A heat exchanger transfers heat between two fluids through a solid wall that keeps their flow paths separate. The hotter fluid gives up thermal energy to one surface of the wall; that energy travels through the material and enters the cooler fluid on the other side. With an intact barrier and properly sealed connections, neither fluid needs to cross into the other circuit.

In a liquid-cooled data centre, this arrangement separates coolant circulating through equipment from water circulating through the building’s cooling infrastructure. Heat passes between the circuits while their water chemistry, pressure and cleanliness remain independently controlled.

The heat crosses three thermal resistances

Inside a plate or tube heat exchanger, heat moves through a sequence of processes. First, convection carries energy from the warmer fluid toward the wall. Conduction then transfers energy through the metal. Finally, convection carries energy from the opposite wall surface into the cooler stream.

Fluid immediately next to a solid surface moves more slowly than fluid farther away. This boundary layer creates resistance to heat transfer. Plate corrugations and suitable flow velocities promote mixing within each separate passage, bringing more fluid into contact with the heat transfer surface. Mixing within a circuit improves heat transfer without mixing the two circuits together.

The wall adds its own resistance. A thinner wall conducts heat more readily, but its thickness must also withstand pressure, corrosion exposure and mechanical loads. Stainless steel is common in plate exchangers because it combines useful corrosion resistance with structural strength. Material selection still requires compatibility with both fluids, cleaning chemicals and operating temperatures.

A highly conductive metal alone does not guarantee strong performance. Poor flow, deposits or an unsuitable channel arrangement can create more resistance than the metal wall itself.

How the passages stay separate

A plate heat exchanger contains a stack of thin metal plates. Ports and seals direct one fluid through alternating channels, while the second fluid travels through the channels between them. Each shared plate acts as a heat transfer surface and a physical barrier.

In gasketed designs, gaskets seal the plate edges and route flow around the appropriate ports. Brazed designs use permanent bonded joints instead of replaceable plate gaskets. Welded constructions provide another way to form sealed passages. These arrangements differ in serviceability and fluid compatibility, but all rely on intact walls and joints to maintain separation.

A shell-and-tube exchanger applies the same principle with different geometry. One fluid travels inside tubes, while the other flows around their outer surfaces within a shell. Tube walls and the joints at the tube sheets separate the streams.

Hydraulic separation does not require equal pressure on both sides. Each circuit can have its own pump and operating pressure, provided the exchanger remains within its specified limits, including allowable pressure difference. The wall carries the pressure load while conducting heat.

Temperature difference determines the direction of transfer

Heat flows locally from the warmer fluid toward the cooler fluid. A conventional liquid-to-liquid exchanger has no compressor and does not raise heat to a higher temperature level. Its ability to cool an equipment loop therefore depends on the temperature of the receiving circuit.

If the facility water entering an exchanger is too warm, increasing exchanger size cannot make the equipment-side outlet colder than that incoming water under steady operating conditions. A larger surface can reduce the temperature gap needed for a given duty, but a finite exchanger needs a temperature difference to transfer heat.

Flow direction affects how that difference is maintained. With parallel flow, both streams enter at the same end, producing a large initial temperature difference that narrows along the passages. Counterflow sends the streams in opposite directions and often sustains a more useful temperature difference over the available surface.

In counterflow, the cooler stream’s outlet can be warmer than the hotter stream’s outlet because those outlets are at opposite ends. This does not reverse heat flow: at each location along the separating wall, the fluid giving up heat remains warmer than the fluid receiving it.

What capacity and temperature readings actually show

For a single-phase liquid with approximately constant properties, the heat transfer rate can be estimated from either circuit:

Heat transfer rate = mass flow rate × specific heat capacity × fluid temperature change.

At steady operation, heat lost by the warmer stream approximately equals heat gained by the cooler stream, allowing for heat exchanged with the surroundings and measurement uncertainty. Equal heat transfer rates do not require equal temperature changes. Different flow rates or fluid properties produce different inlet-to-outlet differences.

This matters when comparing water with a water-glycol mixture. Glycol concentration changes specific heat, viscosity and thermal conductivity. At the same volumetric flow rate, the mixture does not necessarily carry the same heat or produce the same pressure drop as water.

Exchanger performance is also commonly expressed as:

Heat transfer rate = overall heat transfer coefficient × surface area × effective temperature difference.

The overall coefficient combines the resistance of both fluid boundary layers, the wall and any deposits. The effective temperature difference reflects the temperatures and flow arrangement. Consequently, a stated cooling capacity is meaningful only alongside its specified inlet temperatures, flow rates and fluid composition.

A large temperature drop across the hot side alone is insufficient evidence of high cooling capacity. Low flow can produce a large temperature change while carrying relatively little heat. Evaluating performance requires the temperature difference and flow rate together.

Separation inside a CDU

Many cooling distribution units use a liquid-to-liquid heat exchanger between a facility circuit and an equipment circuit. Coolant returning from cold plates delivers heat to the exchanger, then returns toward the equipment at a lower temperature. Facility water absorbs that heat and carries it toward heat rejection equipment or a suitable heat recovery system.

This boundary allows the equipment circuit to maintain coolant quality suited to narrow cold-plate passages and wetted materials. It also prevents normal facility-side pressure fluctuations from directly imposing the same hydraulic conditions on the equipment circuit.

The surrounding pipework must preserve this separation during installation. Liquid Cooling Prefabricated Pipes are designed and manufactured specifically for liquid cooling secondary systems. Their connection locations, cleanliness and circuit identification need to match the CDU layout so that installation does not introduce debris or unintended connections between circuits.

Recovering the transferred heat requires a compatible temperature level at the receiving application. An exchanger transfers available heat; where a higher delivery temperature is required, additional equipment such as a heat pump is needed.

When cooling weakens or the barrier fails

Reduced cooling performance does not automatically indicate that the fluids are mixing. Scale, corrosion products or biological deposits add thermal resistance. Blocked passages reduce effective flow, while trapped gas leaves part of the surface poorly wetted. Similar outlet temperature symptoms can therefore arise from different physical causes.

Pressure drop helps distinguish these conditions when compared at similar flow rates and fluid temperatures. A rising pressure drop alongside declining heat transfer can indicate restricted passages. Reduced heat transfer without a comparable pressure-drop change can point toward deposits on the surfaces, although temperature sensors and actual flow also need verification.

Cross-contamination requires a leakage path, such as a cracked plate, damaged tube or failed joint. If an internal opening develops, pressure difference generally drives fluid from the higher-pressure circuit into the lower-pressure circuit. Some gasket failures instead produce an external leak; the result depends on the seal arrangement and failure location.

Where the consequences of cross-contamination require additional protection, a double-wall exchanger with an appropriate leak detection path offers a way to reveal a barrier failure before the two streams communicate directly. Its extra thermal resistance must be included in sizing. Pressure integrity and heat transfer performance are separate requirements, and both need to match the intended operating conditions.