{"id":"ca1c13a0-5b1f-476d-87f6-ccc06463ba77","arxiv_id":"2608.12092","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A thermoelastic harvester with a protagonist-antagonist seesaw, tunable prestrain, and transversal water flow delivers 366 mW/cm³ of active material at a 66K temperature span, the highest reported power density for sub-100°C solid-state harvesters.","lead":"This paper reports a thermoelastic harvester using shape-memory alloy wires that reaches a directly measured power density of 366 mW/cm³ of active material, outperforming prior thermoelastic, thermomagnetic, pyroelectric, and thermoelectric devices in the sub-100°C range. A generalist should read it because it claims a new record for solid-state conversion of low-grade waste heat, a resource that currently is largely untapped.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Headline outperformance claim compares raw mechanical TEH power to electrical outputs of benchmarks without derating for the stated 90% conversion efficiency.","rationale":"The reader's weakest assumption identifies the same core issue: the measured mechanical output is not directly comparable to the electrical output of benchmarked devices. This is the most load-bearing concern because the paper's headline claim is explicitly comparative ('outperforming thermoelectric generators'), and the numerical comparison rests on treating 190 mW of hydraulic mechanical power as equivalent to electrical watts. The paper even acknowledges the 90% conversion efficiency in the main text, so the omission of any derating in Fig. 4 and the abstract is an internal inconsistency, not merely a difference of convention. The chilled cold side (7.5°C) is a related but secondary aggravator: it inflates the temperature span and hence the measured power. Still, even after derating by 90%, the TEH's power density (329 mW/cm³) remains above the best benchmarked TEG (205 mW/cm³), so the qualitative conclusion might survive a corrected comparison. The proposed concrete test—recomputing with explicit derating and an ambient cold side—would settle whether the outperformance claim is robust or only an artifact of comparison conventions. For these reasons, the reader's CONDITIONAL verdict is appropriate, and no change to the verdict is needed.","tokens_in":121,"tokens_out":4638,"duration_ms":48388,"concrete_test":"Recompute Fig. 4 using P_elec = 0.9 × 190 mW for the TEH, and also compute an ambient-cold-side case (Tcold = 20°C) using the measured linear power–Tcold relation in Fig. 3c to estimate P_mech at ΔT = 53.7 K, then compare with the TEG values. If the adjusted TEH power density (≈329 mW/cm³, and ≈297 mW/cm³ at ambient Tcold) still exceeds every TEG entry, the conclusion is robust; if not, the headline claim is unsupported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central comparative claim treats the directly measured mechanical output of the TEH (190 mW, 366 mW/cm³) as directly comparable to the reported electrical outputs of the benchmarked TEG (205 mW/cm³), PEG (45 mW/cm³), and TMG devices. The paper states in the Results that conversion of mechanical work to electricity can be solved by standard electromagnetic generators with efficiencies of 90% and more, and therefore is not part of this work. Yet Fig. 4 and the abstract report the raw mechanical value without derating. A fair comparison requires at least a 10% reduction to P_elec ≈ 171 mW and ≈ 329 mW/cm³; if realistic generator losses or the energy cost of the 7.5°C chilled cold side are included, the margin over the TEG shrinks further. The conclusion that the TEH outperforms thermoelectric generators below 100°C therefore does not follow from the data as presented; it depends on an unstated equivalence between mechanical and electrical power.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports a thermoelastic harvester based on NiTi shape-memory-alloy wires in a protagonist–antagonist configuration with transversal water flow and a continuously tunable prestrain mechanism. The authors directly measure force and displacement to obtain a time-averaged mechanical output power of 190 mW, corresponding to a power density of 366 mW/cm³ with respect to the active SMA volume. They benchmark this value against previously reported thermoelastic, thermomagnetic, pyroelectric, and thermoelectric generators, claiming that the device outperforms all of them in both power density and power-per-material-cost in the sub-100°C regime. The paper also presents systematic parameter sweeps (prestrain, cycle frequency, hot/cold temperatures, throttle opening) that map the device's operating space and identify the optimum operating point.","tokens_in":17205,"tokens_out":9449,"duration_ms":75976,"significance":"If the reported performance and comparisons are sound, this is an important advance for thermoelastic harvesting, which has lagged behind thermomagnetic and pyroelectric routes. The direct force and displacement measurement of mechanical work is a methodological strength, and the systematic parameter sweeps provide useful design rules for the community. The stated power density would be a new benchmark for bulk thermoelastic harvesters, and the comparison to thermoelectric generators addresses a practically relevant question. However, the central comparative claims currently rest on several comparability assumptions that need to be clarified or corrected before the headline conclusions can be accepted.","major_comments":[{"comment":"The paper compares the directly measured mechanical output power (190 mW, 366 mW/cm³) to the electrical output power of thermoelectric, pyroelectric, and thermomagnetic generators without applying the stated conversion efficiency from mechanical to electrical work. The text states that 'conversion of the extracted mechanical work to electricity can be solved by standard electromagnetic generators with efficiencies of 90% and more, and is therefore not part of this work' (Results, first section), yet Fig. 4 and the abstract report the raw mechanical value. For a fair comparison, the TEH values should be derated by at least a factor of 0.9, yielding approximately 171 mW and 329 mW/cm³. Even after this derating the TEH would still exceed the TEG benchmark of 205 mW/cm³, so the conclusion may survive, but as presented the comparison is between unequal quantities and the headline 'outperforming' claim is not directly supported by the data shown.","section":"Benchmarking and discussion, Fig. 4; Abstract; Introduction"},{"comment":"The title and abstract claim performance 'below 100°C', but the benchmark includes a thermoelectric generator (Ref. [43]) that operates with a hot side up to 140°C, as acknowledged in the text: 'We take 140°C as the upper limit of low-grade waste heat in this benchmark, just enough to include one well-characterized TEG as the strongest competitive reference.' The TEH itself is characterized at Thot = 73.7°C, so the reported comparison is not made within the stated 'below 100°C' range. The authors should either include a TEG benchmark with Thot below 100°C or revise the title and abstract to reflect the actual temperature range of the comparison.","section":"Title; Abstract; Benchmarking and discussion"},{"comment":"The demonstration uses a chilled cold side at Tcold = 7.5°C, which is well below typical ambient conditions of 20–25°C. The paper's own data in Fig. 3c show that output power decreases linearly as Tcold is increased. The benchmark TEGs are likely operated with an ambient cold side, so the comparison is not on an equal thermal basis. The authors should report the device performance at a representative ambient cold-side temperature (e.g., 20°C) or explicitly state that the headline power density requires an artificially cooled cold sink, which is not realistic for waste-heat harvesting applications.","section":"Methods, Fluid circuit and thermal control; Results, 'From actuation to work'; Fig. 3c"},{"comment":"The 'power per material cost' comparison (4.9 W/€ for the TEH vs. 2.78 W/€ for the best TEG) is based on raw element market prices from metal.com, not on the actual cost of the processed functional materials. For NiTi, the raw elements nickel and titanium are inexpensive, but the cost of producing NiTi wire is substantially higher; similarly, Bi2Te3 thermoelectric elements carry significant processing costs. The authors acknowledge this is a 'first-order proxy' and that a full-system cost comparison is premature, but the abstract presents the cost-normalized power as a decisive advantage. Unless the cost basis is made more realistic (e.g., using quoted alloy or wire prices), the cost superiority claim should be softened.","section":"Methods, data analysis; Benchmarking and discussion; Supplementary Table S3"},{"comment":"The paper excludes thermomagnetic microsystems, specifically Ref. [9] with a power density of 118 mW/cm³, because 'microtechnology is not suited to harvesting the bulk of low-grade waste heat'. Although this scope choice is defensible, the abstract's broader statement that the device is 'ahead of every reported thermomagnetic and pyroelectric generator' is then too strong. Since 118 mW/cm³ is still below the TEH value even after derating, this issue does not affect the headline conclusion, but the wording should be qualified to 'bulk' or 'macroscale' devices.","section":"Benchmarking and discussion, Fig. 4a"}],"minor_comments":[{"comment":"The text states that the force and displacement are phase-shifted by approximately 90° 'as expected at the optimum mechanical load.' It would be helpful to quantify this phase shift or provide a brief explanation of why the dissipative-load signature implies a 90° phase shift.","section":"Fig. 3a and Supplementary Fig. S1"},{"comment":"The reported system efficiency, ηsys/ηCarnot = 3.3×10⁻²%, is extremely low. The authors may want to clarify whether this is the efficiency relative to Carnot (as stated) and to note explicitly that the low value stems predominantly from incomplete heat exchange, which is only addressed later in the conclusion.","section":"Results, 'From actuation to work'"},{"comment":"The Savitzky–Golay filtering of the force and displacement signals could in principle bias the computed average power. Since raw signals are retained, a brief sensitivity check (e.g., comparing filtered vs. unfiltered average power at the optimum point) would strengthen confidence in the reported 190 mW value.","section":"Methods, Sensors and data acquisition"},{"comment":"The figure would be more informative if the TEH value were plotted with an error bar or a range reflecting the wire-to-wire spread (about 10%) reported in the frequency series, so that the separation from the TEG benchmark can be assessed.","section":"Benchmarking and discussion, Fig. 4"},{"comment":"The phrase 'in this temperature range' in the abstract is ambiguous: it is not clear whether it refers to the hot-side temperature below 100°C, the temperature span, or the benchmark range up to 140°C. Please make the intended meaning explicit.","section":"Abstract and Introduction"}],"recommendation":"major_revision","confidential_remarks":"The experimental work appears careful and the direct mechanical measurement is a real strength. However, the headline comparative claims currently overreach in three respects: the mechanical-vs-electrical power comparison, the temperature-range mismatch between the title and the benchmark, and the use of a chilled cold side. These are fixable within the scope of a revision, and the central physical result (a high mechanical power density from a thermoelastic harvester) is likely to survive. I would encourage the editor to request a revision that addresses these comparability issues before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe paper is worth taking seriously for the measured power density, but the headline outperformance claim does not survive contact with its own reported numbers.\n\nWhat is actually new: a directly measured 366 mW/cm³ of mechanical power from a thermoelastic harvester using NiTi wires, with a protagonist–antagonist seesaw, tunable prestrain, and transverse flow. The parameter sweeps are careful and the instrumentation is credible. They measure force and displacement directly, no fitted material model, and the time-averaged power is well defined. That is a real contribution: the next-best thermoelastic device was 213 mW/cm³, and this is the first systematic mapping of prestrain, frequency, load, and temperature on output power for this architecture.\n\nThe soft spots are in the benchmarking, not the measurement. The paper claims to outperform thermoelectric generators below 100°C, but the 366 mW/cm³ figure is raw mechanical power. The authors themselves state that mechanical-to-electrical conversion runs at 90% or better, so a fair comparison should derate to about 329 mW/cm³, still ahead of the benchmarked TEG at 205 mW/cm³ but with a thinner margin. More importantly, the comparison uses a cold side at 7.5°C, which is not an ambient sink, and the TEG benchmark set is essentially a single device. They also exclude thermomagnetic microsystems at 118 mW/cm³ on the grounds of market focus, which is defensible but should be stated as a scope choice, not a physical comparison. The efficiency is 0.033% of Carnot, which the paper is honest about but which matters for any real-world claim about harvesting low-grade waste heat.\n\nMinor: the extrapolation to longer wires and more wires is asserted, not demonstrated, and the material-cost comparison uses raw-element prices, which is a fair first-order proxy but not a system cost. The PV comparison (4.9 W/€ vs 1 W/€) mixes material cost with installed system cost, which is apples to oranges.\n\nOverall: the measurement is solid and the device is a genuine step forward for thermoelastic harvesting. The comparative claims need a derating step and a more careful benchmark set before they can be taken at face value. This is a revise-and-resubmit, not a reject. It deserves a serious referee.\n\nWould I bring it to reading group? Maybe, as a case study in how comparison frames decide a headline. I would not cite it myself in the next year, but I would point someone working on waste-heat recovery to the measured data.\n\nRecommendation: send to peer review; ask for the derating and a reworked benchmark.","headline":"Solid measured power density for a thermoelastic harvester, but the 'outperforms thermoelectrics' headline only holds if you compare raw mechanical power to electrical output.","tokens_in":17676,"tokens_out":1962,"would_cite":false,"duration_ms":17482,"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":"A thermoelastic harvester built from NiTi shape-memory-alloy wires reaches 366 mW/cm³ of directly measured mechanical power, and on this measure outperforms reported thermoelectric generators below 100°C.","keywords":["thermoelastic harvesting","shape memory alloys","NiTi","low-grade waste heat","energy harvesting","power density","thermoelectric generators","solid-state energy conversion"],"falsifier":"Run the same harvester with the cold side at 20°C instead of 7.5°C and convert the hydraulic output with a real electrical generator; if the resulting electrical power density falls below the 205 mW/cm³ of the best benchmarked thermoelectric device at the same temperature span, the central outperformance claim would be falsified.","tokens_in":16702,"feed_emoji":"♨️","tokens_out":10525,"duration_ms":86644,"temperature":0.7,"pith_summary":"Low-grade waste heat below 100°C is abundant but has no economically viable solid-state converter. This paper reports that thermoelastic harvesting—using NiTi shape-memory-alloy wires that switch between a stiff high-temperature phase and an easily deformed low-temperature phase—can fill that gap. With a protagonist–antagonist seesaw, a tunable prestrain, and transversal water flow, the device produces a directly measured mechanical power density of 366 mW/cm³ of active material, about 1.7 times the best prior thermoelastic harvester and ahead of reported thermomagnetic and pyroelectric generators. On power per material cost (4.9 W/€) it also beats the thermoelectric generators benchmarked in this temperature range, making thermoelastic harvesting a credible competitor for recovering sub-100°C waste heat.","feed_headline":"366 mW/cm³: shape-memory harvester tops thermoelectrics","feed_subtitle":"A directly measured NiTi device reaches 4.9 W per euro of active material, a new mark for sub-100 °C waste heat.","key_machinery":"The load-bearing mechanism is the protagonist–antagonist thermoelastic harvester: a seesaw connecting two sets of NiTi wires that are alternately heated and cooled, so the contracting hot side does the prestrain work on the cold side. The prestrain at the central bearing sets the operating point on the SMA's stress–strain response, and the enclosed loop on that diagram is the mechanical work delivered per cycle. Transversal water flow decouples the cycle frequency from wire length, and the direct measurement chain—forces from load cells, displacement from a laser, power $\\bar{P} = \\langle F_1 \\dot{x} + F_2 \\dot{x}\\rangle$—is what lets the paper map device performance without material-property estimates.","core_discovery":"The central discovery is that all three historical limitations of thermoelastic harvesters are engineering choices, not material limits, and removing them unlocks competitive power. The device runs a four-stage thermodynamic cycle on the stress–strain plane of NiTi: strain the cold martensitic wire, heat it into stiff austenite, let it contract under load, then cool and repeat. Two wire bundles are coupled by a seesaw so the hot contracting side mechanically prestrains the cold side (protagonist–antagonist recovery), a movable central bearing makes the prestrain a continuously tunable control parameter, and water flows transversally across the wires so heat-exchange time no longer grows with wire length. The authors measure force and displacement directly, compute output power as the time average of $F_i \\dot{x}$ summed over both sides, and report $\\bar{P}=190$ mW mechanical output at a peak of 370 mW; with an active SMA volume of 0.518 cm³ this is 366 mW/cm³. Against published devices with directly measured output, this is the highest power density of any thermoelastic, thermomagnetic, or pyroelectric harvester in the sub-140°C benchmark, and exceeds the benchmarked thermoelectric generators both per active-material volume and per raw-material cost. The electrical step is not built in; the paper assumes standard electromagnetic conversion at 90% or better.","pith_inferences":["If the mechanical power is converted at the stated 90% generator efficiency, the electrical power density would be about 330 mW/cm³; measuring this with a real electrical load at a cold side near 20°C would test whether the outperformance over TEGs survives outside the lab's chilled conditions.","The benchmark excludes thermomagnetic microsystems (118 mW/cm³) and uses a 7.5°C cold side, so the headline outperformance over thermoelectric generators is framed by those choices; a broader or more ambient-condition comparison could change the ranking.","Because NiTi transformation temperatures can be tuned by composition or heat treatment, the same architecture could be adapted to different waste-heat windows, including smaller temperature spans if hysteresis is reduced.","The cost comparison uses raw element prices only, not installed system cost; a full system-level cost analysis would be needed before claiming economic competitiveness with photovoltaics or TEG modules."],"forward_implications":["Because transversal flow removes the length–frequency coupling, output power should scale with both the number and length of SMA wires, so the same device rules transfer to larger active volumes.","The prestrain mechanism turns the force–strain balance into a tunable operating parameter, so a single device can be re-optimized for different hot and cold temperatures and loads instead of being fixed by geometry.","The companion simulation reaching 51.6% of Carnot (near the Curzon–Ahlborn ceiling for maximum power) means the standard NiTi material is not the bottleneck; better heat exchange between water and wire is the remaining lever.","The first-order transition's large hysteresis requires temperature spans of order 60 K for best performance; thermomagnetic generators, with nearly hysteresis-free transitions, remain better suited to spans below 30 K."],"supporting_citations":[{"why":"Establishes the protagonist–antagonist configuration this device builds on, where each wire batch alternately takes hot and cold sides and prestrain energy is recovered mechanically.","marker":"[33]"},{"why":"Reports the previous best thermoelastic power density (213 mW/cm³) that the 366 mW/cm³ result is compared against.","marker":"[13]"},{"why":"Gives an earlier thermoelastic-engine comparison point (37 mW/cm³) used in the benchmark table.","marker":"[12]"},{"why":"Provides the pyroelectric-generator benchmark (45 mW/cm³) included in the ferroic comparison.","marker":"[10]"},{"why":"Provides the thermomagnetic-generator benchmark (0.5 mW/cm³) included in the ferroic comparison.","marker":"[8]"},{"why":"Thermomagnetic microsystem at 118 mW/cm³ whose exclusion from the benchmark defines the bulk-harvester scope of the claim.","marker":"[9]"},{"why":"Serves as the strongest directly measured thermoelectric benchmark below 140°C (205 mW/cm³, 2.78 W/€) that the TEH claims to outperform.","marker":"[43]"},{"why":"Introduces the transversal-flow chambers used here, which decouple heat-exchange time from wire length.","marker":"[36]"},{"why":"Companion simulation that places the material's efficiency near the Curzon–Ahlborn maximum-power ceiling, used to interpret the measured system efficiency.","marker":"[37]"},{"why":"The original endless-wire heat-engine layout whose per-interface transforming volume and fixed prestrain are the design limitations this paper replaces.","marker":"[32]"}],"fun_headline_variants":["Shape-memory harvester trumps thermoelectrics below 100°C","366 mW/cm³: thermoelastic device beats thermoelectric generators","NiTi engineering unlocks record low-grade heat harvesting","Tunable prestrain makes thermoelastic power densest under 100°C"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The claim that the device outperforms thermoelectric generators depends on treating the directly measured mechanical power of the harvester as comparable to the electrical power of the benchmarked generators, because the paper assumes 90% conversion efficiency and uses a 7.5°C cold side that is below typical ambient conditions.","fun_headline_variants_meta":{"raw":{"variants":["Shape-memory harvester trumps thermoelectrics below 100°C","366 mW/cm³: thermoelastic device beats thermoelectric generators","NiTi engineering unlocks record low-grade heat harvesting","Tunable prestrain makes thermoelastic power densest under 100°C"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000203,"raw_usage":{"total_tokens":1434,"prompt_tokens":1045,"completion_tokens":389,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":661,"completion_tokens_details":{"reasoning_tokens":313}},"tokens_in":661,"tokens_out":389,"duration_ms":4343,"temperature":1.0,"reasoning_tokens":313,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T00:16:09.263466+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same harvester with the cold side at 20°C instead of 7.5°C and convert the hydraulic output with a real electrical generator; if the resulting electrical power density falls below the 205 mW/cm³ of the best benchmarked thermoelectric device at the same temperature span, the central outperformance claim would be falsified.","supporting_citations":[{"cited_title":"Working characteristics of a reciprocating-type heat engine using shape memory alloys.JSME International Journal Series B, 41(2): 344–350, 1998","cited_arxiv_id":null,"evidence_quote":"Establishes the protagonist–antagonist configuration this device builds on, where each wire batch alternately takes hot and cold sides and prestrain energy is recovered mechanically."},{"cited_title":"Stewart, Reza Mirzaeifar, Eckhard Quandt, and Shashank Priya","cited_arxiv_id":null,"evidence_quote":"Reports the previous best thermoelastic power density (213 mW/cm³) that the 366 mW/cm³ result is compared against."},{"cited_title":"Characteristics of a new power generation system with application of a shape memory alloy engine.Electrical Engineering in Japan, 165(3):8–15, 2008","cited_arxiv_id":null,"evidence_quote":"Gives an earlier thermoelastic-engine comparison point (37 mW/cm³) used in the benchmark table."},{"cited_title":"Large harvested energy with non-linear pyroelectric modules","cited_arxiv_id":null,"evidence_quote":"Provides the pyroelectric-generator benchmark (45 mW/cm³) included in the ferroic comparison."},{"cited_title":"Memory alloy heat engine and method of operation","cited_arxiv_id":null,"evidence_quote":"The original endless-wire heat-engine layout whose per-interface transforming volume and fixed prestrain are the design limitations this paper replaces."}],"review_version":1}