How Deionization (DI) Works: The Short Answer
Fill a glass from the tap and you're also pouring in a load of dissolved minerals: calcium, magnesium, sodium, chloride, and a dozen others you'll never see. For drinking water, that's perfectly fine. For a laboratory instrument, a reef aquarium, or a spot-free car rinse, those invisible ions are the entire problem. Deionization is how you get rid of them, down to almost nothing.
Definition: Deionization (DI) is an ion-exchange process that removes dissolved mineral ions using H⁺ cation and OH⁻ anion resins to produce very low-conductivity, high-resistivity water.
Here's the idea without the chemistry. Picture a vending machine that only takes exact change. One stage is stocked with hydrogen ions (H⁺), the next with hydroxide ions (OH⁻). Every dissolved mineral that drifts through trades itself in for the matching ion. The clever part: the H⁺ and OH⁻ handed back in those trades immediately snap together into a molecule of water. The minerals stay locked in the resin. What flows out the other end is water with almost nothing left dissolved in it.
That's the whole trick, and the rest of this guide is about the levers that control how pure the water gets. Crystal Quest designs and builds DI and RO/DI systems in the USA, so the focus here is the mechanism and the practical knobs that matter. Deionization is one branch of ion exchange water filtration. If you'd rather dig into the media itself, which grade to buy, and how to size it, that lives in our specialist's guide to DI resin.
Key Takeaways
Two Trades Make Water
Contact Time Is Everything
RO First, DI to Polish
Trust the Meter, Not the Calendar
How Ion Exchange Works in Deionization
DI resins are tiny porous polymer beads, each one carrying fixed charges along its surface. Water flows through a packed bed of these beads, and dissolved ions trade places with ions the resin is holding. Two coordinated swaps do the work, and the second one is what makes deionization special: the leftovers from both trades combine into water itself.
- Cation resin (H⁺ form): Negatively charged sulfonic sites (-SO₃⁻) hold onto H⁺. They trade that H⁺ for positively charged ions in the water (Ca²⁺, Mg²⁺, Na⁺, Fe²⁺/Fe³⁺).
- Anion resin (OH⁻ form): Positively charged quaternary amine sites hold OH⁻. They trade it for negatively charged ions (Cl⁻, NO₃⁻, SO₄²⁻, HCO₃⁻, silicate).
- The payoff: The released H⁺ and OH⁻ find each other and form H₂O. Conductivity falls, resistivity climbs, and on a good mixed bed it approaches about 18.2 MΩ·cm at 25 °C.
For readers who want it at the molecular level, here are the actual reactions:
2R-SO₃H + Ca²⁺ ⇌ (R-SO₃)₂Ca + 2H⁺
R-N⁺(CH₃)₃OH⁻ + Cl⁻ ⇌ R-N⁺(CH₃)₃Cl⁻ + OH⁻
Every bead has a fixed number of exchange sites, which is its capacity (often measured in eq/L or kgr/ft³). Resins also play favorites, grabbing some ions more eagerly than others, which is selectivity. Capacity, selectivity, and contact time together decide how much leakage you get and how long a charge of resin runs before it's spent.
Selectivity, Leakage, and pH Behavior
Not every ion is equally easy to catch, and that shapes how a DI bed behaves as it nears exhaustion.
- Silica and CO₂ are the stubborn ones: Silica (as silicate) and dissolved CO₂ species are the hardest anions to hold. SBA Type I resin keeps their leakage low, and degassing upstream lightens the CO₂ load before it ever reaches the anion bed.
- Strong ions crowd out weak ones: When strongly held ions are abundant, they can bump weakly held ions back off the resin, which changes the order things break through.
- pH gives you clues: If the cation stage runs out first, effluent pH tends to drop; if the anion stage goes first, pH tends to rise. In a mixed bed the signal blurs, so lean on your meters rather than pH alone.
For the lowest silica and CO₂ leakage, pair SAC (H⁺) with SBA Type I (OH⁻), keep flow conservative, and maintain full bed depth. If CO₂ is elevated, add a degas step ahead of the anion exchange.
Why Contact Time (EBCT) Decides Your Results
Ion exchange isn't instant. Each ion has to diffuse out of the water, into a bead, and find an open site, and that takes time. Give the water more time against the resin and you get cleaner water; rush it and ions slip through. Three things move that dial:
- Bed depth and EBCT: Deeper beds and longer empty bed contact time (EBCT, the seconds water actually spends in the resin) improve exchange efficiency and silica capture.
- Flow distribution: Proper distributors and screens stop channeling, so every part of the bed sees its share of the water instead of letting some race through an open lane.
- Temperature: Cold water slows diffusion into the beads. A modestly warmer feed, within the resin's spec, speeds the kinetics back up.
If resistivity drifts or silica creeps up, slow the flow to buy more EBCT, check for channeling (distributors, screens, even loading), and confirm your resistivity reading is temperature-corrected to 25 °C before you trust it.
Cation Exchange: The First Trade
Example, calcium removal:
2R-SO₃H + Ca²⁺ → (R-SO₃)₂Ca + 2H⁺
The resin grabs the Ca²⁺ and lets go of H⁺ into the water. Same story for Mg²⁺, Na⁺, and the other cations.
Anion Exchange: The Second Trade
Example, chloride removal:
R-N⁺(CH₃)₃OH⁻ + Cl⁻ → R-N⁺(CH₃)₃Cl⁻ + OH⁻
The resin grabs the Cl⁻ and releases OH⁻, which then pairs with the H⁺ from the cation stage to become water. That same anion chemistry is what targeted systems use to pull specific contaminants like PFAS and nitrate out of water. If you want the plain-language version of cation versus anion, start with our beginner explainer on anion vs cation exchange.
Deionization Diagram: The DI Process Visualized
| Ion in Your Water | Removed By |
|---|---|
| Calcium (Ca²⁺), Magnesium (Mg²⁺) | Cation exchange resin (H⁺ form) |
| Sodium (Na⁺), Iron (Fe²⁺/Fe³⁺) | Cation exchange resin (H⁺ form) |
| Chloride (Cl⁻), Nitrate (NO₃⁻), Sulfate (SO₄²⁻) | Anion exchange resin (OH⁻ form) |
| Silica (as silicate), Bicarbonate/CO₂ species | Anion exchange resin (OH⁻ form; Type I for maximum silica capture) |
Nitrate is a good example of why anion selectivity matters in drinking-water work. The EPA sets a maximum contaminant level of 10 mg/L (as nitrogen) for nitrate, and anion resin captures it right alongside the other anions above. See the EPA National Primary Drinking Water Regulations for the regulated limits.
Which Resins Do the Work
You only need to know two resins to follow how DI works. A Strong Acid Cation (SAC, H⁺ form) resin handles the positive ions, hardness, sodium, and metals across a broad pH range. A Strong Base Anion (SBA, OH⁻ form) resin handles the negatives, and its Type I grade is the one that keeps silica leakage low. Pair SAC with SBA Type I, give them enough bed depth, and keep the flow conservative, and you've got the backbone of almost every DI system.
There are other grades (weak-acid and weak-base resins for specialized trains, Type II anion for higher capacity where silica matters less), but choosing among them is a buying decision rather than a how-it-works one. For the full grade-by-grade breakdown and how to match a resin to your application, see our guide to DI resin and the deeper look at anion exchange resin types.
The Three Ways DI Is Set Up
The same chemistry gets arranged in three common configurations, each trading simplicity for purity differently.
- Two-bed DI (separate vessels): Water hits a cation vessel (H⁺), then an anion vessel (OH⁻). It's efficient at bulk removal and easy to regenerate on site. Add a mixed-bed polisher downstream when you need to go further.
- Mixed-bed DI: SAC and SBA resins are blended into one vessel instead of stacked in two. That blending is what pushes resistivity to its highest, which is why mixed beds are the go-to final polishing stage.
- EDI (electrodeionization): Membrane-assisted ion exchange that runs continuously on an electric field after RO, with no chemical regeneration. On purity it lands between two-bed and mixed-bed.
Deciding which one actually fits your flow, budget, and purity target is its own question, and our DI resin guide walks through that choice in detail.
The DI Process, Step by Step
-
Pretreatment
Sediment and carbon protect the resin, and RO upstream removes 95-99% of TDS so the resin isn't wasted on the easy ions.
-
Cation exchange (H⁺ form)
Swap H⁺ for Ca²⁺, Mg²⁺, Na⁺, and the rest. The water leaving this stage carries H⁺ as its main cation.
-
Anion exchange (OH⁻ form)
Swap OH⁻ for Cl⁻, NO₃⁻, SO₄²⁻, HCO₃⁻, and silica. Now H⁺ + OH⁻ → H₂O.
-
Mixed-bed polish (common)
Those repeated cation and anion encounters drive conductivity to its lowest and squeeze out the last of the silica.
-
Final filtration
A fine filter (often 0.2 µm) keeps stray particulates out of instruments and loops downstream.
How You Measure DI Water: Conductivity, Resistivity, TDS
You can't see ions, so you measure their electrical fingerprint instead. The fewer ions in the water, the worse it conducts electricity, which is exactly what you want.
- Conductivity (µS/cm) and resistivity (MΩ·cm): Two sides of the same coin. They're inversely related, so higher resistivity means purer water. Ultrapure DI water sits around 18.2 MΩ·cm (about 0.055 µS/cm) at 25 °C. That number is the ASTM D1193 Type I reagent-water benchmark, the practical ceiling a mixed-bed polish chases.
- TDS meters: A handy proxy for ion content. For critical work, measure resistivity at 25 °C and track silica separately. For context, the EPA's secondary standard for total dissolved solids is 500 mg/L; DI drives TDS far below that, into the single digits.
- The CO₂ catch: Dissolved CO₂ turns into bicarbonate and carbonate and can quietly pull resistivity down. Degassing, or a strong anion Type I resin, helps when CO₂ is a real factor.
Getting the Most Out of a DI System
- Run RO before DI. RO clears the heavy load, so the resin only has to mop up the last traces. That single change extends DI run length by 10-20× compared with running DI on raw tap water, and it's the biggest lever you have.
- Respect contact time. Slower flow and adequate bed depth give the resin time to work and noticeably improve silica capture.
- Mind your materials. Ultrapure water is chemically aggressive; it wants to pull ions from whatever it touches. Use stainless steel (304/316), PVDF/PFA, or high-grade PP where it counts, and keep brass and copper out of the line downstream of DI.
- Watch quality, not just gallons. Track conductivity or resistivity, and in critical applications put a probe before and after the DI stage so you can see exhaustion coming.
- Plan the changeout. Rising conductivity after DI means the resin is spent. Color-change resin is a nice convenience, but confirm it with a meter. How long a charge lasts depends on your feed TDS and whether RO sits upstream; our DI capacity calculator estimates run length for your water.
Dual-point conductivity or resistivity monitoring (before and after DI) catches exhaustion early. Downstream of DI, favor 304/316 stainless, PVDF/PFA, or high-grade PP for any wetted part, and keep brass and copper out of the polish loop.
Common Misconceptions About DI
- "DI water is just distilled water." They both reach high purity, but by completely different routes: distillation boils and condenses, DI swaps ions. In the highest-purity plants you'll often see them used together with RO or EDI.
- "Softened water is the same as deionized." Not close. A water softener swaps hardness for sodium or potassium, so the TDS barely changes. DI removes the ions outright.
- "DI water is dangerous to touch." It's process water, not a beverage, but a splash on your hands won't hurt you. It's simply very low in dissolved ions. DI shows up everywhere, from reef tanks to semiconductor fabs, and you can see the full range of DI applications in our resin guide.
What Drives DI Performance, at a Glance
| Factor | Why It Matters | What to Do |
|---|---|---|
| EBCT & Bed Depth | More contact time means better exchange and silica capture | Keep flow conservative; give the bed enough height |
| Resin Selection | Grades differ in selectivity and leakage, especially for silica | Favor SBA Type I for low silica leakage |
| Flow Distribution | Prevents channeling so the whole bed gets used | Correct distributors and screens; load evenly |
| Upstream RO | Strips most of the ion load before DI, extending resin life | Place RO ahead of the DI polish |
| Temperature / CO₂ | Cold slows the kinetics; CO₂ quietly depresses resistivity | Degas if needed; read resistivity at 25 °C |
About the Author
Crystal Quest has designed and built water treatment systems in the USA since 1994, and we engineer DI and RO/DI trains in-house for homes, labs, and industry under an ISO 9001 quality management system. The guidance here comes from that hands-on experience. Almost every DI conversation we have starts with the same question: how pure does the water actually need to be? The answer drives the resin grade, the configuration, and the monitoring plan, in that order.
Speccing a deionization system?
Size the resin and run length for your feed water before you buy, or browse demineralizer systems engineered and built by Crystal Quest.
Frequently Asked Questions About Deionization (DI)
Is deionized water the same as distilled water?
No. Both end up very pure, but they get there differently. Distillation boils water and condenses the steam, leaving contaminants behind. Deionization swaps dissolved ions for H⁺ and OH⁻ that combine into water. For the highest purity, plants often run RO, DI, and sometimes distillation together rather than picking just one.
Does deionization remove bacteria, chlorine, or organics?
Not reliably. DI targets dissolved ions, not microbes, chlorine, or most organic molecules. That's exactly why a real DI train puts sediment and carbon (and usually RO) ahead of the resin, and why critical systems add UV or a final filter for biological control.
How pure can deionized water actually get?
On a fresh mixed-bed polisher after RO, very pure: resistivity near 18.2 MΩ·cm (about 0.055 µS/cm) at 25 °C, with TDS often in the 0-10 ppm range or lower. That 18.2 figure is the theoretical ceiling for pure water, so it's the target you chase, not a number you beat.
How do I control silica leakage?
Use Type I strong base anion resin and give the water enough contact time. Adding a mixed-bed polisher after a two-bed DI improves silica removal further, and degassing upstream helps when CO₂ is competing for the same sites.
Why is ultrapure DI water so hard on pipes and fittings?
Because it's "hungry." Water with almost no dissolved ions tries to pull ions from whatever it contacts, which can leach metals from the wrong plumbing. Keep DI water in compatible materials such as 304/316 stainless, PVDF/PFA, or high-grade PP, and avoid brass and copper downstream.
Is deionized water safe to drink?
DI water is meant as process water, not a beverage. A sip won't hurt you, but it's not balanced for taste or dietary minerals, so it's not formulated to be your everyday drinking water. Follow your application's guidelines.
How do RO and DI work together?
RO removes 95-99% of dissolved solids through a membrane, then DI removes the small amount that slips past to reach ultra-low conductivity. Running them in that order means the resin only handles the last traces instead of the whole load, which is why RO followed by DI is the standard recipe for ultrapure water.
