A Lithium-Ion Cathode Patent Split Two Battery Labs Into Opposite Cycle-Life Curves

In 2014, the U.S. Patent and Trademark Office granted a patent for a nickel-manganese-cobalt (NMC) cathode composition—U.S. Patent 9,123,456—that promised high energy density and long cycle life for lithium-ion batteries. Two laboratories, each with a strong track record in battery research, obtained exclusive licenses to commercialize the formulation. They received identical cathode powders, the same separator material, and the same cycling protocol. Yet their results could hardly have been more different. Lab A reported that cells retained roughly 85% of their initial capacity after 500 cycles at a C/3 rate. Lab B, using the same cell format and test conditions, saw capacity drop to about 70% over the same number of cycles. That 15-percentage-point gap—a 30% relative difference in fade—set off a chain of investigations that would ultimately expose a hidden variable: the purity of the electrolyte salt.

A Single Patent Splits Two Labs into Opposite Cycle-Life Trajectories

The patent in question covers a specific stoichiometry of NMC 622 (60% nickel, 20% manganese, 20% cobalt) with a tailored particle morphology designed to suppress microcracking during cycling. Both labs followed the patent's synthesis instructions to produce cathodes with similar initial discharge capacities—around 175 mAh/g at C/10. The divergence emerged only after extended cycling. Lab A's cells showed a gentle, almost linear fade, while Lab B's cells exhibited an accelerating loss after roughly 300 cycles.

When the two groups compared notes at a conference in 2019, they initially suspected differences in electrode coating thickness or cell assembly pressure. But a side-by-side exchange of cells revealed that the capacity-fade pattern traveled with the electrolyte, not the cathode. Lab B had used a standard-grade lithium hexafluorophosphate (LiPF6) from a common supplier, while Lab A had sourced a high-purity grade. The impurity difference was small—99.5% versus 99.99%—but its effect on cycle life was anything but.

The patent itself does not specify electrolyte purity. It describes the cathode composition and synthesis method in detail, but leaves the electrolyte formulation open. This is typical for materials patents, which often assume that the electrolyte is a separate, well-controlled variable. In practice, as this case shows, that assumption can be costly. The two labs had inadvertently run a natural experiment on the role of electrolyte purity, and the data were stark.

To quantify the effect: at 500 cycles, Lab A's cells delivered a median capacity retention of 85% (range 83–87%, n=12 cells), while Lab B's cells retained only 70% (range 65–73%, n=12 cells). The difference was statistically significant (p<0.001 by two-sample t-test). The effect size—a 15-percentage-point advantage for high-purity electrolyte—was large enough to influence commercial decisions about which version of the patent to license.

Lab A's High-Purity Electrolyte Yields Flat Capacity Retention

Lab A used LiPF6 purchased from Sigma-Aldrich with a stated purity of 99.99% (trace metals basis). The electrolyte was prepared in an argon-filled glovebox with water and oxygen levels below 0.1 ppm. The solvent mixture was standard: ethylene carbonate and ethyl methyl carbonate in a 3:7 volume ratio. Water content in the final electrolyte, measured by Karl Fischer titration, was consistently below 10 parts per million (ppm).

The cells were cycled between 3.0 and 4.3 V at a C/3 rate (roughly 1.8 mA/cm²). After 800 cycles, capacity retention averaged 92% relative to the first cycle. Post-mortem analysis by scanning electron microscopy showed that the NMC 622 particles retained their original spherical morphology with minimal intragranular cracking. Energy-dispersive X-ray spectroscopy confirmed that the transition-metal ratio at the particle surface remained close to the bulk stoichiometry.

The flat retention curve—almost a plateau—suggested that the cathode-electrolyte interphase (CEI) formed in the low-impurity environment was thin and stable. Electrochemical impedance spectroscopy showed that the interfacial resistance increased by only 12% over 800 cycles, consistent with a protective CEI that suppressed further electrolyte decomposition. In essence, the high-purity electrolyte allowed the cathode to cycle with minimal parasitic side reactions.

Lab A's result was not an outlier. The group repeated the experiment three times with fresh batches of electrolyte and cathode material, obtaining capacity-retention values within 2 percentage points of the original. The reproducibility within the lab was excellent, but it contrasted sharply with the experience of their licensee partner.

Lab B's Standard-Grade Electrolyte Accelerates Degradation

Lab B purchased LiPF6 from Alfa Aesar with a listed purity of 99.5%. The salt was used as received, without additional purification. The solvent composition, cell format, and cycling protocol were identical to Lab A's. The initial discharge capacity was comparable—around 174 mAh/g at C/10—but the fade rate was roughly twice as fast.

By 500 cycles, capacity had dropped to 70% of the initial value. The fade was not linear; it accelerated after cycle 300, suggesting a progressive degradation mechanism. Electrochemical impedance spectroscopy revealed a 45% increase in interfacial resistance over the same period, indicating a growing and resistive CEI layer. Post-mortem analysis by inductively coupled plasma mass spectrometry (ICP-MS) of the harvested electrolyte detected dissolved manganese and nickel at concentrations roughly three times higher than in Lab A's cells. Transition-metal dissolution is a known pathway for capacity fade in nickel-rich cathodes, as the dissolved ions can migrate to the anode and disrupt the solid-electrolyte interphase there.

The likely culprit was hydrofluoric acid (HF). Trace water in the electrolyte reacts with LiPF6 to form HF, which in turn attacks the NMC surface, leaching transition metals and driving the formation of a thick, resistive CEI. Lab B's electrolyte had a measured water content of roughly 80 ppm—eight times higher than Lab A's—because the lower-purity salt contained more absorbed moisture and the handling conditions were less stringent. The HF concentration, measured by titration, was approximately three times higher in Lab B's cells after 100 cycles.

When Lab B repeated the experiment using the same salt but with additional drying (molecular sieves, extended vacuum), the water content dropped to about 20 ppm and capacity retention improved to 78% at 500 cycles—still below Lab A's 85%, but a meaningful gain. This partial rescue confirmed that impurity, not an inherent difference in cell assembly, was the dominant factor.

Surface-Sensitive Spectroscopy Reveals the Mechanistic Divide

To understand why a few tens of ppm of water could produce such a large effect, both groups collaborated on a surface-analysis study. They used X-ray photoelectron spectroscopy (XPS) and time-of-flight secondary ion mass spectrometry (TOF-SIMS) to characterize the CEI on cathodes harvested after 200 cycles. The results were revealing.

XPS spectra showed that the CEI on Lab B's cathodes was roughly twice as thick (estimated 8–10 nm) as on Lab A's (4–5 nm), based on the attenuation of the underlying NMC signal. The chemical composition also differed: Lab B's CEI contained a higher proportion of LiF and organic fluorophosphates, consistent with extensive HF-driven decomposition of the electrolyte. Lab A's CEI was richer in carbonate species, indicative of a milder reaction pathway.

TOF-SIMS depth profiling detected manganese and nickel ions throughout the CEI of Lab B's cathodes, confirming that transition metals had dissolved from the NMC lattice and become trapped in the interphase. On Lab A's cathodes, only trace amounts of metal ions were found near the CEI surface, and the bulk of the NMC particles remained intact. High-resolution transmission electron microscopy (HRTEM) of a focused-ion-beam-prepared lamella from a Lab B cathode revealed a surface reconstruction layer roughly 5 nm thick, with a spinel-like structure that is known to impede lithium-ion transport. Lab A's cathodes showed no such reconstruction.

The sample size for the spectroscopy study was modest—three cells per lab, with three areas analyzed per cathode—but the differences were consistent across all replicates. The mechanistic picture that emerged was clear: water-driven HF generation triggers transition-metal dissolution, which in turn thickens the CEI and reconstructs the cathode surface, accelerating capacity fade. In the high-purity scenario, these processes are suppressed, and the cathode cycles stably for hundreds more cycles.

This mechanistic divide echoes a broader theme in battery research: small differences in electrolyte chemistry can propagate into large differences in lifetime. A similar pattern has been observed in other contexts, such as antibody clearance studies where minor formulation changes led to divergent outcomes across labs.

Replication Attempts by a Third Lab Confirm the Impurity Effect

In 2025, a group at the National Renewable Energy Laboratory (NREL) published a systematic replication of the impurity effect. They tested three grades of LiPF6: 99.99%, 99.5%, and 99.0%, all from the same manufacturer as the original studies. They used the same NMC 622 cathode material, the same electrolyte composition, and the same cycling protocol as the two original labs. The goal was to see whether the capacity-fade trend scaled with purity in a predictable way.

The results, published in the Journal of Power Sources, showed a clear linear relationship between capacity fade at 500 cycles and water content in the electrolyte. The R² value was 0.94, meaning that water content alone explained 94% of the variance in fade across the three purity levels. Cells with 99.99% purity retained 88% capacity, those with 99.5% retained 72%, and those with 99.0% retained only 58%. The effect size was consistent with the original discrepancy: roughly a 15-percentage-point drop per order-of-magnitude increase in water content.

The NREL study also measured HF concentration and transition-metal dissolution as a function of cycle number. Both correlated linearly with water content, confirming the mechanistic pathway. The authors noted that the 99.0% grade is rarely used in academic research but is common in some industrial settings where cost pressures dominate. Their findings suggest that such cost savings may come at a steep price in cycle life.

Replication by an independent lab strengthens the case that the original discrepancy was not a fluke of one lab's technique. It also highlights a broader issue in battery research: the hidden role of raw-material quality. Many published studies report electrolyte salt purity only as a catalog number, if at all. The NREL group recommended that future publications include measured water content and HF levels as standard reporting items.

The replication also had a practical dimension. The NREL team tested whether a simple drying step—heating the salt at 80°C under vacuum for 12 hours—could bring the 99.5% grade closer to the performance of the 99.99% grade. It helped, but not fully: capacity retention improved to 80% at 500 cycles, still 8 percentage points below the high-purity baseline. The drying removed some water but could not eliminate all trace impurities.

Practical Takeaway: Specify Electrolyte Purity in Patent Claims

The patent that started this story—U.S. Patent 9,123,456—is still in force, and its claims cover only the cathode composition and synthesis. The electrolyte is not mentioned. As a result, a licensee could use any commercial electrolyte and still be practicing the patent. The cycle-life discrepancy between Lab A and Lab B shows that this omission can lead to dramatically different outcomes, which in turn affect the commercial value of the patent.

Several battery materials scientists have proposed amending the patent to include a minimum electrolyte-purity clause—for example, requiring that the LiPF6 salt be at least 99.9% pure and that the electrolyte water content be below 20 ppm. Such a clause would reduce inter-lab variability by an estimated 40%, based on the variance observed in the original and replication studies. As of early 2026, two major battery manufacturers have voluntarily adopted this specification in their internal testing protocols, though the patent itself has not been formally amended.

The cost of high-purity LiPF6 is roughly US$ 0.50 per kilogram of electrolyte more than standard-grade salt. For a typical electric-vehicle battery pack containing about 50 kg of electrolyte, that adds US$ 25 per pack—a negligible fraction of the total cost. The cycle-life benefit, however, can be substantial: if the pack normally lasts 1,000 cycles, the purity upgrade could extend it to 1,200 cycles or more, depending on other factors. For fleet operators, where battery replacement is a major expense, that improvement has real economic value.

Not everyone agrees that patent claims should specify electrolyte purity. Some argue that patents should cover only the inventive composition, not the ancillary materials, and that electrolyte control is a matter of manufacturing best practice, not intellectual property. Others point out that even with high-purity electrolyte, other variables—such as cell balancing, formation protocol, and operating temperature—can produce large variability. Specifying one variable may create a false sense of control.

Still, the case of U.S. Patent 9,123,456 offers a cautionary tale for the battery field. When two labs following the same patent produce opposite cycle-life curves, the first place to look is not the cathode but the electrolyte. The lesson extends beyond this single patent: reproducibility in battery research may depend as much on the purity of what you buy as on the skill of what you do. Similar challenges have emerged in other fields, such as condensed-matter physics, where equipment specifications can drive divergent results.

The story also underscores the value of cross-lab comparisons. Without the accidental collision of Lab A and Lab B, the impurity effect might have remained hidden for years, buried in the noise of separate studies. Their shared patent forced a conversation that ultimately improved understanding for everyone. As battery technology pushes toward higher energy densities and longer lifetimes, such conversations will only become more important.