High-fidelity entanglement of metastable trapped-ion qubits with integrated erasure conversion

This paper demonstrates that high-fidelity control of metastable trapped-ion qubits enables erasure conversion of spontaneous scattering errors, achieving a SPAM-corrected Bell state fidelity of 98.61% (99.16% after erasure subtraction) to support low-overhead, fault-tolerant quantum computing.

The Gist

The Big Picture: Building a Better Quantum Computer

Imagine you are trying to build a super-computer that solves problems no normal computer ever could. This is a quantum computer. One of the best ways to build one is using trapped ions—basically, tiny atoms (like calcium) that are floating in a vacuum, held in place by invisible electric "jails" (traps).

However, these quantum computers have a major problem: they are incredibly fragile. Like a house of cards in a windstorm, a tiny bit of noise or a stray photon can knock the calculation off track. To fix this, scientists usually have to use a lot of "spare parts" (extra atoms) just to check if an error happened and fix it. This is called overhead, and it makes building a useful computer very hard and expensive.

This paper introduces a clever new trick to make these quantum computers much more efficient. They found a way to turn "mystery errors" into "known errors," which are much easier to fix.

---

The Problem: The "Mystery Error" vs. The "Known Error"

To understand the breakthrough, let's use an analogy of a library.

1. The Old Way (Pauli Errors): Imagine a librarian is organizing books. Suddenly, a book falls off a shelf and lands on the floor. The librarian doesn't know which book fell, or even that it fell. They just notice the story doesn't make sense later. This is a Pauli error. It's a silent mistake. To fix it, the librarian has to check every single book in the library to find the problem. This takes forever (high overhead).
2. The New Way (Erasure Errors): Now, imagine the librarian puts a special alarm on every book. If a book falls, the alarm goes off immediately, and a red light flashes saying, "Book #42 fell!" The librarian knows exactly what went wrong and where. This is an erasure error. Because they know the location, they only need to check that one book. This is incredibly efficient.

The Goal: The scientists wanted to turn those silent "mystery errors" into loud "known errors" (erasures).

---

The Solution: The "Metastable" Trick

The researchers used a specific type of atom (Calcium-40) and a special way of storing information called metastable qubits.

Think of the atom's energy levels like floors in a building:

• The Ground Floor (Normal Qubits): Most quantum computers store data here. But if a photon hits the atom, it might get knocked up to a "forbidden" floor (leakage) and get stuck there, disappearing from the calculation.
• The Metastable Floor (The New Trick): The researchers decided to store their data on a high, safe floor that is far away from the busy "ground floor" traffic.

Why is this better?
When the atom is on this high floor, if it gets knocked off (due to a mistake), it falls all the way down to the ground floor. Because it fell so far, it's easy to spot! They can shine a light on the atom, and if it's on the ground floor, it glows. If it's on the high floor, it stays dark.

• The "Fluorescence Check": This is their "alarm system." After they do a calculation, they shine a light.
  • No Glow: The atom is safe on the high floor. The calculation is good.
  • Glow: The atom fell! We know exactly what happened. We can throw away that specific result and try again, or use a code to fix it easily.

This is called Erasure Conversion. They turned a hidden bug into a visible one.

---

The Experiment: The Dance of the Atoms

The team performed a complex dance with two of these atoms to make them "entangled" (a quantum connection where they act as one unit).

1. The Setup: They used a very powerful, specialized laser system (like a high-tech flashlight) to nudge the atoms without hitting them too hard.
2. The Dance: They made the atoms interact using a "geometric phase gate." Imagine two dancers holding hands and spinning around a pole. If they spin perfectly, they end up in a special synchronized pose.
3. The Result:
   • They achieved a 97.7% success rate on the first try.
   • When they looked at the "mystery errors" and threw away the ones where the atoms fell off the floor (the erasures), the success rate jumped to 99.16%.

This is a record-breaking level of accuracy for this type of quantum computer.

---

Why This Matters: The Road to a Real Quantum Computer

The paper shows that by using this "Metastable" trick:

1. We need fewer spare parts: Because we can catch errors so easily, we don't need as many extra atoms to fix them.
2. We can scale up: It's easier to build a huge, powerful quantum computer if the errors are easy to manage.
3. It's faster: The "alarm system" (checking for errors) is very efficient, so the computer doesn't waste time guessing what went wrong.

In summary: The scientists found a way to make quantum atoms "scream" when they make a mistake, instead of whispering. By listening for that scream, they can fix the problem instantly, paving the way for the first truly useful quantum computers that can solve real-world problems like drug discovery or climate modeling.

Technical Summary

1. Problem Statement

Current state-of-the-art trapped-ion quantum computers face two primary limitations hindering their scalability to utility-scale systems:

• Dual-Species Overhead: Most high-fidelity systems use mixed-species ion crystals (e.g., data qubits + cooling qubits) to separate cooling/readout functions from computation. This introduces significant hardware overhead and complexity.
• Error Correction Efficiency: Standard Quantum Error Correction (QEC) codes are optimized for generic Pauli errors. However, correcting "erasure" errors (errors whose location is known) is significantly more efficient; a code can correct roughly twice as many erasures as Pauli errors for the same code distance.
• The Challenge: While metastable qubits (encoded in long-lived $D$ states) offer a path to "dual-type" architectures (using a single ion species for both cooling and data), they suffer from spontaneous Raman scattering (SRS) and finite state lifetimes. Historically, these errors manifested as undetected Pauli errors or leakage, preventing efficient erasure conversion.

2. Methodology

The authors implemented a system using metastable qubits ($m$-qubits) in $^{40}\text{Ca}^+$ ions to address these issues.

• Qubit Encoding: The qubits are encoded in the $D_{5/2}$ manifold, specifically the $| \uparrow \rangle \equiv |m_J = +5/2\rangle$ and $| \downarrow \rangle \equiv |m_J = +3/2\rangle$ states. This manifold is spectrally isolated from the cycling optical transitions used for laser cooling and fluorescence readout.
• Gate Mechanism:
  • Entangling Gate: A two-ion geometric phase gate was performed using far-detuned (-44 THz) stimulated Raman transitions at 976 nm.
  • Laser System: A compact, injection-locked diode laser system was developed to provide high power (up to 220 mW per beam) with low noise, enabling fast gate times (400 $\mu$s).
  • Control: The gate utilizes a spin-dependent force (SDF) generated by orthogonal 976 nm beams. To mitigate noise, the sequence employs Walsh modulation (4 phase space loops) and a spin-echo pulse.
• Erasure Conversion Scheme:
  • Fluorescence Check (FC): After state preparation and after the entangling gate, the system performs a "fluorescence check." This involves driving the $S_{1/2} \leftrightarrow P_{1/2}$ cycling transition.
  • Detection Logic: If an ion has leaked out of the $D_{5/2}$ manifold (due to decay or SRS), it will scatter 397 nm photons and be detected as "bright." If it remains in $D_{5/2}$, it remains "dark."
  • Conversion: Any detected leakage is flagged as an erasure error. Because the $P_{3/2}$ manifold branches $\sim94\%$ to $S_{1/2}$, most SRS events result in detectable leakage, converting what would be a Pauli error into a known erasure.

3. Key Contributions

• Integrated Erasure Conversion: The paper demonstrates a practical, integrated scheme to convert $\sim94\%$ of spontaneous Raman scattering errors and nearly all decay errors into heralded erasures without disrupting error-free qubits.
• Metastable Architecture Validation: It validates the "omg" (or dual-type) architecture, proving that a single ion species can perform sympathetic cooling, state readout, and high-fidelity logic gates simultaneously.
• Comprehensive Error Budget: The authors provide a detailed breakdown of error sources, distinguishing between non-erasure (Pauli/dephasing) and erasure errors, and offering a roadmap for future improvements.
• Hardware Innovation: Development of a high-power, injection-locked 976 nm laser system specifically tailored for driving Raman transitions on metastable states.

4. Experimental Results

• Bell State Fidelity:
  • Raw Fidelity: $97.73(8)\%$
  • SPAM-Corrected Fidelity: $98.61(8)\%$ (Corrected for State Preparation and Measurement errors).
  • Post-Selected Fidelity (Erasure Corrected): By discarding experimental shots where leakage was detected (0.55% of shots), the fidelity rose to $99.16(7)\%$.
• Error Reduction: The erasure conversion scheme resulted in a 39% reduction in the non-SPAM error rate, a performance competitive with the best neutral atom demonstrations.
• Error Budget Analysis:
  • Dominant Non-Erasure Errors: Motional dephasing ($55 \times 10^{-4}$) and spin dephasing ($26 \times 10^{-4}$).
  • Dominant Erasure Errors: Decay during the fluorescence check ($17.1 \times 10^{-4}$) and Raman scattering ($13.5 \times 10^{-4}$).
  • Total Erasure Rate: $55(3) \times 10^{-4}$.
• Simulation of Optimization: The authors modeled future upgrades (e.g., counter-propagating beams, different qubit states, removing Walsh modulation) and projected a 17-fold reduction in gate time (from 400 $\mu$s to 23 $\mu$s), which would push the Raman scattering error below the $10^{-4}$ threshold required for fault tolerance.

5. Significance

This work represents a critical step toward low-overhead, fault-tolerant quantum computing using trapped ions.

• Scalability: By enabling high-fidelity operations on a single ion species, it removes the need for complex mixed-species crystals, simplifying trap design and scaling.
• QEC Efficiency: The ability to convert a large fraction of gate errors into erasures significantly lowers the resource overhead required for quantum error correction. This makes the path to logical qubits more feasible.
• Benchmarking: The achieved $99.16\%$ post-selected fidelity sets a new benchmark for metastable trapped-ion qubits, demonstrating that they can outperform ground-state qubits in specific error-correction contexts when combined with erasure conversion.

In conclusion, the paper establishes that metastable trapped-ion qubits, when paired with integrated erasure conversion, offer a viable and highly efficient platform for the next generation of quantum computers.