Why Feedwater Conductivity and CO₂ Determine EDI Performance: Thresholds, Root Causes, and How to Validate Your System

Imagine you are responsible for the ultrapure water loop at an electronics fabrication plant. For months, the electrodeionization (EDI) unit has delivered a stable 18 MΩ·cm resistivity. Then, over a few days, the resistivity begins to fall—first to 15 MΩ·cm, then to 10 MΩ·cm. The upstream reverse osmosis (RO) system still shows a permeate conductivity of 4 μS/cm, well within the typical EDI feed specification. You ask yourself: “If the RO is performing fine, why is the EDI failing?”

This scenario repeats across pharmaceutical, power generation, and other high‑purity water installations. The problem often lies in two parameters that are easy to overlook: feedwater conductivity and dissolved carbon dioxide (CO₂). Understanding why they matter—and how to verify whether they are the source of the trouble—is essential for anyone who specifies, operates, or maintains an EDI system.

How EDI Works and Why Feedwater Quality Is Non‑Negotiable

EDI combines ion exchange resins, ion‑selective membranes, and an applied DC electric field to continuously remove residual ions from RO permeate. Unlike conventional mixed‑bed demineralizers, EDI does not need periodic acid‑base regeneration. However, this advantage comes with a strict condition: the feedwater entering the module must stay within a narrow chemical window.

One of the most important design principles is that EDI modules cannot be operated reliably without proper pretreatment and RO upstream. The RO stage removes the bulk of dissolved salts, but the permeate still carries a small ionic load, plus dissolved gases that RO membranes do not reject. If the pretreatment or RO performance drifts, the EDI module receives a higher-than‑designed ionic burden, and its resistivity output begins to suffer.

Feedwater Conductivity: The Obvious but Misunderstood Parameter

Conductivity is a measure of the total ionic load that the EDI stack must handle. When the RO permeate conductivity rises—say from 2 μS/cm to 8 μS/cm—the EDI module must transport more Na⁺, Cl⁻, Ca²⁺, and HCO₃⁻ ions across the resin and membranes. This can lead to three problems:

• Resin overload
• in the concentrating compartments, causing ion leakage into the product water.
• Localized scaling
• on the concentrate side if hardness is present, even at low concentrations.
• A drop in product resistivity
• because the electric field can no longer maintain the required rate of ion removal.

Many EDI manufacturers recommend a feed conductivity of ≤ 20 μS/cm, but this is not a universal “safe” number. The actual limit depends on the module design, the ratio of monovalent to divalent ions, and the target product resistivity. A system aiming for 10 MΩ·cm may tolerate a higher conductivity than one designed for 18 MΩ·cm. The key is to trend RO permeate conductivity over time and correlate it with EDI performance—not simply check a static threshold.

An often‑overlooked cause of rising feed conductivity is the raw water source itself. The conductivity of raw water varies significantly depending on whether it comes from groundwater, surface water, or a municipal supply, and this variation directly affects the RO permeate quality that feeds the EDI. A change in the raw water source, or a seasonal fluctuation, can silently push the EDI beyond its design envelope.

CO₂: The Invisible Ionic Load That Conductivity Meters Miss

Carbon dioxide is a neutral gas that passes through RO membranes almost unchanged. When CO₂ dissolves in water, it forms carbonic acid (H₂CO₃), which partially dissociates into H⁺ and HCO₃⁻ ions. These ions are not detected by a standard conductivity meter as strongly as salts like NaCl, but they are electrochemically active inside the EDI module.

Inside the EDI stack, the weak acid H₂CO₃ dissociates further under the influence of the electric field, consuming current and producing H⁺ and HCO₃⁻ that must be transported through the membranes. This effect:

• Increases the true ionic load without a proportional rise in measured feed conductivity.
• Reduces the effective current efficiency of the EDI, because part of the current is used to split CO₂‑derived species.
• Can cause a noticeable drop in product resistivity even when the feed conductivity appears stable.

For this reason, many EDI system suppliers specify a feedwater CO₂ limit of ≤ 5 mg/L (or ≤ 2 mg/L for the most demanding applications). If the RO feedwater comes from a source with high alkalinity, or if the RO system is not preceded by effective degasification, the CO₂ load can silently degrade EDI performance.

Practical Validation: How to Diagnose Conductivity and CO₂ Issues

When you suspect that feedwater quality is affecting EDI performance, avoid jumping to a module replacement. Instead, follow a structured validation sequence:

1. Trend RO permeate conductivity. If it has increased by more than 30–50% from the baseline, investigate the cause: fouled RO membranes, aging elements, or a change in raw water quality.
2. Measure CO₂ in the feedwater using a titration method or a dissolved CO₂ sensor. Compare the result with the EDI module manufacturer’s recommendation. If the value exceeds 5 mg/L, consider adding a degasification step (membrane degasifier or forced‑draft degasifier) before the EDI.
3. Check the EDI product resistivity at multiple points—immediately after the module, after the polishing loop, and at the point of use. A drop that is seen only after the module but not after a downstream polishing mixed bed can indicate excessive ionic load from the feed.
4. Perform a short‑term “stress test.” Temporarily reduce the feed flow to the EDI stack while keeping the voltage constant. If the product resistivity improves, the module is likely overloaded by the current feed conductivity or CO₂.

Design and Operational Boundaries You Should Respect

Key feedwater parameters for EDI extend beyond conductivity and CO₂. A complete specification also covers hardness, silica, residual chlorine, temperature, and flow rate. None of these can be ignored if you want stable high‑purity water production. For example, even a low level of hardness can cause scaling on the concentrate side, while excessive silica can foul the membranes.

A practical approach is to always design the EDI feedwater quality from the raw water analysis outward. At Chuxin Mingwei, we map the entire treatment chain—from multimedia filtration and activated carbon to RO and EDI—so that the EDI module receives water within the ranges it can reliably handle. This prevents the common mistake of treating the EDI as a stand‑alone “polisher” and forgetting that its performance is entirely dependent on what happens upstream.

For high‑purity applications, we configure the RO stage and pre‑treatment specifically to control conductivity and CO₂, and we can integrate forced‑draft or membrane degasification when source water conditions demand it. The result is a system that maintains stable ultrapure water quality without unpredictable resin consumables or frequent module replacement.

Decision Checklist for Your Team

• [ ] Have you established a baseline for RO permeate conductivity and CO₂ level under normal operating conditions?
• [ ] Is the RO system currently performing within its design rejection rate? (Check normalized permeate flow and salt rejection.)
• [ ] Does the feedwater CO₂ consistently stay below 5 mg/L? If not, has a degasification step been evaluated?
• [ ] Is the EDI module’s current draw within the manufacturer’s recommended range? A high current with low resistivity often indicates an ionic loading problem.
• [ ] Have you ruled out mechanical issues (e.g., internal leaks, damaged membranes) before attributing the problem to feed chemistry?

If you are currently seeing a slow decline in your EDI product water quality, or if you are specifying a new system and want to avoid the common pitfalls of CO₂ and conductivity overload, our team can help you review your water chemistry data and size the right equipment configuration.

Next step: Contact us with your raw water analysis and target water specification. We will provide a technical assessment of the required pre‑treatment and EDI feed conditions—based on your real water, not assumptions.