Arbiter (Step 3): Popa-Simil, "Nano-Hetero Structure for direct energy conversion" (NSTI-Nanotech 2007)

Inputs weighed: facts.json, claims.md, deep/, physicist.md, mathematician.md, contrarian.md, sites.md, research/ notes. Source tags: [read] / [web] URL / [memory] confidence / [calc] my own arithmetic.

Source gap note. None for the paper itself: it has a full text layer and Eq.1 and Figs. 1-2 were read from page images [read: p.2]. But the project's evidence is one 4-page concept paper (a SRIM-2003 simulation, no measurement). Everything about patents, later papers and other groups' experiments comes from the Investigator's files and was read only in part by the reviewers. The primary Russian papers (Anufrienko et al. 2006, 2008; Yershov et al. 2011-2012), the sibling Nanotech 2007 paper (3021, ">90%"), the TMS 2008 and Cleantech 2007 papers, and the USPTO file wrappers have not been read by anyone. Conclusions that rest on those are marked as resting on second-hand reports.

Summary of agreement and disagreement

All three reviewers agree: the paper contains no measurement and no derivation of its headline efficiency, power density or module voltage; Eq.1 has four errors and is decorative (not what SRIM uses); the "84%" of the inventor's patent is, in his own words, "simply the efficiency of ionization energy stopped in the 'C' layer", an upper estimator needing "experimental buildup" [read: US20100061503A1 para. 0121]; no CIci device by the inventor has been measured in the record; and the mechanism (differential secondary-electron emission across thin insulators) is ordinary physics with partial support from other groups.

They differ on labels for the strong form of claim 1 (physicist: Refuted by calculation; mathematician: Internally inconsistent; contrarian: Never properly tested / not testable as stated) and on claim 6, and in some errors of their own which I correct below.

Recomputations where reviewers differ or erred

1. Efficiency bound (claim 1). Electrical work from one collected electron is at most e V. If a fragment of energy E_f creates one free electron per W of deposited energy, the number of electrons is at most E_f / W, and efficiency <= e V / W, whatever the fragment energy. All three reviewers' arithmetic is consistent (physicist 0.1% at W=10 eV; mathematician and contrarian 3e-4 at W=30 eV): 0.01 V / 30 eV = 3.3e-4. Check [calc]: 100 MeV / 30 eV = 3.3e6 electrons; x 0.01 eV = 33 keV; 33 keV / 100 MeV = 3.3e-4. To reach 80% at W = 30 eV, V would have to be 0.8 x 30 = 24 V per collected electron. Two reviewer statements were muddled and are dropped: the physicist's "each electron would carry about 8 V per 10 mV" and the contrarian's "of order 100 V per stack". The correct requirement is about 24 V per electron-crossing (volts, not mV), for W = 30 eV. Web check of W (the physicist and mathematician had it as unverified [memory] medium): W is 3.6 eV in silicon and about 30 eV in gases [web: https://www.nuclear-power.com/nuclear-engineering/radiation-detection/semiconductor-detectors/silicon-based-semiconductor-detectors/ ; search summary]. Metals and insulators in the stack are not covered by those two numbers. Taking the most generous W = 3.6 eV, the bound at 10 mV is 0.28%; at W = 1 eV it would still be 1%. The conclusion does not turn on W. Caveat all three note and I keep: the bound assumes charge is delivered at the paper's stated cell voltage (<10 mV, "given by the breakdown voltage" [read: p.2-3]) with no external field supplying energy. A device that works at volts per crossing (as vacuum-gap fission electric cells do) is not covered by the bound; it is also not what the paper describes.

2. Patent power example (mathematician). 167 MeV x 1.602e-19 = 2.675e-11 W (26.7 pW); / 10 mV = 2.675 nA [calc, confirmed]. That current is 1.67e10 electrons per fission; ionization at 30 eV supplies about 5.6e6. The ratio is about 3000. With only 5.6e6 electrons at 10 mV, power is 9 fW, 0.033% of 26.7 pW [calc, confirmed]. The patent's own "100% at 10 mV" example therefore needs about 3000 times more charge than the energy can liberate. Also confirmed: 5/(5+0.5+1+0.5) = 71%, 5/6 = 83.3%; neither is 84%. Note that 168/200 = 0.84 is the fragment share of fission energy [memory, high; consistent with the physicist's web figure of about 169 MeV: https://www.sciencedirect.com/topics/engineering/fission-fragment], and the match is suggestive but no source says it was derived that way.

3. Claim 6 (power density). 5000 kW/cm3 = 5e6 W/cm3; at 200 MeV (3.204e-11 J) per fission: 1.56e17 fissions/cm3/s [calc]. U-235 metal 4.8e22 atoms/cm3 x 585 b = 28.1 /cm, so thermal flux = 5.6e15 n/cm2/s [calc]; burn-out time = (4.8e22 x 3.2e-11 J = 1.54e12 J/cm3) / 5e6 W = 3.1e5 s = 3.6 days [calc]. All three reviewers agree on these numbers. Reconciliation of the verdicts: the number is not forbidden by energy conservation (contrarian is right), but as a sustained fuel power density it needs the flux of the best research reactors in pure fissile metal and consumes the fuel in days (physicist and mathematician are right). The paper does not say which it means. Correction to the physicist's baseline. The physicist wrote the LWR figure garbled ("100 W/L is 0.1 kW/L ... about 0.1 W/cm3"). The correct figure is: PWR core average about 110 kW/L = 110 W/cm3 [web: summary of https://www.sciencedirect.com/science/article/abs/pii/S0029549321004404 , "several 4-loop PWRs in the U.S. operate at 110 kW/L"]; fuel-pellet averages 22-90 W/cc and sodium fast reactors above about 300 W/cc [web, same search, other results]. The paper's baseline of "0.2-0.5 kW/cm3" (= 200-500 W/cm3) is therefore of the order of fast-reactor core or peak fuel values, higher than a PWR core average. It is roughly right as a peak/fast-reactor fuel number, not a typical-reactor one. The mathematician's "~100 W/cm3" for LWR core is correct. Correction to the physicist's inconsistency ratio. TMS 2008 says existing reactors run "200-500 W/cm3" and the new system could reach "up to 1,000 times higher" [web: https://laro.lanl.gov/esploro/profile/liviu_popasimil , abstract as read by the Investigator]. That is 200-500 kW/cm3. The 2007 paper's 5000 kW/cm3 is 10 to 25 times larger than that (5000/500 = 10; 5000/200 = 25), equivalently 10,000-25,000 times the baseline versus 1,000 times. The physicist's "10 to 1000" and the contrarian's "10x" are both loose; the discrepancy is a factor of 10-25. Unexplained; the TMS paper itself was not read.

4. PuCoRh5 (claim 5). The mathematician could not confirm Tc 8.7 K for "PuCoRh5"; the physicist accepted it. Web check: 8.7 K is the Tc of PuRhGa5, discovered by Wastin et al. (ITU Karlsruhe, 2003) shortly after PuCoGa5 [web: https://www.osti.gov/etdeweb/biblio/20412256 ; search summary]. I found no compound called PuCoRh5. The paper's "PuCoRh5" is most likely a garbled name for PuRhGa5 (and the physicist's remark that its Rh content makes it unusable by the author's own table rests on that name). I state this as likely, not certain: the paper's Table 1 would need re-reading to see which composition the cross-sections were computed for. PuCoGa5 Tc about 18.5 K and the fall of Tc with alpha self-irradiation (about 0.2 K/month initially, slowing) [web: https://www.osti.gov/biblio/917542 , summary quoted by the mathematician, full papers not read] stand.

5. Josephson number [calc, confirmed]: K_J = 2e/h = 483.6 THz/V = 483.6 MHz/microvolt (the paper prints 483 597.879 GHz/V); at 10 mV = 4.84 THz. (Garbled line corrected in the QE revision.)

Corrected claim list with final verdicts

Claims 1-6 as in claims.md, with these wording corrections: claim 1 should read "the abstract says efficiency 'may be higher than 80%'" (a possibility, not a result [read: p.1]); the 80% is not derived anywhere in the paper; the sibling 2007 paper (3021) says ">90%" by "quantum effects" [read: Investigator, grep only]. Claim 5's second compound should be flagged "as named in the paper; Tc 8.7 K matches PuRhGa5". Claim 6 should note the TMS 2008 figure.

Claim 1: direct conversion of fission-product kinetic energy at over 80%

• Agreement: no derivation; 84% is a stopping fraction; the paper's own mV limit makes e V / W tiny; measured efficiencies by others are far lower (10^-5 % Yoshida/Tanabe; 25% Anufrienko, second-hand).
• Disagreement: the label. I side with the physicist's "Refuted by calculation" for the strong form, because it needs only accepted physics plus the paper's own stated voltage (<10 mV), and the mathematician's inconsistency finding (patent's 10 mV / 2.67 nA example) is the same calculation seen from inside the patent. The contrarian's "not testable as stated" is right about the abstract's wording ("may be"), but the claim becomes testable once the paper's own parameters are used, and under them it fails. Weight: the contrarian's point that the bound does not apply to volt-scale, field-assisted or vacuum-gap devices is valid and is why the refutation is limited to the mV cell the paper describes.
• Final: strong form (>80% electrical efficiency at mV cell voltage): Refuted by calculation. Weak form (some useful electrical fraction, order 1-25%, in thin-insulator layered stacks): Never properly tested for CIci; related stacks by others have been tested (10^-5 % to a reported 25%), and the 25% is documented in one published second-hand source (Klassen et al., InTech 2013, pp. 272-273) with the efficiency definition unknown. It was not dismissed without testing; it was never built and measured by the inventor.
• What would settle it: a beam test as in the patent's FIG. 15 reporting open-circuit voltage, load curve, current per incident fragment and electrical energy out / beam energy in; and a deep read of Anufrienko et al. 2008 to learn how its 25% is defined (it would show whether it escapes the e V / W bound by operating at volts, or defines efficiency per electron energy rather than fragment energy).

Claim 2: shower-intensity difference moves electrons across a nm insulator to a low-yield collector

• Agreement (all three): ordinary secondary-electron emission physics; nothing in mainstream physics would change; independent measurements (Klassen/Yershov insulator thinning 200 um to 1 um gives 3 orders; Stepanov 2018 up to 1.15 V in asymmetric supercapacitors under neutrons) support the trend; no net-current model is written down.
• Mathematician's hidden-assumption point stands: Fig. 2 shows bulk ionization, not emission into the gap; the insulator is itself ionized; keV delta electrons cross it regardless.
• Contrarian's point stands: the Russian group cites Popa-Simil, so is not fully independent in concept, though its experiments are its own.
• Final: Plausible, untested as a power source for this design; the general effect is Established at low power. The quantitative "ballistic crossing, net current" picture is Not testable as stated until a net-current model exists.
• Settle it: calculate (SRIM plus electron transport, e.g. a Monte Carlo of secondary electrons) the net electron flux between layers, then measure current vs insulator thickness on a built stack.

Claim 3: reactor would run cold

• Final: Not testable as stated ("cold" is never quantified); weak form (heat falls in proportion to the electrical fraction) Plausible, untested, entirely dependent on claim 1. Floor on remaining heat: about 15% of fission energy (neutrons, gammas, decay, neutrinos escape) [memory, high], and the paper itself says that without conversion "the whole energy becomes heat" [read: p.2]. Strong form follows claim 1's refutation.
• Settle it: a heat balance on a built stack under beam (power in vs electrical out vs temperature rise).

Claim 4: mV modules to grid voltage by DC/AC and AC/AC cascade, possibly Josephson and Andreev

• Physicist: Plausible, untested; mathematician: Never properly tested; contrarian: Never properly tested. Difference is only emphasis. I choose the latter because the paper gives no loss model and the start-up problem (semiconductor switching thresholds of order 0.3-0.7 V against 10 mV input [memory, medium; not web-checked; the verdict does not depend on it]) is unaddressed.
• Josephson: 4.84 THz at 10 mV [calc]; it is a frequency standard/detector, not a power converter. Andreev reflection is a charge-transfer process at a normal/superconductor interface. Neither is supported by an equation in the paper.
• Final: Never properly tested (engineering chain; each stage exists, the loss budget is missing); Josephson/Andreev role: Not testable as stated.
• Settle it: a loss budget for a 10 mV source through a cold-start converter chain; then a bench test.

Claim 5: superconducting actinide compounds with radiation self-repair

• Agreement: Tc values of the discovered compounds are real; self-repair under fission flux is asserted with no rate equation; cryogenic operation suppresses thermal annealing; fission-fragment damage far exceeds alpha self-damage; nobody has tested it in a fission flux.
• Final: Established for the weak form (PuCoGa5 18.5 K; 8.7 K belongs to PuRhGa5 [web above]). Never properly tested for use in a reactor with self-repair. Where tested (alpha self-irradiation), Tc falls with age (about 0.2 K/month initially, saturating) [web, summary only]: that is Tested, not supported for "no loss of superconductivity", but not a test of the paper's dynamic-equilibrium idea under fragments, so the lifetime-repair claim keeps the Never properly tested label. The physicist's "Rh 144 b makes it unusable" point is held until Table 1 is rechecked against the likely PuRhGa5 mislabel.
• Settle it: irradiate a thin PuCoGa5 or surrogate film at cryogenic temperature with a heavy-ion beam and track Tc and resistivity vs fluence.

Claim 6: 0.2-0.5 kW/cm3 to a theoretical 5000 kW/cm3

• Final: Not testable as stated (undefined "theoretical"), with the sustained-fuel reading Refuted by calculation (5.6e15 n/cm2/s thermal flux in pure U-235, burn-out in 3.6 days); Internally inconsistent with the author's TMS 2008 abstract by a factor 10-25; conversion has no bearing on flux, so any increase is a heat-removal/flux claim, not a CIci result. A local power density above the baseline from better heat removal remains Plausible, untested.
• Settle it: the author states the assumptions (flux, spectrum, geometry, duration) behind 5000 kW/cm3; then a neutronics calculation.

Final verdict table

Material kept in the record whatever the verdicts

• Eq.1 errors, the 700 vs 600 nm thickness mismatch, the 9% energy loss in the SRIM stack, and the "1200 electrons/nm3" figure that fits iodine/UO2 but not U or Au (~4400-4700) [read: deep/].
• Patents: US20100061503A1 (with its candid caveat at para. 0121 and a FIG. 15 beam-test design with no results found), US20130121449A1 (extends to fusion/transmutation via "nuclear entanglement and quantum non-localities"; Not testable as stated, no mainstream support), US8067758B2 (shielding). The inventor's later turn to LENR (ILENRS 2012: "the reported phenomenon is real"; "new physics, not in contradiction to the quantum mechanics predictions"; NETS 2013) with no CIci measurements.
• Others' work: Klassen/Yershov (insulator thinning), Anufrienko/Chernov (25%), Yoshida/Tanabe (1e-5 %), Stepanov 2018 (1.15 V), the fission electric cell lineage (Safonov 1957, Krieve 1966, Ursu 1978-88), and the 2026 MDPI review (Gabbar et al.) calling the technology "mainly conceptual ... patent-stage" with 0.85 as an ideal ceiling. The review's characterization is its authors' view, not a test result.
• Interested parties: the inventors own LAVM LLC and the patents, so efficiency figures are their own characterization. The Russian groups have related programmes and cite Popa-Simil.

Absence-of-evidence analysis

Where the record goes dark: from 2007 to 2026 there is a steady stream of conference papers (2007-2015, 2021) and three patent filings, but no published measurement of a built CIci stack, no journal report of a beam test, and both patent applications (US12/384,074; US13/297,243) abandoned; the 2012 roadmap asks for "appropriate funding, equipment and support" [read: research note]. This is an anomaly to explain, not a neutral gap. Competing explanations: 1. Lack of funding/fabrication capability (a small company, nm actinide stacks hard to make, beam time costly). For: the 2012 funding appeal, LAVM LLC's small-company profile, the author's stated need for a "national research program". Against: no grant, SBIR or contract records have been checked either way; the 2007 US Army items cited suggest some contract contact [read: research note], nature unknown. 2. Experiments done but unpublished or negative. For: the paper-only output pattern; the physics budget above predicts a tiny result at mV, which is unattractive to publish. Against: no record of any measurement, claim of one, or statement of failure by the author; later papers keep restating the concept without data. 3. Trade secrecy/commercial reasons. For: patent filings and a company structure. Against: the author publishes openly and the applications were abandoned, so there is nothing secret being protected in the record. 4. Absorption or continuation under another name or classification/suppression. Nothing found for or against; no evidence supports suppression. The research that the Russian groups carried out on similar structures went on in the open, which argues against a field-wide block. 5. The physics budget itself (mV output makes the efficiency tiny; the idea shifted to LENR). For: the author's later pivot. Against: not a documented reason. No explanation is established. Decisive records: USPTO file wrappers (examiner rejections, abandonment reasons), New Mexico Secretary of State filings for LAVM LLC and the Los Alamos Academy of Sciences, SBIR.gov/USASpending/DTIC award lists, Alabama irradiation-centre beam-time logs, and the unread primary Russian papers. The Investigator did not reach these (time limit).

Open questions

1. How is the 25% of Anufrienko et al. defined, and does it escape the e V / W bound by operating at volts?
2. What does "may be higher than 80%" / ">90%" refer to in the 2007 papers? (Answered in the QE revision: nothing new stands behind the 90%; sibling paper 3021 asserts it in one sentence, QE read it in full.)
3. Which compound is "PuCoRh5" in Table 1 (likely PuRhGa5)?
4. What assumptions give 5000 kW/cm3, and why is it 10-25 times the TMS 2008 figure?
5. Why were the two applications abandoned, and was any beam test ever run?
6. Net-current model: what is the electron flux between layers with ionization in every layer, including the insulator?

Web checks made (budget 8, used 3)

• PuRhGa5 Tc 8.7 K (Wastin 2003): confirmed [web: https://www.osti.gov/etdeweb/biblio/20412256].
• PWR core and fuel power density: [web: https://www.sciencedirect.com/science/article/abs/pii/S0029549321004404 ; search summary].
• W values (3.6 eV Si, about 30 eV gases): [web: https://www.nuclear-power.com/nuclear-engineering/radiation-detection/semiconductor-detectors/silicon-based-semiconductor-detectors/ ; search summary]. Not checked: converter start-up voltages, SiO2 breakdown field, electron mean free paths (all [memory] medium; no verdict depends on them). Search summaries only; full papers were not read.

QE revision

Inputs: qe.md (verdict MINOR) and the new deep reads in deep/: the main paper pp.3-4 (added section), Klassen et al. InTech 2013 ch.11, US20100061503A1 (paras 0075-0200), Stepanov et al. 2018, the 2012 "Roadmap to Fusion Battery" (81 pp.) and US20130121449A1. I read the deep-read notes and the QE's notes, not every source page again; page-level claims are tagged to the deep reader's [read: p.N]. [calc] is my arithmetic, rerun with the venv. Web checks added: 2 (running total 5 of 8). QE found no change to claims 1, 2, 4 or 6 as headline verdicts; I confirm that, and add parts to claims 1, 2, 3 and 5.

Table of changes

Claim 1: efficiency theory, the patent's own caveats, and a drift in the number

1. The Klassen ideal-efficiency theory (QE missed item 2). Klassen et al. define eta = W_ext / W_abs and show eta tends to 1 when each emitted electron reaches the collector with zero kinetic energy: the self-bias ephi equals the electron kinetic energy K, so eta = K/(K+W_f), where W_f is the work function [read: Klassen deep read, eqs. 4-8, p.278-279, from page images]. Their model is for gamma-ray electrons, not fission fragments, and it assumes every absorbed photon energy goes into emitted electrons carrying K + W_f. They state the opposite regime too: if K only slightly exceeds the work function, "the efficiency of the system will be always low." Their own measured comparator is far from the ideal: about 1e-5 % in one place and "less than 0.01 %" or "not more than 0.01%" in others (p.279 vs pp.265/267; a factor of 1000 apart inside the same chapter) for the Yoshida et al. patent device, which used gamma pumping [read: deep read]. They give four reasons for the gap (geometry, thick emitter layers, thick insulator with backflow, spread of electron energies). - How it relates to my bound. It does not contradict e V / W. It is the high-voltage branch of the same energy accounting: high efficiency needs a self-bias V near K/e. For fission-fragment delta electrons the maximum energy is T_max = 2.3 keV (light fragment) and 1.05 keV (heavy) [calc, reproduced: 4 m_e/M x E for 100 MeV, A=95 and 67 MeV, A=140; the deep reader got the same; the QE's "1.65 keV" was an average]. Most delta electrons are tens to hundreds of eV. The ideal condition therefore needs about 0.1 to 2 kV across the insulator. Across a 1-10 nm insulator that is 10 to 1000 MV/cm [calc], against about 10 MV/cm intrinsic breakdown of SiO2 [memory, high] and tunnelling leakage from a few MV/cm. A 1-10 nm oxide holds about 1-10 V; with V of a few volts and K of hundreds of eV, eta is about eV/K, a few percent or less before transport losses [deep-reader inference from Klassen's own eq. 5/8, which I accept]. Klassen's chapter names the conflict (insulator thin for transport, thick enough "to prevent breakdown, as well as large leakage current") but gives no numbers [read: p.281]. - Weight.* This is a statement by a group that supports the concept (and is itself an interested party, see below). It sharpens why the mV cell fails and where a working design would have to go: volt-to-kV scale, thicker or vacuum/aerogel insulators (their proposal, p.281), not the nm stack with <10 mV in the 2007 paper. - Minor slip in Klassen: above the optimum K, eq.5 gives eta falling as 1/K squared, not 1/K as the text says [deep reader's algebra, which I checked].
2. The patent's 1 V alternative (QE item 4). Para 0116: 167 MeV "equal with 26.7 pW that supposing is harvested at 10 mV and 2.67 nA or at 1 V and 26.7 pA" [read: p.33, confirmed on page image]. The "pW" is only right as one fission per second (26.7 pJ per fission). Both alternatives assume 100% conversion. At 1 V the bound e V / W is 1/30 = 3.3% at W = 30 eV; the charge needed is 26.7 pA per fission/s but ionization at 30 eV frees 5.6e6 electrons, i.e. 0.89 pA; shortfall about 30 times [calc, reproduced]. The verdict (strong form at the stated mV cell: Refuted by calculation) is not changed; the 1 V case is a lesser failure of the same kind, not a pass. Para 0116 refers to "FIG. 7", but the parallel connections are in FIGS. 11/12 [deep reader; probably a mis-reference]. Para 0087 makes an external circuit "mandatory" because otherwise the voltage passes breakdown and the circuit self-damages [read: p.31]: the patent itself treats insulator breakdown as the voltage ceiling, which is the Klassen trade-off in the inventor's words.
3. What the patent says about its own numbers (deep read, paras 0074-0200). No measured result anywhere [read: p.30-36]. 90% pushed "down to 80%" by insulators (0074); "towards 99%" with magnetic delta layers, no calculation (0096); 84% "simply the efficiency of ionization energy stopped in the 'C' layer", "upper conversion efficiency estimator" needing "experimental buildup" (0121); a 1 micron battery "having a 5% efficiency" (0187); the Pu-238 battery 20 mW is labelled a maximum and is 36% of the 0.56 W/g source power (0122; deep reader's calc, I checked 0.2/0.56). FIG. 10's arithmetic gives 137.5/7 = 19.6 micron and weights 5/7 = 71% (5/6 = 83%), not 84%; the text says Au is "5x the rest taken together" but the weights give 5 vs 2 [read: page image p.11]. FIG. 15 is a proposed beam test (present tense, no species, energy, facility or data) [read: p.34-35]. The 2013 application US20130121449A1 repeats the "84% ... upper limitation" (paras 0203-0204) and likewise reports no measurement [read: pp.45-48].
4. The number drifts with no new derivation. 80% ("may be", abstract, 2007), ">90%" by "quantum effects" (sibling paper 3021, one sentence; QE read it in full, no derivation or figure), 84% and 90%/80% (patent), 99% (patent para 0096), "as high as 95%" for the fusion battery and "20%-95%" on slide 23 (2012 roadmap [read: p.16, 23]). The inventor's own caution (84% is an upper estimator) is in the record next to the larger numbers. This answers open question 2.
5. The "25%" (Anufrienko et al. refs 13, 14). The deep read of the Klassen chapter shows the chapter itself describes these refs three ways: as "experimental and theoretical results" (p.265), as theory ("theoretical estimations show that about 80% ... see, for example ... [13,14, 17-20]", p.267, where the other references are Popa-Simil's patent and papers) and as "experimental studies ... revealed a conversion efficiency of 25 %" (p.281-282). It does not say the radiation type, load, voltage, source or the definition of the efficiency. So the "25%" is documented in one published second-hand sentence, by a group that cites Popa-Simil's 80% as a theoretical estimate. It is evidence, and I keep it in the record. It is not shown to be a measurement, and it does not qualify as a test result for the weak form until the primary papers (Russ. Chem. J. 2006; Nano- i mikrosistemnaya tekhnika 2008, no. 8) are read. Same for "3 orders" from thinning the insulator 200 um to 1 um (refs 49-51): relative only, gamma rays, no starting value, and the authors are the chapter's authors. The bibliographic oddity "Russian Chemical Journal, 2006, v. 1" may be an error [deep reader, memory medium].
6. Final for claim 1. Strong form (>80% at an mV cell): Refuted by calculation, unchanged. Weak form: Never properly tested for CIci. Others' reported numbers: 1e-5 % to 0.01 % (Yoshida, gamma-pumped, internally inconsistent in the chapter that reports it); 25% (second-hand, definition and measurement status unknown). Not dismissed without testing; never built and measured by the inventor.

Claim 2: Stepanov et al. 2018 restated

My first pass listed "up to 1.15 V in asymmetric supercapacitors under neutrons" as support. The deep read (text and Figs. 1-3) changes how that should be described [read: Stepanov deep read, pp.1-4]: - Radiation-induced charge separation was measured: 100 F symmetric cell about 1.24 mV (last plotted point about 1.33 mV) after 50 Gy; asymmetric cell about 0.1 V at the end of irradiation (Fig. 2) and 1.15 V much later (about 1.7e5 s; Fig. 3 shows decay from about 1.1 V). The "1.15 V during irradiation" in the abstract does not match Fig. 2. - No efficiency, power or output energy is defined or reported. The charge (about 5e-3 C) is computed from an earlier value of 1e-8 C/(Gy m2), not measured here. No control run for unirradiated drift, self-discharge or neutron heating is described; the authors say the equilibrium physics of supercapacitors "is not yet completely clear". The neutron source is unnamed. The interested parties are the concept's developers (Chernov, Stepanov) and the capacitor maker's staff. No link to Popa-Simil is stated in the paper. - My rough bound [calc, reproduced]: stored energy 0.5 x 5e-3 C x 1.15 V = 2.9 mJ; absorbed 50 Gy x 3.5 g = 0.175 J; ratio about 1.6%. This assumes the dose is deposited in the carbon only and that Q and U coincide in time (Figs. 2-3 suggest not), so it is a loose figure and the authors make no such claim. - Mechanism: a double-layer capacitor, not a nm insulator crossed by delta electrons. It supports "radiation separates charge in asymmetric layered structures", not the CIci shower mechanism. - Final for claim 2: label unchanged (Plausible, untested as a power source). "Effect Established at low power" is narrowed to "radiation-induced charge separation is Established (Moseley 1913 onward; the Russian reports); the CIci net-current picture is not measured". Contrarian's point that these groups cite Popa-Simil still stands, with the added fact that Stepanov et al. do not.

Claim 3: the author's own mechanism, and two meanings of "cold"

New material, p.3 of the paper, missed by digest, facts, all three reviews and my first pass [read: p.3, quoted by QE and the deep read]: "fast moving fission products are inducing in the insulator Uranium Dioxide ceramics electron showers of few micron range. The moving electrons interact with the lattice heating it. To prevent heating it is necessary to cut the electron showers circuits and drive them through conductive materials in circuits where the total resistance to be significantly smaller." The deep reader notes it is stated as a result of "the analysis" with no calculation, data or citation. - Weighing it (no reviewer did; mine). (a) The micron figure belongs to the fission fragments (range 5-10 um in UO2), not to electrons; delta electrons of up to about 2 keV go tens of nm [web: https://www.sciencedirect.com/science/article/abs/pii/S0022311510000292 and search summary, fragment range and picosecond electron-phonon transfer; the electron range figure is [memory, medium]]. It also conflicts with p.2 of the same paper, where the electron free path is "several nm" [read: p.2]. (b) Standard physics has the energy go from the hot electrons to the lattice by electron-phonon coupling within picoseconds along the track [web, same source]. That deposition happens whatever resistance the external circuit has. Joule heating in a low-resistance return path is a different, smaller term. (c) Energy conservation: lattice heat = deposited energy minus electrical energy delivered. "Preventing heating" by rerouting is therefore the same claim as claim 1, and it is limited by claim 1's bound. The mechanism as written (heating is a circuit-resistance effect of electron showers) is Internally inconsistent with the paper's own p.2 and not supported by accepted energy accounting. The paper's rationale stays in the record as the author's stated reason. - Two meanings of "cold". The 2007 paper both says the reactor would heat less (this mechanism) and (p.4) proposes cryogenic operation below Tc with a superconducting conductor. These pull in opposite directions: the second needs heat removed from a cold zone at an energy cost (claim 5). - "Energetic balance" sentence [read: p.3]: "the electric harvested energy to be smaller than the energy consumed to remove the thermal effects". As printed it reads as a net loss; it is probably a slip for "larger" (QE's and the deep reader's reading, not stated by the author). I keep it quoted both ways. - Final for claim 3: Not testable as stated; weak form (heat falls in proportion to the electrical fraction) Plausible, untested, and bounded by claim 1; the author's resistance-heating mechanism Internally inconsistent. Verdict headline unchanged.

Claim 5: the cryogenic energy balance

QE item 3. Operating below Tc (18.5 K for PuCoGa5, the lower-Tc compound lower still) needs a refrigerator. The Carnot minimum work per joule of heat lifted from T_c to 300 K is (300 - T_c) / T_c: 19 at 15 K, 15.7 at 18 K [calc, reproduced]. Assumptions, stated: the heat load is the part of the fission energy deposited inside the cold stack that is not turned into electricity; fission-product beta energy (about 4% of the 200 MeV [memory, high]) always lands there for structure types (a) and (b); fragments are 84% [memory, high]; everything else is deposited outside the cold zone; a perfect Carnot refrigerator, no heat leaks. Then: - Beta heat alone costs 0.63 to 0.76 of the total fission power in refrigerator work [calc]. - Unconverted fragment energy at 0% conversion costs 13.2 to 16.0 times the fission power [calc]. - Break-even needs an electrical efficiency eta on the fragment energy such that 0.84 eta = c x [0.84 (1 - eta) + 0.04], c the Carnot factor: eta = 98.5% at 18 K and 99.5% at 15 K [calc, reproduced]. Ignoring the beta heat entirely it is 94-95%. Real cryocoolers reach only a fraction of Carnot [memory, medium], which pushes it higher. - So the paper's own "energetic balance" condition is not met by an efficiency of 80%, or even the 84% upper estimator; it fails unless nearly all conversion is perfect. The cryogenic in-core superconducting harvester as a net power source: Refuted by calculation (with the assumptions above; if the heat is mostly deposited outside the cold zone the result is different, but that is not structure types (a) or (b)). The inventor's own sentence shows he knew a balance was required; the paper does not compute it. - Rh question closed: Table 1 has only PuCoGa5 (no Rh row) [read: p.3]. If the paper's second compound is PuRhGa5 (Tc 8.7 K matches), its Rh still falls under the paper's own "Rh = 144 barns makes it unusable" [read: p.4]. PuCoRh5 as printed is probably a naming slip; the paper cites Thompson 2006 for the pair [deep read]. - Table 1 arithmetic [deep read, python; I rechecked the sums]: the Pu-238 column prints 611.68, the sum is 608.93 (+2.75 b, one extra Ga); the Ga value is 2.75 b per atom in the table and 2.5 b in the text; the "absorption decrease" row is really the Pu share of the total (91, 95, 85, 27%): in PuCoGa5 the plutonium dominates the absorption, so the Pu is itself fuel/poison. The text says Ga absorbs "over ten times" more than O and C; real ratios are of order 1000 [memory, high]. Minor, but all inside the paper's own materials argument. - Verdicts otherwise as in the first pass: Tc facts Established; self-repair under fission flux Never properly tested; alpha self-irradiation lowers Tc (Tested, not supported for "no loss" under alpha) [web, first pass]. - Voxel size corrected (QE spot check 1): "mm3 to cm3, while the usage of superconductors may drive this volume up to dm3" [read: p.3]; earlier "up to 1000 dm3" in the digest and facts.json was wrong. The paper gives no derivation (deep read).

Claim 6 and web check

HFIR peak thermal flux is about 2.5e15 n/cm2/s in the flux trap [web: https://ornl.gov/sites/default/files/HFIR%20Main-R1.pdf and search summary]. The 5.6e15 needed for a sustained 5000 kW/cm3 in pure U-235 metal [calc, first pass] is therefore about 2.2 times the peak of the highest-flux research reactor, which closes the QE's process note (previously [memory, medium]). Verdict unchanged. The 2012 roadmap gives a different chain for a fusion battery (1 W = 3e11 reactions/s; "up to about 1 kW/cm3 may be removed", "maximum power density of about 20 kW/cm3", "may reach 1 MW/cc") [read: p.6-7, 3], none of it the 5000 kW/cm3 figure.

New or restored material kept in the record

• The inventor's confirmation claim and his own history. The Physics Procedia 2011 abstract says "some of the ideas being confirmed by collateral experiments" [web: https://www.sciencedirect.com/science/article/pii/S1875389211005943, via Investigator note; full paper not read]. The 2012 roadmap's section "Expected collateral research results and applications" (p.15) lists spin-offs, not experiments, and the deep reader found no CIci test, prototype or beam test in 81 pages [read: roadmap deep read]. The phrase in 2011 is therefore unresolved: the experiments are unidentified, and I make no claim that they exist or not. The roadmap does state two things about the author's own record: slide 48, DARPA 2004 "Liquid Electronics Power Source ... Result = Negative; Issues - low efficiency - low stability", a "liquid electronics effect accidentally observed by my team in 1987" with drastic efficiency drop [read: p.48]; and slide 70, "1984-1995 worked in direct nuclear energy conversion special program - Coated Capacitor structures (gas, Vacuum filled); Solid structures (Se, Al, Au, U)" with, on slide 69, "I have never published because of lack of coherent data" [read: pp.69-70]. These are the inventor's characterizations, single-source, and are evidence both of earlier experimental work and of non-publication.
• Romanian lineage. The roadmap cites Ursu & Popa-Simil, "Synergetics of the fission electric cells" (Int. J. Energy Res. 1980) and a 1978 Institute for Physics and Nuclear Energy cell paper, with Ursu, Badescu-Singureanu, Vamanu and Schachter [read: p.18, refs 51-53]; the Investigator's earlier note puts Popa-Simil at the Institute of Atomic Physics, Magurele (ScholarGPS, [web per research note]). The author's coated-capacitor programme of 1984-95 and these papers are the same lineage as the fission electric cells (Ursu 1978-88). That is a lead, not an established continuity, and the primary papers have not been read.
• Three structure types (a) fissile outside the stack, (b) fissionable mixed in, (c) alpha tiles with no fissile material (closest to existing alpha batteries); the self-assembly route (binary nanoparticle superlattices, Talapin 2006); history refs (Willson 1955, Polansky NERI 1999, DARPA LEAPS) [read: p.1, 3].
• Lesser items from the first pass and QE item 5 (Cleantech 2007 "breakdown voltage of few milivolts", TechConnect 2015 and Granthaalayah 2021 restatements, US Army 2007 Natick items, forward patent citations US10699820B2, WO2011011504A1, FR3018946A1, CN106340684A, JP2022551664A) stay in the record as uptake and leads; forward citations show interest, not proof.
• US20130121449A1. The deep read finds no measurement of any CIci stack or of the Pd-D reaction; the chamber and test circuit (FIG. 20, 22A/B) are procedures in the future tense; the entanglement/non-locality control mechanism is verbal, with one equation I(x,t) = I0 exp(-lambda x) sin(omega t + phi) whose lambda, omega and phi are not defined and which is not derived from the Schroedinger equation [read: p.37]. Verdict unchanged: Not testable as stated; the roadmap gives a long list of cited "new physics" support (deep read section 4b), which I keep as the inventor's claimed support and have not weighed.

Absence-of-evidence analysis: amended

• Explanation 2 (experiments done but unpublished or negative) gains direct statements from the inventor himself: he reports earlier direct-conversion work (1984-95), a negative result in the liquid-electronics test, and "never published because of lack of coherent data" (slide 70) [read: roadmap p.48, 69-70]. These are his own words, single source, not corroborated by a second party; they raise the weight on explanation 2 but do not give data on any CIci stack.
• Explanation 4: the Investigator recorded a Google notice on the Russian-language query ("removed 17 result(s)", legal removal) [read: research note, QE item 5]. Such notices are routine for many topics and the cause here is unknown; it is noted, and it neither supports nor excludes suppression. Items 1, 3 and 5 stand. The decisive records are unchanged (USPTO file wrappers, New Mexico filings for LAVM LLC, SBIR/DTIC awards, beam-time logs, the unread Russian primaries).

Open questions: updates

1. (25% of Anufrienko): still open and now also unanswered whether it was measured; see Claim 1 item 5.
2. Answered (nothing behind the 90%).
3. "PuCoRh5": likely PuRhGa5; Table 1 has only PuCoGa5; closed as a naming slip with the Rh point applying either way.
4. New: could a design at volt-to-kV bias with a thick or vacuum insulator (Klassen's own proposal) reach a useful efficiency, and what energy does the cryogenic stack cost?
5. New: which "collateral experiments" did the 2011 abstract mean?

Corrected claim wording (claims.md)

Claim 1: "The abstract says efficiency 'may be higher than 80%'" (a possibility; not derived). Claim 3: add "the author's stated mechanism is electron-shower currents heating the resistive UO2, remedied by routing them through low-resistance conductors". Claim 5: add "in a cryogenic reactor".

Not yet read (unchanged from the first pass)

Anufrienko 2006/2008, Yershov 2011-2012, TMS 2008, Cleantech 2007, Physics Procedia 2011, the USPTO file wrappers. Conclusions that rest on them stay second-hand.

Web checks in this revision: HFIR flux (done, ORNL), fission-fragment range and electron-phonon transfer (done, ScienceDirect). Total 5 of 8 used.