1.6. Physical & Thermal Properties (BET & DSC/TGA)
Surface Area & Porosity
Technique
Gas adsorption measures how much nitrogen or another adsorptive gas is taken up by a powder at different relative pressures. Brunauer–Emmett–Teller (BET) analysis is commonly used to estimate the specific surface area.
Accessible pore volume and pore-size distribution are calculated from the adsorption data using additional models, such as the Barrett–Joyner–Halenda (BJH) method for mesopores or more advanced density-functional-theory (DFT) methods.
Observation:
Specific surface area - is it within the ideal range?
Pore structure - what pores sizes and pore volumes are detected, are they open or blocked, and are they consistent across batches?
Adsorption-desorption curve - does its shape and hysteresis indicate microporous or mesoporous behaviour? (IUPAC classification: micropores: <2 nm; mesopores: 2-50 nm; macropores: >50 nm)
“Surface area and accessible porosity influence electrolyte contact and interfacial reaction kinetics. Higher values can support rate capability, but excessive surface area may increase side reactions and reduce powder compaction density.”
NCM v LFP
NCM: Dense layered-oxide powders commonly have relatively low BET surface areas. Commercial-grade NMC811 examples are reported near 0.5–1.1 m² g⁻¹, although finer or modified powders may reach several m² g⁻¹. Lower surface area generally helps limit electrolyte contact and high-voltage side reactions, but the appropriate value depends on particle architecture and material grade.
LFP: Pristine LiFePO₄ has very low intrinsic electronic conductivity and relatively slow lithium-ion transport through the solid. Commercial LFP is therefore typically engineered using fine or nanoscale primary particles to shorten lithium-diffusion paths, together with a thin carbon coating to improve electronic conductivity. As a result, carbon-coated LFP often has a higher BET surface area than dense layered oxides, commonly around 5–35 m² g⁻¹. Higher surface area can support electrolyte access and rate capability, but excessive surface area may increase side reactions and reduce powder packing density.
Impact on Cell KPIs
Optimal surface area balances rate performance, safety, and longevity.
Worked example: Interpreting LFP Surface Area and Mesopore Structure
Figure 1. Nitrogen adsorption–desorption isotherms and calculated pore-size distributions of different LFP/C powders (see article 1.3, Fig. 1 for info on these powders). In the main graph, the x-axis shows relative nitrogen pressure while the y-axis shows the amount of nitrogen adsorbed by the powder. Nitrogen uptake across a range of pressures is recorded, and the resulting data is analysed mathematically to calculate specific surface area, accessible pore volume and pore-size distribution.
Image Source: Peng et al. 2023
Increasing adsorption at high relative pressure and the separation between the adsorption and desorption branches indicate accessible mesopores and voids formed between stacked particles. The inset graph shows the pore-size distribution calculated using the BJH method. Most of the detected pore volume lies between approximately 2 and 10 nm, with a strong peak near 2.5 nm.
Using BET analysis of the adsorption data, the authors calculated specific surface areas of 34.18, 34.24 and 32.57 m² g⁻¹ for the LFP/C-0, 60, and 100 powders respectively. Their corresponding accessible pore volumes were approximately 0.090, 0.085 and 0.069 cm³ g⁻¹. These exact values come from calculations applied to the underlying test data rather than being read directly from the plotted curves. The authors describe LFP/C-60 as showing a H4-type hysteresis loop associated with mesoporous voids between stacked nanoparticles.
Note: BET surface area and BJH pore-size analysis describe pores accessible to the adsorbate in the pristine powder. They do not directly measure electrode porosity, electrolyte wetting or lithium-ion diffusion, and the pore network may change during slurry processing and calendering.
Key observations
- The three powders have broadly similar specific surface areas, indicating that surface area alone does not explain their different electrochemical performance.
- Most of the nitrogen-accessible pores lie within the mesopore range.
- The hysteresis is consistent with accessible mesoporous voids formed mainly between stacked primary nanoparticles within the secondary particle structure.
- LFP/C-100 has the lowest accessible pore volume, suggesting a denser or more strongly aggregated particle structure.
- LFP/C-60 appears to provide a better balance of accessible porosity, particle structure and carbon coating rather than simply maximising surface area.
What to look for
- Surface area and pore volume that are consistent with the intended particle size, morphology and application.
- A reproducible isotherm shape and pore-size distribution across representative samples and production batches.
- Sufficient accessible porosity to support electrolyte penetration without creating an excessively low-density powder.
- Results generated using consistent degassing conditions, adsorptive gas and calculation models.
- BET data assessed alongside particle-size distribution, carbon content, tap density, compaction density and electrode testing.
Red flags
- Assuming that higher surface area is always better → excessive electrolyte contact may increase side reactions.
- Large batch-to-batch changes → possible variation in particle size, aggregation, carbon coating or synthesis conditions.
- Very low accessible pore volume → possible pore blockage, particle fusion or dense aggregation.
- Very high pore volume → possible poor powder packing and reduced volumetric performance.
- Comparing results produced using different analysis models → calculated pore sizes may not be directly comparable.
- Treating powder BET data as electrode porosity → electrode structure must be measured after coating and calendering.
Why it matters
- Accessible mesopores can improve electrolyte penetration into particle agglomerates and support active-material utilisation.
- An appropriate surface area can support lower interfacial resistance, better rate capability and improved power performance.
- Excessive surface area may accelerate parasitic reactions, resistance growth and capacity fade.
- Highly porous or loosely packed powders can reduce tap density, electrode density and volumetric energy density.
- Blocked pores or dense aggregation may restrict electrolyte access and increase polarisation.
- Consistent surface area and pore structure support reliable electrode processing and batch-to-batch quality.
Thermal Stability
Technique
Differential scanning calorimetry (DSC) measures heat absorbed or released during heating, while thermogravimetric analysis (TGA) measures changes in sample mass. Used together, they help identify oxidation, decomposition, carbon combustion and other thermal reactions.
Observation
Onset and peak temperatures: At what temperature do thermal reactions begin and reach their maximum?
Heat flow: Are the events exothermic or endothermic, and how much heat is released or absorbed?
Mass change: Does the sample gain or lose mass, and over what temperature range?
Test conditions: What atmosphere, heating rate, sample state and state of charge were used?
Peak sharpness alone is not necessarily good or bad. Onset temperature and total heat release are generally more useful screening parameters.
“Cathode thermal stability is an important contributor to overall cell safety, particularly at high states of charge. However, thermal runaway also depends on the anode, electrolyte, separator, cell design and operating condition.”
NCM v LFP
NCM: In highly charged NCM cathodes, structural decomposition and lattice-oxygen release can become significant at elevated temperatures, commonly around 200–250°C, although the exact onset and severity depend on composition, state of charge, surface treatment and test conditions. Higher-nickel compositions generally become unstable at lower temperatures, while manganese helps stabilise the oxygen framework.
LFP: the olivine phosphate structure holds lattice oxygen much more strongly through its P–O bonds, giving it substantially greater material-level thermal stability and much lower oxygen and heat release than NCM. LFP cells can still enter thermal runaway because the electrolyte and other cell components may react at lower temperatures, so cathode stability alone does not determine full-cell safety.
Impact on Cell KPIs
Thermal onset temperature, heat release and mass change provide screening information about material stability and possible safety risks. Earlier or stronger exothermic reactions can increase the risk of heat generation, gas formation and thermal runaway, but representative charged-electrode and full-cell testing is required to quantify cell-level safety.
Worked example: Interpreting LFP Thermal Oxidation and Carbon Content
Figure 2. Simultaneous DSC–TGA analysis of the carbon-coated LFP/C-0 powder. The sample was heated from 25 to 700°C at 10°C min⁻¹ in flowing air. The black TGA curve, read against the left axis, shows changes in sample mass, while the red DSC curve, read against the right axis, shows heat absorbed or released during heating. The upward direction is exothermic.
Below ~300°C, the sample shows a small initial mass decrease. Between ~300 and 350°C, the strong mass increase and accompanying exothermic DSC feature arise mainly from oxidation of LiFePO₄ by oxygen in the air, producing Li₃Fe₂(PO₄)₃ and Fe₂O₃. At ~430°C and above, combustion of the carbon coating causes mass loss, partially offsetting the earlier oxidation-related mass gain.
The authors use the difference between the expected oxidation-related mass gain for pure LFP (4.19%) and the measured net mass gain of the composite (2.44%) to calculate a carbon content of approximately 1.75 wt% for LFP/C-0. Although the annotations use the word “loss,” the TGA curve undergoes an overall net mass gain because oxygen uptake during LFP oxidation is greater than the mass lost through carbon combustion.
Note: This experiment primarily measures the oxidation behaviour and carbon content of pristine LFP/C powder under flowing air. It does not directly represent a charged electrode containing binder and electrolyte or the thermal-runaway behaviour of a complete cell.
Key observations
- The TGA curve shows competing mass changes: oxygen uptake during LFP oxidation and mass loss from carbon combustion.
- The DSC feature confirms that the oxidation region is accompanied by heat release.
- Comparing the expected mass gain of pure LFP with the measured net mass gain allows the carbon content to be estimated.
- The small initial mass loss may reflect adsorbed moisture or other volatile species.
- In this example, DSC–TGA is primarily used to assess powder oxidation behaviour and carbon content, rather than full-cell thermal safety.
What to look for
- Carbon content consistent with the supplier specification and intended material design.
- A repeatable TGA profile across representative samples and production batches.
- Limited and explainable mass loss at lower temperatures.
- Clearly distinguishable thermal events that can be assigned using complementary analysis.
- Comparisons performed using the same atmosphere, heating rate, sample preparation and temperature range.
- Results considered alongside Raman spectroscopy, elemental carbon analysis and electrochemical testing.
Red flags
- Interpreting every mass loss as decomposition → moisture, residual organics and carbon combustion can produce different events.
- Ignoring mass gain → oxidation in air can increase sample mass through oxygen uptake.
- Unexpected low-temperature mass loss → possible moisture, solvents or residual synthesis products.
- Carbon content outside specification → possible variation in coating or synthesis control.
- Large batch-to-batch shifts in onset or peak temperature → possible compositional, surface or processing inconsistency.
- Using pristine-powder DSC–TGA as proof of cell safety → it does not reproduce charged-electrode or full-cell conditions.
Why it matters
- Carbon content influences electronic conductivity, resistance and rate capability.
- Excess carbon adds inactive mass and can reduce electrode density and energy density.
- Insufficient or inconsistent carbon may cause poor particle-to-particle conduction and variable capacity utilisation.
- Moisture or residual species can interfere with electrode processing and contribute to unwanted cell reactions.
- Reproducible thermal profiles support batch consistency and process control.
- Powder DSC–TGA provides useful screening information, but representative charged-electrode and full-cell testing is required to assess thermal-runaway risk.
Edited by Muthu Krishna
