorporate a decarbonator after the cation unit would be a good candidate for an anion bed consisting of a combination of wea& and strong base resins. "he wea& base resin would be loaded on top of the strong base resin during the initial resin installation. During normal operation, the resin bed would be regenerated prior to bac&washing. "his is necessary to pre#ent the mi$ing of the two resins during bac&wash. In their e$hausted state, the density of the resins is #ery nearly the same. In the regenerated state, their densities are different enough for the resins to separate during bac&washing. 0onsequently, the wea& base resin is always on top. "his is the preferred position, because it insures that the resin with the greatest capacity for the strong acid anions recei#es the effluent from the strong acid cation first. +nother ad#antage of this arrangement is that, li&e the strong base anion resins, the wea& base resins also absorb organicsC howe#er, the

organics absorbed by a wea& base resin are remo#ed during standard regeneration. "his helps protect the strong base resin from irre#ersible organic fouling.

Mixed-Bed Polishers
In #ery high-pressure boiler operations, e#en trace contaminants are e$tremely detrimental. Demineralized water requirements are more stringent than can be produced by the two-bed unit. "o cope with these specifications, many operations follow the two-bed demineralizer with a mixed-bed polisher. "he mi$ed bed polisher consists of a special #essel into which are loaded both strong base anion and strong acid cation resins. "he cation resin anion, being denser, is always on the bottom of the #essel after bac&washing. "he regenerant piping is designed so that caustic enters the top of the #essel, flows down through and regenerates the strong base anion resin, and e$its the side of the #essel near the interface between the anion and cation resins. .imilarly, the sulfuric acid enters the side of the #essel near the resin interface and flows downward through the strong acid cation resin, regenerates it, and flows out of the bottom of the #essel to the sewer. Prior to the addition of the caustic, the #essel is filled with water and &ept under pressure (blocking flow) to insure that no caustic channels down to the cation resin before it lea#es the waste line at the resin interface. +fter regeneration, air is admitted to the bottom of the #essel by opening the #ent. "his is called air mix and ser#es to commingle the resins in a random fashion. +s the anion and cation resins mi$, the result is a #essel filled with small two-bed demineralizers. ;ach cation resin bead and its companion anion resin bead function as a two bed unit. "he combinations on the top that recei#e the water first are successful in remo#ing a great percentage of the contaminants entering the unit. ;ach pair below remo#e, in succession, a high percentage of the contaminants in the water passing the ad6acent pair of resin beads. ;$tremely high-quality effluent water is possible with this arrangement. "he water generally contains silica and sodium in the parts per billion range.

Condensate Polishers
+nother #ariation of the standard demineralizer system is the condensate polisher. 4any high-pressure boiler systems incorporate a #essel loaded with strong acid cation resin or a combination of cation and anion resin to polish the steam condensate coming from turbogenerator surface condensers. %ecause many of these systems return a high percentage of their steam condensate as boiler feed water ma&eup, the polishers ser#e to insure that any contaminants that might enter the condensate (as from the condenser lea&s! are remo#ed prior to reuse. In the past, condensate polishers loaded only with strong acid cation resin were regenerated with salt. %ecause the sodium emanating from these units during the e$change process is detrimental to the control of high pressure boiler water treatment programs, morpholine or ammonia has been substituted in many cases. %efore either of these amines are used for regeneration, the polisher resin must first be put in the hydrogen form by regeneration with sulfuric acid. )e$t, the polisher is regenerated with the amine, which readily e$changes for the hydrogen. During normal operation, the morpholine or ammonia e$changes for the more ob6ectionable cations (0a<<, 4g<<, etc.!. ?egeneration and operation of a mi$ed-bed polisher is identical to that of a standard demineralizer polisher. "he only drawbac& to the operation of a mi$ed-bed polisher is that it remo#es any neutralizing amine fed to the boiler system and therefore loses considerable capacity e#en though no other contaminant e$ists.

Demineralizer 0osts and ?egeneration

General
4ost larger demineralizer systems ha#e been automated to some degree. +utomatic step times and sequencing switches mo#e the units through their regeneration steps. .ome

operators require that manual actuation of each step in the regeneration procedure be made. "his precludes the possibility that a malfunction in the step control automation might send spent regenerant to the treated header instead of the waste system. Initial installation costs for a demineralizer system are quite high. During normal operation (discounting operator labor costs!, resin replacement and regenerants account for most of the system8s e$pense. ?esins are #ery e$pensi#e, with the anion resins being the most costly. ?eplacing an anion resin charge (strong base resin! in a #essel containing 1 ft 1 of resin can run as high as D/ , for the resin alone. + rule-of-thumb estimate of 1= per year is generally used to estimate normal resin loss due to bead brea&age, bac&wash losses, etc. In most cases, >> %e sulfuric acid and 7 = rayon grade caustic are used for regeneration. .ulfuric acid sells for about D 7 per ton and rayon grade caustic sells for about D13 per ton (1AE> dollars!. "he amount of acid and caustic used per regeneration #aries as a function of the desired e$change capacity and the #olume of resin to be regenerated. "ypical regeneration le#els for cation units are 7-E pounds of >> %e sulfuric acid per cubic foot of resin. ?egeneration le#els for strong base anion resins are /-> pounds of caustic per cubic foot of resin. (hen discussing a demineralizer train, the manufacturers generally size the units of the train so that the ser#ice cycle, based on an a#erage ma&eup water analysis, will be about 3/ hours. In some cases, a duplicate train is supplied to allow a rotational sequence to be set up (i.e., one train in ser#ice and the other being prepared for return to ser#ice!. In other cases, a demineralized water holding tan& is used to store water for use during ser#ice cycles. Depending on the size of the demineralizer train, a complete regeneration could ta&e se#eral hours to complete, from the initiation of bac&wash to rinsing the unit down, to specifications prior to standby.

Service Cycle
Strong Acid Cation +s discussed pre#iously, during the normal ser#ice cycle on a strong acid cation unit, the effluent water from the unit contains free mineral acidity (54+! in proportion to the cations in the raw water. "he effluent p* is normally in the 3. -1. range. +s the resin mo#es toward e$haustion, the effluent p* begins to increase (54+ begins to decrease!. "he 9brea&through9 of sodium will occur first, because it is the least preferred cation. "he sodium brea& will precede the brea&through of the cations associated with hardness (0a <<, 4g<<!. 5rom the boiler operator8s standpoint, this is #ery fortunate, because the sodium compounds that lea& through ()a3.'/ )a0l, etc.! are highly soluble and are not a ma6or contributor to scale formation. "he brea&through of sodium is ob6ectionable from the standpoint of its impact on the control of the high pressure boiler water treatment program (it produces high p* in the boiler water!. +s mentioned earlier, most demineralizer systems are anion-limited as far as ser#ice cycle length is concerned. In other words, the anion unit is li&ely to brea& on silica long before the cation unit brea&s on sodium or hardness. %ecause of this, many operators pay little attention to cation 54+ or p* unless serious operational problems de#elop. 4ost of their time is spent worrying about the silica lea&age from the anion. "he ser#ice run on the strong acid cation is terminated when the anion unit reaches the end of its run. %oth units are ta&en off line together and prepared for regeneration. "he cation unit is always regenerated first, because a source of cation free water is required for bac&washing the anion unit. If a water source containing any magnesium is used for bac&washing an anion, the 4g<< can react with unused '*- in the anion and the precipitation of 4g('*! 3 will occur. ;quall# troublesome is the potential for the precipitation of 4g.i' 1, a #ery stic&y, tenacious deposit.

Backwash %ac&washing of the strong acid cation is initiated by re#ersing the flow of raw water and directing it to the outlet from the unit. + bloc& #al#e downstream of the bac&wash line is first closed to pre#ent raw water from entering the anion unit. %ac&washing fluffs up the cation resin bed and aids in releasing any debris that has collected in the bed during the ser#ice cycle. Ideally, bac&washing should e$pand the resin bed by 7 =. %ac&wash flow rates are generally held at 7-> gpmBft3 of #essel cross-sectional area. %ac&washing is normally continued for 7-1 minutes, or until the bac&wash water becomes clear. Acid In ection 'nce bac&washing is completed, diluted sulfuric acid is pumped into the cation #essel through the regenerant distributor. Initially, an acid concentration of 3= or less is started into the unit. "his precludes the possibility of precipitating 0a.'/ in the resin bed. +fter a prescribed period of time, the acid concentration is increased to >= for the balance of the regeneration. +cid flow rates during regeneration are normally .7-1. gpmBft 1 of resin. +cid requirements for regeneration of strong acid cation resins are generally in the 7-E lbBft1 range. "he determination of the e$pected capacity of any strong acid cation resin is a fairly comple$ procedure. %ecause each resin type e$changes the #arious cations found in raw water at different rates, depending upon the cation and its affinity, a careful e#aluation of the raw water chemistry must be made and the manufacturer8s reference data consulted to arri#e at a capacity figure. ?esin e$change capacity is also a function of regeneration le#el. + complete dissertation on the sub6ect of establishing a particular resin8s capacity to demineralize a particular water is beyond the scope of this boo&. Slow !inse Strong Acid Cation. "he slow rinse of a strong acid cation unit is an e$tension of the regeneration acid in6ection step, with the acid feed stopped and only dilution water going to the unit. .low rinsing insures that all of the acid is e#enly distributed through the resin and all of the e$changed cations are purged from the resin bed. .low rinsing is conducted at a flow rate of .7-1. gpmBft 1 of resin for not less than 17 minutes. +s in the acid in6ection step, the effluent from the cation unit during slow rinse is directed to the sewer. "ast !inse Strong Acid Cation. 'nce the slow rinse step is completed, the rinse water flow rate is increased to 1.7-1. gp4Bft1 of resin for 1 -> minutes. 5ast rinsing completely resettles the resin and insures that all spent regenerant and e$changed cations are swept from the #essel to the sewer. 5ast rinsing is generally continued until the 54+ e$iting the unit is 1 -3 = of that e$perienced during the last ser#ice cycle. !et#rn to Service or Stand$y Strong Acid Cation. 'nce the strong acid cation is rinsed down to the desired effluent water characteristics, it is isolated from the system on standby or placed bac& in ser#ice as required. "he regenerated cation unit may be used to supply bac&wash water for its companion anion unit. Service Cycle Strong Base Anion. .er#ice flow rate through strong base anion resins typically runs at 3-1/ gpmBft3 of resin bed cross-sectional area. "he conducti#ity and silica le#el in the treated water

are the parameters normally used to monitor the performance of these units. 4any industrial demineralizer systems are supplied with automatic silica analyzers and on-line conducti#ity analyzers with alarms and ad6ustable setpoints. .trong base anion resins are generally ta&en out of ser#ice for regeneration when the silica le#el in the treated water e$ceeds .1 ppm andBor the conducti#ity e$ceeds 1 micromhos. 4ost strong base anion units are designed to pro#ide specification water for a period of about 3/ hours. ?esin degradation due to organic fouling and resin loss due to attrition are the two main reasons why anion train ser#ice cycles tend to decrease with time. 4ost resin manufacturers quote a useful life of about 1 years for a load of strong base resin. It is not unusual to see a strong base resin lose up to 1 = of its salt-splitting capacity in the first of year operation. 'ther factors (inlet water characteristics, channelling, improper regeneration, etc.! can affect the ser#ice run length of any strong base resin. +ll of these factors must be ta&en into consideration when the production capacity of a strong base anion unit falls below its specified capacity. +s discussed under the section on strong acid cation unit operation, the actual ion e$change capacity of a strong base anion resin depends on a number of factors. .ome of these include regeneration le#el (lbBft1 of resin!, caustic quality, regeneration temperature, the type and quantity of e$changeable anions, etc. +gain, a complete description of the process by which strong base anion resin capacity is determined is beyond the scope of this boo&. Backwash Strong Base Anion. "here is one phenomenon that occurs at the end-of-run on a strong base anion unit that helps the operator determine the proper time to get the unit off-line for regeneration. "his phenomenon is associated with the conducti#ity of the effluent water and is used to catch a unit before it brea&s on silica. Fust prior to silica brea&through, the conducti#ity of the effluent water from the anion units ta&es a noticeable drop, a short time later, it increases sharply, indicating brea&through. "his decrease in conducti#ity is clearly e#ident on a plot of the readings ta&en from the conducti#ity probe in the bottom of the anion #essel. During the normal ser#ice cycle, sodium that lea&s from the cation unit combines with '*- from the anion unit to produce )a'*, which is conducti#e. +s silica first starts to brea& through, the silica combines with the sodium lea&age from the cation to produce sodium silicate, which is less conducti#e than the )a'* it replaces. "herefore, a dip in the effluent conducti#ity occurs. 'nce the anion unit is ta&en out of ser#ice, it is made ready for the regeneration procedures. 5irst, the bloc& #al#es are manipulated to isolate the unit from the treated water system. (ater from the companion cation unit is directed up from the bottom of the unit at rate of 3-1 gpmBft3 of #essel surface area. %ac&washing is continued for 7-1 minutes with the effluent going to the sewer. + bed e$pansion of 7 = is desired. !egeneration Strong Base Anion. "he strong base anion unit is regenerated with a /= solution of rayon grade caustic and cation free water. 0austic (7 =! is pumped from the storage area and mi$ed with dilution water of the required quality. "he /= caustic solution passes through a heater (steam or electric! and on to the regenerant distribution system in the anion #essel. "he effluent from the caustic heater is controlled at 13 5 (/E.EA 0!. "he elution of the e$changed silica from the strong base resin is greatly enhanced by the hot caustic. If the a#ailable time and heater capacity e$ist, it is ad#isable to preheat the anion bed to the desired caustic temperature before caustic in6ection is started. "his insures that the resin bed operates at 13 5 (/E.EA 0! during the full regeneration cycle. If hot caustic is not used for the regeneration of the strong base resin, silica remo#al can be incomplete. "his leads to

reduced capacity on the ne$t ser#ice cycle. .ilica remaining in the resin can also polymerize and permanently foul the resin. 0ommonly, regeneration le#els of /-> lbsBft1 of resin are used for anion units. "he warm caustic is introduced to the unit at a rate of .37-1. gpmBft1 of resin #olume for 1 -E minutes. Slow !inse Strong Base Anion. 'nce the full charge of caustic has been in6ected, the caustic pump is turned off and a slow rinse with dilution water only is started at the same flow rate as that used for caustic in6ection. .low rinsing is conducted for a minimum of 1 7 minutes. .low rinsing mo#es the regenerant solution through the resin in 9plug9 flow fashion and helps remo#e some of the e$changed anions. "ast !inse Strong Base Anion. 'nce the slow rinse cycle is completed, the rinse rate is increased to 1. -3. gpmBft1 of resin for 1-3 hours to scour all of the regenerant and e$changed anions from the resin. "he end of the fast rinse step is generally determined to be the point at which the effluent conducti#ity reaches 1 micromhos. If the anion unit is placed on standby for a significant period of time, it should recei#e a short fast rinse prior to being put bac& in ser#ice. "his insures that any silica or other anions that ha#e eluted from the resin bac& into the stored water are swept to the sewer and not the treated water storage system. !et#rn to Service or Stand$y Strong Base Anion. %ased on need, the anion unit is either returned to ser#ice or placed on standby until needed. %ther Anion !esin &y'es "he regeneration procedures enumerated abo#e apply generally to other anion resin types. "he one notable e$ception is the so-called stratobed anion unit, which has a combination of wea& and strong base resins. (ith this type of unit, regeneration precedes bac&washing to insure that the resins don8t mi$ during bac&washing.

"roubleshooting "ips for the Demineralizer Strong Acid Cation

+. Problem, (igh Sodi#) *eakage +low "MA, . 1. 0hec& ser#ice flow rate... poor flow distribution may result from high or low flow rates. 2. *as the unit been returned to ser#ice from standbyG + short rinse before ser#ice can 3. 4.
help. Is the regeneration le#el satisfactory for the raw water characteristicsG ?efer to resin dataC increase acid by 1 = at ne$t regeneration. Is unit regenerated or bac&washed at least once e#ery 3/ hoursG ;$tended ser#ice runs can result in channelling of flow.

5. (as the unit run to complete e$haustion during the last ser#ice cycleG It should 6. 7.
recei#e a double regeneration before the ne$t ser#ice cycle to achie#e ma$imum capacity. Is a #al#e lea&ingG 2ea&ing bloc& #al#es can introduce a small amount of raw water into the treated water. +re the acid concentrations satisfactoryG *igh acid concentrations during initial phases of regeneration can result in 0a.'/ precipitation in the resin bed. 2ow acid concentrations result in incomplete remo#al of sodium ions from the bottom of the resin bed and sodium lea&ing during the following ser#ice. +re the regeneration flow rates and times satisfactoryG 0hec& the manufacturer s recommendations. ;$cessi#e bac&wash flow rates may result in resin loss while low bac&wash flow rates may lead to channelling during subsequent steps. 2ow flow rates during acid in6ection may result in 0a.'/ precipitation. *igh acid flow rates lead to poor regeneration due to insufficient contact time.

8.

%. Problem, (igh

"ree Mineral Acidity

1. Is the rinse of adequate flow and durationG If the pre#ious rinse was inadequate, the
high 54+ (higher than during pre#ious ser#ice cycle! is due to the incomplete remo#al of spent acid from the resin after the regeneration. +dditional rinsing will bring it down. Does the effluent contain hardnessG +d6ust sample p* to /.1-1 . with 1) )a'* prior to running hardness. If calcium is present, the precipitation of 0a.' / should be suspected. 0hec& acid concentrations during regeneration and flow rates used. +re the acid regenerant #al#es lea&ing acid into the unit during the runG If the 54+ from the unit in ser#ice increases only when a parallel unit is being regenerated, there is a good chance that a #al#e is not holding the pressure of the acid line and acid is entering the operating unit.

2. 3.

0. Problem, *ow

Ca'acity

1. *as the raw water cation loading increasedG 0hec& the raw water total cation loading
against the manufacturer8s design loading. "o determine the total raw water cation loading, add the sodium content to the total hardness. + good estimate can be made in the field by adding the 4 al&alinity of the raw water to the 54+ at the cation effluent. Is the ser#ice flow rate satisfactoryG Poor flow distribution may result from either high or low flow rates. 0hannelling will cause only a small area of the resin bed to recei#e a large portion of the raw water. Is the unit regenerated or bac&washed at least e#ery 3/ hoursG ;$tended ser#ice runs can lead to channelling of the flow. (as the unit o#errun during the pre#ious ser#ice cycleG + unit run to e$naustion must be regenerated twice before full capacity will be restored. +re the acid concentrations satisfactoryG Insufficient acid feed to the unit during regeneration will lead to lower than e$pected capacity. 0hec& the actual acid fed against the manufacturer8s recommendations. +re the regeneration flow rates and times satisfactoryG ;$cessi#e flow rates during regeneration will reduce the acidBresin contact time required for good regeneration and reduced capacity will result.

2. 3. 4. 5. 6.

D. Ins'ect

the #nit

(hen all of the troubleshooting tips are chec&ed out and the problem is not resol#ed, a physical e$amination of the #essel interior will often re#eal a mechanical problem. 0hec& the

raw water inlet distributor for missing spray heads, the regenerant distributor for damaged or missing parts, the surface of the resin for une#enness or signs of resin degradation, etc.

Strong Base Anion

+. Problem, (igh

-..l#ent Cond#ctivity

1. (hat is the cation sodium lea&ageG *igh cation sodium lea&age is indicated by a low
54+ at the cation outlet. ;ach ppm of sodium lea&age from the cation adds about 1 micromhos of conducti#ity to the effluent of the strong base anion unit. "he effluent p* from the strong base anion will also be high (o#er E. !. If high sodium lea&age is the problem, refer to the troubleshooting tips for the strong acid cation unit for a resolution. (as the anion unit rinsed sufficiently at the end of the last regenerationG Hood anion rinse usually requires 1 -17 gallons of rinse water per cubic foot of resin in the unit. 0hec& rinse flow rates and times to confirm. Is a bloc& #al#e lea&ingG 0hec& water samples upstream and downstream of any suspect #al#e. (as 4g('*!3 or 0a0'1 precipitated on the resin during the last regenerationG 0ation treated water should be used for bac&washing the anion resin and in the preparation of caustic regenerant solutions. "he use of raw water for either will result in precipitation of 4g('*!3 and 0a0'1 in the resin bed. ;lution of these compounds during the ser#ice cycle can produce high effluent conducti#ities.

2. 3. 4.

%. Problem,

(igh Silica *eakage

1. Is the ser#ice flow rate satisfactoryG *igh or low ser#ice flow rates may produce
channelling.

2. Is the regeneration le#el satisfactory for the raw water characteristicsG ?efer to the
resin manufacturer8s strong base anion resin lea&age cur#es. Increase the caustic used for regeneration by 1 = to see if the silica lea&age decreases during the following ser#ice cycle. Is the unit regenerated or bac&washed at least e#ery 3/ hoursG ;$tended ser#ice runs can result in channelling of the flow. (as the unit o#errun during the last ser#ice cycleG If the unit ran to e$haustion during the last run (the unit bro&e on conducti#ity as opposed to silica!, a double regeneration will be required to restore full operating capacity. Is the caustic concentration and temperature satisfactoryG In order to insure good regeneration (silica remo#al!, the caustic must be fed at a concentration of /= and a temperature of 13 5 (/E.EA 0!, +re the regeneration flow rates and times satisfactoryG Iariations from the recommended procedures will lead to unsatisfactory operation.

3. 4. 5. 6.

0. Problem, *ow

Ca'acity

1. *as the raw water changed with respect to the design water analysisG *a#e a 2.
thorough analysis made of the raw water and chec& the total e$changeable anion content against that used for the unit design. Determine whether the cation or anion is e$hausting first. If the cation unit e$hausts first, the conducti#ity and p* of the anion effluent water will increase. If the anion unit is e$hausting first, the conducti#ity of the anion effluent water will rise and the p* will drop.

D. Ins'ect

the #nit

1. If all operating #ariables chec& out, an inspection of the strong base anion unit 2.
internals is recommended. "he anion resin should be analyzed periodically to determine if it has become fouled or degraded. "otal and salt-splitting capacities should be determined.

Jseful Data
(ater analyses e$press the le#els of the #arious ions in solution both as the ion (as in analysis by atomic absorption! and as an equi#alent amount of calcium carbonate (0a0' 1!. "he con#ention chosen by resin manufacturers for e$pressing e$change capacity of the #arious ion e$change resins is &ilograinsBcubic foot. "he &ilograms are assumed to be the calcium carbonate equi#alent. "herefore, before any calculations can be made to determine, for instance, the operating capacity of an ion e$change unit, all analyses must first be con#erted to the 0a0'1 equi#alent.

&AB*- /01/ Ion -xchange Conversion "actors Ion "actor 2sed to Convert CaC%3 -4#ivalent 3.7 /.1 3.1E 1. / 1./1 .E1 .E3 .E/ 1.1/ .E1

0alcium, 0a<< 4agnesium, 4g<< .odium, )a< .ulfate, .'/-0hloride, 0l)itrate, )'1%icarbonate, *0'10arbonate, 0'1-0arbon dio$ide, 0 3 .ilica, .i'3

Hrains per gallon $ 1@.1 : parts per million (ppm! Pounds per 1 gallons $ 13 : parts per million (ppm!

&AB*- /015 &reated Water Characteristics "or 6ario#s Ion -xchange Syste)s Sodi#) 7eolite Strong Acid Cation Weak Acid Cation De)in 5Bed Weak Base .1 . . . . Jnaff. 7.> 7-3 . . De)in 5Bed Strong Base .1 . . . . .1 to .3 E-1 7-1 . . De)in Mixed Bed

"* 0a 4 +l& .'/ 0l .i'3 P* 0ond. 54+ 4g

.1 1. Jnaff. Jnaff. Jnaff. Jnaff. .lt. Inc. .lt. Inc. . 1.

.1 . . Jnaff. Jnaff. Jnaff. 3.7-1.7 >.7$"4+ 1. $"4+ .

0a < 4g )0* 0a )0* . Jnaff. Jnaff. Jnaff. /.1 3$"4+ . 4g )0*

. . . . . .1 to .7 @. .7 to .37 . .

"4+ : "otal 4ineral +cidity )0* : )oncarbonate *ardness

&AB*- /013 !egenerant S'eci.ications and Sol#tion Data Salt S'eci.ications 0a 4g .i'3 Insolubles 2ess then .3= 2ess then .1= 2ess then .1= 2ess then .1=

)o oil, grease ?oc& salt not recommended

Salt Sol#tion Data Hal %rineBft1 =)a0l .p Hr )a0l lb per gal .@ 7 .EA/ 1. E@ 3.7A Hal %rine per lbB)a0l 1./3 1.13 .A3 .1A at > lbBft1 E.7 >.@ 7.7 3.17 at E lbBft1 11./ A. @./ 1.1 at 1 lbBft1 1/.3 11.3 A.3 1.A at 17 lbBft1 31.1 1>.E 11.E 7.E

E 1 13 3>K

1. 7> 1. @1 1. E> 1.1A@

K.aturated brine solution

&AB*- /018 S#l'h#ric Acid S'eci.ications and Sol#tion Data S#l'h#ric Acid S'eci.ications "echnical Hrade >> %e acid, free of suspended matter and light in color, should be used. "he sulphuric acid should mi$ completely with water without the formation of any precipitate. +cid with inhibitors should not be used. Grade 0olor 0oncentration Iron )itrogen compounds +rsenic 'rganic contaminants &echnical (ater-white to light brown >> %e 7 ppm ma$imum 3 ppm ma$imum .1 ppm ma$imum . 1= ma$imum

(5S%8 Sol#tion Data Hal .olnBft1

=*3.'/

.p Hr

*3.'/ lb per gal . E1A .1>EA .377 .1/33 ./1 7 .73 .@ 37 1/.E

Hal .oln per *3.'/ 11.A 7.A 1.A 3.A 3.13 1.A3 1./3 . >E

at 3 lbBft1

at 3.7 lbBft1

at 1 lbBft1

at / lbBft1

at > lbBft1

1 3 1 / 7 > E A>-A@K K>> %e

1.

71

31.E 11.E @.E 7.E /.> 1.E 3.E .11>

31.E 11.E A.@ @.3 7.E /.E 1.77 .1@

17.@ 1@.@ 11.@ E.@ @. 7.E /.1 .3

/@.7 31.> 17.> 11.> A.3 @.> 7.> .3@

@1. 17. 31. 1@./ 1/. 11.> E.> ./1

1. 11E 1 .1E/ 1. 37 1. 11A 1. 1E7 1. 773 1.E/

&AB*- /019 (ydrochloric Acid S'eci.ications and Sol#tion Data (ydrochloric Acid S'eci.ications Grade 0olor 0oncentration .ulphuric +cid as .'1 5e 'rganic 0ontaminants &echnical (hite to light yellow 4inimum 1 = *0l ./= ma$imum .1= ma$imum . 1= ma$imum

)ote, +cid should be free of turbidity and contain no inhibitors. Do not use hydrochloric acid obtained as a by-product of chlorinated organic chemicals. *ydrochloric acid obtained by the salt acid or hydrogen chloride process is satisfactory. (Cl Sol#tion Data Hal .olnBft1 =*0l .p Hr *0l Hal .oln at 3 lbBft1 at 1 lbBft1 at / lbBft1 at 7 lbBft1 at > lbBft1

lb per gal 7 1 1 1. 31 1. /@ 1.17 ./3@ .E@/ 3.EE

per *0l 3.17 1.1/ .17 /.@ 3.3E .@ @. 7 1./3 1. 7 A./ /.7> 1./ 11.@7 7.@ 1.@7 1/.1 >.E/ 3.1

&AB*- /01: Ca#stic S'eci.ications and Sol#tion Data Ca#stic Soda +;a%(, )a'* )a3'1 )a0l )a0l'1 5e3'1 .i'3 +l3'1 0a' 4g' 0u 4n Pl )i )a3.'/ +s 3 S'eci.ications +!ayon Grade, AE.37= .@>= .1@= 2ess than 3 ppm 7 ppm 1 ppm 1 ppm 3@ ppm 1E ppm .3 ppm .7 ppm .7 ppm .A ppm ppm . 1 ppm

Ca#stic Soda Sol#tion Data

Hal .olnBft1 =)a'* .p Hr )a'* lb per gal .1@ / .37E1 .1/E1 .7113 .@37> >.1>7 Hal .oln per )a'* 7.E> 1.E> 3.E@ 1.E@ 1.1> .17@ at 3 lbBft1 at 1 lbBft1 at / lbBft1 at 7 lbBft1 at > lbBft1

3 1 / > E 7

1. 3 @ 1. 11E 1. /3E 1. >/E 1. E>A 1.7371

11.@ @.@ 7.@ 1.@ 3.@ .11

1@.> 11.> E.> 7.> /.1 ./@

31./ 17./ 11.7 @.7 7./ .>1

3A.1 1A.1 1/.1 A.1 >.1 .@A

/>.A 1 .A 31. 1/.A 1 .A 1.37