Photon production from gluon splitting and fusion induced by a magnetic field of arbitrary strength in relativistic heavy-ion collisions

Alejandro Ayala; Ana Julia Mizher; Jos\'e Jorge Medina-Serna; Santiago Bernal-Langarica

REVIEW

Photon production from gluon splitting and fusion induced by a magnetic field of arbitrary strength in relativistic heavy-ion collisions

Alejandro Ayala, Santiago Bernal-Langarica, José Jorge Medina-Serna, Ana Julia Mizher

High Energy Physics - Phenomenology · Submitted 21 Aug 2026 · Published version

Pith's verdict: A genuine new vertex calculation, but the photon-yield and v2 claims rest on a fitted subtraction and a cumulative-vs-differential mismatch that make the 'photon puzzle solved' assertion uninterpretable.

T0 · AI review

How this review was made: An AI referee (deepseek-v4-flash) read the paper on August 27, 2026 and answered a fixed list of questions about it. T0 is the starting level every paper gets. The levels describe how deeply Pith has checked the paper, never the authors.

T0 AI review An AI read the paper and answered a fixed list of questions.
T1 Audit Automatic checks that the references and files the paper relies on exist and match.
T2 Compute The paper's code or calculations re-run with the same results.
T3 Formal statement The claims written out in a formal math language a computer can read.
T4 Proof Those claims proved, and the proof checked by computer.

How Pith checks papers →

· 3 major objections · 5 minor · Ask about this paper · The compiled record · Read the full review →

Challenge this review Comment (0) Read on arXiv ↗

Cite ▾

arXiv ID2608.21600 (version 2)

Pith ID6YTJNJ3J, Pith's permanent record of this paper and its review

CitationAlejandro Ayala, Ana Julia Mizher, Jos\'e Jorge Medina-Serna, Santiago Bernal-Langarica. "Photon production from gluon splitting and fusion induced by a magnetic field of arbitrary strength in relativistic heavy-ion collisions." arXiv:2608.21600 (2026). https://pith.science/paper/6YTJNJ3J

Cite this reviewBibTeX for Pith's review of this paper

Claim paper

Why Pith says this

Highlight any text to comment on it.

Editor's note

read the letter

Two things to know. First, the one-loop two-gluon one-photon vertex for arbitrary magnetic field strength is a real advance: previous work from this group covered strong-field and intermediate-field regimes, and the full tensor structure with on-shell reduction to three independent structures is laid out in detail with complete expressions in the appendices. That part deserves serious attention. Second, the phenomenological comparison that carries the paper's headline claim does not hold up.

The v2 problem is internal and decisive. Equation (28) defines v2^mag(omega_q) with an integral over photon energy from 0 to omega_q, so it is the anisotropy of all photons with energy below omega_q, not the differential v2 at that energy. The PHENIX points are differential v2(pT). Plotting a cumulative quantity against differential data, and then mixing it with differential direct-photon v2 via Eq. (29), makes the claimed agreement in shape and strength uninterpretable. This is not a baseline-subtraction issue; it would remain even if the hydro baseline were perfect.

The yield comparison is softer but still weak. The paper states in Sec. IV that Lambda_s, eta, , xi, |eB|, and VT were chosen to give a good description of the PHENIX data, so the 'very good agreement' is a fit, not a prediction. There is no uncertainty quantification. And the subtraction of the hydro calculation of Ref. [47] from the data is ambiguous because that calculation already includes pre-equilibrium photons; the overlap with the magnetic contribution is not quantified.

Credit where due: the paper is transparent about parameters and about its ballpark intent, the matrix element formulas are given in full, and the comparison with the earlier strong-field approximation is informative.

Who is this for? Heavy-ion phenomenologists working on the direct-photon puzzle, and people doing QED/vertex computations in magnetic fields. The vertex calculation likely has value beyond this application.

Recommendation: send it to peer review, but with a clear expectation of major revision. The vertex part can be salvaged; the v2 comparison must be redone (apply the same cumulative definition to the data, or derive a differential v2), the yield comparison should be reframed as a tuned model with systematics, and the baseline overlap must be addressed. As it stands, the paper's conclusion overstates what the evidence shows.

Referee report

Significance

The technical core of the paper is a nontrivial extension of Ref. [75]: the two-gluon-one-photon vertex is kept exact in the magnetic-field strength, reduced to three on-shell tensor structures, and supplemented by explicit trace and scalar expressions in the appendices. The inclusion of both fusion and splitting channels and the comparison of two pre-equilibrium gluon distributions are strengths. If the phenomenological comparison were valid, the mechanism would be an interesting contribution to the photon puzzle. However, the central claim of agreement is weakened by an internal mismatch in the definition of v2, by parameters that are adjusted to the same PHENIX data, and by an incompletely specified baseline-subtraction procedure. These issues are load-bearing for the abstract's assertion of 'very good agreement', so the manuscript needs substantial revision before the phenomenological claim can be accepted.

Major objections (3)

[Sec. IV, Eq. (28)] The quantity v2^mag(omega_q) defined in Eq. (28) integrates the photon energy omega'_q from 0 to omega_q and divides by the total yield Nmag. It is therefore a cumulative, total-normalized Fourier moment of all magnetic photons with energy below omega_q, not the differential v2(p_T) at that energy. The PHENIX points shown in Figs. 4b-10b are differential v2 measurements, and Eq. (29) then combines this cumulative v2^mag with the differential v2_direct(omega_q) and differential yields dNmag/domega_q and dNdirect/domega_q. The plotted agreement for v2 is thus a comparison of unlike quantities, and the same inconsistency enters the weighted average. Even if the baseline subtraction were fully justified, the v2 comparison would need to be redone with a differential v2^mag(omega_q), or with a cumulative version of the experimental data.
[Sec. IV, Figs. 4-10 and Table I] The text states that 'The values of Lambda_s and eta were independently chosen in order to give a good description of the PHENIX data.' The same applies to the other parameters in Table I: <p_T>, xi, |eB|, and VT are varied over ranges to match the same PHENIX yield and v2 data. With at least six parameters adjusted to the same data set and no uncertainty bands, goodness-of-fit statistic, or parameter-sensitivity analysis, the agreement shown in Figs. 4-10 is a fit rather than a predictive test. The abstract's claim of 'very good agreement' is therefore not a falsifiable model test as presented. The authors should either tie these parameters to independent constraints (e.g., saturation scale, multiplicity, magnetic-field evolution) or provide a quantitative chi-square and sensitivity analysis.
[Sec. IV, subtraction of Ref. [47]] The comparison subtracts the full direct-photon calculation of Ref. [47] from the PHENIX data and then restores the magnetic contribution via Eq. (29). Ref. [47] is described as including prompt, thermal, and pre-equilibrium photons. The paper does not specify which pre-equilibrium channels are contained in that baseline, nor does it justify that adding the magnetic yield and v2 linearly does not double count or omit interference terms with the pre-equilibrium photons already included in the baseline. Please state the channel content of the subtracted baseline, quantify the sensitivity of the results to this subtraction, or perform a simultaneous calculation of the pre-equilibrium photon yield that includes both the magnetic and non-magnetic channels.

Minor objections (5)

[Fig. 5 caption and Sec. IV] The Fig. 5 caption lists the dotted curve as <p_T>=1.5 GeV, while the text states <p_T>=1.6 GeV; these values should be reconciled.
[Eq. (28)] There is a parenthesis typo in the splitting term: '1+(n(p_1-q))' should read '1+n(p_1-q)'.
[Table I and Sec. IV] The quantity VT=2155 fm^4 follows from V=(4/3)pi R^3 with R=7 fm and T=1.5 fm, but the table and text should define V and T separately and label the product as VT, since 'space-time volume' alone is ambiguous.
[Numerical methods, Secs. III-IV] The numerical code for the VEGAS/Suave integrations of Eqs. (16) and (28) is not provided; a repository or pseudocode would improve reproducibility of the matrix-element evaluation and the final curves.
[Captions of Figs. 4b-10b] The captions should explicitly state that the blue dash-dotted curve is the hydrodynamic baseline from Ref. [47], and that the other curves include the magnetic contribution, so the reader can distinguish model results from data.

How Pith checks papers → · Full audit of this paper →

The paper

The paper claims

Magnetic gluon fusion and splitting, added to the hydrodynamic baseline, reproduce the measured direct-photon yield and elliptic flow in Au+Au collisions.

The reading

This paper argues that the direct-photon puzzle in relativistic heavy-ion collisions can be addressed by photons radiated from gluon fusion ($gg\to\gamma$) and gluon splitting ($g\to g\gamma$) while the collision fireball still lives in a magnetic field. The authors compute the underlying two-gluon-one-photon vertex at one loop for a magnetic field of arbitrary strength, combine it with two proposed pre-equilibrium gluon distributions, and add the resulting photon yield and elliptic flow $v_2$ to a hydrodynamic baseline. They report good agreement with the measured direct-photon data for Au+Au collisions at $\sqrt{s_{NN}}=200$ GeV in the 20--30% centrality class, including the shape and magnitude of $v_2$. The result matters because it offers a perturbative, early-time source of photon elliptic flow that does not have to borrow its strength from late-stage hydrodynamics.

What carries the argument

The load-bearing object is the one-loop two-gluon-one-photon effective vertex $\Gamma^{\mu\nu\alpha}_{ab}(p_1,p_2,q)$ in a constant magnetic field. Expressed in the Ritus polarization basis and constrained by gauge invariance, gluon exchange symmetry, and CP invariance, it collapses for on-shell bosons to three tensor structures with coefficients $a_1^{++}$, $a_2^{++}$, $a_{10}^{++}$; these coefficients are proper-time and Feynman-parameter integrals that are evaluated numerically without assuming a hierarchy between the field strength and the other scales. This vertex drives both production channels through crossing, and the phase-space integrals over the pre-equilibrium gluon distributions, together with the yield-weighted $v_2$ combination, convert it into the photon yield and elliptic flow that are compared with data.

Extended reading notes

Core claim

The central claim is that magnetic-field-induced gluon fusion and splitting during pre-equilibrium, evaluated with the full one-loop vertex and no restriction on the field strength, explain the direct-photon excess once the hydrodynamic contribution is removed. For on-shell gauge bosons, the vertex reduces to three independent tensor structures built from the Ritus polarization vectors, so the squared matrix element is $|a_1^{++}|^2+|a_2^{++}|^2+|a_{10}^{++}|^2$. Integrating this over the gluon phase space with either a glasma-inspired Bose-Einstein occupation factor or an anisotropic distribution, and combining the resulting magnetic yield and $v_2$ with the hydrodynamic direct photons through a yield-weighted average, reproduces the measured yield and elliptic flow; gluon splitting dominates the yield at intermediate photon energies, and the anisotropic distribution with no longitudinal anisotropy gives the best simultaneous description. The arbitrary-field-strength calculation also improves on the earlier strong-field approximation for the same channels.

Load-bearing premise

The whole comparison rests on subtracting the hydrodynamic calculation [47] from the measured direct-photon data and calling the remainder the magnetic gluon signal, even though Ref. [47] already contains some pre-equilibrium photons that may overlap with that signal.

Editorial extensions

If this is right

Gluon splitting is the dominant magnetic channel for the yield at intermediate photon energies ($1.5\lesssim\omega_q\lesssim4$ GeV), so complete pre-equilibrium photon calculations should include both channels rather than only quark-based processes.
The magnetic channels produce a positive elliptic flow whose shape and magnitude follow the data, so photon $v_2$ can be an early-time, magnetic-field effect rather than a purely hydrodynamic signal.
The arbitrary-field-strength vertex reproduces the data better than the earlier strong-field approximation, and the longitudinal anisotropy of the gluon distribution cannot be much larger than $\xi\simeq1$ without driving the magnetic $v_2$ negative.
Both tested gluon distributions give the right order of magnitude, with the anisotropic distribution of Refs. [54, 84] at $\xi=1$ giving the best simultaneous description of yield and $v_2$.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If this mechanism is correct, direct-photon elliptic flow becomes an early-time observable: the same magnetic field that creates the photons also imprints their azimuthal asymmetry, so $v_2$ need not originate in the hydrodynamic expansion phase.
A natural extension would be to relax the constant-field, fixed-volume approximation and let $|eB|$ decay with the fireball lifetime; this would predict a centrality and beam-energy dependence that could be checked against other centrality classes.
The comparison procedure subtracts the full hydrodynamic yield, which already contains pre-equilibrium photon contributions; quantifying how much of that baseline overlaps with the magnetic signal would be the decisive test of whether the added contribution is double counted.
Other magnetized pre-equilibrium photon channels, such as bremsstrahlung, pair annihilation, and anomaly radiation, are not included here, so a combined calculation would test whether the present agreement survives once all channels are summed without double counting.

Original abstract

We compute the yield and elliptic flow coefficient for photons produced during the pre-equilibrium stage of semi-central relativistic heavy-ion collisions from the gluon fusion and splitting processes induced by the presence of a magnetic field. The calculation is performed for arbitrary values of the field strength. The pre-equilibrium gluon distribution is modeled with a glasma-inspired Bose-Einstein occupation factor, as well as with an anisotropic distribution that accounts for the rapid initial expansion along the beam axis. In both cases, we find a very good agreement between the calculation and the experimental data from PHENIX. Most notably, the shape and strength of the elliptic flow coefficient are described quite well, representing a step towards solving the outstanding photon puzzle.

The compiled record

The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.

Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-09-23: 1 request, 4,280 input tokens.

What the abstract itself states, as probabilities: claims first 0.04 · states limitation 0.12 · human participants 0.02 · controlled comparison 0.07 · randomized assignment 0.01 · states quantitative result 0.15 · compares with prior results 0.47 · mentions code or data availability 0.02.

The abstract, sentence by sentence

arXiv listing abstract v1. Results and the main claim are underlined; show every role.

We compute the yield and elliptic flow coefficient for photons produced during the pre-equilibrium stage of semi-central relativistic heavy-ion collisions from the gluon fusion and splitting processes induced by the presence of a magnetic field. The calculation is performed for arbitrary values of the field strength. The pre-equilibrium gluon distribution is modeled with a glasma-inspired Bose-Einstein occupation factor, as well as with an anisotropic distribution that accounts for the rapid initial expansion along the beam axis. In both cases, we find a very good agreement between the calculation and the experimental data from PHENIX. Most notably, the shape and strength of the elliptic flow coefficient are described quite well, representing a step towards solving the outstanding photon puzzle.

background contribution method result interpretation main claim

What the classifier reports for each sentence. The number is the probability it gave the label.
#	sentence	role	the authors' own work	states the main conclusion	strength
s0	We compute the yield and elliptic flow coefficient for photons produced during the pre-equilibrium stage of semi-central relativistic heavy-ion collisions from the gluon fusion and splitting processes induced by the presence of a magnetic field.	contribution 1.00	0.96	0.13	asserted 0.99
s1	The calculation is performed for arbitrary values of the field strength.	method 0.75	0.90	0.04	asserted 0.98
s2	The pre-equilibrium gluon distribution is modeled with a glasma-inspired Bose-Einstein occupation factor, as well as with an anisotropic distribution that accounts for the rapid initial expansion along the beam axis.	method 1.00	0.79	0.03	asserted 0.99
s3	In both cases, we find a very good agreement between the calculation and the experimental data from PHENIX.	result 0.99	0.94	0.35	asserted 1.00
s4	Most notably, the shape and strength of the elliptic flow coefficient are described quite well, representing a step towards solving the outstanding photon puzzle.	interpretation 0.91	0.91	0.65	asserted 0.86

What kind of paper the classifier reports this is: theoretical result 0.62 · evidence from simulation or computation 1.00.

What was examined, and what was not

sources read

arXiv listing abstract v1

not examined

full text
figures and tables
references and citing sentences
relations to other papers

levels run

P abstract

not run

S sections
T body sentences
C citations
R cross-paper relations

Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. Every number in the measurements table is the authors', checked by code. Record as JSON

Figures

**Figure 1.** FIG. 1. General representation of the two-gluon one-photon vertex.

**Figure 2.** FIG. 2. Diagrams representing the one-loop order contribution to

**Figure 3.** FIG. 3. a) Squared matrix element for the fusion process as a function of the photon energy for different values of one of the gluons energy, for

**Figure 4.** FIG. 4. a) Total magnetic contribution to the photon yield (fusion

**Figure 5.** FIG. 5. a) Total magnetic contribution to the photon yield (fusion

**Figure 7.** FIG. 7. a) Photon yield obtained from gluon fusion (blue dotted

**Figure 9.** FIG. 9. a) Total magnetic photon yield compared to the PHENIX data

Showing 7 of the paper's 9 figures. See all figures in the PDF ↗ Figures reproduced from arXiv:2608.21600 by the authors.

References

How the literature uses it

What this paper does with the arXiv papers it cites

Each citing sentence is taken from the citing paper's LaTeX source and typed by Pith's classifier. A label appears at probability 0.8 or more (0.95 for support), a contest at 0.7 or more; the quoted words are the citing authors'. How each label was checked · JSON

What it cites · 45 sentences without a confident label

Contests 3 papers

These channels were studied in Refs. [ref] under the strong magnetic field approximation, which is only valid for the low transverse momentum region of the spectrum, and using a vertex function [ref] that did not take into account the complete tensor structure and the symmetry properties expected for these processes.

Centrality dependence of photon yield and elliptic flow from gluon fusion and splitting induced by magnetic fields in relativistic heavy-ion collisions · Introduction · contest 0.79
These channels were studied in Refs. [ref] under the strong magnetic field approximation, which is only valid for the low transverse momentum region of the spectrum, and using a vertex function [ref] that did not take into account the complete tensor structure and the symmetry properties expected for these processes.

Anisotropic photon emission from gluon fusion and splitting in a strong magnetic background I: The two-gluon one-photon vertex · Introduction · contest 0.79
These channels were studied in Refs. [ref] under the strong magnetic field approximation, which is only valid for the low transverse momentum region of the spectrum, and using a vertex function [ref] that did not take into account the complete tensor structure and the symmetry properties expected for these processes.

Prompt photon yield and elliptic flow from gluon fusion induced by magnetic fields in relativistic heavy-ion collisions · Introduction · contest 0.79

Uses the method or tool of 5 papers

For the gluon distribution, one can consider, as in Ref. [ref] , an isotropic Bose-Einstein distribution inspired by the kind of occupation number that arises after the glasma is shattered during the first stages of the reaction [equation] where $η$ represents the gluon occupation factor, and $Λ_s$ is the saturation momentum scale.

Prompt photon yield and elliptic flow from gluon fusion induced by magnetic fields in relativistic heavy-ion collisions · Photon yield and v2 · 0.87
This effective vertex is written in terms of the tensor basis developed in Ref. [ref] , which takes into account all its symmetry properties and, for on-shell bosons, reduces to just three independent tensor structures.

Two-gluon one-photon vertex in a magnetic field and its explicit one-loop approximation in the intermediate field strength regime · Summary and conclusions · 0.96
To numerically compute the yield coming from the magnetic field processes, we interpolate the calculated matrix element using the algorithm of Refs. [ref] and evaluate numerically the remaining integrals using the Suave algorithm [ref] .

Concurrent Cuba · Photon yield and v2 · 1.00

Uses data or values from 3 papers

The direct photon contribution of Ref. [ref] is taken as the total contribution, that takes into account prompt, thermal and pre-equilibrium photons.

Multi-messenger heavy-ion collision physics · Photon yield and v2 · 0.86
In Refs. [ref] , the anisotropy coefficient is taken to be of the order of $ξ= 10$.

Pre-equilibrium photons from the early stages of heavy-ion collisions · Photon yield and v2 · 0.93
In Refs. [ref] , the anisotropy coefficient is taken to be of the order of $ξ= 10$.

Isotropization and hydrodynamization in weakly coupled heavy-ion collisions · Photon yield and v2 · 0.93

Compares with 1 paper

The obtained yield and elliptic flow coefficient in both cases give a better description of PHENIX data than the findings of Ref. [ref] , where a strong magnetic field approximation was used.

Prompt photon yield and elliptic flow from gluon fusion induced by magnetic fields in relativistic heavy-ion collisions · Photon yield and v2 · 0.97

Cites as background 51 papers

The work is organized as follows: for the completeness of the discussion, in Sec. [ref] we summarize the findings of Ref. [ref] for the general structure of the two-gluon one-photon vertex in the presence of a constant magnetic field.

Two-gluon one-photon vertex in a magnetic field and its explicit one-loop approximation in the intermediate field strength regime · Introduction · 0.91
To address this puzzle, several approaches have been proposed, including hydrodynamical scenarios [ref] , radiation emitted during the pre-equilibrium stage [ref] , transport models [ref] , intermediate semi-QGP states [ref] , and others, as reviewed in Refs. [ref] .

Event-by-event fluctuations in perturbative QCD + saturation + hydro model: pinning down QCD matter shear viscosity in ultrarelativistic heavy-ion collisions · Introduction · 0.95
To address this puzzle, several approaches have been proposed, including hydrodynamical scenarios [ref] , radiation emitted during the pre-equilibrium stage [ref] , transport models [ref] , intermediate semi-QGP states [ref] , and others, as reviewed in Refs. [ref] .

Photon production and elliptic flow from momentum-anisotropic quark-gluon plasma · Introduction · 0.95

Discussion

Disagree with the review, or know something it missed? Leave a comment, highlight any passage above to comment on it, or challenge this review formally.