{"id":"e76cc333-226b-4983-b000-b97f00224fe4","arxiv_id":"1908.08166","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"Negative muons stopped in natural molybdenum produce radioactive isotopes such as 99Mo/99mTc at rates that the authors estimate could reach tens of millions per second with existing beams.","lead":"This paper shows that ordinary muon capture on a natural molybdenum target produces radioactive niobium isotopes, including the medically relevant chain that yields 99Mo and 99mTc. The authors use a simplified nuclear emission model to estimate production yields and argue that future high-intensity muon beams could make this an efficient isotope production method.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Table 6 RI yields appear to exceed the stated muon budget: the 99Mo yield is larger than the total muons captured on the 100Mo component if the table is for plates A+B, so the measured production-rate claim needs renormalization.","rationale":"The reader's weakest assumption is the transferability of the PNEM parameters to new nuclei and proton-emission channels. That is a real concern for the calculated branching ratios and for extrapolated yields in Tables 2–5 and for the modelled entries in Table 6. However, the most load-bearing problem for the central claim is more immediate: the measured natMo yields in Table 6 appear internally inconsistent with the stated muon stopping numbers. If the table is for plates A+B, the 99Mo count alone exceeds the physical maximum set by the 100Mo abundance, and the summed observed yields leave almost no room for unobserved channels; if it is for all four plates, the text's consistency statement comparing the sum to A+B muons is incorrect. Either way, the absolute normalization underlying the '0.5–0.1 per muon' and 'N_RI≈4e7/s' statements is not yet established. This is a checkable experimental issue, not a matter of theoretical preference, and it should be settled before the efficiency claim is used for future applications. I do not think the paper should be rejected outright: the gamma-ray identifications are a genuine experimental contribution, and the PNEM comparison in Tables 1–5 is useful even if model-dependent. But acceptance should remain conditional on a corrected normalization analysis and, ideally, release of the reduced spectra. Since the reader already recommended CONDITIONAL, I leave the verdict unchanged.","tokens_in":12187,"tokens_out":9773,"duration_ms":95464,"concrete_test":"Re-extract the natMo yields from the raw GM1/GM2 spectra with an absolute efficiency calibration, separating plates A+B from C+D, and compare the summed N(X') for each plate pair to the muons stopped in that pair as obtained from the muonic X-ray spectrum. In particular, test whether N(99Mo) divided by (0.096 × stopped muons in the relevant plates) exceeds unity; if it does, the efficiency or muon-stop normalization is off by at least that factor, and the claimed rate per muon must be revised downward.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The quantitative central claim rests on the natMo measurement summarized in Table 6. The text says the sum of observed RIs is about 3.1e9 while 3.5e9 muons stopped in plates A and B, and that essentially all muons were captured. Taken literally, the identified channels already exhaust about 90% of the muon budget, leaving almost nothing for unobserved stable or long-lived residues; adding the modelled entries for 94Nb, 93Nb and 91Nb (footnote b) gives about 5.1e9, which exceeds the stated muon count. A sharper check: with natural Mo at 9.6% 100Mo, plates A+B contain at most 3.4e8 muons captured on 100Mo, so a 100% branch to 99Mo would give 3.4e8 atoms, yet Table 6 lists (3.8±0.4)e8 for 99Mo via 100Mo(µ,nβ−). If the table instead refers to all four plates (7e9 muons), the 99Mo yield becomes plausible, but then the text's comparison to the 3.5e9 A+B muon count is wrong and the per-plate normalization is not established. This is not a model-extrapolation concern; it affects the measured '0.5–0.1 per muon' and N_RI≈4e7/s numbers directly. The isotope identifications are credible, but the absolute normalization needs to be reconciled before the headline efficiency can be accepted.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports the production of radioactive isotopes by ordinary negative muon capture (MuCIP) on a natural molybdenum target at the RCNP MuSIC beamline. Delayed gamma-ray spectroscopy identifies a number of residual nuclei, including Nb, Zr, and Y isotopes, and the medically relevant 99Mo/99mTc. The measured isotope yields are compared with a pre-equilibrium/equilibrium proton-neutron emission model (PNEM). The paper claims that the production rate per stopped muon is as high as 0.5–0.1, that essentially all stopped muons produce isotopes, and that with a full MuSIC beam one can achieve about 4×10^7 RIs per second for a 1 μA proton beam.","tokens_in":12536,"tokens_out":12319,"duration_ms":104477,"significance":"If the quantitative claims survive scrutiny, the paper demonstrates a new and efficient route to radioactive isotope production, complementary to photon and neutron capture, and provides experimental data on muon-capture residual nuclei that are useful for nuclear structure and applications. The isotope identification from gamma-ray spectra appears credible, and the measured yield patterns for 100Mo, 127I, and 209Bi are broadly consistent with earlier work. However, the central efficiency numbers are tied to a normalization that is internally inconsistent, and the model that supplies several of the calculated yields is not derived in the manuscript. The potential contribution is real, but the current presentation does not yet support the headline production rates.","major_comments":[{"comment":"The normalization of the measured yields is inconsistent with the stated muon budget. The text says that the sum of observed RIs is about 3.1×10^9 while the number of muons stopped in plates A and B was 3.5×10^9, but the delayed-gamma measurement was performed on all four plates and the total muon stop was about 7×10^9. If the yields in Table 6 refer to all four plates, then the measured sum is only about 44% of the muon budget, which contradicts the statement that 'all muons were mainly captured into the Mo nuclei to produce Nb and other isotopes.' If, instead, the yields refer to plates A and B, the 99Mo yield of 3.8×10^8 exceeds the maximum possible number of muons captured on 100Mo, which is 0.096×3.5×10^9 = 3.4×10^8; this is impossible. This inconsistency directly affects the headline numbers: the production rate per muon and N_RI ≈ 4×10^7 s^-1. The authors must clarify which normalization applies and recompute the derived efficiencies accordingly.","section":"Section 4, Table 6"},{"comment":"The PNEM model is not derived in this manuscript. The paper defines the muon capture strength B(μ,E) as a sum of two Lorentzian resonances, but it does not show how B(μ,E) is converted into the branching ratios Br(X') that are tabulated in Tables 1–5 and used for the model entries in Table 6. The parameters (PEQ fraction p = 25%, temperature ratio b ≈ 3, resonance energies E_G1 and E_G2, widths, and relative strengths) are adopted from previous work, and no sensitivity analysis is provided. Since the model entries for 94Nb, 93Nb, and 91Nb in Table 6 are used in the text to argue that essentially all stopped muons produce RIs, the reader cannot evaluate the reliability of those extrapolations. Please present the model equations connecting B(μ,E) to the branching ratios, or explicitly state that these yields are taken from refs. [18,19] and justify their application to the new targets.","section":"Section 3, Eqs. (4)–(5)"},{"comment":"The claim 'The production rate per one μ is as high as 0.5−0.1' is ambiguous and not clearly derived from the data. If it means 0.5 to 0.1 per muon, the abstract's statement of '10^{9-10} per second' is not supported by the demonstrated MuSIC flux of about 4×10^7 muons/s and appears to refer only to hypothetical future beams. The numerical efficiency is also directly tied to the normalization inconsistency described above. The authors should state precisely which quantities are measured, which are modeled, how the per-muon efficiency is defined, and how the extrapolation to 10^9–10^10 RIs/s is justified.","section":"Section 5"}],"minor_comments":[{"comment":"The half-life listed for 93Nb (1.41×10^5 hr) is inconsistent with 93Nb being stable, as the text itself states in Section 4; please correct this entry.","section":"Table 6"},{"comment":"The factor k(Tμ) is not defined; please give its formula, e.g., k(T) = (1 - exp(-λT))/λ, so that the decay correction during irradiation is transparent.","section":"Eq. (2)"},{"comment":"The text refers to 'Table 5' when discussing the natural-Mo results (e.g., 'Figure 6 illustrates ... based on Table 5'); this should be Table 6.","section":"Section 4"},{"comment":"The sentence 'the 66 hr 99Nb' should read 'the 66 hr 99Mo', since 99Nb is not the 66-hour isotope; the text describes the decay product of 99Mo feeding 99mTc.","section":"Section 4"},{"comment":"The formulas for the resonance energies should be typeset as E_G1 = 30 A^{-1/5} MeV and E_G2 = 75 A^{-1/5} MeV to avoid misreading as (A-1)/5.","section":"Section 3"},{"comment":"The activity estimate of '100 GBq' for 10^10 muons/s does not follow immediately from the stated production rate; please show the calculation or qualify the estimate as a rough order-of-magnitude figure.","section":"Section 5"}],"recommendation":"major_revision","confidential_remarks":"The experimental data are a valuable contribution, but the manuscript has a serious normalization inconsistency between Table 6 and the muon budget that must be resolved before the efficiency and rate claims can be accepted. The model section is underdeveloped; it should either be expanded to a self-contained derivation or clearly marked as relying on prior work with an explicit justification of parameter transferability. The abstract's statement of 10^{9-10} RIs per second is not supported by the current measurements and should be moderated or explicitly framed as a projection for future facilities."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the qualitative claim holds up, the quantitative one does not as printed. The experiment really does produce the Nb chain and 99Mo/99mTc from natural Mo by muon capture, and the gamma-ray identifications are credible. But the normalization in Table 6 is internally inconsistent, and the per-muon rates in the abstract and Section 5 inherit that problem.\n\nWhat is actually new: the proton–neutron extension of the group's neutron-emission model is a small step, but the feasibility run on natural Mo is a concrete demonstration of MuCIP producing the 99Mo/99mTc medical pair, which is the first clear application-oriented result from this program. The model is checked against residual mass distributions for five nuclei (100Mo, 107,108Pd, 127I, 209Bi) and the agreement is decent. The authors are also honest about which Table 6 entries are calculated rather than measured.\n\nThe load-bearing problem: the stress-test note is correct. Table 6 lists N(99Mo) = (3.8±0.4)×10^8, attributed to 100Mo(µ,nβ−). At 9.6% abundance, the 3.5×10^9 muons the text assigns to plates A and B can produce at most 3.4×10^8 atoms of the mass-99 chain even at 100% efficiency. The measured 99Mo yield exceeds the entire 100Mo budget. If Table 6 instead covers all four plates (7×10^9 muons), the 99Mo number becomes plausible, but then the text's pairing of 3.1×10^9 observed RIs with the 3.5×10^9 A+B muon count is wrong, and the near-unity per-muon efficiency claim is unsupported. You cannot have both statements. This directly affects the \"0.5–0.1 per muon\" and 4×10^7 s^-1 claims.\n\nLesser soft spots: the branching-ratio derivation from B(µ,E) to the tabulated percentages is deferred to a thesis and an earlier PRC, and the resonance widths and relative strengths are never given, so the model entries are not reproducible from this paper alone. There are also signs of haste: \"66 hr 99Nb\" should be 99Mo; the same table is called Table 1, Table 5, and Table 6 in one section; and the process column for 94Nb (98Mo(µ,1n)) is wrong on its face. Self-citation is heavy but expected given this is a continuing program.\n\nBottom line: send it to a serious referee. The experiment is real and the application is worth pursuing, but the authors need to renormalize Table 6, state explicitly which plates each yield refers to, and put the model parameters on the page. If the normalization is fixed, the production claim may still be strong—it just is not at the level the current text states.","headline":"A real feasibility measurement with credible isotope identifications, but the central normalization in Table 6 is internally inconsistent the printed 99Mo yield exceeds the 100Mo muon budget so the headline per-muon rates need rework before they can be used.","tokens_in":13128,"tokens_out":10968,"would_cite":true,"duration_ms":98043,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["25.30.Mr"],"model":"deepseek-v4-flash","headline":"Ordinary muon capture can turn a stopped muon into a radioactive isotope with near-unity probability, yielding per-muon production rates two orders of magnitude above photon capture, and the paper demonstrates the route on natural…","keywords":["ordinary muon capture","muon capture isotope production","radioactive isotope production","medical isotope","technetium-99m","molybdenum-100","pre-equilibrium emission model","MuSIC muon beam"],"falsifier":"An absolute activation experiment on enriched $^{100}$Mo---counting stopped muons via muonic X-rays and then measuring the activities of $^{99}$Mo and $^{99m}$Tc---would settle the efficiency claim: if the measured number of radioactive atoms per stopped muon falls far below the paper's 0.5--0.1 range, the central claim fails.","tokens_in":11997,"feed_emoji":"⚛️","tokens_out":14559,"duration_ms":126339,"temperature":0.7,"pith_summary":"Ordinary muon capture (OMC) is proposed as an efficient route for producing radioactive isotopes: a stopped negative muon is captured by a medium-heavy nucleus with more than 90% probability, and the excited $Z-1$ nucleus then sheds neutrons, and occasionally a proton, to leave a radioactive residue. The paper claims that on natural molybdenum at the MuSIC beamline, the number of radioactive isotopes produced was close to the number of stopped muons, and that the production rate per muon is 0.5--0.1, about two orders of magnitude higher than photon-capture rates. If true, this makes muon beams a practical complement to reactor and photon sources, including for medical isotopes such as $^{99}$Mo and $^{99m}$Tc. The authors support the yields with a pre-equilibrium/equilibrium proton-neutron emission model that reproduces measured isotope mass distributions for $^{100}$Mo, $^{107}$Pd, $^{108}$Pd, $^{127}$I, and $^{209}$Bi. Why it matters: muon capture shifts the atomic number by one, so MuCIP reaches isotope chains that neutron- and photon-induced reactions do not.","feed_headline":"Stopped muons make one radioactive isotope per capture","feed_subtitle":"Per particle, muon capture beats photon beams by two orders of magnitude, and can make medical 99mTc from molybdenum.","key_machinery":"The load-bearing object is the proton-neutron emission model (PNEM), an extension of the earlier neutron emission model. The muon-capture strength $B(\\mu,E)$ is written as the sum of two Breit--Wigner giant resonances, $B_1$ and $B_2$, with energies $E_{G1}=30A^{-1/5}$ MeV and $E_{G2}=75A^{-1/5}$ MeV; the first-neutron spectrum is $S(E_{n(1)})=k[E_n\\exp(-E_n/T_{\\rm EQ})+p E_n\\exp(-E_n/T_{\\rm PEQ})]$ with the pre-equilibrium fraction $p=25\\%$, $T_{\\rm EQ}=\\sqrt{E_{\\rm ex}/a}$ with $a=A/8$, and $T_{\\rm PEQ}\\approx 3T_{\\rm EQ}$. After the first neutron, the cascade continues as equilibrium evaporation until the residue becomes neutron-bound and decays by gamma emission; proton emission is included only in the energy window where the excitation lies below the neutron separation energy but above the proton separation energy, keeping proton channels at a few percent in medium-heavy nuclei. This machinery turns the resonance strength into branching ratios to each residual isotope, and those branching ratios convert a muon flux into predicted isotope numbers and rates.","core_discovery":"The paper's central claim is that ordinary (non-radiative) muon capture can serve as a high-yield isotope-production reaction. When a negative muon stops in a nucleus with $Z\\ge 20$, capture occurs with more than 90% probability; the muon deposits about 100 MeV of excitation while the emitted neutrino carries off most of that energy, leaving the $Z-1$ daughter excited at roughly 5--50 MeV. De-excitation proceeds mostly by emission of one neutron (about 50--60% of events), with two or more neutrons and a few-percent proton branch, so the final residues are radioactive isotopes of the form $^{A-x}_{Z-1}X$. The measurement on natural molybdenum with the MuSIC beam provides the demonstration: after irradiation with about $3.5\\times 10^9$ stopped muons, delayed gamma rays from many molybdenum-capture products were observed, and the summed radioactive-isotope yield of about $3.1\\times 10^9$ approximates the number of stopped muons. The paper consequently states that the per-muon production rate is 0.5--0.1, two orders of magnitude above photon-capture isotope production, and estimates $N_{\\rm RI}\\approx 4\\times 10^7$ per second for a 1 $\\mu$A proton beam.","pith_inferences":["If near-one-isotope-per-muon holds on all medium-heavy targets, the economics of MuCIP are set by the cost of muon beam time rather than by target material or separation chemistry; the paper does not compare costs with the reactor- and accelerator-based $^{99}$Mo/$^{99m}$Tc supply chain.","The least-tested ingredient is the proton-emission branch, so a dedicated measurement of $^{97}$Zr from $^{98}$Mo$(\\mu,p)$ or $^{89}$Zr from $^{92}$Mo would probe whether the model's few-percent proton channels hold as $Z$ changes.","Because muon capture produces neutron-rich $Z-1$ isotopes, MuCIP could double as a tool for producing exotic neutron-rich nuclei for decay studies, an application the paper only gestures at through its isotope list.","The near-unity yield also suggests MuCIP as a transmutation route for long-lived fission products, converting them into shorter-lived or stable species; the paper mentions transmutation only in passing and gives no quantitative estimate."],"forward_implications":["A 1 $\\mu$A proton beam feeding a MuSIC-like channel is claimed to produce about $4\\times 10^7$ isotopes per second, and an upgraded 10 $\\mu$A beam would scale that to roughly $4\\times 10^8$ per second, so achievable muon intensities translate directly into useful isotope batches.","Because muon capture lowers the atomic number by one, MuCIP produces isotope chains that photon and neutron capture cannot reach, for example $^{100}\\mathrm{Mo}\\to{}^{99}\\mathrm{Nb}\\to{}^{99m}\\mathrm{Tc}/{}^{99}\\mathrm{Mo}$, so medical technetium could be made from natural molybdenum.","The high per-muon yield of 0.5--0.1 implies that the practical bottleneck for MuCIP is muon beam intensity rather than nuclear cross-section, so any future high-intensity muon source becomes a potential isotope-production facility.","Thin targets of order tens of mg/cm$^2$ suffice because negative muons stop by atomic capture, giving high specific activity and allowing the muon momentum to set the depth at which isotopes are deposited.","The model's agreement with measured distributions on five nuclei suggests the one-neutron-dominant branching pattern is general, so isotope yields for other target nuclei can be estimated before any irradiation."],"supporting_citations":[{"why":"Supplies the spectrum formula and the level-density parameter $a=A/8$ on which the emission model is built.","marker":"[16]"},{"why":"Fixes the pre-equilibrium fraction $p=25\\%$ that sets the ratio between the two components of the neutron spectrum.","marker":"[17]"},{"why":"Carries the earlier neutron emission model and the two-giant-resonance form of the muon-capture strength $B(\\mu,E)$ used in the PNEM.","marker":"[18]"},{"why":"Provides the measured $^{100}$Mo$(\\mu,xn)$ branching ratios used to validate the model and to quote the short-lived $^{100}$Nb yield.","marker":"[19]"},{"why":"Describes the MuSIC high-intensity muon beamline that makes the stated muon fluxes and rate estimates realistic.","marker":"[15]"},{"why":"Supplies the delayed-gamma detection method and the characteristic gamma-ray lines used to identify the produced isotopes.","marker":"[7]"},{"why":"Provides measured muon-capture isotope distributions for $^{127}$I and $^{209}$Bi that are used to validate the model.","marker":"[13]"},{"why":"Documents the neutron and charged-particle emission branches following muon capture that motivate the model's channel structure.","marker":"[8]"},{"why":"Gives the resonant photonuclear isotope-transmutation rates used as the baseline when claiming muon capture is two orders of magnitude more efficient per particle.","marker":"[1]"}],"fun_headline_variants":["Muon capture makes one isotope per stopped muon","Muon isotope yield beats photon beams by 100x","Muon capture outperforms photons for isotope yield","Muon capture turns molybdenum into medical isotopes"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The predictions rest on a statistical model whose parameters were fixed using earlier neutron-emission data; if those parameters do not also describe proton emission and new target nuclei, the calculated yields and rates lose their support.","fun_headline_variants_meta":{"raw":{"variants":["Muon capture makes one isotope per stopped muon","Muon isotope yield beats photon beams by 100x","Muon capture outperforms photons for isotope yield","Muon capture turns molybdenum into medical isotopes"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001281,"raw_usage":{"total_tokens":5273,"prompt_tokens":1021,"completion_tokens":4252,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":637,"completion_tokens_details":{"reasoning_tokens":4190}},"tokens_in":637,"tokens_out":4252,"duration_ms":30064,"temperature":1.0,"reasoning_tokens":4190,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:47:55.709559+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"An absolute activation experiment on enriched $^{100}$Mo---counting stopped muons via muonic X-rays and then measuring the activities of $^{99}$Mo and $^{99m}$Tc---would settle the efficiency claim: if the measured number of radioactive atoms per stopped muon falls far below the paper's 0.5--0.1 range, the central claim fails.","supporting_citations":[{"cited_title":"Ejiri, M","cited_arxiv_id":null,"evidence_quote":"Supplies the spectrum formula and the level-density parameter $a=A/8$ on which the emission model is built."},{"cited_title":"Hashim, H","cited_arxiv_id":null,"evidence_quote":"Fixes the pre-equilibrium fraction $p=25\\%$ that sets the ratio between the two components of the neutron spectrum."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Carries the earlier neutron emission model and the two-giant-resonance form of the muon-capture strength $B(\\mu,E)$ used in the PNEM."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Describes the MuSIC high-intensity muon beamline that makes the stated muon fluxes and rate estimates realistic."},{"cited_title":"Ejiri, I","cited_arxiv_id":null,"evidence_quote":"Supplies the delayed-gamma detection method and the characteristic gamma-ray lines used to identify the produced isotopes."},{"cited_title":"Measday, The nuclear physics of muon capture, Phys","cited_arxiv_id":null,"evidence_quote":"Documents the neutron and charged-particle emission branches following muon capture that motivate the model's channel structure."},{"cited_title":"Ejiri, T","cited_arxiv_id":null,"evidence_quote":"Gives the resonant photonuclear isotope-transmutation rates used as the baseline when claiming muon capture is two orders of magnitude more efficient per particle."}],"review_version":1}