Raman
Each peak is a specific molecular vibration mode.
Raman reads bond vibrations from scattered light: each peak is a molecular vibration mode, and for carbon materials the D and G bands form a two-peak fingerprint of structural order that battery teams lean on constantly.
What it measures
A laser illuminates the sample; most light scatters elastically, but a small fraction exchanges energy with molecular vibrations. The Raman shift, Δṽ = 1/λ_laser − 1/λ_scattered, expressed in cm⁻¹, is the vibrational energy, so peak positions identify bonds and structures, complementing FTIR (symmetric and carbon-backbone modes that are weak in the infrared are often strong in Raman).
For battery materials the workhorse readouts are:
- The D band (~1350 cm⁻¹): breathing mode activated by disorder in sp² carbon; it only appears where the graphitic lattice is broken.
- The G band (~1580 cm⁻¹): the in-plane stretch of ordered sp² carbon, present in all graphitic material.
- The I_D/I_G ratio: the standard disorder metric: higher ratio, more defected carbon. For anode graphite, conductive additives, and carbon coatings it is the single most-used Raman number.
How to read the output
Position, width, and ratio, in that order. Peak positions identify what the material is; shifts of a few cm⁻¹ report strain, doping, or lithiation state. Width reports crystallinity: narrow G band, well-ordered graphite; broad merged D/G envelope, amorphous carbon. The I_D/I_G ratio only compares honestly at the same laser wavelength and power, the D band intensity is excitation-dependent, and excess laser power heats or even graphitizes the sample under the beam, manufacturing the change you were looking for. A fluorescence background sloping under everything is common; subtract it before quoting any intensity.
A real use case
An LFP cathode line buys carbon-coated powder from two suppliers, nominally equivalent, but cells built from one supplier’s lots keep landing at the bottom of the rate-capability distribution. Raman on incoming powder separates them in minutes: both show the same phosphate bands, but the slower supplier’s carbon coating runs a higher I_D/I_G with a visibly broader G band, more disordered, less conductive carbon, the resistance sitting exactly where the electrons need to leave the particle. The ratio becomes an incoming spec, and rate capability stops being a lottery over which supplier shipped that month.
Common mistakes
- Comparing I_D/I_G across different laser wavelengths or power settings, the ratio is excitation-dependent and the comparison is meaningless.
- Burning the sample. Carbon spectra change under excessive laser power; if the spectrum evolves during the measurement, the laser is doing chemistry.
- Quoting intensities off an uncorrected fluorescence background.
- Reading one spot as the material. Coatings and blends are heterogeneous; map or sample several locations before declaring a lot good or bad.
- Over-interpreting small shifts without an instrument calibration check, a few cm⁻¹ of spectrometer drift looks exactly like strain.
Spectra compared, bands tracked, with scope stated honestly
Raman support in Niobia today is the comparison workflow: ingest spectra, correct the background, detect peaks and log their positions and widths, compute band intensity ratios like I_D/I_G, and overlay samples or lots so differences are explicit rather than eyeballed, the supplier comparison above is exactly this readout. Full spectral library identification and deconvolution-based quantification are not claims Niobia makes for Raman yet; where a question needs them, it says so and points at the complementary methods it runs in depth, XRD for phase identity, FTIR for functional-group chemistry.
Frequently asked
What does the I_D/I_G ratio actually measure?
The density of disorder in sp² carbon: the D band only activates at defects and edges, the G band reflects ordered graphitic bonds. Higher ratio means more defected carbon, relevant to conductivity of coatings and additives, but it only compares at matched laser wavelength and power.
Raman or FTIR for my sample?
They are complements. Carbon materials, symmetric modes, and aqueous samples favor Raman; polar bonds like C=O and O-H favor FTIR. For battery work: carbons and cathode lattice modes lean Raman, electrolytes and binders lean FTIR.
Why does my spectrum have a huge sloping background?
Fluorescence, common with binders, residues, and some oxides. It is excitation-dependent, so a different laser wavelength often suppresses it; otherwise subtract it explicitly before reading intensities.