{"id":"8b62f6bf-741a-451d-b3cf-d3b4cc5b14b1","arxiv_id":"2411.14366","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"First measurement of low-pT direct photons in 13 TeV pp collisions, showing a 3.2 sigma signal in inelastic and a 1.9 sigma hint in high-multiplicity events, with yield rising with multiplicity.","lead":"ALICE reports the first low-transverse-momentum direct-photon measurement in proton-proton collisions at 13 TeV, using virtual photons decaying to electron pairs. The signal is seen with 3.2 sigma in inelastic events and 1.9 sigma in high-multiplicity events, and the yield grows with charged-particle multiplicity.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The direct-photon claim hinges on an unbiased η-dominated dielectron cocktail; with LF systematics of 16–31% on r and only a 1.9σ HM signal, a closure test on the η template is needed before 'observed in both classes' can stand.","rationale":"The reader's verdict correctly identifies the η-driven cocktail as the weakest assumption, and I agree. The paper's improvement over Ref. [46] is real: the π0 and η input spectra from Ref. [47] reduce the previously dominant uncertainty, and the total systematic on the dielectron yield is improved to 6.5–12%. However, the residual light-flavour cocktail uncertainty on r is still 16–31%, so the direct-photon signal is not yet self-evident. The fit χ2/ndf values (5.9/8, 7.1/8, etc.) show statistical consistency but do not strongly constrain the η normalization, because the same dielectron data are used to normalise the LF template below 0.14 GeV/c2; a correlated error in the η/π0 ratio can migrate into r. The significance calculation is also not documented at the level of a covariance matrix, and the 1.9σ HM point cannot support the word 'observed'; this language should be corrected to 'evidence' or 'hint'. These issues are addressable with additional cross-checks and text revisions, so conditional acceptance, as the reader recommended, remains the appropriate outcome.","tokens_in":31170,"tokens_out":12358,"duration_ms":135848,"concrete_test":"Perform a closure test in the 2–3 GeV/c pT interval: refit the dielectron mass spectrum with r fixed to 0 and the η normalization floated, using a Gaussian penalty from the Ref. [47] η cross-section uncertainty. If the best-fit η yield is more than ~2σ below Ref. [47], or if a fit with r free prefers an η yield shifted by more than the quoted LF systematic, the direct-photon signal is not robust. Equivalently, re-derive r using an m_T-scaled π0-based η estimate instead of the Ref. [47] parameterisation and compare the resulting shift in r with the 0.0254 central value in that bin.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central extraction of the direct-photon fraction r is a one-parameter fit in 0.14 < m_ee < 0.35 GeV/c2, where the light-flavour template is dominated by η Dalitz decays. The η contribution is taken from Ref. [47] and parameterised with a modified Hagedorn function; Tables A.2/A.3 assign the light-flavour cocktail a 16–31% (INEL) and 25–31% (HM) uncertainty on r, despite the text citing only 2–3% uncertainty on the η/π0 ratio. If the η yield or its parameterisation is biased within the allowed freedom, the bias enters r approximately linearly and can mimic a real direct-photon excess. Because the direct-photon yield (Eq. 1) and the multiplicity-scaling exponent are derived from r, this is a direct threat to the central claim, not merely a wording issue. The quoted 3.2σ and 1.9σ significances do not test this: they are calculated with systematic uncertainties treated as fully correlated, and no closure check is presented that would distinguish a true excess from an η-template misnormalisation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This letter reports the first ALICE measurement of direct photons at low transverse momentum (1 < pT < 6 GeV/c, |η| < 0.8) in inelastic (INEL) and high-multiplicity (0–0.072% INEL) proton–proton collisions at √s = 13 TeV, using the virtual-photon method. The direct-photon fraction r is extracted by fitting the dielectron invariant-mass spectrum in 0.14 < mee < 0.35 GeV/c² with a Kroll–Wada direct-photon shape, a light-flavour hadron-decay cocktail, and a fixed heavy-flavour contribution. The paper reports significances of 3.2σ (INEL) and 1.9σ (HM), compares r and the direct-photon yield with NLO pQCD calculations, and fits the integrated 1 < pT < 3 GeV/c yield versus charged-particle multiplicity to a power law with exponent α = 1.17 ± 0.18 (stat) ± 0.21 (syst). The INEL result is found consistent with the pQCD calculations; the HM data are described only after applying empirical pT-independent scaling factors of 3.4–6.6 or 7 to the prompt-photon calculations.","tokens_in":31455,"tokens_out":9190,"duration_ms":85715,"significance":"The measurement is valuable: it uses the same-event-class π0 and η spectra from Ref. [47], roughly four times the statistics of the previous analysis [46], and it provides the first low-pT virtual-photon direct-photon signal in pp collisions at LHC energies. The analysis is documented with fit figures and systematic tables that allow the extraction to be followed in detail, and the use of independent meson measurements and of a one-parameter fit for r are clear strengths. If the inelastic signal holds, it offers a useful low-pT reference for pQCD and for searches for thermal radiation in small systems. However, the strength of the headline claims exceeds the current statistical evidence: the HM signal is only 1.9σ, and the quoted multiplicity-scaling result does not establish a significant super-linear increase.","major_comments":[{"comment":"The abstract and Sec. 4 state that a direct-photon signal is 'observed' in both INEL and HM event classes with significances of 3.2σ and 1.9σ. Under standard HEP conventions a 1.9σ excess is not an observation, and even 3.2σ is usually termed evidence rather than observation. The conclusion that 'the direct photon signal is seen in both event classes' therefore overstates the HM result. Please revise the abstract, Sec. 4, and the summary to use 'evidence' for the INEL class and 'hint' or 'no significant excess' for the HM class unless additional analysis can raise the HM significance.","section":"Abstract; Sec. 4 (Results)"},{"comment":"The text says that the dominant systematic uncertainty on r comes from the η cocktail parameterisation and that the η/π0 ratio uncertainty is only 2–3%, yet Tables A.2 and A.3 assign the light-flavour cocktail an uncertainty on r of 16.3–31% (INEL) and 25–30.7% (HM). Since the light-flavour template in the fit function dσ/dm_ee = r f_dir + (1−r) f_LF + f_HF is dominated by η Dalitz decays in the fitting range, a 2–3% η/π0 uncertainty cannot by itself produce the quoted LF systematic. Please state exactly which variations enter the LF systematic and provide a closure or stress test (for example, an alternative η spectral parameterisation or a fit in a mass window below the η peak) demonstrating that r is not driven by an η-template misnormalisation.","section":"Sec. 3 and Tables A.2/A.3"},{"comment":"The abstract claims a 'significant increase of direct-photon yield with charged-particle multiplicity' based on α = 1.17 ± 0.18 (stat) ± 0.21 (syst). With only two ALICE data points, the fit is equivalent to taking a ratio, and the combined uncertainty is sqrt(0.18² + 0.21²) ≈ 0.28; the deviation from linear scaling (α = 1) is 0.17 ± 0.28, i.e. below 1σ. The paper should either present a fit that includes additional data or explicitly state that the data are consistent with, but do not establish, a super-linear increase.","section":"Sec. 4 and Fig. 4"}],"minor_comments":[{"comment":"The last paragraph of Sec. 1 contains a typo: 'midrapitidy' should be 'midrapidity'.","section":"Sec. 1"},{"comment":"The statement that the unscaled pQCD models are disfavoured by a 'p-value less than 0.1' is not accompanied by a definition of the test statistic or of how systematic uncertainties were included; please provide this information.","section":"Sec. 4 and Fig. 2 (right)"},{"comment":"The comparison in Fig. 4 includes data at different centre-of-mass energies and different pT integration ranges (1–3 GeV/c for the present data versus 1–5 GeV/c for several other points); this should be stated explicitly in the text to avoid over-interpreting the figure as a universal scaling.","section":"Fig. 4"},{"comment":"The entries 'Normalization range' and 'Fitting range' are not self-explanatory; a sentence or footnote defining the exact variations used for those uncertainties would improve readability.","section":"Tables A.2 and A.3"}],"recommendation":"major_revision","confidential_remarks":"I see no reason to question the integrity of the measurement; the concerns are about calibration of the claims relative to the evidence. The requested revisions are substantive but local: soften the 'observation' language for the HM class, clarify and stress-test the η cocktail systematics, and revise the multiplicity-scaling statement. A resubmission after major revision would be appropriate for this journal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing you should know: this is the first low-pT direct-photon measurement in pp at LHC energies, and that's a real step forward. The inelastic result, 3.2σ, is evidence, and the consistency with pQCD is a useful check. But the abstract says 'observed in both event classes' while the HM significance is 1.9σ. That's not observation; that's a hint. The paper should either drop the observation language for HM or quote an upper limit.\n\nWhat the paper does well: it uses four times the statistics of the previous ALICE analysis, and the π0/η yields now come from independent measurements in the same event classes at the same energy (Ref [47]). That cuts the cocktail uncertainty versus the old 50% ballpark, and the fit to the dielectron mass spectrum follows the established Kroll-Wada approach; nothing circular here. The integrated-yield scaling exponent α = 1.17 ± 0.18 ± 0.21 is consistent with the PHENIX value, which is worth noting for small-system physics.\n\nThe soft spots are not fatal, but they need attention.\n\nFirst, the systematics on r are huge: up to 35% (INEL) and 48% (HM). The dominant term is the light-flavour cocktail, specifically the η parameterisation, yet the text says the η/π0 ratio uncertainty is only 2–3%. Those numbers don't obviously hang together. Either the 2–3% is just the ratio and the 16–31% LF term includes other sources (which should be stated explicitly), or the propagation is sensitive to shape and normalisation in a way that deserves a chart. A closure test on the η template in a control region would help a lot. Without it, a small η bias translates directly into r, and the 3.2σ significance – computed with systematics treated as fully correlated – is less reassuring than it looks.\n\nSecond, the HM comparison to theory relies on scaling factors (3.4–6.6 and 7) that are empirical fits, not predictions. The paper acknowledges this, but the conclusion that HM needs 'an additional source of direct photons' is not strongly supported. It's a plausible hint, not a result.\n\nThird, there are no data tables or HEPData deposit. For a measurement paper, that's a reproducibility gap; the referee should ask for it.\n\nOverall: this paper is for the small-systems and direct-photon community, and it deserves a serious referee. The measurement is new, the analysis is careful, and the inelastic result is likely to stand. But the HM claim and the η systematic need to be tightened, the language tempered, and the data released. Send it to peer review with those requests.","headline":"First low-pT direct-photon measurement in pp at 13 TeV, but 'observed in both classes' overstates a 1.9σ HM signal; the η-dominated cocktail systematics and missing closure test need attention.","tokens_in":31958,"tokens_out":4590,"would_cite":true,"duration_ms":40885,"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":"This paper reports the first low-transverse-momentum direct-photon signal in 13 TeV proton-proton collisions, with 3.2σ and 1.9σ significance in inelastic and high-multiplicity events, and a charged-particle-multiplicity scaling exponent…","keywords":["direct photons","virtual photons","dielectron pairs","Kroll-Wada formula","proton-proton collisions","high multiplicity","charged-particle multiplicity scaling","low transverse momentum"],"falsifier":"Measure the direct-photon fraction independently from real photons (calorimeter or external conversion) in the same $1 < p_{\\rm T} < 6$ GeV/$c$ range in 13 TeV proton-proton collisions and compare with the virtual-photon value; a disagreement beyond the quoted uncertainties would falsify the zero-mass extrapolation or the hadron-cocktail decomposition.","tokens_in":30949,"feed_emoji":"⚛️","tokens_out":6463,"duration_ms":58306,"temperature":0.7,"pith_summary":"This paper tries to establish that direct photons—photons not coming from hadron decays—are produced at low transverse momentum in ordinary proton-proton collisions at 13 TeV, and that their yield grows with the number of charged particles produced. Using the virtual-photon method, it reports the first such signal at LHC energies in both inelastic and high-multiplicity event classes, with significances of 3.2σ and 1.9σ. The integrated yield in the range $1 < p_{\\rm T} < 3$ GeV/$c$ rises with charged-particle multiplicity as a power law with exponent $\\alpha = 1.17 \\pm 0.18 (\\text{stat.}) \\pm 0.21 (\\text{syst.})$. If correct, the result shows that high-multiplicity proton-proton collisions produce more direct photons than prompt perturbative-QCD expectations alone, opening the low-momentum window to searches for thermal radiation in small collision systems.","feed_headline":"Direct photons observed at low momentum in proton-proton smashes","feed_subtitle":"First such signal in inelastic and high-multiplicity events; yield grows with charged multiplicity.","key_machinery":"The central object is the Kroll-Wada relation, which links the yield of real photons to the yield of dielectron pairs they internally convert into; for pair transverse momentum much larger than the pair mass, the dielectron spectrum behaves as $\\sim 1/m_{ee}$. This relation supplies the shape of the virtual direct-photon component $f_{\\rm dir}$ and justifies taking the direct-photon fraction $r$ measured at small non-zero mass to be the real-photon fraction in the $m_{ee} \\to 0$ limit. The extraction fits the dielectron cross section with $d\\sigma/dm_{ee} = r\\, f_{\\rm dir}(m_{ee}) + (1-r)\\, f_{\\rm LF}(m_{ee}) + f_{\\rm HF}(m_{ee})$, where $f_{\\rm LF}$ and $f_{\\rm HF}$ are the light- and heavy-flavour hadronic decay templates and $r$ is the only free parameter. The crucial input is the hadronic cocktail, especially the measured $\\pi^0$ and $\\eta$ spectra in the same event classes, which dominates the systematic uncertainty on $r$.","core_discovery":"The paper reports the first observation at LHC energies of a low-transverse-momentum direct-photon signal in proton-proton collisions, seen in both inelastic and high-multiplicity event classes. For $1 < p_{\\rm T} < 6$ GeV/$c$ at midrapidity $|\\eta| < 0.8$, the fraction $r$ of direct to inclusive photons is extracted from a fit to the dielectron invariant-mass spectrum in the range $0.14 < m_{ee} < 0.35$ GeV/$c^2$, where the hadronic-decay background is dominated by $\\pi^0$ and $\\eta$ decays. The signal is observed with 3.2σ significance in inelastic events and 1.9σ in high-multiplicity events, and the direct-photon fraction is compatible between the two event classes. The inelastic results agree with next-to-leading-order perturbative QCD calculations within uncertainties, while in high-multiplicity events the prompt-photon predictions alone do not describe the data; the best description is obtained by scaling the inelastic prompt contribution by a factor of 3.4–6.6 and adding a calculated thermal-photon component. The integrated direct-photon yield in $1 < p_{\\rm T} < 3$ GeV/$c$ increases with charged-particle multiplicity with an exponent $\\alpha = 1.17 \\pm 0.18 \\pm 0.21$.","pith_inferences":["If the high-multiplicity excess survives higher-statistics data, it would strengthen the case that small collision systems can produce thermal-like radiation, although the paper itself does not claim a thermal-photon observation.","The measured $\\alpha \\approx 1.17$ provides a concrete baseline for proton-lead collisions: a testable extension would be to measure the same exponent in p–Pb events, where a value between the proton-proton and lead-lead results would indicate a smooth system-size interpolation.","Because the $\\eta$ yield carries the largest systematic uncertainty on $r$, an independent high-precision measurement of the $\\eta/\\pi^0$ ratio at 13 TeV, or an alternative extraction of $r$ using a different mass window, would directly sharpen the result.","Run 3 and Run 4 data should resolve whether the high-multiplicity excess is a momentum-independent enhancement of prompt photons or a genuinely new component, since the current signal significance in high-multiplicity events is only 1.9σ."],"forward_implications":["Direct photons at low transverse momentum in proton-proton collisions can be measured with the virtual-photon method, providing a new experimental handle in a momentum region where perturbative QCD is difficult.","The inelastic direct-photon fraction and yield are consistent with next-to-leading-order perturbative QCD predictions within uncertainties, so conventional prompt-photon physics describes minimum-bias 13 TeV proton-proton collisions.","In high-multiplicity collisions, prompt-photon-only calculations fail to describe the data, so an additional direct-photon source or an enhanced prompt yield is needed.","The integrated direct-photon yield rises faster than linearly with charged-particle multiplicity, with $\\alpha \\approx 1.17$, in agreement with the trend seen in heavy-ion collisions at lower energies.","The direct-photon fraction itself is similar in inelastic and high-multiplicity events, meaning the multiplicity dependence appears mainly in the overall yield rather than in the fractional composition of the photon field."],"supporting_citations":[{"why":"Establishes the Kroll-Wada relation between real-photon yield and internal-conversion dielectron pairs, the basis for extracting the direct-photon fraction from the mass spectrum.","marker":"[44, 45]"},{"why":"Provides the dielectron analysis technique, hadronic cocktail construction, and systematic framework used here in the same collision systems.","marker":"[46]"},{"why":"Supplies the measured $\\pi^0$ and $\\eta$ spectra in the same 13 TeV event classes, the input that sharply reduces the dominant cocktail uncertainty.","marker":"[47]"},{"why":"Provides the next-to-leading-order perturbative QCD prompt-photon prediction to which the inelastic direct-photon fraction and yield are compared.","marker":"[18]"},{"why":"Provides prompt and thermal photon calculations for small systems, used to interpret the high-multiplicity excess and the empirical scaling factors.","marker":"[35]"},{"why":"Documents the approach of extracting the direct-photon fraction from a fit to dielectron spectra with a hadronic cocktail.","marker":"[67]"}],"fun_headline_variants":["First low-momentum direct photons in pp collisions at LHC","Direct photons at low pT: first signal in LHC pp collisions","Yield of direct photons rises with multiplicity in pp","First direct-photon signal at low momentum in LHC pp","LHC shows direct photons at low momentum in pp collisions"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The result stands on the assumption that the fit to electron-positron pairs correctly splits the spectrum into decays of ordinary particles (mostly pions and etas) plus direct photons, and that the direct-photon share found there is the same as for real, massless photons.","fun_headline_variants_meta":{"raw":{"variants":["First low-momentum direct photons in pp collisions at LHC","Direct photons at low pT: first signal in LHC pp collisions","Yield of direct photons rises with multiplicity in pp","First direct-photon signal at low momentum in LHC pp","LHC shows direct photons at low momentum in pp collisions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000306,"raw_usage":{"total_tokens":1835,"prompt_tokens":1111,"completion_tokens":724,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":727,"completion_tokens_details":{"reasoning_tokens":639}},"tokens_in":727,"tokens_out":724,"duration_ms":6574,"temperature":1.0,"reasoning_tokens":639,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T15:15:35.671646+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the direct-photon fraction independently from real photons (calorimeter or external conversion) in the same $1 < p_{\\rm T} < 6$ GeV/$c$ range in 13 TeV proton-proton collisions and compare with the virtual-photon value; a disagreement beyond the quoted uncertainties would falsify the zero-mass extrapolation or the hadron-cocktail decomposition.","supporting_citations":[{"cited_title":"Polarized and unpolarized prompt photon production beyond the leading order","cited_arxiv_id":null,"evidence_quote":"Provides the next-to-leading-order perturbative QCD prompt-photon prediction to which the inelastic direct-photon fraction and yield are compared."}],"review_version":1}