arXiv:2610.01693
Predicting electric-field noise in ion traps using fluctuation electrodynamics
Markus Teller, Da An, Alberto M. Alonso, Philip C. Holz, Philipp Schindler, Hartmut Häffner, Tracy E. Northup
Abstract
Electric-field noise in microfabricated surface ion traps contributes to gate errors in trapped-ion quantum computers, but this noise has been challenging to predict. Here, we present a method to predict the electric-field noise arising from fluctuations in the bulk of dielectric and metallic materials. The method is valid for arbitrary trap geometries, and the only relevant material property is energy loss. We apply it to an ion trap with an electrically floating electrode and find the contribution from dielectrics to be dominant. Our model predicts the observed noise for this trap within a factor of four. In addition, we apply our model to typical surface trap designs and analyze how the electrode structures shield the noise.
AI Summary
Electric-field noise limits motional coherence and gate fidelity in surface ion traps. This paper addresses how to predict the contribution from bulk dielectric and metallic losses, including shielding by electrodes, for arbitrary trap geometries. The method uses frequency-domain finite-element solutions of Maxwell’s equations for an oscillating test-charge dipole. Dissipated power in each material is converted into electric-field noise through the fluctuation–dissipation theorem, then into ion heating rates. It extends an earlier electrostatic dielectric calculation to include conducting materials and time-dependent shielding. For a room-temperature surface trap containing a floating electrode, measured $^{40}\mathrm{Ca}^{+}$ heating rates are compared with simulations over ion–surface distances of $60$–$100\,\mu\mathrm{m}$. Using a literature dielectric loss tangent of $10^{-3}$, predictions are approximately four times below measurements. Fitting the loss tangent gives $4.0(1)\times10^{-3}$ and reduced $\chi^2=0.89$, reproducing the distance dependence. The floating electrode’s metallic-loss contribution is below $2.2$ quanta/s; dielectric noise dominates. Grounding that electrode is predicted to reduce heating by a factor of 200 at $100\,\mu\mathrm{m}$. Separate geometry sweeps quantify how narrower trenches, undercuts, and sufficient trench depth suppress dielectric noise, with different crossovers to metal-dominated noise for different motional axes. The calculation addresses bulk thermal noise, not a complete model of surface contaminants or other heating sources, and therefore supplies a lower bound. Quantitative predictions depend on sample-specific material losses; the experimental substrate’s loss tangent was not independently measured, and fabrication-related changes or solvent residues are proposed explanations for the discrepancy.
Error found: The abstract overstates the material-input simplification: predicting the fields requires the real dielectric permittivity in addition to dissipative material properties. The printed frequency-normalization equation has its ratio inverted, although Table I uses the correct ratio and therefore does not exhibit the corresponding numerical error. The absolute-noise comparison depends on an unmeasured substrate loss tangent, and fitting that parameter establishes consistency rather than independent validation of the predicted amplitude.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.