From this paper (discussion):
I dare to expect LFP to become dominant (~90% market share) in the li-ion segment also due high recyclability. The LFP particles are extremely durable which allow use of them almost as-is after recycling. The issue is iron phosphate is already abundant and available in millions of tons annually. So recycling value has to be specifically built into the value chain.
Very large electrode cells are so much harder to achieve with LCO/LMO/NMC/NCA that it would most probably make use of them too dangerous. Large electrodes allow also combining the cell case and current collector which allow almost doubling the Wh/l of LFP cells compared to Blade type cells.
One of the key patents for LFP expired in September 2021 in Europe:
Tesla announced it’s going to shift to LFP chemistry in its Megapack ESS in May 2021.
In October 2021, Tesla announced that it will only use iron-based batteries for standard model EVs, globally.
Tesla announced it’s going to shift to LFP chemistry in its Megapack ESS in May.
LFP is safer than NMC:
LFP degrades slower than NMC: typical figures appearing in the literature are 2500-3000 cycles for LFP, 1000-2000 cycles for NMC. @Marcus Ulmefors said could be due to slower electrolyte decomposition. TOFO other influencing factors.
LFP production cost is within 10% lower per kWh of capacity than NMC (source, also pictures above).
The only (?) factor in which LFP is inferior to NMC is energy density, but this is unimportant for energy storage.
However, also @Marcus Ulmefors:
Classic LFP has worse conductivity so needs smaller particle sizes to further increase surface area.
I wonder if this means that LFP can deliver lower peak (a.k.a. burst, pulse) power than NMC (TOFO) and whether this is important for grid balancing energy storage applications (?), cc @Lukas Keller.
Peak power in Grid Applications (Lukas):Roughly, balancing service providers are paid for the power they deliver times the number of hours they can deliver that. This means you can theoretically double your revenue either by doubling your peak power or by doubling your storage capacity (and then bidding for two consecutive hours). The latter is clearly cheaper as your power equipment can remain sized at the same level, you just add more Cores to the bus bar to up the energy capacity and bid for more hours. Same applies for energy arbitrage (exploiting day ahead hourly price differences) where your time horizons are also 6-12 hours.TL;DR: Grid applications generally prefer low power-to-energy ratios
Effect:Yes, in my understanding LFP can deliver lower peak power per cell, but that doesn’t matter in practice for grid applications. Its a fully additive system. You just don’t need more power in relation to energy.
Why do LFP cells have lower peak power? (Lukas' best guess)They can be made thicker relative to to NMC etc, which reduces the cell overhead vs active material. A side effect is that the internal resistance is higher because the distances are longer in the thicker cell and the cell has as a result a lower power to energy ratio. So worse P-to-E ratio isn’t a bug, its a feature or a more energy-centric cell. (Just a hunch, haven’t check how blade cells fit in here)
However, if we look at the picture above (and believe it), LFP blade batteries are better than NMC in both energy density and peak power.