Based on the development and application of energy storage systems, stable and controllable energy dispatch among generation, transmission, distribution, and consumption can be achieved, enabling large-scale utilization of renewable energy, improving the efficiency of conventional power systems and regional energy systems, driving the development of electric vehicles and other end-use electrical products, and reshaping the energy structure of the new era. Lithium-ion batteries with metallic lithium cathodes occupy a major share of new energy storage systems due to their high energy density and long cycle life. With the deployment and operation of energy storage lithium power stations, the safety of lithium-ion energy storage batteries has attracted increasing attention. During the use of lithium-ion batteries, the shuttling of lithium ions generates a significant amount of heat. Even under normal operating conditions, the heat generated by the battery cannot be completely eliminated, especially in high-temperature weather and large battery packs. Numerous research results indicate that the optimal operating temperature range for lithium-ion batteries is 25–40 °C, and the temperature difference within the battery module should be less than 5 °C. Furthermore, thermal propagation is a chain reaction, often accompanied by continuous heat and gas generation. Without temperature regulation, it may eventually lead to thermal runaway of the battery, where heat generation becomes uncontrollable. For large-scale energy storage battery systems, once thermal runaway initiates, the transfer of heat and gas to neighboring battery modules can result in major cascading accidents. Numerical simulation studies have shown that a slight increase in the heat release of a single cell can cause the entire battery pack to enter a thermal runaway state. Battery packs are extremely sensitive to heat accumulation, making the cooling design of energy storage batteries particularly important.
Researchers have extensively studied cooling methods for energy storage batteries, mainly including liquid cooling, air cooling, and phase-change cooling. Liquid cooling has become a research hotspot due to its high thermal conductivity and high specific heat capacity. Liquid cooling can be divided into indirect liquid cooling and direct liquid cooling (also known as immersion cooling) based on whether it directly contacts the battery surface. Numerical modeling and experimental verification on a single 21700 cylindrical battery showed that immersion cooling with dielectric fluids significantly outperformed indirect cooling with cold plates at a 3C discharge rate. Immersion cooling utilizes surface-contact heat transfer, providing a uniform and high-heat-capacity thermal pathway, thereby improving heat transfer efficiency and reducing vertical thermal gradients. Some dielectric fluids have high flash points and flame retardancy, which can effectively prevent battery thermal runaway and its propagation, making them particularly important in the cooling design of energy storage batteries.
Current immersion cooling research mainly focuses on power batteries for electric vehicles. However, the cooling design for energy storage battery packs differs significantly from that for power batteries. First, the system scale is different. Energy storage power stations typically have a larger volume limit for the cooling system. To improve cooling effectiveness, a combination of multiple cooling methods can be considered. When cost permits, the use of dielectric fluids, especially for direct liquid cooling, can be liberalized. Second, the management object is different. In an energy storage power station, thermal runaway of a single cell may cause serious accidents, leading to huge economic losses and casualties. Therefore, preventing thermal runaway from propagating to the entire battery pack is a critical factor in the cooling system design of energy storage power stations. Third, the requirements for normal cooling differ. Power batteries typically require higher charge/discharge rates and thus release more heat during normal use. In contrast, large-capacity energy storage batteries usually operate at standard charge/discharge powers with lower rates, resulting in milder thermal responses and relatively lower performance requirements for the cooling system. Additionally, due to the low charge/discharge rates, the charge/discharge time of energy storage batteries is longer, and thus the cooling time increases accordingly. The immersion cooling system must work over a longer duration to ensure temperature stability and battery safety. Therefore, there are significant differences in cooling strategies between energy storage batteries and power batteries.
This review summarizes the research progress of immersion cooling technology for energy storage battery applications, including single-phase immersion cooling and phase-change immersion cooling. For dielectric fluid selection, four key parameters are proposed: specific heat capacity, heat transfer coefficient, insulation performance, and compatibility with battery materials. Four types of dielectric fluids that can be practically applied in immersion cooling technology are summarized: fluorinated electronic liquids, hydrocarbon-based fluids, silicone oils, and nanofluids. Representative product numbers and related characteristic parameters are introduced and provided. In addition, key aspects for the practical implementation of immersion cooling technology, including system encapsulation and insulating coatings, are discussed and summarized. Finally, the research and progress of immersion cooling technology in suppressing thermal runaway of lithium-ion batteries are explored.
1. Differences in Thermal Management Between Energy Storage Batteries and Power Batteries
The thermal management of energy storage battery systems and power battery systems for electric vehicles has several distinct differences:
| Parameter | Energy Storage Battery | Power Battery (Electric Vehicle) |
|---|---|---|
| System Scale | Large (MW·h level), space for cooling system is relatively ample | Smaller (kW·h to tens of kW·h), limited space, weight-sensitive |
| Charge/Discharge Rate | Usually 0.25C–1C, heat generation rate lower | Often 1C–3C or higher, high heat generation rate |
| Operating Duration | Long (hours), cooling system needs to operate continuously | Short (minutes to an hour), intermittent cooling |
| Thermal Runaway Risk | Large cascading hazard due to many cells; prevention of propagation is critical | Also serious, but vehicle-level safety measures (e.g., firewalls) are used |
| Coolant Cost Sensitivity | Higher cost tolerance, can use advanced dielectric fluids | Cost-sensitive, often use water-glycol or low-cost oils |
These differences must be considered when designing immersion cooling systems for energy storage applications.
2. Immersion Cooling Technology for Energy Storage Batteries
Immersion cooling refers to direct contact between a dielectric fluid and the battery surface, utilizing the sensible heat (single-phase cooling) or latent heat (phase-change cooling) of the dielectric fluid to absorb heat from the battery module. Direct contact eliminates the need for flow channel design in indirect cooling, reduces system complexity, decreases contact thermal resistance, and improves heat exchange efficiency. Based on whether the fluid undergoes a phase change during cooling, immersion cooling is classified into single-phase immersion cooling and phase-change cooling.
2.1 Single-Phase Immersion Cooling
Single-phase immersion cooling uses dielectric fluids with high boiling points. The fluid remains in a stable single-phase state during cooling. Since there is no need to control boiling and recover gaseous dielectric fluid, single-phase immersion cooling systems are simpler. Research in battery immersion cooling has focused on single cells and battery modules, using both experimental studies and numerical simulations. To evaluate performance, natural air cooling and forced air cooling are often used as control groups. Due to differences in physical and chemical properties among various single-phase dielectric fluids—especially viscosity and cost—oil-based dielectric fluids are usually cheaper but have higher viscosity, leading to greater energy consumption during operation. Comparing the cooling effects of different dielectric fluids is also a research direction.
2.1.1 Experimental Studies
There are relatively few purely experimental studies on single-phase immersion cooling; most involve single cells or small battery packs. Controlled experiments comparing immersion cooling with forced air cooling demonstrate that immersion cooling yields better cooling performance and more uniform temperature distribution. For example, using a commercial single-phase immersion fluid (type AC-100) and a 68 Ah battery at 2C charge/discharge, the average temperature remained at 22.5 °C with minimal fluctuation, while forced air cooling resulted in an average temperature of 28.7 °C. In another study using No. 10 transformer oil as the immersion fluid for a pack of five 10 Ah lithium-ion pouch cells (1 mm spacing), the effects of immersion depth and flow rate were investigated. Full immersion provided the best performance, and increasing flow rate improved cooling.
2.1.2 Numerical Simulation Studies
Numerical simulations are more common in immersion cooling research, often validated by experiments. Simulations allow for diverse system designs and parametric studies. For instance, a design using non-metallic sheets to cover battery poles and expose only the terminals in an open container was modeled using AmpCool AC-100. Another study designed a cooling system with auxiliary terminal cooling for pouch batteries, combining air cooling for the terminals and multiple inlets/outlets to enhance heat transfer. Parametric studies on battery module dimensions and cell spacing for electric vehicle cells concluded that wider modules with larger cell spacing yield more uniform temperature. For a module of 32 cylindrical cells (3200 mAh), changing channel height, inlet velocity, and cross-flow demonstrated the immersion cooling effect of HFE6120, validated by experiments with forced air cooling.
Numerical simulation also enables safe study of large battery packs. Using ANSYS, a simulation of 840 18650 cells immersed in dimethyl silicone oil showed that under 1C discharge, the battery pack temperature remained between 17.5 °C and 32.8 °C, with maximum transient temperature difference not exceeding 8.8 °C. Additionally, thermal runaway propagation prevention tests showed that the heat dissipation rate of the immersion system was twice that of an indirect cooling system, reducing the impact of thermal runaway on surrounding cells by 8.9 °C. In another study, a module of 196 cylindrical NCA cells was modeled comparing immersion cooling (Novec 7500) and cold plate cooling (water/ethylene glycol 1:1 mixture). Simulation data showed that at a specific flow rate, immersion cooling’s thermal performance was 2.5 to 3 times that of cold plate cooling.
2.1.3 Comparison of Dielectric Fluids
Thermal conductivity is not the only critical factor affecting immersion cooling performance. Different dielectric fluids have various physical and chemical properties that interact to influence final cooling effectiveness. Specifically, high-viscosity dielectric fluids impair temperature uniformity and reduce pump efficiency, closely relating to system energy consumption. Modeling of a direct cooling system for four 18650 cells (terminals not immersed) comparing deionized water, mineral oil, and AmpCool AC-100 showed that at a mass flow rate of 0.05 kg/s, deionized water limited temperature rise to 2.2 °C; mineral oil and AmpCool AC-100 had similar cooling performance, but pumping power for mineral oil was much higher than for AmpCool AC-100. Experimental comparisons of natural cooling, forced air cooling, and immersion cooling (using mineral oil and thermosensitive oil) indicated that immersion cooling is significantly superior. Low-cost dielectric fluids offer clear advantages over air cooling, while less viscous oils are more advantageous. A study of five fluorinated dielectric fluids covered both single-phase and two-phase cooling modes. A numerical model of a battery pack consisting of 20 cell units (each unit comprising 10 20 Ah cells in parallel) compared water, silicone oil, and air cooling. Although silicone oil required higher pumping power, its low cost and good cooling performance made it a viable choice. The thermal model accounted for flow distribution and was both fast and effective. An experimental study using No. 10 transformer oil for a single 18650 cell found a trade-off between system energy consumption and cooling effect, identifying an optimal volumetric flow rate of 15 mL/min (Re = 0.59).
In summary, single-phase immersion cooling systems are simple and efficient, receiving significant attention in battery cooling. Numerical simulation is a convenient and effective tool for optimizing designs, validated by experiments. Current research focuses on heat transfer but sometimes neglects battery performance and lifespan changes. Future development of multiphysics coupling simulations considering electrochemical reactions, electrolyte flow, and heat conduction is needed.
2.2 Phase-Change Immersion Cooling
In phase-change immersion cooling, also known as passive two-phase cooling, heat transfer strongly depends on nonlinear temperature characteristics because the battery’s heat release cannot be precisely controlled. The boiling point of the liquid can be actively adjusted by regulating system pressure to maintain nucleate boiling at high heat flux. When using phase-change dielectric fluids, the battery is usually fully submerged. As the battery surface temperature increases during charge/discharge and reaches the saturation temperature, the surrounding liquid boils and turns into vapor, which is then condensed externally and returned. Phase-change cooling relies on the significant latent heat during boiling to cool battery surfaces. This heat transfer mechanism and system design are more complex, but due to the effectiveness of evaporative heat absorption, phase-change immersion has broad application prospects in high-heat-flux devices such as nuclear reactors and high-performance computers. For battery systems, phase-change dielectric fluids can achieve the same cooling effect as single-phase fluids in a smaller volume, aiding lightweight design. Research in battery phase-change immersion is limited, mainly experimental.
For example, a boiling cooling experiment using Novec 7000 on a 20 Ah pouch cell showed that continuous flow removed large amounts of heat but caused significant internal temperature differences, especially at high flow rates. An intermittent flow boiling cooling system was proposed and validated. Another study investigated the effect of boiling heat transfer on temperature uniformity, finding that pressure in the boiling chamber is closely related to battery temperature, and by adjusting pressure, nucleate boiling can be actively maintained. Using Novec 7000 (boiling point 34 °C), the battery temperature during 5C discharge can be maintained around 34.5 °C. Numerical simulation of phase-change turbulent heat transfer with HEF-7000 was validated by charge/discharge cooling experiments. The conclusion was that two-phase nucleate boiling offers significant advantages in improving temperature uniformity. A boiling cooling system using SF33 coolant and Sony VTC6 18650 LIB cells recorded bubble growth. At 4C discharge, the temperature rise was only 4.97 °C. It was also found that overly low immersion temperature may cause voltage damage to lithium-ion batteries during discharge.
In summary, the application of phase-change immersion cooling in batteries is rare, but its latent heat absorption offers huge potential. Some researchers believe that due to boiling-induced state changes, phase-change fluids are more suitable for microchannel heat transfer. Therefore, special attention must be paid to dielectric fluid recovery and avoidance of mixing battery exhaust gases with dielectric vapors in phase-change systems. Future research can explore methods to maintain nucleate boiling through structural design and auxiliary devices.
3. Dielectric Fluids
The properties of dielectric fluids directly affect immersion cooling performance. Key parameters for selecting dielectric fluids are specific heat capacity, heat transfer coefficient, insulation performance, and material compatibility. First, specific heat capacity and heat transfer coefficient determine cooling efficiency. Second, for direct immersion cooling, the fluid must have excellent electrical insulation. If batteries are insulated with coatings or have special structures for electrical isolation, the requirement for conductivity can be relaxed. Third, compatibility with battery and system materials directly influences operating life and safety. Additionally, dynamic viscosity, flash point, ozone depletion potential (ODP), global warming potential (GWP), and cost are important evaluation parameters. Currently, common dielectric fluids are classified into four categories based on chemical and physical properties: fluorinated electronic liquids, hydrocarbon-based fluids, silicone oils, and nanofluids.
3.1 Fluorinated Electronic Liquids
Fluorinated liquids are organic compounds in which some or all hydrogen atoms in hydrocarbons are replaced by fluorine atoms. Due to the strong C–F bond, they typically have high dielectric constants. They include various types: chlorofluorocarbons (CFCs), hydrochlorofluorocarbons (HCFCs), perfluorocarbons (PFCs), hydrofluorocarbons (HFCs), hydrofluoroethers (HFEs), and hydrofluoroolefins (HFOs). Their usage restrictions are summarized below.
| Type | Ozone Depletion | Greenhouse Gas | Regulatory Document |
|---|---|---|---|
| CFCs | Yes | Yes | Montreal Protocol |
| HCFCs | Yes | Yes | China Controlled ODS List |
| PFCs | Yes | Yes | Kyoto Protocol |
| HFCs | No | Yes | Kigali Amendment; HFC quota plan (2024) |
| HFEs | No | No | — |
| HFOs | No | No | — |
Only HFEs and HFOs are considered green with no ozone depletion and low GWP. Their electrical insulation and environmental friendliness have attracted much research in electronic cooling. Notably, a major global electronic coolant manufacturer announced cessation of PFAS production by the end of 2025, affecting series such as Fluorinert and Novec. Another major chemical company developed a series of dielectric fluids based on HFOs for thermal management, electronic cleaning, and refrigeration. Example products and their parameters are shown below.
| Product | Boiling Point (°C) | Density at 25 °C (g/cm³) | Dynamic Viscosity at 25 °C (mPa·s) | Thermal Conductivity at 25 °C (W/(m·K)) | Specific Heat at 25 °C (kJ/(kg·K)) | Vapor Pressure at 25 °C (MPa) | Latent Heat (kJ/kg) | Dielectric Constant | Critical Temp (°C) | Critical Pressure (MPa) | ODP | GWP |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| OpteonTM MZ | 33.4 | 1.360 | 0.38 | 0.077 | 1.20 | 0.07 | 166 | 32 | 171.3 | 2.9 | 0 | 2 |
| OpteonTM 2P50 | 48.9 | 1.456 | 0.62 | 0.073 | 1.09 | 0.04 | — | — | 176.0 | 2.7 | 0 | 10 |
3.2 Hydrocarbon-Based Fluids
Hydrocarbon-based fluids consist solely of carbon and hydrogen. Common ones include mineral oils and synthetic hydrocarbon oils. Transformer oil, derived from petroleum distillation and refining, has good insulation and heat dissipation properties at low cost, widely used in transformers. Common issues include oxidation (forming acidic products), high-temperature degradation, insulation failure due to corrosive sulfur, and environmental pollution from aged oils. Typical parameters of transformer oil (e.g., Morunke brand) are:
| Grade | Density at 20 °C (g/cm³) | Kinematic Viscosity at 40 °C (mm²/s) | Flash Point (°C) | Pour Point (°C) | Breakdown Voltage (2.5 mm gap, kV) |
|---|---|---|---|---|---|
| No.10 | 0.895 | <13 | >140 | <-7 | >35 |
| No.25 | 0.895 | <13 | >140 | <-22 | >35 |
| No.45 | 0.895 | <13 | >135 | <-45 | >35 |
Synthetic hydrocarbon oils are more environmentally friendly and safer. For example, AmpCool AC series fluids are sulfur-free, ozone-friendly, and have high biodegradability. Their parameters are listed below.
| Product | Pour Point (°C) | Flash Point (°C) | Density at 15.6 °C (g/cm³) | Kinematic Viscosity at 40 °C (mm²/s) | Thermal Conductivity at 40 °C (W/(m·K)) | Specific Heat at 40 °C (kJ/(kg·K)) | Dielectric Constant | GWP | Biodegradability (%) |
|---|---|---|---|---|---|---|---|---|---|
| AC-110 | -57 | 193 | 0.82 | 8.11 | 0.1359 | 2.2121 | 2.080 | 0 | >93 |
| AC-120 | -62 | 228 | 0.82 | 16.00 | 0.1459 | 2.2060 | 2.100 | 0 | >70 |
| AC-130 | -49 | 268 | 0.82 | 36.10 | 0.1508 | 2.2030 | 2.100 | 0 | >61 |
| AC-140 | -52 | 280 | 0.84 | 67000 | 0.1584 | 2.1912 | 2.220 | 0 | >50 |
| AC-210 | -51 | 194 | 0.82 | 8.61 | 0.1403 | 2.2094 | 2.080 | 0 | >92 |
| AC-220 | -64 | 235 | 0.82 | 17.70 | 0.1459 | 2.2060 | 2.091 | 0 | >70 |
| AC-230 | -44 | 270 | 0.83 | 39.00 | 0.1151 | 2.2030 | 2.100 | 0 | >70 |
| AC-240 | -47 | 282 | 0.84 | 69.87 | 0.1585 | 2.2200 | 2.190 | 0 | >50 |
3.3 Silicone Oils
Silicone oils, specifically organopolysiloxanes with organic side chains, are commonly known as silicone oils. Due to their low toxicity, good electrical insulation, chemical inertness, and hydrophobicity, they are widely used in heat transfer fluids, lubricants, and additives. The most common silicone oil is polydimethylsiloxane (PDMS). Its kinematic viscosity ranges from 0.65 to 20,000,000 mm²/s depending on molecular weight and chain length. Typical parameters for PDMS with viscosities between 0.65 and 1000 mm²/s are shown below.
| Product Code | Kinematic Viscosity (mm²/s) | Viscosity-Temperature Coefficient | Pour Point (°C) | Density (g/cm³) | Thermal Expansion Coefficient (×10⁻⁴/°C) | Thermal Conductivity (W/(m·K)) | Dielectric Constant | Flash Point (°C) |
|---|---|---|---|---|---|---|---|---|
| DMS-T00 | 0.65 | 0.32 | -68 | 0.761 | 13.4 | 0.1005 | 2.20 | -1 |
| DMS-T01 | 1 | 0.37 | -85 | 0.818 | 13.4 | 0.1005 | 2.30 | 39 |
| DMS-T05 | 1.5 | 0.46 | -75 | 0.853 | 13.4 | 0.1047 | 2.39 | 63 |
| DMS-T02 | 2 | 0.48 | -80 | 0.873 | 11.7 | 0.1089 | 2.45 | 79 |
| DMS-T03 | 3 | 0.51 | -70 | 0.898 | 11.4 | 0.1130 | 2.50 | 100 |
| DMS-T05 | 5 | 0.54 | -65 | 0.918 | 11.2 | 0.1172 | 2.60 | 135 |
| DMS-T07 | 7 | 0.55 | -65 | 0.930 | 11.0 | 0.1256 | 2.656 | 150 |
| DMS-T11 | 10 | 0.56 | -65 | 0.935 | 10.8 | 0.1340 | 2.68 | 163 |
| DMS-T12 | 20 | 0.59 | -65 | 0.950 | 10.7 | 0.1424 | 2.72 | 232 |
| DMS-T15 | 50 | 0.59 | -65 | 0.960 | 10.6 | 0.1507 | 2.75 | 285 |
| DMS-T21 | 100 | 0.60 | -65 | 0.966 | 9.3 | 0.1549 | 2.75 | 315 |
| DMS-T22 | 200 | 0.60 | -60 | 0.968 | 9.3 | 0.1549 | 2.75 | 315 |
| DMS-T23 | 350 | 0.60 | -60 | 0.970 | 9.3 | 0.1591 | 2.75 | 315 |
| DMS-T25 | 500 | 0.60 | -55 | 0.971 | 9.3 | 0.1591 | 2.75 | 315 |
| DMS-T31 | 1000 | 0.61 | -50 | 0.971 | 9.3 | 0.1591 | 2.75 | 315 |
Compatibility of dimethyl silicone oil with battery systems has been experimentally verified. For example, a study using dimethyl silicone oil (density 0.873 g/cm³, specific heat 1.8 kJ/(kg·K), thermal conductivity 0.11 W/(m·K)) for direct immersion cooling of single-crystal silicon solar cell arrays showed negligible impact on electronic device performance after 270 days. Chemical modification of silicone oils can enhance specific properties. In immersion cooling, the goal is to maintain insulation while improving thermal conductivity. For example, substituting methyl groups with phenyl groups improves oxidation resistance and thermal stability.
3.4 Nanofluids
To further improve the heat transfer performance of dielectric fluids, especially for oil-based fluids like dimethyl silicone oil and mineral oil that have excellent insulation but relatively low heat transfer ability, high-thermal-conductivity nanomaterials can be added. Nanofluids, first proposed and theoretically estimated for copper nanoparticles in 1995, are fluids containing 1–100 nm nanoparticles dispersed via physical means. They can enhance thermal conductivity and stability. For example, adding SiO₂ nanoparticles to transformer oil improved liquid insulation. Common thermally conductive nanoparticles include Al, Al₂O₃, ZnO, Ag, and graphene. Ag and graphene have excellent thermal conduction but high cost. The mechanism for enhanced thermal conductivity is not fully understood, but a widely accepted explanation is that a nanoscale layer forms on nanoparticle surfaces with higher thermal conductivity. Nanofluids are classified into oil-based and water-based types.
3.4.1 Oil-Based Nanofluids
Oil-based nanofluids primarily use mineral oil or dimethyl silicone oil as base fluids. For example, using a two-step process of mechanical stirring and ultrasonic homogenization, various graphene types were dispersed in dimethyl silicone oil. At 20% volume fraction of graphene oxide (flake diameter 0.5–5 μm, thickness 0.8–1.2 nm), the thermal conductivity reached 0.542 W/(m·K), three times that of base oil. However, volume resistivity measurements showed no clear pattern. Graphene, with extremely high electrical conductivity, can partially degrade insulation. Coating graphene with silicon dioxide can help maintain insulation in graphene nanofluids.
3.4.2 Water-Based Nanofluids
Water-based nanofluids mainly use deionized water or deionized water/ethylene glycol mixtures as base fluids. For example, mixing different ratios of deionized water/ethylene glycol with 0.1%, 0.3%, and 0.5% volume concentrations of Al₂O₃ nanoparticles, after one month of static settling, showed no clear trend in thermal and electrical conductivities. The key issue for water-based nanofluids in direct immersion cooling is electrical insulation. Water-based fluids are inherently non-dielectric, and adding nanoparticles can also affect insulation. Furthermore, even with physical dispersion, complete uniformity is difficult to achieve, leading to inconsistent properties and questionable measurement accuracy. Long-term suspension stability is also a concern.
In terms of practical application, fluorinated electronic liquids and synthetic hydrocarbon fluids are expensive; nanofluid costs vary depending on the nanoparticles; silicone oils and transformer oils are about one-tenth the cost of the former. Considering the construction cost of immersion cooling systems, modification of silicone oils is a direction worth exploring.
4. System Encapsulation and Insulating Coatings
4.1 System Encapsulation
Special encapsulation designs can reduce the amount of coolant, improve economic feasibility, and reduce pump energy consumption. For example, a direct cooling structure with an embedded U-shaped flow channel was designed for pouch cells, which is modular and scalable. Experimental results showed that thermal resistance between fluid and battery was an order of magnitude lower than indirect liquid cooling. Another cooling structure for prismatic batteries uses manifolds for jet impingement and baffles for fluid recovery, achieving high heat transfer efficiency with less dielectric fluid.
4.2 Insulating Coatings
By applying insulating coatings to electronic devices, non-dielectric fluids with excellent thermal conductivity can be used in direct immersion cooling. For example, a 1–25 μm Parylene C coating (dielectric strength 78 V/m, thermal conductivity 0.1 W/(m·K)) applied via chemical vapor deposition provided sufficient electrical protection for circuit boards. Another study used a composite of organic silicone sealant and boron nitride to coat 18650 cells. After immersion in water for charge/discharge tests, the coating showed good waterproofing and insulation, enabling the use of water as a non-dielectric fluid that effectively removed most of the heat.
5. Thermal Runaway Suppression by Immersion Cooling
The thermal runaway mechanism of lithium-ion batteries and their reaction characteristics under various abuse conditions have been widely studied. Thermal runaway is triggered by specific events, with internal short circuit considered a common trigger. When the cathode and anode come into contact, additional heat is released, and side reaction products (e.g., oxygen from cathode above 200 °C, hydrogen from LiH decomposition in the anode solid electrolyte interface) can play a critical role in triggering runaway. Therefore, the only chemical approach to avoid heat accumulation is to find ultimately stable anodes and electrolytes. It has been proven that if reductive attack pathways are cut off, the development of thermal runaway can be easily controlled. Flame-retardant fluorinated electrolytes have been introduced; however, they can still be reduced by lithiated anodes, leading to thermal runaway. Current efforts to eliminate chemical triggers, such as stabilizing cathode crystals or capturing oxygen, only mitigate thermal runaway to a small extent and cannot prevent it. Considering that thermal runaway usually occurs within 0.1 s, safety countermeasures are hard to implement during runaway.
Nearly all thermal runaway triggers have excitation temperatures around 180–250 °C, far beyond normal operating conditions of lithium-ion batteries. However, except for severe collision accidents, there is typically a mild heat accumulation stage lasting minutes to days before runaway, during which mild side reactions occur. This provides an opportunity for thermal runaway control: cutting off the energy release pathway before triggering. When immersion cooling is used, theoretically, the dielectric fluid can fully submerge the battery and rapidly remove heat at high flow rates during the mild heat accumulation stage, preventing heat accumulation. If the battery casing ruptures, immersion cooling can also prevent further short circuits between positive and negative electrodes. Moreover, full immersion can to some extent prevent combustible gases from contacting oxygen in the air; some dielectric fluids can dissolve these combustible gases, inhibiting the formation of explosive mixtures, thereby avoiding the accompanying effect of explosion. Therefore, immersion cooling is very important in suppressing thermal runaway.
Existing research mainly focuses on avoiding heat accumulation to reduce the probability of thermal runaway. For example, in a study of a three-cell pouch battery pack with immersion cooling, the middle cell was overcharged to trigger thermal runaway. The faulty cell temperature rose to a maximum of 183.9 °C and swelled due to gas generation, but under immersion cooling with Novec 649, heat accumulation was controlled, and no explosion or combustion occurred; adjacent cells did not experience thermal runaway. Another study used heating rods to simulate continuous heat sources for thermal runaway suppression tests, demonstrating that fluorinated dielectric fluids can effectively suppress or avoid battery thermal runaway.
6. Conclusion
This review analyzes the differences between cooling systems for energy storage batteries and power batteries. The trend toward large-capacity energy storage battery cells is clear. On the cost side, using larger cells reduces the number of cells, lowers battery management system complexity, and decreases component usage, thus reducing overall cost. On the heat transfer side, compared with small-capacity cells, large cells have a reduced surface area-to-volume ratio, leading to less heat transfer area. Therefore, immersion cooling research for energy storage batteries should be differentiated from that for power batteries and given due attention. This review discusses progress in immersion cooling technology for energy storage batteries, including single-phase immersion cooling, phase-change immersion cooling, dielectric fluids, system encapsulation, and insulating coatings. These studies provide directions and reference data for the future development of immersion cooling technology for energy storage batteries.
In summary, immersion cooling technology, with its efficient thermal management and intrinsic safety characteristics, has become an important development direction in the field of energy storage battery thermal management. Proper selection of dielectric fluids, optimized system design, and effective thermal runaway suppression strategies are critical to practical application. Ongoing research addressing cost, material compatibility, and system integration will further promote the adoption of immersion cooling in large-scale energy storage battery systems.