WorksPhysics and Astronomy
Tracing Vacuum Hadronization with Conserved Currents
Abstract
We show that color triality constrains the nonperturbative states that screen a Wilson-line endpoint, allowing the joint flows of net electric charge, baryon number, and strangeness to probe QCD vacuum hadronization. Whether evaluated from resolved hadrons in a jet initiated by a quark of flavor or from the corresponding charge correlators, these flows satisfy the Gell-Mann–Nishijima relation , where is the third component of the initiating-quark isospin. The net jet baryon number, , directly probes the relative probability of a diquark–antidiquark vacuum excitation, with the diquark-to-quark production ratio. Thus, for , the baryon number carried by the jet is substantially suppressed relative to the initiating-quark value . Likewise, the strange-to-light pair-production ratio, defined by , is encoded in the measured jet strangeness , predicting a nonzero mean net strangeness even in - and -initiated jets. The conserved-current moments appearing in these relations are independent of the renormalization scale. Their simultaneous measurement therefore provides a direct, flavor-resolved probe of vacuum pair production and quantum-number transport during hadronization.
Introduction
Colored partons produced in a QCD hard process [3] [1, 2] must ultimately form color-singlet hadrons through hadronization, long-distance QCD dynamics that remains poorly understood [4]. Fragmentation functions provide the standard description of inclusive identified-hadron production [5], but do not directly reveal how color is screened, flavor is produced, or baryons are formed. Phenomenological string and cluster hadronization models implement these microscopic mechanisms through explicit color-screening degrees of freedom [6, 7], whose production is controlled by parameters such as the relative weights of to light-quark pairs and of diquark–antidiquark to quark–antiquark pairs [8, 9, 10]. Direct experimental constraints on these hadronization parameters remain challenging, hindering discrimination among hadronization mechanisms.
Here we show that joint measurements of the net electric charge , baryon number , and strangeness flows provide a direct probe of microscopic hadronization. For any conserved charge , we denote its flow into a region as , where can be a jet, a hemisphere, or a collinear sector selected by a global event shape. Unlike the momentum-weighted jet charge [11, 12], which is sensitive to momentum sharing, only measures the net charge and its moment is infrared safe [13, 14, 15]. Once its calculable perturbative effects are separated, any shift from the initiating-parton charge therefore isolates nonperturbative charge transfer during hadronization.
The factorization framework makes this separation precise. A gauge-invariant collinear sector, such as the fragmentation function or a jet, is completed by a Wilson line extending to infinity which potentially prevents the physical hadrons assigned to that sector from forming a closed system with respect to global quantum numbers. Indeed, a quark that initiates the sector carries fractional charges, whereas every asymptotic hadron carries integer quantum numbers. Collins and Rogers [16] proposed that the missing quantum numbers reside in a nonperturbative state bound to the Wilson line at vanishing momentum fraction.
In this note, we formulate this state as a defect sector () carrying boundary charges at the Wilson-line puncture. Since the defect charges vanish perturbatively, the shifts from the initiating-parton charges directly measure the quantum numbers carried by the nonperturbative state. By recognizing the color-triality [17] selection rule imposed by confinement, we turn this picture into a predictive framework for classifying the allowed nonperturbative states and their global quantum numbers . Although the Wilson line is itself neutral under the global symmetries, its endpoint carries nonzero color triality and must be screened by dynamical matter of conjugate triality, sharply restricting the quark content and the quantum numbers of the screening state.
On this basis, we establish invertible relations between the joint , , and shifts and the underlying vacuum-hadronization probabilities. The net baryon flow determines the vacuum diquark-to-quark excitation ratio , while the joint flows determine the strange-to-light ratio . These relations predict a nearly vanishing net jet baryon number for , despite of the initiating quark, and a nonzero net strangeness in - and -initiated jets. More generally, QCD isospin symmetry of the screening state yields a jet-level Gell-Mann–Nishijima relation among the conserved flows, providing a closure test that requires fewer assumptions than the probability extractions. Pythia and Herwig simulations reproduce the predicted correlations within their respective hadronization models and allow us to extract effective strange- and baryonic screening parameters. This demonstrates that the charge shifts provide information complementary to inclusive hadron-yield fits.
The remainder of this note is organized as follows. We first define the charge-flow operator associated with a collinear sector and the complementary defect charge required by charge conservation. We then show how the Wilson line completing the collinear sector produces a puncture on the celestial sphere, thereby providing a geometric interpretation of the defect sector. Applying the color-triality selection rule together with a minimal, generic assumption about color screening, we classify the electric charge, strangeness, and baryon number that can be carried by the screening state and derive the resulting nonperturbative shifts to the observed charges, which are renormalization-scale independent at leading power. These shifts are inverted to extract the vacuum-screening probabilities and , leading to the jet Gell-Mann–Nishijima closure relation. Using Pythia 8 and Herwig 7.3, we validate the jet Gell-Mann–Nishijima relation and the extraction of and . The appendices extend these tests to correlators.
Brief Review of Factorization
For brevity, we refer to both a jet with its net electric charge measured and a charge-flow correlation event shape as a charge-measured jet. Throughout the note, we assume that the jet is sufficiently inclusive to contain the entire collinear radiation. The cross section factorizes schematically as
where are parton distribution functions for hadronic collisions and when a hadron beam is replaced by an electron beam. The partonic cross section itself admits a refactorization into the hard, soft functions and collinear matching coefficient associated with the -quark jet. The charge-measured jet function is, at leading power [18],
Here is the -collinear quark field of flavor that initiates the jet, is the collinear Wilson line extending in the conjugate direction, and measures the net electric charge carried by resolved hadrons of the jet.
Collinear Charge Flow Operator
At leading power6,
where , is the electric charge-flow operator [14, 19, 20], with the electromagnetic current and the direction specified by . Here is a small angular disk surrounding the antipodal direction with opening angle . The -collinear representation uses the stereographic coordinate [18], where denotes an -collinear momentum. This coordinate covers except the point , becoming singular as . The charge-flow detector is therefore defined on , while the excluded point is the celestial endpoint of the Wilson line.
Note that the limit is taken only after separating any puncture-supported contribution from the charge measured by the detector; therefore, the charge flow operator need not measure the total net charge of the collinear sector. A fraction of the charge quantum number may be carried away by degrees of freedom localized at the excluded puncture , and hence
where is the conserved total electric charge of the -collinear sector, and denotes the mean charge localized at the puncture that is missed by the charge flow measurement.
Baryon Number and Strangeness Measurement
An identical construction applies to , , as well as to and , with replaced by the baryon–flow operator and the electromagnetic current by the conserved baryon current, or the corresponding operators for strangeness. For definiteness, we formulate the following discussion in terms of electric charge. All statements carry over to baryon number and strangeness under these replacements.
Collinear Wilson Line and the Defect Sector
We now discuss a possible mechanism for generating a nonzero or by examining the degrees of freedom associated with the Wilson-line endpoint. Since the excluded puncture is the celestial endpoint of the collinear Wilson line, quantum numbers not measured by the detector may be carried by its associated line-defect sector.
The future-pointing collinear Wilson line represents an external eikonal color source and therefore defines a line defect [21, 22] of the -collinear theory and its associated defect sector. The line defect extends geometrically along the conjugate direction and terminates at future null infinity on the celestial sphere, as illustrated in Figure 1. The defect sector consists of the defect line together with the QCD excitations that can dress it. The sector can be made explicit by representing as the propagator of an auxiliary field , with action . Together with its QCD dressings, e.g., the charged operators , , , they realize the defect sector.
Figure 1. The coverage of in the collinear sector. The collinear Wilson line defines a line defect and its associated defect sector in the -collinear theory, leaving an area on the celestial sphere of the collinear theory that is not covered by the charge flow detector. The purple insertion schematically denotes charged defect operators generated through the OPE of a quark approaching the Wilson line, . The operators contribute to the puncture charge . Consequently, the charge detector , which covers of the collinear theory, measures only the complementary charge . At leading power, and are scale independent.
Since the puncture location corresponds to , every ordinary emission with lies at a finite and is eventually included in . Consequently, a finite requires charge-weighted support localized at that survives the shrinking-disk limit; any locally integrable contribution around the puncture instead vanishes with . In the factorized -collinear sector, the Wilson line is the only operator extending to this point, so such endpoint support can arise from excitations that dress or bind to the Wilson line.
In perturbation theory, pure virtual corrections cannot change the average charge leaking through the puncture, since all QCD interaction insertions are electrically neutral. Real radiation can accumulate at , for instance, via eikonal singular couplings to gluons, and through , it can bring an anti-quark field onto the Wilson line to form a charged defect operator, and thereby leave the charge flow detector, see Figure 1. The associated charge-weighted contribution cancels, since for every configuration in which a soft quark enters the puncture, there is an equally weighted configuration in which the corresponding antiquark enters it [20]. More precisely, in perturbation theory, charge weighting projects the radiation onto the charge-odd flavor-nonsinglet channel, whose Regge behavior has no nonintegrable singularity [23, 24]. Hence the charge over a small range vanishes as , so perturbative radiation cannot deposit finite charge around . Consequently,
at leading power.
Vacuum Hadronization and the Jet Gell-Mann–Nishijima Relation
Nevertheless, non-perturbatively, hadronization might generate strong correlations between the soft quark distribution and the charged collinear endpoint , see Figure 2 for one example. The conditional distribution in a fixed quark sector need not retain the symmetry responsible for the perturbative cancellation. This does not imply that charge conjugation is broken by QCD, since charge conjugation maps the complete quark sector into the corresponding antiquark sector and requires only
Figure 2. Hadronization may induce a long-range correlation between the initiating quark, shown in black, and the ordering of soft quark-antiquark pairs along the fragmentation chain, shown in red. In the meson-dominated configurations illustrated in [16], the initiating quark can combine with a neighboring , as indicated by the solid green ellipse, whereas the reversed ordering with a neighboring is suppressed, as indicated by the dashed ellipse. The two orderings leave, respectively, a and a onto the Wilson line, contributing and to charges carried by charged defect operators. A probability imbalance between these configurations therefore produces a nonvanishing average dominated by , and analogously a nonvanishing boundary baryon number .
The puncture can therefore acquire a non-vanishing non-perturbative charge expectation value,
The value of , whether 0 or not, is determined by the QCD non-perturbative dynamics, such as the string fragmentation mechanism [6, 7, 16]. Thus the total charge of the sector remains exactly , while the bulk detector measures only the complementary charge . In this sense, characterizes a non-perturbative boundary charge assigned to the celestial puncture generated by the defect sector of the future-pointing collinear Wilson line. Its existence represents a possible non-perturbative redistribution of the conserved charge between the observable flow through and the puncture sector . If the QCD non-perturbative dynamics respects light-flavor symmetry and loses correlation with the initiating flavor, see Figure 2, one would expect that
More precisely, the relation states that in a Monte Carlo simulation, one should expect to see that the charge difference measured over a -jet and -jet should be identical before or after performing hadronization.
We note that the renormalization-group evolution of and is determined by perturbative ultraviolet dynamics. In this domain, Eq. (5) gives , so inherits the vanishing anomalous dimension of the conserved total charge, which then implies that the nonperturbative puncture contribution is also renormalization-scale independent.
Similar discussion applies to the baryon number and other conserved quantum numbers in QCD.
Independently of the particular hadronization configuration illustrated in Figure 2, confinement imposes a general selection rule on the puncture sector. The anti-fundamental Wilson-line endpoint must be screened by dynamical matter carrying the conjugate color triality. Although the open color index of the Wilson line is contracted with its conjugate in the jet function in Eq. (2), this contraction only makes the squared matrix element color invariant. It does not provide a physical screening excitation in an individual hadronization amplitude. After hadronization, the finite-angle bulk consists of color-singlet hadrons and therefore cannot terminate the nonzero-triality color flux carried by the anti-fundamental Wilson-line endpoint. The endpoint must instead be screened by dynamical quarks and antiquarks bound to it. We call the correlated set of dynamical quarks and antiquarks remaining attached to the Wilson-line endpoint at its screening cloud. If this cloud contains endpoint quarks and endpoint antiquarks, color triality [17] requires
Here the is the color triality of the Wilson line , while each quark (antiquark) contributes (). Writing for the net number of quarks of flavor in the screening cloud, the leading screening possibilities and their associated puncture and flow quantum numbers are
| screening cloud | |||||
| higher clouds |
Table 1.
Here is the electric charge of flavor .
Since additional quark–antiquark excitations from the vacuum incur an energetic cost and are therefore disfavored, we conjecture that the puncture is dominated by the minimal one-quark screening cloud, with the anti-diquark cloud providing the leading correction. Define the conditional endpoint ratio
Here denotes the probability of producing a screening cloud. Assuming that vacuum screening probabilities are homogeneous over the celestial sphere, we identify the conditional endpoint ratio with the final vacuum diquark-to-quark screening ratio .
Neglecting higher screening configurations that are generically suppressed, we then obtain
Thus the measured jet baryon number directly probes the probability for anti-diquark screening of the Wilson-line puncture. For small , we expect substantially suppressed measured inside a -quark jet, compared with its naive expectation .
For the electric charge, let the single screening quark be distributed as
Its mean quark charge is . If the two antiquark flavors in the antidiquark cloud are drawn independently from the same distribution, then , and therefore
which immediately gives
Note that by a similar mechanism, can be probed by the strange number through
where the -quark carries . Eqs. (15) and (16) predict linear relations among , , and . We denote their slopes as , and .
Eventually, we have
which is the Gell-Mann–Nishijima relation [25, 26, 27] for the measured jet. Here, is the third component of isospin for the initiating quark. Eq. (17) should be distinguished from the exact measured-flow identity , which follows directly by summing the particle-level Gell-Mann–Nishijima relation over the measured region [13]. The nontrivial dynamical content of Eq. (17) is the initiating-isospin retention relation . Isospin symmetry of the screening cloud makes the contributions from - and -screening configurations cancel on average, yielding and hence the stated retention relation.
These relations provide a way to study non-perturbative strangeness and baryon production in QCD using jet measurements and, within the hadronization framework introduced here, to constrain the parameters and (or, approximately, ). In the next section, we validate these relations using event generators.
Validation against Monte Carlo Simulations
We validate our predictions against Monte Carlo simulations using Pythia 8.312 [8] and Herwig 7.3 [28], which employ different hadronization models. Pythia uses the Lund string model [6], whereas Herwig implements the cluster hadronization model [7]. We consider annihilation and impose a thrust cut, , to select the dijet limit in which the factorization theorem in Eq. (1) applies. Within each selected event, we then measure the conserved quantum numbers flowing into a given hemisphere. We emphasize, however, that our result is not tied to the thrust observable. It applies more generally to any process and measurement, e.g., jets, jet shapes and correlators, for which the factorization theorem in Eq. (1) is valid.
Figure 3 shows the quantum-number flow at , 100, 200, 500, and 1000 GeV measured in PYTHIA 8.312 simulations, restricting the hard process to , , and production. The requirement selects dijet kinematics. The parton-level references are measured at the end of QCD final-state shower (with QED and weak shower off), while the hadron-level results further include hadronization but with decays disabled.
Figure 3. Energy dependence of the measured jet quantum numbers and the jet Gell-Mann–Nishijima relation in events generated with PYTHIA 8.312. Panel (a) shows the flavor-resolved jet charges as functions of . The relevant scales are determined by . Panel (b) tests the jet Gell-Mann–Nishijima relation for the -, -, and -jets at different center-of-mass energies. In both panels, colors distinguish the initiating flavors, while filled and open markers denote hadron- and parton-level results, respectively.
As shown in the left panel, hadronization induces a clear shift in the measured charge for -, -, and -initiated jets. At the parton level, reproduces the charge of the initiating quark, whereas at the hadron level one finds . Despite this hadronization correction, the measured charge is nearly independent of at both levels.
Moreover, remains consistent with over the entire energy range, as predicted by Eq. (8).
The right panel of Figure 3 tests the jet Gell-Mann–Nishijima relation in Eq. (17). The parton- and hadron-level results both follow the predicted diagonal relation over the full energy scan. Figure 4 shows that the same behavior is obtained with the default cluster-hadronization model in Herwig.
Figure 4. Corresponding energy dependence and jet Gell-Mann–Nishijima tests in events generated with Herwig 7.3 using its default hadronization model. The panels and marker conventions are defined as in Figure 3.
Figure 5 tests the -- relations by varying two PYTHIA hadronization parameters. The parameter StringFlav:probStoUD controls the relative probability for producing an pair at a string breaking, while StringFlav:probQQtoQ controls the relative rate of diquark–antidiquark production.
Figure 5. Joint scan of - and -production parameters in PYTHIA 8.312 at GeV with . Colors denote StringFlav:probStoUD = 0.010, 0.217, 0.500, 0.800, 0.950 (default 0.217), while marker shapes denote StringFlav:probQQtoQ = 0.010, 0.081, 0.500, 0.800 (default 0.081). At fixed probStoUD, each line is a linear fit with a free intercept to the four three-flavor cluster centers; its label gives the fitted slope .
We take GeV as our default scattering energy in Figure 5. Here we do not distinguish the -, -, and -initiated points because, for fixed Pythia parameters, they are nearly indistinguishable and form a compact cluster in all panels.
For each fixed value of probStoUD, the clusters organize into 5 nearly linear families. We perform fits with free intercepts to them and compare the values of extracted from the fitted slopes. Table I shows the numerical values with uncertainties propagated from the slope fits. We note that the PYTHIA string model contains additional non-perturbative hadronization settings, nevertheless, we get good agreement . Furthermore, the three projections yield mutually consistent values of within the precision of the scan, supporting the predictions in Eqs. (15) and (16).
probStoUD | ||||||
| 0.010 | −0.4839(44) | −29.8(17) | 0.0161(10) | 0.0217(59) | 0.01105(61) | 0.01080(69) |
| 0.217 | −0.3197(77) | −0.872(27) | 0.3666(29) | 0.273(13) | 0.2766(61) | 0.2785(25) |
| 0.500 | −0.1721(70) | −0.256(11) | 0.6718(28) | 0.560(15) | 0.566(11) | 0.5771(31) |
| 0.800 | −0.0531(24) | −0.0599(28) | 0.8859(42) | 0.8489(64) | 0.8477(61) | 0.8381(57) |
| 0.950 | −0.0161(58) | −0.0167(61) | 0.9644(74) | 0.952(17) | 0.952(16) | 0.948(11) |
Table I. Fitted slopes and the corresponding effective strange-pair ratios for the PYTHIA scan in Figure 5. Each value is obtained from the corresponding fitted slope. Parentheses give the one-standard-error uncertainty in the final quoted digits.
Panels (a) and (c) of Figure 5 provide a complementary test of the baryon-number dependence. At fixed probQQtoQ, the points corresponding to different probStoUD values are nearly vertically aligned. Consequently, is largely independent of the strange-pair parameter, consistent with our prediction. Averaging over the 5 probStoUD settings and 3 initiating flavors yields , , , and for probQQtoQ = , , , and , respectively. Eq. (12) then gives , , , and . These ratios should not be identified with probQQtoQ, which weights diquark production only at eligible steps of the state-dependent fragmentation history. Once a diquark is selected, subsequent steps are constrained, so this local input weight need not equal the final ratio. As a complementary event-level check, we find that agrees with the inferred values within 10%, see Table S-1, supporting the two-channel picture and the extraction in Eq. (12).
As a generator-level cross-check, we repeat the scan with the cluster hadronization model in Herwig. The default Baryonic color reconnection used in Figure 4 does not provide an adjustable baryon-number direction through MesonToBaryonFactor. For the joint scan we therefore use the BaryonicMesonic model, in which MesonToBaryonFactor controls the relative competition between mesonic and baryonic reconnection topologies. The parameters SplitPwtSquark and PwtSquark control strange-quark production during nonperturbative gluon splitting and during cluster fission and decay, respectively.
The corresponding slopes and effective values of are summarized in Table II. Under these settings, Herwig exhibits the same linear relations, the same weak dependence of on the strange-pair parameters, and the same consistent determination of from the three projections as PYTHIA, despite the different microscopic hadronization model.
| PwtSquark = SplitPwtSquark | ||||||
| 0.01 | ||||||
| 0.2 | ||||||
| 0.5 | ||||||
| 0.9 |
Table 2. Fitted slopes and the corresponding effective strange-pair ratios for the Herwig scan in Figure 6. Each value is obtained from the corresponding fitted slope. Parentheses give the one-standard-error uncertainty in the final quoted digits.
Figure 6. Corresponding joint scan of strange- and baryon-production parameters using the Herwig 7.3 BaryonicMesonic cluster-hadronization model at GeV with . Colors denote the matched (PwtSquark, SplitPwtSquark) settings (), (), (), and (), while marker shapes denote MesonToBaryonFactor = , , , . Each parameter point contains generated events. The panels and fits are defined as in Figure 5, and each line label gives its directly fitted slope .
Conclusion
In this note, we have established that the shift in the joint jets/flows of electric charge, baryon number, and strangeness relative to the corresponding quantum number of the initiating parton directly encodes the information on QCD vacuum hadronization. At the perturbative level, each flow moment equals the charge of the initiating parton; any shift observed after hadronization therefore measures the nonperturbative defect charge at the Wilson-line endpoint. Because the underlying currents are conserved, both the flow and defect first moments are renormalization-scale independent at leading power. Gauge invariance fixes the total triality of the Wilson line and its screening cloud to zero, sharply restricting the quark content and the quantum numbers of the screening cloud. Interpreting these shifts in terms of vacuum hadronization probabilities then requires only generic dynamical assumptions: the minimal one-quark and anti-diquark screening clouds dominate and the screening flavors lose memory of the initiating flavor and are sampled approximately independently from a common light-flavor distribution. None of these assumptions prescribes a specific string or cluster algorithm.
Within this framework, the measured baryon number determines the final antidiquark-to-quark endpoint ratio , predicting a near-vanishing net jet baryon number de spite the initiating-quark value , if the vacuum excitation to anti-diquark-to-quark is suppressed and . The same mechanism predicts an -dependent strangeness shift correlated with the baryon deficit . Consequently, - and -initiated jets, although carrying , acquire a nonzero mean net strangeness. The three pairwise relations among the average charge, baryon-number, and strangeness flows then provide mutually consistent determinations of the strange-to-light ratio .
A more general consequence is the jet Gell-Mann–Nishijima relation, . This closure relation only follows from the QCD flavor symmetry in the screening cloud. The joint flows therefore provide both a quantitative probe of vacuum screening and a broadly model-insensitive consistency test of flavor and charge transport during hadronization.
The framework extends naturally to heavy-flavor flows. Since nonperturbative vacuum excitation of heavy flavor pairs is strongly suppressed, the net charm and bottomness of a heavy-flavor jet should retain those of the initiating quark. The remaining charge flows follow the same light-flavor screening as in light jets, implying and for . They must also satisfy the generalized jet Gell-Mann–Nishijima relation including charm and bottomness.
Initial studies with Monte Carlo event generators reproduce the predicted charge-flow relations and recover the corresponding effective screening probabilities, demonstrating that the defect charge encodes fragmentation dynamics beyond those accessible through inclusive hadron yields. The theoretical structure is also insensitive to infrared cutoffs. We find that imposing a fiducial momentum cut of order does not modify the underlying scale invariance. We expect that future experimental measurements will provide stringent tests of these predictions and assess the potential of probing vacuum hadronization. Relevant experimental capabilities have been demonstrated in [29, 30, 31, 32, 33], and the measurements at RHIC and the LHC [34, 35, 36, 37, 38, 39, 40, 41].
Although our discussion primarily concerns hadronization in the vacuum, extending these measurements across collision systems, from high-multiplicity and collisions to heavy-ion collisions (), would be particularly interesting. In hadronic and nuclear collisions, the factorization boundary may be screened by additional mechanisms through color reconnection, interactions with the surrounding matter, or recombination with medium partons [42, 43, 44, 45, 46]. The beam-remnant effects may also induce long-range quantum-number transport through baryon-junction mechanisms, which lead to non-trivial correlation between and [42, 47, 48]. These mechanisms may modify the effective relative weights of light-quark, strange-quark, and baryonic screening channels. In nuclear collisions, the formation of a thermal partonic environment may suggest that, with increasing system size or multiplicity, the redistribution of quantum numbers between the measured flow and the defect sector can gradually approach a thermodynamic limit. This expectation is qualitatively motivated by the observed strangeness enhancement and baryon-to-meson enhancement at intermediate transverse momentum in high-multiplicity and heavy-ion collisions [49, 50, 51, 52, 53, 54, 55]. Therefore, studying these jet quantum-number flows for jets propagating through different collision systems would provide the opportunity to directly probe how vacuum fragmentation is modified by the broader range of hadronization mechanisms that exist in the medium.
Acknowledgments.— We thank Hao Chen, Zhongbo Kang, Kyle Lee, Ian Moult, Manqi Ruan, Zhen Xu and Feng Yuan for discussions. W. K. is supported by National Natural Science Foundation of China under Grant No. 12575140. H. T. L. is supported by the National Natural Science Foundation of China under Grant No. 12275156, No. 12321005 and the Shandong Provincial Department of Science and Technology under Project No. TQ012025001. W. L. and D. Y. S. are supported by the National Natural Science Foundation of China under Grant No. 12275052, No. 12147101, No. 12547102, and the Innovation Program for Quantum Science and Technology under grant No. 2024ZD0300101. W. L. is also supported by the China Postdoctoral Science Foundation under grant No. 2025M783369. X. L. is supported by the National Natural Science Foundation of China under Grant No. 12547109 and Fundamental Research Funds for the Central Universities, Beijing Normal University. X. L. and D. Y. S. would like to thank the Erwin-Schrödinger International Institute for Mathematics and Physics at the University of Vienna for partial support during the Programme ‘New Paradigms for Harnessing Quantum Field Theory at Colliders’, July 27 - August 28, 2026.
References
Appendices
Flavor ratios of the one-point correlator
We verify that our results extend to correlators defined in terms of conserved quantum numbers. As a concrete example, in Fig. S-1, we consider the one-point charge correlator (OPCC) [13, 14, 15, 18, 56, 57, 58, 59], which probes the net charge flow along a given direction . We follow [18] to measure the OPCC within a parton- initiated jet, in which Ref. [18] shows that at the leading power for ,
independent of the direction. Following Eq. (5), at the parton level, the ratio reduces to
Figure S-1. Flavor ratios of the OPCC in events generated with PYTHIA 8.312 at and TeV. Panels (a) and (b) show the parton- and hadron-level results, respectively. Panel (c) shows the hadron-level same-flavor response at TeV for minimum hadron momentum acceptance and GeV relative to . The 5 TeV samples contain events per initiating flavor, and the 10 TeV samples contain events per flavor.
In fig. S-1, we verify this prediction at two center-of-mass energies. At the parton level, the ratios do follow the initiating-quark charges, and , throughout the plotted range.
As discussed in the main text, hadronization shifts the charge away from the corresponding initiating-quark charge, which also modifies the ratio. The corresponding OPCC ratios, obtained by combining the 5 and 10 TeV samples, are
As a comparison, we quote the TeV hemisphere charge result from Figure 3, which gives
in agreement with the OPCC measurement. This agreement shows that the result is not tied to a particular observable definition, but applies more generally to correlators of conserved quantum-number flows. Panel (c) of fig. S-1 compares the hadron-level OPCC for and 1 GeV with the uncut result at TeV through . It is consistently unity and shows that fiducial momentum selection leaves the OPCC angular profile essentially unchanged.
We further test the Gell-Mann–Nishijima relation bin by bin in in Fig. S-2. To properly normalize the correlators, we approximate the normalization factor by the one-point energy correlator, which agrees with the normalization of the one-point correlators up to LL accuracy
and is therefore sufficient for the present analysis. We leave a precise calculation of the normalization factor to future work.
Fig. S-2 shows that the correlators closely follow the relation in Eq. (S-5). This provides a differential test of the jet Gell-Mann–Nishijima relation and confirms that our formalism extends to correlator observables.
Figure S-2. Comparison of the one-point correlators that contribute to eq. (S-6) for , , quark initiated jets at hadron level respectively. No thrust cut was imposed. The lower panels show the ratio , i.e., the two sides of the eq. (S-6). The events are generated with PYTHIA 8.312 for at .
Event-level test of the minimal screening approximation
The extraction of from the first moment assumes that the Wilson-line endpoint is predominantly screened either by a single quark or by an anti-diquark. These two minimal screening clouds carry defect baryon numbers and , respectively, and therefore correspond to and in a quark-initiated hemisphere. To test this two-channel truncation, we retain the event-by-event distribution
The associated topology estimator is defined as
Here we have assumed the homogeneous-vacuum identification adopted in the main text, to identify with .
Table S-1 compares with . At each probQQtoQ value, the probabilities are evaluated from the same selected event sample and combined over the 5 probStoUD settings and three initiating flavors used in the main analysis. The two estimators agree to within 9% over the full scan. This agreement supports the use of the minimal screening approximation, while the systematic difference between the estimators quantifies its residual correction.
| 0.010 | 0.01410(9) | 0.01526(9) |
| 0.081 | 0.1015(3) | 0.1102(3) |
| 0.500 | 0.3869(7) | 0.4215(6) |
| 0.800 | 0.4921(8) | 0.5341(8) |
Table S-1. Event-level test of the two-channel screening approximation. Here , . Agreement between and validates the minimal screening assumption.
References
- [3]J. C. Collins, D. E. Soper, and G. F. Sterman, Adv. Ser. Direct. High Energy Phys. 5, 1 (1989), arXiv:hep-ph/0409313.
- [4]B. R. Webber, Int. J. Mod. Phys. A 15S1, 577 (2000), arXiv:hep-ph/9912292.arxiv.org/abs/hep-ph/9912292
- [5]A. Metz and A. Vossen, Prog. Part. Nucl. Phys. 91, 136 (2016), arXiv:1607.02521 [hep-ex].DOI
- [6]B. Andersson, G. Gustafson, G. Ingelman, and T. Sjostrand, Phys. Rept. 97, 31 (1983).
- [7]B. R. Webber, Nucl. Phys. B 238, 492 (1984).
- [8]C. Bierlich et al., SciPost Phys. Codeb. 2022, 8 (2022), arXiv:2203.11601 [hep-ph].arxiv.org/abs/2203.11601
- [9]S. Gieseke, P. Kirchgaeßer, and S. Plätzer, Eur. Phys. J. C 78, 99 (2018), arXiv:1710.10906 [hep-ph].DOI
- [10]J. M. Campbell et al., SciPost Phys. 16, 130 (2024), arXiv:2203.11110 [hep-ph].DOI
- [11]D. Krohn, M. D. Schwartz, T. Lin, and W. J. Waalewijn, Phys. Rev. Lett. 110, 212001 (2013), arXiv:1209.2421 [hep-ph].DOI
- [12]W. J. Waalewijn, Phys. Rev. D 86, 094030 (2012), arXiv:1209.3019 [hep-ph].DOI
- [13]M. Riembau and M. Son, Phys. Rev. D 111, 014004 (2025), arXiv:2407.12082 [hep-ph].arxiv.org/abs/2407.12082
- [14]H. Cao, X. Liu, and F. Petriello, (2026), arXiv:2605.15280 [hep-ph].arxiv.org/abs/2605.15280
- [15]J. Zhang, K. Lee, I. Moult, Y. Chen, and Y.-J. Lee, (2026), arXiv:2608.14779 [hep-ph].arxiv.org/abs/2608.14779
- [16]J. Collins and T. C. Rogers, Phys. Rev. D 109, 016006 (2024), arXiv:2309.03346 [hep-ph].arxiv.org/abs/2309.03346
- [17]J. Greensite, Prog. Part. Nucl. Phys. 51, 1 (2003), arXiv:hep-lat/0301023.DOI
- [18]W. Li, X. Liu, and D. Y. Shao, (2026), arXiv:2607.16392 [hep-ph].arxiv.org/abs/2607.16392
- [19]D. M. Hofman and J. Maldacena, JHEP 05, 012 (2008), arXiv:0803.1467 [hep-th].DOI
- [20]P. F. Monni, G. Vita, Z. Xu, and H. X. Zhu, Phys. Rev. Lett. 137, 031901 (2026), arXiv:2508.00977 [hep-ph].arxiv.org/abs/2508.00977
- [21]T. D. Brennan, C. Cordova, and T. T. Dumitrescu, (2022), arXiv:2206.15401 [hep-th].arxiv.org/abs/2206.15401
- [22]T. D. Brennan, T. Jacobson, and K. Roumpedakis, JHEP 11, 153 (2025), arXiv:2504.08036 [hep-th].arxiv.org/abs/2504.08036
- [23]T. Gehrmann, A. von Manteuffel, V. Sotnikov, and T.-Z. Yang, (2026), arXiv:2604.09534 [hep-ph].arxiv.org/abs/2604.09534
- [24]A. Mitov, S. Moch, and A. Vogt, Phys. Lett. B 638, 61 (2006), arXiv:hep-ph/0604053.DOI
- [25]T. Nakano and K. Nishijima, Prog. Theor. Phys. 10, 581 (1953).
- [26]K. Nishijima, Prog. Theor. Phys. 13, 285 (1955).
- [27]M. Gell-Mann, Nuovo Cim. 4, 848 (1956).
- [28]J. Bellm et al., Eur. Phys. J. C 76, 196 (2016), arXiv:1512.01178 [hep-ph].DOI
- [29]P. Abreu et al. (DELPHI), Z. Phys. C 67, 543 (1995).
- [30]P. Abreu et al. (DELPHI), Phys. Lett. B 401, 118 (1997).
- [31]P. Abreu et al. (DELPHI), Eur. Phys. J. C 17, 207 (2000), arXiv:hep-ex/0106063.DOI
- [32]K. Abe et al. (SLD), Phys. Rev. D 59, 052001 (1999), arXiv:hep-ex/9805029.DOI
- [33]K. Ackerstaff et al. (OPAL), Eur. Phys. J. C 8, 241 (1999), arXiv:hep-ex/9805025.
- [34]K. Aamodt et al. (ALICE), JINST 3, S08002 (2008).
- [35]J. Adam et al. (ALICE), Eur. Phys. J. Plus 131, 168 (2016), arXiv:1602.01392 [physics.data-an].
- [36]B. I. Abelev et al. (STAR), Phys. Rev. C 79, 034909 (2009), arXiv:0808.2041 [nucl-ex].DOI
- [37]S. Acharya et al. (ALICE), Phys. Lett. B 827, 136984 (2022), arXiv:2105.04890 [nucl-ex].
- [38]S. Acharya et al. (ALICE), Phys. Rev. C 111, 015201 (2025), arXiv:2405.19855 [nucl-ex].
- [39]S. Acharya et al. (ALICE), Phys. Lett. B 833, 137338 (2022), arXiv:2110.06566 [nucl-ex].
- [40]S. Acharya et al. (ALICE), JHEP 09, 102 (2024), arXiv:2308.16706 [hep-ex].
- [41]S. Acharya et al. (ALICE), Phys. Lett. B 807, 135564 (2020), arXiv:1910.14396 [nucl-ex].
- [42]J. R. Christiansen and P. Z. Skands, JHEP 08, 003 (2015), arXiv:1505.01681 [hep-ph].DOI
- [43]R. J. Fries, B. Muller, C. Nonaka, and S. A. Bass, Phys. Rev. Lett. 90, 202303 (2003), arXiv:nucl-th/0301087.DOI
- [44]V. Greco, C. M. Ko, and P. Levai, Phys. Rev. Lett. 90, 202302 (2003), arXiv:nucl-th/0301093.DOI
- [45]R. J. Fries, V. Greco, and P. Sorensen, Ann. Rev. Nucl. Part. Sci. 58, 177 (2008), arXiv:0807.4939 [nucl-th].DOI
- [46]K. C. Han, R. J. Fries, and C. M. Ko, Phys. Rev. C 93, 045207 (2016), arXiv:1601.00708 [nucl-th].DOI
- [47]D. Kharzeev, Physics Letters B 378, 238 (1996).
- [48]B. E. Abelev et al. (STAR), Science 393, 727 (2026), arXiv:2408.15441.DOI
- [49]J. Rafelski and B. Muller, Phys. Rev. Lett. 48, 1066 (1982), [Erratum: Phys.Rev.Lett. 56, 2334 (1986)].
- [50]A. Andronic, P. Braun-Munzinger, K. Redlich, and J. Stachel, Nature 561, 321 (2018), arXiv:1710.09425 [nucl-th].DOI
- [51]B. I. Abelev et al. (STAR), Phys. Rev. Lett. 97, 152301 (2006), arXiv:nucl-ex/0606003.
- [52]B. I. Abelev et al. (STAR), Phys. Rev. C 77, 044908 (2008), arXiv:0705.2511 [nucl-ex].DOI
- [53]B. Abelev et al. (ALICE), Phys. Rev. C 88, 044910 (2013), arXiv:1303.0737 [hep-ex].
- [54]J. Adam et al. (ALICE), Nature Phys. 13, 535 (2017), arXiv:1606.07424 [nucl-ex].
- [55]S. Acharya et al. (ALICE), Phys. Rev. C 99, 024906 (2019), arXiv:1807.11321 [nucl-ex].
- [56]J. Barata, R. Jumanov, and A. V. Sadofyev, (2026), arXiv:2607.06667 [hep-ph].arxiv.org/abs/2607.06667
- [57]H. Cao and F. Petriello, (2026), 10.1103/3rv6-8dvl, arXiv:2606.27454 [hep-ph].DOI
- [58]J. Zhang, Y. Chen, K. Lee, I. Moult, C. Baldenegro, H. Bossi, and Y.-J. Lee, (2026), arXiv:2608.15357 [hep-ex].arxiv.org/abs/2608.15357
- [59]J. Zhang, Y. Chen, K. Lee, I. Moult, C. Baldenegro, H. Bossi, and Y.-J. Lee, (2026), arXiv:2608.15367 [hep-ex].arxiv.org/abs/2608.15367