{"id":"e809f5a7-8704-41d3-b3dc-fbb2775de04f","arxiv_id":"2506.04977","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Hole spin qubits in unstrained bulk germanium should have much lower g-factor anisotropy, faster Rabi oscillations, and broader operating regions than in strained germanium heterostructures.","lead":"This paper uses numerical simulations to show that confining hole spin qubits at an unstrained, bulk germanium interface greatly reduces the magnetic-field anisotropy that plagues strained germanium qubits. If the simulations hold in real devices, the platform could make germanium spin qubits easier to align and scale to many qubits.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Quality-factor advantage hinges on Eq. (4)'s equal-noise assumption; a moderate increase in bulk-Ge charge noise would erase it.","rationale":"The reader's weakest_assumption identifies exactly the same object: Eq. (4)'s equal-noise model. I considered the alternative candidate—the SI's cool-down strains, which change g-factors and Rabi/LSES magnitudes by an order of magnitude—but that concern does not threaten the central claim: the SI maps show the L-gate Rabi frequency and Q2* increase further under realistic strains, so the qualitative conclusion of reduced anisotropy, faster Rabi oscillations, and larger quality factors survives. By contrast, the noise model directly controls the denominator of Q2*. The comparison is clean only if δV_rms is platform-independent, and the paper gives no experimental or microscopic justification for that equality. The Discussion's own caveat about controlling charge disorder confirms that this is the soft spot. A targeted charge-noise measurement, or a crossover-ratio analysis treating δV_rms^bulk/δV_rms^strained as an unknown, would settle whether the advantage persists. Since the manuscript already discloses the conditionality and the reader's CONDITIONAL verdict reflects it, no change to the verdict is needed.","tokens_in":17194,"tokens_out":7122,"duration_ms":91504,"concrete_test":"Measure the 1/f gate-voltage noise (or the equivalent δV_rms at the Rabi frequency) in a bulk Ge/SiGe single-heterojunction device and in a strained Ge/SiGe quantum well using the same gate stack and readout, then recompute the Q2* maps of Figs. 5f and 6f with Eq. (4) using the measured per-gate values. If the bulk-device δV_rms exceeds the strained-device value by a factor of about 2 or more, the maximum L-gate Q2* advantage reported in Figs. 5f and 7b is not robust; if the ratio is below about 1.5, the advantage survives. A purely analytical version of the same check is to compute the crossover ratio η_crit = Q2*_strained/Q2*_bulk from the equal-noise maps and state the maximum tolerable noise degradation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing assumption is not the strain idealization, which the SI shows only increases Rabi frequencies and LSES while preserving the qualitative advantage, but the noise model in Eq. (4). There, the electrical dephasing rate Γ2* is computed by assigning the same rms voltage fluctuation δV_rms to every gate, and the paper assumes this amplitude is identical in the unstrained bulk-Ge device and the strained reference. The headline quality factor Q2* = 2 f_R T2* is then a ratio of computed Rabi frequencies to computed LSES values, with no input from measured noise. Figures 5f and 6f show a bulk-vs-strained Q2* margin of about 2.2 at the L-gate maximum. The bulk device has roughly 3.2x larger Rabi frequency but also larger LSES, so its electrical T2* is shorter for fixed δV_rms. If real bulk-Ge devices—with the thinner 20 nm GeSi barrier and exposed substrate interface discussed in the paper—show rms gate-voltage noise more than about 2x larger than the strained well, the claimed quality-factor improvement inverts. The authors themselves flag this in the Discussion: charge disorder and noise 'must be carefully controlled,' and the advantage is conditional on two-qubit gates not being much slower. No measured noise data or noise-model validation is provided, so the central quantitative promise is not yet secured.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript proposes confining hole spin qubits at the interface between an unstrained bulk Ge substrate (modeled as a 170-nm-thick well) and a thin GeSi barrier, as an alternative to strained Ge/SiGe heterostructures. Using a finite-volume Poisson solver and a finite-difference Luttinger-Kohn Hamiltonian with published Luttinger and deformation-potential parameters, the authors compute dot dimensions, g-factors, LH mixing, Rabi frequencies, LSES, and quality factors Q2*=2f_R T2* for the five-gate test device of Refs. [37,38]. They compare an unstrained bulk device with a strained reference well (L_w=16 nm), finding that the g-factor anisotropy g_⊥/g_∥ drops from about 50 to about 3, that the L-gate Rabi frequency increases from 6.7 to 21.6 MHz/mV (at V_C=-25 mV and f_L=1 GHz), and that the angular maps of Rabi and Q2* are much broader. The SI reports phonon relaxation, hyperfine dephasing, and order-of-magnitude changes of the spin metrics induced by cool-down strains from the gate stack.","tokens_in":17495,"tokens_out":7957,"duration_ms":92473,"significance":"If the calculations are correct, the work gives a concrete, scalable route to reducing the main drawback of Ge hole qubits, namely the extreme sensitivity of spin properties to the magnetic-field orientation. The central physical mechanism (enhanced HH/LH mixing when the biaxial strain is removed) is robust and follows from standard k.p theory with established parameters. The manuscript also ships detailed numerical machinery (Poisson + 3D k.p, g-matrix formalism) and provides falsifiable predictions for g-factors, Rabi frequencies, and LSES maps. The main quantitative caveat is that the Q2* comparison assumes identical gate-voltage noise in the unstrained and strained devices; this is an assumption, not a result, and the paper's own discussion acknowledges the need for noise control. With that caveat addressed, the work would be a useful contribution to the hole-spin-qubit literature.","major_comments":[{"comment":"The quality-factor comparison in Figs. 5 and 6 rests on Eq. (4), which assigns the same rms voltage fluctuation δV_rms to every gate and uses the same value for the unstrained bulk and strained devices. The resulting Q2* = 2 f_R T2* is therefore a conditional statement, not a measured or predicted ratio. The bulk device's thinner 20-nm GeSi barrier and exposed substrate interface could plausibly have larger charge noise; the authors themselves say in the Discussion that noise must be 'carefully controlled.' Because a roughly twofold increase in δV_rms would invert the L-gate Q2* advantage, the paper should provide a sensitivity analysis (for example, Q2* advantage versus the ratio of noise amplitudes) or explicitly temper the abstract's claim of improved quality factors.","section":"II.C, Eq. (4)"},{"comment":"The main-text numbers in Figs. 2, 5 and 7 describe an idealized, perfectly strain-free Ge layer, while the SI shows that cool-down strains from the gate stack change g_∥ from −0.28 to −0.85 and increase the L-gate Rabi frequency from 21.6 to 228.8 MHz/mV (with Q2* also strongly modified). This idealization is disclosed only in the Methods paragraph, not in the abstract or conclusion. The authors should either move the strain-perturbed results (or at least a summary) into the main text, or state explicitly in the abstract and conclusion that the reported numbers are for the ideal strain-free limit and that realistic gates can change them by an order of magnitude. The qualitative trend survives, but the quantitative promise of the platform should not be presented without this caveat.","section":"SI III, Figs. S4-S6"}],"minor_comments":[{"comment":"There are language issues: 'one of the most promising material' should be 'materials,' and 'shall ease' is stilted; please proofread for English style.","section":"Abstract"},{"comment":"References [16] and [48] both list the same arXiv identifier (2310.05902) but are different works; please correct the identifier for one of them.","section":"References"},{"comment":"The caption reads 'as as a function of V_C'; please remove the duplicated 'as.'","section":"Fig. 4 caption"},{"comment":"The notation ℓ_∥ = sqrt(⟨x^2⟩)=sqrt(⟨y^2⟩) is used for the in-plane extension; it would be clearer to define it as the quadratic spread along one in-plane axis, given the dot is quasi-circular.","section":"II.B"}],"recommendation":"major_revision","confidential_remarks":"The manuscript leans heavily on the authors' own previous work (Refs. 31, 32, 37, 38, 44), but the central quantities are computed, not fitted, so this is not a circularity problem. The novelty relative to Refs. [33-35] is mostly quantitative; the proposal of unstrained Ge channels has been made elsewhere, but the detailed device modeling here adds genuine value. The paper fits the journal's scope. My main concern is the unqualified presentation of noise-dependent and strain-dependent figures in the abstract; if the authors add the requested sensitivity analysis and caveats, I would support publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe short version: this is a solid modeling paper that makes a credible case for unstrained bulk Ge as a way around the g-factor anisotropy problem in hole spin qubits. The central qualitative prediction—that g_perp/g_par drops from >50 in strained heterostructures to ~3 in unstrained bulk Ge—is robust and comes from standard k.p calculations with published parameters. That alone is worth refereeing. The quantitative promises about Rabi frequencies and quality factors should be treated more cautiously.\n\nWhat's new: the authors compute g-factors, Rabi frequencies, LSES, and quality factors for a gate-defined dot in unstrained bulk Ge, including maps over magnetic field orientation, well-thickness dependence, and gate-voltage dependence. The idea of unstrained Ge was floating around (refs 33–36), but this is the first set of numbers for a realistic device. The method is consistent with their earlier work, and the essential physics is clearly explained. Credit where due: the paper does not oversell. It discloses that the main-text figures assume zero cool-down strain, and the SI shows that realistic gate-imprinted strains change LSES and Rabi frequencies by an order of magnitude. That is honest.\n\nThe soft spots are real but not fatal, and they are mostly in the quantitative claims. First, the noise model in Eq. (4) lumps all electrical fluctuations into gate-voltage noises with the same rms amplitude on every gate, and assumes that amplitude is the same in strained and unstrained devices. The quality-factor margin at the L-gate maximum is about 2.2, so if bulk Ge devices show even ~2x more charge noise—plausible given the thinner barrier and exposed substrate interface—the advantage inverts. The authors flag this in the Discussion, but they provide no measured noise data or model validation. Second, the cool-down strain issue: the SI shows that inhomogeneous strains can shift the numbers substantially, though the qualitative reduction in anisotropy survives. Since the headline numbers are for an idealized strain-free device, a careful referee should push on which numbers survive realistic strains.\n\nThe citation pattern is fine. There is self-citation, but no circular fitting: the target quantities are computed, not fitted to the same data. The central claim is a prediction, not a post-hoc fit.\n\nWho should read this: anyone working on Ge hole spin qubits, especially groups considering scaling to many qubits. It gives a concrete target for experiments. The paper deserves a serious referee. I would send it out, with referees asked to examine the noise model and the strain idealization. My verdict is close to the reader's: conditional acceptance, with the condition being that the quantitative advantage is framed as dependent on noise assumptions not yet validated.\n\nYours,","headline":"Solid modeling case for unstrained Ge hole qubits; the anisotropy reduction is robust, but the quality-factor advantage depends on noise assumptions that are not yet validated.","tokens_in":18011,"tokens_out":3241,"would_cite":true,"duration_ms":31503,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Confining holes at the interface of unstrained, bulk germanium reduces the g-factor anisotropy from about 50 to about 3 and increases the Rabi frequency and quality factor of electric-dipole spin resonance, easing the scaling of hole-spin…","keywords":["hole spin qubits","germanium","unstrained Ge","gyromagnetic anisotropy","heavy-hole light-hole mixing","Rabi frequency","quality factor","quantum dot spin qubits"],"falsifier":"Grow a gate-defined hole dot at an unstrained Ge/GeSi interface, cool it below 100 mK, and measure g_parallel and g_perp for several magnetic-field angles together with the L-gate Rabi frequency at f_L = 1 GHz; if g_perp/g_parallel stays near 50 rather than dropping to about 3, or if f_R does not exceed the strained-well value at the same drive, the central claim is wrong. A second check is a direct measurement of T2* at fixed Larmor frequency, which tests whether the equal-voltage-noise model underlying the quality-factor gain is realistic.","tokens_in":16956,"feed_emoji":"🧲","tokens_out":7251,"duration_ms":77022,"temperature":0.7,"pith_summary":"Strained germanium is a leading material for hole spin qubits, but the heavy-hole character of the confined states makes the spin response strongly anisotropic: the out-of-plane and in-plane g-factors differ by a factor of about 50, so the Rabi frequency and coherence change by an order of magnitude when the magnetic field tilts by a few degrees from the plane. The paper proposes instead to confine holes at the interface of an unstrained, bulk germanium layer, where the heavy-hole/light-hole splitting is set only by vertical confinement and the light-hole mixing is much stronger. Using numerical simulations of a realistic gate-defined dot, it shows that the g-factor anisotropy drops to about 3 while the Rabi frequency and quality factor for electric-dipole spin resonance increase, making the device much less sensitive to the magnetic field orientation. The authors conclude that this extends the operational range of hole spin qubits and should ease scaling to many-qubit arrays.","feed_headline":"Unstrained Ge shrinks qubit g-factor anisotropy from 50 to 3","feed_subtitle":"Simulations show faster Rabi oscillations and higher quality factors across far more magnetic field angles.","key_machinery":"The argument rests on the heavy-hole/light-hole mixing weight $m^{2}$, which is near zero in strained wells because the biaxial strain term 2 b_v (epsilon_parallel - epsilon_perp) in the heavy-hole/light-hole gap stays around 46 meV, but grows to about 17.7% in the unstrained bulk device where the gap is set only by confinement. This mixing is described by the Luttinger-Kohn Hamiltonian solved with finite differences and by the perturbation formulas g_parallel approximately 3q + (6 m0 / Delta_LH)($\\lambda$ <$p_x^{2}$> - $\\lambda$' <$p_y^{2}$>) and g_perp approximately 6kappa + 27q/2 - 2 gamma_h, which show how the in-plane orbital motion and vertical confinement renormalize the g-factors. Rabi frequencies are computed with the g-matrix formalism, and the dephasing time T2* is obtained from the longitudinal spin electric susceptibility by lumping electrical fluctuations into gate-voltage noises of equal rms amplitude.","core_discovery":"The central claim is that confining a hole at the interface of unstrained, bulk Ge produces a ground state with heavy-hole/light-hole mixing of about 17.7% instead of less than 0.2% in strained Ge quantum wells, reducing the gyromagnetic anisotropy g_perp/g_par from roughly 50 to about 3. In the modeled device, the maximal L-gate Rabi frequency is f_R/V_ac = 21.6 MHz/mV in the bulk device versus 6.7 MHz/mV in a strained 16-nm well, and the quality factor Q2* = 2 f_R T2* is larger and its angular peak much broader (FWHM delta_theta = 12.3 degrees vs 4.8 degrees). The reduction of anisotropy is traced to the closing of the heavy-hole/light-hole bandgap when biaxial strain is absent, which amplifies the heavy-hole/light-hole mixing that mediates both the g-factor corrections and the spin-orbit coupling driving the Rabi oscillations.","pith_inferences":["If the equal-noise assumption in the model gives way and real bulk devices show higher charge noise from the substrate interface, the quality-factor advantage could narrow; a direct noise measurement on a fabricated device would settle this before large arrays are committed.","The tunability of the heavy-hole/light-hole mixing via the GeSi buffer composition suggests a design variable that could trade g-factor anisotropy against dephasing time, possibly allowing different qubits in one array to be matched to different tasks.","Isotopic purification, which the authors note would benefit all germanium qubits, may be even more valuable in unstrained Ge because the hyperfine dephasing time peaks in-plane and is softened by the stronger mixing there.","The sweet lines of zero longitudinal spin electric susceptibility sit well separated from the Rabi hot spots in the bulk device, suggesting that multi-gate driving schemes could operate each qubit near its sweet line without sacrificing speed."],"forward_implications":["A qubit in unstrained Ge can be operated at the same Larmor frequency with a magnetic field pointing anywhere over a much wider range of angles, relaxing the alignment constraints that currently limit many-qubit arrays.","The L-gate Rabi frequency in the modeled bulk device reaches 21.6 MHz/mV compared with 6.7 MHz/mV in the strained well, so single-qubit rotations are about three times faster at the same drive.","The quality factor Q2* = 2 f_R T2* is larger and its angular peak is much broader (12.3 degrees vs 4.8 degrees full width at half maximum) in the bulk device, so the operating point is less fragile to field misalignment and gate variability.","Phonon-limited relaxation times stay above 6.5 ms at 1 GHz despite the stronger spin-orbit coupling, so T1 does not limit operation in the predicted regime.","Inhomogeneous cool-down strains from the gate stack can further enhance L-gate Rabi frequencies and quality factors, so strain control becomes part of the device design space."],"supporting_citations":[{"why":"reports the recent demonstration of a high-mobility hole gas at an unstrained Ge/GeSi interface, the platform this paper models.","marker":"[36]"},{"why":"provides the finite-difference Luttinger-Kohn solver and g-matrix methodology used for the dot wavefunctions and spin dynamics.","marker":"[37]"},{"why":"supplies material parameters and the treatment of inhomogeneous cool-down strains used in the comparison and in the Supplementary Information.","marker":"[38]"},{"why":"introduces the gate-voltage noise model and the geometry of dephasing sweet spots used to compute T2* and quality factors.","marker":"[31]"},{"why":"previous strain-engineering strategy for reducing g-factor anisotropy that the unstrained-bulk route is designed to improve upon.","marker":"[32]"},{"why":"reports experimental strained-Ge qubits with strongly anisotropic g-factors, providing the benchmark the unstrained device is meant to beat.","marker":"[23]"},{"why":"defines the g-matrix formalism from which Rabi frequencies are computed.","marker":"[42]"},{"why":"gives the perturbation expressions for g_parallel and g_perp in terms of heavy-hole/light-hole mixing that explain the anisotropy reduction.","marker":"[44]"}],"fun_headline_variants":["Unstrained Ge cuts spin qubit anisotropy from 50 to 3","Bulk Ge qubits: 3x faster Rabi, 16x lower g-factor anisotropy","Hole qubits in bulk Ge: Rabi up 3x, angle tolerance 2.5x wider","Mixed heavy-light holes in bulk Ge tame qubit g-factor","Unstrained Ge: qubits faster, g-factor flatter, angles broader"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The predicted quality-factor improvement assumes that electrical noise can be described as the same rms voltage fluctuation on every gate in both strained and unstrained devices; if a real bulk germanium device has different charge noise, for example from the thinner GeSi barrier or the exposed substrate interface, the advantage could be reduced or lost.","fun_headline_variants_meta":{"raw":{"variants":["Unstrained Ge cuts spin qubit anisotropy from 50 to 3","Bulk Ge qubits: 3x faster Rabi, 16x lower g-factor anisotropy","Hole qubits in bulk Ge: Rabi up 3x, angle tolerance 2.5x wider","Mixed heavy-light holes in bulk Ge tame qubit g-factor","Unstrained Ge: qubits faster, g-factor flatter, angles broader"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000925,"raw_usage":{"total_tokens":3932,"prompt_tokens":884,"completion_tokens":3048,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":500,"completion_tokens_details":{"reasoning_tokens":2936}},"tokens_in":500,"tokens_out":3048,"duration_ms":25587,"temperature":1.0,"reasoning_tokens":2936,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T10:29:24.507508+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Grow a gate-defined hole dot at an unstrained Ge/GeSi interface, cool it below 100 mK, and measure g_parallel and g_perp for several magnetic-field angles together with the L-gate Rabi frequency at f_L = 1 GHz; if g_perp/g_parallel stays near 50 rather than dropping to about 3, or if f_R does not exceed the strained-well value at the same drive, the central claim is wrong. A second check is a direct measurement of T2* at fixed Larmor frequency, which tests whether the equal-voltage-noise model underlying the quality-factor gain is realistic.","supporting_citations":[{"cited_title":"Bosco, M","cited_arxiv_id":null,"evidence_quote":"reports the recent demonstration of a high-mobility hole gas at an unstrained Ge/GeSi interface, the platform this paper models."},{"cited_title":"Martinez, J","cited_arxiv_id":null,"evidence_quote":"supplies material parameters and the treatment of inhomogeneous cool-down strains used in the comparison and in the Supplementary Information."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"introduces the gate-voltage noise model and the geometry of dephasing sweet spots used to compute T2* and quality factors."},{"cited_title":"Mauro, E","cited_arxiv_id":null,"evidence_quote":"previous strain-engineering strategy for reducing g-factor anisotropy that the unstrained-bulk route is designed to improve upon."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"defines the g-matrix formalism from which Rabi frequencies are computed."},{"cited_title":"(1) [44]","cited_arxiv_id":null,"evidence_quote":"gives the perturbation expressions for g_parallel and g_perp in terms of heavy-hole/light-hole mixing that explain the anisotropy reduction."}],"review_version":1}