Coherent errors from codes to couplers (09/18/26)

TL;DR A Magnus-expansion pulse correction (a series that builds the time-evolution operator order by order) suppresses non-adiabatic coupler leakage using only a single real baseband flux control—no second quadrature as in DRAG (derivative removal by adiabatic gate, which cancels leakage with an orthogonal drive) or transitionless driving (a shortcuts-to-adiabaticity method that cancels non-adiabatic couplings with an extra control). In simulated fast coupler-mediated iSWAP gates (excitation-swapping two-qubit gates) between transmons, the correction cuts ramp-induced leakage by up to three orders of magnitude for realistic device parameters.

• Main result: Analytic shaping of a single baseband flux pulse suppresses non-adiabatic leakage when a tunable coupler is ramped for a fast two-qubit gate, without a second orthogonal control channel.
• Hardware target: The method is demonstrated in simulation for coupler-mediated gates between transmons under baseband flux control—the architecture constraint that normally excludes standard DRAG and transitionless-driving corrections.
• Gate impact: For realistic parameters, leakage in a fast iSWAP is reduced by up to three orders of magnitude relative to uncorrected ramps.
• Mechanism: The Magnus-derived correction does more than notch spectral weight at leakage transitions; it cancels the leading non-adiabatic coupling generated by the coupler ramp in the adiabatic frame.
• Scope: The same single-channel shaping extends across single- and two-excitation leakage pathways and can refine other shaped pulses (e.g., Slepian envelopes) used in cQED coupler control.

TL;DR A positive error-correction threshold exists for coherent $Z$-rotation noise—unwanted unitary phase rotations from imperfect control, not random Pauli flips—on quantum LDPC codes (including the surface code) with a bounded number of logical qubits. Maximum-likelihood Pauli recovery then drives the entanglement infidelity down exponentially in the code distance $d$, up to a prefactor linear in the number of physical qubits, whenever every qubit’s rotation angle stays below a size-independent constant.

• Main result: For CSS quantum LDPC codes with a bounded number of logical qubits, coherent per-qubit $Z$ rotations admit a positive threshold under maximum-likelihood Pauli recovery.
• Infidelity scaling: Below that angle bound, the entanglement infidelity is suppressed exponentially in $d$, with only a prefactor linear in the number of physical qubits.
• Scope: The theorem covers the surface code and other topological codes in that family; prior surface-code evidence for coherent-error thresholds was numerical, not a proof.
• Method: Fourier analysis keeps the interference among coherent-error amplitudes, then abstract polymer models and their cluster expansion control the residual logical failure.
• Significance: The argument gives a statistical-mechanical footing for quantum error correction beyond stochastic Pauli noise.

TL;DR An impedance-matched cavity atomic-frequency-comb (AFC) memory—periodic spectral teeth that rephase and recall a stored optical pulse—in ${\mathrm{Eu}}^{3+}:{\mathrm{Y}}_2{\mathrm{SiO}}_5$ stores coherent pulses at $90.1(5)\%$ efficiency. Brewster-angle crystal mounting and a fully cryogenic Fabry–Pérot cavity cut round-trip intracavity loss to $0.37\%$, so modeling finds remaining headroom is set by AFC preparation and its dispersive response rather than by residual cavity loss.

• Main result: Coherent optical pulses are stored and recalled at $90.1(5)\%$ efficiency in an impedance-matched cavity AFC memory in ${\mathrm{Eu}}^{3+}:{\mathrm{Y}}_2{\mathrm{SiO}}_5$.
• Loss engineering: Brewster-angle crystal geometry plus a fully cryogenic Fabry–Pérot cavity reduce round-trip intracavity parasitic loss to $0.37\%$.
• Why impedance matching matters: Matching input-mirror leakage to the absorber-enhanced cavity loss lets weak rare-earth absorption capture the incident field without free-space reabsorption limits.
• Bottleneck shift: Quantitative modeling shows further gains toward unity efficiency are limited mainly by imperfect AFC preparation and the comb’s associated dispersive response, not by leftover cavity loss.
• QIS stake: High-efficiency rare-earth AFC memories are a photonic–matter interface for quantum networking once optical depth no longer caps recall fidelity.

TL;DR LIGO traps the differential motion of its 40-kg mirrors into a sub-Hz-linewidth optomechanical mode—a radiation-pressure-dominated mechanical resonance—and engineers the quantum state entering the dark port (the normally unused interferometer output) so readout imprecision and measurement back-action (radiation-pressure force noise) partially cancel. Observed motion near resonance drops by ~47%, demonstrating back-action evasion and quantum reservoir engineering (shaping the effective bath via the input optical state) at kilogram scale.

• Main result: Correlations between imprecision and back-action, set by the engineered dark-port state, suppress observed mirror motion near the trapped resonance by ~47%.
• Platform: Differential motion of LIGO’s 40-kg test masses is optically trapped into a sub-Hz-linewidth optomechanical mode in a band where radiation-pressure back-action dominates the dynamics.
• Mechanism: Continuous displacement sensing normally trades readout imprecision against measurement back-action under the Heisenberg limit; engineered imprecision–back-action correlations partially cancel those contributions in the observed spectrum.
• Analysis: A noise decomposition into imprecision, back-action, and correlation terms isolates the origin of the suppression.
• Significance: The experiment shows quantum back-action evasion and quantum reservoir engineering in a macroscopic optomechanical system relevant to gravitational-wave detectors.

TL;DR Preferential nitrogen incorporation along CVD step bunches yields a positionally disordered, quasi-1D chain of optically dark substitutional-nitrogen P1 centers probed by nearby NV centers, establishing a materials route to low-dimensional dipolar spin ensembles and enabling correlation-spectroscopy studies of infinite-temperature dipolar spin transport.

• Main result: Preferential N incorporation along chemical-vapor-deposition step bunches laterally and vertically confines a disordered P1–NV spin chain in diamond and establishes its quasi-1D character.
• Probe platform: Optically dark, unpolarized P1 centers (substitutional nitrogen) form the interacting bath; optically addressable NV centers act as local sensors of that bath.
• Transport probe: Correlation spectroscopy—a two-time NV measurement of local spin autocorrelations—is used to investigate infinite-temperature dipolar spin transport in the dark P1 ensemble.
• Why it matters: The work gives a materials-based path to engineer low-dimensional solid-state spin systems beyond bulk magnetic insulators, with single-defect readout of many-body dipolar dynamics.