Cathode Progress: Small Additions of p-Group Elements Improve Low-Cost Iron-Based Materials

By Kyle Proffitt

May 30, 2024 | A group from Hokkaido University, Tohoku University, and Nagoya Institute of Technology has reported a step improvement in the attributes of low-cost, iron-based cathode materials for use in lithium-ion batteries. The work, led by Professor Hiroaki Kobayashi, is published in ACS Materials Letters (DOI: 10.1021/acsmaterialslett.4c00268). “We developed a cost-effective, high-capacity, and cyclable cathode material for Li-ion batteries,” Kobayashi said.

Despite the push for and continuing advances toward ever-higher energy densities using silicon and lithium metal anodes, solid-state or anodeless designs, and other novel approaches, cost remains a major determinant of the commercial feasibility of battery technology. This explains why LiFePO4 (LFP)-based batteries account for about 30% of market share, despite having a much lower energy density (173 mAh/g) compared to lithium-nickel-manganese-cobalt (NMC) cathode-based materials (around 300 mAh/g).

Iron is plentiful and cheap and does not suffer from the same geopolitical and human rights issues as those surrounding cobalt. In China, the electric vehicle industry is dominated by LFP, with BYD Auto Co. alone accounting for up to 50% of total use, and the standard range versions of many vehicles in the US, such as the Tesla Model 3, using LFP chemistry instead of NMC or nickel cobalt aluminum oxide (NCA). LFP is also the dominant player in grid storage and other areas where size and weight are less limiting.

Of course, improving the energy density with LFP or other low-cost, high-availability materials would marry the best of both options. Kobayashi’s group is making strides toward that outcome using slightly different materials. In 2023, his group reported promising results with a similar but more lithium-rich material, Li5FeO4 (LFO). Unlike LFP, which stores and releases energy by the movement of lithium ions into and out of its crystal lattice structure (intercalation and de-intercalation), the iron and oxygen in LFO both change oxidation state and participate in redox chemistry. During charging, Li5FeO4  is converted to Li2FeO3, Li2O, a lithium ion (now free to move to the anode), and an electron. This is a change of iron from +3 to +4 oxidation state. The Li2O can then undergo further oxidation, forming Li2O2 and releasing another lithium ion and electron. The oxygen in this step goes from -2 oxidation state to -1. The net effect is 2 lithium ions for each molecule of LFO, unlike one lithium ion per LFP.

Structural Shake-Up

The primary challenge with LFO is that if conditions aren’t carefully controlled, molecular oxygen (O2) is released, at which point it can no longer participate in redox chemistry, and battery performance plummets. The cause of this has to do with some structural transitions at the crystal structure level. LFO is normally found in a “defect antifluorite structure with orthorhombic symmetry”: the oxygen atoms form the vertices and the centers of faces of a somewhat imperfect repeating cube. The oxygen is then coordinated with lithium and iron, which occupy sites within this cube. During charging, the structure is converted to a cubic disordered rocksalt, which has a more random arrangement of cations and anions but a more perfect cubic symmetry.

Kobayashi’s group inferred that the transition from the orthorhombic, imperfect cubic symmetry to this cubic rocksalt was responsible for overvoltage—effectively having to put in more voltage than should be necessary to drive the reaction—and this overvoltage is then responsible for the release of molecular oxygen. The solution involved a mechanical treatment, crushing and grinding the LFO powder with small inert balls, shaking up the structure at the molecular level, which actually leads to a more true cubic antifluorite arrangement. In this case, less energy goes into transforming the underlying structure. Kobayashi’s group showed that this treatment enabled creation of cathode material with a capacity of 346 mAh/g, and the material worked well for about 20 cycles without releasing oxygen gas. They further improved the cubic LFO by mixing it with Li6CoO4. The cobalt helped suppress oxygen evolution, and a cell could now undergo 100 cycles while maintaining 300 mAh/g capacity.

More Abundant, Eco-Friendly Dopants

In this new update, Kobayashi’s group made significant further improvements. Reached for comment by email, he explained the impetus for this continuing research, saying “Cobalt is a toxic and expensive element, therefore we would like to avoid using it.” The report explains that the core issue with oxygen release is that the potential for O2- (in Li2O) oxidation to O22- (the species in Li2O2) is too similar to the oxidation potential for O2 gas evolution. Therefore, reversible, solid-state oxygen redox chemistry (the goal) is likely to occur at the same time as irreversible oxygen gas release.

Kobayashi’s group “considered how to stabilize the redox and arrived at the use of p-block elements.” They substituted aluminum, silicon, germanium, phosphorus, and sulfur—all of which form stronger covalent bonds with oxygen than iron does. His group showed that small additions of these elements fit nicely into the cubic antifluorite structure, and by substituting them at just 10%, they drastically changed battery cyclability and oxygen gas evolution.

The group formed 2032 coin-cell-type batteries with each of the new materials and tested them first for 10 cycles. Of note, “the tested cells used Li metal anodes, but the LFO cathode is compatible with graphite anodes, same as commercial lithium-ion batteries,” Kobayashi said. Cubic LFO lacking any dopant starts with an energy density of about 950 Wh/kg, but this reduces to about 650 Wh/kg by the tenth cycle. In contrast, phosphorus- or germanium-doped material maintained nearly 800 Wh/kg at the tenth cycle.

For context, Kobayashi said, “the same calculation for NMC is approximately 600 Wh/kg. Therefore, our materials have higher potential.” All of the dopants improved capacity retention, but phosphorus and germanium showed the highest densities after cycling. “Considering the performance and cost-effectiveness, we selected P as a promising element,” he said.

A little more play with the phosphorus ratio found that replacing 20% of the iron with phosphorus was optimal. The group showed that this blend retained the greatest energy density, 650 Wh/kg, at the 50th cycle, and Kobayashi said that they have actually cycled it 100 times successfully.

Room For Improvement

Kobayashi has plans for continued improvement. His goal, he says, is achieving materials with 99% capacity retention. Some of the main challenges for scaling the technology and creating commercially viable cells have to do with the iron-based cathode materials’ conductivity. “LFO and LFP have less electric conductivity compared to NMC; therefore, effective carbon coating is necessary,” he said. His group is therefore working toward improving this conductivity.

Some of the other dopants also merit more experimentation. Silicon, for instance, did not perform as well as phosphorus, but interestingly, it showed the best capacity retention of tested materials, about 85%, falling from 700 to 600 Wh/kg between the first and tenth cycle. Because it is the second most abundant element in the earth’s crust and retains energy capacity comparable to NMC in these experiments, the economics may favor a solution with silicon. Another abundant element, aluminum, took a bigger capacity hit between the first and tenth cycle, but also retained about 600 Wh/kg. Kobayashi believes that materials using silicon and aluminum can undergo further improvement to maximize this capacity/cost balance and drive toward that 99% capacity retention. “Improvement can be achieved by technical modification and an optimal blend of dopants,” he said.