Scalable Optical Links for Controlling Bosonic Quantum Processors

Superconducting quantum computing has the potential to revolutionize computational capabilities. However, scaling up large quantum processors is limited by the cumbersome and heat-conductive electronic cables that connect room-temperature control electronics to quantum processors, leading to significant signal attenuation. Optical fibers provide a promising solution, but their use has been restricted to controlling simple two-level quantum systems over short distances. Here, we demonstrate optical control of a bosonic quantum processor, achieving universal operations on the joint Hilbert space of a transmon qubit and a storage cavity. Using an array of cryogenic fiber-integrated uni-traveling-carrier photodiodes, we prepare Fock states containing up to ten photons. Additionally, remote control of bosonic modes over a transmission distance of 15 km has been achieved, with fidelities exceeding 95%. The combination of high-dimensional quantum control, multi-channel operation, and long-distance transmission addresses the key requirements for scaling superconducting quantum computers and enables architectures for distributed quantum data centers.

💡 Research Summary

The paper addresses one of the most pressing engineering bottlenecks in superconducting quantum computing: the “wiring bottleneck” caused by thousands of coaxial microwave lines that carry control signals from room‑temperature electronics to millikelvin processors. These lines introduce substantial heat loads, signal attenuation, and occupy valuable space, limiting the scalability of current architectures to only a few thousand qubits—far short of the millions required for fault‑tolerant operation.

To overcome these limitations, the authors develop an optical‑to‑microwave control link based on cryogenic, fiber‑integrated uni‑traveling‑carrier photodiode (UTC‑PD) arrays. At room temperature, an electro‑optic modulator (EOM) encodes microwave‑frequency waveforms onto an optical carrier. The modulated light travels through standard single‑mode fiber, which has negligible thermal conductivity and ultra‑low loss (≈0.2 dB km⁻¹), and is delivered to the 4 K stage of a dilution refrigerator. There, the UTC‑PD chiplets convert the optical signal back into a microwave drive without requiring any bias power; heat is generated only by the absorbed optical power and the resulting photocurrent, dramatically reducing the thermal load compared with traditional coaxial lines.

The experimental platform consists of a three‑dimensional superconducting module that includes a transmon qubit, a readout resonator, and a high‑Q storage cavity. Two independent optical links (A and B) are employed: link A drives and reads out the transmon, while link B controls the storage cavity. The authors first demonstrate that the optical links faithfully reproduce the intended microwave pulse shapes, showing identical waveforms to those generated by conventional coaxial drives. Qubit Rabi experiments reveal a linear dependence of the Rabi frequency on the EOM modulation amplitude, confirming that the optical link operates in a linear regime suitable for precise gate calibration. For the cavity, coherent‑state displacement experiments produce photon‑number distributions that follow the expected Poisson statistics, and the displacement amplitude scales linearly with the modulation amplitude, indicating high‑fidelity, linear control of the bosonic mode. No measurable crosstalk is observed between the two channels, validating the frequency‑selective isolation required for simultaneous multi‑device operation.

Beyond simple displacements, the authors tackle universal control of the joint qubit‑cavity Hilbert space using GRAPE‑optimized pulses. These pulses are tailored to the full system Hamiltonian, which includes dispersive cross‑Kerr couplings (χrq ≈ 1.20 MHz, χsq ≈ 0.417 MHz) and the cavity’s self‑Kerr (Kg ≈ 1.32 kHz). By applying synchronized GRAPE pulses through both optical links, they implement an encoding operation that maps the qubit states |g⟩ and |e⟩ onto cavity Fock states |0⟩ and |n⟩ (n = 1–4) and also generate superpositions |0⟩ + i|n⟩. Measured Wigner functions display clear negative regions, confirming non‑classical states. Process fidelities for the prepared Fock states exceed 95 %, demonstrating that the optical link can support the complex, time‑varying amplitude and phase profiles required for high‑dimensional quantum control.

A key scalability test is the transmission of control signals over 15 km of fiber. Despite the long distance, the same GRAPE‑based gates achieve >95 % fidelity, showing that the optical approach decouples control electronics from the cryogenic environment without sacrificing performance. This result opens the door to distributed quantum computing architectures where processors in separate cryostats or even separate facilities can be synchronized via low‑latency optical links.

Finally, the authors showcase a 2 × 2 UTC‑PD array, providing four independent control channels within a single compact chiplet. The absence of inter‑channel interference suggests that the architecture can be scaled to many more channels by tiling additional chiplets, offering a path toward the dense, low‑heat‑load wiring required for million‑qubit systems.

In summary, the work delivers (1) a practical, low‑thermal‑budget optical control infrastructure, (2) high‑fidelity manipulation of bosonic modes up to ten photons, (3) demonstrated long‑distance (15 km) operation with negligible fidelity loss, and (4) multi‑channel scalability. These achievements collectively address the three critical dimensions—high‑dimensional quantum control, long‑distance transmission, and parallel channel operation—required for scaling superconducting quantum processors and for building distributed quantum data centers. Future directions include expanding the channel count, integrating cryogenic optical amplifiers to further reduce required optical power, and exploring quantum‑coherent microwave‑optical transduction for entanglement distribution across the same fiber network.