Spectroscopy, at its core, is the study of how matter interacts with light and other forms of electromagnetic radiation. That deceptively simple idea has branched into dozens of distinct techniques, each tuned to reveal something different about a sample: its chemical composition, its molecular structure, the distances between atoms, or even its behavior on timescales shorter than a trillionth of a second. From identifying pollutants in city air to characterizing the atmospheres of planets orbiting distant stars, spectroscopy is one of the most versatile toolkits in modern science. What makes it so powerful is that every molecule and atom absorbs, emits, or scatters light in a characteristic way, producing patterns as unique as fingerprints.
Infrared Spectroscopy and Molecular Fingerprints
When molecules absorb infrared light, their chemical bonds vibrate in ways that depend on the types of atoms involved and the bonds linking them. Fourier transform infrared spectroscopy, commonly called FTIR, exploits this by passing a broad beam of infrared light through a sample and recording which wavelengths get absorbed. The resulting spectrum reveals which functional groups are present, such as hydroxyl groups, carbonyls, or amines. FTIR is one of the most widely used analytical tools in chemistry, though interpreting its spectra by hand can be slow and requires considerable expertise.1PubMed. Functional Group Identification for FTIR Spectra Using Image-Based Machine Learning Models
That bottleneck has driven recent work pairing FTIR with machine learning. Researchers have built deep-learning models trained on thousands of spectra that can predict functional groups automatically. One such model, trained on roughly 8,700 spectra, used attention-based neural networks to predict 17 different functional groups, outperforming older computational approaches in both per-group accuracy and overall compound-level precision.2PubMed. Fcg-Former: Identification of Functional Groups in FTIR Spectra Using Enhanced Transformer-Based Model For industries that need rapid quality control, from pharmaceutical manufacturing to food safety, this kind of automation is a practical game-changer.
Raman Spectroscopy
Raman spectroscopy works by a different mechanism than infrared absorption, but it answers many of the same questions. When laser light hits a sample, most photons scatter without changing energy. A tiny fraction, however, exchange energy with molecular vibrations, shifting to slightly different wavelengths. Those shifts map directly to the bonds and structures in the sample. Because Raman and infrared methods probe molecular vibrations through complementary physical processes, the two are often described as a matched pair: infrared spectroscopy detects changes in a molecule’s dipole moment, while Raman detects changes in its polarizability. Together, they provide a more complete picture of a sample’s chemistry than either one alone.3PubMed Central. Molecular Fingerprint Detection Using Raman and Infrared Spectroscopy Technologies for Cancer Detection: A Progress Review
One of the most exciting frontiers for Raman is in medicine. Over the past decade, Raman scattering has been widely adopted for cancer screening, diagnosis, and even real-time surgical guidance, because it can detect chemical differences between healthy and cancerous tissue without staining or labeling the sample.4PubMed Central. Raman Spectroscopy and Imaging for Cancer Diagnosis Researchers have engineered a handheld Raman probe that fits into a standard brain biopsy system, letting surgeons analyze tissue in situ before extracting it. In validation studies, high-wavenumber Raman spectroscopy detected dense brain cancer tissue during surgery with a sensitivity of 80% and specificity of 90%.5Scientific Reports. A new method using Raman spectroscopy for in vivo targeted brain cancer tissue biopsy That is not yet a replacement for conventional pathology, but it offers surgeons a real-time second opinion at the boundary between tumor and healthy brain.
Surface-Enhanced Raman Spectroscopy
Normal Raman scattering is inherently weak. Only about one in every ten million photons undergoes the inelastic scattering that produces a Raman signal. Surface-enhanced Raman spectroscopy, or SERS, overcomes this limitation by placing analyte molecules near metallic nanostructures, typically gold or silver, whose surface plasmons amplify the electromagnetic field and boost the signal by many orders of magnitude. The amplification is strongest in the tiny gaps between nanoparticles, called “hot spots.”
Engineering those hot spots with precision has been a major research challenge. One approach uses a molecular spacer to link a silver nanoparticle to a metal substrate, creating a nanogap where a single target molecule can be positioned. In experiments using this strategy, 96% of hot spots contained exactly one analyte molecule, allowing reproducible single-molecule measurements. The researchers also mapped how the signal dropped off with distance from the center of the hot spot, finding that the enhancement decreased as the square of the distance.6PubMed. Smart SERS Hot Spots: Single Molecules Can Be Positioned in a Plasmonic Nanojunction Using Host-Guest Chemistry
More recent work has pushed SERS toward detecting specific chemical modifications on biomolecules. A plasmonic particle-in-pore sensor was used to trap a gold nanoparticle inside a gold nanopore, generating a single ultrasmall hot spot capable of detecting both aromatic and non-aromatic amino acids in individual peptides. This system discriminated phosphorylation, a key post-translational modification involved in cell signaling, in the amino acids serine and tyrosine with over 95% and 97% accuracy, respectively.7The Journal of Physical Chemistry Letters. Single-Molecule SERS Detection of Phosphorylation in Serine and Tyrosine Using Deep Learning-Assisted Plasmonic Nanopore Detecting a single phosphorylation event on a single peptide is remarkable; it opens doors to studying diseases driven by signaling errors at a resolution that bulk methods cannot reach.
Nuclear Magnetic Resonance
Nuclear magnetic resonance spectroscopy, or NMR, works on a completely different principle from the optical methods above. Certain atomic nuclei behave like tiny magnets. When placed in a strong external magnetic field and pulsed with radio waves, they absorb and re-emit energy at frequencies that depend on their local chemical environment. The resulting spectrum encodes detailed information about molecular structure, the distances between atoms, and the flexibility of different parts of a molecule.
NMR has become one of the primary methods for determining the three-dimensional structures of proteins, alongside X-ray crystallography. Chemical shifts, the most readily measured NMR parameter, are highly sensitive to protein conformation. Using chemical shifts as structural constraints in combination with molecular mechanics calculations, researchers have determined protein structures to a resolution of two angstroms or better across representative proteins containing up to 123 amino acid residues.8PubMed Central. Protein structure determination from NMR chemical shifts Unlike X-ray crystallography, NMR can study proteins in solution rather than in crystal form, making it especially valuable for examining flexible or disordered regions that resist crystallization.
X-ray Techniques
X-ray crystallography has long been the most commonly used technique for determining the structures of proteins and other biological macromolecules.9PubMed Central. x ray crystallography The method works by shining X-rays through a crystal and analyzing the diffraction pattern produced as the X-rays bounce off atoms in the crystal lattice. Some of the most celebrated achievements in structural biology, including the first structures of DNA, hemoglobin, and thousands of drug targets, came from X-ray crystallography.
Beyond structure determination, X-ray spectroscopy encompasses analytical techniques like X-ray fluorescence (XRF) and X-ray photoelectron spectroscopy (XPS). XRF identifies elements in a sample by measuring the characteristic X-rays they emit when excited, while XPS probes the chemical states of elements at a surface. A combined XRF/XPS instrument was developed to analyze the surface layers of silicon spheres, measuring oxygen mass deposition in nanograms per square centimeter and quantifying the ratios of oxygen, silicon, and carbon at the surface.10Acta IMEKO. Development of a combined XRF/XPS surface-analysis system for the surface-layer quantification of 28Si spheres This kind of surface analysis matters for semiconductor manufacturing, thin-film engineering, and metrological standards.
Fluorescence and FRET
Fluorescence spectroscopy detects light emitted by molecules that have been excited by absorbing photons of a shorter wavelength. When two fluorescent molecules are close enough together, typically within a few nanometers, the excited one can transfer its energy to the other without emitting a photon at all. This process, Förster resonance energy transfer (FRET), acts as a molecular ruler: the efficiency of energy transfer depends steeply on the distance between the two fluorophores, so measuring FRET efficiency reveals how far apart two tagged molecules or parts of a molecule sit.
FRET is widely used to study protein-protein interactions and structural changes inside living cells. By measuring the fluorescence lifetime of a donor molecule, researchers can calculate the distance to a nearby acceptor molecule with nanometer spatial resolution and nanosecond time resolution.11PubMed. Nanosecond fluorescence resonance energy transfer-fluorescence lifetime imaging microscopy to localize the protein interactions in a single living cell Using FRET between fluorescently labeled DNA dyes, scientists have even visualized how tightly chromatin, the complex of DNA and protein that packages the genome, is compacted at the nanoscale inside living cells.12PubMed Central. Chromatin nanoscale compaction in live cells visualized by acceptor-to-donor ratio corrected Förster resonance energy transfer between DNA dyes
The technique is tricky in practice, though. Artifacts from variable fluorophore concentrations or environmental conditions can mimic or mask real FRET signals. Fluorescence lifetime imaging microscopy (FLIM) helps by measuring the time a fluorophore spends in its excited state, a parameter that is independent of concentration and excitation intensity. Adding environmental controls for temperature and CO₂ during live-cell experiments eliminated nonspecific FRET signals in one system, producing a sevenfold improvement in measurement precision and yielding the only statistically significant evidence of specific molecular interactions in the live cells studied.13PubMed Central. Physiological fluorescence lifetime imaging microscopy improves Förster resonance energy transfer detection in living cells
Atomic Emission and Trace Element Analysis
While molecular spectroscopy reveals bonds and structures, atomic spectroscopy identifies and quantifies individual elements. Inductively coupled plasma optical emission spectrometry, or ICP-OES, works by injecting a liquid sample into a plasma torch hot enough to break molecules apart and excite their constituent atoms, which then emit light at wavelengths characteristic of each element. It is especially powerful for multielement analysis, measuring trace and ultratrace concentrations across a wide range of sample types. Applications span environmental monitoring, food analysis, and medical diagnostics.14PubMed. How ICP-OES changed the face of trace element analysis: Review of the global application landscape Water treatment plants, for instance, routinely use ICP-OES to screen for heavy metals at parts-per-billion levels in a single run.
Mass Spectrometry Hybrids
Strictly speaking, mass spectrometry measures mass-to-charge ratios rather than light, so it sits at the edge of what most textbooks call spectroscopy. In practice, though, it is so often coupled with spectroscopic separations that it belongs in the conversation. Liquid chromatography paired with tandem mass spectrometry (LC-MS/MS) is a workhorse of metabolomics, the large-scale study of small molecules in biological systems. The technique separates complex mixtures by their physical and chemical properties, then identifies and quantifies individual metabolites with high sensitivity.15PubMed Central. Metabolite identification and quantitation in LC-MS/MS-based metabolomics Clinical labs use it to measure drug levels in blood, diagnose inborn errors of metabolism, and screen newborns for dozens of genetic conditions simultaneously.
Monitoring the Atmosphere
Spectroscopy scaled to the open atmosphere is the basis for air-quality monitoring systems around the world. Differential optical absorption spectroscopy, or DOAS, works by shining a broad-spectrum light beam across several kilometers of open air and analyzing the absorption features imprinted by pollutant gases along the path. By fitting the measured spectrum to laboratory reference spectra, the system evaluates the path-averaged concentrations of gases like nitrogen dioxide, sulfur dioxide, ozone, ammonia, and nitrous acid.16Applied Optics. Differential optical absorption spectroscopy (DOAS) system for urban atmospheric pollution monitoring
Field deployments of DOAS have covered heavily polluted cities. In Milan, a DOAS system measured SO₂, NO₂, O₃, and nitrous acid over a horizontal light path of 1.7 km during a two-week campaign.17Applied Optics. Differential optical absorption spectrometer for measurement of tropospheric pollutants A pulsed variant tracked temporal changes in NO₂ concentration and aerosol loading along a 5.5 km optical path in the lower troposphere.18Journal of Quantitative Spectroscopy and Radiative Transfer. Pulsed differential optical absorption spectroscopy applied to air pollution measurement in urban troposphere Because DOAS measures an entire air column rather than point samples, it captures the spatial average that point monitors miss, making it a valuable complement to ground-level monitoring stations.
Spectroscopy in Astronomy
The very origins of spectroscopy are intertwined with astronomy, where splitting starlight into its component wavelengths revealed the chemical compositions of stars long before anyone could physically visit them. Today, transmission spectroscopy is a primary method for probing the atmospheres of exoplanets. When a planet transits its host star, starlight filters through the planet’s atmosphere, and specific molecules absorb characteristic wavelengths, imprinting their signatures on the spectrum.
Bayesian retrieval methods applied to these spectra can uniquely constrain the mixing ratios of absorbing gases and even spectrally inactive components like molecular nitrogen, provided the observations capture both broadband transit depths in absorption features and the slope of molecular Rayleigh scattering.19The Astrophysical Journal. ATMOSPHERIC RETRIEVAL FOR SUPER-EARTHS: UNIQUELY CONSTRAINING THE ATMOSPHERIC COMPOSITION WITH TRANSMISSION SPECTROSCOPY The practical reality remains difficult. High-resolution observations of the rocky exoplanet GJ 486b, taken with three different spectrographs, searched for signals from a wide variety of molecular species but found no robust detections. The data were sensitive enough, however, to rule out a clear hydrogen-helium atmosphere and a pure water-vapor atmosphere on that planet.20The Astronomical Journal. High-resolution Transmission Spectroscopy of the Terrestrial Exoplanet GJ 486b Non-detections like this are still informative: they narrow the range of possible atmospheric compositions and push comparative planetology into the terrestrial regime.
As upcoming space missions are expected to generate massive volumes of spectral data, machine-learning approaches are being developed to accelerate the computationally intensive retrieval process, aiming to make biosignature detection on habitable-zone planets more feasible.21ESS Open Archive. Accelerating Transmission Spectroscopy of Exoplanets for Biosignature Detection Using Machine Learning
Terahertz Spectroscopy
Terahertz radiation sits between infrared and microwave on the electromagnetic spectrum, a region that was historically difficult to generate and detect but has become increasingly accessible. Terahertz time-domain spectroscopy is sensitive to large-scale molecular motions, particularly the collective vibrations involving whole groups of atoms, hydrogen-bond networks, and lattice modes in crystalline materials. This makes it useful for distinguishing between polymorphs, different crystal forms of the same chemical compound, which can have dramatically different properties even though their molecular formulas are identical.
In pharmaceutical research, terahertz spectroscopy has been used to differentiate polymorphs of a drug cocrystal. Two crystal forms of a nifedipine-isonicotinamide cocrystal displayed distinct spectral fingerprints in the terahertz range. Temperature-dependent measurements revealed that hydrogen bonds between the drug and the coformer produce vibrations below 1.5 THz, while methyl group rotations spread across 1.5 to 4.0 THz. These features help stabilize the cocrystal and prevent oxidation of the drug.22PubMed. The fingerprints of nifedipine/isonicotinamide cocrystal polymorph studied by terahertz time-domain spectroscopy For drug developers, the ability to identify polymorphs quickly and non-invasively helps ensure that the correct crystal form makes it into the final product, which affects everything from shelf life to how efficiently the body absorbs the drug.
Machine learning is being paired with terahertz spectroscopy to handle mixtures. Methods like non-negative matrix factorization and self-modeling mixture analysis have been shown to extract individual component spectra from mixed-sample terahertz data, matching well with spectra measured on pure reference materials.23PubMed. Component spectra extraction and quantitative analysis for preservative mixtures by combining terahertz spectroscopy and machine learning
Ultrafast Spectroscopy
Chemical reactions and energy transfer processes in molecules often unfold on femtosecond timescales, where one femtosecond is a millionth of a billionth of a second. To capture events that fast, you need light pulses that are even shorter. Ultrafast pump-probe spectroscopy works by hitting a sample with a brief “pump” pulse to trigger a process, then probing the sample with a second pulse after a precisely controlled delay. By varying the delay, researchers build a frame-by-frame movie of the reaction.
Real-time pump-probe imaging implemented on a single-shot basis has been demonstrated in the femtosecond time regime for studying photochemical reaction dynamics in biological and organic materials.24Japanese Journal of Applied Physics. Femtosecond Real-Time Pump-Probe Imaging Spectroscopy Implemented on a Single Shot Basis This kind of temporal resolution is essential for understanding how plants capture sunlight during photosynthesis, how vision begins when a photon hits a retinal molecule, and how charge carriers move through solar cell materials in the first moments after light absorption.
Mössbauer Spectroscopy and Mineral Chemistry
Mössbauer spectroscopy occupies a specialized niche. It works by measuring the absorption of gamma rays by specific nuclei, most commonly iron-57, embedded in a solid. Because the gamma-ray energy required for absorption is exquisitely sensitive to the electronic and magnetic environment of the nucleus, Mössbauer spectroscopy can distinguish between different oxidation states and coordination environments of iron in minerals, catalysts, and other solid materials.
In geology, this is invaluable for understanding how iron behaves under the extreme pressures and temperatures of subduction zones. Mössbauer measurements on serpentine minerals from subduction-related rocks found that antigorite, a high-pressure serpentine, maintains a remarkably constant ratio of ferric to total iron, ranging from about 0.17 to 0.23 across samples formed under different pressure-temperature conditions and geologic ages. By contrast, lizardite, a lower-pressure polymorph, gave a ratio of 0.49, nearly half its iron in the oxidized state.25European Journal of Mineralogy. Iron oxidation state in serpentines and magnesian chlorites of subduction-related rocks Those differences in iron chemistry affect how rocks store and release water and oxidizing agents as tectonic plates descend into the mantle, with implications for volcanism and the deep cycling of elements between Earth’s surface and interior.