🔬 atomic physics

Inverse identification of thermal-radiative parameters for reduced effective-temperature uncertainty in an ion optical clock

This paper presents a hybrid inversion framework combining infrared thermography, finite-element modeling, and surrogate-based optimization to accurately identify thermal-radiative parameters, thereby reducing the effective-temperature uncertainty and the associated blackbody-radiation shift in room-temperature single-ion optical clocks to a frequency-shift uncertainty of 1.647 × 10⁻¹⁹.

Original authors: Jie Zhao, Yani Zuo, Shaoyang Dai, Chaowei Wang, Ang Huang, Jian Song, Guangyao Wang, Fangfei Wu, Haotian Shi, Zhengyong Li, Fang Fang

Published 2026-10-14

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Jie Zhao, Yani Zuo, Shaoyang Dai, Chaowei Wang, Ang Huang, Jian Song, Guangyao Wang, Fangfei Wu, Haotian Shi, Zhengyong Li, Fang Fang

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Time, as we measure it with the most precise instruments on Earth, is not a steady river but a series of incredibly rapid ticks. These ticks come from atoms, specifically ions trapped in a vacuum and held in place by invisible electric fields. When scientists look at these atoms, they see a frequency so stable that it could define the second itself, replacing the current standard based on cesium atoms. However, even in a vacuum, these atoms are not isolated from their surroundings. They are bathed in a faint, invisible glow of heat radiation emitted by everything around them, including the walls of the chamber and the metal parts of the trap holding the atom. This glow, known as blackbody radiation, pushes on the atom ever so slightly, changing the speed of its tick. For the most advanced clocks, this tiny push is the largest source of error, a whisper of noise that must be silenced to hear the true time.

To silence this noise, scientists must know exactly how hot the atom feels. It is not enough to know the temperature of the room; the atom sits in a complex environment where metal parts heat up differently due to the electric fields holding it. Some parts might be slightly warmer than others, and the way these parts radiate heat depends on their surface texture and material. If a scientist guesses the temperature wrong by even a fraction of a degree, the clock's timekeeping will drift. The challenge is that the atom is trapped in a place where you cannot stick a thermometer without destroying the electric field that holds it. You cannot simply touch the trap to measure its heat, nor can you easily see the heat radiating from every tiny corner of the metal structure.

In a recent study, a team of researchers developed a new way to solve this puzzle without ever touching the trap. They built a detailed digital twin of the ion trap, a computer model that simulates how heat moves through the metal and ceramic parts. But a computer model is only as good as the numbers fed into it, and many of those numbers—such as how well the metal surfaces emit heat—are difficult to know precisely for a specific, assembled device. Instead of guessing these values, the researchers used a clever trick involving infrared cameras and a mathematical process called inverse identification. They took pictures of the trap with an infrared camera, which sees heat as light, while the trap was running and while it was off. By comparing the difference in the heat patterns, they could see exactly how much the electric fields warmed the trap.

The researchers then used a sophisticated computer program to work backward from these heat pictures. The program adjusted the unknown numbers in their digital model—such as the heat-emitting properties of the metal surfaces and the electrical resistance of the insulating parts—until the model's predicted heat pattern matched the real pictures taken by the camera. To make this process fast enough to be practical, they created a streamlined version of their complex computer model that could run thousands of times in the blink of an eye, learning the relationship between the material properties and the resulting heat. This allowed them to find the exact set of material properties that explained the real-world heat they observed.

The result was a highly accurate map of the temperature at the exact spot where the atom sits. The team found that the electric fields caused the effective temperature at the atom's location to rise by about 0.43 degrees Celsius above the surrounding environment. More importantly, they could state with great confidence that this measurement was accurate to within 0.155 degrees Celsius. This level of precision is significant because it reduces the uncertainty in the clock's timekeeping caused by heat radiation to a tiny fraction of a billionth of a billionth. When combined with other known factors, the total uncertainty of the clock's frequency shift due to heat is now calculated to be roughly 2.85 times 10 to the power of negative 19.

This achievement matters because it proves that scientists can now determine the thermal environment of a trapped ion with high precision without needing to install sensors inside the vacuum. The method relies on observing the heat from the outside and using physics to deduce what is happening inside. The researchers also discovered that if they had simply guessed the material properties instead of measuring them, the error in their temperature calculation could have been large enough to ruin the clock's accuracy. By letting the real heat patterns guide their model, they removed the guesswork. This approach offers a reliable path forward for building even more precise optical clocks, ensuring that the next generation of timekeepers can keep time with a stability that was previously out of reach.

Technical Summary: Inverse Identification of Thermophysical Parameters for Reduced Effective-Temperature Uncertainty in an Ion Optical Clock

Problem Statement
The blackbody radiation (BBR) shift remains the leading systematic uncertainty for room-temperature single-ion optical clocks. Accurate evaluation of this shift requires determining the effective temperature ( $T_{eff}$ ) at the ion's position. However, $T_{eff}$ is difficult to measure directly due to RF-induced heating, spatially varying surface emissivities, low emissivity of metallic trap components, and the inaccessibility of trap surfaces to direct contact thermometry. Existing methods often rely on literature values for thermophysical parameters (e.g., emissivity, dielectric loss) or discrete point measurements, which propagate significant uncertainty into the $T_{eff}$ evaluation. For the $^{171}\text{Yb}^+$ E3 transition, reducing the temperature uncertainty to approximately 0.10 K is necessary to lower the temperature-related BBR uncertainty contribution to the $10^{-19}$ level, making it comparable to or smaller than the atomic-response contribution.

Methodology
The authors developed a hybrid inversion framework that integrates calibrated broadband infrared thermography with a finite-element (FE) model and a reduced-order surrogate to identify uncertain thermophysical parameters.

Experimental Setup and Radiometry:
- An end-cap $^{171}\text{Yb}^+$ ion trap was operated at 20.408 MHz.
- Infrared thermograms were acquired using a calibrated FLIR T1040 camera through a ZnSe viewport.
- Measurements were taken in three states: RF-off (baseline), 500 V (for parameter identification), and 485 V (for independent validation).
- To constrain the inversion, specific regions of the trap were black-painted (emissivity $\epsilon \approx 0.82$ ) to serve as high-emissivity radiometric anchors, while metallic regions (Cu, Au, Mo) provided sensitivity to unknown emissivities.
- Radiance changes ( $\Delta L$ ) were calculated by subtracting RF-off from RF-on thermograms, correcting for reflected background radiation using a silver mirror measurement.
Coupled Physical and Surrogate Models:
- A coupled electromagnetic-thermal FE model simulated the RF-induced temperature rise, accounting for Joule heating in conductors and dielectric losses in insulators (alumina, fused silica).
- Because iterative parameter identification requires thousands of forward calculations, a sensitivity-enriched reduced-order surrogate was trained on 5,400 FE solutions.
- The surrogate combines a quadratic regression baseline with a residual multilayer perceptron (MLP) and a fixed spatial basis (Proper Orthogonal Decomposition) that includes modes specifically enriched with parameter sensitivity information. This ensures the surrogate preserves weak sensitivity directions critical for inversion while maintaining millisecond evaluation speeds.
Inverse Identification:
- The inverse problem was formulated in the radiance domain to decouple emissivity from temperature.
- A trust-region reflective least-squares optimization identified five unknown thermophysical parameters: emissivities of stainless steel ( $\epsilon_{SS}$ ), copper ( $\epsilon_{Cu}$ ), and gold ( $\epsilon_{Au}$ ), and dielectric loss tangents for alumina ( $\tan \delta_{Al_2O_3}$ ) and fused silica ( $\tan \delta_{FS}$ ).
- The identified parameters were then "frozen" and reintroduced into the full FE model to calculate the final $T_{eff}$ and its uncertainty.

Key Contributions

Radiance-Domain Inversion: The paper introduces a method to identify thermophysical parameters directly from spatially resolved infrared radiance data, rather than relying on literature values or discrete point sensors.
Sensitivity-Enriched Surrogate: The development of a reduced-order surrogate that explicitly preserves parameter sensitivity directions allows for efficient and accurate inverse identification without sacrificing the fidelity of the underlying physics.
Hybrid Constraint Strategy: The use of black-painted reference regions provides fixed emissivity anchors that break the degeneracy between heating parameters and unknown metallic emissivities.
Cross-Condition Validation: The identified parameter set was validated against an independent dataset at a lower RF voltage (485 V) without refitting, demonstrating the robustness and transferability of the identified parameters.

Results

Identified Parameters: The inversion yielded specific values for the thermophysical parameters (e.g., $\epsilon_{SS} = 0.2552$ , $\epsilon_{Cu} = 0.09358$ , $\tan \delta_{Al_2O_3} = 1.059 \times 10^{-4}$ ), which differed from nominal literature values, highlighting the necessity of experimental identification for assembled traps.
Effective Temperature: The RF-induced effective-temperature rise at the ion position was determined to be $\Delta T_{eff} = 0.4345$ K.
Uncertainty Budget: The combined standard uncertainty for $T_{eff}$ $T_{e f f}$ was evaluated as $u_c(T_{eff}) = 155.0$ $u_{c} (T_{e f f}) = 155.0$ mK.
- The dominant contributions arose from the transfer of thermal boundary conditions between the characterization and operational systems (Route III, ~59% of variance) and the copper-rod thermal boundary measurement (Route I, ~27%).
- The inverse identification process itself contributed ~16% to the total variance.
BBR Shift Impact: The temperature-related fractional BBR uncertainty was calculated as $1.647 \times 10^{-19}$ . When combined with the atomic-response uncertainty ( $2.326 \times 10^{-19}$ ), the total fractional BBR uncertainty is $2.850 \times 10^{-19}$ . The RF-induced change in the fractional BBR shift is $-4.607 \times 10^{-19}$ , which exceeds the combined uncertainty.

Significance and Claims
The paper claims that this framework provides a reproducible and traceable route for evaluating $T_{eff}$ in room-temperature ion traps without requiring in-vacuum temperature sensors or dedicated cryogenic designs. By replacing poorly constrained prior parameter values with experimentally identified values, the method reduces the reliance on literature data. The authors note that while the current uncertainty is dominated by the transfer of thermal boundary conditions between the test and operational setups, the radiance-domain inversion successfully constrains the model parameters to a level where residual non-uniqueness (demonstrated by a 10.67 mK difference in $T_{eff}$ between alternative solutions) is the limiting factor of the identification process itself. The method offers a generalizable approach for suppressing BBR-limited uncertainty budgets in optical clocks.

More like this