{"id":"f1dfb6c9-7fec-4efd-8567-d87fe2d0fca2","arxiv_id":"2508.17036","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":8,"one_line_summary":"A single diradical molecule acting as a particle-exchange heat engine shows up to 53% of the Curzon-Ahlborn efficiency limit, with power and efficiency enhanced by Kondo correlations.","lead":"Researchers built a working heat engine from a single molecule a few nanometers across, using heat from one electrode to push electrons through the molecule and generate electric power. The engine works better when electron interactions (the Kondo effect) are active, reaching roughly half the ideal efficiency for such devices.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 53%-of-CA efficiency claim is a model estimate based on J_Q = εI and an Onsager relation that the authors themselves state fails in the Kondo regime; it is not a direct measurement.","rationale":"After reading the paper, I find the strongest claim—that Kondo correlations push a single-molecule engine to 53% of the Curzon–Ahlborn limit—is exactly as load-bearing as the reader's weakest assumption. The efficiency denominator is not measured; it is assigned as ε per transferred electron, and the Carnot normalization is tied to the Onsager relation. The paper admits Onsager symmetry fails in the very regime where the headline number is obtained. The argument that nonzero conductance at zero thermocurrent is negligible addresses only the fit quality, not the heat current: in linear response J_Q contains a thermal-conductance term that survives at I = 0 and a co-tunneling contribution not proportional to ε. Equation 7's neglect of passive heat flow makes it an upper-bound toy model, not a measurement. I do not think this warrants rejection: the directly measured thermocurrent and conductance maps are credible, the Kondo identification (T_K = 4.3 K, g ≈ 2, singlet-triplet excitation) is supported by independent zero-bias and magnetic-field data, and the power enhancement with Kondo correlations is a robust qualitative observation. But the abstract's efficiency number is a model inference under a violated assumption. The internal inconsistency between '2.3%' and '0.53 η_CA' and the admitted systematic error from restricting the conductance fit (SI, last paragraph) further reduce confidence in the quantitative headline, though they are secondary to the heat-flow issue. The appropriate disposition remains CONDITIONAL: accept the qualitative power enhancement, but require either a revised analysis with the Onsager assumption removed and all data disclosed, or a direct heat-flow measurement, before the 53%-of-CA claim can be used.","tokens_in":14812,"tokens_out":10135,"duration_ms":110807,"concrete_test":"Recompute Q_H and ΔT/T at the B = 0, R_load = 2 MΩ operating point from the full Anderson-impurity heat current, using NRG or a numerically exact method with the independently fitted Γ_L, Γ_R, U, α, and T_K = 4.3 K, and substitute the resulting η into the power-versus-efficiency curve. If the corrected η_Pmax falls below roughly half the reported 0.53 η_CA, or if the η-versus-σ trend changes, the Kondo-regime efficiency claim fails. A complementary check is to integrate a local electron thermometer (SNS/SINIS) into a second device and measure J_Q directly, removing the Onsager-based denominator entirely.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The headline efficiency (0.53 η_CA at R_load = 2 MΩ, and the claimed Kondo enhancement) is computed as η = P/Q_H, with Q_H = α(V_g − b)I_m (SI, 'Data analysis and fitting', step 7), and the Carnot normalization ΔT/T is obtained from the amplitude ratio a/a_G using the Onsager relation L = εG/T (SI Eqs. 2, 10-11, step 6). The authors explicitly state in SI 'Device description' that Onsager symmetry does not hold in the Kondo regime and that transport then proceeds by co-tunneling rather than through the level at ε. Their defense—that deviations only matter where I_th ≈ 0—is insufficient: heat conduction at zero particle current still degrades efficiency, and co-tunneling contributions to J_Q need not vanish where the net thermocurrent is small. Equation 7 (η/η_C = GR/(1+GR)) deliberately omits passive heat flow, so it cannot validate the Kondo-regime number. Since the device design does not permit direct heat-flow measurement, the 53% value rests on an assumption the paper itself flags as violated. The power-enhancement observation is less affected, but the efficiency and 53%-of-CA claim are not established.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports a single-molecule particle-exchange heat engine based on an all-organic diradical SMe-2OS molecule in an electromigrated break junction. Using simultaneous AC conductance and thermocurrent measurements, the authors identify a Kondo resonance, estimate a Kondo temperature of 4.3 K, and introduce an asymmetry parameter σ to quantify the thermocurrent line shape. They calculate output power P = R_load I_th² and efficiency η = P/Q_H with Q_H estimated from εI, and claim that Kondo correlations enhance both power and efficiency, reaching about 53% of the Curzon–Ahlborn limit at R_load = 2 MΩ. The power-enhancement observation is based on directly measured thermocurrent, while the efficiency and the σ-dependence of performance rely on a single-level Onsager model that the authors themselves state is not valid in the Kondo regime.","tokens_in":15130,"tokens_out":3187,"duration_ms":36593,"significance":"If the efficiency claim were established, this would be an important experimental demonstration that many-body Kondo correlations can improve thermoelectric energy conversion at the single-molecule scale, bringing molecular engines close to the performance of semiconductor quantum-dot heat engines. The paper has clear strengths: simultaneous measurement of conductance and thermocurrent, magnetic-field control to tune the Kondo regime, a fitting protocol for the asymmetry parameter, explicit discussion of the model's limitations, and a data/code availability statement. However, the central efficiency number is not directly measured but derived from a model whose key assumptions (Onsager symmetry, single-level sequential tunneling, negligible passive heat flow) are acknowledged to break down in the Kondo regime, and the model is also used to extract the very parameter σ whose performance-enhancing role is then claimed. The manuscript therefore needs substantial revision to either justify the efficiency estimate in the Kondo regime or reframe the headline claim as a model-based indication rather than an experimental measurement.","major_comments":[{"comment":"The 53%-of-Curzon–Ahlborn efficiency claim is not a measured efficiency: η is computed as P/Q_H with Q_H = α(V_g−b)I_m (SI 'Data analysis and fitting', step 7), and the Carnot normalization ΔT/T is obtained from a/a_G using the Onsager relation L = εG/T (SI Eqs. 2 and 10–11, step 6). The authors explicitly state in SI 'Device description' that Onsager symmetry does not hold in the Kondo regime and that transport then proceeds via co-tunneling rather than through the single level at ε. Their defense that the deviations matter only where I_th ≈ 0 is insufficient: heat conduction at zero particle current still degrades efficiency, and co-tunneling contributions to J_Q need not vanish where the net thermocurrent is small. Equation (7), η/η_C = GR/(1+GR), deliberately omits passive heat flow, so it cannot validate the Kondo-regime number. The device design does not permit direct heat-flow measurement, so the 53% value rests on an assumption the paper itself flags as violated. I recommend either measuring or bounding the heat current independently, using a Kondo-valid transport model for J_Q, or clearly reframing the efficiency as a model estimate with explicit caveats and removing the unqualified 'demonstrate' language.","section":"SI 'Device description'; main text Eq. (7)"},{"comment":"There is a circularity in using the same model both to extract the asymmetry parameter and to compute the performance enhancement attributed to it. The thermocurrent traces are fitted with Eq. (12), whose σ is then inserted into the conductance fit (Eq. 13), and the same functional forms (Eqs. 10–11) are used to derive P and η via Eqs. (5)–(7). Since the power P = R_load I_m² is directly measured, the correlation between larger σ and larger P is partly tautological: σ is extracted from the thermocurrent asymmetry that directly enters P. The manuscript should quantify how much of the claimed σ-enhancement is independent of this fitting procedure, for example by testing the model's predictive power at different load resistances or by comparing against an independent determination of σ from conductance and Mott-relation analysis.","section":"SI 'Data analysis and fitting', steps 6–8"},{"comment":"The conductance fit, which is used to determine a_G and hence the Carnot normalization via a/a_G = αΔT/T, excludes data outside the thermal broadening range left of the conductance peak because the Kondo conductance plateau deviates from Eq. (13). The authors acknowledge that this can lead to systematic errors and that the errors are visible in Fig. 2. Since the extracted ΔT/T directly enters the reported η/η_C, this systematic error must be propagated into all efficiency numbers, especially the 53% claim at R_load = 2 MΩ, where the Kondo effect is strongest. Without such an uncertainty analysis, the quantitative efficiency values are not supported by the data.","section":"SI 'Data analysis and fitting', step 8"},{"comment":"The claim that Kondo correlations 'significantly enhance' both power and efficiency conflates two different evidential standards. The power enhancement is directly observed in the measured thermocurrent under magnetic-field variation. The efficiency enhancement, however, is computed from the same model that defines σ, and the model's heat-flow estimate is not independently validated. The Discussion's attribution of the enhancement to a sharp asymmetric Kondo resonance is plausible but not established by the data, as the paper provides no direct spectroscopic or theoretical evidence that the resonance asymmetry is the operative mechanism. I would ask the authors to clearly separate the directly measured power enhancement from the model-dependent efficiency statement, and to temper the conclusion accordingly.","section":"Main text, Fig. 3 and Discussion"}],"minor_comments":[{"comment":"The phrase 'Theroretical considertions' contains two typographical errors and should read 'Theoretical considerations'.","section":"SI 'Supporting Information Available'"},{"comment":"The caption contains the stray text '1 /uni03BCm' that appears to be a rendering issue; it should display as '1 μm'.","section":"SI Methods, Fig. 4 caption"},{"comment":"The parametric power-versus-efficiency curves in Fig. 3a and 3d would benefit from error bars or shaded uncertainty regions, since the efficiency values are derived quantities and the fitting procedure is acknowledged to have systematic errors.","section":"Main text, Fig. 3"},{"comment":"The notation switches between ε, α(V_g−b), and ε − Tσ without a single consistent definition of the detuning variable; defining one symbol, e.g., x = (α(V_g−b) − μ)/k_BT, and using it throughout would improve readability and reduce the risk of confusion in the fitting protocol.","section":"Main text, Eq. (1) and SI Eqs. (8)–(11)"},{"comment":"The oscillations around 4 T and 6.6 T in the I_th+/I_th− ratio are noted as unexplained; a brief comment on possible origins (e.g., molecular configuration switching or measurement artifacts) would help the reader judge whether they affect the extracted σ trend.","section":"SI 'Thermocurrent' section"}],"recommendation":"major_revision","confidential_remarks":"The paper is potentially interesting, and the direct power-enhancement observation appears solid. My main concern is the mismatch between the headline efficiency claim and the model-dependence acknowledged in the SI. The authors may be able to fix this by rephrasing the abstract and conclusions to state clearly that the efficiency is a model-based estimate under assumptions that fail in the Kondo regime, and by providing a propagation-of-error analysis for the systematic fit deviations. If they cannot do so within the manuscript's scope, the 53%-of-CA claim should be removed or substantially downgraded. I would also encourage the editor to ask the authors to reconcile the stated data-availability commitment with the absence of a repository link in the manuscript text."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear [Colleague],\n\nThe thing to know: this is a real experimental step for single-molecule heat engines. They operate a diradical molecule as a particle-exchange engine, show via gate maps and magnetic-field dependence that it's in the Kondo regime (TK≈4.3 K), and measure directly that the electrical power, P=I_th^2 R_load, is larger at zero field than when Kondo is suppressed by 8.8 T. That part is credible and new: previous molecular engines didn't run in the Kondo regime, and the power enhancement is a direct measurement, not a fit.\n\nWhat they do well: the device work is careful (simultaneous AC conductance/thermocurrent, load resistors, magnetic field sweeps), and they are upfront about where the model is being stretched. The asymmetry parameter σ is a useful phenomenological way to quantify the line-shape asymmetry of the thermocurrent, and the numerical comparison linking larger σ to better engine performance is plausible. The SI is thorough and the data are promised open.\n\nNow the soft spots. The headline claim—efficiency up to 53% of the Curzon–Ahlborn limit—is not a measured efficiency. Heat flow J_Q is never measured; it's estimated as ε I_th, with ε = α(V_g−b), and the Carnot normalization ΔT/T comes from the ratio a/a_G using the Onsager relation L=εG/T. The authors themselves state in the SI that Onsager symmetry does not hold in the Kondo regime and that transport proceeds by co-tunneling. Their counter—that deviations only matter where I_th≈0—doesn't rescue the efficiency estimate, because heat conduction at zero particle current still degrades η and co-tunneling contributes to J_Q even when net thermocurrent is small. So the 53% number is model-dependent in exactly the regime they care about. The stress-test note lands. Also, the conductance fitting excludes points outside the thermal-broadening range; they acknowledge this causes systematic errors, but it means the σ values used in the performance analysis are fitted under a restricted window. That's a mild circularity (σ is both fitted and used to compute η), though the power enhancement doesn't hinge on σ.\n\nOne more editorial annoyance: they mix Carnot and Curzon–Ahlborn normalizations. In the text they quote 2.3% of CA at 100 kΩ and 0.53η_CA at 2 MΩ; the theory curve is plotted as η/η_C.\n\nWho is this for? People working on molecular thermoelectrics and quantum-dot heat engines. It deserves a serious refereeing: the experiment is novel, the direct power evidence is real, and the efficiency claim can be fixed by reframing it as an estimate with clear caveats or, ideally, by a direct heat-flow measurement. I'd send it out.\n\nBest, [Your name]","headline":"Solid molecular thermoelectric experiment, but the 53%-of-CA efficiency is a model estimate that inherits the Onsager assumption the authors concede fails in the Kondo regime.","tokens_in":15689,"tokens_out":4660,"would_cite":true,"duration_ms":42975,"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":"A single diradical molecule converts heat to work with efficiency boosted by Kondo correlations to 53 percent of the Curzon-Ahlborn limit.","keywords":["single-molecule heat engine","Kondo effect","particle-exchange heat engine","thermoelectric power factor","Curzon-Ahlborn limit","diradical molecule","thermocurrent spectroscopy"],"falsifier":"Measure the heat current leaving the hot electrode directly, for example with a superconductor–normal-metal thermometer integrated on the hot side, while running the engine at the optimal load; if the measured heat flow disagrees with the value implied by $L = \\varepsilon G/T$ and the fitted level position, then the 53 percent efficiency is not a directly measured efficiency.","tokens_in":14590,"feed_emoji":"⚡","tokens_out":15658,"duration_ms":140781,"temperature":0.7,"pith_summary":"This paper reports a working particle-exchange heat engine built from a single diradical molecule, only a few nanometers across, bridging two gold electrodes at cryogenic temperature. It claims that Kondo correlations—the many-body screening of the molecule's unpaired spin by conduction electrons—sharpen the molecular orbital's energy filter and roughly double the maximum power output, raising efficiency at maximum power to about 53 percent of the Curzon-Ahlborn limit. The claim matters because it turns strong electron-electron interactions, usually viewed as harmful to molecular transport, into a resource for nanoscale thermoelectric energy conversion. The paper's own supporting information notes that the heat-flow estimate behind this efficiency relies on a linear-response Onsager relation that is not strictly valid in the Kondo regime.","feed_headline":"Kondo effect lifts a single-molecule heat engine to 53% of ideal","feed_subtitle":"At 1.8 K, the molecule's Kondo resonance sharpens the energy filter, rivaling quantum-dot engines.","key_machinery":"The load-bearing object is the molecular particle-exchange heat engine: a single SMe-2OS diradical molecule, tunnel-coupled to hot and cold gold electrodes, with a back gate tuning the energy $\\varepsilon$ of the transport level through $\\varepsilon = \\alpha(V_g - b)$. Transport is energy-selective single-electron exchange, and the engine performance is captured by the self-consistent load equations $P = \\varepsilon^2 G^2 R\\,(\\Delta T)^2/(T^2(1+GR)^2)$ and $\\eta/\\eta_C = GR/(1+GR)$, obtained from the Onsager relation $L = \\varepsilon G/T$. The Kondo resonance—the many-body antiferromagnetic screening cloud that forms around the molecule's local spin below $T_K \\approx 4.3$ K—is the mechanism that sharpens and asymmetrizes the transmission window. Its effect is quantified by the asymmetry parameter $\\sigma$, introduced by fitting the thermocurrent to $I_{\\mathrm{th}}(V_g) = A\\,\\varepsilon\\,\\Delta T\\,T^{-2} f(\\varepsilon - T\\sigma)(1 - f(\\varepsilon))$; $\\sigma$ is treated as a fitting asymmetry rather than a thermodynamic entropy in the Kondo regime. The Curzon–Ahlborn efficiency $\\eta_{\\mathrm{CA}} = 1 - \\sqrt{T_c/T_h}$ is the benchmark for the engine's efficiency at maximum power.","core_discovery":"The central claim is that a static, particle-exchange heat engine can be realized in a single SMe-2OS diradical molecule and that Kondo correlations improve, rather than degrade, its thermodynamic performance. The molecule is gate-tuned so its neutral charge state has a spin-$1/2$ ground state, giving a Kondo resonance with $T_K \\approx 4.3$ K. Under a temperature difference between the electrodes, this resonance acts as a narrow, slightly asymmetric energy filter: thermocurrent, power factor, and output power all increase, and a magnetic field that suppresses the Kondo effect cuts the maximum power by roughly half. Efficiency at maximum power grows with the asymmetry parameter $\\sigma$ and reaches about 53 percent of the Curzon-Ahlborn limit at an optimized load resistance of 2 M$\\Omega$, a performance the authors compare with the best quantum-dot heat engines. The heat flow entering this efficiency is computed from the Onsager relation $L = \\varepsilon/T\\,G$ and the assumption that each transferred electron carries heat $\\varepsilon$, rather than measured directly.","pith_inferences":["Inference: The same device should operate as a heat pump when the load is reversed, with the asymmetry parameter $\\sigma$ controlling the cooling coefficient of performance; this follows from the same line-shape model but is not tested in the paper.","Inference: If the heat flow were measured directly rather than estimated from $L = \\varepsilon G/T$, the 53 percent figure could shift, but the Kondo-induced enhancement of output power would remain a separate, robust observation from the raw thermocurrent data.","Inference: Synthesizing closed-shell analogues of SMe-2OS with identical anchor groups would test whether the thermocurrent asymmetry and efficiency gain disappear with the spin, separating the Kondo mechanism from geometric or coupling asymmetries."],"forward_implications":["Molecular-scale particle-exchange engines can reach the same thermodynamic performance class as semiconductor quantum-dot heat engines, with efficiency at maximum power around 53 percent of the Curzon-Ahlborn limit.","Kondo correlations become a design resource for low-temperature thermoelectric devices: they boost power output and efficiency instead of degrading coherent transport.","Load resistance provides a practical control knob, since efficiency at maximum power peaks at an optimal value (here 2 M$\\Omega$) and declines for larger loads.","The asymmetry parameter $\\sigma$, extracted from a single thermocurrent gate trace, can serve as a compact device figure of merit for comparing molecular heat engines.","Tuning spin ground states and exchange couplings by chemical design or by magnetic field offers a direct route to optimizing nanoscale energy filtering."],"supporting_citations":[{"why":"Defines the gate-dependent thermocurrent form and the asymmetry parameter $\\sigma$ used throughout the analysis.","marker":"[11]"},{"why":"Supplies the magnetic-field scaling $B_{\\mathrm{th}} \\approx B_c = 0.75 k_B T_K/(g\\mu_B)$ used to extract $T_K = 4.3$ K.","marker":"[14]"},{"why":"Provides the quantum-dot heat engine benchmark whose near-Curzon-Ahlborn performance the molecule is compared with.","marker":"[8]"},{"why":"Reports thermoelectric signatures of the Kondo resonance in nanowire quantum dots, supporting the enhanced, Mott-violating thermocurrent observed here.","marker":"[13]"},{"why":"Documents the thermopower line shapes of a Kondo-correlated quantum dot and the deviations from Mott's relation that the molecular data reproduce.","marker":"[25]"},{"why":"Gives Anderson-impurity-model calculations in which an asymmetric Kondo resonance amplifies the Seebeck coefficient and thermoelectric power.","marker":"[27]"},{"why":"Establishes the simultaneous AC heating and lock-in thermocurrent measurement technique that the experiments rely on.","marker":"[22]"}],"fun_headline_variants":["Kondo boosts single-molecule heat engine to 53% of ideal","Single-molecule heat engine hits 53% of ideal via Kondo","Kondo sharpens energy filter in single-molecule heat engine","Single diradical molecule runs heat engine with Kondo boost"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The reported efficiency assumes each transferred electron removes heat $\\varepsilon$ from the hot reservoir, through the linear-response relation $L = \\varepsilon G/T$, even though the paper acknowledges that Onsager symmetry—which this relation presupposes—breaks down in the Kondo regime.","fun_headline_variants_meta":{"raw":{"variants":["Kondo boosts single-molecule heat engine to 53% of ideal","Single-molecule heat engine hits 53% of ideal via Kondo","Kondo sharpens energy filter in single-molecule heat engine","Single diradical molecule runs heat engine with Kondo boost"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000278,"raw_usage":{"total_tokens":1607,"prompt_tokens":853,"completion_tokens":754,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":469,"completion_tokens_details":{"reasoning_tokens":679}},"tokens_in":469,"tokens_out":754,"duration_ms":7820,"temperature":1.0,"reasoning_tokens":679,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T17:08:47.026370+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the heat current leaving the hot electrode directly, for example with a superconductor–normal-metal thermometer integrated on the hot side, while running the engine at the optimal load; if the measured heat flow disagrees with the value implied by $L = \\varepsilon G/T$ and the fitted level position, then the 53 percent efficiency is not a directly measured efficiency.","supporting_citations":[],"review_version":2}