Resistivity is the spec
Deionization passes water through cation and anion exchange resins that swap sodium, calcium, iron and copper for hydrogen, and chloride, sulfate and bicarbonate for hydroxide, leaving water whose resistivity is read in megohm-centimeters. A mixed bed with polishing reaches 18.2 megohm, which is 0.055 microsiemens and the theoretical limit. That water will restore its equilibrium by dissolving whatever it flows over, and an exchanger is a long narrow pipe with a lot of surface.
The parent page explains pitting in low-conductivity water. What a DI loop adds is leaching: metal that goes into solution with no visible attack and shows up as falling resistivity at the loop return and as copper on a wafer. This page is about that, resin temperature limits and loop velocity.
Copper is the first thing a DI loop finds. A copper-brazed plate, a bronze valve body or a brass gauge fitting puts copper into 18 megohm water within hours, and copper on a wafer or in a cell culture is a ruined run. Nickel-brazed plates are better and still not right: the braze is unpassivated and its edges are crevices. The rule is stainless or better everywhere the water touches: 316L tubes, tubesheets and heads, electropolished and passivated, or titanium where the site's metals specification excludes stainless.
Elastomers leach too. Nitrile gives up plasticizers and silicone gives up siloxanes into low-TOC water. HeatX uses EPDM or PTFE-encapsulated gaskets on the utility side only and no gasket at all in the DI path on a straight-tube unit.
Hot DI water for parts cleaning or hot rinse is heated after the resin beds, never before, because anion resin loses capacity and sheds amines above about 140 F while cation resin tolerates 250 F. The heater sits on the loop supply downstream of the mixed bed and the final filter; the return to the tank goes through a cooler so the bed never sees hot water. On a loop with a UV sterilizer, the heater sits after the UV as well, because the lamp's transmittance is rated at ambient.
A DI loop is kept moving at 3 to 5 ft/s so a biofilm cannot establish. The exchanger is part of that loop: tubes sized for that velocity, no shell-side pockets, and a low-point drain so the unit empties for hot water or peroxide sanitization. The loop cooler removes pump heat, roughly 2,500 Btu/h per horsepower; on a 200 gpm loop with a 20 hp pump that is a 50,000 Btu/h cooler holding 72 F on 45 F chilled water.
Shell diameters run from 2 in for a lab loop to 48 in and 65 ft long for electronics rinse trains, in straight tube, U-tube, multi-pass and double tubesheet forms.
A copper-brazed plate bought as a loop cooler because it is compact and cheap. Resistivity at the return drops from 18 to 15 megohm within a week, the mixed bed is blamed and changed, and the number keeps falling because the copper source is downstream of the bed. A 316L straight-tube cooler costs more on the day and is the last time anyone thinks about it.
Call and talk it through with an engineer: 1-805-484-2992
Resistivity, velocity and wetted materials are all fixed in the rows, since deionized water attacks whatever it touches. Compare your loop requirements and send them with your quote request.
| Condition | Detail |
| Fluid | Deionized water, 1 to 18.2 megohm; TOC under 50 ppb on semiconductor and lab loops |
| Flow range | 5 to 100 gpm laboratory and pharma loops; 100 to 1,500 gpm semiconductor and electronics rinse |
| Temperatures in / out | Loop held 68 to 77 F; hot DI at 140 to 160 F for parts cleaning; cold DI at 40 to 50 F for rinse |
| Hold | None; return to the storage tank |
| Utility | Chilled water or glycol on coolers; hot water or steam through a hot water loop on heaters |
| Approach | 5 to 8 degrees on chilled water; 10 on hot water |
| Construction | Straight tube in 316L, DI water in the tubes; double tubesheet where the utility is tower water or plant steam; titanium tubes where the metals spec excludes stainless |
| Finish / class | 316L, 15 to 20 Ra electropolished and passivated; polymer-lined or PVDF-shell options where required |
| Velocity | 3 to 5 ft/s in the tubes, loop never stagnant |
| CIP / SIP | Hot water sanitization at 176 F, hydrogen peroxide or ozone on schedule; gaskets rated for the sanitizer |
Flush it to drain until the resistivity at its outlet matches the supply, instead of letting the polishing bed absorb the start-up load. A properly degreased and passivated 316L unit settles in minutes at loop velocity; if the outlet keeps lagging the inlet after the flush, look for a brass guage fitting or a nitrile seal in the new piping before blaming the passivation. Then sanitize the whole loop, because the new unit arrived clean but not sterile.
Because resistivity depends strongly on temperature and the meter compensates to 25 C, so a probe where the water is still changing temperature reads wrong. Ultrapure water that reads 18.2 megohm at 77 F reads about 24 megohm at 68 F and about 14 at 86 F with compensation off, so an uncompensated probe at a cooler outlet that drops 10 degrees shows a jump that is not purity. Check that both probes compensate, that their temperature elements are in the flow and not in a stub, and that the sample is at steady temperature before comparing. If compensated readings at steady temperature still differ accross the cooler, then the unit is adding ions and the braze, gaskets or passivation are the suspects.
Yes, from fabrication oils, from elastomers in the water path and from a cleaning step that was not rinsed out. On a loop held under 50 ppb TOC a drawing lubricant left inside a tube is visible for weeks. Ask for oil-free fabrication, a degreasing step before passivation, and a straight-tube design with no gasket in the DI path, which takes elastomers off the suspect list entirely. Where the site has a TOC analyzer at the return, run it continuously through the first week after installation and keep the trace with the commissioning record.
Not the metal; passivated 316L handles both at loop sanitization strengths. The parts that fail are the instrument seals and any clamp gasket touching the water, and residual ozone has to be destroyed by UV before it reaches the use points or the exchanger downstream of them. Tell us the sanitizer, its concentration and contact time so every seal is specified for it; EPDM that is fine for peroxide is not the right pick for continuous ozone. Send the sanitizer data with the loop data.
Cooling and heating water that has to stay sterile after the still, from the hot loop to the last valve.
+ Learn More
Feed heaters that respect the membrane limit, and permeate exchangers for water that is acidic and empty.
+ Learn More
Heating makeup for the boiler, cooling water for the resin, and cooling the regeneration waste for the drain.
+ Learn More