XPS
Surface chemistry and oxidation state, not just elements.
XPS reads the top few nanometers of a surface and reports not just which elements are there but what chemical state they are in, the difference between an oxide and a hydroxide, a carbonate and a fluoride. For batteries, that surface is where the chemistry that decides lifetime happens.
What it measures
X-rays eject core electrons; their kinetic energy obeys KE = hν − BE − φ, so the measured spectrum maps binding energies (BE) that are element- and chemical-state-specific. Three layers of analysis pull information out:
- Survey scan: every element except H and He announces itself with characteristic peaks; this is the elemental inventory of the top ~5-10 nm.
- High-resolution regions: within one element’s peak, chemical shifts of a few eV separate oxidation states and bonding environments. The envelope is deconvolved with Gaussian-Lorentzian components on a Shirley background.
- Quantification: peak areas divided by relative sensitivity factors (RSF) and normalized give atomic percentages; depth profiling extends the composition into a near-surface stack.
How to read the output
Check the charge correction first, spectra from insulating samples shift wholesale, and every assignment downstream depends on the reference (usually adventitious C 1s). Then read the fitted components, not the raw envelope: a C 1s region on a battery electrode routinely hides C-C, C-O, C=O, and carbonate components inside one lump. Trust a fit when component positions sit at literature values, widths are physically reasonable, and the residual is structureless. A fit with five overlapping components of arbitrary width can match anything and prove nothing.
A real use case
Two lots of NMC811 powder behave differently in pouch cells, one gasses noticeably during formation. Survey scans look identical; the difference is in the C 1s and O 1s high-resolution regions, where deconvolution quantifies the residual lithium species on the particle surface: the gassing lot carries roughly twice the Li₂CO₃/LiOH component fraction. That number, a surface chemical-state ratio invisible to XRD or EDS bulk composition, sends the lot back for re-calcination and adds a residual lithium spec to incoming powder QC.
Common mistakes
- Skipping or botching charge correction. A 1-2 eV uncorrected shift silently converts one chemical state into another in every assignment.
- Fitting without constraints. Unconstrained component positions and widths can fit any envelope; constrain to known chemistry and judge the residual.
- Using a linear background where the physics calls for a Shirley, the background choice moves quantified areas by tens of percent.
- Reading XPS atomic % as bulk composition. The sampling depth is nanometers; surface contamination and segregation dominate.
- Ignoring overlaps (e.g., transition-metal Auger lines crossing O 1s or F 1s regions) when assigning small components.
From raw spectrum to a quantified surface inventory
Give Niobia an XPS dataset and it applies charge correction, identifies elements from the survey scan, subtracts Shirley backgrounds, and fits high-resolution envelopes with Gaussian-Lorentzian components, assigning chemical states and quantifying atomic percentages through relative sensitivity factors. Depth-profile series are handled as profiles, not as disconnected spectra. The output is the full visual set: survey plot, fitted high-resolution regions, elemental composition bars, and chemical-state pies, so a lot-to-lot comparison like the carbonate ratio above is a readout, not an afternoon of manual peak fitting.
Frequently asked
How deep does XPS actually see?
Photoelectrons escape from roughly the top 5-10 nm. That makes XPS the right tool for SEI layers, surface residual lithium, and coatings, and the wrong tool for bulk stoichiometry, where EDS or ICP applies.
What makes an XPS peak fit trustworthy?
Component positions anchored to literature chemical shifts, physically reasonable and consistent widths, a correct background model, and a structureless residual. If a fit needs a mystery component at an arbitrary position to close, the fit is the problem.
Why do battery teams care about residual lithium on cathode powder?
Li₂CO₃ and LiOH on the particle surface drive slurry gelation and gas generation during formation. They form during storage and calcination, sit only on the surface, exactly the depth XPS samples, and differ lot to lot even when bulk composition is identical.