---
title: Clock-Line-Mediated Sisyphus Cooling
url: https://www.emergentmind.com/topics/clock-line-mediated-sisyphus-cooling
type: topic
---

# Clock-Line-Mediated Sisyphus Cooling

Clock-line-mediated Sisyphus cooling is a dissipative laser-cooling scheme in which the ultranarrow optical clock transition is used as the entry channel into a spatially structured cooling cycle. In the realization demonstrated with alkaline-earth-like \(^{171}\mathrm{Yb}\), atoms are excited on the \(^{1}\mathrm{S}_{0}\rightarrow{}^{3}\mathrm{P}_{0}\) clock line at \(578~\mathrm{nm}\), while a near-resonant \(1388~\mathrm{nm}\) standing wave on \(^{3}\mathrm{P}_{0}\rightarrow{}^{3}\mathrm{D}_{1}\) creates a spatially periodic light shift of the \(^{3}\mathrm{P}_{0}\) clock state. The resulting position-dependent excitation and decay sequence implements a Sisyphus cycle that reaches sub-recoil temperatures, enhances loading into a \(759~\mathrm{nm}\) magic-wavelength one-dimensional optical lattice, and can be operated in pulsed or continuous form [2406.13782].

## 1. Definition and scope

Clock-line-mediated Sisyphus cooling denotes a specific subset of Sisyphus-cooling protocols: the coherent excitation step occurs on the optical clock transition itself, while dissipation is supplied through a higher-lying dipole-allowed transition that dresses the clock state and creates a spatially varying potential. Related experiments in strontium and lithium use the same broad Sisyphus principle—conversion of kinetic energy into potential energy followed by irreversible spontaneous decay—but they do not all use the clock transition as the mediating line [2406.13782] [2506.19701] [1811.06014] [1308.1935].

| Scheme | Mediating transition | Distinctive feature |
|---|---|---|
| Clock-line-mediated Sisyphus cooling in \(^{171}\mathrm{Yb}\) | \(^{1}\mathrm{S}_{0}\rightarrow{}^{3}\mathrm{P}_{0}\) clock line | \(1388~\mathrm{nm}\) standing wave dresses \(^{3}\mathrm{P}_{0}\) |
| Narrow-line-mediated Sisyphus cooling in metastable \(^{88}\mathrm{Sr}\) | \(^{3}\mathrm{P}_{2}\rightarrow{}^{3}\mathrm{D}_{3}\) at \(2.92~\mu\mathrm{m}\) | Non-clock narrow line in metastable manifold |
| Off-magic Sisyphus cooling in clock-magic Sr tweezers | \(^{1}\mathrm{S}_{0}\leftrightarrow{}^{3}\mathrm{P}_{1}\) at \(689~\mathrm{nm}\) | Cooling compatible with \(813.4~\mathrm{nm}\) clock-magic trapping |
| Polarization-gradient Sisyphus cooling in lithium | D1 and D2 lines | Not a true clock-transition scheme |

This terminology matters because several distinct architectures are often grouped together. A common misconception is that any Sisyphus cooling performed in a clock-magic trap is automatically clock-line-mediated. The strontium tweezer work at \(813.4~\mathrm{nm}\) is a counterexample: the trap is magic for \(^{1}\mathrm{S}_{0}\leftrightarrow{}^{3}\mathrm{P}_{0}\), but the cooling occurs on \(^{1}\mathrm{S}_{0}\leftrightarrow{}^{3}\mathrm{P}_{1}\), not on the clock transition [1811.06014]. Likewise, the metastable-strontium experiment uses a narrow infrared line and reproduces the same qualitative Sisyphus physics without involving the clock line [2506.19701].

## 2. Dissipative mechanism on the clock transition

The defining mechanism is a dressed-state Sisyphus process built around the long-lived clock state. In \(^{171}\mathrm{Yb}\), atoms are first excited from the nearly unperturbed \(^{1}\mathrm{S}_{0}\) ground state to \(^{3}\mathrm{P}_{0}\) by the \(578~\mathrm{nm}\) clock laser. A near-resonant \(1388~\mathrm{nm}\) standing wave on \(^{3}\mathrm{P}_{0}\rightarrow{}^{3}\mathrm{D}_{1}\) then produces a spatially periodic light shift of \(^{3}\mathrm{P}_{0}\), so that atoms entering the clock state experience a position-dependent potential landscape. As they climb that landscape, kinetic energy is converted into potential energy; scattering on the \(1388~\mathrm{nm}\) transition pumps them to \(^{3}\mathrm{D}_{1}\), after which they decay spontaneously back to \(^{1}\mathrm{S}_{0}\), mostly through \(^{3}\mathrm{P}_{1}\), thereby removing motional energy from the sample [2406.13782].

The supplemental dressing-picture model writes the atom-light Hamiltonian as \(H = H_a + H_{LA}\). Diagonalization gives the spatially dependent ac-Stark shift of the dressed \(^{3}\mathrm{P}_{0}\) state. In the lin\(\parallel\)lin geometry, the \(1388~\mathrm{nm}\) field forms an intensity standing wave; in lin\(\perp\)lin, it forms a polarization-gradient lattice displaced by \(\lambda/4\). In both cases the dressed \(^{3}\mathrm{P}_{0}\) energy oscillates in space, and the scattering rate is largest near the maxima of the potential, which is the condition required for a Sisyphus cycle [2406.13782].

The cycle is nearly closed because the chosen \(^{3}\mathrm{D}_{1}\) state has favorable branching. The supplemental material gives \(^{3}\mathrm{D}_{1}\rightarrow ({}^{3}\mathrm{P}_{0},{}^{3}\mathrm{P}_{1},{}^{3}\mathrm{P}_{2}) = (65\%,\,34\%,\,<1\%)\). The \(\sim 1\%\) leakage to \(^{3}\mathrm{P}_{2}\) is therefore a practical limitation rather than a dominant failure mode [2406.13782].

A central feature of the mechanism is a spatial dark region for clock excitation. Atoms are excited preferentially near minima of the dressed \(^{3}\mathrm{P}_{0}\) potential, where the clock transition is closest to resonance. Away from those minima, the position-dependent light shift detunes the clock transition, so atoms become effectively decoupled from the laser in position space. The authors identify this dark-region physics as the basis of the sub-recoil regime [2406.13782].

## 3. Realization in \(^{171}\mathrm{Yb}\)

The reported implementation starts from \(^{171}\mathrm{Yb}\) atoms cooled on the \(556~\mathrm{nm}\) narrow-line MOT, \(^{1}\mathrm{S}_{0}\rightarrow{}^{3}\mathrm{P}_{1}\), to around \(20~\mu\mathrm{K}\). The \(1388~\mathrm{nm}\) field is delivered as two orthogonal pairs of counter-propagating beams, each with roughly \(1~\mathrm{mm}\) \(1/e^2\) diameter, in the plane transverse to the axis of a \(759~\mathrm{nm}\) magic-wavelength one-dimensional optical lattice. The main implementation uses the lin\(\parallel\)lin geometry. Because the \(759~\mathrm{nm}\) lattice is magic for the clock transition, clock excitation is not perturbed by the trapping light to first order, which is important both for clock operation and for clean interpretation of the cooling dynamics [2406.13782].

In free space, the authors report optimal cooling with the \(1388~\mathrm{nm}\) beam blue-detuned by about \(+50~\mathrm{MHz}\) from the \(^{3}\mathrm{P}_{0}(F=3/2)\) transition. Under those conditions the dressed-state potential depth exceeds \(U/h > 50~\mathrm{kHz}\), corresponding to \(U/k_B > 2.4~\mu\mathrm{K}\). The measured \(1388~\mathrm{nm}\) light shift, inferred through clock spectroscopy, is on the order of \(0.5\)–\(1~\mathrm{MHz}\) at the highest intensities and agrees with the simple model to within a factor of a few [2406.13782].

The free-space cooling signature is already substantial. After \(7~\mathrm{ms}\) of Sisyphus cooling on atoms initially at \(\sim 20~\mu\mathrm{K}\), the measured time-of-flight temperature of the cooled fraction is about \(3~\mu\mathrm{K}\), comprising roughly \(20\%\) of the sample. When the same cooling is applied during the last \(10~\mathrm{ms}\) of the narrow-line MOT and for \(10~\mathrm{ms}\) after the MOT is turned off, the number of atoms loaded into the \(759~\mathrm{nm}\) magic-wavelength lattice increases by about a factor of 4. The enhancement persists across lattice depths from about \(10\,E_r\) to \(100\,E_r\) [2406.13782].

These results define the method not merely as a free-space sub-Doppler stage, but as a clock-compatible loading protocol. The clock line is simultaneously the spectroscopic resource and the entrance to the dissipative cycle, which is the distinctive architectural feature of the scheme [2406.13782].

## 4. Sub-recoil operation and dynamical regimes

Once atoms are trapped in the one-dimensional magic lattice, the method reaches the sub-recoil regime in weakly confined directions. For one-dimensional cooling along a weakly confined axis at lattice depth \(60\,E_r\), stronger clock excitation, with \(\Omega_{\text{clock}}/2\pi = 2.7~\mathrm{kHz}\), yields faster cooling with an exponential time constant of \(36(5)~\mathrm{ms}\), but a higher steady-state temperature of around \(300~\mathrm{nK}\). Weaker excitation, with \(\Omega_{\text{clock}}/2\pi = 0.760(14)~\mathrm{kHz}\), cools more slowly, with a time constant of \(123(6)~\mathrm{ms}\), yet reaches \(165(5)~\mathrm{nK}\) [2406.13782].

Both temperatures are below the recoil temperature associated with the cascaded spontaneous-emission process,
\[
k_B T_r = \frac{\hbar^2 (k_{D-P}^2 + k_{P-S}^2)}{2m} \approx 410~\text{nK},
\]
where \(k_{D-P}\) and \(k_{P-S}\) are the wave numbers for the \(^{3}\mathrm{D}_{1}\rightarrow{}^{3}\mathrm{P}_{1}\) and \(^{3}\mathrm{P}_{1}\rightarrow{}^{1}\mathrm{S}_{0}\) decays. This is the principal quantitative marker of the sub-recoil claim [2406.13782].

The authors interpret the speed–temperature trade-off through the spatial dark region created by the periodic ac-Stark shift. Larger clock Rabi frequency makes it more likely that atoms are excited out of the dark region, which accelerates cooling but raises the steady-state temperature. The scheme also remains effective under substantial imbalance between counter-propagating \(1388~\mathrm{nm}\) beam powers, which the authors present as evidence of robustness to realistic optical imperfections [2406.13782].

Two-dimensional operation extends the method beyond the simplest geometry. With two orthogonal standing waves applied in the plane transverse to the one-dimensional lattice axis, the experiment infers radial temperatures near the recoil limit and also observes cooling along the longitudinal lattice axis, even though the Sisyphus potential is applied only transversely. The longitudinal effect arises because the spatially varying \(^{3}\mathrm{P}_{0}\) light shift can move a motional sideband into resonance with the clock laser, enabling red-sideband cooling. The measured longitudinal temperature is about \(0.8~\mu\mathrm{K}\) even when the lattice depth is reduced from \(\sim 107\,E_r\) to \(\sim 55\,E_r\) [2406.13782].

The method can also be combined with energy-selective blue-sideband excitation. Repeating a “selection + Sisyphus + red-sideband return” sequence ten times for a total of \(81~\mathrm{ms}\) lowers the radial temperature from \(451(35)~\mathrm{nK}\) to \(314(37)~\mathrm{nK}\). Continuous operation is likewise demonstrated: with the clock laser held on the selection frequency and the Sisyphus light left on, the experiment still observes subrecoil radial temperatures and \(\bar{n}_z < 0.32(5)\) [2406.13782].

## 5. Relationship to other Sisyphus architectures

The broad Sisyphus principle long predates clock-line mediation, and comparison with neighboring schemes helps delimit what is specific to the clock-line variant. In lithium, the demonstrated mechanism is canonical polarization-gradient Sisyphus cooling in a lin \(\perp\) lin standing wave. The cooling uses the D1 and D2 lines, not a forbidden clock transition; it reaches temperatures as low as \(40~\mu\mathrm{K}\) in 1D and \(100~\mu\mathrm{K}\) in 3D, with cooling on a millisecond timescale, and the authors explicitly frame it as evidence that unresolved excited-state hyperfine structure does not preclude standard sub-Doppler molasses in lithium [1308.1935].

A second neighboring class is narrow-line-mediated but non-clock cooling. In magnetically trapped \(^{88}\mathrm{Sr}\) prepared in the \(5s5p\,^{3}\mathrm{P}_{2}\) metastable state, cooling occurs on the narrow \(5s5p\,^{3}\mathrm{P}_{2}\rightarrow 5s4d\,^{3}\mathrm{D}_{3}\) line at \(2.92~\mu\mathrm{m}\), with linewidth \(\Gamma_{\rm IR}\approx 57~\mathrm{kHz}\). A \(641~\mathrm{nm}\) standing wave blue-detuned from \(5s4d\,^{3}\mathrm{D}_{3}\rightarrow 5p4d\,^{3}\mathrm{F}_{4}\) creates a dissipative optical lattice in the \(^{3}\mathrm{D}_{3}\) state, modeled as
\[
U_{\rm S}(z)=U_0\cos^2\!\left(\frac{2\pi z}{\lambda_{641}}\right),
\]
with \(U_0/k_{\rm B}\sim 380\,\mu\mathrm{K}\) in the main demonstration. The authors report a cooling time reduced by about an order of magnitude, a \(z\)-axis temperature of about \(26\,\mu\mathrm{K}\) versus \(\sim 74\,\mu\mathrm{K}\) for Doppler cooling alone after the same observation window, and an \(88(10)\%\) increase in atom-loading efficiency into a moving optical lattice, corresponding to nearly a factor-of-two improvement in atom number [2506.19701].

A third neighboring class couples clock-compatible trapping to a different cooling line. In \(^{88}\mathrm{Sr}\) clock-magic tweezers at \(813.4~\mathrm{nm}\), the trap is magic for \(^{1}\mathrm{S}_{0}\leftrightarrow{}^{3}\mathrm{P}_{0}\), but imaging is stabilized by attractive narrow-line Sisyphus cooling on the \(^{1}\mathrm{S}_{0}\leftrightarrow{}^{3}\mathrm{P}_{1}\) intercombination line at \(689~\mathrm{nm}\). The excited state is more strongly trapped than the ground state, and the cooling works with a single non-retroreflected \(689~\mathrm{nm}\) beam in the radial direction. This protocol enables \(2000\)-times repeated imaging, a single-image survival probability of \(0.99932(8)\), a detection fidelity of \(0.99991(1)\), and temperatures below \(5~\mu\mathrm{K}\) [1811.06014].

Taken together, these comparisons show that clock-line-mediated Sisyphus cooling is neither synonymous with Sisyphus cooling in general nor reducible to the use of a clock-magic trap. Its defining feature is the use of the clock transition itself as the coherent gateway into a dissipative, state-dressed potential landscape [2406.13782].

## 6. Metrological relevance, limitations, and directions

The immediate application emphasized in the ytterbium work is optical lattice clocks. Lower temperatures facilitate the use of shallower magic lattices with reduced light shifts while retaining large atom numbers to reduce quantum projection noise. The paper identifies this as directly relevant to pushing clock performance toward and below the \(10^{-18}\) level. Because the method can be pulsed or continuous, and because it does not require switching off the trap or using time-varying magnetic fields, it is also positioned for continuous or high-bandwidth sensing and may help avoid aliasing in continuous clock schemes [2406.13782].

The broader applicability follows from the alkaline-earth-like level structure: a long-lived clock state together with a higher-lying dipole-allowed state that can dress that clock state. The authors explicitly identify Cd, Hg, Mg, and Ra as candidate systems. They further note relevance to atom interferometry, optical tweezer arrays, site-selective cooling and imaging, continuous sensors, and even (anti)hydrogen cooling ideas. This suggests a general platform strategy in which narrow clock coherence and engineered dissipation are combined rather than separated into distinct stages [2406.13782].

A practical limitation is the \(\sim 1\%\) branching to \(^{3}\mathrm{P}_{2}\) per cycle in the ytterbium implementation, which causes atom loss unless a repumper is added. The authors nevertheless argue that the loss is manageable because only a few cycles are needed to reach the recoil-limited regime. In the strontium metastable implementation, the same general logic appears in a different form: faster cooling moves atoms into the loading region before background-gas collisions or light-assisted losses accumulate during a multi-second cooling stage, thereby increasing transfer into the moving optical lattice [2506.19701].

For tweezer-based clock platforms, the strontium \(813.4~\mathrm{nm}\) result demonstrates that clock-magic trapping and low-loss cooling can coexist, even when the cooling line is not the clock line itself. That coexistence enabled clock-state-resolved detection with average state detection fidelity \(0.981(1)\) and average survival probability \(0.996(1)\), together with repeated interrogation-compatible imaging [1811.06014]. A plausible implication is that future architectures may combine clock-line-mediated cooling for loading and state preparation with intercombination-line or metastable-manifold cooling for imaging, transport, or continuous outcoupling, depending on which transition best matches the required dissipation pathway.

Source: https://www.emergentmind.com/topics/clock-line-mediated-sisyphus-cooling