When water evaporates, the dissolved and suspended solids it contained are left behind. When that water is used in boilers or as a rinse for electronic circuits, those residuals can be very damaging, causing losses in heat efficiency or unwanted electrical conductivity. Therefore, they must be removed.
The process of removing the minerals from water is called demineralization or deionization. The former refers to the removal of suspended as well as soluble materials from the water, whereas the latter means the removal of the soluble ions from water. In the past, the process of demineralization was primarily by distillation. However, a combination of reverse osmosis and ion exchange or ion exchange alone is now used
The Demineralization Process
The removal of ionic materials from water by ion exchange is straightforward chemistry. Cations are replaced by hydrogen (H+) ions, shown in Reaction 1, and anions are replaced by hydroxide (OH-) ions, shown in Reaction 2. The result is that a salt such as sodium chloride (NaCl) will be replaced by H+ + OH- (HOH), or H2O. So we remove the salt and create pure water in its place.
Reaction 1 (de-cationization):
NaCl + OH+ → ONa+ + HCl (All salts become acids.)
Reaction 2 (neutralization):
HCl + OOH- → OCl- + HOH (H2O) (All acids are neutralized.)
These reactions are nearly complete at about 99.5 percent. That means with a feed water of 250 parts per million (ppm), you can expect a residual of about 1.25 ppm at a pH of 9 or greater. The residual will consist primarily of sodium hydroxide with some bicarbonate and silicates, which results in the high pH. Due to the higher conductivity of the high pH water, this will test as about 6 micro siemens (mS) water, about 165,000 ohms. That’s with 8 pounds of acid and caustic used in the regeneration process with co-flow configuration. With harder regenerations (i.e., more chemical), the typical two-bed quality will be 250,000 ohms to 350,000 ohms. That will satisfy some of the deionized water needs, but not all. To achieve higher quality, a mixed bed would be required.
The process of removing the minerals from water is called demineralization or deionization. The former refers to the removal of suspended as well as soluble materials from the water, whereas the latter means the removal of the soluble ions from water.
Mixed-Bed Efficiency
Mixed beds are so effective for deionization because the two-bed process of cation/anion exchange is repeated hundreds of times. Mixed-bed water will routinely test at 15 megohms to 18 megohms or 0.056 mS (about 0.03 ppm).
Historically, many portable exchange deionization plants evolved from the simple in-tank, two-bed manifold regeneration plants to the larger and more sophisticated bulk regeneration facilities. But employing mixed-bed regenerations to handle the more demanding part of the potential business entails additional work. Mixed-bed plants require a lot of extra equipment and higher-purity water to facilitate the increased quality. The question is, are there any alternatives to the mixed bed for making higher-purity water?
Two-Bed Quality Limits
If we look at the equilibrium relationship equation, we can see that the amount of sodium (Na) in the water and the level of hydrogen ion on the resin (shown in the numerator) pushes the reaction forward. What slows it down and pushes back is the amount of H+ ion in the water and the sodium, or chloride, level on the resin (shown in the denominator).
Equilibrium relationship:
An increase in the amount of hydrogen ions in the water ([H’]W) is going to result in a drop in the pH. In other words, the more acidic the environment, the slower the cation reaction occurs. We can generalize for both the strong acid cation (SAC) and the strong base anion (SBA) that, as we move away from pH 7, the reaction slows down. As we move down the bed of cation in service, the amount of sodium in the water decreases, as does the hydrogen on the resin.
Meanwhile, the amount of sodium on the resin increases, as does the amount of hydrogen in the water. This, then, becomes the limiting factor in how far this reaction will proceed. The sodium that is not removed by the SAC then enters the SBA bed, where all the acids are neutralized by replacement with hydroxide ions. The residual sodium is thus converted to sodium hydroxide, resulting in an elevated pH. Again, as the pH rises, the reaction slows and cannot go to completion. We are left with an almost but not completely deionized feed water. A mixed bed will easily polish that out. But how can we do it without the complications of the mixed bed?
Regeneration Technique Can Improve Quality
The above scenario pertains to co-current regeneration. With co-current, the regenerant is introduced in the same direction as the service flow. Let’s assume this is down flow. As the chemical passes through the bed, it strips the exhausted resin of the salt ions and replaces them with either the hydrogen or hydroxide. The salt concentration gets higher in the water phase as the strength of the regenerant gets lower and lower. Salts are deposited on the portion of the bed that was previously the least exhausted, the bottom. Unless very high levels of regenerant are used, the bottom of the bed will be left with a heel of salt that will bleed off in the next service cycle, resulting in a background leakage. This further limits the quality of the two-bed cycle.
Counterflow Regeneration
By reversing the direction of the regenerant flow (counterflow), the regenerant will first see the cleanest portion of the exchange bed: the bottom. This leaves a stronger chemical to pass further into the bed and produces a slightly higher capacity for a given level of regenerant.
Counterflow regeneration also avoids the drawback of having the regenerant chemical diluted by the head space in the resin tank. This boosts the capacity and furthers the regeneration process. Furthermore, the exit portion of the bed in service will be highly polished, producing a lower leakage and higher-quality water. As a general rule, counterflow regeneration will produce a water with one-third the leakage and 10 percent higher capacity for any level of regeneration. Deionization proceeds to 99.8 percent. Leakage is 0.5 ppm as sodium hydroxide. Resistivity is 1 mS to 1.5 mS or 0.6 megohm to 1 megohm. To gain the benefits of counterflow regeneration, the tanks must be completely filled, with no freeboard. Figure 1 illustrates the basic difference between co-flow and counterflow with respect to the cleanliness of the regenerated bed. 
Alternative to the Alternative
If counterflow packed beds are not possible because of your plant design, you can approach the same quality using standard co-flow tanks. Since the total dissolved solids (TDS) from a co-flow two-bed is mainly sodium hydroxide, you can follow the two-bed with an additional cation polisher and produce a near-neutral effluent with a TDS of about 0.5 ppm. Conductivity will be comparable to the counterflow two-bed system. If you have an in-tank or a funnel regeneration system, give a three-bed configuration (SAC/SBA/SAC) a try. One megohm is possible.
The Ultimate Alternative
With highly polished exit zones, counterflow regenerated tanks will produce very good quality. There will still be slight leakage from the primary cation that can be polished out with an additional SAC unit following the SBA. With this three-bed configuration, it is possible to produce 10 megohm water. There are special precautions for the regeneration.
Counterflow regeneration is not simply a matter of switching hoses around. There would be no backwash because the polishing zone at the exit of the bed cannot be disturbed. The regenerant has to be introduced at a flow rate low enough to not expand the bed. For a 14-inch portable exchange deionization (PEDI) tank (one square foot surface area), that is a flow of 1 gallon to 1.5 gallons per minute (gpm) at 70 degrees Fahrenheit. With 3.5 cubic feet (cu ft) of resin in the vessel, this will be a regenerant flow rate of 0.33 gpm to 0.5 gpm per cubic foot. For the SBA, the flow would be 0.75 gpm to 1.1 gpm. This flow recommendation is for in-tank regenerations and does not apply to bulk regeneration designs.
Is It Possible to Produce 18 Meg Water Without a Mixed Bed?
It has been well demonstrated that properly regenerated vessels in a counterflow sequence will polish feed water to megohm quality. Using uniform bead components with a twin two-bed series followed by an SAC polisher, it is possible to make 18 megohm water (SAC/SBA/SAC/SBA/SAC), as shown in Figure 2. 
Systems Design
If you are seeking to advance your PEDI business by venturing into the realm of higher-purity water, you probably will not be able to use your old float, which, after years of service, has likely become hopelessly fouled. High-purity water requires high-purity resins and quality chemicals. Start with new resin. Obtain uniform bead cation and anions, and purchase them in the regenerated form. Fill the tanks completely. The resin will shrink upon exhaustion.
Use good prefiltration to avoid having to backwash your vessels. Backwashing puts contaminating ions on the polishing end of the beds, which will require excessive rinsing to bring back to quality. All steps should be carried out with deionized (DI) water.
If you are seeking to advance your PEDI business by venturing into the realm of higher-purity water, you probably will not be able to use your old float, which, after years of service, has likely become hopelessly fouled.
Chemical dosing and slow rinse are counterflow. Fast rinse is co-flow. Hydrochloric acid is preferred. Buy a salt-acid grade at 30 percent hydrochloric acid. Dilute to 4 percent to 6 percent with DI water for regeneration. Buy 50 percent sodium hydroxide. Rayon grade should be sufficient. Dilute to 4 percent to 5 percent with DI water for regeneration. Start your regeneration process at 8 pounds per cubic foot, and fine-tune it from there. You can save your used chemical for regenerating your standard float. Use only the last half of the regenerant.
Acid and caustic regenerant levels are measured as active weights. In other words, 100 percent solids 30 percent hydrochloric acid contains 2.877 pounds of 100 percent hydrochloric acid per gallon. A dose at 8 pounds per cubic foot is 2.8 gallons of the 30 percent liquid.
Fifty percent caustic contains 6.364 pounds of 100 percent sodium hydroxide. You need 1.25 gallons of the 50 percent liquid. A 5 percent hydrochloric acid solution will have a specific gravity of 1.025 and contain 0.4275 pound of 100 percent hydrochloric acid per gallon. A 5 percent sodium hydroxide solution will have a specific gravity of 1.0538 and contain 0.4397 pound of 100 percent sodium hydroxide per gallon.
Eight pounds of dilute acid will have a volume of 18.7 gallons. At a draw rate of 0.5 gpm/cu ft, it will take 36 minutes to draw it in, and that would be followed by a 50-minute slow rinse. Eight pounds of diluted caustic will have a volume of 18.2 gallons. At a draw rate of 0.33 gpm/cu ft, it will take 55 minutes to draw it in, and that would be followed by a 60-minute slow rinse. Contact times are important to strip everything off the resins. Fast rinses are done on the separate tanks at 1.5 gpm/cu ft. SAC resins are rinsed with DI water to a pH of 5. SBA are rinsed with DI water to a pH of 9. Blow the water out with air before transporting to avoid allowing the beds to mix. A typical manifold setup is shown in Figure 3.
Conclusions
In-tank counterflow regeneration is an easy technique to adopt for higher quality. Even if the setup is not perfect, there will be a greatly improved quality. Figure 4 illustrates the quality possible.
About the author
C.F. “Chubb” Michaud, MWS, is the technical director and CEO of Systematix Company of Buena Park, CA,which he founded in 1982. He has served as chair of several sections, committees, and task forces within the Water Quality Association (WQA), as well as served as a past director and governor. He served on the Pacific Water Quality Association (PWQA) board, chairing the Technical and Education Committees for 12 years. Michaud is a proud member of both the WQA and PWQA Halls of Fame, has been honored with the WQA Award of Merit, and is a two-time recipient of the PWQA Robert Gans Award. A frequent and well-published author and speaker, Michaud has contributed over 100 original papers on water-treatment techniques and holds four U.S. patents on ion-exchange technologies. He holds a BS and an MS degree from the University of Maine.