Single Phase & 2 Phase D2C Liquid Cooling ?
What Is Two-Phase Liquid Cooling and How Does It Work?
Liquid cooling is increasingly used to manage the heat generated by high-power CPUs, GPUs, and AI accelerators. As chip power and heat density continue to rise, the way a cooling fluid absorbs and transfers heat becomes more important.
Single-phase and two-phase liquid cooling use different heat-transfer processes. Understanding this distinction helps data center operators evaluate cooling systems for current equipment and future increases in heat load.
This article explains how two-phase liquid cooling works, how it differs from single-phase cooling, and where it can be applied in data centers.
What Is Two-Phase Liquid Cooling?
Two-phase liquid cooling is a heat-transfer method in which the coolant changes between liquid and vapor as it moves through the cooling cycle. The liquid absorbs heat and vaporizes near the heat source. It later releases the heat, condenses, and returns to liquid form.
The term two-phase describes the two physical states of the coolant. It does not describe a specific cooling-system layout. Two-phase cooling can be used in direct-to-chip systems, where coolant flows through cold plates, or in immersion systems, where dielectric fluid surrounds the electronic components.
How Does Two-Phase Liquid Cooling Work?
A two-phase cooling system moves the coolant through a continuous evaporation-and-condensation cycle:
Liquid coolant → heat absorption and vaporization → vapor flow → condensation → liquid coolant
The cycle begins when liquid coolant reaches the heat source. Depending on the system architecture, this may occur inside a cold plate or around components submerged in dielectric fluid.
As the coolant absorbs heat, it reaches its boiling point and begins to vaporize. The boiling point depends on the coolant and the pressure inside the system. During vaporization, the coolant absorbs latent heat without requiring a similarly large rise in temperature.
The vapor then moves away from the heat source to a condenser. It releases the collected heat and changes back into liquid. The heat is transferred to the next part of the cooling system and eventually released outside the data center.
Finally, the liquid coolant returns to the heat source, and the cycle begins again.
Single-Phase vs. Two-Phase Liquid Cooling
Both systems use a fluid to collect heat and carry it away from server components. The main difference is what happens to the coolant as it absorbs heat.
Single-Phase Liquid Cooling
In single-phase cooling, the coolant remains liquid throughout the cycle. It absorbs sensible heat and becomes warmer as it passes the heat source. To carry away a higher heat load, the system may need to increase the coolant flow rate or allow a greater temperature rise.
Two-Phase Liquid Cooling
In two-phase cooling, part of the coolant changes from liquid to vapor as it passes the heat source. It absorbs latent heat during this phase change, allowing it to take in more heat with a relatively small temperature rise. To carry away a higher heat load, more coolant can vaporize, reducing the need to rely on a higher flow rate or greater temperature rise.
Single-phase cooling | Two-phase cooling | |
Coolant state | Remains liquid | Changes between liquid and vapor |
Main heat-absorption process | Coolant temperature rises | Part of the coolant vaporizes |
Response to a higher heat load | A higher flow rate or greater coolant temperature rise may be needed | More of the liquid coolant vaporizes to absorb the additional heat |
Where Can Two-Phase Liquid Cooling Be Used?
Two-phase describes how a coolant absorbs heat, not where the coolant is placed. In data centers, this heat-transfer process can be used in both direct-to-chip and immersion cooling systems.
Two-Phase Direct-to-Chip Cooling
Two-phase direct-to-chip cooling uses cold plates attached to CPUs, GPUs, or other high-heat components. Liquid coolant flows through channels inside each cold plate and vaporizes as it absorbs heat from the chip.
The vapor then leaves the cold plate and flows to a condenser. After releasing the collected heat, it returns to liquid and circulates back to the cold plates.
Two-Phase Immersion Cooling
Two-phase immersion cooling places server components in a tank filled with a dielectric fluid. The fluid directly surrounds the hardware and begins to boil when it absorbs enough heat from the component surfaces.
The resulting vapor rises to a condenser inside or above the tank. After releasing its heat, it condenses and returns to the liquid bath.
Two-phase direct-to-chip cooling | Two-phase immersion cooling | |
Coolant location | Inside cold plates and connecting lines | Inside an immersion tank |
Contact with electronics | Coolant remains inside the cold plates | Coolant directly surrounds the hardware |
Where vaporization occurs | Inside the cold plates | Around the submerged components |
Where condensation occurs | In a separate condenser or condensation module | At a condenser inside or above the tank |
Factors to Consider When Choosing a Two-Phase Cooling System
Two-phase cooling systems differ in their cooling fluids, operating conditions, system architecture, and facility requirements. When comparing solutions, consider the following factors.
Coolant Properties
Different coolants have different boiling points and absorb different amounts of heat during vaporization. The coolant must also be compatible with the pipes, seals, cold plates, tanks, and other materials used in the system.
Operating Pressure
A coolant’s boiling point changes with pressure. Lower pressure generally lowers the boiling point, while higher pressure raises it. The system must maintain the intended pressure so that vaporization occurs within the required temperature range.
Vapor Flow and Condensation
After vaporization, the vapor must be able to move from the heat source to the condenser. The condenser must have enough capacity to release the collected heat and return the vapor to liquid form.
System Integration
The cooling system must fit the server hardware, rack layout, available space, and facility heat-rejection equipment. Its cold plates for tanks, piping, pumps, controls, and monitoring equipment must also be accessible for installation and maintenance.
How Does CreAlien W2PC Use DI-Water Two-Phase Cooling?
CreAlien W2PC applies two-phase cooling through a sealed, negative-pressure DI-water loop. The system uses cold plates to collect heat directly from high-power CPUs and GPUs.
Under normal atmospheric pressure, water boils at approximately 100ºC. W2PC lowers the pressure inside the loop so that the DI water begins to vaporize at approximately 45ºC. This allows phase change to occur within the system’s intended chip-cooling temperature range.
As the DI water absorbs heat inside the cold plates, part of it changes into vapor. The vapor moves to the condensation side, releases the collected heat, and condenses back into liquid. The liquid water then returns to the cold plates to continue the cycle.
By using latent heat during vaporization, W2PC can respond to increasing chip heat loads without relying only on a higher coolant flow rate or a greater temperature rise. Its DI-water cooling medium is also PFAS-free, providing an alternative to two-phase systems that use fluorinated fluids.
Contact CreAlien to explore DI-water two-phase cooling for your AI server deployment.