# FTIR spectroscopy: how it works, how to read an infrared spectrum and what it tells you about materials

> FTIR spectroscopy explained: infrared absorption by bond vibrations, the interferometer and Fourier transform, wavenumber regions, ATR and reading bands.

URL: https://www.labcanvas.com/techniques/ftir

Fourier-transform infrared (FTIR) spectroscopy measures how much infrared light a sample absorbs at each wavenumber. Bonds whose dipole moment changes as they vibrate absorb at characteristic wavenumbers (Baker et al. 2014), and an FTIR instrument records all of them at once with a Michelson interferometer, the spectrum being the Fourier transform of the interferogram (Lagos et al. 2018). The result is a fingerprint of the bonds present: S–H, C–H, N–H and O–H stretches between about 2,550 and 3,500 cm⁻¹, and bending and skeletal vibrations in the fingerprint region from about 600 to 1,450 cm⁻¹ (Baker et al. 2014).

Characterisation techniques Published 27 October 2026 Last updated 27 October 2026 7 references

![FTIR: Functional groups from infrared absorption (A = log₁₀(I₀/I))](https://www.labcanvas.com/materials/ftir/ftir-cover.webp)

**Contents**

**Contents**

- Fourier-transform infrared (FTIR) spectroscopy at a glance
- Fourier-transform infrared (FTIR) spectroscopy instrument drawing (high-resolution SVG)
- Fourier-transform infrared (FTIR) spectroscopy data: an example plot to edit
- How FTIR works
- What the data looks like and how to read it
- Sample preparation
- What FTIR tells you about materials
- Common mistakes and limits
- How to draw the FTIR setup and data in LabCanvas
- FTIR compared with Raman and NMR
- Related techniques
- More on LabCanvas
- Frequently asked questions
- References

## Fourier-transform infrared (FTIR) spectroscopy at a glance

| Probe | Broadband infrared light, usually from a globar (silicon carbide) source on benchtop instruments (Baker et al. 2014) |
| --- | --- |
| Instrument | A Michelson interferometer replaces the dispersive analyser; the spectrum is the Fourier transform of the recorded interferogram (Lagos et al. 2018) |
| Signal measured | Absorbance: the base-10 logarithm of the ratio of the background and sample single-beam spectra, related to concentration by Beer’s law (Baker et al. 2014) |
| Selection rule | Vibrations that change the bond’s electric dipole moment are IR active (Baker et al. 2014) |
| Main regions | Fingerprint 600–1,450 cm⁻¹; amide I/II 1,500–1,700 cm⁻¹; 2,550–3,500 cm⁻¹ for S–H, C–H, N–H and O–H stretches (Baker et al. 2014) |
| Sampling modes | Transmission, transflection and attenuated total reflection (ATR) (Baker et al. 2014) |
| ATR depth | Evanescent wave penetrates ~1–2 μm in the 1,800–900 cm⁻¹ region, with ~5% intensity still at 3 μm (Baker et al. 2014) |
| ATR crystals | High-refractive-index internal reflection elements: diamond, ZnSe, Ge or Si (Baker et al. 2014) |
| Spatial resolution | Diffraction-limited; a globar gives 20–100 μm apertures at the sample, a synchrotron (100–1,000× brighter) 10–20 μm (Baker et al. 2014) |
| Signal-to-noise | Rises with the square root of the number of co-added scans; MCT detectors usually beat DTGS detectors (Baker et al. 2014) |
| Interferences | Water vapour (1,350–1,950 and 3,600–3,900 cm⁻¹) and CO₂; purge with dry air or nitrogen (Baker et al. 2014) |

## Fourier-transform infrared (FTIR) spectroscopy instrument drawing (high-resolution SVG)

![Fourier-transform infrared (FTIR) spectroscopy instrument drawn as a LabCanvas Library object](https://www.labcanvas.com/materials/ftir/ftir-drawing.webp)

An FTIR spectrometer as a LabCanvas Library instrument: the infrared source, Michelson interferometer, sample and detector. Change colours, labels and view in the figure editor.

[Edit the drawing in LabCanvas](https://www.labcanvas.com/editor.html?add=ch-ftir-path)[Download SVG](https://www.labcanvas.com/materials/ftir/ftir-drawing.svg)[Download PNG (3000 px)](https://www.labcanvas.com/materials/ftir/ftir-drawing.png)

## Fourier-transform infrared (FTIR) spectroscopy data: an example plot to edit

![Fourier-transform infrared (FTIR) spectroscopy data plot drawn as a LabCanvas Library object](https://www.labcanvas.com/materials/ftir/ftir-plot.webp)

An FTIR spectrum (graphene oxide) as an editable LabCanvas plot: transmittance against wavenumber. Edit the data, labels and colours in the figure editor.

[Edit the plot in LabCanvas](https://www.labcanvas.com/editor.html?add=gr-ftir-go)[Download SVG](https://www.labcanvas.com/materials/ftir/ftir-plot.svg)[Download PNG (3000 px)](https://www.labcanvas.com/materials/ftir/ftir-plot.png)

## How FTIR works

Infrared light makes chemical bonds vibrate. A bond whose electric dipole moment changes as the atoms move is **IR active**, and it absorbs light at the wavenumber of its vibration (Baker et al. 2014). Measuring which wavenumbers are absorbed, and how strongly, therefore identifies the bonds in a sample without labels and without changing it (Baker et al. 2014).

A dispersive spectrometer would scan one wavelength at a time. A Fourier-transform spectrometer instead sends the whole infrared beam through a **Michelson interferometer**, whose moving mirror produces an **interferogram**; the spectrum is the **Fourier transform** of that interferogram (Lagos et al. 2018). Weighting the interferogram with an apodization function before the transform smooths the spectrum but can cost signal-to-noise (Baker et al. 2014).

To get absorbance, the instrument records a background single-beam spectrum without the sample, then a single-beam spectrum with it. The base-10 logarithm of their ratio is the absorbance, A = log₁₀(I₀/I), which is proportional to concentration by the Beer–Lambert law (Baker et al. 2014). FTIR is one of several [characterisation techniques](https://www.labcanvas.com/techniques) for chemical bonding; its closest partner is [Raman spectroscopy](https://www.labcanvas.com/techniques/raman).

## What the data looks like and how to read it

An FTIR spectrum plots absorbance (or transmittance) against wavenumber in cm⁻¹. Read it from high to low wavenumber:

- **2,550–3,500 cm⁻¹: stretching vibrations** of S–H, C–H, N–H and O–H bonds (Baker et al. 2014). Atmospheric water vapour also absorbs between 3,600 and 3,900 cm⁻¹ (Baker et al. 2014).
- **1,500–1,700 cm⁻¹:** in proteins and other amides, the amide I and amide II bands (Baker et al. 2014). Water vapour also absorbs across 1,350–1,950 cm⁻¹, so check for it before assigning bands here (Baker et al. 2014).
- **600–1,450 cm⁻¹: the fingerprint region**, dominated by bending and carbon-skeleton vibrations. Together, these regions make up the spectral fingerprint of the sample (Baker et al. 2014).

Practical processing steps are a baseline correction (for example a rubber-band baseline), normalisation to a reference band, and, to separate overlapping bands, a first or second derivative, which narrows bands but amplifies noise (Baker et al. 2014). Compare spectra before and after a treatment, or against a reference compound, rather than reading one spectrum in isolation.

## Sample preparation

- **ATR (attenuated total reflection).** Press the sample onto the internal reflection element (diamond, ZnSe, Ge or Si). The evanescent wave probes only ~1–2 μm, so the sample must be in direct contact with the crystal and ideally three to four times thicker than the penetration depth; there is no maximum thickness (Baker et al. 2014). ATR needs little sample preparation, and liquids can be placed directly on the crystal (Baker et al. 2014).
- **Transmission.** The beam passes through the sample on an IR-transparent window such as BaF₂ or CaF₂; ordinary glass absorbs and has its own mid-IR signature (Baker et al. 2014). If the sample is too thick, the response becomes nonlinear and the Beer–Lambert law no longer applies (Baker et al. 2014).
- **Windows and water.** KBr and NaCl are water-soluble, so they are unsuitable for wet samples (Baker et al. 2014).
- **Background and purging.** Purge the instrument with dry air or nitrogen to remove water vapour and CO₂, and record a background before every sample or whenever the atmosphere changes (Baker et al. 2014).
- **Signal-to-noise.** Co-add scans: the signal-to-noise ratio grows with the square root of the number of co-additions (Baker et al. 2014).

## What FTIR tells you about materials

**Functional groups and surface ligands.** FTIR shows which bonds are present and how they change when a material is made, activated, loaded or reacted. In the metal–organic framework Hf-UiO-66, infrared spectra showed formate bound to two types of sites on the metal oxide cluster nodes, defect sites and μ-OH sites, and followed how water switches the formate between bidentate and monodentate bonding (Xiao et al. 2024). In [UiO-66](https://www.labcanvas.com/mof/uio-66), vibrational spectroscopy with solid-state NMR showed that CO₂ is chemisorbed as bicarbonate on the Zr-cluster hydroxyl groups as well as physisorbed (Nadol et al. 2026). The same before-and-after comparison can be used to follow linker functionalisation in other MOFs such as [ZIF-8](https://www.labcanvas.com/mof/zif-8) or [HKUST-1](https://www.labcanvas.com/mof/hkust-1), and to oxygen groups on [graphene oxide](https://www.labcanvas.com/2d-materials/graphene-oxide), where a reference spectrum of the parent material makes new bands easy to spot.

**Complement to Raman.** In materials with a centre of inversion, a vibration is either IR active or Raman active, not both (the mutual exclusion rule), so FTIR and Raman see different modes; without inversion symmetry a mode can show up in both (Jones et al. 2019). Water absorbs strongly in the infrared, which is a reason to use [Raman](https://www.labcanvas.com/techniques/raman) for aqueous samples (Jones et al. 2019).

**Mapping.** With a microscope and a focal-plane-array detector, FTIR records thousands of spectra at once and builds chemical images, at a diffraction-limited spatial resolution (Baker et al. 2014).

For which atoms are bonded to which, use [NMR](https://www.labcanvas.com/techniques/nmr); for surface composition and oxidation states, [XPS](https://www.labcanvas.com/techniques/xps); for crystal phase, [XRD](https://www.labcanvas.com/techniques/xrd).

## Common mistakes and limits

- **Water vapour and CO₂ bands mistaken for sample bands.** Purge, and take a fresh background; computational subtraction of water vapour works only partly (Baker et al. 2014).
- **Stale background.** A background should be taken before every sample, and again if the room atmosphere changes, for example when a door is opened (Baker et al. 2014).
- **Poor ATR contact or too-thin samples.** The sample must cover the crystal and be several penetration depths thick, or the spectrum is weak and the substrate shows through (Baker et al. 2014).
- **Too-thick transmission samples.** Saturated bands break the Beer–Lambert law, which ruins quantitative comparisons (Baker et al. 2014).
- **Comparing ATR with transmission spectra directly.** Recorded intensities depend on the sampling mode, sample morphology and substrate, not only on chemistry (Baker et al. 2014).
- **Spatial resolution.** FTIR imaging is diffraction-limited, so features much smaller than the infrared wavelength are not resolved (Baker et al. 2014).

## How to draw the FTIR setup and data in LabCanvas

The LabCanvas figure editor Library has the FTIR spectrometer as an editable object, including a cutaway of the light path from the source through the Michelson interferometer (beamsplitter, fixed and moving mirrors) to the sample or ATR crystal and the detector, and an editable data plot for the spectrum. Add your spectrum with a wavenumber axis that runs from high to low, label the bands with their assignments (O–H, C=O, fingerprint region), stack before-and-after spectra, and export a high-resolution image for a methods figure or graphical abstract.

## FTIR compared with Raman and NMR

| Technique | What it measures | Typical sample | Depth / resolution |
| --- | --- | --- | --- |
| FTIR | Infrared absorption by vibrations that change the dipole moment (Baker et al. 2014) | Films, sections and liquids; ATR needs contact with the crystal (Baker et al. 2014) | ATR ~1–2 μm deep; diffraction-limited imaging (Baker et al. 2014) |
| Raman | Inelastic scattering of laser light; vibrations that change the polarisability (Jones et al. 2019) | Solids and liquids, including water-rich samples (Jones et al. 2019) | Laser spot at least half the wavelength, ~250 nm with visible light (Jones et al. 2019) |
| NMR | Nuclear spin resonances: chemical shift, coupling, integrals (Pauli et al. 2014) | Solutions, or powders under magic-angle spinning (Brunner & Rauche 2020) | Bulk average; atom-level local structure (Brunner & Rauche 2020) |

## Related techniques

[![Raman: Vibrations from inelastically scattered laser light (Raman shift Δν̃ = ν̃₀ − ν̃ₛ (cm⁻¹))](https://www.labcanvas.com/materials/raman/raman-cover.webp) Technique Raman shift Δν̃ = ν̃₀ − ν̃ₛ (cm⁻¹) Raman spectroscopy Raman spectroscopy explained: inelastic laser scattering, Raman shift, Stokes lines, IR complementarity, graphene D, G and 2D bands, lasers, fluorescence.](https://www.labcanvas.com/techniques/raman)

[![NMR: Atomic-level structure from nuclear spins (Chemical shift δ (ppm))](https://www.labcanvas.com/materials/nmr/nmr-cover.webp) Technique Chemical shift δ (ppm) Nuclear magnetic resonance (NMR) spectroscopy NMR spectroscopy explained: nuclear spins in a magnetic field, chemical shift in ppm, ¹H and ¹³C, coupling, integration, solid-state MAS and MOF digestion NMR.](https://www.labcanvas.com/techniques/nmr)

[![XPS: Surface elements and chemical states, top ~10 nm (BE = hν − KE)](https://www.labcanvas.com/materials/xps/xps-cover.webp) Technique BE = hν − KE X-ray photoelectron spectroscopy (XPS) XPS explained: binding energy, 2–10 nm surface sensitivity, survey vs high-resolution spectra, chemical shifts, C 1s charge referencing and depth profiling.](https://www.labcanvas.com/techniques/xps)

[![UV–vis: Absorbance, concentration and optical band gap (A = εcl)](https://www.labcanvas.com/materials/uv-vis/uv-vis-cover.webp) Technique A = εcl UV–vis spectroscopy UV–vis spectroscopy explained: absorbance and the Beer–Lambert law, band gap from a Tauc plot, diffuse reflectance for powders and gold plasmon bands.](https://www.labcanvas.com/techniques/uv-vis)

[![TGA / DSC: Thermal stability, composition and phase transitions (Mass and heat flow vs temperature)](https://www.labcanvas.com/materials/tga-dsc/tga-dsc-cover.webp) Technique Mass and heat flow vs temperature TGA and DSC (thermal analysis) TGA and DSC explained: mass loss and DTG vs temperature, atmosphere, MOF thermal stability, and glass transition, melting and crystallisation in DSC.](https://www.labcanvas.com/techniques/tga-dsc)

## More on LabCanvas

- [All characterisation techniques](https://www.labcanvas.com/techniques): XRD, XPS, BET, SEM, TEM, NMR, FTIR, Raman and more, explained with cited sources
- [Graphical abstract sizes and templates](https://www.labcanvas.com/graphical-abstracts): ACS, Wiley, Angewandte, RSC, Elsevier and Cell Press, with templates at the right size
- [Materials explained](https://www.labcanvas.com/blog/materials): MOFs, COFs, 2D materials, perovskites, batteries and zeolites, with cited facts
- [Journal figure size guide](https://www.labcanvas.com/blog/journal-figure-size-guide): Column widths and resolutions for 38 journals

## Frequently asked questions

### What is FTIR used for?

FTIR identifies functional groups and chemical bonds by the infrared wavenumbers they absorb, and follows how they change when a material is made, modified or reacted. It is label-free and does not alter the sample (Baker et al. 2014).

### How does an FTIR spectrometer work?

It sends broadband infrared light through a Michelson interferometer and records an interferogram; the spectrum is the Fourier transform of that interferogram (Lagos et al. 2018). Absorbance is then the log₁₀ ratio of a background spectrum to the sample spectrum (Baker et al. 2014).

### How do you read an FTIR spectrum?

Work from high to low wavenumber: 2,550–3,500 cm⁻¹ holds the S–H, C–H, N–H and O–H stretches, 1,500–1,700 cm⁻¹ the amide I/II bands of amides, and 600–1,450 cm⁻¹ the fingerprint region of bending and skeletal vibrations (Baker et al. 2014).

### What is the fingerprint region in IR spectroscopy?

The region from about 600 to 1,450 cm⁻¹, dominated by bending and carbon-skeleton vibrations. Together with the other regions it forms the spectral fingerprint that is used to tell samples apart (Baker et al. 2014).

### What is ATR-FTIR?

Attenuated total reflection FTIR shines infrared light through a high-refractive-index crystal (diamond, ZnSe, Ge or Si); an evanescent wave reaches about 1–2 μm into a sample pressed against it. It needs little sample preparation (Baker et al. 2014).

### FTIR vs Raman: what is the difference?

FTIR measures infrared absorption by vibrations that change the dipole moment, Raman measures inelastic scattering by vibrations that change the polarisability. In centrosymmetric materials a mode is active in one or the other, and water absorbs strongly in the IR but is much less of a problem for Raman (Baker et al. 2014; Jones et al. 2019).

### Why do I see sharp noisy peaks around 1,600 and 3,700 cm⁻¹?

These are often water vapour lines, which fall between 1,350–1,950 and 3,600–3,900 cm⁻¹. Purge the instrument with dry air or nitrogen and record a fresh background (Baker et al. 2014).

### Why use CaF₂ or BaF₂ windows instead of glass?

Glass absorbs mid-infrared light and has its own spectral signature, while CaF₂ and BaF₂ are IR-transparent; KBr and NaCl are also transparent but dissolve in water (Baker et al. 2014).

## References

- M. J. Baker et al.. Using Fourier transform IR spectroscopy to analyze biological materials. *Nat. Protoc.* 2014, 9, 1771–1791. [doi:10.1038/nprot.2014.110](https://doi.org/10.1038/nprot.2014.110)
- M. Lagos, R. Paredes, C. Retamal. Fourier transform spectrometry with Fourier analysis of the interferogram as just an optional tool. *ACS Omega* 2018, 3, 18258–18262. [doi:10.1021/acsomega.8b02914](https://doi.org/10.1021/acsomega.8b02914)
- R. R. Jones, D. C. Hooper, L. Zhang, D. Wolverson, V. K. Valev. Raman techniques: fundamentals and frontiers. *Nanoscale Res. Lett.* 2019, 14, 231. [doi:10.1186/s11671-019-3039-2](https://doi.org/10.1186/s11671-019-3039-2)
- Y. Xiao, B. C. Gates, D. Yang. Chemistry of formate and water ligands on metal oxide cluster nodes of metal–organic framework hcp Hf-UiO-66: keys to understanding reactivity of paired μ-OH and defect sites. *ACS Appl. Mater. Interfaces* 2024, 16, 52445–52454. [doi:10.1021/acsami.4c11541](https://doi.org/10.1021/acsami.4c11541)
- A. Nadol et al.. Insights into the local adsorption of CO₂ in UiO-66. *Solid State Nucl. Magn. Reson.* 2026, 144, 102095. [doi:10.1016/j.ssnmr.2026.102095](https://doi.org/10.1016/j.ssnmr.2026.102095)
- G. F. Pauli et al.. Importance of purity evaluation and the potential of quantitative ¹H NMR as a purity assay. *J. Med. Chem.* 2014, 57, 9220–9231. [doi:10.1021/jm500734a](https://doi.org/10.1021/jm500734a)
- E. Brunner, M. Rauche. Solid-state NMR spectroscopy: an advancing tool to analyse the structure and properties of metal–organic frameworks. *Chem. Sci.* 2020, 11, 4297–4304. [doi:10.1039/d0sc00735h](https://doi.org/10.1039/d0sc00735h)

## Draw Fourier-transform infrared (FTIR) spectroscopy figures for your paper

LabCanvas is a browser workspace for scientific figures: instruments with their beam and light paths, editable data plots, crystal structures, graphs and graphical abstracts at journal sizes.

[Draw Fourier-transform infrared (FTIR) spectroscopy in LabCanvas](https://www.labcanvas.com/editor.html?add=ch-ftir-path)

**Contents**

- Fourier-transform infrared (FTIR) spectroscopy at a glance
- Fourier-transform infrared (FTIR) spectroscopy instrument drawing (high-resolution SVG)
- Fourier-transform infrared (FTIR) spectroscopy data: an example plot to edit
- How FTIR works
- What the data looks like and how to read it
- Sample preparation
- What FTIR tells you about materials
- Common mistakes and limits
- How to draw the FTIR setup and data in LabCanvas
- FTIR compared with Raman and NMR
- Related techniques
- More on LabCanvas
- Frequently asked questions
- References