Laser cooling and qubit measurements on a forbidden transition in neutral Cs atoms

This paper experimentally demonstrates high-fidelity, hyperfine-level-selective measurements of individual neutral cesium atoms by combining simultaneous laser cooling on a forbidden transition with background-free imaging, achieving a detection fidelity of 0.9993 while enabling repeated, low-loss state measurements.

The Gist

Imagine you are trying to take a photograph of a single, tiny, invisible firefly floating in a dark room. You want to know two things: Is the firefly there? and What color is it? (Is it a "red" firefly or a "blue" firefly?).

In the world of quantum computing, these fireflies are atoms, and their colors represent the "qubits" (the basic units of information). The problem is that taking a picture usually involves shining a bright light on them. If the light is too bright or the wrong kind, you might accidentally scare the firefly away (losing the atom) or change its color (destroying the information) before you can take the picture.

This paper describes a new, clever way to take a "perfect" photo of a single Cesium atom without scaring it away or changing its color. Here is how they did it, explained simply:

1. The "Forbidden" Flashlight

Usually, scientists take pictures of atoms using a very common, bright "flashlight" (a laser) that makes the atom glow brightly. But this glow is so intense that it heats the atom up, causing it to jitter and fly out of its trap.

The researchers used a "forbidden" transition. Think of this like trying to open a door that is usually locked. It's very hard to open, so the atom doesn't react as violently. Specifically, they used a special laser (685 nm) that nudges the atom into a state it doesn't usually visit easily. Because this "door" is hard to open, the atom glows much more softly and calmly. This allows them to keep the atom cool and trapped while they look at it.

2. The "Background-Free" Camera

Imagine trying to hear a whisper in a room where a loud fan is humming. It's hard to tell if you're hearing the whisper or just the fan.

In previous experiments, the light used to take the picture often scattered off the glass windows or the lenses, creating a "fog" of background noise that made it hard to see the atom clearly.

The researchers used a trick: they looked for the atom's glow at a different color than the light they used to excite it.

• They shined a red laser to wake the atom up.
• They took a picture of the blue light the atom emitted.
• They used special filters to block all the red light.

This is like wearing sunglasses that block the sun but let the moonlight through. The result is a crystal-clear image with zero background noise. They could see the atom perfectly, distinguishing between "bright" (atom is there) and "dark" (atom is gone) with 99.93% accuracy.

3. The "Cooling" Blanket

Taking a picture usually takes time. If you hold a camera still for too long, your hand shakes. In this experiment, the "shaking" is the atom moving around due to heat.

To solve this, they didn't just take a picture; they cooled the atom while taking the picture. They used a 3D "molasses" of lasers (a sticky, cold trap) that slowed the atom down to a temperature of just 5.3 micro-Kelvin. That is colder than outer space! This kept the atom still and safe inside its trap, allowing them to take repeated pictures without losing it.

4. The Speed Problem and the "Turbo" Button

Even with this perfect setup, the "forbidden" door was too hard to open. The atom glowed very slowly, meaning the researchers had to wait about 200 milliseconds (0.2 seconds) to get a clear picture. While that sounds fast to us, for a quantum computer, it's like watching paint dry. It's too slow to keep up with the computer's calculations.

The paper proposes a solution: Quenching.
Imagine the atom is a slow, sleepy firefly. The researchers suggest adding a second "helper" laser (an auxiliary field) that acts like a turbo button. This helper laser nudges the atom to release its energy faster, making it glow much brighter and faster.

• Current speed: ~200 milliseconds.
• Projected speed with the "turbo": ~60 microseconds (0.00006 seconds).

This would make the measurement 3,000 times faster while keeping the accuracy just as high.

The Bottom Line

The team successfully demonstrated a way to take a high-definition, noise-free photo of a single atom without losing it or changing its state. They proved it works with incredible accuracy (99.93% fidelity) and very low loss.

While the current method is a bit slow because the "forbidden" transition is so gentle, their theoretical analysis shows that by adding a helper laser to speed things up, they could make the process nearly instantaneous. This is a crucial step toward building faster, more reliable quantum computers that can correct their own errors in real-time.

Technical Summary

Technical Summary: Laser cooling and qubit measurements on a forbidden transition in neutral Cs atoms

Problem Statement
Scalable quantum error correction requires fast, high-fidelity measurement of qubit states with minimal atom loss. While neutral atom qubits have achieved rapid gate operations, state measurements in free-space arrays typically require integration times of several milliseconds—orders of magnitude slower than gate times—to achieve high fidelity. Existing methods that achieve faster measurements often fail to resolve internal hyperfine states or result in significant atom loss. Furthermore, integrating atom arrays with optical cavities to enhance measurement speed remains a scalability challenge. The authors address the need for a free-space measurement technique that simultaneously offers high fidelity, low atom loss, and the potential for rapid measurement.

Methodology
The authors experimentally demonstrate a background-free, hyperfine-level-selective measurement scheme for neutral Cesium (Cs) atoms using the electric-quadrupole (E2) transition from the ground state $|6s_{1/2}\rangle$ to the excited state $|5d_{5/2}\rangle$ (685 nm).

• Experimental Setup: The system utilizes a 2D MOT to load atoms into a 3D MOT, followed by transfer to a single-site 803 nm "bottle-beam" optical trap created via an imaged hole array. Cooling is performed in two stages: initial cooling on the standard 852 nm D2 line, followed by "second-stage" cooling on the 685 nm quadrupole transition.
• Measurement Principle: The measurement relies on detecting fluorescence at 852 nm (from the $|5d_{5/2}\rangle \to |6p_{3/2}\rangle$ decay) while exciting the atom with 685 nm light. Because the excitation wavelength (685 nm) differs from the detection wavelength (852 nm), the method is background-free, eliminating stray light from the excitation laser.
• State Selectivity: The $|5d_{5/2}\rangle$ state possesses a large ratio of hyperfine splitting to radiative linewidth ($y = \Delta_{hf}/\Gamma \approx 1000$). This large $y$ factor suppresses Raman scattering events that would change the hyperfine ground state, enabling state-selective cycling between $|6s_{1/2}, f=4\rangle$ and $|5d_{5/2}, f'=6\rangle$ without depumping to the $f=3$ ground state.
• Cooling and Trapping: The 685 nm light provides simultaneous 3D cooling to prevent heating-induced trap loss during the measurement. The 803 nm trap is "magic" for the $f=4$ and $f=6$ states, minimizing differential light shifts.
• Theoretical Extension: The authors perform numerical simulations using the Lindblad master equation to model a proposed speed-up technique. This involves "quenching" the long-lived $|5d_{5/2}\rangle$ state by stimulating the dipole-allowed transition to $|6p_{3/2}\rangle$ using an auxiliary 3491 nm field, thereby increasing the photon scattering rate.

Key Results

• Experimental Performance: The authors achieved a state detection fidelity of $F = 0.9993(4)$ and an atom retention probability of $0.9954(5)$ during a 200 ms integration time. The atom temperature was cooled to $5.3 \, \mu\text{K}$, and Ramsey coherence measurements indicated a dephasing time $T_2^* = 7.6(3)$ ms.
• Noise Characterization: The background noise was dominated by spectrally broad scattering from the objective lens and vacuum cell windows, contributing approximately two-thirds of the total noise.
• Loss Mechanisms: Atom loss was primarily attributed to vacuum collisions and hyperfine depumping. The measured depumping lifetime was $\tau_{1/e} \approx 42$ s, and the trap lifetime was $\tau_{1/e} \approx 43$ s.
• Theoretical Projections: Simulations incorporating the auxiliary quenching field predict that the measurement time can be reduced to $\sim 60 \, \mu\text{s}$ while maintaining a projected fidelity of $\sim 0.9995$. This represents a scattering rate increase of over 64 times compared to the experimental radiative limit.

Significance and Claims
The paper claims to demonstrate some of the highest fidelity values reported for neutral atom qubits in free space, and to the authors' knowledge, the highest fidelity demonstrated with an alkali atom. The significance of the work lies in:

1. Background-Free Imaging: Successfully implementing a state-selective measurement that avoids the background noise typically associated with resonant excitation.
2. Simultaneous Cooling and Measurement: Demonstrating that high-fidelity readout can be performed while actively cooling the atom in three dimensions, preventing trap loss.
3. Scalability Potential: Providing a free-space alternative to cavity-based measurements.
4. Path to Speed: Identifying that the current speed limitation is due to the narrow linewidth of the $5d_{5/2}$ state, and theoretically validating that excited-state quenching can overcome this bottleneck without sacrificing fidelity.

The authors note that while the current demonstration is limited by long integration times ($\sim 200$ ms) and technical noise, the combination of the demonstrated approach with the proposed quenching scheme, multi-atom repetition codes, and machine learning-enhanced image processing has the potential to reduce measurement times to well under $10 \, \mu\text{s}$.