Mixing Secrets: How Carbon Black Distribution Shapes Solid-State Battery Performance

Solid-state lithium-metal batteries have long been heralded as the successor to the lithium-ion cells that have dominated the market since 1991. By replacing the flammable organic liquid electrolyte with a solid polymer electrolyte, these batteries promise improved safety, the possibility of using a lithium metal anode with a theoretical capacity of 3860 mAh per gram compared to 372 mAh per gram for graphite, and consequently higher gravimetric and volumetric energy density. Yet a stubborn obstacle has kept them out of commercial devices: the relatively low ionic conductivity of polymer electrolytes, which is typically addressed by incorporating highly dissociated lithium salts such as LiTFSI into a poly(ethylene oxide) host. Now, a team of researchers led by L. Fernandez-Diaz and M.C. Morant-Miñana, publishing in Results in Engineering, has shown that one of the most decisive factors in solid-state battery performance is not the chemistry of the materials at all, but something far more mundane: how long and how fast the electrode slurry is mixed.

The composite cathode of a solid-state battery is a dense, non-porous blend of active material, conductive additive and a portion of solid electrolyte known as the catholyte. Unlike conventional lithium-ion cells, where a liquid electrolyte seeps into every pore, the solid-state version has no pore-filling liquid to rescue poor contacts. Every particle must touch its neighbour intimately, and both ions and electrons need homogeneous, percolating pathways through the electrode. The conductive additive in this study is Super C65, a carbon black from Imerys, which must form an unbroken electronic network threading through the cathode. The active material is carbon-coated lithium iron phosphate, or LFP, whose own carbon coating enhances intrinsic electronic conductivity and interfacial lithium transfer. The researchers were careful to distinguish between these two carbons: the factory-applied coating on the LFP particles and the C65 introduced during electrode processing behave very differently under the forces of mixing.

To isolate the effects of processing, the team prepared slurries under dry-room conditions with a dew point below minus 60 degrees Celsius, using a bead mill to first dissolve LiTFSI in acetonitrile and then mix in high molecular weight PEO to form the catholyte. The LFP and C65 powders were dry-mixed in a planetary vacuum mixer at 200, 750 or 1500 revolutions per minute, and the resulting powder blend was then wet-mixed into the catholyte for 60, 120 or 180 minutes. Five slurries were produced in total, each cast onto carbon-coated aluminium current collectors and dried under vacuum. Crucially, the researchers deliberately skipped calendering, the mechanical compaction step common in electrode manufacturing, so that the morphological fingerprints left by the mixing process itself would remain visible in the finished electrodes.

One of the study’s most elegant methodological contributions is the introduction of a relative viscosity parameter. Batch-to-batch variability in the PEO raw material, stemming from differences in molecular weight distribution, meant that absolute viscosity measurements could not be compared reliably across samples. By normalizing the slurry viscosity to the viscosity of the corresponding catholyte at the end of its own mixing, the team could decouple the true effects of mixing time and speed from uncontrolled fluctuations in the polymer feedstock. Validation experiments with four separate preparations, including three different PEO batches, confirmed that this normalization substantially reduced variability, offering a practical tool that other laboratories can adopt to make their slurry comparisons meaningful.

The rheological results revealed a counterintuitive trend. Longer wet mixing times and higher dry mixing speeds both reduced slurry viscosity, and the researchers traced this not to further polymer degradation but to the progressive deagglomeration of the carbon black. Gel permeation chromatography showed that the PEO molecular weight decreased exponentially toward a plateau, with an apparent scission-rate constant of 7.49 times ten to the minus seven per minute, but since total mixing times stayed below the point where degradation becomes negligible, the viscosity drop had to come from elsewhere. As C65 agglomerates break apart, the number of carbon particles rises, promoting immobilization of polymer chains and reducing chain entanglement, a mechanism previously described by Hoffman and colleagues. More effective deagglomeration also increases interparticle volume, allowing more solvent molecules to occupy the spaces between carbon particles and thereby lowering the suspension viscosity.

These rheological differences translated directly into electrode microstructure. Scanning electron microscopy of surfaces and argon ion-milled cross-sections, analysed with ImageJ, showed that longer wet mixing produced electrodes with lower porosity, higher active material loading and fewer surface lumps, because lower viscosity allows better particle rearrangement and compaction during drying. In the dry mixing series, the sample processed at 1500 rpm displayed the most homogeneous morphology, while the 200 rpm sample exhibited a highly cracked surface, suggesting that large C65 agglomerates weaken the mechanical integrity of the composite cathode and hinder uniform polymer distribution. Porosity values ranged from about 1.65 percent for the longest wet mixing time to nearly 3 percent for the mildest conditions, differences that proved consequential for transport properties.

Electronic and ionic conductivity measurements told a more nuanced story. Reference pellets containing as little as 5 weight percent C65 achieved electronic conductivities of roughly ten to the minus two siemens per centimetre, confirming that this modest carbon fraction is sufficient for percolation. Yet after full electrode processing, bulk electronic conductivity dropped by nearly two orders of magnitude at both 25 and 70 degrees Celsius. Wet mixing time had little effect on this bulk conductivity, but in-plane resistivity measurements revealed that excessive wet mixing stresses the carbon particles and degrades the conductive network. In the dry mixing series, the intermediate 750 rpm condition delivered the lowest interfacial resistance and the highest ionic conductivity, while the 1500 rpm sample, despite its denser packing, showed signs of carbon particles being over-stressed and unable to form an optimal network.

Full-cell testing in monolayer pouch cells with lithium metal anodes, cycled at 70 degrees Celsius between 2.6 and 3.7 volts, exposed the trade-offs with striking clarity. At the gentle C/20 rate, all cells delivered similar capacities between 156 and 158 mAh per gram, but at higher rates the mildest processing conditions won out: the 60-minute wet mix and the 200 rpm dry mix, both with the highest slurry viscosities and highest porosities, achieved the highest capacities at C/5. The researchers attribute this to the higher viscosity limiting conductive additive redistribution, thereby preserving a favourable electronic network, while mud cracking in the 200 rpm electrode may have created preferential channels for solid electrolyte infiltration during cycling. However, these high-capacity electrodes also showed the poorest cycle life, with the 200 rpm sample lasting only 65 cycles before coulombic efficiency fell below 98 percent, compared to 120 cycles for the balanced 750 rpm condition.

The degradation mechanism was revealed through X-ray computed micro-tomography of intact pouch cells, a technique that avoids the impossible task of disassembling strongly adhesive cycled cells. Cross-sections of a pristine cell showed homogeneous layers with only minor pores at the cathode-electrolyte interface, but after 130 cycles the lithium metal anode had thinned from 500 to 470 micrometres, voids had multiplied in the cathode, and dark regions appeared within the solid electrolyte layer, signatures of lithium dendrites growing from the anode through the polymer. Grain boundaries in the electrolyte, attributed to solid electrolyte interphase formed by soft dendrites, were also visible. Impedance spectroscopy of the aged cell confirmed diminished ionic diffusion, tying the loss of coulombic efficiency directly to dendrite propagation.

The broader lesson of this work is that in solid-state batteries, processing is not a secondary consideration but a first-order design variable. The optimal wet mixing time of 120 minutes balanced electrode compaction against preservation of the conductive network, while an intermediate dry mixing speed of 750 rpm achieved the best cyclability without compromising lithium-ion transport. Even a small amount of residual porosity, the authors suggest, may serve as stress-relief space accommodating the volume changes of lithium insertion and extraction in a mechanically rigid system. As solid-state batteries edge toward commercialization, this study demonstrates that the path to better cells may run through the mixing vessel, where the fate of every carbon black agglomerate ultimately determines whether electrons and ions can find their way home.

Subject of Research: Effect of slurry mixing parameters on carbon black distribution and performance of solid polymer electrolyte composite cathodes

Article Title: Carbon black distribution in low viscosity polymer slurries

Article References: Carbon black distribution in low viscosity polymer slurries. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: solid-state batteries, carbon black, polymer electrolyte, slurry rheology, composite cathode, lithium metal anode, PEO, LiTFSI, electrode processing, lithium dendrites, X-ray tomography, LFP