Mølmer-Sørensen Gate on Superconducting Quantum Computers Achieves Competitive Fidelity with Hardware Efficiency

NISQ Architecture Gate Set Expansion

The effective implementation and complete benchmarking of the Mølmer-Sørensen (MS) entangling gate on a superconducting quantum processor is a significant quantum hardware engineering accomplishment. MS gates, a key operation in trapped-ion systems, has 92.47% process fidelity on IBM Quantum hardware. This performance is quite comparable with the device's 93.02% inherent controlled-NOT (CX) gate integrity.

This illustrates that hardware-aware compilation can optimize non-native entangling gates to work like native processor architectural operations. The discoveries are crucial for Noisy Intermediate-Scale Quantum (NISQ) computing because they expand the gates available for algorithm construction on fixed-architecture processors.

The Mølmer-Sørensen Gate (MS): A New Superconductor Tool

The Mølmer-Sørensen (MS) gate is a crucial entangling quantum logic gate in trapped-ion quantum computing.

This high-fidelity entangling gate generates an Ising-like interaction between qubits, a R XX rotation on two qubits. As a universal gate, it can be used in any quantum circuit using single-qubit rotations.

Its principal use is trapped-ion quantum computers.

The ions are exposed to a bichromatic laser field with two laser tones. This field links the ions' vibrational modes (phonons) to their qubits. At the end of the procedure, the gate separates the collective motion from the qubits, restoring the motional state to its initial state.

Its resistance to ions' early motional condition is a major benefit. Its lack of ion cooling to their motional ground state is a major advantage over earlier gate technologies.

Scalability: A global MS gate applied to a chain of trapped ions can entangle many qubits concurrently (all-to-all connectivity).

Quantum computation relies on high-fidelity two-qubit entangling gates to generate entangled states for quantum advantage. However, physical platforms sometimes limit two-qubit gate selection.

Superconducting quantum processors like IBM's use the controlled-Z (CZ) or CNOT gate for entanglement. However, trapped-ion quantum computing introduced the MS gate. It was designed to overcome experimental problems including thermal mobility sensitivity to perform reliably even without the strict thermal ground state conditions of prior techniques. The MS gate's usefulness and theoretical stability in trapped-ion systems made it a promising cross-platform translator.

As the quantum sector moves toward hardware-agnostic programming, quantum compiler optimization must effectively compile non-native gates across platforms.

Hardware-Efficient Implementation and Characterization

Researchers compiled the MS gate's abstract unitary onto the actual device's native gate set to adapt it to superconducting conditions. The implementation was made "hardware-efficient," decreasing circuit depth, to prevent fault buildup. This optimization yielded a single-CNOT gate breakdown. The MS gate, a maximally entangling operation, is locally equivalent to the CNOT gate but has unique algebraic properties that may benefit certain algorithms.

Full Quantum Process Tomography (QPT) and direct state measurements were employed to characterize the gate's performance.

Input-specific Direct State Measurements confirmed the operation's rationale. When applied on the starting state (the “00” state), the gate should prepare a Bell state. Execution on actual hardware verified its functionality with a 94.2% empirical success probability in the correct two-dimensional Bell state region. The 5.8% state preparation infidelity was caused by population leaking into fake states.

Quantum Process Tomography (QPT): Reconstructing and characterizing the process matrix provided a complete gate quality benchmark. With 96.86% process fidelity, the QPT verified the circuit's mathematical accuracy in simulation.

Strong Operations and Benchmarking

The key finding is that the ibm_nairobi superconducting processor had 92.47% experimental fidelity. This is a successful, nearly parity implementation and matches the device's native CX gate's 93.02% fidelity. Device-specific noise causes decoherence and control errors, as seen by the 4.4 percentage point fidelity reduction between hardware execution and noiseless simulation.

Despite hardware changes during the studies, the MS gate maintained competitive fidelity. The device has qubit readout errors of 1.8% to 9.9% and T1 coherence durations of 63 to 144 microseconds. The consistent performance of the MS gate under non-ideal, heterogeneous conditions makes it a practical and trustworthy high-performance entangling primitive.

Future Compiler Designs and Algorithm Flexibility

The Mølmer-Sørensen gate's low fidelity penalty-effective adaptability impacts future compiler optimization and quantum algorithm design.

We prove high-fidelity alternatives for the native gate to broaden the practical gate set for superconducting architectures. This gives algorithm designers more alternatives because different gates have distinct algebraic properties and entanglement structures for enhancing algorithmic primitives, quantum simulations, or error mitigation measures.