# Neutral-Atom Arrays: Reconfigurable Quantum Hardware **Source**: https://quantumsequrity.com/blog/neutral-atom-arrays-explained **Category**: Quantum Computing Fundamentals --- [← Back to Blog](../../blog.html) Quantum Computing Fundamentals # Neutral-Atom Arrays: Reconfigurable Quantum Hardware 13 min read For most of the 2010s, the leading approaches to quantum computing felt locked into specific tradeoffs. Superconducting circuits were fast but small. Trapped ions had high fidelity but did not scale easily. Then, around 2020, a third approach started moving rapidly. Neutral atom arrays use clouds of laser cooled atoms held in place by optical tweezers, with hundreds and now over a thousand atoms operating as qubits in a single experimental system. Two companies, Atom Computing and QuEra, have driven the commercial momentum, and a 2024 Atom Computing announcement of a 1180 atom array signaled that neutral atom platforms had genuinely arrived as a serious contender. This article explains how neutral atom quantum computers work, the unique advantages of optical tweezer arrays, and what the recent atom count milestones actually mean for the path to fault tolerance. ## A Plain English Picture Imagine you wanted to play three dimensional chess with thousands of pieces, but you wanted to move pieces freely without picking them up. Magnets could push the pieces around, but they would interfere with each other. Instead, you build a magic table that uses focused beams of light to grab each piece individually, hold it in place, and move it anywhere you want. The pieces never touch you or each other except when you choose. Want to rearrange a hundred pieces at once? The light beams handle it. Neutral atom quantum computers work like this. The pieces are individual atoms, typically rubidium or strontium. The light beams are tightly focused laser tweezers that grab atoms and hold them at specific positions in a vacuum chamber. The arrangement is fully reconfigurable. Atoms can be moved, swapped, brought close together to interact, and then separated. The flexibility is unlike any other quantum computing platform. When two atoms need to interact, they are pushed close enough that an effect called Rydberg interaction takes over. One or both atoms are excited to a high energy state where the electron is far from the nucleus. The interactions between Rydberg atoms are strong and precisely controllable, which is what enables fast two qubit gates. ## How Neutral Atom Arrays Work A neutral atom quantum computer has several key components. ### The Atom Source A magneto optical trap captures atoms from a vapor source, typically a few hundred atoms at a time. The atoms are cooled to microkelvin temperatures using laser cooling, where the atoms slow down by absorbing and reemitting photons in a way that always reduces their kinetic energy. ### The Optical Tweezer Array Once cooled, atoms are loaded into an array of optical tweezers, individual focused laser beams that each trap one atom at a specific position. Modern systems use spatial light modulators to create programmable arrays with thousands of tweezer positions. The spacing between tweezers is typically a few micrometers, small enough that Rydberg interactions can be activated by lasers but large enough that the atoms are well isolated when not interacting. The position of each tweezer can be moved during the computation. Atoms can be picked up from one position and dropped at another. This is fundamentally different from other platforms where qubits are fixed in place. ### Qubit Encoding The qubit is encoded in two long lived atomic states, typically two hyperfine ground states. These states have very long coherence times, similar to those in trapped ion systems, often seconds or longer. ### Rydberg Gate A two qubit gate is performed by exciting one or both atoms to a Rydberg state using precisely tuned lasers. Two Rydberg atoms within a few micrometers of each other interact strongly, producing a controlled phase shift that implements the gate. Gate times are typically microseconds, much faster than trapped ion gates and only slightly slower than superconducting gates. ### Readout Measurement is done by laser induced fluorescence. Atoms in one qubit state scatter many photons, atoms in the other state scatter few. A camera images the array and the photon counts at each tweezer position determine the measured state of each atom. ## Why Reconfigurability Matters The ability to move atoms during computation is a unique feature that has substantial practical consequences. ### Effective All to All Connectivity In fixed qubit platforms like superconducting chips, two qubits that are not adjacent require routing operations to interact. On a neutral atom array, two qubits that need to interact can simply be moved next to each other. This provides effective all to all connectivity without requiring SWAP overhead. ### Algorithm Specific Layouts Different algorithms have different connectivity requirements. A neutral atom system can be reconfigured for each algorithm, putting qubits in the optimal layout for that specific problem. This is impossible on fixed layout platforms. ### Mid Circuit Measurement and Reset Some atoms can be measured mid circuit and replaced with fresh atoms while others continue computing. This enables protocols that require measurement during computation, including various error correction schemes. ### Modular Architecture Multiple atom arrays can be connected, with atoms shuttled between them. This is a natural path to scaling beyond a single array. ## Recent Milestones The neutral atom field has moved exceptionally quickly through the early 2020s. | Year | Group | Result | |---|---|---| | 2020 | Lukin group, Harvard | 256 atom programmable simulator | | 2022 | QuEra | 256 atom commercial system, Aquila | | 2023 | Lukin group, Harvard | 280 logical qubits demonstrated, fault tolerant operations | | 2024 | Atom Computing | 1180 atom array announced | | 2024 | QuEra | Logical qubit demonstrations on commercial hardware | The 1180 atom result from Atom Computing in late 2024 was a milestone in scale. While the physical qubit count is large, the practical capability depends heavily on gate fidelity and error correction overhead. Logical qubit counts are still much smaller, but the trajectory is favorable. The 2023 Harvard demonstration of 280 logical qubits and fault tolerant operations was particularly significant. The team showed that logical qubits encoded in surface codes outperformed the underlying physical qubits, and that logical operations could be performed with high fidelity. This was the first demonstration of fault tolerant quantum computation with significant scale on any platform. ## QuEra and Atom Computing QuEra, founded in 2018 and spun out from Harvard's Lukin group, was the first commercial neutral atom company. Their Aquila system, accessible through Amazon Braket, provides 256 atom programmable computation. The company has announced roadmaps targeting hundreds of logical qubits using neutral atom hardware combined with surface code error correction. Atom Computing, founded in 2018, has emphasized scale and announced the 1180 atom array as a step toward million qubit systems. Their architecture uses strontium atoms, which have favorable properties for some quantum computing tasks. Both companies are still relatively early stage commercially. The neutral atom platform has been successful in research demonstrations but has not yet matched superconducting or trapped ion systems in commercial deployment scale. ## Tradeoffs Versus Other Platforms Neutral atom systems have a distinctive set of strengths and weaknesses. | Property | Neutral atom | Superconducting | Trapped ion | |---|---|---|---| | Largest reported system | 1180 atoms | Around 1100 qubits | About 56 ions | | Connectivity | Effectively all to all via reconfiguration | Nearest neighbor | All to all | | Two qubit gate time | Microseconds | Tens of nanoseconds | Microseconds to milliseconds | | Coherence time | Seconds | Tens of microseconds | Seconds to minutes | | Operation temperature | Room temp vacuum | Millikelvin | Room temp to cryogenic vacuum | | Reconfigurability | Yes, in real time | No | Limited via shuttling | The combination of large scale, reconfigurability, room temperature operation, and Rydberg gate fidelities make neutral atom systems a strong candidate for fault tolerant quantum computing. The remaining challenges are gate fidelity improvements and engineering reliability for sustained commercial operation. ## What This Means for Cryptographic Timelines The fast pace of neutral atom progress has shifted some assessments of CRQC timelines. The 2023 logical qubit demonstrations and the 2024 atom count milestones are real progress that did not appear in roadmaps a few years ago. Neutral atom systems may not be the first platform to reach CRQC, but they are now plausible contenders. The right defensive posture is not to bet on which platform wins. It is to deploy [post-quantum cryptography](../what-is-post-quantum-cryptography.html) ahead of any plausible CRQC arrival regardless of platform. QNSQY ships [NIST standardized](../nist-fips-guide.html) [ML-KEM](../ml-kem-explained.html) and [ML-DSA](../mldsa-vs-slhdsa.html) algorithms today, in [hybrid encryption](../hybrid-encryption.html) with classical schemes for transition safety. The protection is platform agnostic. ## Engineering Challenges That Remain Despite the rapid progress, neutral atom systems face several engineering challenges that will determine whether the platform reaches its potential. ### Atom Loading Efficiency Loading individual atoms into specific tweezer positions is probabilistic. A typical loading attempt fills only about half the available tweezers. To get a fully populated array, the system iterates the loading process, repositioning atoms to fill empty positions. This rearrangement step takes time and is one of the bottlenecks for system throughput. Recent work has improved loading efficiency to over 95 percent for the targeted atom positions, but the remaining failures still require rearrangement. Scaling this process to thousands of atoms while maintaining throughput is an active engineering area. ### Two Qubit Gate Fidelities at Scale While single atom Rydberg gates have demonstrated fidelities above 99.5 percent in dedicated experiments, achieving these fidelities consistently across an entire array of hundreds or thousands of atoms is harder. Variations in laser intensity across the array, fluctuations in atom positions, and crosstalk between simultaneous gates all degrade fidelity. Closing the gap between single gate demonstrations and array wide consistent performance is essential for scaling. ### Crosstalk and Parallel Operations To run algorithms efficiently, multiple gates need to execute in parallel across the array. The Rydberg interaction range and laser focus spot determine how close two simultaneous gates can be without interfering. Engineering low crosstalk parallel gate operations is a significant challenge that becomes harder as arrays scale. ### Sustained Operation Many neutral atom demonstrations are short, lasting from microseconds to milliseconds before the array needs to be reloaded. Sustained continuous operation is more challenging and is a focus of commercial development. Atom Computing has demonstrated longer continuous operation than typical academic experiments, which is part of what positions them as a commercially serious player. The combination of these challenges does not invalidate the platform's promise but does establish that real engineering work is required between current capabilities and the million qubit visions on company roadmaps. ## Frequently Asked Questions ### Why is reconfigurability such a big deal? Most quantum algorithms have specific patterns of qubit interactions that depend on the problem structure. On platforms with fixed layouts, executing these patterns requires routing operations that consume time and introduce errors. On neutral atom arrays, the qubits can be physically moved into the optimal layout for each algorithm. This effectively eliminates the routing overhead and gives algorithms more operations within their coherence window. ### Are 1180 atoms actually 1180 usable qubits? The atom count and the usable qubit count are different things. 1180 atoms means 1180 individual trapped atoms, but not all of them may have the calibration quality needed for high fidelity operations. The actual usable qubit count for state of the art demonstrations is somewhat smaller, though the trajectory is favorable. ### How are neutral atoms different from trapped ions? Trapped ions are charged atoms held in place by oscillating electric fields, while neutral atoms are uncharged atoms held by optical tweezers. The physics of trapping is different, the gate mechanism is different, and the practical engineering tradeoffs are different. Neutral atoms scale more easily and have all to all connectivity through reconfiguration. Trapped ions have demonstrated higher fidelities and longer maturity. ### Could neutral atoms reach a million qubits? The platform's manufacturing scalability is genuinely good. Optical tweezer arrays can in principle hold thousands of atoms in a single experimental system, and multiple arrays can be networked. Whether million atom systems become commercially deployable depends on engineering progress on laser systems, vacuum technology, and control electronics. Atom Computing has explicitly targeted this scale. ### Does QNSQY's defense depend on which quantum platform wins? No. QNSQY uses post-quantum algorithms whose security rests on mathematical hardness assumptions independent of the specific quantum hardware. Whether the eventual CRQC is built from neutral atoms, superconducting circuits, trapped ions, or photons, the algorithms remain protected. ## Sources 1. [QuEra Aquila system documentation](https://www.quera.com/aquila) 2. [Atom Computing 1180 atom announcement](https://atom-computing.com/news) 3. [Bluvstein et al, Logical quantum processor based on reconfigurable atom arrays, Nature 2023](https://www.nature.com/articles/s41586-023-06927-3) 4. [NIST post-quantum cryptography standards](https://csrc.nist.gov/projects/post-quantum-cryptography) 5. [Saffman, Quantum computing with neutral atoms, IEEE Spectrum review article](https://arxiv.org/abs/1605.05207) 6. [NSA Commercial National Security Algorithm 2.0 advisory](https://media.defense.gov/2022/Sep/07/2003071834/-1/-1/0/CSA_CNSA_2.0_ALGORITHMS_.PDF) ## Related Articles - [Qubit basics for security professionals](../qubit-basics-for-security-professionals.html) - [Logical qubits vs physical qubits](../logical-qubits-vs-physical-qubits.html) - [What is post-quantum cryptography](../what-is-post-quantum-cryptography.html) - [CRQC meaning explained](../crqc-meaning-explained.html) - [Q-Day: when quantum breaks encryption](../q-day-when-quantum-breaks-encryption.html) --- ### Protect Your Data Before Q-Day Arrives QNSQY's NIST-standardized post-quantum encryption protects files against both current and quantum-era threats. [Try QNSQY](../../pricing.html)