Lab notebooks rarely make it into marketing. That is a pity, because the interesting failures live there. This piece is written the way the work actually went: a plausible idea, a closer look, an unexpected result, and a better specification at the end. The subject is the number every buyer now asks for first in glycol liquid cooling — conductivity — and the protective film it can quietly destroy when it is chased too far. The notebook below belongs to the formulation team at Glacier Coolant, and the conclusion it reaches now sits inside products like LM-425D-YS.

The hypothesis everyone starts with

A cold-plate secondary loop sits close to live electronics. Water-based coolant conducts. Therefore, the reasoning goes, the best data center coolant is the one that conducts least. Purchasing writes a number into the RFQ — as low as achievable — and suppliers oblige on paper.

Half of that reasoning is sound. Conductivity does set the scale of damage when fluid reaches a board, and it does gate the galvanic current that flows between dissimilar metals in the loop. Those facts are why low conductivity earns its place on every datasheet, including ours for LM-425D-YS. The unsound half is the word “therefore”.

The closer look: what ions are doing in there

Coolant is not water pretending to be clean. It is a working solution full of things that must be there: corrosion inhibitors, pH buffers, biocides. Every one of those functional additives arrives with ions attached. The ionic content of a healthy coolant is not contamination. It is the delivery system for protection.

The film-forming inhibitors deserve the closest look. Their job is to adsorb onto metal surfaces and build a dense, persistent layer — the barrier between glycol solution and copper, aluminium or steel. That chemistry needs ionic partners in the fluid to proceed. Pull the ion content toward zero and the first casualty is not the conductivity reading. It is the film.

The first row of that table has a name in our lab notes: running bare. A fluid can post a beautiful conductivity figure and protect nothing. The spec was met; the loop was not.

The unexpected result: protection without purity

Walk the logic of a real loop and the paradox resolves. The fluid circulates past copper plates and aluminium manifolds. Two dissimilar metals, one electrolyte — a galvanic pair the moment current finds a path. Conductivity sets how easily that current flows, which is why it belongs in the specification.

But corrosion protection is a surface event, not a bulk event. What shields the plate is the film on the plate. A fluid that holds a controlled ionic band while keeping film-forming chemistry active beats a fluid that wins a conductivity contest and leaves metal naked. The band is narrow. It sits low enough to limit leak damage, high enough to feed the film, and steady enough to trend honestly. Products are engineered into that window; they do not drift into it. Narrow windows need watching. That is what the monitoring plan is for. The finding that reshaped our formulation targets is old and unglamorous: the goal is not the lowest number. It is a moving equilibrium — free ions held in a managed range while active inhibitor keeps repairing the film.

What actually threatens the film

A second lab-note theme: the loop is not a beaker on a shelf. Metal surfaces give up trace ions slowly. Pump impellers whip in bubbles and shear the boundary layer. Flow scours. Joints micro-leak and make-up water enters with its own mineral load. Each of these wears the film in miniature.

Notice what is missing from that table: a conductivity alarm. By the time bulk conductivity moves, local chemistry has often already turned. The bulk number is an average, and averages are polite about local disasters. A pinhole site reads the same reservoir as the healthy loop around it, right up until the pinhole wins. This is why a point of local film breakdown matters more than the average reading. Where the film fails, a micro corrosion cell forms, local ion concentration spikes, and pitting starts — a pinhole factory running inside an in-spec loop.

Writing the trade-off into a spec

The specification line that captures all of this is short. It should read something like: “Coolant shall maintain conductivity within the supplier’s defined band for the fill life, with a film-forming inhibitor package verified on copper, aluminium and carbon steel, and residuals tested on a schedule.” Three commitments live in that sentence: a band rather than a floor, chemistry verified on your metals, and monitoring that keeps both honest.

Bands need a supplier willing to publish one. We publish the starting point for LM-425D-YS and define the warning lines in the monitoring plan that ships with the first drums. The operating-side numbers and their cadence are covered in the conductivity red line; the galvanic side of the story, where two metals meet one electrolyte, is treated in our corrosion piece.

A note on the word “dielectric”

Search lists for liquid cooling are full of dielectric coolant fluid and dielectric liquid. The words mean something precise for engineered immersion fluids, and something looser in glycol discussions. A water-based secondary coolant is not dielectric and should not be sold as one; what it can be is low-conductivity, clean, and monitored — which is what actually limits damage in a cold-plate loop. If a datasheet implies dielectric behaviour for a glycol fluid, ask what is meant. Precision here is a sign of a supplier worth having.

The conclusion the notebook ends on

The hypothesis survived in weakened form. Low conductivity matters. As a sole target it dismantles the protection that matters more. The specification that works names a band, a film, and a schedule. Formulate to the equilibrium, monitor the trend, and the number takes care of itself.

Four clauses, one page, no poetry. Every one of them is checkable at delivery and re-checkable in year five, which is what separates a specification from a wish.

Glacier Coolant has worked this equilibrium since 1994 — the M3 super-film rust inhibition technology and the Modify2000 compounding technology both exist to hold films in real flowing loops, and the Box7 full-scope test method checks a candidate against the loop’s duty before delivery. Box7 is the name for that set — metals, temperatures and flow are checked together instead of one property at a time. More than 30 invention patents and 4,000 customers later, the lab-note conclusion has not changed.

FAQ

Is lower conductivity always better for coolant?

No. Below the level where film-forming inhibitors can work, lower conductivity strips protection. The fluid then meets the spec and fails the loop. Specify a supplier-defined band with active inhibition, not a floor.

Why does coolant conductivity matter in liquid cooling?

It sets the scale of damage if fluid reaches live electronics and gates the galvanic current between dissimilar metals. Both effects scale with how well the fluid conducts, so the number belongs on the datasheet.

Can coolant be too pure?

Yes. Deionized water is aggressive toward metals and carries no protection. A coolant stripped of ions cannot support the surface films that actually prevent corrosion. Purity is a starting condition, not a protection strategy.

What conductivity should glycol coolant have?

Whatever band the supplier defines and stands behind, with a documented starting point and warning lines for service drift. Ask for the band in writing with the corrosion data behind it.

What is a micro corrosion cell?

A localized corrosion site where the protective film has broken down. The spot becomes an anode against the surrounding protected surface, and pitting can perforate thin cold-plate walls while the bulk fluid still tests in spec.

Does pre-diluted coolant help conductivity control?

It helps consistency. Factory pre-dilution removes make-up water variation, one of the main drivers of conductivity drift in service. The drum-to-drum chemistry stays where the spec put it.

Do dielectric fluids solve this?

Dielectric fluids belong to specific applications, such as engineered immersion tanks. A cold-plate secondary loop runs water-based coolant by design; the working levers there are the conductivity band, the inhibitor package and monitoring.

How do I monitor the film if conductivity lags the damage?

Test inhibitor residuals and pH on schedule alongside conductivity, and treat two consecutive adverse readings as an action line. The film’s health shows in residuals before it shows in the bulk number.

Send the spec, get the band

Tell us your conductivity target, your wetted-metal list and your fill-life expectation. We will return a one-page spec clause with the band, the film verification data and the monitoring cadence. Start from the product collection, or read how the operating team keeps the number honest in the conductivity red line.

About Glacier Coolant

Glacier Coolant is the industrial coolant brand of Chaoyang Guangda Chemical Co., Ltd., Since 1994 the company has built secondary refrigerants and industrial cooling fluids, with the M3 super-film rust inhibition technology, the Modify2000 compounding technology, the Box7 full-scope test method, more than 30 invention patents, and over 4,000 customers served.