Exploring liquid cooling technologies

The industry currently recognizes two basic types of liquid cooling technologies: cold plate and immersive, both which are available in single-phase and two-phase processes. While cold plate technology is well-established in the market and is the predominant type of liquid cooling used today, precision liquid cooling has emerged as a third, unique, and differentiated technology.

The implications of which solution to choose are far-reaching. The choice of a cooling solution not only influences the immediate efficiency and performance of data centers but also wields a profound impact on the pursuit of environmental responsibility and sustainability goals. As the data center industry grapples with the imperative to curb energy consumption and conserve water resources, the decision between liquid cooling methodologies takes on heightened significance.

To gain a comprehensive grasp of these technologies, let's explore the various liquid cooling methods to unravel the strengths and weaknesses of each and their implications for data center operations.

Cold-plate cooling

Cold-plate cooling is a specialized liquid cooling method that involves the use of cold plates, typically made from metal, to dissipate heat generated by electronic components within IT equipment. The cold plates are strategically designed to come into direct contact with particular equipment components. As the electronic components heat up during operation, the cold plates conduct and absorb the heat, quickly transferring it to the cooling fluid — usually water or a dielectric liquid.

The cooling fluid, carrying the absorbed heat, flows through a closed-loop system to a heat exchanger or chiller, where the heat is dissipated into the ambient environment or directed to another cooling system for further heat removal. This continuous cycle of heat absorption and dissipation maintains the desired operating temperatures of the electronic components.

Cold-plate cooling excels in efficient heat transfer by establishing direct contact between the cold plates and heat sources, which minimizes thermal resistance. However, cold-plate cooling has its limitations. Implementing cold-plate cooling systems can be complex, often necessitating custom designs to accommodate various IT equipment configurations. Cold-plate configurations utilizing rigid metal tubing, such as copper or aluminum, exhibit superior cooling capabilities when compared to other designs. However, this advantage comes at the cost of increased maintenance complexity, because copper or aluminum necessitates meticulous removal for IT equipment access. In the context of high-compute applications like AI, the substantial copper requirements render these setups considerably more expensive than conventional air-cooled systems.

Furthermore, adapting cold-plate cooling for large-scale data centers may pose challenges, especially when dealing with diverse hardware setups. Servicing IT components equipped with cold-plate cooling systems can be more intricate than traditional air cooling, potentially mandating specialized training and tools. These drawbacks and limitations should be taken into account when evaluating the suitability of cold-plate cooling for specific data center applications.

Immersion cooling

Immersion cooling enables direct contact between electronic components and the cooling fluid — IT equipment, including servers, are submerged into a specially designed dielectric liquid or coolant. As the IT equipment operates, it generates heat, which is efficiently transferred to the surrounding coolant. The heated coolant rises to the surface, where it can be directed to a cooling system, such as a heat exchanger, to dissipate the accumulated heat. This continuous cycle maintains the equipment at optimal operating temperatures.

Another critical industry development worth noting: 3M, a major manufacturer of fluids used in two-phase immersion cooling and some cold-plate systems, announced its exit from per- and polyfluoroalkyl substances (PFAS) chemical manufacturing by the end of 2025. This shift is significant, considering 3M accounts for approximately 80% of the market for fluids used in two-phase liquid cooling. As a result, liquid manufacturers will need to explore alternative formulations that do not contain PFAS chemicals to ensure the viability of this technology at scale. However, the challenge lies in finding formulations that are both effective and do not introduce safety or health concerns, adding a layer of complexity to the adoption of immersion cooling in data centers.

Precision liquid cooling

Precision liquid cooling is an advanced method of cooling high-performance computing systems, by using a carefully controlled flow of dielectric liquid coolant. Unlike traditional air cooling, where fans dissipate heat, precision liquid cooling involves direct contact between the liquid coolant and CPUs and GPUs within the hardware. The ability to provide precise, targeted cooling exactly where it's needed most ensures no overcooling of less heat-intensive parts, which can waste energy and resources.

This technology offers several notable advantages that significantly impact data center efficiency and sustainability. One of the most prominent benefits is the operational cost savings it delivers. By employing precision liquid cooling, data centers can achieve remarkable reductions in their overall operational expenses. But, precision liquid cooling doesn't just contribute to cost savings; it also plays a pivotal role in conserving water resources. Through its innovative design and cooling mechanisms, water usage is substantially reduced compared to traditional cooling methods. This reduction in water consumption aligns with the growing focus on sustainability and responsible resource management, making precision liquid cooling an environmentally conscious choice for modern data centers.

Additionally, precision liquid cooling is designed in a familiar chassis-based form factor that seamlessly integrates cooling within the IT equipment, ensuring uniform and efficient cooling across every component within the server blade. Precision liquid cooling also excels in environments demanding high compute densities, making it a superb choice for edge applications. No specialized training is necessary for servicing IT equipment, and in the event of component failures, quick and seamless hot swapping can be performed right on the data center floor or at edge locations.

Precision liquid cooling also employs a significantly lower volume of coolant, roughly 90% less than traditional single-phase immersive cooling technologies, meaning reinforced floors or major retrofitting are not needed.

The future of liquid cooling is now

Developing an understanding of liquid cooling technologies is crucial for making informed decisions about the future of our data centers, especially in light of rising IT equipment demand and evolving sustainability requirements. By addressing climate change challenges, we can all contribute to a better, more sustainable planet.

Global electricity generation has experienced growth in the utilization of renewable energy sources, rising from 27% in 2019 to 29% in 2020, but substantial progress remains to be seen. . By embracing liquid cooling solutions, organizations can enhance operational efficiency, reduce energy consumption, and adhere to emerging sustainability standards.

In comparison with other types of liquid cooling solutions, precision liquid cooling technology addresses several inherent limitations, including sustainability concerns and serviceability challenges.

This article was originally posted on www.missioncriticalmagazine.com.