{"id":"d0e4c8fb-96a8-4cd0-819f-0c9aa07d0b97","arxiv_id":"1908.04757","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Neighboring Bose-Fermi doublons decay through a triplon-singlon channel, and small connected clusters accurately describe the short-time dynamics of dilute doublon gases.","lead":"This paper analyzes how boson-fermion pairs (doublons) in an optical lattice break apart and move after a sudden quench. It derives compact decay laws and a cluster-expansion method for the short-time dynamics, benchmarked against exact simulations.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The connected-cluster truncation rests on an unverified quantum-Zeno assumption; the only exact test has JD/Gamma about 0.3 and stops before the predicted separated-pair decay time.","rationale":"The reader's weakest assumption identifies exactly the quantum-Zeno truncation of disconnected clusters, and my reading agrees that this is the load-bearing support for Eq. (5). I considered whether the apparent inconsistency in Eq. (A29), where W is written with sqrt(pi/L) although the final Gamma in Eq. (A32) corresponds to W proportional to sqrt(2/L), is a more serious threat. Re-deriving the Bixon-Jortner calculation with the normalized sine states used in Eq. (A24) gives Gamma = 8 JB^2/JF sqrt(...), so the displayed W is best read as a typographical slip rather than a flaw in the central rate. The substantive gap is the unquantified condition JD << Gamma. The paper is careful to describe the Zeno argument as making the cluster expansion plausible, and the t-DMRG benchmark is exactly what is needed, but it covers one parameter set and a time window too short to see the predicted separated-pair decay. Since the analytical few-doublon results are parameter-free and the 1D benchmark supports the short-time connected-cluster picture, the correct disposition remains conditional: the concern warrants an explicit parameter scan and a stated validity region, but it does not overturn the main message.","tokens_in":17712,"tokens_out":14239,"duration_ms":163400,"concrete_test":"Run t-DMRG for two doublons initially on sites 0 and 2 (one empty site between) in 1D, scanning UBF/JF = 5, 10, 20, JB/JF = 0.05, 0.1, 0.2, and UBB/JF = 0.5, 1.0, 1.8, with L at least 200 to suppress boundary effects. Extract the long-time decay rate Gamma_sep by fitting P2(t) from t = 20/JF to t = 500/JF and compare it with the connected-only prediction (constant at the single-doublon saturation value) and with JD^2/Gamma. Then recompute Eq. (5) for f = 5% and f = 20% with Gamma_sep included for initially separated pairs at distance 2. If this shifts P2 by more than about 1% at t = 20/JF or 5% at t = 100/JF, the connected-cluster truncation underpredicts the decay and the central short-time claim must be restricted to smaller fillings or shorter times.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central predictive claim is Eq. (5): the doublon fraction of a dilute random initial state is a weighted average over evolutions of small connected clusters only. This truncation requires that initially separated doublons remain effectively stable on the timescale of interest. The stability argument in Sec. III.2 is that two doublons with an empty site between them can decay only after they first tunnel together at the effective rate JD = 2 JF JB / UBF, and the resulting neighboring pair then decays at the rate Gamma of Eq. (A32); for JD << Gamma the separated-pair decay rate is about JD^2/Gamma. The problem is that this condition is asserted rather than systematically verified. The only exact demonstration, Fig. 3(b), uses JB = 0.1 JF, UBB = JF, UBF = 10 JF, for which JD = 0.02 JF and Gamma = 8(0.1)^2 sqrt(3/4) JF about 0.069 JF, so JD/Gamma is about 0.29 rather than much smaller than 1. The displayed time window ends near t = 20 JF, while the predicted suppression scale 1/(JD^2/Gamma) is about 170 JF. The observed non-decay is therefore consistent with the Zeno argument but does not verify it. In addition, Gamma vanishes at the band edge UBB = 2 JF, so JD << Gamma cannot hold uniformly throughout the decay regime. At filling f = 20%, pairs separated by one empty site are abundant; if their actual decay rate is anywhere near JD^2/Gamma or JD, the cluster sum underpredicts P2(t) at times only modestly beyond the benchmarked window. Because Fig. 3(c,d) tests one parameter set and no code or data are released, this truncation is the least secure support for the central claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript studies the non-equilibrium dynamics of a dilute Bose-Fermi mixture in an optical lattice initialized in a product state of boson-fermion doublons. It derives analytical expressions for the decay of isolated doublons and of small clusters of two and three neighboring doublons, identifies triplon-singlon formation as the dominant decay channel of neighboring doublons when UBB < 2dJF, and proposes a cluster expansion in which the total doublon fraction is a weighted sum over connected few-doublon clusters. The cluster expansion is benchmarked against t-DMRG simulations in 1D, and experimental parameters for 40K-87Rb mixtures are discussed.","tokens_in":17993,"tokens_out":9305,"duration_ms":95582,"significance":"The 1D few-body results are genuinely useful: they are derived from the microscopic Hamiltonian, involve no fitted parameters in the central formulas, and are tested against independent t-DMRG simulations. The identification of the triplon-singlon decay channel and the hierarchy of time scales is physically interesting and potentially relevant for ongoing experiments with ultracold molecules. The main limitation is that the connected-cluster truncation, which is the basis of the predictive Eq. (5), relies on a quantum-Zeno stability assumption for separated doublons that is neither derived nor systematically verified; the only exact test uses a parameter set with JD/Gamma about 0.3 and a time window well below the predicted Zeno timescale.","major_comments":[{"comment":"The connected-cluster truncation in Eq. (5) rests on the claim that initially separated doublons are stable due to a quantum Zeno effect, but the paper's own parameters do not satisfy the stated condition JD << Gamma. For the only exact test shown in Fig. 3(b)-(d), with JB=0.1JF, UBB=JF, and UBF=10JF, one has JD = 2JFJB/UBF = 0.02JF and Gamma = 8JB^2/JF sqrt(1-(UBB/(2JF))^2) ≈ 0.069JF, giving JD/Gamma ≈ 0.29 rather than a strong inequality. The numerical time window ends near t ≈ 20 JF^{-1}, while the predicted Zeno-suppressed separated-pair decay timescale Gamma_eff^{-1} = Gamma/JD^2 ≈ 170 JF^{-1}; the observed non-decay is therefore only consistent with, not a verification of, the Zeno assumption. Moreover, Gamma vanishes at the band edge UBB = 2JF, so JD << Gamma cannot hold uniformly throughout the decay regime. Since pairs separated by one empty site are abundant at filling f=20%, the predictive power of Eq. (5) for times beyond the benchmarked window is not established. A systematic test across parameters or a quantitative bound on the disconnected-cluster contribution is needed.","section":"Section III.2/III.3"}],"minor_comments":[{"comment":"The caption lists UBF/JF values of -10 and -30, which is inconsistent with the convention in Eq. (1) where UBF is positive for attractive interspecies interactions; please correct the sign or clarify the convention.","section":"Fig. 2(c) caption"},{"comment":"With JD defined as 2JFJB/UBF, the values JB=0.1JF and UBF=10JF give JD^{-1} = 50 JF^{-1}, not 100 JF^{-1} as stated in the caption; this inconsistency should be corrected because it affects the quantitative Zeno timescale estimate.","section":"Fig. 3(b) caption"},{"comment":"It would help to state explicitly whether the cluster evolutions P2[c](t) used in Fig. 3(c,d) are the analytical expressions from Appendix A or numerically obtained cluster evolutions, since the analytical three-doublon expression (A23) is derived in the perturbative regime UBB > 2JF and is not directly applicable to the parameter set UBB=JF used in the figure.","section":"Section III.3 and Fig. 3(c,d)"},{"comment":"The claim of analytical expressions for small doublon clusters is qualified in the appendix: in dimensions d>1 the two-doublon decay rate is extracted from numerical fits (Fig. 5) rather than derived. A brief qualification in the abstract or introduction would avoid overstating the analytic content.","section":"Abstract and Introduction"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within the journal's scope. The main concern is the unverified Zeno assumption behind the connected-cluster truncation; extending the t-DMRG simulations to longer times or providing a quantitative bound on disconnected-cluster contributions would substantially strengthen the paper. The sign typo in Fig. 2(c) should also be checked against the numerical data to ensure the reported parameters are correct."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The thing you should know: this is a careful, well-benchmarked theory paper on doublon decay in Bose-Fermi mixtures, and the genuinely new piece is the triplon-mediated decay of neighboring doublons plus a cluster expansion for dilute doublon gases. The central short-time claim holds up. I would not desk-reject this.\n\nWhat it does well: the analytical derivations for one-, two-, and three-doublon clusters are clean and parameter-free in 1D, and the benchmark against t-DMRG is honest and quantitative. The two-doublon exponential decay rate, Eq. (A32), follows from a Bixon-Jortner quasi-continuum treatment and matches the numerics where shown. The cluster expansion reproduces the exact t-DMRG dynamics for low filling and short times, which is exactly the regime the paper claims. The experimental parameter table for 40K-87Rb is useful and makes the connection to cold-molecule experiments concrete. Credit is deserved for the triplon mechanism itself: the conversion of two neighboring doublons into a singlon plus a triplon is a real process that is absent for fermionic doublons, and the authors identify it clearly.\n\nThe soft spots, in order of severity. First, the cluster expansion truncates to connected clusters because separated doublons are assumed stable via a quantum Zeno argument (JD << Gamma makes the effective separated-pair decay rate ~JD^2/Gamma). The only exact demonstration, Fig. 3(b), uses parameters where JD/Gamma is about 0.29, not much less than 1, and the time window stops at tJF = 20, much shorter than the predicted 1/(JD^2/Gamma) ~ 170 JF. So the non-decay is consistent with the Zeno argument but does not verify it. At filling 20%, pairs separated by one empty site are abundant, and if their actual decay rate is appreciable, the cluster sum could underpredict P2 at times only modestly beyond the benchmarked window. This is a real gap in the evidence, but it does not break the paper's stated scope of short-time, low-filling dynamics. Second, the higher-dimensional decay rates in Fig. 5 are fitted, not derived; the paper says so, but the fit is presented without error bars or a discussion of how the density-of-states f(u) changes with d. That is acceptable, but should be labeled more prominently. Third, there is a sign inconsistency in the Fig. 2 caption: UBF/JF is listed as -10 and -30, while the text defines UBF positive with the interaction term written as -UBF. Minor, but confusing. Fourth, no code or data are released, which makes it unnecessarily hard to extend the benchmark.\n\nWho this is for: people working on doublon dynamics, quantum simulation with polar molecules, or lattice Bose-Fermi mixtures. It is an incremental but solid extension of the single-doublon literature, not a field-changer. I would send it to a serious referee, with the request that the authors either extend the Zeno check (longer times, smaller JD/Gamma, more parameter sets) or soften the cluster-expansion claim, and ideally release the code and data. As it stands, it deserves peer review and, with those requests, likely publication.","headline":"Solid short-time doublon-decay paper with a genuinely new triplon decay channel and a cluster expansion that is benchmarked against t-DMRG; the connected-cluster truncation is plausible but not fully verified, so accept with requests for a broader Zeno check and code/data release.","tokens_in":18570,"tokens_out":2156,"would_cite":true,"duration_ms":25000,"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":"Adjacent Bose-fermion pairs decay exponentially","keywords":["doublon decay","Bose-Fermi mixture","optical lattice","cluster expansion","triplon formation","quantum Zeno effect","relaxation dynamics","Hubbard model"],"falsifier":"Take two doublons separated by a single empty site with $J_B=0.1J_F$, $U_{BB}=J_F$, $U_{BF}=10J_F$ in 1D and track $P_2(t)$ up to $t\\approx 300 J_F^{-1}$; the cluster expansion predicts decay only at the Zeno-suppressed rate $\\Gamma_{\\rm eff}\\sim J_D^2/\\Gamma\\approx 10^{-2}J_F$. An observed decay at a rate comparable to $J_D$ (or any rate linear in $J_D$) would falsify the connected-cluster truncation.","tokens_in":17476,"feed_emoji":"⚛️","tokens_out":9628,"duration_ms":89915,"temperature":0.7,"pith_summary":"This paper asks how a dilute gas of boson-fermion pairs (doublons) placed on random lattice sites relaxes under the competition between tunneling and on-site interactions. It claims that at short times the relaxation is governed by the decay of small connected clusters of doublons, with the dominant microscopic process being the conversion of two neighboring doublons into a triplon (a site containing one fermion and two bosons) plus a mobile singlon (a lone fermion). The paper derives closed-form decay laws for single, double, and triple doublon clusters, including the exponential pair-decay rate $\\Gamma = (8J_B^2/J_F)\\sqrt{1-(U_{BB}/(2J_F))^2}$ in one dimension, and assembles them into a cluster expansion whose predictions match numerically exact matrix-product-state simulations. If the picture holds, it turns short-time relaxation in strongly correlated Bose-Fermi mixtures into a few-body problem that experiments with ultracold molecules can directly benchmark.","feed_headline":"Adjacent Bose-fermion pairs decay exponentially","feed_subtitle":"A cluster expansion reproduces the short-time doublon decay in optical lattices and matches exact 1D simulations.","key_machinery":"The central object is the connected doublon cluster, whose two-doublon version reduces to a single 'bound' initial state coupled to a continuum of singlon-triplon scattering states of bandwidth $4dJ_F$. The mechanism producing exponential decay is the Bixon-Jortner quasi-continuum: for $J_B\\ll J_F$, the bound state couples with nearly constant matrix element $W$ to an effectively equally spaced continuum with spacing $\\Delta$, and the survival amplitude follows $e^{-\\Gamma t/2}$ with $\\Gamma=2\\pi W^2/\\Delta$, which evaluates to the closed form $\\Gamma = (8J_B^2/J_F)\\sqrt{1-(U_{BB}/(2J_F))^2}$ in 1D. For isolated doublons the same bound-to-continuum structure instead yields an oscillatory saturation with $P_{\\rm sat}=U_{BF}^2/(U_{BF}^2+4J_F^2)$. A secondary mechanism, doublon tunneling at effective rate $J_D=2J_FJ_B/U_{BF}$, is what makes separated doublons stable on accessible timescales and is what the cluster expansion drops.","core_discovery":"The central claim is that the short-time doublon fraction $P_2(t)$ of a dilute lattice Bose-Fermi mixture is controlled by the decay of connected doublon clusters. For an isolated doublon in 1D the survival probability saturates at $P_{\\rm sat}=U_{BF}^2/(U_{BF}^2+4J_F^2)$ after an oscillation at frequency $\\sqrt{4J_F^2+U_{BF}^2}$. Two neighboring doublons open an additional channel: the boson tunnels away and, together with the fermion of the neighboring site, forms a triplon, while the leftover fermion (singlon) moves freely. When the boson-boson interaction $U_{BB}$ lies inside the singlon-triplon band ($U_{BB}<2dJ_F$), the pair population decays exponentially with rate $\\Gamma = (8J_B^2/J_F)\\sqrt{1-(U_{BB}/(2J_F))^2}$ in 1D; for $U_{BB}>2dJ_F$ the pair is stable. Doublons separated by empty sites decay only through slow doublon tunneling at rate $J_D=2J_FJ_B/U_{BF}$, which is suppressed by a quantum Zeno effect because $J_D\\ll\\Gamma$; this justifies a cluster expansion restricted to connected clusters. Exact t-DMRG calculations at filling fractions of 5% and 20% support the expansion at short times.","pith_inferences":["Extension: increasing $J_D/J_F$ (e.g., by lowering $U_{BF}$) should switch separated doublons from Zeno-stable to decaying; the crossover rate as a function of $J_D/\\Gamma$ is a quantitative prediction the paper leaves open.","Extension: because the triplon-singlon channel requires two bosons on one site, it is absent for fermion-fermion doublons; a matched Bose-Fermi versus Fermi-Fermi doublon experiment would isolate the boson-statistics contribution.","Extension: in 2D and 3D the decay rate should take the form $\\Gamma=(J_B^2/J_F)f(U_{BB}/J_F)$ with $f$ fixed by the singlon-triplon density of states; computing $f$ would extend the analytical predictions beyond 1D.","Extension: the cluster expansion's error as a function of filling fraction and time has not been mapped; a systematic comparison of Eq. (5) with exact dynamics for $f=10\\%,20\\%,30\\%$ would define the regime of practical applicability."],"forward_implications":["Two neighboring doublons decay exponentially with rate $\\Gamma\\propto J_B^2/J_F$ when $U_{BB}<2dJ_F$, and remain stable when $U_{BB}$ exceeds the band edge.","The doublon fraction of a dilute random gas is a weighted average over single-, double-, and triple-doublon cluster dynamics; this reproduces exact 1D t-DMRG results at short times for fillings of 5% and 20%.","Doublons separated by empty sites persist over timescales of order $1/J_D$ or longer, so the short-time relaxation of a dilute gas is almost entirely due to connected clusters.","The separation of time scales (single-doublon decay at rate $J_F$, pair decay at rate $\\Gamma$, Zeno-suppressed decay at $J_D^2/\\Gamma$) gives experiments distinct windows in which each process can be observed.","For $^{40}$K-$^{87}$Rb mixtures in a 1064 nm lattice, the parameters fall in the predicted regimes, so time-resolved doublon-fraction measurements can directly test the cluster-expansion predictions."],"supporting_citations":[{"why":"Supplies the Green's function method used to solve the single-doublon problem in 1D and the analytic saturation $P_{\\rm sat}$.","marker":"[32]"},{"why":"Provides the Bose-Fermi doublon scattering-state formalism from which the single-doublon decay and the two-doublon mapping are built.","marker":"[42]"},{"why":"The KRb experiment that motivates the doublon initial state and observable, and the source of the experimental parameter table for $^{40}$K-$^{87}$Rb.","marker":"[45]"},{"why":"The Bixon-Jortner quasi-continuum model used to derive the exponential two-doublon decay rate $\\Gamma$.","marker":"[53]"},{"why":"Supplies the t-DMRG method used to numerically benchmark the cluster expansion in 1D.","marker":"[7]"},{"why":"Together with [7] provides the time-dependent DMRG algorithm used for the exact finite-lattice simulations.","marker":"[44]"}],"fun_headline_variants":["Neighboring doublons induce exponential decay in optical lattices","Isolated doublons saturate, neighboring doublons decay exponentially","Cluster expansion captures short-time doublon decay in 1D","Doublon pairs decay via fermion-triplon conversion mechanism"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper's predictive scheme rests on the assumption that separated doublons never decay on the timescales of interest, a quantum Zeno suppression that is argued heuristically and shown for one parameter set rather than derived or scanned across parameters.","fun_headline_variants_meta":{"raw":{"variants":["Neighboring doublons induce exponential decay in optical lattices","Isolated doublons saturate, neighboring doublons decay exponentially","Cluster expansion captures short-time doublon decay in 1D","Doublon pairs decay via fermion-triplon conversion mechanism"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000405,"raw_usage":{"total_tokens":2153,"prompt_tokens":1040,"completion_tokens":1113,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":656,"completion_tokens_details":{"reasoning_tokens":1037}},"tokens_in":656,"tokens_out":1113,"duration_ms":11507,"temperature":1.0,"reasoning_tokens":1037,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:01:25.817038+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take two doublons separated by a single empty site with $J_B=0.1J_F$, $U_{BB}=J_F$, $U_{BF}=10J_F$ in 1D and track $P_2(t)$ up to $t\\approx 300 J_F^{-1}$; the cluster expansion predicts decay only at the Zeno-suppressed rate $\\Gamma_{\\rm eff}\\sim J_D^2/\\Gamma\\approx 10^{-2}J_F$. An observed decay at a rate comparable to $J_D$ (or any rate linear in $J_D$) would falsify the connected-cluster truncation.","supporting_citations":[{"cited_title":"Ohashi, Phys","cited_arxiv_id":null,"evidence_quote":"Provides the Bose-Fermi doublon scattering-state formalism from which the single-doublon decay and the two-doublon mapping are built."},{"cited_title":"Gr¨ obner, P","cited_arxiv_id":null,"evidence_quote":"The Bixon-Jortner quasi-continuum model used to derive the exponential two-doublon decay rate $\\Gamma$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Together with [7] provides the time-dependent DMRG algorithm used for the exact finite-lattice simulations."}],"review_version":1}