FTIR or Raman: Which Technique Fits Your Sample?

Fourier-transform infrared spectroscopy (FTIR) and Raman spectroscopy both probe the vibrations of molecules and crystal lattices, so both produce spectra whose peaks correspond to bonds and structural units in the material. They reach that result by different routes. FTIR measures absorption: infrared light passes through the sample or reflects from it, and the instrument records which wavenumbers are absorbed. A vibration absorbs only if it changes the dipole moment of the molecule, which is why bonds between atoms of different electronegativity, such as O–H, C=O, N–H and C–O, give strong bands.

Raman spectroscopy measures inelastic scattering. A monochromatic laser illuminates the sample, and a very small fraction of the scattered photons has exchanged energy with a vibration, so its energy is slightly shifted. That shift, expressed in cm⁻¹, is the Raman shift. A vibration is Raman active only if it changes the polarizability, the ease with which the electron cloud is distorted. Bonds between similar atoms, such as C=C, S–S and C–C, and symmetric stretching modes tend to scatter strongly.

The two selection rules are complementary rather than interchangeable. In molecules with a centre of symmetry, the mutual exclusion rule applies: a vibration that is Raman active is infrared inactive, and vice versa. Most real samples are not that symmetric, so many bands appear in both spectra, but with very different intensities. In practice this means that neither technique sees everything, and a band that is weak in one may be prominent in the other.

FTIR has three common sampling modes. ATR (attenuated total reflectance) presses the sample against a crystal, often diamond, and records the evanescent wave that penetrates from a fraction of a micrometre to a few micrometres into the surface; it needs almost no preparation and suits films, pellets, powders, pastes and many liquids. Transmission sends the beam through the sample, so the sample must be thin enough or be thinned, for example as a KBr pellet, a cast film or a liquid cell. Reflectance modes, specular or diffuse, are used for coatings on reflective substrates, rough powders, or large objects where contact is undesirable.

Raman is usually performed as micro-Raman: the laser is focused through a microscope objective to a spot on the order of a micrometre, and the same objective collects the scattered light. Little or no preparation is needed, and measurements can often be made through a glass slide or a transparent container. A motorised stage lets you map composition across a surface or locate a single particle or inclusion. The choice of laser wavelength and power is part of the method and, as the next sections show, directly affects the result.

FTIR is the most dependable tool for polar functional groups such as O–H, N–H and C=O. Polymers, organic compounds, adhesives, oils, additives and organic coatings are its natural territory. Identifying a polymer type, characterising a surface contaminant, or seeing signs of oxidation such as a growing carbonyl band during ageing are routine FTIR tasks. Measurements are fast, the method is robust, and extensive reference libraries exist.

Raman stands out for non-polar or symmetric bonds such as C=C, C≡C and S–S. In carbon materials, the graphitic G band (near 1580 cm⁻¹), the defect-related D band (near 1350 cm⁻¹) and their ratio say a great deal about structural disorder. Lattice modes of inorganic materials such as oxides, sulfides and carbonates appear at low wavenumbers and are well suited to Raman; telling the anatase and rutile phases of TiO₂ apart is a classic example. Because water is a weak Raman scatterer, aqueous solutions and hydrated samples can be measured. Through a microscope the spatial resolution approaches one micrometre, whereas in an FTIR microscope the long infrared wavelength limits resolution to a few micrometres at best.

The best-known problem of Raman is fluorescence. The Raman signal is a tiny fraction of the excitation photons, so if the sample, or an impurity, dye or aged additive in it, fluoresces, that broad background can completely bury the weak Raman peaks. A longer-wavelength laser often reduces fluorescence, but it affects signal strength and detection conditions. The second issue is laser heating: strongly absorbing samples such as carbon black, some pigments and dark minerals can burn or change phase even at modest power, and the spectrum then reflects the damaged spot rather than the original material. For dark or heat-sensitive samples the power is therefore kept low and the spectrum is checked for changes during the measurement.

The best-known difficulty of FTIR is the strong infrared absorption of water; the O–H stretch near 3400 cm⁻¹ and the bending band near 1640 cm⁻¹ mask sample bands in the same regions. Aqueous samples are measured by ATR and the water contribution is usually subtracted, or the sample is dried. ATR demands good contact, and hard, rough surfaces give weak spectra. Because the penetration depth grows with wavelength, low-wavenumber bands look stronger in an ATR spectrum than in a transmission spectrum, and a layer thinner than the penetration depth is measured together with the substrate beneath it. Depending on the crystal and detector there is also a low-wavenumber cutoff, and many inorganic lattice modes lie in that region.

For ATR-FTIR it is enough that the sample is clean, dry and able to make flat contact with the crystal; powders are pressed down and a drop of liquid is sufficient. Solvent residue and moisture appear as bands, so drying matters. If a KBr pellet is to be made for transmission, the sample must be finely ground and dry. For Raman, preparation is often unnecessary, although a flat, clean surface makes focusing easier. Fluorescent contaminants such as fingerprints, tape or label adhesive should be avoided. It is useful to say beforehand whether the sample is dark, heat-sensitive or hazardous.

A spectrum is a plot of intensity against wavenumber (cm⁻¹). The range of roughly 4000 to 1500 cm⁻¹ is the functional group region: O–H, N–H and C–H stretches appear roughly between 3600 and 2800 cm⁻¹ and the C=O stretch mostly between 1800 and 1650 cm⁻¹, telling you which groups are present. Below about 1500 cm⁻¹ lies the fingerprint region, where many bands overlap and the resulting pattern is characteristic of the compound as a whole. The same logic applies to Raman, where the low-wavenumber range also contains lattice modes.

Identification is usually done by comparing the measured spectrum with reference library spectra. That works well for pure, known compounds and common polymers. It has limits: in mixtures the spectra overlap, fillers and additives bring extra bands, crystallinity and orientation change band intensities, and ATR and transmission spectra of the same material differ in appearance. A high match score is therefore a pointer, not proof. Unknown composites need interpretation and sometimes a second technique.

For a polymer film, start with ATR-FTIR: polymer type, additives and oxidation show up this way. Raman can complement it for backbone C=C bonds, crystallinity, or when the film is very thin. For graphene oxide, Raman gives information on disorder through the D and G bands, while FTIR shows the oxygen-containing groups such as O–H, C=O and C–O; the laser power is kept low on this dark material. For pharmaceutical powders both techniques help: Raman is strong for the active ingredient, for polymorphs, and for measuring through transparent packaging, while ATR-FTIR gives fast identification and hydrogen-bonding information; fluorescent excipients can make Raman difficult.

For a thin coating the choice depends on thickness and substrate. Organic coatings on metal suit FTIR in reflectance or ATR mode; for inorganic coatings or a cross-section examination, micro-Raman is more convenient. If the coating is thinner than the ATR penetration depth, the substrate contributes to the spectrum. For aqueous solutions Raman is the clear choice, and FTIR is possible only with ATR and water subtraction. For minerals, micro-Raman is the first choice for identifying individual grains and inclusions, while FTIR (KBr or ATR) is also very informative for hydroxyl, carbonate and clay minerals.

If you are unsure, you do not need to settle the choice in advance. NAKTEM offers FTIR and Raman spectroscopy analysis; our team helps select the technique and agrees the sample preparation with you. In your request, state what the material is, what form it is in (film, powder, solution) and the question you want answered. One caveat: vibrational spectra show bonds and functional groups, not elemental composition directly, so if your real question is about elements, another technique may be needed.