By Mi-Young Park, CEO and Co-Founder
In early 2024 we ran accelerating rate calorimetry on LFP and NMC cells laminated into our standard 2-mm CFC layup. LFP reached self-heating onset at 271 degrees Celsius, while NMC reached onset at 196 degrees. That 75-degree separation is not a marginal benefit in a structural application. It separates a cell that may tolerate a localized thermal event inside a composite member from one that may not. In a structural battery floor panel, thermal runaway damages more than the pack. It can compromise the vehicle's load-bearing structure.
The Energy Density Tradeoff
Energy density is the usual starting point for comparing LFP with NMC, and that is valid within its limits. At the material level, LFP cathode chemistry provides 160 to 185 mAh per gram, translating to roughly 170 to 200 Wh per kilogram at the cell level in pouch or prismatic formats. NMC 811 provides 200 to 220 mAh per gram and roughly 260 to 300 Wh per kilogram at the cell level. When every gram of cell mass in a conventional pack contributes only energy capacity, that difference is significant.
In a structural cell, the cell also carries load. The relevant measure is energy per kilogram of the integrated structure, not cell alone. An NMC pack keeps its energy advantage but still needs a housing, structural frame, and thermal management hardware, all parasitic mass. A structural LFP cell in a composite member provides energy capacity while replacing structural mass with no energy value. The system-level gap narrows considerably and can reverse, depending on structural overhead avoided.
Thermal Stability Under Load
Thermal stability carries more weight here than it does in a conventional LFP versus NMC comparison. Those comparisons generally assume that the cell can be accessed and replaced. In a structural battery floor panel, the cell is laminated into the vehicle chassis and cannot be removed without taking apart the vehicle structure. A thermal runaway in that setting is not merely a pack failure. It is a chassis failure.
LFP cathode material is intrinsically more thermally stable than NMC because the iron-oxygen bonds in its olivine structure are stronger than the nickel-oxygen bonds in layered oxide. During abuse, LFP releases oxygen more slowly and at much higher temperatures than NMC. The accelerating rate calorimetry results above come from cells at 100 percent state of charge under adiabatic conditions, representing the worst practical abuse case for a cell already installed in a vehicle. The margin is greater during normal operation, but the direction of the advantage remains the same.
Carbon-fiber composite retains structural integrity up to roughly 150 to 180 degrees Celsius, depending on the resin system. Above that, the matrix softens. An NMC thermal runaway in a CFC laminate can degrade the matrix before containment. An LFP event, if it reaches runaway, is more likely to stay below that threshold. This is a probabilistic argument, not an absolute safety guarantee, but the probability distribution matters for structural design.
Cycle Life for Structural Service
Cycle life is the second major argument for LFP structural cells. In flat-cell tests at 0.5C charge and 1C discharge, LFP cells typically reach 80 percent capacity retention at 3,000 to 4,000 cycles. At the same rates, NMC cells reach 80 percent at 800 to 1,200 cycles. Under equivalent discharge conditions, LFP offers roughly 3x the calendar-cycle life.
For a passenger EV with a structural battery floor panel, a daily transportation profile of 1.5 to 2 full equivalent cycles per day is realistic. At 2 cycles per day, 3,000 LFP cycles correspond to approximately 4.1 years of service. At 4,000 cycles, the figure is 5.5 years. NMC reaches the same 80 percent threshold at 1,000 cycles in roughly 1.4 years at that usage rate. When a structural member cannot be replaced without chassis disassembly, a 3x to 4x cycle life difference is a serviceability requirement, not a minor improvement.
Combined mechanical loading causes both chemistries to degrade faster than flat-cell tests indicate, as covered in the cycle life testing article published earlier. LFP's relative advantage narrows somewhat under combined load because delamination and electrode cracking are structural degradation mechanisms shared by both chemistries. Even so, LFP retains a meaningful cycle life lead in combined-load testing. Its avoided electrochemical degradation adds to the structural degradation both chemistries experience.
The Flat OCV Plateau
LFP has a known SoC estimation disadvantage: its OCV plateau is nearly flat across 20 to 80 percent of the usable range. In conventional packs with prismatic or pouch cells, this makes SoC estimation more dependent on coulomb counting accuracy and less on OCV recalibration, introducing error over long cycle counts. This is a real engineering challenge and why many EV pack designers prefer NMC's more sloped OCV curve, which enables better SoC tracking from terminal voltage alone.
In our structural cells, we address this issue with the strain-conditioned estimator described in a separate technical note. The flat OCV plateau is more difficult in a structural cell because limited OCV information is further affected by the OCV shift caused by bending. We are not minimizing the problem: SoC estimation is harder for structural LFP cells than for structural NMC cells. The point is not to ignore that difficulty but to solve it in software rather than accept NMC's thermal and cycle life penalties in the physical layer. That choice reflects our product direction and may not suit every application.
When NMC Still Fits
NMC still has a place in structural battery applications. On short-duty-cycle platforms with a total lifetime cycle count well below 1,000, its cycle life disadvantage is irrelevant. If the structural member can be accessed and replaced at mid-life, the serviceability constraint supporting LFP does not apply. Where maximum range from a fixed structural volume is the primary constraint and the thermal environment is well controlled, NMC's energy density advantage may justify its tradeoffs.
The LFP case is strongest when three conditions are present: the structural member is permanently bonded into the chassis, projected active use exceeds 5 to 7 years, and a thermal event in the structure would have catastrophic consequences. Long-range EV floor panels, aerospace mobility applications, and high-cycle robotic platforms fit that profile. Outside those conditions, NMC deserves equal consideration before the chemistry is fixed at the structural design stage.