A commercial thermal imaging camera with a 600 × 600 pixel array and 20 mK thermal sensitivity is sufficient to reconstruct the complete internal thermal state and full-aperture wavefront errors of a large laser optic, according to results published in Classical and Quantum Gravity (L. Tao, P. Goodarzi, and J. Richardson, Class. Quantum Grav., 43, 14, 5008, 2026). Using this off-the-shelf hardware, the researchers determined absorbed power levels from individual heating sources with relative error below 0.1% and located the laser beam centroid to within 0.5 mm.

The problem the method addresses begins with coating absorption. Even the most tightly specified optical coatings absorb roughly 0.5 parts per million of incident radiation. At megawatt-scale circulating power, that fractional absorption deposits more than a watt of steady-state heating into each optic. The resulting temperature gradients drive thermo-elastic surface deformations in reflection and thermo-refractive substrate lensing in transmission — aberrations that directly limit system performance.

LIGO provides the driving application. The observatory aims to operate its 4-kilometer arm cavities at 1.5 MW of laser power in the coming years. Combined with squeezed-light quantum enhancement, the higher power reduces quantum noise and expands the accessible volume of the universe for gravitational-wave observation. Thermally induced aberrations work against this: they cause power losses and degrade the injected squeezed field, capping interferometer sensitivity.

The interferometers already correct low-order aberrations on their main 40-kilogram mirrors with ring heaters. A newly developed actuator, the FROnt Surface Type Irradiator (FROSTI), will project tailored annular heating patterns onto each mirror's periphery to address higher-order aberrations (L. Tao et al., Phys. Rev. Lett., 134, 5, 1401, 2025; T. Rosauer et al., Optica, 12, 10, 1569, 2025).

Deploying these actuators has hit a sensing gap. Hartmann wavefront sensors and phase cameras require a reference or probe beam, and the probe beam's transverse extent is practically limited. In LIGO, such sensors monitor only the central region of large optics and remain blind to wavefront errors near the edges — precisely where annular actuators such as FROSTI operate. The actuators lack a full-aperture error signal to set their power levels.

The new technique fills that gap with front-surface thermal imaging. A thermal camera captures the two-dimensional temperature profile across the entire optic surface. Calibration takes advantage of the region where the fields of view of LIGO's Hartmann sensors and the thermal camera overlap: the central-zone measurements convert raw thermal fluxes into a calibrated temperature map.

The physics behind the reconstruction is a linear superposition argument. In thermal steady state, the optic's surface temperature profile decomposes into a linear combination of the individual temperature responses generated by each distinct heating source — in LIGO's case, the main laser beam, the ring heater, and FROSTI. The method computes 2D spatial overlap integrals between the measured temperature map and finite element analysis (FEA) unit-heating maps for each source. Each integral yields the power absorbed from that source. Iterating over different beam positions jointly identifies the laser beam's centroid on the optic. With the complete thermal state pinned down, the FEA model reconstructs the actual wavefront errors in reflection and transmission over the full aperture.

One operational constraint matters: the technique requires thermal steady state. Optical systems can maintain that condition with an offline heating system that holds the optics' thermal state when the primary laser beam is absent. LIGO plans to incorporate such a system in a future upgrade.

The measured performance numbers — sub-0.1% relative error on absorbed power and 0.5 mm beam-position accuracy — come from the authors' demonstration with the 600 × 600, 20 mK camera; the FEA decomposition itself carries the accuracy claim under stated steady-state conditions, not under arbitrary transient operation.

The authors, led by Jonathan W. Richardson, associate professor at the University of California, Riverside, argue the method transfers broadly to any field running large optics under extreme thermal loads from high-power continuous-wave radiation. External thermal imaging replaces full-aperture wavefront sensor coverage, delivering the precision error signals that large-radius adaptive optics actuators need for continuous optimization of their heating profiles. The work was supported by the U.S. National Science Foundation under Award No. 2409496.

For facilities now qualifying thermal compensation systems, the development raises a practical adoption question: will steady-state operation and central-zone calibration — both prerequisites for this technique — fit their duty cycles, or does full-aperture thermal wavefront sensing remain confined to observatories like LIGO that can afford offline heating infrastructure?