{"id":"9b5d691f-1f6c-4f58-bd2e-f486276803a4","arxiv_id":"2608.06194","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"First-detection quantum engines obey modified Jarzynski relations whose correction is the logarithm of the mean first-detection time of the time-reversed process.","lead":"This paper derives two new thermodynamic fluctuation relations for quantum devices that are controlled by the first time a measurement clicks. The relations link the work produced by such a device to the average first-detection time of a time-reversed measurement process, and they yield bounds on extractable work.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Central proof is recoverable, but Fig. 2 likely uses the forward mean first-detection time where eq. (22) requires the backward quantity; appendix index/PDF typos also need correcting.","rationale":"The reader's CONDITIONAL verdict is appropriate. I read the proof of eq. (18) as recoverable: the omitted U_tau in eq. (A2) cancels in the exponential average, and the index error in eq. (A11) is compensated by the sum starting at n=0 in the main text. The time-reversal identification in eq. (A10) is standard; the traced quantity is real because it can be written as Tr[Pi A M A^dagger] with M >= 0, so the antiunitary conjugation does not introduce an extra phase. Thus I do not see a fatal flaw in the central closed-system claim. The real weakness is the only concrete check: the backward initial state and Hamiltonian are not specified for the worked example, and the inset quantity appears to be the forward thermal first-detection time. If that is true, Fig. 2 validates a different bound, not eq. (22). This is addressable by rerunning the numerics with rho_tilde_Delta and by correcting the two appendix typos. The open-system generalization in Section IV is asserted without derivation, but it is not the central claim and does not change the conditional recommendation. The proof and example can be fixed without altering the main result, so I keep the reader's verdict unchanged.","tokens_in":10323,"tokens_out":33681,"duration_ms":348009,"concrete_test":"Implement the two-spin model with an explicit time-reversal convention, e.g. theta = i sigma_y K on each spin. Form H_tilde_0 = theta H_0 theta^{-1}, H_tilde_1 = theta H_1 theta^{-1}, Pi_tilde = theta Pi theta^{-1}, and rho_tilde_Delta = Pi_tilde e^{-beta(H_tilde_1-H_tilde_0)} Pi_tilde / Z_Delta. Compute <n>_{rho_tilde_Delta} using eqs. (5)-(6) on the same (tau,T,J,h) grid as Fig. 2, and plot the right side of eq. (22). Independently evaluate the left side by direct TPM simulation: sample an initial H_0 eigenstate, evolve with U_tau, apply Pi_perp until the first Pi click, quench to H_1, sample the final H_1 eigenstate, and average W_ex = epsilon_0 - epsilon_1 over detection events. Check the inequality. Then repeat using the forward thermal <n>_{rho_0}; the inequality should fail if the figure currently uses that wrong quantity.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central closed-system relation (18) is probably correct, but the manuscript does not currently prove or validate it cleanly. In the appendix, eq. (A2) writes the first-detection amplitude as <e1|Pi G^{n-1}|e0>, omitting the final U_tau that appears in eq. (5); this cancels in the exponential average, but the displayed work PDF is not the physical one. More seriously, eq. (A11) states f_n = (Z1/Z0) S_n[rho_tilde_1]; with the paper's definition S_n = 1 - sum_{k=1}^n P_k this would give sum_n f_n = (Z1/Z0)(<n>-1), contradicting eq. (18). The correct recurrence gives f_n = (Z1/Z0) S_{n-1}[rho_tilde_1], matching the sum from n=0 in eq. (18), so this is an index typo. The load-bearing check is the two-spin example: eq. (22) requires the mean first-detection step of the backward process generated by exp(-i tau H_tilde_0) and Pi_tilde, starting from rho_tilde_Delta. For spin-1/2 the time reversal flips sigma_z, so H_tilde_0 differs from H_0 and rho_tilde_Delta is not the thermal state rho_0. The paper never specifies these backward quantities, and the Fig. 2 inset appears to plot the forward mean detection time from eq. (6) with the thermal initial state; if so the dashed curve is not the bound proved.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper derives two modified quantum Jarzynski equalities for a first-detection protocol: Eq. (18) for the total energy change of a measurement-and-quench sequence and Eq. (20) for the work extracted by the quench triggered by the first detection event. The claimed corrections to the standard quantum Jarzynski equality are logarithmic in the mean first-detection step of a time-reversed protocol. The central proof in the appendix maps forward work weights to backward survival probabilities, and the paper illustrates the bounds on a two-spin model, additionally introducing a first-detection-averaged ergotropy and sketching a generalization to a system coupled to a bath.","tokens_in":10590,"tokens_out":13792,"duration_ms":130332,"significance":"If the central relations are correct, the work is significant: it provides an experimentally accessible, parameter-free quantum fluctuation relation for measurement-triggered work extraction in which the only protocol-dependent correction is the mean first-detection time of the time-reversed process. The derivation is essentially self-contained, uses no fitted parameters, and the algebraic steps behind Eq. (18) are explicit and recoverable once the index errors in the appendix are corrected. The numerical illustration, however, does not currently validate the backward quantity that enters the extracted-work bound, so the example as printed does not yet demonstrate the proved inequality.","major_comments":[{"comment":"The work PDF in Eq. (A2) is not the physical first-detection work PDF. From the definition in Eq. (5), the detected amplitude at step n from an initial eigenstate |epsilon0,k0> is <epsilon1,k1|Pi G^{n-1} U_tau |epsilon0,k0>, not <epsilon1,k1|Pi G^{n-1} |epsilon0,k0>. The missing unitary U_tau cancels in the thermal exponential average that leads to Eq. (18), so the final equality is not affected, but the displayed PDF should be corrected to match the protocol.","section":"Appendix, Eq. (A2)"},{"comment":"Equation (A11) contains an index error: the last equality should read f_n = (Z1/Z0) S_{n-1}[rho_tilde_1], not S_n, because the survival probability is defined in Section II as S_n = 1 - sum_{k=1}^{n} P_k. With the printed S_n, summing over n gives (Z1/Z0)(<n_tilde> - 1), contradicting Eq. (18); the summation in Eq. (A12) actually uses the correct n-1 form, so the proof is recoverable once this typo is fixed.","section":"Appendix, Eq. (A11)"},{"comment":"The definition of the shifted extracted work is inconsistent with the exponential in Eq. (20). The text defines w'_ex = w_ex + <H0>0, but the factor exp[-beta(H1 - H0)] exp[beta H0(0)] in Eq. (20) corresponds to the random variable W'_ex = -(W_ex + H0), i.e. w'_ex = -w_ex - <H0>0. With the printed definition, Jensen's inequality applied to Eq. (20) gives a lower bound on w_ex, not the upper bound in Eq. (22). The bound in Eq. (22) is the one that follows from the corrected definition, so the sign error is load-bearing and must be fixed.","section":"Section IV, Eq. (20)"},{"comment":"The bound in Eq. (22) involves the mean first-detection step of the time-reversed protocol generated by exp(-i tau H_tilde_0) and Pi_tilde, starting from rho_tilde_Delta. For the two-spin example, time reversal flips sigma_z, so H_tilde_0 differs from H_0 and rho_tilde_Delta is not the thermal state rho_0. The manuscript never specifies these backward quantities for the example. The inset of Fig. 2 is labeled only as <n> of Eq. (6), which is the forward mean first-detection time, and appears to be computed with the forward thermal initial state. If so, the dashed curves in Fig. 2 are not the bound proved in Eq. (22). Please specify the backward protocol explicitly and recompute Fig. 2, or clearly state that the forward quantity is used for illustration only.","section":"Section IV, Fig. 2"}],"minor_comments":[{"comment":"The phrase \"fluctuation relations and bonds\" should read \"fluctuation relations and bounds\", and the Fig. 2 caption contains the same typo in \"upper bond\".","section":"Section IV, text near Eq. (15)"},{"comment":"The sentence \"The maxima w_dot_ex follows closely the minima of <n>” mixes singular and plural; it should read \"the maxima follow\" or \"the maximum follows\".","section":"Section IV, text after Eq. (23)"},{"comment":"The word \"sightly\" should be \"slightly\".","section":"Section IV, text near Fig. 2"},{"comment":"The displayed equation for f_n in Eq. (A9) has a line break that obscures the algebra; rewriting the second term as (Z1/Z0) times the survival-probability expression would improve readability.","section":"Appendix, Eq. (A9)"},{"comment":"The statement that Eq. (18) holds without modification for a system coupled to a bath is asserted rather than derived; since the proof in the appendix assumes the projective measurement acts on the full Hilbert space of the thermal state, a sentence explaining the commutativity or trace-cyclicity argument would be helpful, even if the full analysis is deferred.","section":"Section IV, bath discussion"}],"recommendation":"major_revision","confidential_remarks":"The stress-test concern about Fig. 2 is confirmed by the caption: the inset is labeled only with Eq. (6) and no backward protocol is defined for the spin example. The sign/shift inconsistency around Eq. (20) is also present and needs to be corrected before the extracted-work bound can be considered established."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Alberto's paper is worth a serious referee seat, but it needs some cleanup before I'd trust the numbers. The new content is real: two exact fluctuation relations for first-detection quantum engines, generalizing the qubit-specific continuous-monitoring theorem of Manikandan et al. to arbitrary-dimensional stroboscopic projective measurements. The central trick—mapping forward work weights to backward survival probabilities—looks right, and the correction term involving the mean first-detection time of the time-reversed process is a nice, experimentally accessible way to close the Jarzynski equality. The second relation for extracted work and the first-detection-averaged ergotropy are also sensible additions.\n\nThe proof in the appendix has two typos that need fixing. Eq. (A2) drops the U_tau that appears in eq. (5) in the work PDF; the subsequent calculation uses the correct operator, so it's likely just a display error. More substantively, eq. (A11) defines f_n with a survival probability S_n, but the recurrence actually gives S_{n-1}. With that index error, summing f_n would contradict eq. (18). Once you shift the index, the derivation goes through. Neither typo is fatal, but they will confuse anyone trying to verify the algebra.\n\nThe bigger concern is the two-spin example. Eq. (22) requires the mean first-detection step of the backward process generated by exp(-iτ \\tilde{H}_0) and \\tilde{Π}, starting from \\tilde{ρ}_Δ. For spin-1/2, time reversal flips σ_z, so \\tilde{H}_0 differs from H_0 and the backward initial state is not the thermal state. The manuscript never specifies these backward quantities, and the inset of Fig. 2 appears to plot the forward mean detection time from eq. (6) with the thermal initial state. If that's right, the dashed bound in Fig. 2 is not the bound proved. That's a mismatch that needs to be resolved—either recompute the backward quantity or clarify what was plotted.\n\nThe open-system extension is one paragraph and reads as an assertion; it should be flagged as such or removed.\n\nOverall, the central closed-system result is probably correct and is a genuine contribution to quantum stochastic thermodynamics. The paper needs a revision that fixes the typos, specifies the backward dynamics in the example, and verifies Fig. 2. I'd send it to peer review.","headline":"New and likely correct fluctuation relations for first-detection quantum engines, but the appendix has fixable typos and the example's backward dynamics need clarification.","tokens_in":11146,"tokens_out":7360,"would_cite":true,"duration_ms":62151,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["03.65.-w","05.70.Ln"],"model":"deepseek-v4-flash","headline":"First-detection engines obey a modified Jarzynski equality","keywords":["quantum fluctuation relations","first-detection process","Jarzynski equality","feedback control","information engine","time-reversed dynamics","ergotropy","repeated projective measurements"],"falsifier":"Simulate the backward first-detection process for the two-spin model with the time-reversed Hamiltonian $\\widetilde H_0 = -J\\sigma_x^a\\sigma_x^b + h(\\sigma_z^a + \\sigma_z^b)$, starting from $\\tilde\\varrho_1$, and compare its mean first-detection step to the forward $\\langle n\\rangle$ plotted in the inset of Fig. 2; if the two differ, the bound in eq. (22) as written is not the bound the proof establishes.","tokens_in":10067,"feed_emoji":"⏱️","tokens_out":9868,"duration_ms":82264,"temperature":0.7,"pith_summary":"This paper derives two exact work fluctuation relations for quantum devices in which a repeated projective measurement triggers a feedback operation the first time a prescribed outcome occurs. For a system starting in thermal equilibrium, the detection-conditioned average of the exponentiated total work obeys a modified quantum Jarzynski equality, $\\langle e^{-\\beta W_{\\rm tot}}\\rangle_{\\rm det} = e^{-\\beta(\\Delta F - T \\log \\langle \\tilde n\\rangle_{\\tilde \\rho_1})}$, where the only protocol-dependent correction is the mean first-detection time of the time-reversed dynamics. A parallel relation holds for the work extracted by the final feedback quench alone, with a free-energy difference built from $\\Delta H = H_1 - H_0$. Jensen's inequality converts these equalities into second-law-like bounds on both the total work cost and the extracted work. The paper also generalizes the total-work relation to a device coupled to an external bath and introduces a first-detection-averaged ergotropy as an independent upper bound on extracted power.","feed_headline":"First-detection engines obey a modified Jarzynski equality","feed_subtitle":"A single experimentally accessible number, the mean reversed detection time, sets the new bound on extracted work.","key_machinery":"The load-bearing object is the survival operator $G_\\tau = U_\\tau \\Pi^\\perp$, which propagates the state only through null measurement outcomes, together with the detection-conditioned expectation value $\\langle A\\rangle_{\\rm det} = \\sum_n \\mathrm{tr}[A \\Pi G_\\tau^{n-1} U_\\tau \\rho_{\\rm in} U_\\tau^\\dagger (G_\\tau^\\dagger)^{n-1}\\Pi]$ that sums over first-detection times. The proof's central identity, eq. (A10), rewrites the forward work weight of a trajectory detected at step $n$ as the first-detection probability at step $n-1$ of the time-reversed process generated by $\\exp(-i\\tau \\widetilde H_0)$ and projector $\\widetilde\\Pi$, starting from $\\widetilde\\varrho_1$. Summing these weights converts the detection-conditioned Jarzynski average into $(Z_1/Z_0)\\langle \\tilde n\\rangle_{\\tilde \\varrho_1}$, which is exactly the source of the logarithmic correction in the final equality.","core_discovery":"The central claim is that the first-detection ensemble, meaning the set of measurement trajectories that end at the first click of a stroboscopic projective detector, obeys its own quantum work fluctuation relations. If the initial state is thermal with Hamiltonian $H_0$ at inverse temperature $\\beta$, and a positive detection at step $n$ is followed by an instantaneous quench $H_0 \\to H_1$, then $\\langle e^{-\\beta (H_1(t_{\\rm det}) - H_0(0))}\\rangle_{\\rm det} = (Z_1/Z_0)\\sum_n S_n[\\tilde \\varrho_1] = e^{-\\beta(\\Delta F - T\\log \\langle \\tilde n\\rangle_{\\tilde \\varrho_1})}$. The forward work statistics are thereby controlled by the survival probabilities of a backward first-detection process that starts from the time-reversed, detection-conditioned thermal state $\\tilde \\varrho_1$. An analogous identity, eq. (20), holds for the extracted work defined through $H_1 - H_0$. When $\\Pi = \\mathbb{I}$, the mean reversed detection step is one and the standard quantum Jarzynski equality is recovered. The logarithmic correction $-T\\log\\langle \\tilde n\\rangle$ acts as an entropic reduction of the effective free-energy cost, which the paper reads as the thermodynamic value of the information gathered by the repeated measurements.","pith_inferences":["The same first-detection machinery should also produce a Crooks-type detailed fluctuation relation for the ratio of forward and backward work probabilities, a distribution-level statement the paper does not write down.","Because the bound is set by the backward mean detection time, power optimization could be phrased as minimizing the time-reversed detection time over the sampling interval $\\tau$, a strategy the paper does not explicitly propose.","The open-system extension is only given for the total-work relation; the extracted-work relation is deferred, with the finiteness of the bath Hilbert space flagged as a technical requirement, so the thermodynamic limit of a large bath remains an open question.","In the worked two-spin example the ergotropy bound is looser than the fluctuation-relation bound, suggesting that the information-theoretic correction is the tighter certificate for first-detection engines, though only one model is tested."],"forward_implications":["The total work cost of a first-detection engine satisfies $w_{\\rm tot} \\ge \\Delta F - T\\log\\langle \\tilde n\\rangle_{\\tilde \\varrho_1}$, so the detection protocol effectively lowers the free-energy barrier that the engine must overcome.","The extractable work satisfies $w_{\\rm ex} \\le -(\\Delta F_\\Delta - T\\log\\langle \\tilde n\\rangle_{\\tilde \\varrho_\\Delta}) - \\langle H_0\\rangle_0$, giving a sampling-interval-dependent upper bound on the output power $\\dot w_{\\rm ex}$.","All correction terms are experimentally accessible: they require only the mean first-detection time of the time-reversed process, not the full distribution of forward and backward trajectories.","Setting $\\Pi = \\mathbb{I}$ reproduces the standard quantum Jarzynski equality, so the new relations extend, rather than replace, the usual two-projective-measurement fluctuation theorem.","For a device coupled to a bath, the total-work relation (18) remains valid when the quench and the monitored observable act only on the device subsystem."],"supporting_citations":[{"why":"Supplies the quantum first-detection formalism used throughout: survival probabilities, dark states, and the mean first-detection time.","marker":"[20]"},{"why":"Gives the compact two-time-correlation form of the quantum Jarzynski equality that eq. (18) generalizes.","marker":"[27]"},{"why":"Provides the two-projective-measurement work distribution that the detection-conditioned work PDF in the appendix extends.","marker":"[25]"},{"why":"Provides the antilinear time-reversal operator relations used in the key identity (A10).","marker":"[29]"},{"why":"Supplies the classical first-passage information engine whose quantum counterpart is the setup analyzed here.","marker":"[11]"}],"fun_headline_variants":["First-detection measurements alter quantum work relations","First-click statistics modify Jarzynski equality","Mean reversed detection time sets new work bound","Quantum work relations for first-detection protocols","First-detection engines follow a corrected Jarzynski"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof's central step (eq. A10) identifies the forward work weight with the survival probability of a backward first-detection process starting from the time-reversed state $\\tilde\\varrho_1$; if the backward dynamics used in practice differ from this time-reversed process, the logarithmic correction term changes and the bounds do not hold.","fun_headline_variants_meta":{"raw":{"variants":["First-detection measurements alter quantum work relations","First-click statistics modify Jarzynski equality","Mean reversed detection time sets new work bound","Quantum work relations for first-detection protocols","First-detection engines follow a corrected Jarzynski"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000279,"raw_usage":{"total_tokens":1643,"prompt_tokens":914,"completion_tokens":729,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":530,"completion_tokens_details":{"reasoning_tokens":661}},"tokens_in":530,"tokens_out":729,"duration_ms":7547,"temperature":1.0,"reasoning_tokens":661,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T12:49:07.733431+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Simulate the backward first-detection process for the two-spin model with the time-reversed Hamiltonian $\\widetilde H_0 = -J\\sigma_x^a\\sigma_x^b + h(\\sigma_z^a + \\sigma_z^b)$, starting from $\\tilde\\varrho_1$, and compare its mean first-detection step to the forward $\\langle n\\rangle$ plotted in the inset of Fig. 2; if the two differ, the bound in eq. (22) as written is not the bound the proof establishes.","supporting_citations":[{"cited_title":"Friedman, D","cited_arxiv_id":null,"evidence_quote":"Supplies the quantum first-detection formalism used throughout: survival probabilities, dark states, and the mean first-detection time."},{"cited_title":"De Chiara and A","cited_arxiv_id":null,"evidence_quote":"Gives the compact two-time-correlation form of the quantum Jarzynski equality that eq. (18) generalizes."},{"cited_title":"Talkner, E","cited_arxiv_id":null,"evidence_quote":"Provides the two-projective-measurement work distribution that the detection-conditioned work PDF in the appendix extends."}],"review_version":1}