{"id":"dea3d095-b0ca-40fb-90f5-50bb496908ee","arxiv_id":"2507.08634","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Rb2's ionization energy is measured as 31497.3 ±0.6 cm-1 by extrapolating field-ionization onsets to zero field, 149.3 cm-1 above the 2013 upper bound.","lead":"An experiment measures the ionization energy of the rubidium dimer Rb2 as 31497.3 ±0.6 cm-1 by extrapolating pulsed electric-field ionization onsets to zero field, 149.3 cm-1 above the 2013 Bellos value. The result matters because precise dimer ionization energies are needed to model ultracold ion-atom collisions and to calibrate molecular Rydberg spectroscopy.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"PFI-onset extrapolation is uncalibrated for Rb2: the fitted sqrt(F) slope (0.767) is ~8x smaller than the hydrogenic value (6.12) and the paper admits slew-rate dependence, so the F=0 intercept may be biased well beyond its quoted 0.5 cm-1 error.","rationale":"The paper is a careful measurement with a transparent analysis, and the central number is the intercept of a six-point linear regression. The weakest link is exactly the mapping from the observed half-rise of the PFI step to the field-free ionization threshold. The reader already identified this as the least secure assumption, and I agree. The concern is not internal inconsistency or bad faith; it is a calibration gap. The factor-of-eight discrepancy between the fitted slope and the hydrogenic value is a red flag, and the paper's own citation of slew-rate dependence means the effective field entering the √F law is not independently established. The manual choice of Imin/Imax in the logistic fit and the subtraction of three Gaussian peaks add further model uncertainty that is absent from the error budget. A calibration test on a known atomic PFI threshold, or an equivalent re-extraction of the v=1, 2, 3 thresholds in Fig. 3 with the same analysis, would settle whether ν0(0) is robust. If that test passes, the measurement stands; if not, the 149.3 cm-1 discrepancy with Bellos is likely a methodological artifact. For these reasons the reader's CONDITIONAL verdict should be kept; no adjustment is proposed beyond requiring the calibration or an equivalently decisive validation.","tokens_in":11651,"tokens_out":10577,"duration_ms":138533,"concrete_test":"Calibrate the PFI analysis on a known threshold in the same apparatus: record the pulsed-field ionization spectrum of Rb atoms (e.g., 5s→np Rydberg series converging to Rb+ 5s) using the same 20 ns rise-time field and apply the identical logistic + √F extrapolation. If the extracted Rb ionization energy is not recovered to within ±1 cm-1 of the well-known value 33690.81 cm-1, the slope anomaly and slew-rate dependence indicate a calibration bias rather than a molecular effect, and the Rb2 value must be corrected before publication.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section V determines Ei from the F=0 intercept of a linear fit of ν0(F) versus √F. The load-bearing assumption is that the logistic half-rise of the residual PFI step (after subtracting P1, P2, P3) tracks the field-lowered ionization threshold, with a shift proportional to √F. This assumption is not calibrated for Rb2. The fitted slope is 0.767 cm-1/(V/cm)^{1/2}, whereas the hydrogenic field-ionization slope used in the Mg+ work is 6.12, and the paper itself notes that the slope depends on the field slew rate [25]. Thus the observed onset may be controlled by the pulsed-field time profile and by which high-n Rydberg states survive to be ionized, not simply by the static saddle-point threshold. With the lowest field at 18 V/cm, the implied shift is only 3.25 cm-1, so a small nonlinearity in the √F law, a peak-versus-effective-field offset, or a bias from the manual Gaussian subtraction and the fixed Imin/Imax values in Eq. (2) could move ν0(0) by several times the stated ±0.5 cm-1. Because the claim against Bellos is a 149.3 cm-1 offset, even a few cm-1 of unmodeled bias would change the central conclusion, and the quoted error does not include these model choices.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports a measurement of the ionization energy of 85Rb2 using resonantly enhanced two-photon ionization (RE2PI) in a supersonic beam. Molecules are first excited via the X^1Σg+ (v=0) → B^1Πu (v=2) transition at 14753 cm−1, and a second photon is scanned over 16720–16755 cm−1 while Rb2+ ions are detected after pulsed-field ionization. For six electric field strengths between 18 and 180 V/cm, the onset of the direct photoionization step is modeled by subtracting three Gaussian peaks (P1, P2, P3) and fitting the residual signal with a logistic function; the half-rise wavenumbers ν0(F) are then linearly extrapolated versus √F to F = 0, yielding Ei = 31497.3 ± 0.6 cm−1 and, via Eq. (3), D0(X^2Σg+) = 6158.2 ± 0.6 cm−1. The paper also assigns the prominent peaks and additional structure to molecular Rydberg series converging to vibrational levels of Rb2+, and compares the result with previous experiments and quantum-chemistry calculations, concluding that the 2013 Bellos et al. value is an upper bound 149.3 cm−1 below the present result.","tokens_in":11998,"tokens_out":7553,"duration_ms":85653,"significance":"If the measurement is correct, it provides a long-missing accurate ionization energy for Rb2 and a benchmark for theory, reducing the uncertainty from the previous ~400 cm−1 level to 0.6 cm−1 and improving on the 0.6 cm−1 upper bound of Bellos et al. The experiment is clearly described and the raw normalized spectra at six field strengths are displayed. The paper is honest in acknowledging that the fitted √F slope (0.767) differs strongly from the hydrogenic value used in the Mg+ calibration and that the slope depends on the pulsed-field slew rate; this is a significant limitation. The accompanying quantum-chemistry calculations are useful, but the claimed agreement with the measured D0 is not quantitative and needs to be restated.","major_comments":[{"comment":"The central result depends on the assumption that the logistic half-rise of the residual PFI signal tracks the static field-lowered ionization threshold, with a shift linear in √F. This assumption is not calibrated for Rb2. The fitted slope is 0.767 cm−1/(V/cm)^{1/2}, roughly eight times smaller than the hydrogenic value 6.12 used for Mg+ in Ref. [30], and the paper itself notes that the slope depends on the time derivative of the pulsed field [25]. With the lowest field at 18 V/cm, the implied shift is only about 3.3 cm−1, so an unmodeled offset or nonlinearity of a few cm−1 could change the F = 0 intercept by an amount comparable to the claimed 149.3 cm−1 discrepancy with Bellos et al. Please provide a calibration of the onset-extrapolation method for Rb2, or a quantitative uncertainty budget that includes the effect of the finite slew rate and the non-hydrogenic slope.","section":"Section V, Eq. (2)"},{"comment":"The quoted uncertainty ν0(0) = 16744.3 ± 0.5 cm−1 does not include systematic contributions from the manual choice of Imin and Imax in Eq. (2), the Gaussian subtraction of P1–P3, the measured field inhomogeneity (≈11% axial and ≈3% radial, Section III), or the choice of the √F functional form. The horizontal error bars are described as ±5σ with σ = 0.09 cm−1 derived only from the laser linewidths; this is not a measure of the onset-fitting error. An estimate of these systematic effects is needed before the ±0.6 cm−1 total uncertainty on Ei and D0 can be accepted.","section":"Section V and Fig. 4"},{"comment":"The statement that the measured D0(X+) = 6158.2 ± 0.6 cm−1 \"agrees with our theoretical determination\" is overstated. Table I lists two present calculations: t12 gives D0 = 6200.63 cm−1 and t11 gives D0 = 6098.74 cm−1, which differ from the measurement by +42.4 cm−1 and −59.5 cm−1, respectively, i.e., roughly 70–100 times the quoted experimental uncertainty. The paper should either soften this claim to an approximate agreement at the 1% level or discuss the systematic errors in the quantum-chemistry calculations that would account for these differences.","section":"Table I and Section VI"},{"comment":"The assignment of P1–P3 to Rydberg levels such as R2(35), R1(49), and R1(50) relies on fitted quantum defects δ1–δ5 that vary strongly with v (1.17, 0.69, 0.58, 0.28, 0.07) and on an invoked but unobserved doubly-excited autoionizing perturber. Because these peaks are subtracted before the onset analysis in Section V, a misassignment or an incorrect Gaussian model for them could bias the half-rise ν0(F). The paper should demonstrate that the extracted ν0(0) is robust to reasonable variations in the peak-subtraction procedure.","section":"Section IV and Fig. 2(b)"}],"minor_comments":[{"comment":"The finite-mass Rydberg correction has the wrong sign: the reduced-mass Rydberg constant should be R_M = R∞/(1 + me/md), not R∞(1 + me/md). Although the effect is small at the quoted level, the formula as written would increase the Rydberg energy rather than decrease it.","section":"Eq. (1)"},{"comment":"The red laser wavenumber is given as 14753 cm−1 in Section III but later as 14753.0 ± 0.03 cm−1 in Section V; please clarify how the 0.03 cm−1 uncertainty is obtained given the stated wavemeter accuracy of 0.1 cm−1.","section":"Sections III and V"},{"comment":"The vertical error bars are said to \"mirror the relative uncertainty of F estimated at ±10%\"; please specify whether ±10% is a 1σ value or a conservative bound, and how it is propagated to the error bar on ν0.","section":"Fig. 4 caption"},{"comment":"The supplemental material reference still contains the placeholder \"[url]\"; a working link or a more complete description should be provided.","section":"Ref. [23]"},{"comment":"The phrase \"We modeled the onset of the ionization signal\" would be more precise as \"We analyzed the onset\" or \"We modeled and measured the onset\", since the onset position is extracted from the data rather than solely modeled.","section":"Abstract"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of the journal and addresses an important quantity. The main risk is the uncalibrated PFI-onset extrapolation; I recommend that the editor seek a referee with specific expertise in pulsed-field ionization and Rydberg-state threshold measurements. The overstatement of agreement with the authors' own quantum-chemistry calculations should be corrected in revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe headline: this paper reports the first precise field-free ionization energy of 85Rb2, Ei = 31497.3 ± 0.6 cm-1, and a corresponding D0(X+) = 6158.2 ± 0.6 cm-1, from RE2PI with pulsed-field ionization of molecular Rydberg states. If correct, it completes the alkali dimer set and improves input for ultracold ion-neutral collision models. It also challenges the 2013 Bellos upper bound, which is 149.3 cm-1 lower.\n\nWhat is genuinely new: the method—an extrapolation of the PFI onset versus sqrt(F)—has been used for Mg+ and MgKr+, but this is the first application to a neutral alkali dimer. The measurement is not a parameter scan or restatement of theory. The data are shown at six field values, the fits are visible, and the analysis is transparent. D0 is obtained from Ei plus two external measured constants, so there is no circularity.\n\nThe soft spots are real but not fatal. The load-bearing assumption is that the half-rise of the residual ion signal tracks the field-lowered threshold with a shift proportional to sqrt(F). The fitted slope is 0.767 cm-1/(V/cm)^{1/2}, about eight times smaller than the hydrogenic slope (6.12) used in the Mg+ work. The paper notes the slope depends on the time derivative of the pulsed field. That means the extrapolation to F = 0 is not calibrated for this molecule, and the ±0.5 cm-1 intercept error excludes model choices like the manual Gaussian subtraction and the fixed Imin/Imax in the logistic fit. In absolute terms, that could be several cm-1. It would not erase the 149.3 cm-1 gap to Bellos, but it makes the stated ±0.6 cm-1 questionable.\n\nSecond, the paper says its measured D0 'agrees' with its own quantum chemistry values, but Table I shows t11 D0 = 6098.74 and t12 D0 = 6200.63 cm-1, i.e. differences of -59 and +42 cm-1 from the measured 6158.2. That is not agreement at the claimed precision. The Rydberg assignment is also called unambiguous, but it relies on fitted quantum defects δ1–δ5 that vary strongly with v (1.17 down to 0.07), and the explanation invokes an unobserved perturbing doubly-excited state. That is plausible but speculative.\n\nWho will get value: molecular spectroscopy and ultracold chemistry people, and anyone relying on the Bellos value. It deserves a serious referee, but the referee should push for a systematic error budget on the PFI calibration, a direct comparison of the fitted slope with a Rydberg-state simulation or a calibrated measurement, and a toned-down claim about the theory agreement. I would accept it for review but not as a definitive replacement of the benchmark in its current form.","headline":"A likely right new Rb2 ionization energy, but the absolute calibration of the PFI extrapolation needs scrutiny before it replaces the Bellos benchmark.","tokens_in":12615,"tokens_out":2378,"would_cite":true,"duration_ms":25899,"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":"The paper establishes the field-free ionization energy of $^{85}$Rb$_2$ as $31497.3 \\pm 0.6$ cm$^{-1}$ and derives the Rb$_2^+$ dissociation energy of $6158.2 \\pm 0.6$ cm$^{-1}$.","keywords":["ionization energy","rubidium dimer","pulsed-field ionization","Rydberg states","RE2PI","dissociation energy","quantum defect","supersonic beam"],"falsifier":"A decisive check would be an independent measurement of $E_i$ by PFI-ZEKE photoelectron spectroscopy or rotationally resolved two-photon ionization; a value outside $31497.3 \\pm 0.6$ cm$^{-1}$ would show the half-rise extrapolation is biased. A cheaper internal test is to repeat the same analysis with the half-rise defined at 10% and 90% of the logistic step and with different Gaussian subtraction widths: if the $\\sqrt{F}$ extrapolations shift by more than about 1 cm$^{-1}$, the quoted uncertainty understates the model error.","tokens_in":11419,"feed_emoji":"⚛️","tokens_out":12368,"duration_ms":116268,"temperature":0.7,"pith_summary":"The paper reports a measurement of the ionization energy of the $^{85}$Rb$_2$ molecule using resonantly enhanced two-photon ionization with pulsed-field ionization detection. By modeling the onset of the ion signal as a function of applied electric field strength, the authors extract $E_i = 31497.3 \\pm 0.6$ cm$^{-1}$, which is $149.3$ cm$^{-1}$ larger than the previously reported upper bound [12]. The same value yields a dissociation energy $D_0(X^2\\Sigma_g^+) = 6158.2 \\pm 0.6$ cm$^{-1}$ for the Rb$_2^+$ ground state, in line with current quantum chemistry calculations. A simple quantum-defect model assigns the unevenly spaced peaks in the spectrum to molecular Rydberg series converging to the lowest vibrational levels of Rb$_2^+$. If correct, the result converts a long-uncertain constant for the rubidium dimer into one of the best-determined among alkali dimers.","feed_headline":"Rb2 ionization energy is pinned at 31497.3 ± 0.6 cm-1","feed_subtitle":"The new value also fixes the Rb2+ bond strength and puts the 2013 upper bound 149.3 cm-1 too low.","key_machinery":"The load-bearing mechanism is pulsed-field ionization: a positive high-voltage pulse applied $1.13\\,\\mu$s after the laser pulse lowers the ionization threshold of molecular Rydberg states by an amount proportional to $\\sqrt{F}$. The analysis chain is to subtract the Gaussian fits to the Rydberg peaks P1-P3, fit the remaining spectrum with the logistic function of Eq. (2), extract the half-rise $\\tilde{\\nu}_0(F)$, and extrapolate $\\tilde{\\nu}_0$ versus $\\sqrt{F}$ linearly to $F = 0$. A second piece is the Rydberg-series model of Eq. (1), a reduced-mass-corrected Rydberg formula with optimized quantum defects $\\delta_v$ (from 1.17 to 0.07 for $v = 1$ to 5), which assigns the observed unevenly spaced lines to series $R_v(n)$ converging to the lowest vibrational levels of the Rb$_2^+$ ground state.","core_discovery":"The central claim is that the field-free ionization threshold of $^{85}$Rb$_2$ lies at $E_i = 31497.3 \\pm 0.6$ cm$^{-1}$, obtained by extrapolating the half-rise wavenumber $\\tilde{\\nu}_0$ of the residual ion signal linearly in $\\sqrt{F}$ to $F = 0$. For each electric field strength between 18 and 180 V/cm, the paper subtracts three Gaussian peaks assigned to Rydberg levels and fits the remaining step with a logistic function; the regression of $\\tilde{\\nu}_0$ against $\\sqrt{F}$ gives the second-photon threshold $16744.3 \\pm 0.5$ cm$^{-1}$, which added to the first photon at $14753.0$ cm$^{-1}$ yields $E_i$. Using the known $^{85}$Rb ionization limit and the measured $D_0(X^1\\Sigma_g^+)$, the paper derives $D_0(X^2\\Sigma_g^+) = 6158.2 \\pm 0.6$ cm$^{-1}$ for Rb$_2^+$ and argues that the 2013 upper bound [12] was too low because stray electric fields of at least 300 V/cm ionized Rydberg molecules in that experiment, shifting its apparent onset.","pith_inferences":["A rotationally resolved or ZEKE measurement would test whether the half-rise analysis is biased by unresolved Rydberg structure; the quoted 0.6 cm$^{-1}$ assumes the three-Gaussian subtraction removes all structure near the step.","Because the extrapolation slope depends on the pulsed-field slew rate, repeating the measurement with rise times other than 20 ns would give a direct handle on the systematic error not captured by the quoted uncertainty.","The fitted quantum defects vary from 1.17 to 0.07 across $v=1$ to 5, which is unusually strong for levels near the potential minimum; a multichannel quantum-defect treatment including the doubly-excited state might reassign some of the weaker lines.","If the new $E_i$ holds, the earlier photoassociation-based bound should be interpreted as a field-affected onset, which may require revisiting other molecular constants derived from that spectrum."],"forward_implications":["The Rb$_2$ ionization energy becomes known to 0.6 cm$^{-1}$, bringing the rubidium dimer in line with other alkali dimers and anchoring the Rb$_2^+$ potential curve experimentally.","The derived $D_0(X^2\\Sigma_g^+) = 6158.2 \\pm 0.6$ cm$^{-1}$ gives a direct benchmark for potential curves used in modeling cold Rb$^+$ + Rb three-body recombination and molecular-ion formation.","The 2013 upper bound [12] is reinterpreted as a field-affected onset lying 149.3 cm$^{-1}$ below the true threshold, which changes how that photoassociation spectrum should be used.","The assigned Rydberg series with strongly $v$-dependent quantum defects point to a perturbing doubly-excited autoionizing state, giving a target for higher-resolution Rydberg spectroscopy.","The same PFI-RE2PI approach can be applied to other alkali dimers, opening a route to complete and precise the set of dimer ionization energies."],"supporting_citations":[{"why":"Supplies the earlier upper bound to $E_i$ that the new value exceeds by 149.3 cm$^{-1}$; the main comparison target.","marker":"[12]"},{"why":"Provides the logistic-function and $\\sqrt{F}$ extrapolation procedure for PFI-ZEKE onsets that is adapted to the molecular spectra.","marker":"[30]"},{"why":"Gives the pulsed-field ionization mechanism by which the electric field lowers the observed threshold.","marker":"[24]"},{"why":"Documents the dependence of selective field ionization on the pulsed-field slew rate, cited for the regression slope.","marker":"[25]"},{"why":"Demonstrates PFI-ZEKE spectroscopy on Mg$_2^+$ and is the methodological precedent for molecular ions.","marker":"[26]"},{"why":"Provides the experimental X and B state spectroscopic constants, including the $v_X=0 \\to v_B=2$ transition used for the first photon.","marker":"[13]"},{"why":"Fixes the $^{85}$Rb ionization energy used in Eq. (3) to convert $E_i$ into $D_0(X^+)$.","marker":"[19]"},{"why":"Provides the model-potential Rb$_2^+$ ground-state potential used in the theoretical comparison of $D_0$.","marker":"[16]"}],"fun_headline_variants":["New Rb2 ionization energy: 31497.3 cm-1, 149.3 cm-1 above 2013","Precise Rb2 ionization energy measured, revises 2013 upper bound","Rb2 ionization and Rb2+ bond strength pinned via field ionization","Electric-field ionization yields Rb2 threshold and Rb2+ dissociation energy","Rb2 ionization energy corrected: 31497.3 cm-1, offsets 2013 by 149 cm-1"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The half-rise of the residual ion signal, after subtracting the three Gaussian peaks, is assumed to track the field-lowered ionization threshold with a shift proportional to $\\sqrt{F}$, so that linear extrapolation to $F = 0$ recovers the field-free threshold; if unresolved Rydberg structure, imperfect peak subtraction, or the pulsed-field time profile biases the half-rise, the quoted $\\pm 0.6$ cm$^{-1}$ uncertainty would be too small.","fun_headline_variants_meta":{"raw":{"variants":["New Rb2 ionization energy: 31497.3 cm-1, 149.3 cm-1 above 2013","Precise Rb2 ionization energy measured, revises 2013 upper bound","Rb2 ionization and Rb2+ bond strength pinned via field ionization","Electric-field ionization yields Rb2 threshold and Rb2+ dissociation energy","Rb2 ionization energy corrected: 31497.3 cm-1, offsets 2013 by 149 cm-1"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000652,"raw_usage":{"total_tokens":3092,"prompt_tokens":1150,"completion_tokens":1942,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":766,"completion_tokens_details":{"reasoning_tokens":1823}},"tokens_in":766,"tokens_out":1942,"duration_ms":15101,"temperature":1.0,"reasoning_tokens":1823,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T18:14:52.160634+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive check would be an independent measurement of $E_i$ by PFI-ZEKE photoelectron spectroscopy or rotationally resolved two-photon ionization; a value outside $31497.3 \\pm 0.6$ cm$^{-1}$ would show the half-rise extrapolation is biased. A cheaper internal test is to repeat the same analysis with the half-rise defined at 10% and 90% of the logistic step and with different Gaussian subtraction widths: if the $\\sqrt{F}$ extrapolations shift by more than about 1 cm$^{-1}$, the quoted uncertainty understates the model error.","supporting_citations":[{"cited_title":"Bellos, R","cited_arxiv_id":null,"evidence_quote":"Supplies the earlier upper bound to $E_i$ that the new value exceeds by 149.3 cm$^{-1}$; the main comparison target."},{"cited_title":"Genevriez, D","cited_arxiv_id":null,"evidence_quote":"Provides the logistic-function and $\\sqrt{F}$ extrapolation procedure for PFI-ZEKE onsets that is adapted to the molecular spectra."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the pulsed-field ionization mechanism by which the electric field lowers the observed threshold."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Documents the dependence of selective field ionization on the pulsed-field slew rate, cited for the regression slope."},{"cited_title":"G´ en´ evriez, M","cited_arxiv_id":null,"evidence_quote":"Demonstrates PFI-ZEKE spectroscopy on Mg$_2^+$ and is the methodological precedent for molecular ions."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the experimental X and B state spectroscopic constants, including the $v_X=0 \\to v_B=2$ transition used for the first photon."},{"cited_title":"Aymar and O","cited_arxiv_id":null,"evidence_quote":"Fixes the $^{85}$Rb ionization energy used in Eq. (3) to convert $E_i$ into $D_0(X^+)$."},{"cited_title":"Aymar, S","cited_arxiv_id":null,"evidence_quote":"Provides the model-potential Rb$_2^+$ ground-state potential used in the theoretical comparison of $D_0$."}],"review_version":1}