A December 2025 demonstration (Morrison et al., arXiv:2512.08921) realized the first autonomously operating multi-ion optical clock in which every clock operation — cooling, state preparation, interrogation, and detection — is performed with light delivered by waveguides integrated in a surface-electrode ion trap, reaching a fractional frequency instability of 3.14(5)×10⁻¹⁴/√τ with four ¹⁷¹Yb⁺ ions. The demonstration, however, is limited to short-coherence Rabi interrogation of the 435 nm quadrupole clock transition, without sideband cooling, with ion lifetimes of about one minute during operation, and it attributes these limits to five hardware constraints: outgassing fiber feedthroughs and epoxy-coated acrylate fibers degrading the vacuum, roughly −27 dB total on-chip optical loss at 435 nm dominated by fiber-to-chip coupling, elevated motional heating (mean occupation n̄≈30 quanta) linked in part to photo-induced charging of the dielectric grating output windows, and control-electronics bandwidth and interlock limitations. The open problem is to reach long-coherence clock operation in this architecture — resolved-sideband cooling and Ramsey interrogation of the clock transition using on-chip-delivered light — the two capabilities the demonstration did not implement and traces to its optical-loss and motional-heating budgets.

Background

Optical atomic clocks lock a laser to a narrow atomic reference transition at optical frequencies; interrogating ensembles of N atoms in parallel reduces the quantum projection noise (the fundamental statistical scatter of projecting each atom's superposition onto a measurement outcome) and can improve clock stability by up to √N or, with interleaved probe durations, even exponentially in the number of ensembles (Rosenband and Leibrandt, arXiv:1303.6357). Trapped-ion clocks, in which a single laser-cooled ion is confined by radio-frequency (RF) electric fields, hold the best reported systematic uncertainties, and multi-ion operation is the natural route to better stability — provided ion–ion field shifts and motional effects are controlled. For ¹⁷¹Yb⁺ the reference is the electric-quadrupole (E2) transition ²S₁/₂→²D₅/₂ at 435 nm; interrogation is either by Rabi pulses (a single resonant pulse, limited in coherent duration by motional heating and the available optical power) or by Ramsey's method of separated oscillatory fields (two short π/2 pulses separated by a long dark time, which allows much longer coherence and hence narrower linewidths). Resolved-sideband cooling — driving the red motional sidebands of a narrow transition to bring the ion close to the motional ground state — raises contrast and coherence time but demands substantially more optical power at the ion than Doppler cooling alone. Surface-electrode traps confine ions tens of micrometres above a planar microfabricated chip; in integrated-photonic traps, the cooling, repump, and clock beams are delivered by waveguides fabricated in the trap chip and emitted toward the ion by diffraction gratings, replacing bulk free-space optics. This is essential for portable clocks and for scaling trapped-ion quantum computers (the QCCD, quantum charge-coupled device, architecture in which ions are shuttled between zones), but it introduces system-level difficulties that free-space systems avoid: fiber-to-chip edge coupling at violet wavelengths (369 nm and 435 nm for Yb⁺) is lossy; the packaging (fiber feedthroughs and epoxies) must be ultra-high-vacuum (UHV) compatible; ions sit close to dielectric grating windows that accumulate photo-induced charge and shift the ion off the RF null (the point of zero RF drive field, where excess micromotion and amplifier-noise sensitivity are minimized); and anomalous motional heating from nearby surfaces is aggravated — electric-field noise from dielectrics is a documented heating source (Teller et al., Phys. Rev. Lett. 126, 230505 (2021)). Prior demonstrations established component-level functionality: multi-site integrated optical addressing of Yb⁺ ions including Rabi spectroscopy of the E2 transition (Kwon et al., Nature Communications 15, 3709 (2024)); coherent control of a Ca⁺ quadrupole transition in a passively phase-stable standing wave emitted by integrated photonics (Vasquez et al., Phys. Rev. Lett. 130, 133201 (2023)); and a Ramsey sequence using integrated 729 nm light in a two-zone Ca⁺ QCCD device, where the optical-qubit Ramsey fringes were phase-scrambled and the sequence had to be mapped to ground-state (Zeeman) qubits to close with 0.989 contrast (Mordini et al., Phys. Rev. X 15, 011040 (2025)). Very recently, ground-state cooling via electromagnetically induced transparency (EIT) was demonstrated with integrated-photonics-delivered ultraviolet light in a Ca⁺ surface trap (Xing et al., arXiv:2512.10900 (2025)), and gold-coating exposed silicon in a (non-photonic) silicon trap chip was shown to suppress semiconductor charging and thereby re-enable sideband cooling (Chung et al., Quantum Science and Technology 10, 035014 (2025)). None of these, however, constitutes an optical clock operated with long-coherence Ramsey interrogation of the clock transition using on-chip-delivered light. Against this backdrop, Morrison et al. (arXiv:2512.08921) demonstrated the first autonomously operating multi-ion optical clock with entirely on-chip light delivery: four ¹⁷¹Yb⁺ ions in a room-temperature multi-site integrated-photonics surface trap, Rabi interrogation of the E2 transition at 435 nm, automated shuttling and reloading that overcame roughly one-minute ion lifetimes, and a short-term instability of 3.14(5)×10⁻¹⁴/√τ with more than two hours of continuous unattended operation. The authors explicitly identify the unresolved limitations that stand between this demonstration and a clock with useful accuracy, stability, and uptime: (i) vacuum and materials — tiny leaks in the fiber feedthroughs and out-gassing of the epoxy-coated acrylate fibers have degraded the vacuum and capped ion lifetimes, so UHV-compatible materials and feedthroughs are required; (ii) optical coupling — the 435 nm clock path suffers about −27 dB of total on-chip loss dominated by fiber-to-trap input coupling, so too little power reaches the ion for sideband cooling and short probe durations; (iii) motional heating and stray charging — the mean motional occupation is n̄≈30 quanta, attributed to imperfect ion placement on the RF null, RF/amplifier noise coupling, fabrication imperfections, and photo-induced charging of the dielectric grating windows by the 369 nm beam (with hours-timescale compensation-voltage drifts), for which transparent conductive coatings such as indium tin oxide are proposed but not yet implemented; (iv) interrogation methods — Ramsey interrogation and sideband cooling were not implemented and require the heating and optical-power improvements; and (v) control — the ion-presence reporting is limited to every 20 clock cycles by receiving-electronics communication, and software interlocks abort unattended runs when laser-unlock conditions cannot be discriminated. The authors present these as near-term improvements toward better clock performance; physically they are related, since available on-chip power sets the achievable cooling and probe durations, charging displaces ions from the RF null and thereby raises heating, and heating caps the coherent interrogation time that Ramsey operation exists to exploit.

Problem Statement

Enable long-coherence operation of the autonomous multi-ion ¹⁷¹Yb⁺ optical clock with on-chip integrated photonic light delivery: demonstrate, in a multi-site integrated-photonic surface-electrode trap in which all clock light is delivered by on-chip waveguides, (a) resolved-sideband cooling of the clock ions using on-chip-delivered light and (b) Ramsey (separated-oscillatory-field) interrogation of the 435 nm electric-quadrupole (E2) clock transition. These are the two capabilities the demonstrated clock did not implement: sideband cooling was time-prohibitive in part because of the low fiber-to-trap optical coupling (about −27 dB total on-chip loss at 435 nm), and the elevated motional heating rate (mean occupation n̄≈30 quanta) is what the authors identify as preventing higher clock resolution with the Ramsey method. The limitations acknowledged by Morrison et al. (arXiv:2512.08921) identify the subsystems through which progress is expected, without prescribing particular fixes: improved fiber-to-chip edge coupling, which the authors note would allow sideband cooling and increase both contrast and coherence time; reduced motional heating, including mitigation of the photo-induced charging of the dielectric grating windows (the authors suggest transparent conductive coatings such as indium tin oxide) and placement of ions on the RF null while preserving beam overlap; and UHV-compatible fiber feedthroughs and fiber materials to stop the vacuum degradation that caps ion lifetimes at roughly one minute during clock operation — a lifetime the authors describe as too short to operate as an optical clock, and whose improvement they note would reduce collisional decoherence and allow longer Ramsey sequences that lower quantum projection noise.

References

  1. Morrison, Tharon D.; Kwon, Joonhyuk; Delaney, Matthew A.; Gehl, Michael; Leibrandt, David R.; Stick, Daniel; McGuinness, Hayden J. (2025). Autonomous multi-ion optical clock with on-chip integrated photonic light delivery. arXiv:2512.08921, DOI 10.48550/arXiv.2512.08921, https://arxiv.org/abs/2512.08921
  2. Kwon, Joonhyuk; Setzer, William J.; Gehl, Michael; Karl, Nicholas; Van Der Wall, Jay; Law, Ryan; Blain, Matthew G.; Stick, Daniel; McGuinness, Hayden J. (2024). Multi-site integrated optical addressing of trapped ions. Nature Communications 15, 3709, DOI 10.1038/s41467-024-47882-5; arXiv:2308.14918, https://arxiv.org/abs/2308.14918
  3. Mordini, Carmelo; Vasquez, Alfredo Ricci; Motohashi, Yuto; Müller, Mose; Malinowski, Maciej; Zhang, Chi; Mehta, Karan K.; Kienzler, Daniel; Home, Jonathan P. (2025). Multizone trapped-ion qubit control in an integrated photonics QCCD device. Physical Review X 15, 011040, DOI 10.1103/PhysRevX.15.011040; arXiv:2401.18056, https://arxiv.org/abs/2401.18056
  4. Vasquez, Alfredo Ricci; Mordini, Carmelo; Vérnière, Chloé; Stadler, Martin; Malinowski, Maciej; Zhang, Chi; Kienzler, Daniel; Mehta, Karan K.; Home, Jonathan P. (2023). Control of an atomic quadrupole transition in a phase-stable standing wave. Physical Review Letters 130, 133201, DOI 10.1103/PhysRevLett.130.133201; arXiv:2210.02597, https://arxiv.org/abs/2210.02597
  5. Xing, Zhenzhong; Rivy, Hamim Mahmud; Natarajan, Vighnesh; Kolhatkar, Aditya Milind; Beck, Gillenhaal; Mehta, Karan K. (2025). Rapid multi-mode trapped-ion laser cooling in a phase-stable standing wave. arXiv:2512.10900, https://arxiv.org/abs/2512.10900
  6. Teller, Markus; Fioretto, Dario A.; Holz, Philip C.; Schindler, Philipp; Messerer, Viktor; Schüppert, Klemens; Zou, Yueyang; Blatt, Rainer; Chiaverini, John; Sage, Jeremy; Northup, Tracy E. (2021). Heating of a trapped ion induced by dielectric materials. Physical Review Letters 126, 230505, DOI 10.1103/PhysRevLett.126.230505; arXiv:2103.13846, https://arxiv.org/abs/2103.13846
  7. Chung, Daun; Choi, Kwangyeul; Lee, Woojun; Kim, Chiyoon; Shon, Hosung; Park, Jeonghyun; Cho, Beomgeun; Lee, Kyungmin; Kim, Suhan; Yoo, Seungwoo; Jung, Eui Hwan; Jung, Changhyun; Kang, Jiyong; Kim, Kyunghye; Berkis, Roberts; Northup, Tracy; Cho, Dong-Il Dan; Kim, Taehyun (2025). A silicon-based ion trap chip protected from semiconductor charging. Quantum Science and Technology 10, 035014, DOI 10.1088/2058-9565/add04c; arXiv:2411.13955, https://arxiv.org/abs/2411.13955
  8. Grimpe, Carl-Frederik; Lüßmann-Sorokina, Anastasiia; Du, Guochun; Sah, Pragya; Sauer, Steffen; Jordan, Elena; Thomas, Rijil; Gehrmann, Pascal; Lipkin, Maksim; Suckow, Stephan; Lemme, Max C.; Kroker, Stefanie; Mehlstäubler, Tanja E. (2026). Benchmarking dual-polarization silicon nitride photonic integrated circuits for trapped-ion quantum technologies. Journal of the European Optical Society – Rapid Publications 22, 56, DOI 10.1051/jeos/2026049; arXiv:2603.22578, https://arxiv.org/abs/2603.22578
  9. Rosenband, T.; Leibrandt, D. R. (2013). Exponential scaling of clock stability with atom number. arXiv:1303.6357, https://arxiv.org/abs/1303.6357