How coolant quality affects data center liquid cooling reliability

Searching for “liquid cooling for data center reliable” solutions? Start with the coolant itself. Fluid chemistry affects heat transfer, corrosion, fouling and the long-term performance of cooling distribution units (CDUs), manifolds and heat exchangers. A correctly sized cooling system can still develop restricted flow or damaged components if the fluid is incompatible or contaminated.

For operators, the challenge is not simply buying clean coolant. It is keeping that coolant within an approved operating envelope through commissioning, maintenance, top-ups and equipment changes. That requires clear acceptance limits, representative sampling and an understanding of what each test can—and cannot—tell you.

Define the circuit before specifying the fluid

In direct-to-chip installations, a CDU commonly separates the facility-water circuit from the technology cooling circuit through a heat exchanger. These circuits may have different materials, operating conditions and water-quality requirements. Facility water that is acceptable on one side is not automatically suitable for cold plates on the other.

This distinction also prevents a common misunderstanding: direct-to-chip cooling and immersion cooling do not share a universal coolant specification. Immersion fluids require separate evaluation of dielectric performance and material compatibility. Here, the focus is on water-based, single-phase cooling circuits, including approved water-glycol mixtures.

The starting document should identify every wetted material: metals, seals, hoses, adhesives and quick-disconnect components. A fluid approved for a CDU alone may not be approved for the complete rack loop.

The properties that deserve operational attention

Coolant quality is a combination of chemistry and cleanliness. No single reading proves that a loop is healthy.

• pH and inhibitor condition: Keep both within the fluid and equipment suppliers’ approved ranges. An acceptable pH does not prove that corrosion inhibitors remain effective.
• Conductivity: Use it as a chemistry indicator, not a universal pass/fail measure. Inhibited fluids can have higher conductivity than untreated purified water by design.
• Hardness, chloride and other dissolved constituents: Depending on concentration, temperature and materials, these can contribute to deposits or corrosion. Limits must come from the approved system specification.
• Particles: Construction debris, corrosion products and degraded material can obstruct cold-plate passages or accumulate at strainers.
• Glycol concentration: Where glycol is required, verify the approved concentration. Excess glycol increases viscosity and can reduce heat-transfer performance; insufficient concentration may compromise the intended protection.
• Microbiological condition: Closed circuits are not automatically immune to biological growth. Assess the risk with the fluid specialist, particularly after contamination or prolonged shutdown.

Deionized water is not a complete corrosion-control strategy. Once introduced into a system, it interacts with wetted materials and contaminants. Its suitability depends on the equipment specification and the complete treatment program, not its initial purity alone.

Why small chemistry problems become cooling problems

Corrosion does more than threaten leakage. The products of corrosion can travel downstream and collect in narrow passages. Deposits then create thermal resistance or reduce flow distribution, potentially leaving individual devices hotter even when the CDU’s overall supply temperature appears normal.

Operators should therefore read chemistry results alongside hydraulic and thermal trends. Rising differential pressure across a filter may indicate loading. Higher component temperatures at comparable load, inlet temperature and flow warrant investigation. Neither observation proves a coolant problem on its own: valves, pumps, sensors and air pockets can produce similar symptoms.

The energy consequence matters too. Additional pumping or lower supply-temperature settings may temporarily mask deteriorating performance while increasing operating demand. For energy-efficient infrastructure and projects using renewable electricity, protecting heat-transfer surfaces remains necessary; a cleaner power supply does not remove avoidable cooling losses.

Turn guidance into a project-specific acceptance specification

ASHRAE liquid-cooling guidance can support system planning, but it should not be treated as a substitute for current equipment and coolant documentation. There is no defensible universal pH, conductivity or particle limit for every data center liquid-cooling loop.

Before filling, obtain a written specification agreed by the relevant equipment and fluid suppliers. It should state the approved formulation, dilution-water requirements, material compatibility, chemistry limits, filtration requirements, sampling methods and corrective actions. Resolve conflicting requirements before commissioning rather than selecting whichever limit is easiest to meet.

For equipment spanning several interfaces, ownership needs equal attention. Shandong Liangdi Energy Saving Technology Co., Ltd., based in Changqing Industrial Park in Jinan, develops, designs, manufactures and services CDUs, water distribution manifolds, data center cold storage tanks, heat exchanger units and water supply units. Across this equipment scope, a useful handover requirement is a consolidated fluid-compatibility schedule covering each supplied component—not disconnected instructions for individual units.

Keep building water supply separate from coolant qualification

Building services may use equipment such as a Non-Negative Pressure Variable Frequency Water Supply Unit. Intended for municipal-network pressure boosting in residential communities, office buildings and hospitals, this type of unit supports stable, energy-efficient water delivery and water-quality protection. It does not, by itself, qualify municipal water for a technology cooling loop. Any connection used for filling or makeup still requires appropriate treatment, verification and backflow protection under applicable local requirements.

Build monitoring around changes, not just calendar dates

Commissioning establishes the baseline. Flush and clean using procedures compatible with the equipment, verify cleanliness against agreed criteria, and sample after the final coolant has circulated sufficiently to become representative. Record the fluid batch, dilution method, laboratory results and initial hydraulic conditions.

Set routine sampling intervals with the equipment supplier and treatment specialist. Increase scrutiny after hose replacement, rack expansion, significant top-up, overheating or unexpected pressure changes. Use consistent sampling points and methods so that trends remain meaningful.

An unexplained conductivity increase should trigger investigation, not automatic dosing. Possible causes include contamination, concentration changes or fluid degradation. Likewise, replacing a blocked filter without examining the captured material may remove the symptom while leaving corrosion or construction debris unresolved.

Never mix formulations merely because both are labelled “propylene glycol” or “data center coolant.” Additive packages can differ. Obtain written compatibility confirmation or follow an approved drain, cleaning and refill procedure.

Make the next maintenance decision traceable

A useful coolant record links each result to an operating condition and an action. Record what changed, who approved the response and how recovery was verified. If the loop has no current fluid specification, begin there before changing chemicals or adjusting alarm thresholds. Reliable operation depends less on a single “good” sample than on keeping a known, compatible fluid within documented limits throughout its service life.