Hard feed water and a purification skid

A purification skid does not meet a generic "lab water." It meets a particular stream at the first fitting. That stream is the feed, and two feed properties decide whether the membranes live: hardness, meaning calcium and magnesium and the scale they form, and oxidant residual, meaning chlorine and chloramine. This is the workflow for reading those facts and branching when the skid objects. Product grades and the tests they do not cover are in laboratory water types and where they fail.

Downstream, that product is only as useful as the buffer you make with it. The pH steps sit in preparing a buffer and checking pH. Chemicals for buffers, not improvised membrane cleaners, come from the reagents and chemicals catalogue. Collect product water in a known vessel from beakers and flasks. When you are specifying a system, put the feed data on the quotation request.

Who this workflow is for

Use it when you are commissioning a skid, writing the feed limits into a method, or explaining a decline in permeate quality. The decision is whether this feed is inside the envelope the manufacturer published, and what class of pretreatment stands between the feed and a thin-film membrane. It is not a licence to mix a cleaner from memory, and it is not a substitute for the skid manual.

If facilities deliver pretreated water that is already soft and dechlorinated, your feed is that pretreated water, and you still sample it at your inlet. A municipal number quoted from another building is not your feed.

What hardness and oxidant do to the membrane

Hard feed water means the calcium and magnesium load is high enough to matter for scale. On a reverse-osmosis membrane the salts concentrate on the reject side. Calcium carbonate and related magnesium scales block area, raise differential pressure, and let more salt into the permeate. The product then looks ionically worse, and the cause is a solid on the membrane, not a mystery about ionic strength in the beaker you have not made yet.

Chlorine and chloramine are oxidants. Thin-film polyamide reverse-osmosis membranes are damaged by them. The damage is chemical, not a scale you can backflush as if it were silt. Once the barrier is oxidised, salt rejection falls. Carbon pretreatment is the usual class of answer in laboratory skids: a carbon bed, sometimes a catalytic carbon where chloramine is the residual the manufacturer calls out, sized for the contact the manual requires. Some industrial designs use a chemical reductant instead. That is a different class, with its own residue and its own safety data sheet. Choose the class the skid was built for. Do not invent a dose.

Particulates and iron are neighbouring feed problems. A filter ahead of the membrane is a class of protection against fouling that is not hardness and not chlorine. Name it if the manual names it. Do not make one pretreatment word cover three mechanisms.

The inlet measurements worth writing down

Sample at the pretreatment inlet. For hardness, record calcium and magnesium, or hardness expressed in a unit you define, by a method the laboratory can repeat. A single word "hard" does not travel. For disinfectant, record free chlorine and total chlorine if both are in your kit, because chloramine hides in the gap between a free-chlorine test and a total-chlorine test. Note the temperature of the feed and the time of day if the building supply swings.

Write the manufacturer's feed envelope beside those numbers: maximum hardness, maximum oxidant ahead of the membrane, required carbon type, and any softener or antiscalant the design includes. Softening by ion exchange, which trades calcium and magnesium for sodium, reduces scale risk and changes the ionic mix the membrane then sees. Antiscalant is a chemical class the manufacturer doses. Either can be legitimate. Neither is universal. The manual picks.

Product release stays on product tests. Conductivity or resistivity at 25 °C reports ions in the permeate or at the polisher. TOC, if specified, reports oxidisable organic carbon. A microbial or endotoxin line is a further test. Feed hardness can be unacceptable while today's product still looks polished, which means you are consuming membrane life or pretreatment capacity. Feed hardness can be acceptable while product resistivity is poor, which means you look at cartridges, reject flow, and leaks rather than blaming the municipal calcium you have already measured.

The operating sequence and the failed-control branch

Confirm the pretreatment is in the flow path you think it is. A bypassed softener or a valved-around carbon bed is a different skid. Verify that carbon-outlet oxidant is at or below the membrane limit before you celebrate the product meter. Verify that differential pressure and permeate conductivity are inside the trend you recorded at commissioning.

Then dispense product for the assay only after the point-of-use reading meets the method. Rinse the collection flask. Cap it if it must wait, and accept that waiting is a new process. Buffers made from that water still need their own pH and salt records.

Branch as soon as a control moves. If oxidant appears at the carbon outlet, stop production through the membrane. Replace or regenerate carbon by the procedure the manufacturer gives, then retest. Do not "dilute" breakthrough by opening the reject valve and hoping. If permeate conductivity rises and hardness scale is plausible, follow the cleaning or replacement path in the manual. A home-made acid wash with an arbitrary concentration is how membranes and people both get hurt. If differential pressure rises after a change in feed silt, the particle filter is the first hardware to inspect. If product resistivity is excellent and the assay blank is not, leave the hardness hypothesis and move to organics, the vessel, or the assay. Hardness does not explain a nuclease.

How the failure shows up later in a buffer

A scaled or oxidised membrane that still limps along may deliver water that is "good enough" for a rinse and poor for a blank. The buffer you then prepare inherits that water. pH adjustment will not remove a leachate, and ionic strength calculations assume the salts you weighed, not a variable drip of hardness ions. If a buffer conductivity trend drifts without a recipe change, include the water system in the investigation beside the balance. The recipe check still comes first: hydrate form and a missed salt explain more buffer failures than a municipal hardness swing.

Reject water is a concentrate. It is not a spare feed for a sensitive assay, and it is not a casual drain substitute for a reagent bottle. Route it as the installation specifies.

Safety and institutional limits

Pretreatment chemicals, membrane cleaners, and oxidant-quenching agents are hazardous until the safety data sheet says otherwise. Acid, caustic, and sulfite-class reductants need eye protection, the right materials of construction, and the manufacturer's procedure. This overview does not supply a concentration, a soak time, or a neutralisation recipe.

The skid does not set biosafety policy. If process water could have contacted biological waste, the institution decides containment and disinfection. Product water for a research reagent is still a research material. It is not a diagnostic device and not an approval to drink the dispense.

Feed water in a warm building

Heat changes saturation and makes a closed plant room a hard place to trust an unlabelled jug. Sample the feed you have this month. Disinfectant residual and hardness both move when the supply changes, including after maintenance you do not see. A power cut that stops a booster pump or a softener regeneration can leave you with a different inlet the next morning. After the restart, confirm oxidant ahead of the membrane and product conductivity before you resume the fussy assay. Write the observations. Skip the temptation to describe the whole region as hard or soft. The inlet analysis is the fact.

What belongs on the skid enquiry

State the feed analyses you already have, with units and methods: calcium, magnesium or hardness, free chlorine, total chlorine, and any iron or silt note the manufacturer requested. State the product specification by document and test, the volume per day, and the assays downstream, including buffers that need a defined water at the point of use. Ask which pretreatment class is included for scale and for chloramine, and what residual must be shown at the membrane inlet. Ask what the manual requires after an idle period. Those lines can be checked. A promise that the water will be "pure enough for any buffer" cannot.