{"id":"0fa40807-043f-4750-a581-cdea8b43f31b","arxiv_id":"2509.00041","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Laser spectroscopy of thorium-229 in calcium fluoride reveals four distinct dopant sites with characteristic electric field gradients, and identifies the atomic structures of the two dominant sites.","lead":"Scientists used a vacuum ultraviolet laser to probe thorium-229 nuclei inside calcium fluoride crystals and found that the thorium atoms sit in four different types of spots, each with a different surrounding electric field. The work provides key data for designing solid-state nuclear clocks and demonstrates a new optical form of Mössbauer spectroscopy.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The four-site inventory is not fully closed: the X2 fit is acknowledged as possibly missing a fifth site, and the free per-crystal Qe/Qg fit (0.563 for X2 vs 0.57003(1) in ref. [10]) indicates model misspecification that could bias the extracted weak-site EFGs.","rationale":"The reader's conditional verdict is well supported, and the stress test converges on the same load-bearing assumption: the four-site decomposition of Eq. (1) is the foundation for the claim of four distinct doping sites, but the model is exactly where the evidence is weakest. The X2 spectrum, the only one with enough statistics to constrain the weak sites, is the one the paper itself flags as poorly described by four sites. The free Qe/Qg drift to 0.563 in X2 provides an internal diagnostic of this misspecification: since the quadrupole energies depend on the product QVzz, a missing component or an incorrect line shape can be partially absorbed into both Qe/Qg and the site EFGs. The small fractional contributions of sites 3 and 4 make their extracted Vzz and eta particularly sensitive to unmodeled overlapping peaks, so the 'four sites with characteristic EFGs' claim is not fully closed even though sites 1 and 2 are on firmer ground. The proposed test directly checks whether the weak-site EFGs survive a more complete model. Because this concern does not overturn the identification of the two dominant sites and the reader already assigned a conditional verdict, the appropriate recommendation is to keep the verdict unchanged.","tokens_in":13785,"tokens_out":6021,"duration_ms":63997,"concrete_test":"Refit the X2, C10, and C13 spectra with Qe/Qg fixed to the independently measured value 0.57003(1) from ref. [10], and repeat the fit adding a fifth Gaussian site (or allowing per-site linewidths), comparing models by a likelihood-ratio test or BIC. If the extracted Vzz/eta for sites 3 or 4 shift by more than the spread in Table 3, or if systematic residuals remain, the four-site description should be treated as provisional rather than established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim is that the observed spectra are explained by four distinct doping sites. This rests on the global fit in Eq. (1), which assumes exactly four sites, a single Gaussian linewidth sigma for all lines and sites, transition intensities r_i fixed by the quadrupole Hamiltonian, and a Qe/Qg ratio that is left free per crystal. The weakest point is the tension between this model and the data at the highest concentration. Fitted Qe/Qg values are 0.574 (C10), 0.570 (C13), and 0.563 (X2), while ref. [10] determined Qe/Qg = 0.57003(1); the X2 value is the kind of offset expected when the model is missing spectral components. This is not speculation: Appendix C states that for X2 'the agreement between the data and the fitted spectrum is lower than for the other two crystals, which might indicate the presence of a fifth or further doping sites.' Because Vzz enters the level splitting together with Qe/Qg (Eqs. 2-6), a missing site or linewidth mismatch will bias the extracted Vzz and eta. The bias matters most for sites 3 and 4, whose relative contributions are only 0.4-3.6%; their Vzz values (-320 and 260 V/A^2) and eta values are determined from weak, partially overlapping lines. Table 1 also shows poor run-to-run reproducibility for the X2 site-4 fraction (3.0(12)), consistent with an incompletely specified model. The identification of sites 1 and 2 is more robust because site 2 is independently measured in ref. [10] and site 1 is an unresolved central line with near-zero EFG; the four-site inventory as a closed set is the claim that is not yet fully supported.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports vacuum-ultraviolet laser excitation of the 229Th nuclear clock transition in three 229Th:CaF2 crystals with different dopant concentrations. The authors observe many spectral lines and fit them with a model that assumes four distinct thorium doping sites, each characterized by an electric field gradient (Vzz, eta), a common Gaussian linewidth, and a per-crystal ratio of nuclear quadrupole moments Qe/Qg. They extract Vzz values of approximately 0, 106-107, -319 to -320, and 258-261 V/Å^2 for sites 1-4, with eta about 0.6 for site 2, and they report site-selective isomeric lifetimes of about 630 s together with site-dependent laser-induced quenching rates. Combining DFT calculations and STEM imaging, the authors assign the dominant site 1 to a highly symmetric ThCa defect with distant charge compensation and site 2 to a nearest-neighbor pair of thorium atoms without local compensating defects.","tokens_in":14174,"tokens_out":3835,"duration_ms":40405,"significance":"If the four-site interpretation is correct, the work resolves previously unassigned resonances in 229Th:CaF2, provides the first multi-site EFG inventory for this system, and demonstrates a laser-based Mössbauer approach with site selectivity. A notable strength is that the DFT calculation for site 2 predicts Vzz = 95 V/Å^2 and eta = 0.6 without fitting to the measured EFG, in reasonable agreement with the experimental values Vzz = 106-107 V/Å^2 and eta = 0.59-0.61, and the site-2 EFG is consistently reproduced across three crystals and agrees with refs. [7,10]. The lifetime and quenching measurements add practical information for solid-state nuclear clock design. The main risk is that the completeness of the four-site inventory is not fully closed, as discussed below.","major_comments":[{"comment":"The central claim that the spectra are explained by exactly four doping sites is weakened by the acknowledged poorer fit for the highest-concentration crystal X2. Appendix C states that the agreement between data and fit is lower for X2 and 'might indicate the presence of a fifth or further doping sites.' Because the fit function in Eq. (1) sums over exactly four sites, any fifth site will bias the extracted Vzz and eta, especially for the weak sites 3 and 4 whose relative contributions are only 0.4-3.6% (Table 1). The authors should either include a fifth site in the X2 fit, provide a quantitative upper limit on the amplitude of a fifth site, or explicitly restrict the 'four sites' claim to the two dominant sites.","section":"§3 and Appendix C"},{"comment":"Qe/Qg is left as a free parameter per crystal and the fitted values are 0.574 (C10), 0.570 (C13), and 0.563 (X2), while ref. [10] determined Qe/Qg = 0.57003(1). Since Qe/Qg is a nuclear property, the 0.563 value for X2 is a red flag for model misspecification, and because the level splittings scale with Qe/Qg * Vzz (Eqs. 2-6), a missing spectral component or an incorrect linewidth model can bias the extracted EFGs. The authors should test whether fixing Qe/Qg to the precise literature value changes the site 3 and 4 parameters, and should report the resulting systematic shifts.","section":"§3.1, Table 3, Eqs. (2)-(6)"},{"comment":"The structural assignment of site 1 to a ThCa defect with distant charge compensation is supported only indirectly: Table 2 shows that moving a calcium vacancy farther away drives Vzz toward zero, but the total energy increases with distance, and the manuscript explicitly states that no explicit charge-compensation mechanism is postulated. Since the abstract claims identification of the microscopic structure of the two dominant configurations, this assignment should be presented more cautiously, or additional evidence (e.g., concentration-dependent or annealing-dependent data) should be provided to support the invoked kinetic trapping mechanism.","section":"§3.3 and Table 2"}],"minor_comments":[{"comment":"Typo: 'dominanted' should be 'dominated'.","section":"Appendix C"},{"comment":"The phrase 'where no magnetic field is not applied' contains a double negative; it should read 'where no magnetic field is applied'.","section":"Appendix A"},{"comment":"The statement that values in parentheses are 'differences of fitting results of the two spectra' is ambiguous; please clarify whether these are half-differences, standard deviations, or full spreads, and note that they are not statistical uncertainties.","section":"Table 3 caption"},{"comment":"The sentence 'For η = 0, the transition between m = ±5/2 → m' = ±1/2 is asymptotically forbidden' would be clearer as 'has zero intensity at η = 0'.","section":"§3.3"},{"comment":"In the DFT methods paragraph, 'enables efficient simulation evaluation at the Γ-point only in reciprocal space' should be rephrased, for example as 'enables efficient evaluation using only the Γ-point in reciprocal space'.","section":"Appendix B"}],"recommendation":"major_revision","confidential_remarks":"The X2 fifth-site caveat is the main technical weakness. It is fixable within the scope of the manuscript by reanalyzing the X2 spectrum with Qe/Qg fixed to the literature value, by adding a fifth site in the fit, or by reporting an upper limit on the fifth-site amplitude and showing the robustness of the sites 3 and 4 EFGs. I would be happy to reconsider after such a revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThis is the paper that finally sorts out the mess of unassigned lines in 229Th:CaF2. The main new result is a four-site inventory with distinct EFGs, extracted from consistent fits across three crystals. Site 2 matches the earlier precise Vzz and eta from refs. [7,10], which is a good check. Site 1 is a near-zero EFG central line and the paper gives a plausible structural assignment to isolated ThCa with distant charge compensation. The DFT prediction for the site-2 cluster (ThCa-4e) gives Vzz=95 V/Ų and eta=0.6, matching the measured 106–107 and 0.59–0.61 without fitting those numbers. That is a genuine, parameter-free consistency argument. The site-selective lifetime and quenching measurements are also useful and directly relevant for solid-state nuclear clock design.\n\nSoft spots, in proportion. The weakest point is the completeness of the four-site model for the highest-concentration crystal X2. The paper itself admits in Appendix C that the fit quality is worse there and that a fifth site might be present. The fitted Qe/Qg for X2 is 0.563, which is a noticeable offset from the accepted 0.57003(1); that is exactly the kind of shift you expect when the model is missing spectral components. Since Vzz and eta are extracted together with Qe/Qg, a systematic miss at X2 could bias the weak sites 3 and 4, whose contributions are only 0.4–3.6% of the signal. I would not call the whole inventory wrong, but a sceptic can reasonably say sites 1 and 2 are solid while sites 3 and 4 are tentative until the possible fifth site is either resolved or shown not to matter. The paper should report uncertainties on Qe/Qg and show how the extracted EFGs shift if the model is extended or the common-linewidth assumption is relaxed.\n\nOn the citation pattern and method: references are appropriate, no red flags. The fitting assumptions are stated clearly (Eq. 1); the common Gaussian width and free per-crystal Qe/Qg are the main assumptions to scrutinize. The lifetime numbers are internally consistent and the quenching effect is site-dependent, which is a nice observation.\n\nWho this is for: the nuclear clock community, and anyone working on Th-doped crystals or EFG measurements. The paper deserves a serious referee. My recommendation: send it to review, but the referee should push for proper uncertainty quantification, a discussion of the fifth-site scenario, and a sensitivity analysis for sites 3 and 4. With those addressed, the four-site claim would be much closer to closed.","headline":"Four-site inventory is a real advance, but the completeness claim is weakest at highest concentration and needs uncertainty reporting.","tokens_in":14873,"tokens_out":1575,"would_cite":true,"duration_ms":15389,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper reports that a vacuum-ultraviolet laser can resolve the nuclear quadrupole spectrum of thorium-229 in calcium fluoride into four distinct doping sites, each with its own electric field gradient, and assigns the two dominant…","keywords":["thorium-229","laser Mössbauer spectroscopy","nuclear clock","electric field gradient","CaF2 crystal","quadrupole splitting","doping sites","vacuum ultraviolet laser"],"falsifier":"Cool the X2 crystal to liquid-helium temperature and scan with a sub-megahertz-linewidth VUV source, with long integration to expose weak lines; if the residuals near sites 3 and 4 then require a fifth Gaussian site to fit, the four-site inventory is incomplete.","tokens_in":13586,"feed_emoji":"⚛️","tokens_out":12281,"duration_ms":111129,"temperature":0.7,"pith_summary":"The paper aims to establish that a vacuum-ultraviolet laser can perform Mössbauer-style nuclear spectroscopy on thorium-229 inside calcium fluoride, and that the resolved spectra reveal four distinct thorium doping sites, each with its own electric field gradient. If this holds, it explains the unassigned lines left over from earlier measurements and provides a site-resolved picture of where thorium sits in the host crystal. That matters because the local site determines the line shape, the isomeric lifetime, and the laser-quenching behavior of the nuclear clock transition, which are the data needed for designing a solid-state nuclear clock.","feed_headline":"VUV laser maps four thorium sites in a nuclear-clock crystal","feed_subtitle":"Site-selective field gradients and lifetimes provide the material data needed to build solid-state nuclear clocks.","key_machinery":"The central mechanism is nuclear quadrupole splitting: the thorium nucleus's electric quadrupole moment $Q$ interacts with the electric field gradient (EFG) tensor of the crystal, splitting the nuclear ground state ($I_g = 5/2$) into three sublevels and the isomeric state ($I_e = 3/2$) into two, so each site contributes up to six transitions. The paper's instrument is a pulsed vacuum-ultraviolet laser at $\\sim$148 nm with a $\\sim$30 MHz linewidth, narrow enough to resolve those splittings. The load-bearing analysis is a common-width four-site Gaussian fit to the summed spectra, governed by the quadrupole Hamiltonian with site parameters $V_{zz}$, $\\eta$, and the fixed moment ratio $Q_e/Q_g$, with relative line intensities fixed by the transition matrix elements.","core_discovery":"On the paper's own terms, the central discovery is that the vacuum-ultraviolet spectrum of $^{229}$Th:CaF$_2$, recorded with a 30 MHz linewidth pulsed laser, is a superposition of six-line quadrupole patterns from four distinct thorium sites. The fit yields $V_{zz} \\approx 0$, $+110$, $-320$, and $+260$ V/Å$^2$ for sites 1–4, with site 2 matching the previously characterized site and the other three sites accounting for the transitions earlier work could not assign. Site 1 has a vanishing electric field gradient and thus a locally near-cubic environment; the paper assigns it to a thorium atom on a calcium site with distant charge compensation (Th$_{Ca}$−2e), and site 2 to a nearest-neighbor thorium pair (Th$_{Ca}$−4e), supported by DFT calculations and STEM images showing thorium clustering. The radiative lifetime is about 630 s for all sites, but laser-induced quenching is site-dependent and strongest for site 1, making that high-symmetry site the preferred candidate for a future solid-state nuclear clock.","pith_inferences":["If the four-site inventory is complete, crystal growth conditions (dopant concentration, annealing, cooling rate) should be able to shift the site population balance, potentially yielding crystals dominated by a single site for simpler clock operation.","The concentration scaling of sites 3 and 4 suggests they are also thorium clusters; a growth series varying only concentration could turn site-resolved spectroscopy into a fingerprinting tool for cluster geometries in rare-earth-doped fluorites.","The site-dependent quenching efficiency implies the nuclear decay path is modulated by the local electronic structure; comparing the same site's quenching across temperatures could separate phonon-assisted from direct electronic mechanisms.","Going below the current 30 MHz laser linewidth at low temperature might reveal whether site 1's field gradient is exactly zero or merely unresolved, distinguishing a true cubic site from a slightly distorted one and sharpening the DFT-based assignment."],"forward_implications":["The unassigned transitions in earlier $^{229}$Th:CaF$_2$ spectra are accounted for by three additional doping sites, so future spectra can be analyzed as a sum of four known site patterns.","Site-selective excitation makes it possible to interrogate each of the four sites independently; the radiative lifetime is the same (about 630 s) for all sites, while the laser-induced quenching rate differs by site, with site 1 quenching most efficiently.","The structural assignments of the two dominant sites (isolated thorium on a calcium site with distant compensation, and a nearest-neighbor thorium pair) provide the microscopic geometry needed to model how the crystal environment responds to stress, temperature, and strain.","The technique can be transferred to other VUV-transparent host materials and, with conversion-electron detection, to opaque materials, extending laser Mössbauer spectroscopy beyond $^{229}$Th:CaF$_2$.","Site 1's near-zero field gradient and strong quenching make it a leading candidate for a solid-state nuclear clock, with site 2 as a built-in cross-check for systematic effects."],"supporting_citations":[{"why":"Earlier spectrum that reported one thorium site and left several transitions unassigned; the four-site model in this paper explains them.","marker":"[7]"},{"why":"Precise values for site 2's Vzz, eta, and Qe/Qg; used as the consistency check for the new fits.","marker":"[10]"},{"why":"Previous laser-induced quenching study with a broad laser; this paper resolves the effect per site and explains the concentration dependence.","marker":"[9]"},{"why":"Theoretical EFG estimates for thorium defect configurations, used to rule out oxygen- and nearest-neighbor-compensation schemes for site 2.","marker":"[17]"},{"why":"Multiconfigurational calculation that attributed the same isolated-defect geometry to earlier VUV absorption features, supporting the site 1 assignment.","marker":"[23]"},{"why":"Characterization establishing thorium in the 4+ charge state and ruling out interstitial thorium, motivating the substitutional ThCa model.","marker":"[31]"},{"why":"Supplies the fixed ground-state quadrupole moment Qg = 3.11 eb used as an anchor in the fits.","marker":"[30]"},{"why":"The annealing and growth procedure (CF4 atmosphere at 1250 C) that produced the VUV-transparent crystals measured here.","marker":"[20]"}],"fun_headline_variants":["Three new thorium sites found in CaF2 for nuclear clocks","Four thorium sites mapped by VUV laser in clock crystal","Site-selective laser excitation reveals four thorium environments","VUV laser probes four thorium sites with distinct field gradients","Thorium-229 in CaF2: four sites for a future nuclear clock"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole inventory rests on assuming that every observed line belongs to exactly four sites, each producing Gaussian lines of the same width; if a fifth site (which the paper itself suggests for the highest-concentration crystal) or site-dependent linewidths exist, the extracted field gradients for the weaker sites would be biased.","fun_headline_variants_meta":{"raw":{"variants":["Three new thorium sites found in CaF2 for nuclear clocks","Four thorium sites mapped by VUV laser in clock crystal","Site-selective laser excitation reveals four thorium environments","VUV laser probes four thorium sites with distinct field gradients","Thorium-229 in CaF2: four sites for a future nuclear clock"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001118,"raw_usage":{"total_tokens":4630,"prompt_tokens":897,"completion_tokens":3733,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":513,"completion_tokens_details":{"reasoning_tokens":3644}},"tokens_in":513,"tokens_out":3733,"duration_ms":28772,"temperature":1.0,"reasoning_tokens":3644,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T17:07:41.632432+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Cool the X2 crystal to liquid-helium temperature and scan with a sub-megahertz-linewidth VUV source, with long integration to expose weak lines; if the residuals near sites 3 and 4 then require a fifth Gaussian site to fit, the four-site inventory is incomplete.","supporting_citations":[{"cited_title":"Zhang et al., Frequency ratio of the ^ 229m Th nuclear isomeric transition and the ^ 87 Sr atomic clock","cited_arxiv_id":null,"evidence_quote":"Earlier spectrum that reported one thorium site and left several transitions unassigned; the four-site model in this paper explains them."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Precise values for site 2's Vzz, eta, and Qe/Qg; used as the consistency check for the new fits."},{"cited_title":"Schaden et al., Laser-induced quenching of the Th-229 nuclear clock isomer in calcium fluoride","cited_arxiv_id":null,"evidence_quote":"Previous laser-induced quenching study with a broad laser; this paper resolves the effect per site and explains the concentration dependence."},{"cited_title":"Dessovic et al., ^ 229 thorium-doped calcium fluoride for nuclear laser spectroscopy","cited_arxiv_id":null,"evidence_quote":"Theoretical EFG estimates for thorium defect configurations, used to rule out oxygen- and nearest-neighbor-compensation schemes for site 2."},{"cited_title":"Nalikowski, V","cited_arxiv_id":null,"evidence_quote":"Multiconfigurational calculation that attributed the same isolated-defect geometry to earlier VUV absorption features, supporting the site 1 assignment."},{"cited_title":"Takatori et al., Characterization of the thorium-229 defect structure in CaF _2 crystals","cited_arxiv_id":null,"evidence_quote":"Characterization establishing thorium in the 4+ charge state and ruling out interstitial thorium, motivating the substitutional ThCa model."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the fixed ground-state quadrupole moment Qg = 3.11 eb used as an anchor in the fits."},{"cited_title":"Beeks et al., Optical transmission enhancement of ionic crystals via superionic fluoride transfer: growing VUV-transparent radioactive crystals","cited_arxiv_id":null,"evidence_quote":"The annealing and growth procedure (CF4 atmosphere at 1250 C) that produced the VUV-transparent crystals measured here."}],"review_version":2}