{"id":"89e8485d-7f81-419a-b4b3-321609139126","arxiv_id":"2608.12055","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Tuning the temperature of an atomic rubidium reservoir changes the microscopic collision rates that transfer heat to a cesium working medium, allowing a quantum Otto engine's power to be optimized at constant efficiency.","lead":"Scientists built a tiny heat engine from 40 cesium atoms inside a cloud of rubidium atoms and showed that warming the rubidium cloud changes how heat flows into the engine, letting them speed up the engine cycle and maximize power without losing efficiency. The result is a direct experimental handle on the microscopic interaction that transfers heat, a knob theorists had predicted but no one had turned in an atomic engine before.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Cyclic closure condition G_h=G_c is not sufficient for a true Otto cycle; finite-time partial equilibration leaves the working medium displaced, so Eq. (13) overestimates work and the reported optimum may be an artifact.","rationale":"The central claim is that tuning the kinetic temperature of the atomic reservoir controls the microscopic heat transfer law and thereby optimizes power at fixed efficiency. The experimental population dynamics and the collision-rate model are credible, and the open data are a real asset. The quantitative engine figures, however, all pass through Eq. (13), which converts measured heat into work under the assumption that the working medium returns to its initial state after the cycle. The reader identified this assumption as the weakest point. My analysis of the triangular rate matrices sharpens it: equal thermalization factors are not merely unproven as a closure condition; for finite-time partial equilibration they are generally inconsistent with exact periodicity, because the cooling dynamics reaches its ground-state fixed point only asymptotically. The residual unreturned internal energy is omitted from the work balance, so the reported optimal power and optimal temperature are not established by the data as presented. The concrete full-cycle test would settle whether the qualitative power-enhancement conclusion survives a correct periodic-state calculation. Because the reader's conditional verdict already requires this kind of check, my read does not move the verdict; it confirms and strengthens the condition under which the paper should be accepted.","tokens_in":18667,"tokens_out":14957,"duration_ms":147847,"concrete_test":"Use the open dataset and the stated rate matrices (Eqs. (7)-(9)) to compute, for each (τ_h,τ_c) pair used in Fig. 6, the periodicity defect δ=∥exp(K_c τ_c) exp(K_h τ_h)e0 - e0∥. If δ is not below the atom-number shot-noise floor, recompute the engine figures of merit using the true periodic steady state P* satisfying P*=exp(K_c τ_c)exp(K_h τ_h)P*, and evaluate Q_h, Q_c, W, and P from P*. Then check whether P_max,opt=28.0 nK/ms and T_Rb,opt=1410(5) nK survive; if they shift or disappear, the quantitative power-enhancement claim is not supported by the current analysis.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is the assertion in Sec. IV, after Eq. (13), that cyclic operation follows from equal thermalization factors G_h(τ_h)=G_c(τ_c). For the directed triangular rate matrices of Eqs. (7)-(8), this is not a periodicity condition. Starting from the ground state e0, heating gives P_h=exp(K_h τ_h)e0, and cooling gives P_D=exp(K_c τ_c)P_h; periodicity requires P_D=e0. Because K_c is triangular with all nonzero eigenvalues negative and a single fixed point e0, P_D approaches e0 only asymptotically; for any finite τ_c, P_D≠e0 whenever P_h has any excited population. With the actual-initial-state normalization implied by Q_i∞=-Σ c_k d_k in Eq. (11), G_c is the fraction of the available excitation removed, so G_h=G_c leaves a residual mean excitation 6G_h(1-G_c) rather than zero. If instead Q_c∞ is held at the full-inversion value -6χB_c, equality forces complete removal of excitation, which is unattainable at finite τ_c unless G_h=1. In either reading, the equality G_h=G_c does not certify return to the initial state for the finite, incomplete equilibration shown in Fig. 7(b), where G is approximately 0.8-1.0. Equation (13) therefore omits the unreturned internal energy and overstates |W|, the power P, and the position and magnitude of P_max,opt=28.0 nK/ms at T_Rb,opt=1410(5) nK. The experiment's raw population data and the parameter-free collision model are valuable, and control of relaxation timescales is plausible; the unsupported step is specifically the conversion of measured heats into cyclic work.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports an experimental quantum Otto engine in which seven Zeeman states of ultracold 133Cs atoms serve as the working medium and an ultracold 87Rb cloud serves as the reservoir. Heat exchange is mediated by inelastic s-wave spin-exchange collisions with energy-dependent cross sections, producing asymmetric heating and cooling dynamics. By tuning the Rb kinetic temperature, the authors control the relative durations of the isochoric strokes and thereby the total cycle time, and they report a power maximum at fixed efficiency, with a predicted optimal point T_Rb,opt = 1410(5) nK and P_max,opt/k_B = 28.0 nK/ms. The experimental population dynamics are compared with a parameter-free model based on coupled-channel scattering cross sections, plus a finite-Rb-lifetime correction with parameters fitted from separate lifetime measurements.","tokens_in":19082,"tokens_out":8180,"duration_ms":83571,"significance":"If the cycle-closure issue is resolved, this would be a valuable demonstration: direct microscopic control of a system-reservoir interaction is used to modify the heat transfer law and to optimize finite-time power at fixed efficiency. The strengths of the paper include time-resolved population measurements, a spin-dynamics model with no free parameters for the collision rates, cross sections from independent coupled-channel calculations, and a public data availability statement. The quantitative peak claim, however, depends on an unproven cyclic-closure condition and on model extrapolation beyond the measured temperature range, so the central numerical result is not yet established.","major_comments":[{"comment":"The statement that cyclic operation follows from equal thermalization factors G_h(τ_h)=G_c(τ_c) is not established and is in fact not a periodicity condition for the triangular rate matrices in Eqs. (7)-(8). Starting from e0, the state after heating is exp(K_h τ_h)e0, and after cooling it is exp(K_c τ_c)exp(K_h τ_h)e0; because all nonzero eigenvalues of K_c are negative and e0 is its unique fixed point, this product equals e0 only in the limit τ_c→∞. Equal thermalization factors fix only a scalar heat fraction, not the population vector, so the internal energy of the unreturned population is omitted from Eq. (13). At the reported optimum the thermalization factor is close to but below unity (Fig. 7(b)), so the omission is first order in 1-G. Please provide either a derivation of full state return or a direct measurement of the population vector after a complete cycle, and recompute |W|, P, and η from the actual return state; until then the extracted power and the optimum at T_Rb,opt=1410(5) nK are not established as cycle-averaged quantities.","section":"Section IV, Eq. (13)"},{"comment":"The identification Q_c,∞=-6χB_c is valid only if the cooling stroke starts from full population inversion. Because Q_c,∞=-Σ c_k d_k in Eq. (11) depends on the initial population vector, and the cooling stroke starts from the partially inverted state reached after finite τ_h, the normalization used in G_c overestimates the available heat whenever G_h<1. This is not a negligible correction in the regime shown in Fig. 7(b), where the thermalization factor is below unity at higher T_Rb. Please evaluate Q_c,∞ from the measured state at the start of cooling, or explicitly justify that the initial state for cooling is the full inversion within experimental uncertainty.","section":"Section III, Eqs. (11)-(13)"},{"comment":"The headline optimum T_Rb,opt=1410(5) nK and P_max,opt/k_B=28.0 nK/ms are not direct measurements: the data are reported for T_Rb = 679, 858, and 1140 nK, and 1410 nK lies outside the measured range. The optimum therefore comes from numerical model extrapolation, including the finite-lifetime correction with parameters fitted in Appendix C. Please state explicitly that the optimum is a model prediction, and support the stated uncertainty either by measuring at or near T_Rb ≈ 1410 nK or by propagating the uncertainties of the fitted lifetime parameters into the predicted peak position and height.","section":"Section V and Fig. 7(c)"}],"minor_comments":[{"comment":"The claim that the numerical model is parameter-free should be qualified: the bare collision model in Eq. (6) is parameter-free for the scattering rates, but the curves that include the finite Rb lifetime use five fitted parameters (T_Rb,ss, T_Rb,0, γ_ht, N_Rb,0, γ_ev) from Appendix C. Please state clearly which curves are purely parameter-free and which include the lifetime correction.","section":"Section III and Section V"},{"comment":"The heat leak Q_l is introduced in Eq. (14) but never defined; please give its explicit expression or cite the derivation in Ref. [33] so that the efficiency formula can be checked directly.","section":"Section IV, Eq. (14)"},{"comment":"The uncertainty quoted for T_Rb,opt = 1410(5) nK is not explained in the text; please report how this error bar is obtained (for example, from the covariance of the fitted model parameters or from bootstrap resampling of the measured data).","section":"Section V"},{"comment":"In Fig. 4(c) the cooling heat is magnified by a factor of 20, which makes the cooling dynamics difficult to assess at the same scale as the heating dynamics; a separate panel with identical axes would clarify the comparison.","section":"Figure 4"}],"recommendation":"major_revision","confidential_remarks":"The experimental platform and the microscopic scattering model are strong, and the data availability is a clear asset. The main risk to the central claim is the unsupported cycle-closure condition in Eq. (13); if the authors can provide full state-return data or a corrected derivation, the paper would be a strong contribution. In the meantime, the quantitative optimum should be presented as a model prediction."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my read for your files.\n\nThe new thing is the control knob. By tuning the Rb kinetic temperature they change the relative rates of exothermal and endothermal s-wave spin-exchange collisions, which shifts the time allocation between heating and cooling strokes and makes the total cycle duration a non-monotonic function of T_Rb. That mechanism is experimentally demonstrated and is genuinely new relative to their earlier collision-driven engines (Refs. 33, 34) and to Pancotti et al. (Ref. 18), which predicted the speed-up but didn't supply a physical handle. The parameter-free spin model is a real asset: it reproduces the population dynamics, and the scattering cross sections come from independent coupled-channel calculations. The constant efficiency 0.489(2) versus the predicted 0.491 is a good consistency check.\n\nNow the soft spots.\n\nFirst, the cyclic closure condition. Section IV asserts that G_h=G_c guarantees the working medium returns to its initial state. For seven-level triangular rate matrices this is not proven and, as the stress-test note explains, it is generally false. If G_c is normalized to the full-inversion heat, a partially heated state cannot be fully cooled in finite time; if G_c is normalized to the actual initial excitation, equality leaves residual internal energy 6χB_c G(1-G). In either reading, Eq. (13) overstates work unless the engine operates very close to full inversion. The data in Fig. 7(b) show G around 0.8–1.0, so the error may be small, but the paper neither proves nor measures state return. This is the load-bearing quantitative step and it needs a direct check.\n\nSecond, the headline optimum T_Rb,opt = 1410(5) nK is a model extrapolation beyond the measured range (679–1140 nK), and the 5 nK uncertainty does not include the 100+ nK temperature uncertainties. That precision is not supported by the data.\n\nThe citation pattern looks honest; the self-citations are to their prior platform papers and the theory proposal they extend. Data are open.\n\nOverall, this is a worthwhile paper. The qualitative claim—reservoir temperature controls the microscopic heat transfer law and can be used to optimize power at fixed efficiency—is plausible and the underlying dynamics are well characterized. The quantitative work extraction and the optimum need fixing, not the concept. I'd send it to peer review with a request for a direct measurement of the final population vector after a full cycle (or a quantitative bound on the unreturned energy) and a proper uncertainty propagation on T_Rb,opt. If the residual is as small as I suspect, the main message will survive. I'd bring it to the reading group; it's a good paper to argue about.","headline":"Solid experimental mechanism, but the cycle-closure step G_h=G_c is asserted without proof; the headline optimum is a model extrapolation, so the quantitative claims need a referee's attention while the qualitative insight is worth keeping.","tokens_in":19599,"tokens_out":8830,"would_cite":true,"duration_ms":92175,"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":"This paper claims that a quantum Otto engine's power can be optimized by tuning the kinetic temperature of its atomic reservoir, which reshapes the heat transfer law while efficiency stays fixed.","keywords":["quantum heat engine","Otto cycle","ultracold atoms","spin-exchange collisions","system-reservoir coupling","heat transfer law","power optimization","cesium-rubidium mixture"],"falsifier":"Measure the full seven-Zeeman-state population vector after one complete Otto cycle at the operating points of Figures 5 and 6 and compare it with the vector at the start of the cycle; if populations have drifted, the extracted work, power, and efficiency do not describe a true cyclic engine. A second test is to run the engine at $T_{\\mathrm{Rb}}=1410$ nK and above to see whether the measured peak power occurs where the model predicts.","tokens_in":18474,"feed_emoji":"⚛️","tokens_out":9835,"duration_ms":89022,"temperature":0.7,"pith_summary":"The paper presents a quantum Otto engine whose working medium is seven Zeeman states of ultracold cesium atoms immersed in a rubidium spin reservoir, and claims that tuning the reservoir kinetic temperature gives direct microscopic control over the heat transfer law between engine and reservoir. Because exothermal and endothermal spin-exchange collisions have different energy-dependent scattering cross sections, raising $T_{\\mathrm{Rb}}$ slows the heating stroke while accelerating the cooling stroke, shifting the cycle time allocation and making the total cycle duration non-monotonic in temperature. The measured performance at three bath temperatures, matched by a parameter-free numerical model, shows a power maximum at a finite reservoir temperature, $T_{\\mathrm{Rb,opt}} = 1410(5)$ nK, with peak power $P_{\\mathrm{max,opt}}/k_B = 28.0$ nK/ms and efficiency fixed at $\\eta = 0.489(2)$ independent of stroke durations. If this claim holds, it validates the idea that a bath's microscopic relaxation spectrum, not just its temperature, is a controllable resource for finite-time quantum thermal machines.","feed_headline":"Tuning reservoir temperature boosts a quantum engine's power","feed_subtitle":"Ultracold cesium–rubidium collisions let a seven-level Otto engine adjust stroke timing and raise peak power at fixed efficiency.","key_machinery":"The load-bearing object is the triangular rate matrix $K_i(T_{\\mathrm{Rb}},B_i)$ of the collisional master equation $\\dot P=K_i P$, whose state-dependent rates $\\Gamma_{m,i}=n(r,T_{\\mathrm{Rb}})\\,\\langle\\sigma_{m,i}(E_{\\mathrm{col}},B_i)\\,v_{\\mathrm{col}}\\rangle$ come from thermal averages over energy-dependent s-wave scattering cross sections. Because the matrix is triangular, its eigenvalues are just its diagonal entries, and the heat transferred in a stroke is a multi-exponential sum over those modes, summarized by the thermalization factor $G_i(t)=Q_i(t)/Q_{i,\\infty}$. The physical asymmetry doing the work is the opposite energy scaling of exothermal versus endothermal cross sections: heating rates fall with $T_{\\mathrm{Rb}}$ while cooling rates rise, which lets the reservoir temperature reallocate time between strokes and thereby shape the power–duration trade-off.","core_discovery":"The central discovery is that the heat transfer law of this engine is set at the microscopic level by energy-dependent s-wave spin-exchange scattering, and that this law can be reshaped by changing the kinetic temperature of the atomic reservoir. The rate matrices for the heating and cooling strokes are both triangular, with diagonal entries equal to the state-dependent rates $\\Gamma_{m,i}(T_{\\mathrm{Rb}},N_{\\mathrm{Rb}},B_i)$, and heat evolves as a weighted sum of exponentials $Q_i(t)=\\sum_k c_k d_k[\\exp(\\lambda_k t)-1]$ rather than as a single exponential. Exothermal cross sections decrease monotonically with collision energy while endothermal cross sections show threshold behaviour, so the heating rates fall and the cooling rates rise as $T_{\\mathrm{Rb}}$ increases in the explored range; the paper exploits this asymmetry to control the relative duration of the two isochoric strokes. Imposing the cyclic-closure condition $G_h(\\tau_h)=G_c(\\tau_c)$ links the stroke durations, and the resulting total cycle duration is non-monotonic in $T_{\\mathrm{Rb}}$, which yields a well-defined optimum at finite temperature. The paper also shows that replacing the multi-exponential dynamics with a single average rate overestimates maximum power by factors of 1.52–1.71 and underestimates the optimal cycle duration, so the usual phenomenological single-rate heat transfer law is questioned.","pith_inferences":["Inference: the predicted optimum at $T_{\\mathrm{Rb,opt}}=1410(5)$ nK sits above the highest measured temperature of 1140 nK, so a direct run at 1410 nK and higher would test whether the power peak is truly there rather than an extrapolation from the model.","Inference: the same asymmetric-rate control could be exercised by other knobs that shift the relative heating-to-cooling rate ratio, such as magnetic field near Feshbach resonances, bath density, or choice of collision partner; the paper fixes magnetic fields and density.","Inference: if the cyclic-closure assumption is verified by a full state-return measurement, the approach suggests a design principle, shaping a reservoir's relaxation-time spectrum is an engine resource, that could be tested in engineered solid-state or spin-bath platforms."],"forward_implications":["Tuning the kinetic temperature of an atomic reservoir is a demonstrated control knob for the heat transfer law of a quantum engine, independent of modifying the system Hamiltonian.","Power can be optimized at fixed efficiency: in the measured window the efficiency stays $\\eta=0.489(2)$ while the maximum power changes with $T_{\\mathrm{Rb}}$.","The optimum is reached in the low-fluctuation regime with sub-Poissonian relative power fluctuations, so high power does not come at the cost of stability.","Microscopic multi-exponential relaxation should replace single-exponential phenomenological laws when modelling such engines; the mono-exponential reference model overestimates power by 52–71% in this system.","The mechanism is expected to transfer to other systems with multiple relaxation channels, including spin-bath and central-spin environments, where the relaxation spectrum rather than the bath temperature controls performance."],"supporting_citations":[{"why":"Supplies the theoretical prediction that controlling system–bath coupling can speed up thermal machines, which this experiment claims to realize.","marker":"[18]"},{"why":"Previous realization of the atomic-collision quantum Otto engine in the same platform; provides the engine scheme, heat-leak term, and magnetic-field calibration used here.","marker":"[33]"},{"why":"Frames the cesium–rubidium system as a physical implementation of a collision model, grounding the repeated-interaction description of equilibration.","marker":"[40]"},{"why":"Provides the spin-bath reservoir model used to treat the rubidium cloud as an atomic spin reservoir.","marker":"[41]"},{"why":"Defines the quantum Otto cycle and its reversible efficiency form, against which the engine's cycle and efficiency expression are set.","marker":"[43]"},{"why":"Supplies the master-equation description of hyperfine spin dynamics and the spin-selective fluorescence readout used to extract populations and heat.","marker":"[46]"},{"why":"Gives the thermal average formula for rate coefficients used to connect collision-energy distributions to $T_{\\mathrm{Rb}}$-dependent transition rates.","marker":"[47]"}],"fun_headline_variants":["Tuning atomic collisions boosts quantum engine power","Microscopic heat control keys quantum engine optimization","Reservoir temperature tunes quantum engine stroke timing","Scattering rates shape quantum engine power at fixed efficiency","Control heat exchange to boost quantum engine output"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The claim depends on assuming that matching how much heat is absorbed and released in the two strokes is enough to bring the seven-level working medium back to its starting configuration each cycle; equal heat amounts do not automatically return every population to its initial value when heating and cooling follow different rate structures, and the paper does not show a direct measurement of full state return.","fun_headline_variants_meta":{"raw":{"variants":["Tuning atomic collisions boosts quantum engine power","Microscopic heat control keys quantum engine optimization","Reservoir temperature tunes quantum engine stroke timing","Scattering rates shape quantum engine power at fixed efficiency","Control heat exchange to boost quantum engine output"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000394,"raw_usage":{"total_tokens":2119,"prompt_tokens":1045,"completion_tokens":1074,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":661,"completion_tokens_details":{"reasoning_tokens":1004}},"tokens_in":661,"tokens_out":1074,"duration_ms":8710,"temperature":1.0,"reasoning_tokens":1004,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T00:18:29.749524+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the full seven-Zeeman-state population vector after one complete Otto cycle at the operating points of Figures 5 and 6 and compare it with the vector at the start of the cycle; if populations have drifted, the extracted work, power, and efficiency do not describe a true cyclic engine. A second test is to run the engine at $T_{\\mathrm{Rb}}=1410$ nK and above to see whether the measured peak power occurs where the model predicts.","supporting_citations":[{"cited_title":"Pancotti, M","cited_arxiv_id":null,"evidence_quote":"Supplies the theoretical prediction that controlling system–bath coupling can speed up thermal machines, which this experiment claims to realize."},{"cited_title":"Kosloff and Y","cited_arxiv_id":null,"evidence_quote":"Defines the quantum Otto cycle and its reversible efficiency form, against which the engine's cycle and efficiency expression are set."},{"cited_title":"Schmidt, D","cited_arxiv_id":null,"evidence_quote":"Supplies the master-equation description of hyperfine spin dynamics and the spin-selective fluorescence readout used to extract populations and heat."},{"cited_title":"Cannoni, Relativisticσv rel in the calculation of relics abundances: A closer look, Phys","cited_arxiv_id":null,"evidence_quote":"Gives the thermal average formula for rate coefficients used to connect collision-energy distributions to $T_{\\mathrm{Rb}}$-dependent transition rates."}],"review_version":1}