{"id":"e0b41f74-5e29-4392-b15c-01b33eef2880","arxiv_id":"2505.15878","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Readout back-action in spin qubits from g-tensor modulation is minimized when the magnetic field is oriented so the static Zeeman field is parallel to the sensor-induced Zeeman fluctuation (gB parallel to g'B), a condition that is always achievable.","lead":"A theoretical model of charge-sensing qubit readout shows that residual tunneling and electric-field-induced g-tensor changes introduce back-action errors, and identifies a magnetic-field orientation, called the readout sweet spot, that suppresses the g-tensor-induced errors. The result gives spin-qubit device designers a concrete rule for making readout closer to an ideal projective measurement, which matters for error-corrected quantum computing.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Reference-dot g-tensor modulation is omitted from Eq. (8); universal sweet-spot existence proof for the measured dot does not guarantee a common field direction suppressing leakage from the reference spin.","rationale":"The single-dot algebra is sound: any real 3x3 matrix (g)^-1 g' has at least one real eigenvector, and along that direction the right-dot static Zeeman field and the QPC-induced fluctuation commute. The numerical QMQ results provide legitimate independent support for that single-dot statement. My concern is the step from that statement to the paper's 'universal' device-level claim. The physical readout uses two dots, and the QPC electric field can modulate the left reference spin as well; the text itself says it modulates the g-tensors of the spins in the DQD. Eq. (8) keeps only the right-dot term -s_R · Delta ⊗ tau_x. Once the analogous left-dot term is included, leakage from |↓↓> to |↑↓> is controlled by g_L B parallel g'_L B at the same magnetic field. There is no reason two generic real matrices share a real eigenvector, so the proof of existence for one dot does not prove existence for the full device. This is a model-completeness issue rather than an internal inconsistency: within the truncated single-dot model the leakage-elimination claim holds, but the universal design rule needs qualification unless the left-dot coupling is negligible or can be synchronized by gate-voltage tuning. The reader's identified perfect-sensing idealization is a legitimate limitation but does not threaten the leakage-elimination argument itself; the omitted reference-dot term does, which is why I propose CONDITIONAL rather than an unchanged ACCEPT.","tokens_in":938,"tokens_out":4365,"duration_ms":276360,"concrete_test":"Extend the QMQ model by adding a left-dot spin-charge coupling term -s_L · Delta_L ⊗ tau_x with Delta_L = mu_B g'_L B/2 to Eq. (8). For a representative strong-spin-orbit DQD model, e.g. a k·p simulation using parameters from Ref. [70], compute g_L, g'_L, g_R, g'_R and check whether (g_L)^-1 g'_L and (g_R)^-1 g'_R share a real eigenvector. If no common real eigenvector exists, no single magnetic field direction can make both local Zeeman fluctuations parallel to their static fields, so the universal leakage-free sweet spot does not survive inclusion of the reference-dot backaction.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim that leakage is eliminated when g_R B is parallel to g'_R B, and that such a direction always exists, is proven for a model in which the QPC-induced spin-charge coupling acts only on the right dot, namely the s_R · Delta term in Eq. (8). But the readout uses a two-dot device, and the introduction states that the QPC field modulates the g-tensors of the spins in the DQD. If the left reference spin has its own response g'_L, then leakage from |↓↓> to |↑↓> is generated unless the same field also satisfies g_L B parallel g'_L B. Two real 3x3 matrices do not in general share a real eigenvector, so the 'always exists' construction for the measured dot does not automatically give a leakage-free readout for the full DQD. This is not a question of sensing efficiency or parameter accuracy; it challenges the universality of the design rule as stated, because the model Hamiltonian omits the left-dot backaction term.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a 'qubit measures qubit' (QMQ) model of QPC-based charge-sensing readout for charge qubits and single-spin qubits in double quantum dots, and uses it to quantify readout errors beyond simple infidelity. It identifies two back-action mechanisms: residual tunnel coupling during readout, which causes measurement-induced relaxation and mixed post-measurement states, and g-tensor modulation by the fluctuating QPC field, which causes leakage from the computational subspace in spin-orbit-coupled devices. For the spin-qubit case, the paper derives analytical rates for measurement, relaxation, and leakage, and shows that leakage is eliminated when the static Zeeman field in the measured dot is parallel to the QPC-induced Zeeman-field fluctuation. It argues that such a direction always exists as a real eigenvector of g^{-1}g', identifies a 'readout sweet spot' for the magnetic-field orientation, and proposes an experimental protocol to measure leakage.","tokens_in":30788,"tokens_out":12084,"duration_ms":109956,"significance":"If the central claim holds, the readout sweet spot is a simple, parameter-free design rule for spin qubits with strong spin-orbit interaction, and the QMQ framework provides a useful way to evaluate post-measurement purity and leakage alongside fidelity. The paper's strengths include the analytically derived rates that are checked against numerical QMQ simulations (Eqs. 3, 4, and 10), the rigorous eigenvector argument for the existence of the sweet-spot field direction, and a concrete leakage-detection protocol. The main risk to the universality claim is the model's restriction of g-tensor modulation to the measured dot only, since the readout protocol is inherently a two-dot procedure with a reference spin.","major_comments":[{"comment":"The interaction Hamiltonian in Eq. (8) contains only the right-dot g-tensor modulation term s_R·Δ. Because the readout uses a two-electron DQD and the introduction states that the QPC field modulates the g-tensors of the spins in the DQD, the model should either include a left-dot term s_L·Δ_L or explicitly justify its absence. Without such a term, leakage from |↓↓⟩ to |↑↓⟩ due to left-spin flips is not addressed, and the proof that g_R^{-1}g'_R has a real eigenvector does not imply the existence of a field direction that simultaneously satisfies g_L B ∥ g'_L B for two unrelated 3x3 matrices. The universality claim therefore currently exceeds the model; please add a quantitative justification (e.g., field localization, parameter estimates) or explicitly limit the claim to the measured dot.","section":"Spin qubit, Eq. (8)"},{"comment":"The leakage rate is written as 2 Δ_x² δγ²/(Z_R² Δτ) sin²(Z_R Δτ/ℏ), which contains an extra factor δγ² relative to the main-text Eq. (10) and has dimensions of energy²/time rather than 1/time. The two equations are mutually inconsistent, and the supplementary version would not match the numerical data in Fig. 3f if used as written. Please correct this formula and state explicitly which expression is used in the inset comparison.","section":"SM S4, Eq. (S45)"}],"minor_comments":[{"comment":"The sentence 'The red arrow in Figs. 3c,d show the readout sweet spot' appears to refer to the panels that actually display the sweet-spot orientation; please check the figure panel numbering and correct the caption references.","section":"Fig. 3 and main text"},{"comment":"The components Δ_x and Δ_z are first used in Fig. 3 without being defined in the main text; please define them explicitly as components of Δ = μ_B g'_R B/2 before their first use.","section":"Spin qubit section"},{"comment":"The statement that current sensing is assumed to be perfect is an important limitation for the quantitative fidelity comparison between measurement and leakage rates; a sentence in the main text discussing the expected impact of inefficient sensing would help readers.","section":"SM S2"},{"comment":"The abstract lists 'charge noise of the sensor' as a modeled error mechanism, but the QMQ model does not include classical charge noise; please either add a brief explanation of how charge noise enters the model or reword the abstract to match the model content.","section":"Abstract"}],"recommendation":"major_revision","confidential_remarks":"The paper is well within the journal's scope and the core idea is attractive. My recommendation of major_revision is driven by two fixable issues: the missing left-dot g-tensor backaction, which undermines the 'universal' phrasing rather than the mathematical result for a single dot, and the inconsistency between Eq. (10) and Eq. (S45). If the authors can justify the left-dot approximation quantitatively or extend the sweet-spot condition to the full DQD, and correct the supplementary formula, I would support acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read this if you work on spin qubit readout. The QMQ model is a genuinely useful simplification, and the analytical rates in Eqs. (3), (4) and (10) are checked against exact numerics over the relevant parameter range. The sweet-spot condition for the measured dot — gB parallel to g'B — follows directly from commutation of H_spin and H_int, and the argument that such a direction always exists via a real eigenvector of g^{-1}g' is correct. That is a real design rule, not a fit.\n\nWhat the paper does not do is close the loop for the two-dot device it actually describes. Equation (8) includes only the right-dot g-tensor modulation, s_R · Δ. But the readout uses a reference spin in the left dot, and the introduction says the QPC modulates both spins' g-tensors. If the left dot has its own response g'_L, then a single field direction B must satisfy g_L B ∥ g'_L B as well as g_R B ∥ g'_R B. Two arbitrary real 3x3 matrices do not generally share a real eigenvector, so the existence proof for the measured dot does not guarantee a leakage-free configuration for the full DQD. The paper never models the left-dot term, so the headline claim that leakage is 'eliminated' is too strong. This is a modeling gap, not a numerical error. Some devices may indeed couple the QPC to one dot much more strongly than the other; the paper should say that and restrict the universality claim accordingly.\n\nThe other limitations are stated honestly: perfect current sensing, and a factor 2–5.5 difference from the stochastic master equation rates. Those are acceptable for a minimal model, and the authors flag them. The citation pattern is fine.\n\nOverall: the single-dot result is solid and worth publishing; the universal two-dot design rule needs a caveat or a proper treatment of the second spin. The paper is useful for theorists and experimentalists who want a quick estimate of readout back-action rates. Send it to review, but ask the authors to either generalize Eq. (8) to include the left-dot modulation or explicitly scope the claim to devices where that term is suppressed.","headline":"Clean single-dot readout sweet spot, but the universality claim stumbles on the reference spin's g-tensor modulation.","tokens_in":31391,"tokens_out":4208,"would_cite":true,"duration_ms":38054,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Spin-qubit readout leakage vanishes at a universal field direction.","keywords":["spin qubits","charge sensing readout","quantum point contact","spin-orbit interaction","g-tensor modulation","quantum measurement back-action","leakage","readout sweet spot"],"falsifier":"Scan the external magnetic-field direction for a single spin qubit with known $g$ and separately characterized $g'$, measure readout infidelity, post-measurement purity, and leakage rate at fixed integration time, and check that all three are simultaneously optimized at a direction satisfying $gB \\parallel g'B$; finding a well-isolated better direction outside the predicted eigenvector set would falsify the claim.","tokens_in":30381,"feed_emoji":"🧲","tokens_out":7121,"duration_ms":59478,"temperature":0.7,"pith_summary":"This paper models charge-sensing readout of spin qubits in double quantum dots and separates the readout problem into three error channels: residual tunnel coupling between the dots, charge noise of the sensor, and g-tensor modulation caused by the fluctuating electric field of the charge-sensing point contact. The central result is a configuration called a readout sweet spot: for devices with strong spin-orbit interaction and electrically tunable g-tensors, choosing the magnetic-field direction so that the static Zeeman field on the measured dot is parallel to the Zeeman-field fluctuation produced by the sensor eliminates readout-induced leakage and keeps the post-measurement state pure. The paper proves such a direction always exists, because it is a real eigenvector of the matrix $g^{-1}g'$, and argues the recipe is universal, insensitive to the microscopic details of spin-orbit interaction. If true, this turns a nuisance (sensor back-action) into a tunable resource for high-fidelity mid-circuit measurement.","feed_headline":"Spin-qubit readout leakage vanishes at a universal field direction","feed_subtitle":"Leakage stops when the static and sensor-induced Zeeman fields align — and such a direction always exists.","key_machinery":"The load-bearing object is the $3\\times3$ matrix $g^{-1}g'$, built from the dot's $g$-tensor $g$ (static Zeeman response) and the sensor-induced $g$-tensor modulation $g'$. Its real right eigenvectors are exactly the magnetic-field directions satisfying $gB \\parallel g'B$, the readout sweet spot; because a real matrix has either one or three real eigenvalue–eigenvector pairs, at least one physical field direction always exists. The argument runs through a 'qubit measures qubit' (QMQ) model in which the quantum point contact is represented by a single two-level meter, so the entire readout is a sequence of weak indirect measurements; the measurement operations for the final inferred outcome are then computed exactly, and the sweet-spot condition follows from demanding that the interaction Hamiltonian commute with the spin Hamiltonian in the computational subspace.","core_discovery":"The paper's claim is that a spin-qubit readout can be made effectively projective by satisfying two conditions at once: switch off the inter-dot tunnel coupling during charge sensing, and orient the external magnetic field along a direction where the static Zeeman field in the measured dot is parallel to the local Zeeman-field fluctuation caused by the charge sensor. In that geometry the sensor still sees the qubit state through the charge configuration, but the g-tensor-modulation term commutes with the spin Hamiltonian, so no incoherent transitions are activated between the computational state and the leaked state. The commutation is captured by the vector equation $gB \\parallel g'B$, and the paper shows this equation always has a physical solution: the matrix $g^{-1}g'$ has at least one real eigenvalue with a real eigenvector, which supplies the required field direction. When the direction is chosen instead with a perpendicular modulation component $\\Delta_x$, leakage appears with rate $\\Gamma_{\\rm leak} = 2\\Delta_x^2\\delta\\gamma^2 \\sin^2(Z_R\\Delta\\tau/\\hbar)/(Z_R^2\\Delta\\tau)$, saturating at leakage probability $1/2$ for long integration times.","pith_inferences":["If current sensing is inefficient or noisy rather than perfect, the measurement rate would fall while the dephasing and relaxation rates stay fixed, so the fidelity advantage at the sweet spot would shrink; the paper flags this assumption in its supplementary material but does not quantify the degradation.","The same $gB \\parallel g'B$ criterion should apply to any charge-sensing technique whose fluctuating electric field modulates the $g$-tensor, not only a quantum point contact, because the geometry of the commutation condition is independent of the sensor's microscopic implementation.","A direct experimental test would be to fix a strong-spin-orbit qubit, measure $g$ and $g'$ independently, then sweep the field direction and check that the minimum of readout-induced leakage and the maximum of post-measurement purity occur at the predicted eigenvector direction.","The sweet-spot field direction could be found without knowing $g'$ in advance by measuring leakage as a function of field orientation and looking for the zero; that measured direction would then give an indirect characterization of $g^{-1}g'$."],"forward_implications":["With the tunnel coupling switched off and the field at the sweet spot, readout infidelity tends to zero as the integration time grows, and the post-measurement state stays pure.","When the field is misaligned, leakage appears at a rate proportional to the square of the perpendicular modulation component $\\Delta_x$, so orienting the field to satisfy $gB \\parallel g'B$ suppresses that error channel entirely.","The parallel component $\\Delta_z$ increases the measurement rate, so among the (at most three) sweet-spot directions the one with the largest eigenvalue of $g^{-1}g'$ is optimal.","In multi-qubit devices, the $g$-tensor parameters can be tuned by gate voltages, so the sweet-spot directions of different qubits can in principle be synchronized."],"supporting_citations":[{"why":"Supplies the quantum-measurement formalism, including measurement operations, used to define readout infidelity, post-measurement mixedness, and leakage.","marker":"60"},{"why":"Provides the earlier stochastic-master-equation description of QPC charge sensing that the qubit-measures-qubit model simplifies and compares against in the supplementary material.","marker":"62"},{"why":"Supplies the experimental g-tensor and g'-tensor parameters for a strong-spin-orbit device used to plot the readout sweet spot.","marker":"70"},{"why":"Establishes that the real $3\\times3$ matrix $g^{-1}g'$ has at least one real eigenvalue-eigenvector pair, guaranteeing a physical field direction satisfies the sweet-spot condition.","marker":"75"}],"fun_headline_variants":["Sweet spot for spin-qubit readout: parallel Zeeman fields","Spin-qubit leakage vanishes at aligned field direction","Projective readout via readout sweet spot: align Zeeman fields","Universal field direction makes spin-qubit readout projective"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that the quantum point contact current is sensed perfectly, with no measurement inefficiency or added noise; if real amplifiers and detectors are slower or noisier, the measurement rate drops relative to the relaxation and leakage rates, and the fidelity gain at the sweet spot may be smaller than the model predicts.","fun_headline_variants_meta":{"raw":{"variants":["Sweet spot for spin-qubit readout: parallel Zeeman fields","Spin-qubit leakage vanishes at aligned field direction","Projective readout via readout sweet spot: align Zeeman fields","Universal field direction makes spin-qubit readout projective"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000274,"raw_usage":{"total_tokens":1638,"prompt_tokens":946,"completion_tokens":692,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":562,"completion_tokens_details":{"reasoning_tokens":623}},"tokens_in":562,"tokens_out":692,"duration_ms":6152,"temperature":1.0,"reasoning_tokens":623,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T15:11:05.749621+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Scan the external magnetic-field direction for a single spin qubit with known $g$ and separately characterized $g'$, measure readout infidelity, post-measurement purity, and leakage rate at fixed integration time, and check that all three are simultaneously optimized at a direction satisfying $gB \\parallel g'B$; finding a well-isolated better direction outside the predicted eigenvector set would falsify the claim.","supporting_citations":[{"cited_title":"\\ Goan , author G","cited_arxiv_id":null,"evidence_quote":"Provides the earlier stochastic-master-equation description of QPC charge sensing that the qubit-measures-qubit model simplifies and compares against in the supplementary material."},{"cited_title":"Sen , author G","cited_arxiv_id":null,"evidence_quote":"Establishes that the real $3\\times3$ matrix $g^{-1}g'$ has at least one real eigenvalue-eigenvector pair, guaranteeing a physical field direction satisfies the sweet-spot condition."}],"review_version":1}