The Two Coolant Samples You Can Only Take Once

What the paperwork can prove

This is an illustrative example, not a customer case.

In our example, the delivery note lists two batch numbers. Besides the product data sheet, the envelope holds a safety data sheet, but no certificate of analysis.

A data sheet describes the product in general. Dow's guide for its DOWFROST LC coolants heads its property table “typical product specifications”. A footnote then says the values are not to be construed as specifications.⁠^[2]

A certificate of analysis, the supplier's record for one batch, says more. If you get one, check which batch it covers, whether it reports measured results or only states conformity, and which methods and concentration basis it uses. Even then, it describes the batch at dispatch, not your drums after transport and storage.

The label needs the same care. As VeriFluid's fluid intelligence glossary notes, “PG25” by itself does not define the product, the concentration basis, the inhibitor package (the corrosion-control additives) or the condition of the fluid.

The basis catches people out. Glycol can be stated as a percentage of volume or of weight, and the two differ: Dow's guide puts 25 percent by weight at 24.5 percent by volume. Its acceptable range for LC 25 is 24 to 27 percent by volume in the text, and 24 to 28 percent by weight in the in-use table.⁠^[2] Half a percentage point matters near the bottom of a range, so check the basis before comparing any two documents.

In-use limits are the wrong yardstick for new fluid

In-use limits are written for fluid that has already worked in a loop, so they can be much looser than what fresh product typically achieves. The table shows how much, for one product.⁠^[2]

Reserve alkalinity is a measure of the coolant's alkaline buffer, its capacity to neutralise acid. It is the volume of 0.1 N hydrochloric acid needed to bring a 10 mL sample down to pH 5.5.⁠^[3] ASTM adds two cautions: a higher number does not by itself mean a better coolant, and reserve alkalinity is not a dependable measure of corrosion protection or remaining life.⁠^[3] Compare it only with the same product's own figures. And because the pH range is the same in both columns, pH alone cannot show that a drum is as good as new.

Typical figures cannot fail a drum on their own, and passing in-use limits says little about fluid that should be fresh. Judge new fluid against the supplier's new-fluid or sales specification. The filling procedure, for instance, says not to use fresh fluid whose pH is outside the sales specification, but does not print it.⁠^[1] In the example, the team asks the supplier for that specification and a certificate for each batch.

Where the drums waited

Castrol's guide and its newer datasheet both ask for the drums to stay sealed, out of sunlight and below 30 °C.⁠^[1]⁠^[4] The guide adds a 36-month shelf life, after which it recommends testing before use.⁠^[1] For filling, the newer datasheet points to a technical bulletin and control plan,⁠^[4] and other suppliers set their own conditions, so ask for current documents with the order.

In the example, the site is in Chennai, the racks slip by three weeks, and the drums wait in a covered store without air conditioning. At the city's Nungambakkam observatory, the India Meteorological Department's 1991 to 2020 averages put the mean daily maximum above 30 °C in ten months of the year. It peaks at 37.3 °C in May.⁠^[5] Outdoor air is not store temperature. Still, it is reason enough to record where the drums waited and for how long, and to keep a simple min-max thermometer in the store.

Fill day: quick checks first

Start with three quick checks on each batch: pH, glycol concentration and a look at the fluid. The filling procedure opens with the pH check, and the newer datasheet asks for all three checks before use whenever its storage conditions cannot be met.⁠^[1]⁠^[4] Three weeks in a store that may have passed 30 °C calls for them.

Two simple instruments cover this. A hand-held refractometer gives an approximate glycol concentration and freezing point. If it reads refractive index, convert the reading with the fluid maker's table. If it reads percent glycol, calibrate it first and use the propylene glycol scale; the wrong scale gives wrong results.⁠^[2]⁠^[6] For pH, use a meter. The procedure repeats the hot product flush if the pH afterwards differs by more than 0.5, and pH paper calibrated to 0.5 units is too coarse to show that.⁠^[1]⁠^[2] Write the fluid temperature beside every reading. Refractive index changes with temperature, and later readings will be compared with this one.⁠^[2]

Keep the Day 0 bottle

The Day 0 bottle comes from a sealed drum, drawn with a clean tube just before the fluid is warmed and filled. Take one from each batch: two, in this example. The filling procedure asks for it to be marked “Day 0”, with batch and date.⁠^[1]

Draw enough for two portions. One goes for a new-fluid test, since reserve alkalinity, chloride, sulfate and hardness need a laboratory. VeriFluid's glossary calls that result the acceptance baseline. It arrives after the fill, so the on-site checks carry the fill decision and the laboratory confirms it. The other portion stays sealed, for a question nobody has asked yet. Agree the volume and container for each portion with the laboratory before fill day.

Drawn on fill day, the Day 0 bottle records the product after storage, so keep the storage record with it.

The same details, with storage, flush, fill-day, Day 1 and follow-up lines, are in the receiving record template (CSV).

Day 1: the first bottle from the filled loop

Day 1 comes after the fill and the on-site checks. If those checks find glycol low, the filling procedure allows a top-up with glycol concentrate, followed by a second air purge four hours later. It sets no other waiting time before Day 1.⁠^[1] For its own fluid, Dow asks for at least 24 hours of circulation before the baseline sample, drawn away from dead legs (pipe branches with little or no flow).⁠^[2] In the example, the team waits 24 hours, splits Day 1 into a laboratory portion and a sealed retain, and writes the hours on the label. VeriFluid's glossary calls the laboratory result the commissioning baseline.

Both guides want debris flushed out of new pipework first,⁠^[1]⁠^[2] and Dow warns that flush water left in the loop over-dilutes the fluid.⁠^[2] In the example, the contractor flushed with purified water before the product flush. So if Day 1 reads lower in glycol than Day 0, leftover water is the first suspect, not the delivery.

Six months later

Six months on, the routine on-site check shows glycol concentration below the in-use range. Was the delivery weak, was it diluted at fill, or has something changed in service?

The numbers on file cannot settle it. They came from a hand-held refractometer on fill day, a laboratory after it and now another site instrument, each at a different time and temperature. A small gap may come from the methods, not the fluid.⁠^[2]⁠^[6] Our Insight When the Sensor and the Lab Report Disagree shows how to check what each number measures before choosing which to trust.

Castrol's guide asks for a confirmation sample before acting on an out-of-specification result.⁠^[1] Send it to the laboratory with the retained Day 0 and Day 1 bottles, to be tested together by one method and on one basis. If the new sample reads in range, check the site instrument and the conditions of the earlier reading before deciding why they differed. If not, look at the bottle results. A low Day 0 points to the supplier, with the storage record alongside. A good Day 0 with a low Day 1 points to the fill. Two good bottles point to a change in service.

In service, look at top-ups first. Reliability Engine's guide to keeping PG25 ready for the rack shows why. In its illustrative 100 L loop, a 10 L water top-up takes glycol from 25 to about 22.7 percent. Log every top-up in the same record as the bottles.

The bottles also answer a harder question: is the fluid in the loop still the product that was delivered? The infrared (FTIR) fingerprint on our PG25 coolant testing page screens for a change in identity or formulation by comparing a sample with matched new-fluid and commissioning references. Here, those are the Day 0 and Day 1 bottles.

Store the retained bottles as the supplier asks for the product itself, and tell the laboratory how they were kept. With a 30 °C limit, that means an air-conditioned room in Chennai, not the drum store. Without the bottles, the team is left arguing about drums that were emptied months ago.

Common questions

Can we take a Day 0 sample later from a spare drum?

Only if a sealed drum from the same batch was kept back. Even then, it shows the product after more months in store, not as it went into the loop. Day 1 cannot be taken later at all: once the loop has run, its starting condition is gone.

How long should we keep the retained bottles?

For as long as the fluid stays in the system, which Castrol's guide expects to be 5 to 12 years, depending on system design and maintenance. That is longer than the 36-month shelf life it gives the sealed product,⁠^[1] so the bottles age too. Agree the container and storage with the laboratory, and tell it how old each bottle is.

Do we need another reference sample when the loop is topped up?

We suggest one whenever the top-up comes from a different batch, or is glycol concentrate rather than premix. Label it like a Day 0 bottle, and record which loop received it and how much.

Does the same approach work for immersion fluid?

The idea does, with different checks. Castrol's checklist for its single-phase immersion fluids asks for samples after the tank is filled and again after the IT equipment goes in. It asks for none from the delivery container, so a Day 0 bottle would be your addition.⁠^[7] One useful test on it is breakdown voltage, the voltage at which the fluid stops insulating; IEC 60156:2025 covers acceptance testing of unused liquids at delivery.⁠^[8]

References

Sources reviewed on 3 and 4 October 2026, with every link rechecked on 6 October 2026. The ASTM D1121-25 and D3321-26 summaries were first checked on 25 September 2026. Manufacturer documents are cited for their own named products, and their values are not general limits for other fluids. ASTM and IEC references use the publicly available scope and abstract pages, not the full purchased standards. Citing a method does not establish VeriFluid's accreditation or availability for a particular sample.

1. Castrol: ON Direct Liquid Cooling PG 25 operating guide (v2)Undated; PDF file created October 2025. Section 5 (pipework flushing), Table 3 (conditioning, filling, Day 0 and Day 1 retention samples) and sections 4 and 6 (storage, shelf life, expected lifespan, confirmation sample). Not linked from Castrol's current customer hub; the newer datasheet (reference 4) omits the fill procedure.
2. Dow: DOWFROST LC engineering and operating guideForm 176-01641-01-0623, June 2023, distributor-hosted copy. Tables 1, 4 and 8 and sections 2.4, 4.1, 5.1, 5.3 and 5.4: typical versus in-use values, concentration basis, flushing, field checks and the 24-hour baseline. The guide does not mention certificates of analysis.
3. ASTM D1121-25: Reserve alkalinity of engine coolants and antirustsPublic scope summary. Definition of reserve alkalinity and what it indicates (sections 5.1, 5.2, 5.5 and 5.6), and the cautions in sections 5.2 and 5.7 against reading it as a measure of coolant quality, corrosion protection or remaining life. Engine-coolant scope.
4. Castrol: ON Direct Liquid Cooling PG 25 extended technical datasheetUndated; PDF file created July 2026. Storage conditions, on-site testing when they cannot be met, and a pointer to a technical bulletin and control plan for filling and monitoring. No shelf life or expected lifespan.
5. India Meteorological Department: Chennai (Nungambakkam) extreme weather events and climatological table1991 to 2020 normals: mean daily maximum by month, outdoor air at one observatory. The extreme-events section at the top of the sheet covers a single month (still September on 6 October 2026); only the normals table is cited.
6. ASTM D3321-26: Handheld refractometers for freeze point and percent glycol of aqueous engine coolantsPublic scope summary. Covers ethylene and propylene glycol coolants in cooling systems; approximate field result; calibrate before use and use the correct glycol scale.
7. Castrol: ON single-phase immersion cooling fluids, checklist and general troubleshootingUndated; PDF file created October 2024. Samples after filling and after submersion, filtration and testing before IT equipment, label and delivery-note check. No sample from the delivery container.
8. IEC 60156:2025: Insulating liquids, determination of the breakdown voltage at power frequency, test methodPublished 31 January 2025. Public abstract: acceptance testing of unused liquids at delivery and samples in service. A test method, not limits; the abstract does not mention immersion coolants.

Related-site note: VeriFluid's liquid-cooling training is offered through Reliability Engine. The Reliability Engine article above is linked for its worked top-up example; it is not evidence of a VeriFluid integration.