Research
Building a useful quantum computer will require quantum error correction: encoding each logical qubit redundantly across many noisier physical degrees of freedom. In conventional qubit-based architectures this redundancy carries a steep hardware cost: every physical qubit needs its own control and readout lines, all thermalized inside a dilution refrigerator.
Our group develops a hardware-efficient alternative, inspired by the memory hierarchy of classical computers. Just as a classical processor operates on a small, fast register while most data resides in denser, cheaper memory, we build quantum processors in which a single superconducting circuit controls a register of long-lived microwave cavity modes, a random-access quantum memory. Cavity photon lifetimes reach milliseconds to tens of milliseconds, one to two orders of magnitude longer than superconducting qubits, and their harmonic-oscillator structure is naturally suited to hardware-efficient bosonic error correction and to quantum simulation of systems with continuous degrees of freedom.
Our research is organized around four interconnected thrusts.
I. Processor development: random-access quantum memories
We build successive generations of multimode cQED processors in which one nonlinear circuit provides universal control over many cavity modes. Our current devices include a ten-mode bosonic memory with fast, error-resilient sideband control [PRX 16, 011058 (2026)] and a cascaded random-access quantum memory, in which a buffer cavity acts as a cache between the processor and a multimode storage cavity, with quantum information shuttled between memory layers by a flux-biased SNAIL coupler [arXiv:2503.13953]. We have demonstrated random-access SWAP operations between the buffer and any of seven storage modes in a few microseconds, at fidelities approaching 99%, while maintaining millisecond cavity coherence. In parallel, we are developing fluxonium qubits as high-coherence ancillas for 3D circuit QED (with Angela Kou, UIUC).
Our cascaded random-access quantum memory: the layered architecture (left), and the MMC2 device with the measured coherence times of its cavity modes (right).
II. Bosonic control and quantum error correction
Controlling a high-Q cavity through a weakly coupled ancilla, fast and without inheriting the ancilla’s errors, is the central control problem of cavity-based quantum computing. We developed fast sideband control of weakly coupled multimode memories, and used it to demonstrate the first universal Jaynes–Cummings control of an oscillator: arbitrary unitaries compiled from ancilla rotations and JC interactions, with leakage suppressed by construction and ancilla relaxation detectable by design [arXiv:2605.18658]. With the SQMS Center at Fermilab, we extended this control framework to ultracoherent niobium SRF cavities with photon lifetimes of 15–20 ms and dephasing times beyond 40 ms, the highest-coherence multimode bosonic platform to date, preparing Fock states up to N = 20 and demonstrating two-mode entanglement with 99.9% post-selected fidelity [arXiv:2506.03286]. We are using these tools for bosonic quantum error correction in a multimode system.
Error-resilient control of multimode memories: fast sideband and universal Jaynes–Cummings control of a ten-mode bosonic memory.
III. Mitigating decoherence and quantum sensing
Superconducting circuits are exquisite sensors of their environment and of the materials they are built from. We use this sensitivity to identify, understand, and mitigate the microscopic sources of qubit decoherence, and to probe new quantum materials. With the Gershenson group (Rutgers), we used charge-parity monitoring to probe nonequilibrium quasiparticles in transmon qubits, and developed a gap-engineering technique that suppresses quasiparticle poisoning by three orders of magnitude, preserving charge parity for minutes rather than sub-millisecond timescales, with a single additional evaporation step [arXiv:2309.02655]; the approach has since been validated by several groups. We are now measuring how cryogenic microwave materials (Teflon, Eccosorb, YIG) thermalize at millikelvin temperatures, using a multimode cavity as an in-situ thermometer, working to understand and mitigate a common source of qubit decoherence beyond the T1 loss set by materials limitations. The photon-counting techniques behind this work previously set a three-orders-of-magnitude improved exclusion limit on hidden-photon dark matter. With Eva Andrei’s group (Rutgers), we are also integrating moiré materials into superconducting resonators to probe the superfluid stiffness and quantum geometry of flat-band superconductors.
Sensing and mitigating decoherence: gap-engineered junctions that suppress quasiparticle poisoning (left), a multimode probe cavity with a dielectric dipper for in-situ thermometry of cryogenic materials (center), and a moiré-cQED device for probing flat-band superconductors (right).
IV. Quantum simulation of strongly correlated matter
Many problems in quantum chemistry and quantum field theory involve bosons, rotors, and large spins that are inefficient to encode in qubits but map naturally onto cavity modes. We are developing hybrid continuous-variable/discrete-variable simulations of strongly coupled rotational–vibrational molecular dynamics (with Yuan Liu, NC State), quantum machine learning and constrained optimization for chemistry on bosonic hardware (with Victor Batista, Yale), and digital simulations of lattice gauge theories on co-designed, application-specific processors. On the theory side, we showed that non-Gaussian resources restore measurement-induced phase transitions in multimode bosonic systems (with Jed Pixley, Rutgers), and identified a novel monitored bosonic phase with no qubit analog [arXiv:2603.13125].
Hybrid CV-DV quantum simulation: encoding rotors and anharmonic vibrations in pairs of cavity modes (left), and the monitored bosonic circuit used to study measurement-induced phase transitions (right).
Summaries of our recent papers, with figures, are collected on the Highlights page. Earlier results from the group and from Vatsan’s postdoctoral work (the quantum flute, multimode photon blockade, the random-access processor, heavy fluxonium, and dark matter detection) are described on the Past Research page. For the full list of papers, see Publications.