• What happened: A research team from Tohoku University and UACJ used manganese-doped aluminum foil as the anode of a lithium battery and achieved a first-cycle Coulombic efficiency of 81%. In a separate durability test, the foil maintained its capacity beyond 85 cycles.
  • Why it matters: A single sheet of aluminum foil can serve as both the anode active material that stores lithium and the current collector. The study also suggests that the boundaries of the fine crystal grains created by adding manganese may act as pathways for lithium to move.
  • What to watch next: Capacity continues to decline in cells paired with a cathode. A key challenge is whether a protective coating can suppress the pulverization and degradation that progress at the foil surface.

A research team from Tohoku University's Institute for Materials Research, including Assistant Professors Yuki Nakata and Hongyi Li and Professor Tetsu Ichitsubo, together with UACJ and others, has developed a technique that adds manganese to 50 µm-thick aluminum foil for use as the anode of a lithium battery. The foil released 81% of the charge first stored in it, a first-cycle Coulombic efficiency nearly equal to the 80% of pure aluminum foil. In a separate test that repeated charging and discharging under different conditions, pure aluminum foil stopped functioning within 40 cycles, whereas the manganese-doped foil maintained its capacity beyond 85 cycles.

The anode of a typical lithium-ion battery is made by coating an active material that stores lithium onto copper foil, the current collector that carries current to the external circuit. Aluminum foil, by contrast, forms an alloy with lithium by itself, so a single sheet can serve as both the anode active material and the current collector. The difficulty is that lithium taken deep into the foil is hard to extract repeatedly, which has been an obstacle to practical use.

What makes the result interesting is that calculations show manganese actually impedes lithium movement inside the aluminum crystal. So why does adding manganese make lithium easier to extract? Nakata and colleagues observed cross-sections of aluminum foil partway through lithium extraction to examine the structural changes occurring inside. The results were published in the journal Communications Materials on October 23, 2026, and Tohoku University released details in Japanese on October 26. UACJ also announced the results of the joint research on October 15.

AD

Comparing first-cycle Coulombic efficiency across three types of aluminum foil

When lithium is extracted from pure aluminum foil, a layer of aluminum with low lithium concentration forms first near the surface. This layer acts as a barrier that makes it hard to extract the lithium remaining inside. If cracks or pores form in the foil, the electrolyte can penetrate deeper, but reactions with the electrolyte proceed on the newly exposed surfaces and cause the foil to crumble into powder. Aluminum foil doped with silicon showed improved durability against repeated charging and discharging, but had the problem that a lithium-aluminum compound (LiAl) was left behind in dense regions.

The team prepared aluminum foil with a purity of 99.99% by weight, along with alloy foils to which silicon or manganese had been added at a nominal 1 atomic percent. All three types of foil were cold-rolled to the same thickness and punched into discs 16 mm in diameter, which were used directly as electrodes. This structure eliminates the need for the binders and conductive additives used in ordinary electrode manufacturing. X-ray diffraction analysis estimated that the proportion of the additive dissolved into the aluminum crystal lattice to form a solid solution was 0.64 ± 0.05 atomic % for manganese and 0.49 ± 0.07 atomic % for silicon.

In the first test, the team built "half cells" using metal lithium as the counter electrode, at room temperature with a carbonate-based electrolyte. In Figure 5b of the paper, the current density was set at 0.6 mA/cm², lithium equivalent to an areal capacity of 3 mAh/cm² was loaded into the foil, and lithium was then extracted until the voltage reached 1.0 V. The first-cycle Coulombic efficiency is the ratio of the charge extracted to the charge initially loaded.

Foil used as anode First-cycle Coulombic efficiency
Pure aluminum foil 80%
Aluminum-silicon alloy foil 56%
Aluminum-manganese alloy foil 81%

First-cycle Coulombic efficiency was 80% for pure aluminum, 56% for the silicon alloy, and 81% for the manganese alloy. The manganese alloy's 81% is nearly equal to pure aluminum, and the difference is not large enough to be judged a significant improvement. What matters is that the manganese alloy avoided the sharp drop in first-cycle efficiency seen with the silicon alloy.

The cycle life mentioned at the outset, meanwhile, comes from a separate durability test with the current density changed to 1 mA/cm², the capacity limit to 2 mAh/cm², and the cutoff voltage for lithium extraction to 1.5 V. Under these conditions, the silicon alloy also maintained its capacity up to about 65 cycles. The manganese alloy combined a first-cycle Coulombic efficiency close to that of pure aluminum with durability against repeated charging and discharging.

Manganese refines aluminum's crystal grains and promotes lithium movement

Manganese's effect is hard to explain by lithium movement inside the aluminum crystal alone. The team replaced one atom in a crystal model of 108 aluminum atoms with an additive element and calculated the energy barrier for lithium moving nearby. The barrier was 0.40 eV in pure aluminum and 0.34 eV near silicon, but rose to 0.89 eV near manganese. In other words, looking only at local movement inside the crystal, manganese works to hinder lithium diffusion.

However, real aluminum foil is made up of many crystal grains with different crystal orientations, and the boundaries between them, known as grain boundaries, can also serve as pathways for lithium. When the team examined the foil after the first charge and discharge under an electron microscope, silicon had formed clusters smaller than 100 nm. In the silicon alloy, the aluminum crystal grains were about 500 nm even at the small end and exceeded 1 µm at the large end. In the manganese alloy, by contrast, manganese remained uniformly distributed and the aluminum grains stayed within about 100 to 300 nm.

The authors believe that adding manganese suppressed grain growth during charging and discharging. Smaller grains shorten the distance lithium inside must travel to reach a grain boundary. Grain boundaries per unit volume also increase, which may create more pathways for lithium to move from the interior of the foil to the outside.

Clues supporting this interpretation also came from cross-sections of foil from which lithium had been partly extracted. First, when lithium equivalent to 3 mAh/cm² was loaded, the reacted region was about 18 µm thick. When 1 mAh/cm² of lithium had been extracted from the manganese alloy, aluminum crystal grains with a different orientation from the original structure formed by rolling appeared at the boundary between the reacted region and the unreacted aluminum deeper in. The authors believe this structure shows that lithium leaves LiAl and returns to aluminum even deep inside the foil, and that this change proceeds from the interior toward the surface. In a separate experiment examining the electrochemical response, the point at which resistance associated with diffusion in the solid rose sharply came later for the manganese alloy than for the silicon alloy.

Differences between materials also appeared in cross-sections of foil after lithium extraction was complete. In pure aluminum, the reacted region was 13 to 16 µm thick and contained many voids. In the silicon alloy, cracks and LiAl remained in a layer 10 to 12 µm thick, whereas the manganese alloy's reacted region retained a dense structure and had shrunk to about 9 µm. This thickness is close to what is estimated to be the thickness the reacted region occupied in the 50 µm foil before lithium was loaded.

The team also performed atomic-level calculations on LiAl that had already taken up lithium. They created vacant lattice sites where lithium atoms had left and tracked atomic motion for 2 ns at 900 K. With manganese added, both lithium and aluminum diffused faster, with aluminum movement especially promoted. The authors also point out that as the difference in diffusion rates between the two elements narrows, a layer that impedes lithium movement may be less likely to form at the surface.

However, these calculations targeted LiAl crystals after lithium uptake. The 0.89 eV value given earlier represents the local movement barrier in aluminum crystals before lithium uptake. Neither is a direct measurement of lithium moving along grain boundaries in actual aluminum foil at room temperature.

AD

Capacity maintained after 200 cycles, but decline continues

Manganese's effect was also confirmed in tests that varied the amount of lithium loaded. When the initial loading in half cells was increased from 1 mAh/cm² to 10 mAh/cm², the silicon alloy's first-cycle Coulombic efficiency fell from about 65% to below 50%. The manganese alloy, by contrast, fell only from about 88% to about 80%.

In another test, the lithium loading was fixed at 3 mAh/cm² and the current density was raised in steps every five cycles. Comparing each material's final cycle at the highest current density of 7.5 mA/cm², Coulombic efficiency was 78% for pure aluminum, 82% for the silicon alloy, and 98% for the manganese alloy. Even at high current density, the manganese alloy was shown to extract loaded lithium efficiently.

However, in half cells that use metal lithium as the counter electrode, the amount of lithium the cathode can supply is not limited as it is in a real battery, so it remains unclear how much capacity can be retained in practice. The team therefore also built full cells combining a lithium iron phosphate cathode with an areal capacity of 1.5 mAh/cm² and an aluminum foil anode. In this test, no pre-lithiation treatment was applied to either electrode, and a high-concentration ether-based electrolyte was used. Separate half-cell tests confirmed that for all three types of foil, an ether-based electrolyte allowed longer repeated charging and discharging than a carbonate-based one.

In the full-cell test, pure aluminum foil showed an initial capacity of about 0.9 mAh/cm² but stopped functioning within a few cycles. The silicon alloy's capacity also dropped sharply within the first 10 cycles and thereafter stayed below 0.2 mAh/cm². The manganese alloy, by contrast, showed an initial capacity comparable to pure aluminum and could still deliver about 0.25 mAh/cm² after 200 cycles. Capacity nevertheless continues to decline with charging and discharging.

The manganese alloy's average Coulombic efficiency was about 99.3%. This is the average ratio of charge extracted to charge loaded in each cycle, and differs from capacity retention, which indicates how much capacity remains relative to the initial discharge capacity.

The main text of the paper and the relevant figure captions also do not state the number of cells tested independently under each condition or the variation in measurements between cells. Reproduction of the results by other research groups has not been confirmed, and the study did not evaluate lifetime or manufacturing cost as a commercial battery.

The major remaining challenge is degradation at the foil surface. When the team examined cross-sections after 5 and 20 cycles, even foil that retained a dense internal structure showed pulverization progressing from the surface and a buildup of films produced by reactions with the electrolyte. Even if adding manganese makes lithium easier to extract from the interior of the foil, surface degradation where the foil meets the electrolyte must be suppressed for it to withstand long-term charging and discharging. The team plans to explore improvements to the alloy composition as well as the introduction of a protective coating on the surface.