During high-drain cylindrical cell assembly, electrolyte filling and wetting are frequently the throughput-limiting unit operations once electrode areal capacities exceed 3.0 mAh cm⁻² and separator thickness falls below 16 µm. EMC-rich carbonate blends are selected for these cells because ethyl methyl carbonate reduces the room-temperature dynamic viscosity to 0.65–0.85 mPa·s at 25 °C and lowers the freezing point of the solvent mixture, but this formulation choice introduces competing risks: high EMC content raises the vapor pressure, depresses the flash point in some blend ratios, and reduces the dielectric constant available for lithium salt dissociation. On production-scale rotary vacuum filling equipment, the dose volume is typically 3.8–5.2 mL per 21700 cell, and the fill head must dispense against a chamber pressure ramp from −0.085 MPa to −0.095 MPa relative to atmosphere. The electrolyte is drawn into the wound jelly roll by a combination of external pressure differential and capillary pressure generated at the separator/electrode interfaces. Wetting defects that remain after the pressure-cycling step are not immediately detectable by open-circuit voltage screening; they become visible only during formation cycling as local lithium plating near the core or as anomalous cell-to-cell impedance scatter above 0.5 mΩ in end-of-line testing.

Representative property ranges for EMC-rich carbonate electrolyte blends and separator wetting indicators

Parameter	Standard / method	Representative range	Process relevance
Kinematic viscosity at 25 °C	ASTM D445-21	0.65–0.85 mm²/s	Washburn penetration rate
Density at 25 °C	ASTM D4052-22	1.05–1.18 g/cm³	Shot-weight verification
Surface tension	ISO 19403-4:2017	28.5–31.5 mN/m	Capillary pressure
Advancing contact angle on uncoated polypropylene separator	ISO 19403-2:2017	52°–64°	Separator wetting
Advancing contact angle on alumina-coated polyethylene separator	ISO 19403-2:2017	35°–45°	Ceramic separator wetting
Flash point	ASTM D3278-21	23–35 °C	Filling-room safety classification
Ionic conductivity at 25 °C with 1.0 mol L⁻¹ LiPF₆	impedance spectroscopy / IEC 62660-1:2022	6.5–8.0 mS/cm	High-rate performance

Why Does EMC-Rich Blend Viscosity Control Capillary Imbibition in Separator Pores?

The Washburn equation provides the first-order relationship for one-dimensional wetting into a cylindrical pore: L² = (γ cos θ r t)/(2η), where L is penetration length, γ surface tension, θ contact angle, r pore radius, t time, and η dynamic viscosity. In EMC-rich blends the viscosity term is favorable, but the contact angle on untreated polyolefin separators is less favorable because polyethylene and polypropylene surfaces have low polar surface energy and interact weakly with carbonate solvents. Typical reported contact angles on uncoated polyethylene separators fall in the range 52°–64° for EMC-rich electrolytes, while ceramic-coated polyethylene surfaces reduce the advancing contact angle to 35°–45° when the coating contains alumina with a mean particle size below 0.8 µm. The effective capillary pressure is 2γ cos θ/r; for a pore radius of 0.15 µm, the ceramic-coated separator can generate a capillary pressure near 0.38–0.42 MPa, whereas an uncoated separator with the same geometry generates approximately 0.26–0.30 MPa. These pressure differences determine whether vacuum filling alone is sufficient or whether externally applied pressure cycling is required to displace trapped gas. The viscosity of EMC-rich blends measured by ASTM D445-21 is not directly equivalent to the dynamic viscosity required in the Washburn equation unless the density is accounted for; for production troubleshooting, the kinematic viscosity must be multiplied by the density from ASTM D4052-22 to obtain the dynamic viscosity in mPa·s. Because carbonate electrolyte blends behave as Newtonian liquids up to shear rates above 1000 s⁻¹ in cone-and-plate measurements, single-point viscosity data are generally sufficient for filling-line calculations.

In vacuum-assisted filling of 21700 cells, the chamber pressure is reduced to approximately −0.095 MPa relative to atmospheric before the electrolyte dose is injected through the central void of the wound jelly roll. The external pressure gradient across the jelly roll is superimposed on the capillary pressure at the separator surface. A production-scale six-station rotary vacuum filler typically operates with a dry-room dew point below −40 °C and uses pressure-assisted dwell cycles of 0.4–0.8 MPa nitrogen or argon for 120–300 s to force electrolyte into residual gas-occupied pores. The choice of EMC-rich blends shortens the necessary dwell time when viscosity is the controlling factor, but it lengthens the time to degas the liquid because the higher vapor pressure of EMC requires deeper vacuum before the first pressure cycle. If the chamber pressure is not reduced below the vapor pressure of EMC at the fill temperature, the dose stream cavitates inside the metering valve, producing variable shot volumes. Online mass-flow verification between the storage tank and the dosing needle is commonly specified to control shot-weight variation to ±0.5%; for a 4.0 mL dose, this corresponds to ±0.020 mL. Residual gas bubbles trapped near the mandrel remain a documented failure mode during teardown of high-rate cells: the negative electrode in the innermost winding layer shows dull grey patches rather than uniform golden graphite, and lithium plating is observed after formation at 1C rates. The operational boundary for this filling sequence is that EMC-rich blends containing more than 70 vol% EMC should not be held under deep vacuum for extended periods because preferential evaporation shifts the solvent ratio and can increase the local EC concentration near the liquid–vapor interface, raising viscosity at the wetting front rather than lowering it.

When Electrode Calendering Pushes Pore Throats Below 120 nm, Wetting Time Scales Nonlinearly with EMC Content

Electrode calendering is used to set the electronic percolation and reduce the ionic tortuosity penalty, but it simultaneously narrows the pore-throat distribution. Mercury intrusion porosimetry per ASTM D4284-12 indicates that a high-drain NMC811 cathode calendered to a density of 3.30–3.45 g cm⁻³ exhibits a median pore-throat diameter near 100–180 nm, depending on the carbon-black and polyvinylidene fluoride binder network. At pore throats below 120 nm, the characteristic wetting time for a fixed pressure difference increases steeply because the permeability of the porous electrode scales roughly with the square of the pore diameter. EMC-rich blends mitigate the viscosity contribution to this rise, but they do not address the capillary pressure plateau caused by the pore geometry. If the blend is diluted with EMC beyond 60 vol%, the reduced dielectric constant of the solvent mixture, typically ε ≈ 7–9 relative to 19–23 for EC-rich formulations, lowers the ionic conductivity of the bulk electrolyte to 6.5–8.0 mS cm⁻¹ at 25 °C with 1.0 mol L⁻¹ LiPF₆. This creates a processing conflict for high-drain cells: the wetting rate improves with EMC content, but the high-rate discharge capacity after formation may fall below the specification if the bulk ionic conductivity is too low. Production-scale data from calendering trials show that a reduction in cathode pore-throat diameter from 160 nm to 110 nm can increase the required vacuum-pressure-cycling time by approximately 40–70% when the same EMC-rich electrolyte is used. This is not a linear effect; the added dwell time scales with the inverse of the permeability, which is a function of both the porosity and tortuosity of the electrode stack. For high-drain cells, the calendering window is often specified at ±0.1 g cm⁻³ around the target density because crossing that boundary changes both the adhesion and the wetting-time distribution. If the cathode is under-calendered, electrode adhesion to the current collector degrades and high-rate cycling delaminates the coating. If the cathode is over-calendered, the electrolyte cannot reach the full depth of the electrode before the formation charge begins, producing irreversible lithium plating on the anode. The operational boundary is therefore narrower than the mechanical tolerance of the calender; the wetting-time constraint, not the tensile strength of the coating, controls the upper density limit. Published data for the exact interaction between EMC content and pore-size distribution in high-drain cylindrical cells is limited, but the general trend is reproduced across multiple industrially relevant cathode loadings.

Separator Surface Energy, Ceramic Coating Architecture, and Electrolyte Spreading

The separator is the first liquid-contacting surface in the jelly roll and the most common origin of incomplete wetting. Polyolefin separators have surface energy well below the surface tension of carbonate electrolytes, producing a positive spreading coefficient only when the surface is modified. Ceramic-coated separators with alumina or boehmite layers are used in high-drain cells because the ceramic layer increases the polar component of the surface energy and provides a thermal-shutdown margin. ISO 19403-2:2017 describes a contact-angle method for polymer surfaces; it is adapted in production quality control to monitor separator surface treatment after corona or plasma exposure. Owens–Wendt analysis of the same contact-angle data separates the surface energy into dispersive and polar components; an uncoated polypropylene separator typically has a polar component below 5 mN m⁻¹, while an alumina-coated polyethylene separator can exhibit a polar component from 15 mN m⁻¹ to 25 mN m⁻¹. The higher polar component promotes wetting by carbonate solvents with a polar surface tension contribution from the carbonate carbonyl groups. Coating architecture matters as much as coating chemistry: a dense ceramic layer can block through-plane gas flow and create a bubble point above 0.5 MPa, while a discontinuous or particulate coating may allow gas to escape but provide uneven wetting paths. For high-drain cells, the separator must simultaneously satisfy the shutdown temperature requirement and the fill-line wetting requirement; these two targets can conflict. A separator with a thicker ceramic layer improves wetting but increases the ionic transport path length and lowers the cell’s pulse-power response. In cylindrical cells, the separator is wound under tension; the wound tension changes the effective pore radius because the separator is compressed against the electrode surface. Winding tension above 1.5 N per 10 mm width can reduce the measured separator thickness by 5–8%, which reduces the electrolyte retention volume but improves contact between layers. The formation of dry regions is therefore not simply a separator property; it is a system-level outcome of separator surface energy, winding tension, electrode pore structure, and the chosen electrolyte solvent ratio.

Typically, high-drain cylindrical cells with tabless current collectors demand faster filling but exhibit lower tolerance to residual dry regions because the current collection paths are shorter and local current densities can exceed 25 mA cm⁻² during pulse discharge. X-ray computed tomography after vacuum filling shows that the innermost windings and the shoulder region near the positive terminal are the two locations where gas entrapment persists most frequently. In a production-scale failure analysis, cylindrical cells with incomplete wetting displayed an impedance distribution that broadened by 0.3–0.6 mΩ across the cell population, while the mean open-circuit voltage remained within specification. This is why wetting defects are often classified as latent formation defects rather than fill-line rejects. To address the shoulder-region defect, some filling systems rotate the cell during electrolyte dosing to distribute the liquid circumferentially before the pressure cycle. The rotation speed is kept low, typically 60–120 rpm, to avoid foaming of the EMC-rich electrolyte. Foaming is a specific risk because the lower surface tension of EMC-rich blends stabilizes a higher population of small bubbles during rapid pressure changes. A de-foaming step with a slow pressure release ramp of 0.01 MPa s⁻¹ is often added when the EMC fraction exceeds 50 vol%. If the release is too fast, dissolved argon or nitrogen supersaturates and forms new bubbles inside the pores, reversing the wetting progress achieved during the pressure dwell. This operational rule is derived from filling-line observations rather than from a single standardized test; published wetting-performance standards specific to lithium-ion cylindrical cells are limited, so fill-line qualification commonly combines ASTM D445-21 viscosity data, ISO 19403-2:2017 contact-angle data, and internal pressure-cycling trials.

Electrolyte Additive Partitioning Modifies the Advancing Contact Angle on Ni-Rich Cathodes

Even when the separator and anode wet readily, the electrolyte must penetrate the porous cathode, where the surface chemistry of Ni-rich oxides differs significantly from carbonaceous anodes. The cathode surface contains residual lithium species, transition-metal oxides, and polyvinylidene fluoride binder domains. Electrolyte additives such as vinylene carbonate and fluoroethylene carbonate partition at the liquid–solid interface before formation, changing the advancing contact angle dynamically. In EMC-rich blends, the lower dielectric constant can cause additive dimerization or clustering at the wetting front, producing a contact-angle hysteresis that is not captured by single-drop equilibrium measurements. Time-resolved contact-angle measurements show an initial advancing angle that is higher on NMC811 than on graphite by roughly 8°–15°, but the receding angle remains higher as well, indicating that the electrolyte wets the cathode less reversibly. This hysteresis becomes process-relevant during pressure cycling: the liquid front advances during the pressure dwell, partially retreats when the pressure is released, and must re-advance during the next cycle. If the receding contact angle is too high, the hysteresis loop leaves isolated pores that never re-fill. The addition of 0.5–2.0 wt% of a low-molecular-weight carbonate-compatible surfactant can reduce the receding contact angle on the cathode, but the surfactant must later be electrochemically stable on the negative electrode; otherwise it degrades and consumes lithium inventory during formation. For high-drain cells, the formation protocol is usually designed to avoid high voltage before full wetting is achieved. A common formation sequence holds the cell at 1.5 V for an initial rest period of 6–24 h, then charges at C/20 to 3.6 V, then holds at 3.6 V for 12–24 h. This sequence provides time for slow capillary redistribution of the EMC-rich electrolyte into the remaining dry pores. If the cell is instead ramped directly to 4.2 V, the anode may polarize locally in dry regions and plate metallic lithium at current densities that would be benign in a fully wetted cell. The incompatibility here is not chemical but kinetic: EMC-rich blends wet quickly where a continuous liquid pathway exists, but the low dielectric constant reduces the ionic conductivity in partially filled pores, so the first formation current can find alternative low-resistance paths and initiate plating.

Storage and handling of EMC-rich blends require a dry-room environment with a dew point below −40 °C because the water solubility of the blend increases as the EMC fraction rises; water levels above 20 ppm in the finished electrolyte promote LiPF₆ hydrolysis and reduce the wetting repeatability by generating surface-active decomposition products. The filling line itself should be purged with dry nitrogen, and the electrolyte storage tank must be blanketed with argon or nitrogen to maintain moisture below 15 ppm and to prevent evaporation of the volatile EMC fraction. Separator rolls must be pre-dried at 60 °C under vacuum for at least 12 h if they have been stored at ambient relative humidity above 60% for more than 30 min; the ceramic coating can adsorb moisture and reduce wetting repeatability. Flash-point testing per ASTM D3278-21 is used to classify the blend for filling-room safety; blends with more than 70 vol% EMC can fall below the 35 °C flash-point threshold used in some regional storage codes, requiring explosion-proof dosing valves and vapor extraction. In addition, EMC-rich blends should not be combined with amine-containing cleaning agents because residual amines can initiate transesterification and form colored high-molecular-weight byproducts that alter the wetting front and reduce separator permeability by blocking pores. This incompatibility is often overlooked during filling-line changeovers, when the same dosing system is flushed with solvent cleaners. The safer cleaning solvent for EMC-rich carbonate systems is a dry carbonate mixture rather than an amine-based degreaser.