As computing hardware produces more heat within smaller spaces, liquid-based cooling can provide a direct method of transferring heat away from electronic components.

Unlike conventional air cooling, which relies primarily on airflow, liquid cooling uses a fluid to absorb and transport heat. Depending on the design, the liquid may circulate through cold plates attached to processors or directly contact electronic components inside specialized immersion systems.

What Are Data Center Liquid Cooling Systems?

Data center liquid cooling systems use a circulating liquid to capture heat generated by computing equipment and transfer that heat to another cooling medium or heat-rejection system.

A typical system can include:

• Cold plates
• Coolant distribution units
• Pumps
• Heat exchangers
• Manifolds
• Supply and return piping
• Coolant reservoirs
• Sensors
• Control systems
• Heat-rejection equipment

The exact configuration depends on rack density, processor thermal load, facility design, coolant type, and cooling architecture.

Why Liquid Cooling Is Used in Data Centers

Modern computing equipment can produce substantial heat, particularly in high-performance computing, artificial intelligence, machine learning, scientific computing, and other computationally intensive applications.

Liquid cooling can provide several technical advantages:

• Direct heat transfer from high-heat components
• Higher heat-transfer capacity than air in many applications
• Support for high-density server configurations
• Reduced dependence on large volumes of room air
• More controlled processor temperatures
• Integration with facility water or heat-rejection systems

Liquid cooling does not necessarily replace all air cooling. Many data centers use hybrid arrangements in which liquid cooling handles major processor loads while air cooling manages remaining components.

Types of Data Center Liquid Cooling Systems

Direct-to-Chip Liquid Cooling

Direct-to-chip cooling places a cold plate directly against heat-generating components such as CPUs, GPUs, or other accelerators.

Coolant flows through channels inside the cold plate and absorbs heat from the component. The heated fluid then travels to a heat exchanger or coolant distribution unit.

This approach can be integrated into conventional server racks while adding a dedicated liquid circuit.

Single-Phase Immersion Cooling

In single-phase immersion cooling, servers or selected electronic components are immersed in a non-conductive liquid.

The liquid remains in the same physical phase during operation. Pumps or natural convection move heated fluid toward a heat exchanger where the heat is transferred to another cooling circuit.

Two-Phase Immersion Cooling

Two-phase immersion cooling uses a specially selected dielectric fluid that changes from liquid to vapor as it absorbs heat.

The vapor rises toward a condenser, where it releases heat and returns to liquid form. The liquid then returns to the immersion tank.

Rear-Door Heat Exchangers

Rear-door heat exchangers are installed at the back of server racks. They capture heat from the exhaust air and transfer it to a liquid cooling circuit.

This approach can reduce the amount of heat released into the data center room without requiring liquid to circulate directly through individual server components.

Hybrid Liquid Cooling

Hybrid systems combine liquid cooling with conventional air cooling. Liquid circuits can handle high-heat processors, while fans and room-level systems manage memory, storage, networking equipment, power components, and other heat sources.

Main Components of Liquid Cooling Systems

How Data Center Liquid Cooling Works

1. Heat Generation

Processors and other electronic components generate heat during operation. High-performance processors can produce substantial thermal loads.

2. Heat Transfer to the Coolant

In a direct-to-chip system, heat moves from the processor through a thermal interface material into a cold plate.

The coolant flowing through the cold plate absorbs the heat.

3. Heated Coolant Movement

The heated coolant leaves the server through the return circuit. Pumps maintain the required flow through the cooling network.

4. Heat Exchange

The coolant passes through a heat exchanger or coolant distribution unit. Heat is transferred from the technology cooling loop to a facility water loop or another heat-rejection circuit.

5. Heat Rejection

The facility cooling system removes the transferred heat through equipment such as dry coolers, cooling towers, chillers, or other heat-rejection technologies.

6. Coolant Recirculation

The cooled fluid returns to the server racks, completing the cooling cycle.

Coolants Used in Data Center Systems

Different liquid cooling architectures use different fluids.

Water-Based Coolants

Water has high heat capacity and is widely used in indirect cooling circuits. Treated water or water-glycol mixtures may be selected according to system requirements.

Dielectric Fluids

Immersion cooling systems use electrically non-conductive fluids that can come into direct contact with electronic components.

The fluid must be compatible with electronics, seals, plastics, metals, and other materials within the cooling system.

Factors Affecting System Design

Rack Density

Higher rack power density generally increases the amount of heat that must be removed from a specific area.

Processor Thermal Design

CPU, GPU, accelerator, and other processor characteristics influence cold-plate design and coolant requirements.

Coolant Flow Rate

Flow rate affects the amount of heat that can be transported through the liquid circuit. Pumps and piping must be sized according to the required flow.

Coolant Temperature

Supply temperature affects the temperature difference available for heat transfer. The appropriate temperature range depends on equipment specifications and facility conditions.

Facility Infrastructure

Liquid cooling may require dedicated piping, pumps, heat exchangers, water treatment, leak detection, and heat-rejection equipment.

Applications of Data Center Liquid Cooling

Artificial Intelligence and Machine Learning

AI workloads can use large numbers of high-performance GPUs or specialized accelerators. Liquid cooling can help manage the thermal loads associated with dense computing configurations.

High-Performance Computing

Scientific simulations, engineering calculations, weather modeling, and other high-performance workloads can generate substantial heat.

Cloud Computing

Cloud infrastructure can use liquid cooling in selected high-density server environments where conventional cooling approaches are insufficient or less suitable.

Enterprise Data Centers

Enterprise facilities can adopt liquid cooling for specific high-power racks or specialized computing clusters without necessarily converting the entire facility to immersion cooling.

Edge Computing

Some compact computing environments have limited space for conventional cooling infrastructure. Liquid-based thermal management can be considered where equipment density and environmental conditions require it.

Monitoring and Control

Data center liquid cooling systems use sensors and control equipment to monitor operating conditions.

Common measurements include:

• Coolant supply temperature
• Coolant return temperature
• Flow rate
• Pressure
• Pump speed
• Heat exchanger temperature
• Rack temperature
• Leak detection status

Automated controls can adjust pump speed, valves, cooling capacity, and other parameters according to the thermal load.

Data can be integrated with data center infrastructure management systems to provide centralized visibility of cooling performance.

Maintenance Requirements

Regular maintenance is important for maintaining liquid cooling performance.

Cooling loops should be inspected for leaks, corrosion, deposits, and changes in pressure or flow. Filters and strainers should be checked according to system requirements.

Coolant quality may also need periodic testing for contamination, chemical balance, conductivity, or other relevant properties.

Pumps, valves, heat exchangers, manifolds, and cold plates should be inspected according to manufacturer specifications. Leak detection sensors should also be tested periodically.

Safety Considerations

Liquid cooling introduces additional considerations compared with conventional air cooling. A leak can potentially affect electronic equipment, infrastructure, or personnel, depending on the coolant and system design.

Important measures can include:

• Leak detection sensors
• Automatic isolation valves
• Appropriate piping materials
• Pressure monitoring
• Coolant compatibility checks
• Proper electrical isolation
• Spill containment where applicable
• Preventive inspection

For immersion systems, coolant handling procedures should account for the specific properties of the dielectric fluid.

Advantages and Limitations

Liquid cooling can support higher thermal loads, but implementation requires careful integration between IT equipment and facility infrastructure.

FAQs

What are data center liquid cooling systems used for?

They are used to remove heat from servers, CPUs, GPUs, accelerators, and other high-density computing equipment using circulating liquid.

What are the main types of liquid cooling?

Common approaches include direct-to-chip cooling, single-phase immersion cooling, two-phase immersion cooling, rear-door heat exchangers, and hybrid liquid cooling.

What is direct-to-chip cooling?

Direct-to-chip cooling uses cold plates attached to heat-generating processors. Coolant flows through the cold plates and carries heat away from the components.

Does liquid cooling completely replace air cooling?

Not necessarily. Hybrid systems often use liquid cooling for high-heat processors while air cooling handles other components and residual room heat.

How are liquid cooling systems monitored?

Sensors can measure coolant temperature, pressure, flow, pump operation, and leaks. Automated controls can adjust cooling equipment according to changing thermal loads.

Conclusion

Data center liquid cooling systems provide a method for managing the increasing thermal loads associated with high-density computing. Direct-to-chip cooling, immersion cooling, rear-door heat exchangers, and hybrid systems use different approaches to transfer heat from IT equipment into a liquid cooling circuit.

System design depends on rack density, processor characteristics, coolant properties, flow requirements, facility infrastructure, and heat-rejection capacity. Proper monitoring, leak detection, coolant management, and maintenance are important for maintaining reliable operation.