Atomic spectroscopy has quietly gone from counting elements in bulk to counting single nanoparticles, imaging tissue at the micrometer scale, and reading Martian rocks from several meters away. The instruments got smaller, the software got smarter, and the questions they can now answer got a lot more interesting.
For a field built on more than a century of physics, atomic spectroscopy is having a strikingly modern moment. Tandem mass spectrometry now strips away interferences that once made trace arsenic or lead isotope work a guessing game. Single-particle ICP-MS counts and sizes individual nanoparticles rather than reporting a blurred average, opening a direct line into how engineered nanomaterials behave in the environment and the body. Laser-induced breakdown spectroscopy has left the bench entirely, riding aboard a rover across Jezero Crater and beyond and, closer to home, helping to classify tumor tissue in near-real time. Laser ablation ICP-MS now maps trace elements at micrometer resolution, and artificial intelligence is quietly rewriting how all of these data get interpreted, calibrated, and trusted. This article surveys what is genuinely new across these fronts, drawing on recent literature and trade coverage, and asks where the field is headed next.
Atomic spectroscopy rarely makes headlines the way genomics or AI drug discovery does, yet it underpins decisions that touch nearly everyone: whether tap water is safe, whether a supplement contains what the label claims, whether a Mars rock once sat in water. In recent years, three forces have converged to change what the field can do. First, tandem and multi-quadrupole mass spectrometry has made interference removal routine for many matrices that used to defeat single-quadrupole instruments.1 Second, single-particle and imaging modes have shifted the fundamental unit of measurement from the bulk sample to the individual particle or cell.2 Third, artificial intelligence and machine learning have moved from a novelty into the standard toolkit for calibration, classification, and real-time decision-making.3 Layered on top of all this is a hardware story: instruments that once filled a lab bench now fit in a backpack, and one is currently roving across the surface of Mars.4 None of this is science fiction. It is what analytical chemists are actually publishing and building right now, and it is worth pausing to look at what is genuinely new rather than incremental.
For decades, trace-level analysis by inductively coupled plasma–mass spectrometry (ICP-MS) meant living with spectral interferences: polyatomic ions and isobaric overlaps that could inflate or mask a true signal, especially for tricky elements such as arsenic, selenium, and the low-abundance isotopes of lead.5 The addition of a second quadrupole ahead of the reaction cell, creating so-called ICP-MS/MS or triple-quadrupole systems, changed the calculus. Because the first quadrupole filters the ion beam before it ever reaches the reaction cell, gas-phase chemistry inside the cell becomes far more consistent from one sample matrix to the next.5 Work applying oxygen reaction-cell chemistry in MS/MS mode demonstrated accurate arsenic quantitation, with selenium measured under the same conditions, across certified reference materials ranging from plant and animal tissues to hair, using a single set of instrument conditions,5 and new reaction gases and applications for ICP-MS/MS continue to emerge.6 The practical effect is that laboratories doing routine environmental, food-safety, and clinical trace analysis can now report defensible parts-per-trillion data on a much shorter timeline and with far less method-development overhead than was once required.
Perhaps the most conceptually interesting shift in atomic spectroscopy is single-particle ICP-MS (spICP-MS), which does not measure a sample's average elemental concentration at all. Instead, it detects the transient ion pulse produced as each individual nanoparticle flashes through the plasma, allowing simultaneous measurement of particle number concentration, size, and size distribution, together with the dissolved-ion background, all without any prior separation step.7 Coupling this approach with tandem mass spectrometry has pushed multielement single-particle analysis into practical, rapid workflows, with up to 16 elements measured in a single run across tap, river, and seawater samples.8 Time-of-flight instruments, which capture the full elemental makeup of each particle, are helping researchers distinguish engineered nanoparticles from naturally occurring ones in complex environmental samples.2 That distinction matters enormously for understanding the environmental fate of nanoscale titanium dioxide, silica, and other manufactured nanomaterials that regulators are only beginning to grapple with. What used to be an averaged, faintly abstract number is now a particle-by-particle census.
Laser-induced breakdown spectroscopy (LIBS) fires a laser pulse at a target, vaporizes a microscopic volume into plasma, and reads the elemental fingerprint in the emitted light, all without physical contact with the sample. That contactless, essentially prep-free quality has made it the standout technique for two very different frontiers. On Mars, the SuperCam instrument aboard the Perseverance rover has used LIBS to interrogate rock and soil chemistry from several meters away, with recent plasma-diagnostic work refining exactly how elemental abundances are extracted under Martian atmospheric conditions that differ sharply from those on Earth.4 Closer to home, LIBS combined with machine learning classifiers is being explored for distinguishing malignant from healthy tissue and for classifying calcified tissue composition, pointing toward a potential real-time, label-free aid to intraoperative tumor classification.9 Machine learning is doing much of the heavy lifting here: deep neural networks, transfer learning, and other machine learning methods are increasingly used to improve the reliability and accuracy of predictions from LIBS spectra, which are notoriously sensitive to matrix effects and poor shot-to-shot repeatability.10 The same combination is also being adapted for on-site food-safety screening, including rapid detection of toxic elements and pesticide residues without the days-long turnaround of conventional wet chemistry.11
Laser ablation coupled to ICP-MS (LA-ICP-MS) has matured from a bulk solid-sampling technique into a genuine imaging modality. By scanning a focused laser across a tissue section and recording elemental signal line by line, researchers can now build quantitative, spatially resolved maps of trace and ultratrace elements. Pairing laser ablation with time-of-flight mass analyzers, which can record a complete mass spectrum every 30 µs, has pushed acquisition speeds high enough for rapid multielement imaging at micrometer-scale pixel sizes, with laser spots as small as about 1 µm approaching the dimensions of individual cells.12 That resolution is opening applications from tracking metal-based drug distribution in tissue to mapping toxic or essential element deposition across tissue architecture, applications that were simply not achievable with earlier bulk-digestion approaches.
Two trends are converging on the instrument itself. The first is miniaturization: advances in transmission gratings, free-space optical coupling, batteries, displays, and smartphone-based data systems have produced handheld X-ray fluorescence (XRF) analyzers and other portable spectrometers that bring laboratory-style analysis into a device that fits in one hand, enabling real-time decisions at crime scenes, riverbeds, and other points of need rather than days later in a central lab.13 The second is intelligence: artificial intelligence and modern chemometric methods, including deep learning and explainable-AI frameworks, are increasingly embedded in sensor systems for real-time analytical decision support, automating feature extraction and nonlinear calibration.3 Coverage of 2025–2026 product launches shows embedded AI-based chemometric models, AI-assisted spectral search, and cloud connectivity appearing across a growing range of instruments, especially portable platforms.14 Meanwhile, isotope ratio analysis, long a niche capability, is being folded into broader authentication and provenance workflows, from tracing the geographic origin of food products to flagging adulteration, using stable isotope fingerprints that are very difficult to fake.15 Taken together, the instrument on the bench today is faster, smaller, and considerably better at explaining itself than the one from just a few years ago.
The common thread across these developments is resolution, in every sense of the word. Tandem mass spectrometry resolves interferences that used to be invisible obstacles. Single-particle ICP-MS resolves a population of nanoparticles into individual, countable events. LIBS and LA-ICP-MS resolve spatial detail down to a rock face on another planet or micrometer-scale features in a tissue section. And artificial intelligence resolves ambiguity in spectra that a human analyst would need far more time, and far more training data, to interpret confidently. None of these advances required a wholly new physical principle; atomic emission, absorption, and mass spectrometry rest on the same quantum mechanics they always have. What changed is the instrumentation, the software layered on top of it, and the willingness of the field to point these tools at problems—cancer diagnostics, Martian geochemistry, nanomaterial pollution—that would have seemed out of reach for atomic spectroscopy a decade ago.
Expect the next five years to push further on autonomy and integration rather than on entirely new detection principles. Fully autonomous analytical laboratories, cloud-based predictive maintenance, and real-time quality control already appear in recent product coverage as emerging trends,14 and reinforcement learning is being explored for adaptive calibration and autonomous spectral optimization.3 Single-particle and single-cell techniques will likely extend toward routine multiplexed analysis, tracking dozens of elements per particle or per cell rather than a handful. LIBS is poised to expand further into clinical and point-of-need settings as machine learning models become better validated and more explainable, a prerequisite for any regulatory acceptance in medicine.9 And as instruments keep shrinking while keeping pace with laboratory-grade performance, atomic spectroscopy will keep moving out of the central lab and into the field, the clinic, the delivery dock, and, quite literally, other planets. The recent story of atomic spectroscopy has been less about discovering new spectral lines and more about finally being able to see, count, and trust what those lines have been telling us all along.
(1) Evans, E. H.; Pisonero, J.; Smith, C. M. M.; et al. Atomic Spectrometry Update: Review of Advances in Atomic Spectrometry and Related Techniques. J. Anal. At. Spectrom. 2025, 40 (5), 1136–1157. DOI:
(2) Olesik, J. W. Single Particle ICP-MS: From Engineered Nanoparticles to Natural Nanoparticles. Spectroscopy 2020, 35 (7).
(3) Workman, J., Jr. Recent Research in Chemometrics and AI for Spectroscopy, Part I: Foundations, Definitions, and the Integration of Artificial Intelligence in Chemometric Analysis. Spectroscopy Online, November 3, 2025.
(4) Manelski, H. T.; Wiens, R. C.; Bousquet, B.; et al. LIBS Plasma Diagnostics with SuperCam on Mars: Implications for Quantification of Elemental Abundances. Spectrochim. Acta, Part B 2024, 222, 107061. DOI:
(5) McCurdy, E.; Woods, G. ICP–MS-MS Delivers Accurate Trace-Level Arsenic Analysis in Complex Samples. Spectroscopy 2012, 27 (s10).
(6) Cairns, W. R. L.; Braysher, E. C.; Butler, O. T.; et al. Atomic Spectrometry Update: Review of Advances in Environmental Analysis. J. Anal. At. Spectrom. 2026, 41 (1), 16–70. DOI:
(7) Thomas, R.; Stephan, C. Single-Particle ICP-MS: A Key Analytical Technique for Characterizing Nanoparticles. Spectroscopy 2017, 32 (3).
(8) Soffey, E.; Jones, C.; Kelinske, M. Rapid Multielement Nanoparticle Analysis Using Single-Particle ICP-MS/MS. Spectroscopy 2019, 34 (5).
(9) Dastageer, M. M.; Siraj, K.; Pedarnig, J. D.; et al. From Fundamentals of Laser-Induced Breakdown Spectroscopy to Recent Advancements in Cancer Detection and Calcified Tissues Analysis: An Overview (2015–2025). Molecules 2025, 30 (21), 4176. DOI:
(10) Wang, Z. New Trend: Application of Laser-Induced Breakdown Spectroscopy with Machine Learning. Chemosensors 2025, 13 (1), 5. DOI:
(11) Mehdizadeh, M.; Omidi, A.; Ikegwu, T. M.; et al. Laser-Induced Breakdown Spectroscopy (LIBS) for Rapid Food Safety Monitoring: Advancements, Applications, and Future Directions. Food Saf. Health 2026, 4 (1), 50–66. DOI:
(12) Tanner, M.; Bussweiler, Y.; Borovinskaya, O. Laser Ablation and Inductively Coupled Plasma–Time-of-Flight Mass Spectrometry—A Powerful Combination for High-Speed Multielemental Imaging on the Micrometer Scale. Spectroscopy 2017, 32 (5).
(13) Wetzel, W. The Rise of Portable and Handheld Spectroscopy. Spectroscopy Online, April 16, 2026.
(14) Workman, J., Jr. New Product Advances in Vibrational and Atomic Spectroscopy (2025–2026). Spectroscopy 2026, 41 (2), 16–21.
(15) Terro, C.; Modic, R.; Ogrinc, M.; et al. IsoFoodTrack: A Comprehensive Database and Management System Based on Stable Isotope Ratio Analysis for Combating Food Fraud. Front. Nutr. 2025, 12, 1516521. DOI:
