FTIR Troubleshooting

What you see

Characteristic sharp, closely spaced lines appear at 3500–3950 cm⁻¹ (O-H stretching) and 1300–1900 cm⁻¹ (H-O-H bending). These are atmospheric water vapor absorptions, not from the sample.

Possible causes

• Ambient humidity in the instrument sample compartment
• Background and sample spectra collected under different humidity conditions
• Inadequate or absent purge gas

Solutions

1. 1Purge the spectrometer with dry nitrogen (N₂) or dry air — the most effective solution
2. 2Ensure background and sample scans are collected close together in time under the same conditions
3. 3Use atmospheric compensation software (available in most FTIR software packages) to subtract water vapor lines
4. 4Desiccant packs in the sample compartment can help in non-purged instruments

What you see

A sharp doublet near 2350 cm⁻¹ (asymmetric stretch) and a weaker band near 670 cm⁻¹ (bending mode). These are the signature absorptions of atmospheric carbon dioxide.

Possible causes

• Ambient CO₂ in the optical path
• Mismatch between CO₂ levels during background and sample collection
• Door opening, breathing near the instrument, or HVAC changes between scans

Solutions

1. 1Purge with dry N₂ or dry air (eliminates both CO₂ and water vapor)
2. 2Collect background immediately before sample measurement
3. 3Use atmospheric compensation software
4. 4Avoid breathing near the sample compartment during measurements

What you see

The baseline drifts upward or downward across the spectrum, or curves in unexpected ways. Peaks appear to sit on a tilted or curved background rather than a flat baseline near 100% transmittance.

Possible causes

• Scattering — particles in the sample are comparable in size to IR wavelengths
• Detector drift or warm-up instability
• Optical misalignment
• Stale background — conditions changed significantly since the background was collected

Solutions

1. 1Rerun the background immediately before the sample measurement
2. 2Check alignment — many instruments have an automatic alignment procedure
3. 3Reduce particle size (grind finely, use ball mill) if using KBr pellet or DRIFTS
4. 4Switch to ATR — the evanescent wave does not pass through the sample, so particle scattering does not distort the baseline (grind coarse powders anyway for good crystal contact)
5. 5Allow the instrument to warm up fully (typically 15–30 minutes) before collecting data

What you see

Random fluctuations in the baseline make it difficult to identify small peaks or accurately determine peak positions. The signal-to-noise ratio is poor.

Possible causes

• Too few scans averaged
• Weak IR source (aging or failing globar)
• Degraded detector performance
• Low sample concentration or short pathlength
• Electronic noise from nearby equipment

Solutions

1. 1Increase the number of scans — SNR improves as √n (e.g., 64 scans gives 8× the SNR of a single scan)
2. 2Check and replace the IR source if output has decreased
3. 3For MCT detectors: ensure liquid nitrogen dewar is filled and detector is properly cooled
4. 4Increase sample concentration or pathlength to produce stronger absorption peaks
5. 5Reduce resolution if high resolution is not needed — 4 cm⁻¹ resolution produces lower noise than 1 cm⁻¹

What you see

The strongest peaks in the spectrum appear flat or clipped at their maxima rather than coming to a point. In absorbance mode, peaks exceed A ≈ 1.5–2.0 and flatten out.

Possible causes

• Too much sample — almost no light reaches the detector at the band centre, so noise and detector nonlinearity dominate
• Sample too thick (transmission mode) or too concentrated
• Long pathlength with concentrated solutions

Solutions

1. 1Reduce sample amount — use a thinner KBr pellet, dilute the solution, or reduce the pathlength
2. 2Switch to ATR — the fixed, shallow penetration depth prevents oversaturation
3. 3If using ATR, try a crystal with smaller depth of penetration (Ge instead of ZnSe or diamond)
4. 4For solutions, dilute and remeasure — quantitative work requires the peaks used for quantitation to be in the linear range (A < 1.0 ideally)

What you see

A sinusoidal oscillation is superimposed on the spectrum baseline. The fringes are regular and evenly spaced in wavenumber.

Possible causes

• Internal reflections within a thin, uniform-thickness film
• The film acts as an etalon — constructive and destructive interference between front and back surface reflections

Solutions

1. 1Tilt the sample slightly to change the angle of incidence
2. 2Use a different thickness of sample
3. 3Apply fringe removal software available in most FTIR packages
4. 4Use ATR instead of transmission for thin film samples