An infrared (IR) spectroscopy machine works by shining infrared light through or onto a sample and measuring which wavelengths the sample absorbs. Because every molecule absorbs infrared light at characteristic frequencies tied to the vibrations of its chemical bonds, the resulting absorption pattern acts as a molecular fingerprint. The most common form of this instrument today is the Fourier transform infrared (FTIR) spectrometer, which has largely replaced older designs and become a workhorse in chemistry labs, pharmaceutical plants, forensic units, and even space missions.
What the Machine Actually Does
When infrared radiation passes through a material, specific wavelengths are absorbed depending on which chemical bonds are present and how they vibrate. Stretching, bending, rocking, and twisting motions of bonds between atoms each absorb at particular frequencies. The instrument records how much light makes it through (or bounces back from) the sample at each wavelength, producing a spectrum: a graph of absorption versus frequency. Peaks in that spectrum correspond to specific bond types, so a trained analyst or a software algorithm can read off what functional groups are present, how pure a substance is, or whether a sample matches a known reference.
The infrared region of the electromagnetic spectrum spans wavelengths from roughly 700 nanometers to 1 millimeter, sitting between visible light and microwaves. In practice, most laboratory IR spectroscopy focuses on the mid-infrared range (roughly 2.5 to 25 micrometers), where the most diagnostically useful molecular vibrations occur. Near-infrared (roughly 700 to 2,500 nanometers) and far-infrared instruments exist too, each with their own strengths for different kinds of analysis.
Inside the Box
At its simplest, an IR spectrometer has three main parts: a light source that emits broadband infrared radiation, a way to sort that light by wavelength, and a detector to measure what comes out the other end. The specifics of each component vary depending on the instrument type.
For light sources, the choice depends on which part of the infrared spectrum you need. A globar (a silicon carbide rod heated to high temperature) is the classic mid-infrared source. Tungsten halogen lamps work well in the near-infrared range. More specialized instruments may use laser diodes, which offer broad spectral coverage from the visible range well into the infrared and tend to have good spectral stability.1Optical Materials: X. Recent progress on broadband near-infrared phosphors-converted light emitting diodes for future miniature spectrometers In research settings, synchrotron radiation provides an exceptionally bright and tightly focused infrared beam, enabling imaging at much finer spatial resolution than a benchtop source can deliver.
The detector side has its own variety. Pyroelectric detectors respond to changes in temperature caused by incoming infrared photons. Semiconductor-based photoconducting detectors are more sensitive but often need to be cooled, sometimes with liquid nitrogen. The choice of detector affects both the sensitivity and the speed of the instrument, and manufacturers pair detectors with specific spectral ranges to optimize performance.
Why Nearly Every Modern Instrument Uses Fourier Transform
Older dispersive IR spectrometers used prisms or diffraction gratings to split infrared light into individual wavelengths and scanned through them one at a time. This was slow and wasted most of the light, since only a narrow slice hit the detector at any moment. The Fourier transform approach, which became commercially dominant from the 1980s onward, works differently. An interferometer (usually a Michelson interferometer with a moving mirror) sends all wavelengths to the detector simultaneously, producing an interference pattern called an interferogram. A mathematical operation, the Fourier transform, converts this raw signal into the familiar absorption spectrum.
The practical payoff is speed and sensitivity. Because every wavelength contributes to the signal at every instant, FTIR spectrometers collect data much faster and with a better signal-to-noise ratio. You can average many scans in the time it would take a dispersive instrument to complete one, which makes weak spectral features easier to detect. FTIR instruments also have precise internal wavelength calibration, typically provided by a helium-neon (HeNe) laser beam that travels alongside the infrared beam and tracks the mirror position. This built-in reference helps keep the wavelength axis accurate over time.
How Samples Meet the Beam
One of the most practically important choices when using an IR spectroscopy machine is how to present the sample to the infrared beam. The sampling technique affects what kinds of materials you can analyze, how much preparation is needed, and what depth of the sample you are probing.
- Transmission: The simplest approach. Infrared light passes straight through the sample. Works well for thin films, gases, and liquids in transparent cells, but solids often need to be ground into a fine powder and pressed into a pellet with an IR-transparent material like potassium bromide (KBr).
- Attenuated total reflection (ATR): The sample is pressed against a crystal with a high refractive index (diamond, germanium, or zinc selenide are common). Infrared light bouncing inside the crystal creates an evanescent wave that penetrates a short distance into the sample surface. ATR is popular because it requires almost no sample preparation: you can press a powder, a piece of fabric, or even a paste directly onto the crystal and get a spectrum in seconds.
- Diffuse reflectance (DRIFTS): Infrared light is scattered by a rough or powdered sample surface, and the diffusely reflected light is collected and directed to the detector. Useful for powders and rough solids where making a pellet or getting good crystal contact is impractical.
- Specular reflectance: Used for smooth, reflective surfaces like coatings on metal substrates. The beam reflects off the sample surface at a defined angle.
ATR in particular has become the default for many routine applications because of its convenience. The depth of penetration into the sample can be varied by changing the angle at which the infrared beam hits the crystal, which lets analysts probe different layers of a material.2PubMed. Micro-Attenuated Total Reflection Fourier Transform Infrared (Micro ATR FT-IR) Spectroscopic Imaging with Variable Angles of Incidence For pharmaceutical tablets, forensic trace evidence, and quality control on production lines, ATR has largely replaced older KBr pellet methods because there is less chance of altering the sample during preparation.
The Persistent Problem of Moisture
If there is one annoyance that every IR spectroscopist knows, it is water. Water vapor in the air absorbs infrared light strongly, and its absorption bands overlap with some of the most important regions of a spectrum, particularly the ranges where proteins and lipids show up. The instrument’s optical path runs through ambient air, so any fluctuation in humidity can leave unwanted water features stamped on your data.
Two main factors drive this problem. First, even small temperature changes in the HeNe laser used for wavelength calibration can introduce a slight systematic shift between the background and sample measurements, which makes simple spectral subtraction of the water signal unreliable.3Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy. A new approach to removing interference of moisture from FTIR spectrum Second, the actual concentration of water vapor in the light path fluctuates moment to moment, so the water absorption changes between the background scan and the sample scan. Researchers have developed computational methods to address both issues, including building large databases of background spectra and matching them to each sample measurement based on spectral similarity metrics, and using advanced correlation techniques to subtract the residual moisture signal.4PubMed. Comprehensive modified approaches to reducing the interference of moisture from an FTIR spectrum and the corresponding second derivative spectrum
In biological applications, the problem is even trickier because water is part of the sample, not just the atmosphere. When researchers use FTIR to classify bacteria, for example, the infrared absorption of water overlaps heavily with the protein region (roughly 1200 to 1800 wavenumbers) and the lipid region (roughly 2800 to 3700 wavenumbers). One group found that the polysaccharide and nucleic acid region below 1200 wavenumbers was the only spectral window that remained credible for bacterial typing once water interference was accounted for, and restricting analysis to that window actually improved classification accuracy.5PubMed. The strategy for correcting interference from water in Fourier transform infrared spectrum based bacterial typing Many high-end instruments address moisture by purging the optical compartment with dry nitrogen or dry air, but this adds cost and complexity. Some portable or field instruments simply accept a noisier spectrum in the water-affected regions.
Keeping the Machine Honest
An IR spectrum is only useful if the wavelength axis is accurate and the intensity readings are reproducible. Calibration and verification are handled through reference standards, and the most widely used one is remarkably low-tech: a thin film of polystyrene plastic. The National Institute of Standards and Technology (NIST) offers a standard reference material (SRM 1921) consisting of a polystyrene film about 38 micrometers thick, whose absorption peaks fall at well-characterized wavelengths. Running a scan of this film and comparing the measured peak positions to the certified values tells you whether the instrument’s wavelength calibration has drifted.6Applied Spectroscopy. Determination of Wavelength Accuracy in the Near-Infrared Spectral Region Based on NIST’s Infrared Transmission Wavelength Standard SRM 1921 For near-infrared work, a thicker polystyrene sample is needed because the absorption features in that region are weaker, but the same basic principle applies.
Beyond wavelength accuracy, labs typically verify photometric accuracy (are the intensity readings correct?), noise levels, and resolution on a regular schedule. Regulated industries like pharmaceuticals often require documented calibration records, and instrument software usually automates the tests so a technician can run them in a few minutes.
IR Microscopy and Chemical Imaging
Coupling an FTIR spectrometer to a microscope creates a tool that can record infrared spectra from spots just a few micrometers across. This is invaluable when you need chemical information at the scale of individual cells or thin material layers. Spatial resolution in infrared microspectroscopy is fundamentally limited by the wavelength of the light being used, and since mid-infrared wavelengths are several micrometers long, the achievable resolution is coarser than what you get with visible-light microscopy.7PubMed. Spatial resolution in infrared microspectroscopic imaging of tissues Still, it is fine enough to map the distribution of proteins, lipids, and other molecular components across a tissue section or a polymer film.
Focal plane array (FPA) detectors accelerated this field by allowing thousands of spectra to be collected simultaneously across a two-dimensional area, rather than scanning point by point. When paired with the intense, tightly focused infrared beam from a synchrotron light source, FPA-based imaging achieves subcellular spatial resolution, with pixel spacing two orders of magnitude finer than conventional benchtop systems. This combination has been used for label-free biochemical imaging of neurons and other delicate biological structures.8PubMed. Biochemical label-free tissue imaging with subcellular-resolution synchrotron FTIR with focal plane array detector Synchrotron-based IR imaging is not something you find in a routine lab, but it pushes the technique’s boundaries in biomedical research.
Pharmaceuticals and Drug Quality
The pharmaceutical industry is one of the largest users of IR spectroscopy machines. The technique is fast, generally nondestructive, and sensitive to the solid-state form of a drug, not just its chemical composition. That last point matters more than you might expect. Many drug molecules can crystallize in multiple arrangements (polymorphs), and different polymorphs can have dramatically different dissolution rates and bioavailability. An IR spectrum can distinguish between these forms because the crystal packing subtly changes the vibrations of the bonds.
Researchers have shown that ATR-FTIR combined with statistical pattern recognition can quickly assign the polymorph present in a finished tablet, with minimal sample preparation. For mebendazole, an antiparasitic drug that exists in multiple crystal forms with very different therapeutic activity, an ATR-FTIR approach was found to be a fast, efficient, and nondestructive way to identify which solid-state form was present in drug products.9PubMed. Mebendazole crystal forms in tablet formulations. An ATR-FTIR/chemometrics approach to polymorph assignment Far-infrared spectroscopy, which probes lower-frequency vibrations sensitive to the crystal lattice itself, has also been explored as a process analytical tool for catching polymorph changes during manufacturing.10PubMed Central. Far-Infrared Spectroscopy as A Probe for Polymorph Discrimination
On production lines, near-infrared (NIR) spectrometers are commonly embedded directly into manufacturing equipment for real-time monitoring of moisture content, blend uniformity, and tablet coating thickness. NIR has the advantage that it can penetrate deeper into a sample than mid-IR, though its spectral features are broader and harder to interpret without computational help.
Forensic Applications
Crime labs use IR spectroscopy to identify unknown substances and link physical evidence to a source. Fibers, paint chips, drugs, explosives, and inks all have distinctive infrared signatures. A fiber recovered from a crime scene, for instance, can be compared against a database of known materials. In one study of synthetic textile fibers using FTIR and machine-learning classification, over 97% of test samples were correctly classified by fiber type at a significance level of 5%.11PubMed Central. Forensic Analysis of Textile Synthetic Fibers Using a FT-IR Spectroscopy Approach
FTIR microspectroscopy (often written µ-FTIR) adds spatial resolution, which forensic analysts use to image heterogeneous evidence. Paint chips, for example, are multilayered, and an IR microscope can record a spectrum from each layer individually, revealing whether the binder is acrylic, alkyd, or polyurethane. Similarly, latent fingerprints can be chemically imaged with µ-FTIR to detect trace compounds like drugs embedded in the fingerprint residue.12Journal of Forensic Science Education. Chemical imaging of latent fingerprints, paint chips, and fibers using µ-FTIR: An experiment for forensic chemistry and instrumental analysis courses The technique is nondestructive, which matters in a forensic context where the evidence may need to be preserved for court.
Coupling IR with Other Instruments
An IR spectrometer often works best when it is not alone. Hyphenated techniques connect it to another instrument so that the output of one feeds into the other, giving you complementary information in a single experiment. One of the most established pairings is thermogravimetric analysis coupled with FTIR (TGA-FTIR). A TGA instrument heats a sample at a controlled rate and records its mass loss, while the gases driven off are piped directly into an FTIR spectrometer, which identifies them in real time. This combination lets you watch a material decompose step by step and know exactly which gases are released at each temperature. It has been used to study the decomposition mechanisms of energetic materials, polymers, fuels, and biomass, among other things.13Thermochimica Acta. Insights into decomposition mechanism and gas evolution characteristics of energy-rich nitrochitosan by hyphenated TGA-FTIR analysis
Gas chromatography–FTIR (GC-FTIR) is another pairing, where a gas chromatograph separates a complex mixture into individual components that are then identified by their infrared spectra as they elute. Liquid chromatography can be coupled similarly, though it is trickier because the liquid solvents themselves absorb infrared light. These hyphenated setups are common in environmental analysis, petrochemical characterization, and any application where you need both separation and identification.
Machine Learning Meets IR Spectra
Interpreting an infrared spectrum has traditionally been a manual skill: experienced spectroscopists learn to recognize the characteristic peaks of different functional groups. But the volume of spectra generated by modern instruments, especially imaging systems that produce thousands of spectra per sample, has driven growing interest in automated interpretation. Researchers have trained convolutional neural networks on images of FTIR spectra from thousands of organic molecules to identify functional groups automatically. One group scraped intensity-frequency data for nearly 9,000 gas-phase organic molecules from the NIST spectral database, converted them to images, and trained models for 15 of the most common functional groups.14ChemRxiv. Functional Group Identification for FTIR Spectra Using Image-Based Machine Learning Models These models can then be applied to previously unseen spectra, potentially accelerating routine identification tasks.
Machine learning also shows up in classification problems, like the forensic fiber study mentioned earlier, and in pharmaceutical quality control, where chemometric models built on principal component analysis or partial least squares regression help extract quantitative information from overlapping spectral features. The trend is clearly toward software doing more of the interpretation, though human oversight remains essential for edge cases and for validating that the models are not learning artifacts rather than genuine chemistry.
Environmental and Atmospheric Monitoring
IR spectroscopy does not have to happen in a laboratory. Open-path FTIR (OP-FTIR) instruments send an infrared beam across an open span of atmosphere, sometimes hundreds of meters, and measure the absorption caused by gases along that path. This approach has been evaluated for large-scale monitoring of greenhouse gases, including real-time detection of carbon dioxide leaks at carbon capture and storage facilities.15Energy Procedia. Ground-based Remote Sensing with Open-path Fourier- transform Infrared (OP-FTIR) Spectroscopy for Large-scale Monitoring of Greenhouse Gases Because the technique can identify and quantify multiple gas species simultaneously along the beam path, it gives a spatial picture of gas concentrations that point sensors (which sample a single location) cannot provide.
OP-FTIR has also been used to monitor volcanic emissions, industrial fence-line concentrations, and wildfire smoke composition. The instruments are ruggedized for outdoor conditions and can run unattended, making them suitable for long-term monitoring campaigns. Meanwhile, satellite-based infrared spectrometers operate on the same principles at a planetary scale, measuring atmospheric gases and surface compositions from orbit.
IR Spectroscopy in Space
Some of the most striking applications of infrared spectroscopy happen far from any laboratory. Interstellar clouds contain icy mantles of frozen molecules clinging to dust grains, and the composition of those ices is studied almost entirely through their infrared absorption spectra. The Infrared Space Observatory (ISO) provided spectra in the 2 to 20 micrometer range covering the principal solid-state features of condensed matter in interstellar clouds, opening a window into the inventory of carbon-, nitrogen-, and oxygen-bearing material present during the formation of planetary systems.16PubMed. Interstellar ices studied with the Infrared Space Observatory
To interpret those astronomical spectra, laboratory scientists create “ice analogues” by freezing mixtures of molecules like water, carbon dioxide, methane, ammonia, and methanol at cryogenic temperatures and recording their FTIR spectra under conditions that mimic interstellar environments.17Astronomy & Astrophysics. Infrared spectra of complex organic molecules in astronomically relevant ice mixtures: V. Methyl cyanide (acetonitrile) These reference spectra are then matched against the astronomical data to identify what molecules are present in a given cloud. With the James Webb Space Telescope now delivering infrared spectra of interstellar ices at unprecedented resolution and sensitivity, the demand for high-quality laboratory reference spectra has surged. Researchers have also begun investigating how varying the angle of incidence affects the appearance of ice spectra, since icy grains in space are oriented randomly and astronomers observe them along different lines of sight.18Monthly Notices of the Royal Astronomical Society. Infrared spectroscopy of astrophysical ice analogues at oblique angles
Portable and Handheld Instruments
For decades, IR spectrometers were heavy benchtop instruments confined to climate-controlled labs. That has changed. Handheld FTIR and near-infrared spectrometers now weigh under two kilograms and can run on battery power. Customs officials use them to screen suspicious powders at border crossings. Art conservators carry them into museums to identify pigments and binders on paintings without removing a sample. Recycling plants use NIR sensors on conveyor belts to sort plastics by polymer type in real time.
Miniaturization comes with tradeoffs. Smaller optics mean lower throughput and coarser spectral resolution, and the detectors in handheld units are usually uncooled, which limits sensitivity. The spectral range is often narrower than a benchtop instrument, focusing on the most diagnostically useful windows. Still, for many identification tasks, “good enough in the field” beats “perfect back at the lab.” The software embedded in portable instruments typically runs spectral matching against an onboard library and returns a material identification within seconds, so the user does not need to interpret the spectrum visually at all.
The gap between portable and benchtop performance continues to narrow as detector technology improves and computational power gets cheaper. Micro-electromechanical systems (MEMS) based interferometers, which replace the bulky moving mirror of a traditional Michelson interferometer with a tiny chip-scale device, represent one frontier. Another is the development of broadband LED-based infrared sources that could replace incandescent emitters with more stable and efficient solid-state alternatives. Whether you are checking the identity of a raw material on a loading dock or mapping the chemistry of interstellar ices from a space telescope, the underlying physics is the same: infrared light, molecular vibrations, and a detector patient enough to catch what comes through.