{"id":"1a72b904-293a-496a-be7b-cc59be0ae0a5","arxiv_id":"2607.27067","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":5,"one_line_summary":"A germanium two-hole singlet–triplet qubit is made fully degenerate at idle and driven with all-electrical orthogonal X and Z gates at 99.53% average physical fidelity.","lead":"Researchers built a two-hole germanium singlet–triplet qubit that can sit idle with both control knobs off, then turn on pure X or pure Z rotations using only DC voltage pulses. That removes always-on magnetic gradients and unwanted idle precession, which matter for scaling semiconductor spin processors.","discovery_kind":"new_method","skeptic_critique":{"model":"grok-4.5","headline":"Residual idle Hamiltonian and approximate virtual-gate orthogonality remain the soft underbelly of the degenerate-idle claim.","rationale":"The reader correctly isolates the load-bearing assumption: calibrated virtual gates plus a stable g-matched idle. Main-text data (spectroscopy, orthogonal Rabi maps, Fig. 3e null line at t_Z=π/2, RB) make a credible single-qubit demonstration and keep correctness risk low; the fidelity extraction (p=0.973, 2.83 phys/Clifford → F_gate=99.53(9)%) is standard. What is not sealed is a tight experimental bound on residual idle splitting and on cross-axis leakage under the exact pulse amplitudes used in RB, plus the prospective character of two-qubit scaling. That is enough to keep the verdict CONDITIONAL rather than ACCEPT, and not enough to push toward REJECT. No stronger internal inconsistency or hidden assumption in the Hamiltonian (Eq. 1) or RB protocol was found. Public data release and an explicit idle-Ramsey bound would be the natural next steps to upgrade the verdict.","tokens_in":18363,"tokens_out":700,"duration_ms":44996,"concrete_test":"Perform a Ramsey-style free evolution at the nominal idle point (V_J=V_ΔEZ=0) for τ up to ≳5 µs after both S and ↑↓ initialization; fit any residual oscillation. If a coherent frequency ≳100–200 kHz appears (or if echo-revivable static offset is comparable to 1/T*_2,X), the degenerate-idle claim is quantitatively weakened and RB should be re-run with idle-phase tracking or tighter virtualization.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim requires that at the calibrated idle point both J and ΔEZ are simultaneously negligible, and that the virtual controls V_ΔEZ=(−VB12,VB1) and V_J=(VB12,0.5·VB1) generate pure X and Z rotations under the amplitudes and sequential pulsing used in RB (~100 ns gates; Methods E, Extended Data Figs. 2–4). Evidence is spectroscopy (Fig. 2c–d), single-axis Rabi maps (Fig. 3a–d), a two-axis checkerboard (Fig. 3e), and contrast under complementary init/readout bases. That evidence is supportive but incomplete: the compensation ratio is quoted as “approximately 0.5,” positive V_ΔEZ is acknowledged to turn on finite J, and no direct long-time free-evolution / Ramsey trace at nominal idle (V_J=V_ΔEZ=0) is reported in the main text to bound residual coherent splitting. If a static or slowly drifting residual ω_idle ≳ few×100 kHz remains, unwanted phase accumulates during idles and the advertised advantages—true degeneracy, no rotating-frame tracking, fully orthogonal axes—degrade even while single-axis Rabi contrast and RB numbers can still look acceptable. Multi-qubit leakage suppression via engineered ΔE_ST12_Z (Methods G) is purely theoretical and not required for the single-qubit claim, but the same g-tuning knobs that set degeneracy are what would have to keep inter-qubit detuning large; that tension is untested.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The manuscript reports a degenerate singlet–triplet (DST) qubit encoded in two hole spins in a Ge/SiGe double quantum dot. By electrically tuning the anisotropic g-factors of the two holes, the authors identify a magnetic-field and gate-voltage regime in which both the exchange J and the Zeeman difference ΔEZ vanish, so that |S⟩ and |T0⟩ are degenerate at idle. Baseband virtual-gate pulses then independently activate ΔEZ (X rotations) and J (Z rotations). They demonstrate single-axis Rabi maps, a two-axis sequence confirming axis alignment, coherence times T*2,X = 1.55(1) μs and T*2,Z = 2.32(4) μs, and randomized benchmarking yielding an average physical single-qubit gate fidelity of 99.53(9)% for ~100 ns gates. They further show electrical tuning of the degenerate point over a range of field orientations, with enhanced T*2,X near in-plane B, and discuss multi-qubit leakage suppression via engineered inter-qubit Zeeman detuning.","tokens_in":18713,"tokens_out":1747,"duration_ms":40905,"significance":"If the results hold, this is a clear advance for semiconductor spin qubits. Conventional ST qubits suffer from a fixed, always-on ΔEZ that prevents true orthogonal two-axis control and forces rotating-frame phase tracking. Achieving a voltage-tunable degenerate idle with fully baseband orthogonal X/Z control in a two-dot encoding—without micromagnets or three-spin EO overhead—addresses a long-standing control bottleneck and is directly relevant to scaling under a shared global field. The reported RB fidelity is competitive with state-of-the-art ST work, the g-factor spectroscopy and virtual-gate calibration are concrete and reproducible in principle, and the field-angle tunability of the degenerate point is a useful practical knob. The multi-qubit leakage discussion (Methods G) is theoretical but points to a coherent scaling path. Overall this is a solid experimental contribution appropriate for a high-profile condensed-matter/quantum-device venue.","major_comments":[{"comment":"Central claim of a degenerate idle (Abstract; §II.B–C; Eq. 1): the manuscript asserts that both J and ΔEZ vanish at the operating point, yet the main text does not report a direct free-evolution / Ramsey measurement at nominal idle (VJ = VΔEZ = 0) that bounds any residual coherent splitting ω_idle. Spectroscopy (Fig. 2b–d) and complementary init/readout contrast (Methods E; Extended Data Figs. 2–4) are supportive, and Hahn-echo behavior (Extended Data Fig. 6) is consistent with ΔEZ-dominated noise, but a quantitative upper bound on residual |J| and |ΔEZ| during idling (e.g., from long idle free induction or a calibrated phase-accumulation sequence) is needed to substantiate “degenerate” and “no unwanted state rotations during idling” at the level claimed. Without it, residual ω_idle ≳ few×100 kHz could still accumulate phase over multi-gate sequences even if single-axis Rabi and RB look","section":"§II.B–C, Eq. 1; Methods E"},{"comment":"Virtual-gate orthogonality under the amplitudes used for gates and RB (§II.C; Methods E): VΔEZ = (−VB12, VB1) and VJ = (VB12, 0.5·VB1) are calibrated with a compensation ratio quoted as “approximately 0.5” (Extended Data Fig. 4), and the text acknowledges that positive VΔEZ turns on finite J (Extended Data Figs. 2–3). For the claim of “fully orthogonal” X and Z axes at tπ ≈ 100 ns, the paper should state the residual cross-axis angle or residual unwanted Hamiltonian component under the actual pulse amplitudes used in Fig. 3 and in the Clifford set, not only in the spectroscopy/low-amplitude limit. A short quantification (e.g., residual oscillation contrast in the “wrong” init/readout basis at the RB operating point, or fitted axis tilt from the Fig. 3e checkerboard) would close this gap.","section":"§II.C; Methods E; Extended Data Figs. 2–4"},{"comment":"Fidelity interpretation (§II.D, Fig. 3f): Fgate = 99.53(9)% is extracted from Clifford RB with a shared depolarizing parameter and 2.83 physical gates per Clifford (Methods F). The text does not discuss what limits this fidelity (T*2 vs. pulse calibration vs. SPAM vs. residual non-orthogonality) or whether idle intervals between physical gates in the RB sequence were zeroed or phase-tracked. A brief error budget or at least a statement of idle handling during RB is load-bearing for interpreting the number as evidence of a clean degenerate idle rather than of short, high-contrast gates alone.","section":"§II.D, Fig. 3f; Methods F"}],"minor_comments":[{"comment":"Fig. 2d inset and positive-VΔEZ regime: the finite-ΔEZ splitting attributed to VB12 when J is off is important for understanding cross-talk; consider marking the VΔEZ = 0 and VJ = 0 loci more explicitly on the Rabi maps in Fig. 3a,c.","section":"Fig. 2d; Fig. 3a,c"},{"comment":"Eq. 1 uses H = −(J/2)σz + (ΔEZ/2)σx. Confirm sign conventions relative to the Bloch-sphere drawing in Fig. 2a and to the pulse polarities in Fig. 3 so that “X” and “Z” labels are unambiguous for readers reconstructing the sequences.","section":"Eq. 1; Fig. 2a"},{"comment":"Methods G (two-qubit leakage) is useful but purely theoretical; a sentence in the main Discussion clarifying that no two-qubit experiment is reported would avoid over-reading the scaling claim in the Abstract (“route towards multi-qubit scaling”).","section":"Abstract; §III; Methods G"},{"comment":"Extended Data Fig. 1 and readout calibration: state the typical single-shot assignment fidelity or separation SNR used when converting |S11| to singlet probability, since RB and T*2 fits depend on that calibration.","section":"Methods C; Extended Data Fig. 1"},{"comment":"Typographical/consistency: “adegeneratequbit” spacing in §II.C; ensure ΔEZ vs. ΔEZ and VB12 notation is uniform across main text and Extended Data; arXiv date stamp “July 30, 2026” is fine but check journal submission metadata.","section":"§II.C"},{"comment":"Fig. 4c: the claim of no clear gate-fidelity advantage below ~40 mT is important; if RB or quality-factor data exist at 24 mT (Extended Data Fig. 9), a one-line comparison in the main text would strengthen that conclusion.","section":"§II.E; Fig. 4c"}],"recommendation":"minor_revision","confidential_remarks":"The work is experimentally solid and well matched to a top condensed-matter/quantum-information journal. The residual-idle and virtual-gate-orthogonality points are the only issues that touch the central “degenerate + fully orthogonal” claim; they look fixable with existing apparatus (a Ramsey/idle bound and a residual-tilt number) without new devices. I would not require a two-qubit demonstration for acceptance. Novelty relative to recent Ge ST and EO work appears genuine on the simultaneous J=ΔEZ=0 idle with baseband orthogonal control; no citation red flags stood out."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"Punchline: they actually ran a two-hole Ge singlet–triplet qubit at a point where both J and ΔEZ are electrically off, then drove pure baseband X and Z with ~100 ns gates and 99.53% physical RB. That is the thing worth knowing.\n\nWhat is new is not g-anisotropy or ST encoding—those are established—but the operating point. Prior ST work lives with always-on ΔEZ; EO gets degeneracy with three spins. Here they use voltage-tunable hole g-tensors to find (and retune) a g-matched angle where spectroscopy shows the two resonances coincide, kill J with the barrier, and define virtual gates V_ΔEZ and V_J that give independent axes. Fig. 2 spectroscopy, Fig. 3 Rabi maps, the two-axis checkerboard, complementary init/readout contrast, and joint singlet/triplet RB are the right evidence stack. Coherence split (T2,X* vs T2,Z*) and the in-plane sweet-spot scan are honest; they even say low-|B| T2 gains do not yet buy gate quality factor. Citations look appropriate.\n\nSoft spots, in proportion. The idle claim rests on calibrated virtual gates and a compensation ratio quoted as “approximately 0.5.” Positive V_ΔEZ turns on some J; they acknowledge it. A direct long free-evolution / Ramsey bound at nominal idle (VJ = VΔEZ = 0) is not front-and-center in the main text—that would tighten the residual-Hamiltonian story the stress-test worries about. I do not think that sinks the result: single-axis contrast, axis-alignment sequence, and RB still constrain residual splitting, but a referee will ask for a cleaner idle bound. Multi-qubit leakage via engineered inter-qubit ΔEST12_Z is Methods theory only; fine as outlook, not a demonstrated claim. Data-to-Zenodo is promised, not yet public.\n\nWho it is for: anyone building baseband spin control or global-field Ge arrays. Not a general QC audience. Math is standard ST Hamiltonian; data quality is solid experimental mesoscopics.\n\nI would send this to peer review. Engage if you care about ST/EO control tradeoffs; skip if you only track two-qubit fidelities.","headline":"Real experimental advance: two-hole Ge ST qubit with electrically zeroed idle and baseband orthogonal X/Z, backed by spectroscopy and RB, not just a rebrand of known g-tuning.","tokens_in":19455,"tokens_out":584,"would_cite":true,"duration_ms":14285,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["73.21.La","03.67.Lx","85.35.Be","71.70.Ej"],"model":"grok-4.5","headline":"A germanium two-hole qubit can idle with both exchange and Zeeman difference off, then use only baseband voltages for independent X and Z rotations at 99.53% gate fidelity.","keywords":["singlet-triplet qubit","degenerate idle","hole spins","germanium quantum dots","all-electrical control","orthogonal rotations","g-factor tuning","baseband pulses"],"falsifier":"Measure the idle evolution of a prepared equatorial state for many microseconds with both control voltages held at the claimed zero point; any coherent precession or contrast loss beyond the reported T2* would show residual J or ΔEZ and falsify true degeneracy and orthogonality.","tokens_in":19217,"feed_emoji":"⚛️","tokens_out":885,"duration_ms":13816,"temperature":0.7,"pith_summary":"Standard singlet–triplet qubits leave a fixed Zeeman energy difference always on, so the idle point keeps rotating and the two control axes are not independent. This paper shows that the electrically tunable, anisotropic g-factors of holes in a germanium double quantum dot can be set so that both the exchange interaction J and the Zeeman difference ΔEZ vanish at once. From that true degenerate idle, ordinary baseband voltage pulses alone turn J and ΔEZ on and off independently, giving clean Z and X rotations on the Bloch sphere. Randomized benchmarking reports an average physical single-qubit gate fidelity of 99.53% for roughly 100 ns gates, and the same electrical knobs move the degenerate point across magnetic-field angles into a quieter coherence regime. A sympathetic reader cares because the scheme removes always-on idle rotations, needs no micromagnets or RF drives for single-qubit control, and points toward arrays that share one global field while still allowing local electrical tuning.","feed_headline":"Hole spins idle with both controls off, then rotate on pure voltage","feed_subtitle":"Germanium double-dot qubit reaches 99.53% single-qubit fidelity with fully orthogonal baseband X and Z gates","key_machinery":"The degenerate singlet–triplet (DST) idle point: the voltage-tuned operating regime in which both J = 0 and ΔEZ = 0, so the computational states |S⟩ and |T0⟩ are energetically degenerate and the only Hamiltonian during idle is the identity; virtual gates VΔEZ and VJ then restore each term independently.","core_discovery":"By electrically tuning the anisotropic g-factors of two hole spins in a Ge double quantum dot, both the exchange J and the Zeeman difference ΔEZ can be set simultaneously to zero, creating a degenerate S–T0 idle from which baseband voltage pulses alone implement fully orthogonal X (ΔEZ) and Z (J) rotations, with randomized-benchmarking average physical single-qubit fidelity 99.53(9)% at ~100 ns gate duration.","pith_inferences":["If g-tensor uniformity improves with materials, the same electrical knobs could still be used deliberately to create controlled inter-qubit Zeeman offsets for leakage-free two-qubit gates.","The demonstrated electrical steering of the degenerate point suggests a calibration loop that could keep many DST qubits degenerate under slow global-field drift without per-qubit magnets.","Combining the DST idle with existing hole-spin shuttling would give a baseband-only architecture from initialization through multi-qubit operations."],"forward_implications":["Single-qubit gates need only baseband pulses, removing on-chip RF delivery and heating for this encoding.","The idle point can be parked where hyperfine noise is weaker (in-plane low field), lengthening coherence without sacrificing two-axis control.","Neighboring DST qubits can be electrically g-matched under one global B while still keeping inter-qubit Zeeman contrast large enough to suppress exchange leakage.","No fixed micromagnet gradients are required, simplifying fabrication of larger arrays.","Gate quality is already competitive with the best reported singlet–triplet devices while adding true degeneracy."],"fun_headline_variants":["Ge hole spins reach degenerate S-T0 idle with J and ΔEZ both zero","All-electrical orthogonal X and Z gates on degenerate singlet-triplet qubit","Baseband voltages alone drive fully orthogonal rotations in DST hole qubit","Tunable g-factors set both controls to zero for 99.53% fidelity gates","Degenerate idle point enables pure-voltage orthogonal control of hole spins"],"cache_read_input_tokens":128,"weakest_assumption_plain":"That the calibrated voltage combinations that turn J and ΔEZ on and off stay cleanly independent under the pulse sizes and idle times used for gates, without leftover always-on couplings or drift that would reintroduce a finite idle Hamiltonian.","fun_headline_variants_meta":{"raw":{"variants":["Ge hole spins reach degenerate S-T0 idle with J and ΔEZ both zero","All-electrical orthogonal X and Z gates on degenerate singlet-triplet qubit","Baseband voltages alone drive fully orthogonal rotations in DST hole qubit","Tunable g-factors set both controls to zero for 99.53% fidelity gates","Degenerate idle point enables pure-voltage orthogonal control of hole spins"]},"model":"grok-4.5","effort":"low","cost_usd":0.004209,"raw_usage":{"total_tokens":1338,"prompt_tokens":846,"num_sources_used":0,"completion_tokens":102,"cost_in_usd_ticks":42088000,"prompt_tokens_details":{"text_tokens":846,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":390,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":846,"tokens_out":102,"duration_ms":7435,"temperature":1.0,"reasoning_tokens":390,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-30T12:09:35.904852+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Measure the idle evolution of a prepared equatorial state for many microseconds with both control voltages held at the claimed zero point; any coherent precession or contrast loss beyond the reported T2* would show residual J or ΔEZ and falsify true degeneracy and orthogonality.","supporting_citations":[],"review_version":1}