Water for Injection Heat Exchangers

Joint, finish, rouge, endotoxin and the source of the water

The tube-to-tubesheet joint is the whole argument

Every WFI exchanger has a wall between compendial water and plant steam or cooling water, and the wall is a tube 0.049 in thick. The joint where the tube enters the tubesheet is the weak point. A rolled joint weeps after enough thermal cycles. A welded single tubesheet still has one face on the steam side, and a crack there is invisible until a sample fails. HeatX builds WFI units with two tubesheets 1 in apart, tubes welded to both, the gap open at the bottom so a leak drips out where an operator sees it.

The product-side tubesheet face is a compendial-water surface too: tube welds are ground flush and polished to the tube-bore finish, because a weld crown is a crevice the borescope will find.

Finish, and what electropolishing actually does

Mechanical polishing to 15 to 20 Ra leaves smeared metal and embedded abrasive. Electropolishing removes about 0.0005 in of that layer, takes the peaks off and leaves a chromium-rich surface that passivates better and holds rouge off longer. ASME BPE SF4 calls for 15 Ra maximum mechanically polished then electropolished, and that is the finish HeatX quotes for hot WFI unless the site standard says otherwise. Every readable surface is measured with a profilometer and logged by location, because the validation package needs numbers.

Rouge is not a defect, but it is your problem

Hot WFI on stainless grows a red, blue or black iron oxide film. Class I rouge migrates from a pump or weld; Class II grows in place; Class III is the magnetite of a pure steam system. None fails the water on conductivity or TOC, but a heavy film sheds particles, and a QA group that finds red in a filter housing wants it gone. The exchanger is built so a chelant derouging solution and a passivation reach every surface and drain: vertical or a 3 degree slope, no horizontal tube, a drain at the true low point. A unit that cannot be derouged in place gets pulled, and that is a shutdown.

Endotoxin, and why bacteria are the real target

Endotoxin is the cell-wall fragment of Gram-negative bacteria; it survives 121 C and no exchanger removes it. The exchanger's job is to refuse to grow the bacteria that make it: no place for a biofilm to start, which is the finish and the welded joints; no water standing still, which is drainability and loop velocity; and the 80 C hold. A WFI exchanger that runs cold for long periods, such as a use-point cooler left full between draws, is where endotoxin excursions come from. The common mistake is validating the still and treating the exchanger as pipe.

Still-fed and membrane-fed systems

A multi-effect or vapor-compression still delivers WFI at 90 to 98 C, and its exchangers are the feed preheater, the condenser and the distillate cooler to storage temperature. Membrane WFI, allowed under the monograph with a validated RO, EDI and ultrafiltration train, delivers ambient water, so the system needs a heater rated to lift the whole make-up flow to 80 C plus a loop heater to hold it. Both are double tubesheet duties; the membrane-system heater is the larger unit because it heats every gallon made.

Send the make-up rate, storage temperature, steam pressure and finish specification. Call and talk it through with an engineer: 1-805-484-2992 The condenser on a WFI still is covered on our WFI condenser pages.