REVIEW 3 major objections 4 minor 21 references
Evidence of the $e^+ e^-\to f_1(1285)$ reaction with the CMD-3 detector at VEPP-2000
T0 review · 3 major / 4 minor · reviewed 2026-08-05 · deepseek-v4-flash
Pith's one-line read A 26-event excess at 1282 MeV is claimed as 4.5σ evidence for e+e− → f1(1285) production, the first statistically significant direct production of a C-even resonance in electron-positron collisions.
desk verdict A credible, careful first measurement of f1(1285) direct production, but the 4.5σ needs a trial-correction check and the wording needs to respect the earlier chi_c1 observation. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The analysis selects events with six or more photons, identifies two π0's and two unconstrained photons, and applies a 6C kinematic fit (constraining total energy and momentum to the initial state and both photon pairs to the π0 mass). It then retains events near the a0(980) and η bands in the 2D plane of m(π0η) versus m(γγ). Signal events are counted by an unbinned likelihood fit of the m(γγ) projection with a double-Gaussian signal shape fixed from simulation plus a second-order polynomial background.
What would settle it
Collect additional off-peak data at energies just below and above 1282 MeV with the same event selection and fit the same 2D region under a peaking-background hypothesis; if the apparent signal at 1282 MeV does not scale as a resonance (i.e., a comparable excess appears in off-peak control samples) or if the signal yield falls below about 2σ when a third-order polynomial or an ωπ0 reflection component is included, the evidence claim would not survive.
Extended reading notes
Core claim
The central claim is that a clean excess of 26.5 ± 7.4 events over a smooth background in the two-dimensional distribution of m(π0η) versus m(γγ) at Ec.m.=1282 MeV constitutes 4.5σ evidence for the reaction e+e− → f1(1285) → ηπ0π0, produced via two-photon annihilation. The extracted peak cross section is 0.074 ± 0.021 ± 0.010 nb, and the implied e+e− branching fraction of the f1(1285) is (8.3 ± 2.3 ± 1.1)×10−9. The paper emphasizes that a C-even resonance cannot be produced by a single photon, so observation fixes the production mechanism as two-photon exchange and provides a direct estimate of the f1's leptonic coupling, complementing (and consistent with) the earlier two-event SND result a
Load-bearing premise
The load-bearing premise is that the background in the selected m(π0η)–m(γγ) region is smooth and well described by a second-order polynomial; the paper states it cannot prove the background is uniform, so a locally peaking background of even a few events would weaken the signal.
Editorial extensions
If this is right
- If the signal is real, this is the first statistically significant observation of C-even resonance production in e+e− collisions.
- The measured B(f1(1285)→e+e−) = (8.3 ± 2.3 ± 1.1)×10−9 can be compared with theoretical predictions and supports the two-photon production mechanism.
- The result is consistent with the SND two-event measurement and with the theoretical estimate, strengthening the case for the inverse-reaction approach to measuring leptonic widths of C-even mesons.
- The absence of a signal in the nearby-energy scan data indicates that the excess is peaking at the f1 mass, as expected for a resonance.
Reading between the lines
- If confirmed, the same analysis method could be applied at higher luminosity to search for other C-even axial-vector or tensor mesons (e.g., f1(1420), a1, f2) in e+e− data, probing their leptonic widths.
- The measured e+e− width, when combined with hadronic light-by-light models, could provide a more precise constraint on the contribution of axial-vector mesons to the muon anomalous magnetic moment.
- A direct test of the two-photon production mechanism would be to measure the angular distribution of the f1 decay products; if the excess is produced via two virtual photons, it should match the predicted helicity structure.
- The paper's reliance on a smooth background shape could be checked by collecting more control-sample data at energies immediately below and above 1282 MeV; a similar excess there would indicate a non-resonant background.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper reports a search for the reaction e+e− → f1(1285) with the CMD-3 detector at VEPP-2000. Using 51.7 pb−1 collected at Ec.m.=1282 MeV, the authors select ηπ0π0 final states in a two-dimensional m(ηπ0) vs m(γγ) plane and find 26.5±7.4 signal events in the γγ invariant-mass projection, corresponding to a statistical significance of 4.5σ. A control sample from an energy scan (57.4 pb−1, Ec.m.=1.20–1.36 GeV) shows no significant signal. The authors interpret the excess as evidence for the two-photon production of the C-even f1(1285) resonance and derive a peak cross section σ0=0.074±0.021±0.010 nb and a branching fraction B(f1→e+e−)=(8.3±2.3±1.1)×10−9.
Significance. If correct, this is the first statistically significant direct observation of a C-even resonance in e+e− annihilation via the inverse process, and it provides a competitive measurement of the small f1→e+e− branching fraction. The result is relevant for the light-by-light contribution to (g−2)μ. The analysis has notable strengths: a dedicated on-peak data sample, a full GEANT4-based simulation with a channel-specific matrix-element generator, multiple cross-check procedures (constant-background fit, sideband counting, control scan at off-peak energies), and use of external PDG inputs for the branching fractions. The extraction of σ and B from the fitted yield is standard and does not involve circular reasoning. The main fragility is statistical: the evidence-level significance is not corrected for the data-driven optimization of the signal region, and the background model is constrained by only 41 events in the signal region.
major comments (3)
- [Fig. 3(a), Fig. 4(a); 'To maximize the signal-to-background ratio...'] The 2D signal region is selected by maximizing FOM = S/√B on the same on-peak data, and the quoted 4.5σ significance is then computed from a fit to the m(γγ) projection inside that same region. No trial factor, look-elsewhere correction, or cross-validation is described. This is a load-bearing issue: the likelihood-ratio statistic is not a pre-specified test statistic, and a positive background fluctuation in the optimized window is counted as signal. The control scan (Fig. 3(b), Fig. 4(c)) is not a substitute, because it has different luminosity per point and does not quantify the number of equivalent trials in the 2D optimization. Please provide a trial-factor estimate, e.g. by background-only pseudo-experiments that repeat the full 2D optimization and fit, or by an alternative pre-defined signal region.
- [Background model and systematics; text: 'We cannot prove that the background has uniform distribution'] The paper itself concedes that the background shape cannot be proven uniform. The nominal second-order polynomial fit, the constant-background fit, and the sideband counting give consistent yields (26.5, 28.9, 27.7), and the difference is absorbed into a 1–2 event systematic. However, the 2D selection was chosen after inspecting the data, and a local peaking background of even a few events would reduce the signal yield and significance. The control scan has only 2±2 signal events with small statistics, so it has limited power to exclude a narrow background bump. Please quantify the robustness against a locally peaking background, for example by adding a second narrow Gaussian at the signal mass with variable amplitude and reporting the likelihood profile, or by a data-driven background estimate from a sideband in both dimensions.
- [Significance statement and systematic uncertainties] The quoted significance is only the statistical likelihood ratio from the nominal fit. The systematic band includes selection variations and background subtraction, but it is not propagated into the significance. Given that the claim is 'evidence' at 4.5σ, the authors should state how systematics affect the significance, or at least emphasize that the 4.5σ is the statistical significance before systematic uncertainties. This is a presentation/interpretation point but it affects how the evidence claim is read.
minor comments (4)
- [General] There are several typos: 'two-dimetional' and 'two-dimentional' should be 'two-dimensional'; 'dash-dotted line' appears inconsistently; 'Ec.m.' is sometimes written as 'E_c.m.' and sometimes 'Ec.m.'.
- [References] Reference [20] is not cited in the text. Please either cite it where relevant or remove it from the list.
- [Fig. 1 and 2 captions] The captions describe distributions but do not always state the data sample (e.g., scan vs on-peak) clearly. For reproducibility, please add the energy and luminosity for each figure.
- [Formula for σ0] The definition of (1+δ) is not explicit. Please state whether (1+δ)=0.8 is the radiative correction factor that multiplies the Born cross section, and clarify the direction of the correction in the text, since it enters the cross section formula directly.
Circularity Check
No circularity found: the derived cross section and branching fraction follow from the fitted event yield via standard formulae with external inputs.
full rationale
The paper presents a counting experiment. Candidate events are selected in a 2D region, then a signal yield is extracted from a fit to the di-photon invariant mass distribution. The cross section is computed as sigma0 = N_signal / [L * epsilon * (1+delta) * B(f1 -> eta pi0 pi0)], and the e+e- branching fraction follows from a standard theoretical relation B(f1 -> e+e-) = sigma0 * m_f1^2 / (12*pi*C). None of these steps is circular: the fitted parameter N_signal is not forced to a target value, the inputs (L, epsilon, delta, B from PDG) are external, and the theoretical formula is independent of the measurement. The only data-dependent selection is the FOM = S/sqrt(B) optimization of the 2D region, which is a statistical trial-factor issue, not a logical reduction of the prediction to an input. The paper also candidly states 'We cannot prove that the background has uniform distribution,' but this is a statistical uncertainty, not circularity. No load-bearing self-citations, uniqueness imported from the authors' own work, or renaming of known results are present. The central claim is based on a bona fide fit to data, so the analysis is self-contained and not circular.
Assumptions & free parameters
free parameters (3)
- Number of signal events =
26.5 +- 7.4
- Background polynomial coefficients =
not quoted
- 2D selection window and chi2 thresholds =
not applicable (chosen by hand)
assumptions (6)
- standard math Standard Breit-Wigner relation between peak cross-section and branching fraction
- domain assumption f1(1285) parameters from PDG (mass, width, B(f1->eta pi0 pi0)=17.3%)
- domain assumption Signal Monte Carlo decay model: 73% a0(980)pi0 and 27% f0(500)eta
- domain assumption Radiative correction (1+delta)=0.8 is accurate
- domain assumption Background in the signal region is smooth and described by a second-order polynomial
- domain assumption The observed signal is attributed to f1(1285) rather than another C-even state
Cite this review
Pith. "Pith review of Evidence of the $e^+ e^-\to f_1(1285)$ reaction with the CMD-3 detector at VEPP-2000." pith.science (2026). https://pith.science/paper/TBOCEDMZ
@misc{pith2026260803417,
author = {Pith},
title = {Pith review of: Evidence of the $e^+ e^-\to f_1(1285)$ reaction with the CMD-3 detector at VEPP-2000},
year = {2026},
howpublished = {\url{https://pith.science/paper/TBOCEDMZ}},
note = {Machine review of arXiv:2608.03417}
}
abstract
A search for the $e^+ e^-\to f_1(1285)$ reaction is performed with the CMD-3 detector at the VEPP-2000 collider. Data from a scan in the center-of-mass energy range Ec.m.=1.20--1.36 GeV with 109.1 pb$^{-1}$ of the integrated luminosity is used. A dedicated 51.7 pb$^{-1}$ have been collected at the $f_1(1285)$ resonance peak energy Ec.m.=1282 MeV, where $26.5 \pm 7.4$ events have been found in the $e^+ e^-\to\eta\pi^0\pi^0$ reaction. No indication of the signal is observed in the scan data. We interpret the result as 4.5$\sigma$ evidence of the C-even resonance production in the $e^+ e^-\to f_1(1285)\to\eta\pi^0\pi^0$ reaction via the two-photon interaction. The production cross section is found to be $\sigma(e^+ e^-\to f_1) = 0.074 \pm 0.021 \pm 0.010$ nb, and the corresponding branching fraction to the $e^+ e^-$ pair is $B(f_1(1285)\to e^+e^-) = (8.3 \pm 2.3 \pm 1.1)\times 10^{-9}$.
Figures
Reference graph
Works this paper leans on
-
[1]
P. V. Vorob’ev et al. (ND Collaboration), Sov. J. Nucl. Phys. 48, 273 (1988) [Yad. Fiz. 48, 436 (1988)]
work page 1988
-
[2]
M. N. Achasov et al. (SND Collaboration), Phys. Lett. B 492, 8 (2000)
work page 2000
-
[3]
M. N. Achasov et al. (SND Collaboration), Phys. Rev. D 91, 092010 (2015)
work page 2015
-
[4]
M. N. Achasov et al. (SND Collaboration), Phys. Rev. D 98, 052007 (2018)
work page 2018
-
[5]
M. Ablikim et al. (BESIII Collaboration), Phys. Lett. B 749, 414 (2015)
work page 2015
-
[6]
M. Ablikim et al. (BESIII Collaboration), Phys.Rev.Lett. 129 (2022) 12, 12
work page 2022
-
[7]
M. N. Achasovet al.(SND Collaboration), Phys.Lett.B 800 (2020) 135074
work page 2020
-
[8]
A. S. Rudenko, Phys. Rev. D 96, 076004 (2017)
work page 2017
Show all 21 references
-
[9]
A. I. Milstein and A. S. Rudenko, arXiv:1909.07938 [hep- ph]
1909 arXiv
-
[10]
Navaset al.(Particle Data Group), Phys
S. Navaset al.(Particle Data Group), Phys. Rev. D110, 030001 (2024) and 2025 update
2024
-
[11]
Danilovet al., Proceedings EPAC96, Barcelona, p.1593 (1996); I.A
V.V. Danilovet al., Proceedings EPAC96, Barcelona, p.1593 (1996); I.A. Koop, Nucl. Phys. B (Proc. Suppl.) 181-182(2008) 371
1996
-
[12]
Khazin, Nucl
B.I. Khazin, Nucl. Phys. B (Proc. Suppl.)181-182, (2008) 376
2008
-
[13]
Grancagnoloet al., Nucl
F. Grancagnoloet al., Nucl. Instr. Meth. A623(2010) 114
2010
-
[14]
Anisyonkovet al., Nucl
A.V. Anisyonkovet al., Nucl. Instr. Meth. A598(2009) 266
2009
-
[15]
Epifanov (CMD-3 Collaboration), J
D. Epifanov (CMD-3 Collaboration), J. Phys. Conf. Ser. 293(2011) 012009
2011
-
[16]
Akhmetshinet al., Nucl
R.R. Akhmetshinet al., Nucl. Phys. B (Proc. Suppl.) 225-227(2012) 69
2012
-
[17]
Agostinelliet al.(GEANT4 Collaboration), Nucl
S. Agostinelliet al.(GEANT4 Collaboration), Nucl. In- str. Meth. A506(2003) 250
2003
-
[18]
Kuraev and V.S
E.A. Kuraev and V.S. Fadin, Sov. J. Nucl. Phys.41, (1985) 466; S. Actiset al., Eur. Phys. J. C66(2010) 585
1985
-
[19]
Abakumovaet al., Phys
V. Abakumovaet al., Phys. Rev. Lett.110(2013) 140402
2013
-
[20]
Gribanov and A.S
S.S. Gribanov and A.S. Popov, JINST18(2023) P05030
2023
-
[21]
M. N. Achasovet al.(SND Collaboration), Phys. Rev. D94(2016) 11, 112001
2016
Reviewed August 5, 2026 · model on record in the stance chip above.
Discussion (0). Sign in to comment.