Energy storage systems (ESS) are pivotal for stabilizing power grids, integrating renewable sources, and enabling the widespread adoption of electric vehicles. Among various storage technologies, lithium-ion batteries dominate due to their high energy density and long cycle life. However, thermal management remains a critical challenge, especially for large-scale energy storage cells. Unlike automotive batteries, which experience high discharge rates and short operational cycles, energy storage cells typically operate at lower C-rates but over extended durations, demanding distinct cooling strategies. The optimal temperature range for lithium-ion cells is 25–40 °C, with temperature differences across modules kept below 5 °C to prevent capacity fade and thermal runaway. In this review, we focus on immersion cooling—a direct liquid cooling technique where dielectric fluids contact the cell surface—as a promising solution for energy storage cell thermal management. We examine key aspects including single-phase and two-phase immersion cooling, dielectric fluid selection criteria, system packaging, insulation coatings, and thermal runaway suppression. Our aim is to provide a comprehensive reference for designing efficient and safe cooling systems for energy storage cells.
Immersion cooling distinguishes itself from indirect liquid cooling by eliminating contact thermal resistance and enabling uniform heat extraction. In single-phase immersion cooling, the dielectric fluid remains liquid, absorbing sensible heat. In two-phase (phase-change) immersion cooling, the fluid boils, absorbing latent heat, which can be more effective for high heat fluxes but requires complex vapor management. The following figure illustrates a typical immersion cooling arrangement for a battery module.
For energy storage cells, the system volume is larger than in automotive applications, allowing greater flexibility in fluid selection and system design. Additionally, preventing thermal runaway propagation is paramount, as a single cell failure can cascade into catastrophic events. Immersion cooling offers inherent safety advantages: the dielectric fluid can absorb heat from a failing cell, disrupt oxygen supply to combustible gases, and even dissolve some flammable species.
Single-Phase Immersion Cooling
Single-phase immersion cooling employs high-boiling-point dielectric fluids that remain in liquid state throughout operation. This simplicity reduces system complexity because no vapor handling or condensation loop is needed. Experimental studies have demonstrated that single-phase immersion cooling outperforms forced air cooling. For example, using a commercial dielectric fluid (AmpCool AC-100) with a 68 Ah cell at 2C charge/discharge resulted in an average temperature of 22.5 °C, while forced air cooling gave 28.7 °C. Numerical simulations further enable parametric studies on flow rate, immersion depth, and channel geometry. The heat transfer in single-phase immersion cooling is governed by convection:
$$
\dot{Q} = h A (T_s – T_f)
$$
where \( \dot{Q} \) is heat transfer rate, \( h \) is convective heat transfer coefficient, \( A \) is surface area, \( T_s \) is cell surface temperature, and \( T_f \) is fluid temperature. The Nusselt number for forced convection over cylindrical cells can be correlated as:
$$
Nu = C \, Re^m \, Pr^n
$$
where \( Re = \frac{\rho v D}{\mu} \) is the Reynolds number, \( Pr = \frac{c_p \mu}{k} \) is the Prandtl number, and \( C, m, n \) are empirical constants depending on flow regime and geometry. For natural convection in immersion systems, correlations differ, but the overall heat transfer coefficient is typically lower than forced flow. Studies show that increasing fluid velocity improves cooling but also raises pumping power, so an optimal flow rate exists (e.g., 15 mL/min for a 18650 cell in transformer oil).
Table 1 summarizes typical single-phase dielectric fluids and their key thermophysical properties for energy storage cell cooling applications.
| Fluid Type | Density @25°C (kg/m³) | Dynamic Viscosity @25°C (mPa·s) | Thermal Conductivity @25°C (W/(m·K)) | Specific Heat Capacity @25°C (kJ/(kg·K)) | Dielectric Strength (kV/mm) | Flash Point (°C) |
|---|---|---|---|---|---|---|
| AmpCool AC-100 | 820 | 8.11 | 0.136 | 2.21 | ≥20 | 193 |
| Transformer Oil #10 | 895 | 13 (40°C) | 0.12 | 1.8 | ≥35 (2.5mm gap) | 140 |
| Silicone Oil (PDMS, 10 cSt) | 935 | 9.35 | 0.134 | 1.6 | ≥15 | 163 |
| Novec 7500 | 1614 | 0.77 | 0.065 | 1.19 | ≥18 | None (non-flammable) |
Two-Phase (Phase-Change) Immersion Cooling
Two-phase immersion cooling exploits the latent heat of vaporization, enabling very high heat transfer coefficients, especially in the nucleate boiling regime. The heat transfer during boiling is given by:
$$
\dot{Q} = h_{nb} A (T_s – T_{sat})
$$
where \( h_{nb} \) is the nucleate boiling heat transfer coefficient, which can be orders of magnitude larger than single-phase convection. The critical heat flux (CHF) marks the upper limit of efficient boiling. For energy storage cells operating at moderate heat fluxes, maintaining nucleate boiling is key. The saturation temperature of the fluid can be actively controlled by adjusting system pressure. For instance, using Novec 7000 (boiling point 34 °C at 1 atm), a 20 Ah pouch cell discharged at 5C maintained a surface temperature of approximately 34.5 °C. However, two-phase systems require vapor condensation and recycling, adding complexity. Intermittent flow strategies have been proposed to mitigate temperature non-uniformity caused by vapor blanketing. Table 2 lists selected two-phase dielectric fluids suitable for energy storage cell immersion cooling.
| Fluid | Boiling Point @1 atm (°C) | Latent Heat of Vaporization (kJ/kg) | Liquid Density @25°C (kg/m³) | Liquid Thermal Conductivity @25°C (W/(m·K)) | Dielectric Constant | GWP |
|---|---|---|---|---|---|---|
| Novec 7000 | 34 | 142 | 1400 | 0.075 | 7.4 | 530 |
| Novec 649 | 49 | 88 | 1600 | 0.059 | 1.8 | 1 |
| Opteon MZ | 33.4 | 166 | 1360 | 0.077 | 32 | 2 |
| SF33 | 33 | 155 | 1350 | 0.074 | 8.5 | 4 |
Dielectric Fluid Selection Criteria
We propose four key parameters for selecting a dielectric fluid for energy storage cell immersion cooling: specific heat capacity, heat transfer coefficient, dielectric strength, and material compatibility. Additional considerations include viscosity, flash point, ozone depletion potential (ODP), global warming potential (GWP), and cost. The following subsections detail the four main categories of dielectric fluids with practical application potential.
Fluorinated Electronic Liquids
Fluorinated liquids (e.g., perfluorocarbons, hydrofluoroethers, and hydrofluoroolefins) offer excellent dielectric properties and chemical inertness. However, many are regulated due to environmental impacts (see Table 3). Recent developments focus on low-GWP fluids such as HFO-based dielectrics (e.g., Opteon MZ, Opteon 2P50). Their thermophysical properties are listed in Table 4.
| Fluid Type | Ozone Depletion? | Greenhouse Gas? | Key Regulation |
|---|---|---|---|
| CFCs | Yes | Yes | Montreal Protocol |
| HCFCs | Yes | Yes | Chinese ODS List |
| PFCs | Yes | Yes | Kyoto Protocol |
| HFCs | No | Yes | Kigali Amendment |
| HFEs | No | No | None |
| HFOs | No | No | None |
| Product | Boiling Point (°C) | Density @25°C (g/cm³) | Viscosity @25°C (mPa·s) | Thermal Conductivity @25°C (W/(m·K)) | Specific Heat @25°C (kJ/(kg·K)) | Dielectric Constant | GWP |
|---|---|---|---|---|---|---|---|
| Opteon MZ | 33.4 | 1.360 | 0.38 | 0.077 | 1.20 | 32 | 2 |
| Opteon 2P50 | 48.9 | 1.456 | 0.62 | 0.073 | 1.09 | — | 10 |
Hydrocarbon-Based Fluids
Hydrocarbon fluids include mineral transformer oils and synthetic hydrocarbons (e.g., AmpCool AC series). They are cost-effective but may suffer from oxidation, high viscosity, and limited biodegradability. Transformer oils typically have flash points above 140 °C and dielectric strengths >35 kV (2.5 mm gap). Table 5 provides typical properties of transformer oils used for energy storage cell immersion cooling.
| Grade | Density @20°C (g/cm³) | Kinematic Viscosity @40°C (mm²/s) | Flash Point (°C) | Pour Point (°C) | Breakdown Voltage (2.5 mm gap, kV) |
|---|---|---|---|---|---|
| #10 | 0.895 | <13 | >140 | <-7 | >35 |
| #25 | 0.895 | <13 | >140 | <-22 | >35 |
| #45 | 0.895 | <13 | >135 | <-45 | >35 |
Synthetic hydrocarbons like AmpCool AC-110 offer improved environmental profiles (biodegradability >93%, GWP=0). Their thermal conductivities range 0.136–0.159 W/(m·K) and specific heats around 2.2 kJ/(kg·K).
Silicone Oils (Polydimethylsiloxane, PDMS)
Silicone oils are chemically inert, have low toxicity, and excellent dielectric properties. Their viscosity can vary widely (0.65–1,000,000 mm²/s) depending on chain length. Table 6 lists properties of commercial PDMS oils with viscosities from 0.65 to 1000 cSt.
| Product Code | Kinematic Viscosity (mm²/s) | Viscosity-Temperature Coefficient | Pour Point (°C) | Density (g/cm³) | Thermal Conductivity (W/(m·K)) | Dielectric Constant | Flash Point (°C) |
|---|---|---|---|---|---|---|---|
| DMS-T00 | 0.65 | 0.32 | -68 | 0.761 | 0.1005 | 2.20 | -1 |
| DMS-T01 | 1 | 0.37 | -85 | 0.818 | 0.1005 | 2.30 | 39 |
| DMS-T11 | 10 | 0.56 | -65 | 0.935 | 0.1340 | 2.68 | 163 |
| DMS-T21 | 100 | 0.60 | -65 | 0.966 | 0.1549 | 2.75 | 315 |
| DMS-T31 | 1000 | 0.61 | -50 | 0.971 | 0.1591 | 2.75 | 315 |
PDMS has been validated for long-term immersion of electronic devices. Its compatibility with battery materials is generally good, though chemical modification (e.g., phenyl substitution) can enhance thermal stability and oxidation resistance.
Nanofluids
Nanofluids—nanoparticle suspensions in base fluids—can significantly improve thermal conductivity. For oil-based nanofluids, adding graphene, Al₂O₃, or SiO₂ nanoparticles increases k by up to 3×. However, conductive nanoparticles may compromise dielectric strength. Coating particles with insulating layers (e.g., SiO₂ on graphene) helps maintain electrical insulation. Water-based nanofluids are usually non-dielectric and require additional cell insulation. Table 7 compares thermal conductivity enhancement of selected nanofluids for energy storage cell immersion cooling.
| Base Fluid | Nanoparticle | Volume Fraction (%) | Thermal Conductivity Ratio (k_nf / k_base) | Dielectric Strength Change |
|---|---|---|---|---|
| Dimethyl silicone oil | Graphene oxide (C1) | 20 | 3.0 | Not significantly reduced (with coating) |
| Transformer oil | SiO₂ | 0.1–0.5 | 1.15–1.25 | Enhanced |
| Water-EG (50:50) | Al₂O₃ | 0.5 | 1.10 | Increased conductivity (not dielectric) |
Stability and uniformity remain challenges for nanofluids in long-term energy storage cell applications. Future work should focus on ensuring suspension stability and maintaining dielectric integrity.
System Packaging and Insulation Coatings
To minimize coolant volume and pumping power, special encapsulation designs have been proposed. For prismatic cells, a manifold-jet impingement structure enables direct cooling with minimal fluid inventory. For pouch cells, U-shaped channels integrated into the cell housing reduce thermal resistance by an order of magnitude compared to indirect liquid cooling. Additionally, insulating coatings allow the use of non-dielectric coolants (e.g., water) for higher heat transfer performance. Parylene C coatings (1–25 μm thick) provide electrical isolation without significant thermal penalty. Composite coatings of silicone sealant and boron nitride have been demonstrated to protect 18650 cells immersed in water, achieving excellent cooling with negligible leakage current.
Thermal Runaway Suppression via Immersion Cooling
Lithium-ion battery thermal runaway is typically triggered by internal short circuits or abusive conditions. The pre-runaway stage involves gradual heat accumulation over minutes to days, offering a window for intervention. Immersion cooling can effectively extract heat during this stage, preventing temperature rise to critical levels (~180–250 °C). Moreover, the dielectric fluid acts as a physical barrier, isolating the failing cell from oxygen and reducing the risk of fire or explosion. Some fluids can even dissolve flammable gases released during decomposition. Experiments using Novec 649 with a three-cell pouch module showed that when the middle cell was overcharged to trigger thermal runaway, its temperature peaked at 183.9 °C, but no fire or propagation to adjacent cells occurred due to boiling immersion cooling. Numerical simulations of an 840-cell module in silicone oil demonstrated that the heat dissipation rate of immersion cooling was twice that of indirect cooling, reducing the temperature rise in neighboring cells by 8.9 °C during a simulated thermal runaway event.
The chemical pathways to runaway involve exothermic reactions between cathode, anode, and electrolyte. While fluorinated electrolytes can slow reactions, complete mitigation remains elusive. Immersion cooling, however, provides a robust thermal barrier that can arrest runaway before it spreads. The effectiveness depends on fluid boiling point, heat capacity, and flow rate. We recommend that for energy storage cell systems, the dielectric fluid be selected with a boiling point above the normal operating range but low enough to activate evaporative cooling during pre-runaway heating.
Conclusion
Immersion cooling is a promising thermal management technology for energy storage cells, offering high heat transfer rates, uniform temperature distribution, and intrinsic safety benefits. Single-phase systems are simpler and suitable for moderate heat loads, while two-phase systems excel at high heat fluxes with careful vapor management. Dielectric fluid selection must balance thermophysical performance, electrical insulation, environmental impact, and cost. Fluorinated fluids provide excellent inertness but face regulatory scrutiny; hydrocarbons are cost-effective but may degrade; silicone oils offer good dielectric properties and flexibility; nanofluids enhance conductivity but require stability and insulation assurance. System encapsulation and insulation coatings enable further optimization and even allow the use of non-dielectric coolants. Crucially, immersion cooling can suppress thermal runaway by absorbing heat and isolating cells from oxygen, making it a key technology for large-scale energy storage safety. Future research should focus on long-term fluid compatibility, multi-physics modeling coupling electrochemistry and fluid dynamics, and cost-effective system designs for commercial energy storage cell installations.