Package modes, altermagnet qubits, modular codes (09/22/26)

TL;DR Four commercial-solver-compatible recipes extract bare electromagnetic couplings between qubits and package modes—the cavity-like standing waves of the chip enclosure—via avoided level crossing (minimum eigenmode splitting at resonance), the energy participation ratio (fraction of mode energy in a junction), induced electromotive force, and the impedance matrix. On a $10\times 10$ transmon array they agree to $<5\%$, giving a practical path to map package-mediated crosstalk at hundred-qubit scale.

• Problem: High-fidelity control needs accurate qubit–package electromagnetic couplings, but extraction must span Josephson-junction nanometers to centimeter-scale packages.
• Toolkit: The paper adapts four extraction routes—avoided level crossing, energy participation ratio, induced electromotive force, and impedance-matrix fitting—to commercial 3D electromagnetic solvers.
• Benchmark: On a $10\times 10$ transmon array, all four methods return consistent couplings to standing package modes, with maximum relative difference $<5\%$.
• Physics payoff: The comparison clarifies how package-mode field structure sets qubit couplings, which is the lever for package-mediated crosstalk and Purcell pathways in cQED hardware.

TL;DR Altermagnets—collinear magnets with zero net magnetization but momentum-dependent spin splitting—make the Josephson potential of a superconductor–altermagnet–superconductor junction programmable by junction design. That engineered potential yields a transmon-like qubit whose coherent two-Cooper-pair tunneling (a $2\varphi$ Josephson term) adds large anharmonicity and parity protection against charge noise, with flux bias available for further noise suppression.

• Platform: An S–AM–S junction (superconductor–altermagnet–superconductor) uses altermagnetic spin splitting without net magnetization to reshape the current–phase relation by geometry and barrier parameters rather than by conventional ferromagnetism.
• Main result: Inverse design of the Andreev spectrum produces prescribed Josephson potentials, including a dominant $2\varphi$ junction where the supercurrent prefers period $\pi$ in phase $\varphi$ via coherent tunneling of two Cooper pairs.
• Qubit: Capacitively shunting that AM Josephson element gives a transmon-like circuit with large anharmonicity while parity protection from the $2\varphi$ term suppresses charge-noise sensitivity.
• Control: External magnetic flux tunes the qubit spectrum; under suitable bias the architecture further suppresses both charge and flux noise relative to a standard single-junction transmon.
• cQED takeaway: Altermagnets become a materials knob for Josephson-potential engineering—an alternative route to high-coherence, high-anharmonicity, flux-tunable superconducting qubits without needing exotic topological Josephson elements.

TL;DR A cascade of traveling-wave parametric amplifiers (TWPAs)—a short 736-element Superconducting Nonlinear Asymmetric Inductive eLement (SNAIL) stage for low added noise, then a long 1632-element stage for gain—delivers nearly 30 dB over a tunable ~1 GHz band while keeping added noise near the quantum limit ~½ℏω. For cQED readout chains, that combination pushes HEMT contribution out of the noise budget without trading away broadband, quantum-limited first-stage performance.

• Architecture: A short SNAIL TWPA (736 elements) leads a long conventional SNAIL TWPA (1632 elements), splitting the usual length-versus-dissipation tradeoff so the first stage sets noise and the second supplies gain.
• Main result: The cascade reaches nearly 30 dB total gain over a tunable bandwidth of approximately 1 GHz.
• Noise: Added noise stays close to the quantum limit, reported as ~½ℏω, despite the high cascade gain.
• Compression: The 1 dB compression point is −101 dBm, comparable to best-in-class single TWPAs that typically offer only ~20 dB gain.
• Hardware stakes: Large pre-HEMT gain suppresses amplitude and phase noise from the following HEMT stage, which matters for multiplexed qubit readout and microwave entanglement experiments.

TL;DR On a trapped-ion processor, the authors implement for the first time a full modular fault-tolerant toolkit for the perfect $[[5,1,3]]$ code—the smallest quantum error-correcting code that corrects any single-qubit error and is non-CSS (not a Calderbank–Shor–Steane stabilizer code with separate $X$/$Z$ checks). That suite covers logical teleportation between modules plus magic-state injection for universal logical control, showing compact non-CSS codes can host modular FT computation.

• Main result: The experiment demonstrates every logical primitive needed for modular fault-tolerant universal control of the $[[5,1,3]]$ code, including state preparation, real-time-feedback QEC, logical measurement, magic-state preparation (resource states that enable non-Clifford gates), logical entangling operations, and magic-state injection.
• Modular link: Logical state teleportation moves encoded information between qubit registers, realizing the quantum channel that stitches modules without requiring a larger CSS code.
• Spectator errors: Idling logical qubits pick up measurable logical spectator error while other registers run operations, quantifying crosstalk cost in a multi-module layout.
• Tomography: A logical Pauli quantum process tomography protocol characterizes logical channels while cutting the sampling overhead of full logical process tomography.

TL;DR A unified entropic converse pins down the direct static capacity region of a bipartite state—the asymptotic Alice-to-Bob trade-off among classical communication, quantum communication, and entanglement generated or consumed from that shared state alone. A supporting-hyperplane formula then yields exact single-letter regions for erased pure states, symmetrically extendible states, and locally flagged pure-state mixtures, while showing that a nonorthogonal discarded quantum memory can beat every efficient local instrument.

• Main result: One finite-block converse, built from no-signalling, the chain rule, strong subadditivity, and continuity of conditional entropy, derives all three rate inequalities from a single local instrument (a measurement-plus-residual-state map on Alice’s share).
• Capacity formula: A supporting-hyperplane characterization produces the static capacity formula, which decides single-letterization of the region after regularization.
• Erased pure states: For a pure state of entanglement entropy $h$ erased with probability $p$, the complete region in classical–quantum–entanglement rate space is the convex hull of the origin and $(0,-ph,(1-p)h)$, plus the unit-resource cone (the free trade-offs among the three resources once one ebit, qubit, or cbit is available).
• Other exact families: Symmetrically extendible states and locally flagged pure-state mixtures also admit complete characterizations; maximally correlated states reduce the instrument optimization to an exact matrix program with an additive outer bound.
• Surprise & open gaps: An explicit qubit example shows a nonorthogonal discarded quantum memory can outperform every efficient instrument and every instrument with conditionally commuting discarded states; full single-letter forms for general Hadamard states and remaining dephased-state trade-offs stay open.