Is the Moving Mirror Obsolete? Six Innovations Redefining FT-IR Instrumentation

Abstract

For sixty years, Fourier-transform infrared (FT-IR) spectroscopy has depended on a single mechanical compromise: a mirror sliding back and forth inside a Michelson interferometer. That compromise is now being challenged on every front. Frequency-comb lasers are replacing the moving mirror entirely, delivering resolution and speed once reserved for atomic clocks. Photothermal probes are shattering the diffraction limit that has confined infrared microscopy to blurry, micrometer-scale images. Silicon photonics is shrinking the interferometer itself onto a fingernail-sized chip, while artificial intelligence is teaching instruments to read spectra faster and more reliably than trained spectroscopists. None of these advances make FT-IR obsolete, but together they are rewriting what the technique can measure, where it can travel, and how small it can see.

Introduction

Fourier-transform infrared spectroscopy has spent six decades as one of chemistry's most dependable workhorses, built on the same core architecture that Michelson interferometry made possible in the 1950s and 1960s: a beamsplitter, a fixed mirror, and one mirror that physically moves back and forth to generate an interferogram. That architecture has proven remarkably durable, but it has also imposed hard limits on speed, spatial resolution, and portability. Over the past several years, those limits have started to give way. A cluster of separate technologies spanning laser physics, nanoscale probes, silicon photonics, and machine learning are converging to reshape what an infrared instrument can be.¹ The following six developments represent the most consequential shifts now reaching research laboratories and, in some cases, factory floors.

From Moving Mirrors to Light Combs: Dual-Comb Spectroscopy

The most direct challenge to the classic Michelson design comes from dual-comb spectroscopy, which eliminates the moving mirror altogether. Two mutually coherent frequency combs, each a laser emitting thousands of precisely spaced spectral lines, are mixed together to generate a radio-frequency interferogram without any mechanical scanning.² Because both combs can be built from quantum cascade lasers operating directly in the mid-infrared fingerprint region, the resulting spectrometers dispense with the cryogenically cooled detectors and bulky external cavities that older tunable-laser systems required. The performance gap relative to conventional FT-IR is substantial: dual-comb systems have demonstrated spectral resolution improved by roughly four orders of magnitude, down to the megahertz level, while cutting acquisition times from tens of seconds to well under a millisecond.² That combination of speed and precision opens applications that were previously impractical for infrared absorption measurements, including tracking transient reaction intermediates, monitoring combustion chemistry in real time, and performing open-air trace-gas sensing from a mobile platform. Because the technique replaces mechanical scanning with frequency-referenced electronics, instrument stability and reproducibility improve as a side effect, addressing a chronic pain point in calibration transfer between FT-IR units.

Breaking the Diffraction Barrier: Optical Photothermal Infrared (O-PTIR)

Conventional infrared microscopy has always been hostage to diffraction: because mid-infrared wavelengths run from roughly 2.5 to 25 micrometers, spatial resolution is capped at several micrometers no matter how good the optics are. Optical photothermal infrared (O-PTIR) spectroscopy sidesteps that limit by using a pulsed, tunable infrared laser to heat a sample locally and a co-aligned visible probe beam, rather than the infrared beam itself, to detect the resulting thermal expansion.³ Because spatial resolution is now set by the shorter visible wavelength rather than the infrared one, O-PTIR delivers roughly 30 times better resolution than conventional FT-IR or discrete-frequency infrared imaging, reaching spot sizes near 450 nanometers while preserving spectra that overlay closely with standard transmission FT-IR references.⁴ The technique also collects Raman scattering from the same visible probe simultaneously, so a single measurement yields co-located, submicron-resolved infrared and Raman spectra, a pairing that has proven valuable for identifying microplastics, characterizing pharmaceutical contaminants, and mapping protein aggregates inside individual cells without the extensive sample preparation that transmission FT-IR microscopy typically demands.⁵ For failure-analysis and life-science laboratories that have struggled to resolve chemically heterogeneous features below a few micrometers, O-PTIR effectively hands FT-IR-quality spectra to a confocal microscope.

Chemistry at the Nanoscale: AFM-IR

Where O-PTIR pushes resolution into the sub-micrometer range, atomic force microscopy–infrared spectroscopy (AFM-IR) goes further still, using the sharp tip of an AFM cantilever as the infrared absorption sensor itself. As a pulsed infrared source heats an absorbing region of a sample, the resulting thermal expansion launches a ringdown oscillation in the cantilever that is captured and Fourier-analyzed to reconstruct a local absorption spectrum, achieving chemical mapping down to roughly 10 nanometers, two to three orders of magnitude finer than diffraction-limited FT-IR.⁶ Recent development work has extended AFM-IR from polymer science into life-science applications, including subcellular protein-conformation mapping, biomineral characterization, and single-particle chemical imaging, while tapping-mode implementations have improved sensitivity enough to detect monolayer-level surface chemistry.⁶ For materials scientists chasing nanoscale heterogeneity, buried interfaces, and single-particle contamination, AFM-IR increasingly functions as the infrared complement to electron microscopy.

FT-IR on a Chip: MEMS and Silicon Photonics

Perhaps the least visible but most disruptive change is happening at the level of the interferometer itself. Silicon photonics and microelectromechanical systems (MEMS) fabrication are shrinking the Michelson interferometer from a benchtop instrument into a device that fits in the palm of a hand, and eventually onto a fingernail-sized chip. In one demonstrated approach, the moving mirror is replaced by an electrostatically actuated MEMS platform capable of ±250-micrometer displacement, monolithically aligned during fabrication so that no manual optical alignment is required, a step that has historically limited FT-IR to trained operators.⁷ A related computational-spectrometer architecture forgoes a moving element entirely, instead cycling a bank of individually addressable, on/off MEMS-tunable waveguide couplers across a mid-infrared silicon photonic chip and reconstructing the spectrum digitally, an approach that is inherently insensitive to voltage drift and well suited to mass production.⁷ Because these chip-scale instruments consume little power and require no optical realignment, they are being positioned for downhole chemical sensing, gas-leak detection, point-of-care diagnostics, and other field-deployable roles that a full-size benchtop FT-IR could never reach.

Teaching Infrared to Think: AI and Deep Learning

Hardware is only half the story. Across the field, artificial intelligence and deep learning are being layered onto infrared instruments to automate tasks that once required an experienced spectroscopist: baseline correction, peak assignment, spectral-library search, and multivariate calibration.⁸ A convolutional neural network trained on FT-IR spectra, for example, has been shown to classify 17 functional groups and 72 vibrational coupling modes directly from raw spectra, collapsing a task that traditionally demanded manual interpretation into a near-instantaneous prediction.⁹ Similar deep-learning frameworks are now being applied to strip baseline distortion and etalon artifacts out of dual-comb absorption spectra, improving quantitative accuracy in gas-sensing applications without additional hardware.⁹ The broader implication is that infrared instruments are shifting from passive data generators into active interpreters, a change that is particularly consequential for portable and field-deployed FT-IR systems, where an on-board model, rather than a remote expert, must make the call in real time.

The New Instrument Shelf: Commercial Platforms Catch Up

Instrument manufacturers have moved quickly to commercialize pieces of this shift. Recent FT-IR product cycles have emphasized vacuum optics and sub-0.1 cm⁻¹ resolution for combined mid- and far-infrared measurement, cloud-connected software with AI-assisted spectral search, compact mid-wave sensors with an order-of-magnitude gain in responsivity, and high-sensitivity attenuated total reflectance microscopy platforms aimed at pharmaceutical and polymer quality control.¹⁰ These releases indicate that the laboratory-bench FT-IR is absorbing many of the same pressures driving the more exotic technologies described above: manufacturers are racing to add automation, connectivity, and higher sensitivity even as chip-scale and photothermal alternatives chip away at applications the conventional instrument once owned outright.¹¹ The result is a broader, more stratified FT-IR ecosystem than existed five years ago, spanning benchtop research-grade systems, cloud-connected quality-control instruments, and pocket-sized field sensors.

Summary and Conclusions

Taken together, these six developments do not represent a single breakthrough so much as a coordinated erosion of every constraint that has historically defined FT-IR: the moving mirror, the diffraction limit, the benchtop footprint, and the dependence on expert interpretation. Dual-comb sources and MEMS photonics attack the interferometer itself; O-PTIR and AFM-IR attack the resolution ceiling; and artificial intelligence attacks the interpretation bottleneck.^1,8 None of these approaches has fully displaced conventional FT-IR, and for routine bulk analysis the classic Michelson-based bench instrument remains the most cost-effective choice. But for applications that demand sub-micron resolution, field portability, or real-time gas-phase kinetics, the instrumentation landscape now looks meaningfully different than it did five years ago.

Future Outlook

Expect the boundaries between these technologies to blur further. Chip-scale photonic spectrometers are likely to absorb AI-based spectral reconstruction directly on-chip, while O-PTIR and AFM-IR platforms continue to add simultaneous Raman and fluorescence channels that turn a single measurement into a multimodal chemical fingerprint.^4,6 Dual-comb architectures, still largely confined to specialized research and industrial gas-sensing applications, should become more affordable as quantum cascade laser comb sources mature and displace bulkier external-cavity designs.² Meanwhile, as chemometric and machine-learning tools mature alongside these hardware gains, infrared instruments are likely to increasingly function as autonomous sensing nodes within larger digital manufacturing and diagnostic networks rather than as standalone laboratory devices.^8,10

References

1. Workman, J., Jr. The Most Important Vibrational Spectroscopy Trends of 2025. Spectroscopy, Dec 8, 2025. https://www.spectroscopyonline.com/view/the-most-important-vibrational-spectroscopy-trends-of-2025 (accessed 2026-07-31).
2. Hayden, J.; Geiser, M.; Gianella, M.; Horvath, R.; Hugi, A.; Sterczewski, L.; Mangold, M. Mid-Infrared Dual-Comb Spectroscopy with Quantum Cascade Lasers. APL Photonics 2024, 9 (3), 031101. DOI: 10.1063/5.0159042.
3. Prater, C. B.; Kansiz, M.; Cheng, J.-X. A Tutorial on Optical Photothermal Infrared (O-PTIR) Microscopy. APL Photonics 2024, 9 (9). DOI: 10.1063/5.0219983.
4. Prater, C.; Cheng, J.-X.; Kansiz, M. An Introduction to Optical Photothermal Infrared (O-PTIR) Spectroscopy. Spectroscopy, Sept 22, 2025 (updated Feb 23, 2026). https://www.spectroscopyonline.com/view/an-introduction-to-optical-photothermal-infrared-o-ptir-spectroscopy (accessed 2026-07-31).
5. Prater, C.; Bai, Y.; Konings, S. C.; Martinsson, I.; Swaminathan, V. S.; Nordenfelt, P.; Gouras, G.; Borondics, F.; Klementieva, O. Fluorescently Guided Optical Photothermal Infrared Microspectroscopy for Protein-Specific Bioimaging at Subcellular Level. J. Med. Chem. 2023, 66 (4), 2542–2549. DOI: 10.1021/acs.jmedchem.2c01359.
6. dos Santos, A. C. V. D.; Hondl, N.; Ramos-Garcia, V.; Kuligowski, J.; Lendl, B.; Ramer, G. AFM-IR for Nanoscale Chemical Characterization in Life Sciences: Recent Developments and Future Directions. ACS Meas. Sci. Au 2023, 3 (5). DOI: 10.1021/acsmeasuresciau.3c00010.
7. Qiao, Q.; Liu, X.; Ren, Z.; Dong, B.; Xia, J.; Sun, H.; Lee, C.; Zhou, G. MEMS-Enabled On-Chip Computational Mid-Infrared Spectrometer Using Silicon Photonics. ACS Photonics 2022, 9 (7), 2367–2377. DOI: 10.1021/acsphotonics.2c00381.
8. Workman, J., Jr. AI, Deep Learning, and Machine Learning in the Dynamic World of Spectroscopy. Spectroscopy, Dec 2, 2024. https://www.spectroscopyonline.com/view/ai-deep-learning-and-machine-learning-in-the-dynamic-world-of-spectroscopy (accessed 2026-07-31).
9. Workman, J., Jr. AI-Based Neural Networks Revolutionize Infrared Spectra Analysis. Spectroscopy, 2024. https://www.spectroscopyonline.com/view/ai-based-neural-networks-revolutionize-infrared-spectra-analysis (accessed 2026-07-31).
10. Workman, J., Jr. New Product Advances in Vibrational and Atomic Spectroscopy (2025–2026). Spectroscopy 2026, 41 (2), 16–21. DOI: 10.56530/spectroscopy.mm4170h4.
11. Miseo, E. V. Review of Spectroscopic Instrumentation. Spectroscopy 2025, 40 (4), 26–31. DOI: 10.56530/spectroscopy.nx9188m9.