How Does FTIR Spectroscopy Work?
At the heart of every FTIR instrument is a Michelson-type interferometer — an optical device that splits and recombines infrared light to create an interference pattern containing information about all wavelengths simultaneously.
The Michelson Interferometer
An FTIR spectrometer built around a Michelson interferometer has four essential parts: an infrared source, the interferometer itself (a beamsplitter plus one fixed and one moving mirror), the sample, and a detector. The beamsplitter sits at 45° to the incoming beam and divides the IR light into two paths. Many commercial instruments use Michelson variants (cube-corner, wishbone or rotating-mirror designs) that work on the same principle.
One beam reflects off the fixed mirror, while the other reflects off a mirror that translates back and forth along a precision track. When the two beams recombine at the beamsplitter, the variable path difference creates constructive and destructive interference — the intensity at the detector rises and falls as the mirror moves.
Each point in the resulting signal contains information about all infrared wavelengths simultaneously. This multiplexed encoding is what makes FTIR fundamentally different from dispersive spectrometers. Once the Fourier transform converts this data into a spectrum, you can begin interpreting the peaks.
Interactive Michelson Interferometer
The moving mirror creates a variable optical path difference, producing an interference pattern at the detector.
From Interferogram to Spectrum: The Fourier Transform
The detector records a plot of intensity versus optical path difference (twice the distance the moving mirror has travelled) — this raw data is called an interferogram. It exists in the “time domain” (or more precisely, the “path difference domain”). To produce a conventional spectrum of intensity versus frequency, the interferogram must be converted using the Fourier Transform.
The concept is intuitive: a single infrared frequency produces a simple cosine wave in the interferogram. Multiple frequencies each contribute their own cosine wave at different periods. The actual interferogram is the sum of all these cosine components. The Fourier Transform reverses this process — it decomposes the complex interferogram back into its individual frequency components, revealing the spectrum.
From Interferogram to Spectrum
Select a signal complexity to see how the Fourier Transform decomposes the interferogram into its frequency components:
Interferogram
Spectrum
A single frequency produces a simple cosine interferogram. Multiple frequencies produce a complex pattern. The Fourier Transform decomposes this back into the individual frequency components (spectrum shown schematically).
B(ν) = ∫−∞∞ I(δ) cos(2πνδ) dδ
Where B(ν) is the spectral intensity at wavenumber ν, I(δ) is the interferogram intensity at optical path difference δ (twice the mirror displacement, in cm when ν is in cm⁻¹). In practice, the integral is computed with a fast Fourier transform (FFT) algorithm, first published by Cooley and Tukey in 1965; the FFT and affordable minicomputers are what made routine FTIR practical.
Key Components
IR Source
- Mid-IR (4000–400 cm⁻¹)
- Globar (silicon carbide rod, heated to ~1100°C)
- Near-IR (12000–4000 cm⁻¹)
- Tungsten-halogen (quartz-halogen) lamp
- Far-IR (<400 cm⁻¹)
- Mercury arc lamp or synchrotron
Beamsplitter
- Mid-IR
- Germanium-coated KBr (potassium bromide) — most common
- Near-IR
- CaF₂ (calcium fluoride) substrate
- Far-IR
- Mylar (polyethylene terephthalate) film or, in many modern instruments, a silicon beamsplitter
Detectors
- DTGS
- Deuterated triglycine sulfate — pyroelectric, room temperature, general purpose. Adequate for most routine measurements.
- MCT
- Mercury cadmium telluride — photoconductive, usually cooled with liquid nitrogen (77 K). 10–100× more sensitive than DTGS. Essential for microscopy, GC-IR, and low-concentration work. Narrow-band MCT detectors stop responding below about 700 cm⁻¹ and become nonlinear at high light levels; thermoelectrically cooled versions avoid liquid nitrogen at some cost in sensitivity.
HeNe Reference Laser
- Purpose
- A helium-neon laser (632.8 nm) runs alongside the IR beam. Its known, stable wavelength provides precise tracking of the moving mirror position — this is the source of Connes' advantage.
- Precision
- Enables wavenumber precision of about 0.01 cm⁻¹, far better than typical dispersive instruments.
Resolution vs. Scan Time
Spectral resolution in FTIR is set by the maximum optical path difference between the two beams. Because light travels to the moving mirror and back, the optical path difference is twice the mirror displacement:
Nominal resolution (cm⁻¹) ≈ 1 / Maximum optical path difference (cm)
To achieve 4 cm⁻¹ resolution, the maximum optical path difference must be about 0.25 cm, which means the mirror moves about 0.125 cm. For 0.5 cm⁻¹, the optical path difference must reach 2 cm (1 cm of mirror travel), which takes longer per scan and needs more precise mechanics. The exact resolution achieved also depends on the apodization function applied to the interferogram.
| Resolution | Max. Optical Path Difference | Mirror Displacement | Typical Use |
|---|---|---|---|
| 4 cm⁻¹ | 0.25 cm | 0.125 cm | Standard for solids and liquids |
| 2 cm⁻¹ | 0.5 cm | 0.25 cm | Sharper bands, e.g. crystalline solids |
| 1 cm⁻¹ | 1 cm | 0.5 cm | Gas phase: resolves rotational lines of light molecules (HCl, CO) |
| 0.5 cm⁻¹ | 2 cm | 1 cm | Gas phase: closely spaced lines (H₂O, CO₂); research instruments go below 0.001 cm⁻¹ |
Improving Signal-to-Noise Ratio
Multiple scans are averaged to reduce random noise. The signal-to-noise ratio improves as the square root of the number of scans: SNR ∝ √n. Doubling the number of scans improves SNR by about 41%. To double the SNR, you need 4× as many scans.
Frequently Asked Questions
What does FTIR stand for?
FTIR stands for Fourier Transform Infrared. It describes a spectroscopy technique that uses a Michelson interferometer to collect infrared absorption data and a mathematical Fourier Transform to convert the raw interferogram into a usable spectrum.
Why is FTIR better than dispersive IR spectroscopy?
FTIR offers three key advantages: the Fellgett (multiplex) advantage (all wavelengths are measured simultaneously, which improves signal-to-noise in a given measurement time when detector noise dominates), the Jacquinot advantage (higher throughput from circular apertures instead of slits), and the Connes advantage (precise wavenumber calibration from a HeNe reference laser).
How long does an FTIR scan take?
A single scan at 4 cm⁻¹ resolution takes about 0.5–1 second. Most measurements average 16–64 scans for better signal-to-noise ratio, taking roughly 10 seconds to 1 minute in total. Higher resolution or more averaged scans increase the measurement time proportionally.
References
- Griffiths, P. R.; de Haseth, J. A. Fourier Transform Infrared Spectrometry, 2nd ed. Wiley, 2007.
- Smith, B. C. Fundamentals of Fourier Transform Infrared Spectroscopy, 2nd ed. CRC Press, 2011.
- Skoog, D. A.; Holler, F. J.; Crouch, S. R. Principles of Instrumental Analysis, 7th ed. Cengage, 2017.