Nikhef 2025-010

Detecting Traces of Light-Quark Yukawa Couplings

to the Higgs Boson in Fragmentation Products

Johannes K. L. Michel1, 2, ∗
1
Institute for Theoretical Physics Amsterdam and Delta Institute for Theoretical Physics,
University of Amsterdam, Science Park 904, 1098 XH Amsterdam, The Netherlands
2
Nikhef, Theory Group, Science Park 105, 1098 XG, Amsterdam, The Netherlands
(Dated: August 8, 2025)
We point out that Yukawa interactions of light quarks with the Higgs boson are imprinted as
unique azimuthal modulations in the density of fragmentation hadrons relative to the Higgs pT .
We introduce Yukawa Fragmentation Asymmetries (YFAs), interference observables that are lin-
early proportional to real (Standard Model) or CP-odd Yukawa couplings, respectively. The chiral
suppression is lifted nonperturbatively by chiral-odd multi-hadron fragmentation functions. As a
case study, we consider V H production with a tagged target fragmentation hadron at the HL-LHC.
We estimate that the YFAs from this single example process are already competitive with current
arXiv:2508.05914v1 [hep-ph] 8 Aug 2025

projected HL-LHC sensitivities for second-generation couplings, and outperform them by a factor
of 2-3 for first-generation couplings. Our results point to deep synergies between precision studies
of confinement and the in-depth exploration of the Higgs sector.

I. INTRODUCTION tromagnetic repulsion in the (uud) proton [7], the (ddu)

neutron would be the lightest baryon, making chemistry
Thirteen years after the discovery of the Higgs bo- impossible, and thus Life. Empirically confirming that
son [1, 2], evidence for many of its predicted couplings this curious fact is indeed rooted in the relative size of
to leptons and quarks remains elusive. While the most the u and d Yukawa couplings to the Higgs field, as the-
exotic fermions in the Standard Model (SM) of particle orized in the SM, remains a pressing unsolved problem
physics were the first to have their Yukawa interactions for 21st -century particle physics.
with the Higgs boson confirmed [3, 4], owing precisely to A variety of experimental probes of first and second-
the large masses (and thus large couplings) that made generation Yukawa couplings at the upcoming high-
them hard to discover in the first place, it is a great luminosity (HL) LHC have been proposed, ranging from
irony that such an observation is extremely challenging exclusive [8–11] and inclusive decays [12–14], to off-shell
to make for some of the lightest, most mundane, and most rates [15–18] and differential distributions in (associated)
abundant matter particles, the up and down quarks that Higgs production [19–25]. Nonetheless, experimental
form the protons and neutrons of everyday life [5]. Cu- measurements to date have only been able to set the
riously, within the first fermion generation the up quark, expected upper limits [26–35], and even HL-LHC pro-
with charge Qu = 2/3, is significantly lighter than the jections range from tantalizingly close to a long way off
down quark with Qd = −1/3 [6]. If the roles were re- from discovery: At 95% CL, one expects to be able to
versed, as in the other two generations, or if the mass constrain [36]
difference were too small to counter the stronger elec- |yq /yqSM | < 560 , 260 , 13 , 1.2 , (1)

for q = u, d, s, c. This would leave much room for

nonstandard, beyond-the-SM (BSM) Yukawa interac-
tions [37], and calls for innovative approaches to bring
down the projected limits at the HL-LHC and future
colliders [38–44]. Experimental searches for first and
second-generation Yukawa interactions face three key
challenges: (1) At its most basic, they contend with
the smallness of the light-quark Yukawa couplings, mak-
ing for tiny signal rates. (2) In all known cases, one
faces irreducible backgrounds from other SM processes
and the Yukawa couplings of heavier quarks, which re-
quire complicated perturbative calculations. (3) While
FIG. 1. Factorization for pp collisions with a tagged target light quarks interact perturbatively with the Higgs bo-
fragmentation hadron. As an example of a hard scattering son at high energies, they originate from, and fragment
diagram (blue), we illustrate the interference of the Yukawa into, strongly bound hadronic states, requiring rigorous
interaction with standard SM V H production and indicate a factorization statements to separate the Yukawa signal
representative helicity configuration. from the nonperturbative physics of confinement.
 2

In this letter we propose the measurement of Yukawa target fragmentation plane

Fragmentation Asymmetries (YFAs), a novel class
of observables that resolve all three challenges in one
stroke. They are interference observables that are (1) lin- lab frame
H π± ϕ
early proportional to the Yukawa couplings of interest, in y
particular those of the abundant valence u and d quarks x
in protons and pions, which improves sensitivity. They z
p p
(2) feature powerful, symmetry-protected mechanisms
that stabilize them against radiative corrections and the
contributions of heavier sea quarks. Most notably, they V
turn challenge (3) into a virtue and — within the factor- Higgs plane
ization paradigm — make use of the nontrivial properties
of the QCD confinement transition itself to overcome (1) FIG. 2. Kinematics of V H production in pp collisions with
and (2). a tagged target fragmentation hadron in the lab frame (mo-
The remainder of this letter is structured as follows: menta not to scale). The YFA for CP-even, SM-like Yukawa
We first present an accessible introduction to the physical interactions is formed by comparing rates of hadrons above
mechanism behind YFAs, focusing on the simple exam- (sin φ > 0) and below the Higgs plane.
ple of V H production with a tagged target-fragmentation
hadron h at the LHC. We next list some of the distin-
guishing all-order properties of YFAs. (Derivations and insertion of a mass term mq ψ̄q ψq and leads to an overall
additional field-theoretical background are presented in suppression by yq mq /Q ∼ yq2 , with Q2 = (pV + pH )2 the
a companion paper [45].) We then present sensitivity es- hard scale, on par with the Yukawa amplitude squared.
timates for YFAs in V Hh production at the HL-LHC for One important insight is that the helicity configuration
first and second-generation quarks. We close by listing where one interferes λq = − with + is, in fact, physical,
applications to other processes and at future colliders. and corresponds to transverse quark polarization. In-
deed, in a basis of helicity eigenstates |±⟩, the spin den-
sity matrix of a transversely polarized quark involves a
II. PHYSICAL MECHANISM linear combination of offdiagonal Pauli matrices σ1 and
σ2 . If the incoming fermion can be prepared in such a
We are interested in constraining the parameters of the state, the chiral suppression of the interference diagram
following effective Lagrangian coupling the Higgs boson in Fig. 1 is lifted. The key question is whether a nonzero
H to quark fields ψq [10], transverse Bloch vector is physically allowed.
yq iỹq This is most clearly the case for incoming electrons,
L ⊃ − √ H ψ̄q ψq − √ H ψ̄q γ5 ψq . (2) where recently a proposal was made [46] to tune positron
2 2
√ beams and transversely polarized electron beams to the
Within the SM, yqSM = 2 mq /v, where mq is the quark Higgs resonance to improve the sensitivity to the electron
mass and v is the Higgs vacuum expectation value, while Yukawa coupling compared to the unpolarized rate [47].
CP-odd interaction terms vanish, ỹqSM = 0. This proposal can be adapted to light quarks by giving
A key challenge in constructing interference observ- one of the incoming nucleons a net transverse polariza-
ables linear in the quark Yukawa couplings yq and ỹq is tion, which would then enter the partonic collision modu-
that interference diagrams with other SM processes (as lated by the quark transversity parton distribution func-
shown in blue in Fig. 1) are chirally suppressed for mass- tion (PDF). This, of course, comes at the significant cost
less quarks in unpolarized partonic collisions. In stan- of having to polarize the beam while maintaining high
dard V H production with V = Z, W ± as shown to the energies and luminosities.
right of the cut, the helicities of the quark and antiquark We instead propose a different route that is viable also
must be opposite because an SM interaction term like in unpolarized hadronic collisions, and is based on the
ψ̄q′ γ µ ψq = ψ̄q′ L γ µ ψqL + ψ̄q′ R γ µ ψqR (which annihilates powerful theoretical observation [48–54] that transverse
these helicity states) only couples field components of parton polarization is nonperturbatively imprinted on
identical handedness to each other, i.e., it is chiral even. the distribution of fragmentation products in the trans-
The situation is different on the left because Eq. (2) is verse plane, leading to e.g. the experimentally established
chiral odd and couples left to right-handed fields, thus Collins [55–59] and Collins-Artru effects [60]. For our
a nonzero amplitude for a negative-helicity antiquark re- example, assume that we tag on a final-state hadron
quires quark helicity λq = − (red). However, in unpolar- h that is observed at forward rapidity and small trans-
ized collisions one only sums over identical helicity config- verse momentum phT ∼ ΛQCD , and thus at leading power
urations in both the amplitude and the conjugate ampli- originates from the remnant of the struck proton mov-
tude. Getting back λq = + (blue) requires an additional ing in the forward direction, as illustrated in Fig. 2.
 3

In this case, the hadron carries information about the make them particularly powerful probes of light-quark
“quark hole” left by the hard-scattering quark, whose Yukawa couplings:
opposite transverse polarization in particular breaks the
azimuthal symmetry of the proton remnant around the (A) Forming an asymmetry linear in phT at the nonper-
beam axis. This means that a linear correlation between turbative scale, which is an intrinsically chiral-odd,
transverse quark spin and the direction of phT becomes al- fermionic phenomenon, uniquely picks out a single
lowed at leading twist. The correlation strength is known chiral-odd contribution at the hard scale to all or-
as the quark Boer-Mulders fracture function (BMFrF) ders in perturbation theory and at leading power in
h⊥ 1
qh [51, 53, 54], and is a chiral-odd nonperturbative ma-
ΛQCD /Q. The only such term in the SM is Eq. (2).
trix element related to chiral symmetry breaking by the
QCD vacuum. Formally, it appears as the coefficient (B) The quark and antiquark terms feature a relative
of a Dirac structure ip h
/T p
/q in the decomposition of the sign. This property is phenomenologically critical,
FrF correlator Φqh/p in Fig. 1, which becomes allowed in since it means that sea (anti)quark contributions
addition to the unpolarized term p can cancel, leaving behind the valence quark contri-
/q after breaking the
butions of interest. It is likewise stable to all orders,
azimuthal symmetry.
and can be understood by noting that the SM La-
Returning to Fig. 1, we therefore expect to observe,
grangian (approximately) preserves CP, while the
at the level of the hadronic cross section, modulations
sin φ modulation is P odd.
in hadron yield that based on Lorentz invariance are ei-
ther proportional to pH h
T · pT ∝ cos φ or the contraction (C) Forming an asymmetry ỸFAh,V with respect to pxh
ϵ(Pa , Pb , pH , ph ) ∝ sin φ with the Levi-Civita symbol,
instead, one arrives at a signal prediction of the
where φ ≡ φh − φH is the signed azimuthal separation
same form as Eq. (4) and with identical coefficient
between H and h, see Fig. 2. The maximal violation of TU
Cqq ′ V , but with the yq replaced by ỹq . Therefore
parity by the weak interaction makes the latter the dom-
inant one multiplying CP-even (real) Yukawa couplings all of our sensitivity estimates for yq /yqSM in this
yq . This modulation leads to a difference between hadron letter immediately carry over to possible CP-odd
+
rates Nh,V −
observed at phy > 0 compared to the rate Nh,V ỹq in units of yqSM .
at negative phy < 0 in a righthanded coordinate system (D) While the BMFrF is an a priori unknown non-
chosen as in Fig. 2. We therefore propose to measure the perturbative matrix element, it can readily be ex-
asymmetry observable tracted using azimuthal correlations between two
+
Nh,V −
− Nh,V ∆σh,V target fragmentation hadrons, one in each beam
YFAh,V ≡ + − ≡ , (3) direction, with chiral-even hard probes. These
Nh,V + Nh,V σh,V
include high-statistics SM baseline processes like
where the denominator is given by the total single- Drell-Yan production. In light of the present work,
inclusive cross section for V H production with a tagged we strongly encourage our experimental colleagues
hadron in a suitable forward acceptance volume. The to initiate differential measurements of such az-
signal cross section in the numerator is predicted to be imuthal hadron-hadron correlations. Pending clar-
Z   ification of its universality [68], the BMFrF may
e3 X TU ⊥ ⊥ also be extracted in a clean fashion at the future
∆σh,V = dΦ C ′ yq h̄ f q̄ ′ − yq ′ h̄ ′ fq
2 qq V qh q̄ h
4πEcm ′ EIC [69].
q,q
(4)
in terms of a perturbatively calculable coefficient IV. SENSITIVITY ESTIMATES
TU ⊥
Cqq ′ V [45], the (anti)quark BMFrFs h̄ih (xa ) integrated

over the acceptance volume, standard PDFs fj (xb ), and

We now derive sensitivity estimates for YFAs in V H
the Yukawa coupling of the respective transversely polar-
production at the HL-LHC with an integrated luminos-
ized parton.
ity of 3 ab−1 at Ecm = 14 TeV [70]. We begin with the
impact of charged-pion YFAs on yu and yd . We envision
III. KEY PROPERTIES a scenario where valence BMFrFs have been measured
to sufficient accuracy from baseline processes, while the
smaller sea quark BMFrFs and their potential asymme-
Eqs. (3) and (4), apart from their great experimental
tries are still poorly known (since they are not enhanced
simplicity, have several theoretical properties [45] that
in the baseline). The latter then become the dominant
source of systematic theoretical uncertainty. We focus
on the statistical component of the experimental uncer-
1 For (extended) fracture functions in general, see Refs. [61–67]. tainty, leaving experimental systematics to future work.
 4

200 30
pp → V Hπ ± pp → V HK ±/D ±
150 14 TeV 14 TeV
W +π − 20 W ∓D±
R R
L = 3 ab−1 L = 3 ab−1
100
10
50 W −π +
W ±K ∓
yd/ydSM

ys/ysSM
0 0

−50
Zπ ± −10
ZK ±
−100
∆χ2 = 1 −20 ∆χ2 = 1
−150 ∆stat ∆stat
∆stat ⊕ ∆syst ∆stat ⊕ ∆syst
−200 −30
−300 −200 −100 0 100 200 300 −1.0 −0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0
yu/yuSM yc/ycSM

FIG. 3. Projected ∆χ2 = 1 contours from YFA measurements in V Hh production at the HL-LHC. Left: Constraints on up
and down Yukawa couplings from h = π ± . Right: Constraints on charm and strange Yukawa couplings from h = K ± , D± .

To evaluate the YFAs, we simply assume here that A valuable principle to estimate the relative sizes of
multi-hadron fragmentation matrix elements is the de-
h̄⊥
ih (x) = Pih fi (x) , (5) gree to which they are favored in terms of valence fla-
vor [49]. Applying this principle, together with a sup-
where Pih is an effective degree of transverse polarization pression by fπ /mp for the most favored contributions (as
(DOP) that we take to be independent of the partonic appropriate for a chiral-odd matrix element), we can es-
momentum fraction x.2 Using Eq. (5), the coefficients timate all DOPs entering Eq. (5) for h = π ± [71]. For
multiplying yq Pqh or yq′ Pq̄′ h in Eq. (4) are readily eval- q = s, c, b we assign separate, conservative uncertain-
uated in terms of standard PDFs [71]. We also need to ties for the symmetrized and sea asymmetry DOPs. For
estimate σh,V in the denominator of Eq. (3), which can q = c, b we further work to leading power in ΛQCD /mq to
be recast as the product of the total V H cross section σV sharpen our estimates, using techniques from Ref. [73].
and the average yield nh of hadrons of type h per V H The projected ∆χ2 = 1 contours in the (yu , yd ) plane
event. Using Pythia 8.3 [72] with the default tune, we that arise from the above statistical and systematic the-
find nπ+ ≈ 8.65 and nπ− ≈ 8.59 within an acceptance of ory uncertainties are shown in the left panel of Fig. 3.
3 ≤ ηh ≤ 5 and 0.2 ≤ phT ≤ 2 GeV.3 This results in the We account for the full systematic covariance matrix
following statistical uncertainty estimate for ∆σh,V in a of experimental channels that use the same final-state
given channel: hadron (and thus are subject to the same unknown sea-
r quark BMFrFs and sea asymmetries). If we let ∆κq ≡

nh σV
δstat ∆σh,V = , (6) yq /yqSM − 1 and convert to 95% CL, the projected limits
2L AV H
on the individual parameters of interest are
where we included a rough estimate AV H = 0.3 of the
|∆κu | < 172stat ⊕ 89syst , |∆κd | < 95stat ⊕ 63syst . (7)
joint acceptance of V and H decay products, and a factor
of 2 for the two possible beam directions in which h may We observe that even under our extreme assumptions on
be reconstructed. unknown sea-quark BMFrFs, the inherent cancellation of
heavy-quark contributions in Eq. (4) is active, in particu-
lar for ZH production, and leads to a subdominant sys-
2
tematic uncertainty. Another feature of the observable
Any actual x dependence revealed by baseline data can of course
be included, and possibly even exploited to enrich the signal.
apparent from Fig. 3 is that combining different final-
3 These results are independent of the vector boson type to good state hadrons and boson types provides a unique handle
approximation. A more refined treatment in terms of integrated to tease apart the individual Yukawa couplings.
unpolarized FrFs f¯qh (x) is again possible with baseline data. We also consider the prospects of constraining yc and
 5

modulations are proportional to (a) ϵ(P, kf , pH , RT ),

with P the total dihadron momentum and kf the mo-
mentum of the second jet (or final-state lepton), and
(b) ϵ(P, q, pH , RT ) with q = (Ecm , 0) in the center-of-
mass frame. We leave phenomenological studies of this
rich new field of Higgs precision physics to future work.
We point out that one can also recast our results in the
language of (nuclear) Energy-Energy Correlators [87–93],
FIG. 4. Examples of YFAs that leverage dihadron fragmen- use reconstructed forward jets instead of hadrons, ex-
tation at the HL-LHC and future colliders. We suppress PDF plore structurally similar [94] (but Sudakov-suppressed)
correlators for unpolarized partons for simplicity.
observables involving transverse momentum-dependent
PDFs and FFs, or extend them to (pseudo)scalar BSM
ys from V H YFAs involving reconstructed kaons K ± states interacting with quarks as in Eq. (2), which may
and charm mesons D± .4 The relevant yields are nK ± ≈ lead to novel search strategies.
2.67645 and nD± = αs (mc )/(2π), where the latter is also
well in line with the result from Pythia 8.3. The re-
sulting ∆χ2 = 1 contours are shown in the right panel of
VI. CONCLUSIONS
Fig. 3. Converting to 95% CL, we find

|∆κc | < 1.3stat ⊕ 0.49syst , |∆κs | < 14stat ⊕ 13syst . (8) We have introduced Yukawa Fragmentation Asymme-
tries (YFAs), a new class of observables that combine
linear sensitivity, theoretical control, and experimental
V. GENERALIZATIONS AND EXTENSIONS simplicity into a uniquely powerful probe of light-quark
Yukawa couplings to the Higgs boson. The observ-
Beyond the basic example of pp → V Hh that we chose able pulls out all the stops of the Standard Model to
here to keep things simple and analytically tractable, un- approach these challenging measurement targets, draw-
suppressed YFAs featuring sea-quark cancellations can ing on confinement, chiral symmetry breaking, the va-
be constructed for any Higgs production process with lence/sea structure of hadrons, and weak parity violation.
a source of parity violation at present and future col- As part of recent renewed interest in interdisciplinary ap-
liders. A straightforward generalization is to tag tar- plications of multi-hadron fragmentation functions [87–
get fragmentation hadrons in other such processes, in- 93, 95, 96], our results strongly motivate the experimen-
cluding vector boson fusion (VBF), with much higher tal exploration of these nonperturbative matrix elements,
rates than V H. A more powerful extension is to ex- estimating them from models [97], studying their higher-
ploit the same physical mechanism in dihadron fragmen- twist counterparts [98], and rigorously analyzing them
tation q, q̄ → h1 h2 X, where a closely related chiral-odd in effective field theories of QCD [73, 99–102]. We ex-
correlation Hh∢1 h2 /q exists between the (anti)quark trans- pect that our results can become a new cornerstone of a
verse polarization and the relative transverse momentum nascent field [103–108] applying confinement physics to
RTµ = phT1 − phT2 ∼ ΛQCD of the two hadrons with respect improve SM measurements and BSM searches. We look
to the fragmentation axis [50, 52, 74–83]. This interfer- forward to their future experimental realization that may
ence dihadron fragmentation function (IFF) has already help solve the riddle of first-generation Yukawa couplings.
been determined from data [84, 85]; a further advan-
tage is that these pairs of nearby hadrons can typically Acknowledgments: The author gratefully acknowl-
be reconstructed in the central detector volume. The edges encouraging discussions with Ankita Budhraja,
IFF induces a YFA in weak Higgs production processes Eric Laenen, Juraj Klarić, Rebecca von Kuk, Ian Moult,
featuring final-state quarks or antiquarks, examples of Piet Mulders, Tristan du Pree, Zhiquan Sun, and Wouter
which are shown in Fig. 4. They range from (a) VBF Waalewijn, and would like to thank Gavin Salam for a
at the LHC (f = q, q̄), to the cleaner environment and Nikhef Colloquium in March 2024 emphasizing the mys-
high rates of VBF at a possible LHeC (f = e− ) [86], tery of the first-generation Yukawa couplings, which led
and to the pristine conditions of (b) V H production the author to doodle Fig. 4 (b) on the margin of his notes
with a hadronic V decay at a future e+ e− collider. For from that talk.
real, SM-like Yukawa couplings, the relevant azimuthal The author wishes to thank the CERN Theoretical
Physics Department for hospitality while part of this
work was carried out. The author was supported by the
D-ITP consortium, a program of NWO that is funded by
4 We generically use K ± (D± ) for all hadrons with strangeness the Dutch Ministry of Education, Culture and Science
S = ±1 (charm C = ±1). (OCW).
 6

[27] M. Aaboud et al. (ATLAS), Phys. Lett. B 786, 134

(2018), arXiv:1807.00802 [hep-ex].
[28] A. M. Sirunyan et al. (CMS), JHEP 03, 131 (2020),
∗
[email protected] arXiv:1912.01662 [hep-ex].
[1] G. Aad et al. (ATLAS), Phys. Lett. B 716, 1 (2012), [29] G. Aad et al. (ATLAS), Eur. Phys. J. C 82, 717 (2022),
arXiv:1207.7214 [hep-ex]. arXiv:2201.11428 [hep-ex].
[2] S. Chatrchyan et al. (CMS), Phys. Lett. B 716, 30 [30] A. Tumasyan et al. (CMS), Phys. Lett. B 842, 137534
(2012), arXiv:1207.7235 [hep-ex]. (2023), arXiv:2206.03525 [hep-ex].
[3] A. Tumasyan et al. (CMS), Nature 607, 60 (2022), [Er- [31] G. Aad et al. (ATLAS), JHEP 08, 027 (2022),
ratum: Nature 623, (2023)], arXiv:2207.00043 [hep-ex]. arXiv:2202.00487 [hep-ex].
[4] G. Aad et al. (ATLAS), Nature 607, 52 (2022), [Erra- [32] A. Hayrapetyan et al. (CMS), Phys. Rev. Lett. 132,
tum: Nature 612, E24 (2022)], arXiv:2207.00092 [hep- 121901 (2024), arXiv:2310.05164 [hep-ex].
ex]. [33] G. Aad et al. (ATLAS), JHEP 02, 045 (2025),
[5] G. P. Salam, L.-T. Wang, and G. Zanderighi, Nature arXiv:2407.15550 [hep-ex].
607, 41 (2022), arXiv:2207.00478 [hep-ph]. [34] A. Hayrapetyan et al. (CMS), Phys. Lett. B 862, 139296
[6] M. Tanabashi et al. (Particle Data Group), Phys. Rev. (2025), arXiv:2410.18289 [hep-ex].
D 98, 030001 (2018). [35] V. Chekhovsky et al. (CMS), (2025), arXiv:2502.05665
[7] S. Borsanyi et al. (BMW), Science 347, 1452 (2015), [hep-ex].
arXiv:1406.4088 [hep-lat]. [36] J. de Blas et al., JHEP 01, 139 (2020), arXiv:1905.03764
[8] G. T. Bodwin, F. Petriello, S. Stoynev, and M. Velasco, [hep-ph].
Phys. Rev. D 88, 053003 (2013), arXiv:1306.5770 [hep- [37] B. A. Erdelyi, R. Gröber, and N. Selimovic, JHEP 05,
ph]. 189 (2025), arXiv:2410.08272 [hep-ph].
[9] A. L. Kagan, G. Perez, F. Petriello, Y. Soreq, [38] J. Gao, JHEP 01, 038 (2018), arXiv:1608.01746 [hep-
S. Stoynev, and J. Zupan, Phys. Rev. Lett. 114, 101802 ph].
(2015), arXiv:1406.1722 [hep-ph]. [39] J. Duarte-Campderros, G. Perez, M. Schlaffer, and
[10] M. König and M. Neubert, JHEP 08, 012 (2015), A. Soffer, Phys. Rev. D 101, 115005 (2020),
arXiv:1505.03870 [hep-ph]. arXiv:1811.09636 [hep-ph].
[11] D. d’Enterria and V. D. Le (2025) arXiv:2508.00466 [40] R. Li, B.-W. Wang, K. Wang, X. Zhang, and Z. Zhou,
[hep-ph]. Phys. Rev. D 100, 053008 (2019), arXiv:1905.09457
[12] G. Perez, Y. Soreq, E. Stamou, and K. Tobioka, Phys. [hep-ph].
Rev. D 92, 033016 (2015), arXiv:1503.00290 [hep-ph]. [41] Q. Bi, K. Chai, J. Gao, Y. Liu, and H. Zhang, Chin.
[13] G. Perez, Y. Soreq, E. Stamou, and K. Tobioka, Phys. Phys. C 45, 023105 (2021), arXiv:2009.02000 [hep-ph].
Rev. D 93, 013001 (2016), arXiv:1505.06689 [hep-ph]. [42] M. Knobbe, F. Krauss, D. Reichelt, and S. Schumann,
[14] L. M. Carpenter, T. Han, K. Hendricks, Z. Qian, Eur. Phys. J. C 84, 83 (2024), arXiv:2306.03682 [hep-
and N. Zhou, Phys. Rev. D 95, 053003 (2017), ph].
arXiv:1611.05463 [hep-ph]. [43] J. F. Kamenik, A. Korajac, M. Szewc, M. Tammaro,
[15] Y. Zhou, Phys. Rev. D 93, 013019 (2016), and J. Zupan, Phys. Rev. D 109, L011301 (2024),
arXiv:1505.06369 [hep-ph]. arXiv:2306.17520 [hep-ph].
[16] A. Falkowski, S. Ganguly, P. Gras, J. M. No, K. To- [44] H. Liang, Y. Zhu, Y. Wang, Y. Che, C. Zhou, H. Qu,
bioka, N. Vignaroli, and T. You, JHEP 04, 023 (2021), and M. Ruan, Phys. Rev. Lett. 132, 221802 (2024),
arXiv:2011.09551 [hep-ph]. arXiv:2310.03440 [hep-ex].
[17] N. Vignaroli, Symmetry 14, 1183 (2022), [45] J. K. L. Michel, “On the Formal Properties of Yukawa
arXiv:2205.09449 [hep-ph]. Fragmentation Asymmetries,” (2025).
[18] E. Balzani, R. Gröber, and M. Vitti, JHEP 10, 027 [46] R. Boughezal, F. Petriello, and K. Şimşek, Phys. Rev.
(2023), arXiv:2304.09772 [hep-ph]. D 110, 075026 (2024), arXiv:2407.12975 [hep-ph].
[19] I. Brivio, F. Goertz, and G. Isidori, Phys. Rev. Lett. [47] D. d’Enterria, A. Poldaru, and G. Wojcik, Eur. Phys.
115, 211801 (2015), arXiv:1507.02916 [hep-ph]. J. Plus 137, 201 (2022), arXiv:2107.02686 [hep-ex].
[20] F. Bishara, U. Haisch, P. F. Monni, and E. Re, Phys. [48] J. C. Collins, Nucl. Phys. B 396, 161 (1993), arXiv:hep-
Rev. Lett. 118, 121801 (2017), arXiv:1606.09253 [hep- ph/9208213.
ph]. [49] X. Artru and J. C. Collins, Z. Phys. C 69, 277 (1996),
[21] Y. Soreq, H. X. Zhu, and J. Zupan, JHEP 12, 045 arXiv:hep-ph/9504220.
(2016), arXiv:1606.09621 [hep-ph]. [50] J. C. Collins, S. F. Heppelmann, and G. A. Ladinsky,
[22] F. Yu, JHEP 02, 083 (2017), arXiv:1609.06592 [hep-ph]. Nucl. Phys. B 420, 565 (1994), arXiv:hep-ph/9305309.
[23] L. Alasfar, R. Corral Lopez, and R. Gröber, JHEP 11, [51] D. Boer and P. J. Mulders, Phys. Rev. D 57, 5780
088 (2019), arXiv:1909.05279 [hep-ph]. (1998), arXiv:hep-ph/9711485.
[24] J. A. Aguilar-Saavedra, J. M. Cano, and J. M. No, [52] D. Boer, R. Jakob, and M. Radici, Phys. Rev. D
Phys. Rev. D 103, 095023 (2021), arXiv:2008.12538 67, 094003 (2003), [Erratum: Phys.Rev.D 98, 039902
[hep-ph]. (2018)], arXiv:hep-ph/0302232.
[25] L. Alasfar, R. Gröber, C. Grojean, A. Paul, and [53] D. Sivers, Phys. Rev. D 79, 085008 (2009),
Z. Qian, JHEP 11, 045 (2022), arXiv:2207.04157 [hep- arXiv:0811.2388 [hep-ph].
ph]. [54] D. Sivers, Phys. Rev. D 81, 034029 (2010),
[26] M. Aaboud et al. (ATLAS), JHEP 07, 127 (2018), arXiv:0910.5420 [hep-ph].
arXiv:1712.02758 [hep-ex]. [55] K. Abe et al. (Belle), Phys. Rev. Lett. 96, 232002
(2006), arXiv:hep-ex/0507063.
 7

[56] R. Seidl et al. (Belle), Phys. Rev. D 78, 032011 [84] A. Courtoy, A. Bacchetta, M. Radici, and A. Bianconi,
(2008), [Erratum: Phys.Rev.D 86, 039905 (2012)], Phys. Rev. D 85, 114023 (2012), arXiv:1202.0323 [hep-
arXiv:0805.2975 [hep-ex]. ph].
[57] J. P. Lees et al. (BaBar), Phys. Rev. D 90, 052003 [85] C. Cocuzza, A. Metz, D. Pitonyak, A. Prokudin,
(2014), arXiv:1309.5278 [hep-ex]. N. Sato, and R. Seidl (JAM), Phys. Rev. D 109, 034024
[58] J. P. Lees et al. (BaBar), Phys. Rev. D 92, 111101 (2024), arXiv:2308.14857 [hep-ph].
(2015), arXiv:1506.05864 [hep-ex]. [86] F. Ahmadova et al., (2025), arXiv:2503.17727 [hep-ex].
[59] H. Li et al. (Belle), Phys. Rev. D 100, 092008 (2019), [87] Y. Guo, X. Liu, F. Yuan, and H. X. Zhu, Research
arXiv:1909.01857 [hep-ex]. 2025, 0552 (2025), arXiv:2406.05880 [hep-ph].
[60] A. Vossen et al. (Belle), Phys. Rev. Lett. 107, 072004 [88] K.-B. Chen, J.-P. Ma, and X.-B. Tong, JHEP 08, 227
(2011), arXiv:1104.2425 [hep-ex]. (2024), arXiv:2406.08559 [hep-ph].
[61] J. Collins, Foundations of perturbative QCD (Cam- [89] K. Lee and I. Stewart, (2025), arXiv:2507.11495 [hep-
bridge Univ. Press, New York, NY, 2011). ph].
[62] L. Trentadue and G. Veneziano, Phys. Lett. B 323, 201 [90] Y. Guo, F. Yuan, and W. Zhao, (2025),
(1994). arXiv:2507.15820 [hep-ph].
[63] A. Berera and D. E. Soper, Phys. Rev. D 53, 6162 [91] C.-H. Chang, H. Chen, X. Liu, D. Simmons-Duffin,
(1996), arXiv:hep-ph/9509239. F. Yuan, and H. X. Zhu, (2025), arXiv:2507.15923
[64] M. Grazzini, L. Trentadue, and G. Veneziano, Nucl. [hep-ph].
Phys. B 519, 394 (1998), arXiv:hep-ph/9709452. [92] Z.-B. Kang, A. Metz, D. Pitonyak, and C. Zhang,
[65] M. Anselmino, V. Barone, and A. Kotzinian, Phys. (2025), arXiv:2507.17444 [hep-ph].
Lett. B 699, 108 (2011), arXiv:1102.4214 [hep-ph]. [93] E. Herrmann, Z.-B. Kang, J. Penttala, and C. Zhang,
[66] X. P. Chai, K. B. Chen, J. P. Ma, and X. B. Tong, (2025), arXiv:2507.17704 [hep-ph].
JHEP 10, 285 (2019), arXiv:1903.00809 [hep-ph]. [94] A. Bacchetta, M. Radici, and L. Rossi, Phys. Rev. D
[67] K. B. Chen, J. P. Ma, and X. B. Tong, JHEP 11, 038 108, 014005 (2023), arXiv:2303.04314 [hep-ph].
(2021), arXiv:2108.13582 [hep-ph]. [95] K. Cheng and B. Yan, Phys. Rev. Lett. 135, 011902
[68] J. C. Collins, Phys. Rev. D 57, 3051 (1998), [Erratum: (2025), arXiv:2501.03321 [hep-ph].
Phys.Rev.D 61, 019902 (2000)], arXiv:hep-ph/9709499. [96] R. von Kuk, K. Lee, J. K. L. Michel, and Z. Sun,
[69] R. Abdul Khalek et al., Nucl. Phys. A 1026, 122447 (2025), arXiv:2503.22607 [hep-ph].
(2022), arXiv:2103.05419 [physics.ins-det]. [97] P. Schweitzer, M. Strikman, and C. Weiss, JHEP 01,
[70] I. Zurbano Fernandez et al., 10/2020 (2020), 163 (2013), arXiv:1210.1267 [hep-ph].
10.23731/CYRM-2020-0010. [98] K. B. Chen, J. P. Ma, and X. B. Tong, Phys. Rev. D
[71] See Supplemental Material, which contains Refs. [109, 108, 094015 (2023), arXiv:2308.11251 [hep-ph].
110], for the numerically integrated coefficient cross sec- [99] J.-W. Chen and X.-d. Ji, Phys. Rev. Lett. 87, 152002
tions, the estimated degrees of transverse polarization, (2001), [Erratum: Phys.Rev.Lett. 88, 249901 (2002)],
and the final χ2 distributions in numerical form. arXiv:hep-ph/0107158.
[72] C. Bierlich et al., SciPost Phys. Codeb. 2022, 8 (2022), [100] L. Dai, C. Kim, and A. K. Leibovich, JHEP 05, 002
arXiv:2203.11601 [hep-ph]. (2024), arXiv:2310.19207 [hep-ph].
[73] R. von Kuk, J. K. L. Michel, and Z. Sun, JHEP 09, [101] R. von Kuk, J. K. L. Michel, and Z. Sun, JHEP 07,
205 (2023), arXiv:2305.15461 [hep-ph]. 129 (2024), arXiv:2404.08622 [hep-ph].
[74] X.-D. Ji, Phys. Rev. D 49, 114 (1994), arXiv:hep- [102] M. Copeland and T. Mehen, (2024), arXiv:2412.07717
ph/9307235. [hep-ph].
[75] R. L. Jaffe, X.-m. Jin, and J. Tang, Phys. Rev. Lett. [103] Y. Furletova and S. Mantry, Eur. Phys. J. A 57, 315
80, 1166 (1998), arXiv:hep-ph/9709322. (2021), arXiv:2111.03912 [hep-ph].
[76] A. Bianconi, S. Boffi, R. Jakob, and M. Radici, Phys. [104] R. Boughezal, A. Emmert, T. Kutz, S. Mantry,
Rev. D 62, 034008 (2000), arXiv:hep-ph/9907475. M. Nycz, F. Petriello, K. Şimşek, D. Wiegand,
[77] A. Bianconi, S. Boffi, R. Jakob, and M. Radici, Phys. and X. Zheng, Phys. Rev. D 106, 016006 (2022),
Rev. D 62, 034009 (2000), arXiv:hep-ph/9907488. arXiv:2204.07557 [hep-ph].
[78] M. Radici, R. Jakob, and A. Bianconi, Phys. Rev. D [105] J. L. Zhang et al., Nucl. Instrum. Meth. A 1053, 168276
65, 074031 (2002), arXiv:hep-ph/0110252. (2023), arXiv:2207.10261 [hep-ph].
[79] A. Metz and A. Vossen, Prog. Part. Nucl. Phys. 91, 136 [106] R. Boughezal, D. de Florian, F. Petriello, and
(2016), arXiv:1607.02521 [hep-ex]. W. Vogelsang, Phys. Rev. D 107, 075028 (2023),
[80] M. A. Ebert, A. Gao, and I. W. Stewart, JHEP arXiv:2301.02304 [hep-ph].
06, 007 (2022), [Erratum: JHEP 07, 096 (2023)], [107] X.-K. Wen, B. Yan, Z. Yu, and C. P. Yuan, (2024),
arXiv:2112.07680 [hep-ph]. arXiv:2411.13845 [hep-ph].
[81] D. Pitonyak, C. Cocuzza, A. Metz, A. Prokudin, [108] D. Curtin, S. Dreyer, M. F. Costa, S. Heim,
and N. Sato, Phys. Rev. Lett. 132, 011902 (2024), G. Kasieczka, L. Moureaux, D. Rousso, D. Shih, and
arXiv:2305.11995 [hep-ph]. M. Sommerhalder, (2025), arXiv:2506.11192 [hep-ph].
[82] T. C. Rogers, M. Radici, A. Courtoy, and T. Rainaldi, [109] M. A. Ebert, J. K. L. Michel, I. W. Stewart, and F. J.
Phys. Rev. D 111, 056001 (2025), arXiv:2412.12282 Tackmann, JHEP 04, 102 (2021), arXiv:2006.11382
[hep-ph]. [hep-ph].
[83] D. Pitonyak, C. Cocuzza, A. Metz, A. Prokudin, and [110] J. Butterworth et al., J. Phys. G 43, 023001 (2016),
N. Sato, (2025), arXiv:2502.15817 [hep-ph]. arXiv:1510.03865 [hep-ph].
 8

SUPPLEMENTAL MATERIAL

In this supplemental material we collect explicit numerical results for all asymmetry coefficients and total cross
sections. We also list the estimated degrees of transverse polarization and provide the final χ2 distributions of our
sensitivity analysis in numerical form. This material is not required in any way to follow the presentation in the main
text, but is provided purely for the sake of reproducibility and the reference of the reader.

A. Integrated signal and background cross sections

The total unpolarized V H cross section to leading order in perturbation theory is given by
Z X h i
e4 UU
σV = 2
dΦ Cqq ′V fq (xa ) fq̄′ (xb ) + fq̄′ (xa ) fq (xb ) , (S1)
16πEcm ′ q,q

where Ecm is the hadronic center-of-mass energy, dΦ ≡ dYV dYH dpH T is the hard production phase space, and the
PDFs are evaluated at fractions xa,b of the initial-state proton momenta carried by the V H final state. Using Eq. (5),
the signal cross section in Eq. (4) simply becomes

2 X yq  
∆σh,V = Pqh ∆σq,V − Pq̄h ∆σq̄,V , (S2)
π q e

where the coefficients ∆σq,V and ∆σq̄,V are defined as

Z X Z X
e4 TU e4
∆σq,V ≡ 2
dΦ Cqq ′V fq (xa ) f (xb ) ,
q̄ ′ ∆σq̄,V ≡ 2
dΦ CqT′ U
qV fq̄ (xa ) fq ′ (xb ) . (S3)
16πEcm 16πEcm
q′ q′

These coefficients are readily evaluated by performing a numerical integral over the PDFs and coefficients, with
numerical results collected in Table S1. We use the PDF4LHC15 nnlo 100 PDF set [110] and set the factorization scale
to µF = Q. Our electroweak inputs follow Ref. [6, 109]. We use the complete CKM matrix including offdiagonal
UU TU
entries. Analytic expressions for the required hard coefficient functions Cqq ′ V and Cqq ′ V are given in Ref. [45]. Note

that for V = Z we have ∆σq,Z = ∆σq̄,Z by construction because we scaled out the Yukawa couplings and everything
related to the target fragmentation process, leaving only the PDFs, so the a ↔ b symmetry is restored. To numerically
assemble the signal prediction, one also requires — in addition to the estimated DOPs quoted below — the values
of the quark Yukawa couplings at the reference scale µ = mH . These are obtained from the MS quark masses at
the input scale [6], using for simplicity the leading-logarithmic renormalization group evolution of the quark Yukawa

V Z W+ W−
σV [fb] 744.05 827.09 532.49
∆σd,V [fb] 416.42 0 899.39
∆σu,V [fb] 375.45 1566.9 0
∆σs,V [fb] 95.497 0 179.50
∆σc,V [fb] 30.463 169.64 0
∆σb,V [fb] 23.395 0 0.1516
∆σd,V
¯ [fb] 416.42 1516.3 0
∆σū,V [fb] 375.45 0 883.43
∆σs̄,V [fb] 95.497 221.65 0
∆σc̄,V [fb] 30.463 0 194.94
∆σb̄,V [fb] 23.395 0.1591 0

TABLE S1. Integrated total cross sections and signal coefficient cross sections for V H production at Ecm = 14 TeV.
 9

 
q mq (µ0 ) µ0 yq (µ0 ) yq ≡ yq (µ = mH ) αs mq (mq )
d 4.67 MeV 2 GeV 2.68231 × 10−5 1.72783 × 10−5 −
−5 −6
u 2.16 MeV 2 GeV 1.24064 × 10 7.99169 × 10 −
−4 −4
s 93.4 MeV 2 GeV 5.36462 × 10 3.45567 × 10 −
−3 −3
c 1.27 GeV mc 7.29451 × 10 4.32926 × 10 0.30678
−2 −2
b 4.18 GeV mb 2.40087 × 10 1.72720 × 10 0.21217

TABLE S2. Input MS quark masses, input scales, and Yukawa couplings at the input scales (µ0 ) and the reference scale
µ = mH = 125.09 GeV, as used for the numerical analysis in the main text. For reference, we also provide the value of the LL
running coupling at the scale of the heavy quark mass for q = c, b.

couplings and the strong coupling. The numerical values are given in Table S2. We recall that we have scaled out the
Yukawa couplings and DOPs from the ∆σq,V (which massively reduce the latter), and that at the same time the total
semi-inclusive hadron production cross section is still larger than σV by a factor of nh , so there is nothing unphysical
about individual asymmetry coefficients ∆σq,V being larger than the total σV .

B. Estimated degrees of transverse polarization

For the light/light DOPs, we have two doubly-favored flavor combinations Pdπ+ = Puπ− = Pfav2 where the “quark
hole” has the quantum numbers of the pion valence antiquark and the pion valence quark is already present in the
proton. For these cases we take Pfav2 = 1. We stress that even for an estimate of Pqh = 1, the ratio h⊥ qh /fqh ∼
Pqh /nh ≈ 0.1 is still far from its positivity bound ≤ 1, and well below ≲ fπ /mp . We further have two singly-favored
cases Pūπ+ = Pdπ¯ − = Pfav = 0.5, where only the “quark hole” quantum numbers agree, but an antiquark is required
from the remnant. Other combinations of u and d (anti)quarks with π ± are disfavored, Pdis = 0.1.
For heavier quarks Q = s, c, b, it is useful to separately consider the total (symmetrized) DOP P̄Qh ≡ (PQh +PQ̄h )/2
and the sea-quark asymmetry ∆PQh ≡ (PQh − PQ̄h )/2. For the strange/pion DOP we take P̄sh = Pdis (1 ± 3) and, to
be conservative, the extreme case ∆Psh = (0 ± 1)P̄sh . To estimate the heavy/pion DOPs with Q = c, b, we work to
leading power in ΛQCD ≪ mQ and find, using techniques from Ref. [73],

(3) (3)
Λgπ± αs (mQ ) Λgπ±
P̄Qπ± = (1 ± 3) , ∆PQπ± = (0 ± 5) , (S4)
mQ π mQ

(3)
where Λgπ± = 0.3 GeV is a twist-3 gluon/pion FrF. We note that while the matching onto this twist-3 FrF will involve
an additional power of αs (mQ )/π, the latter drops out since the DOPs are defined relative to fQ ∼ (αs /π) fg . The
estimate for the sea asymmetry follows by noting that perturbatively, heavy quarks and antiquarks only become
distinguished by higher-order color structures like dabc dabc .
For the case of reconstructed kaons and D mesons, we simply set the first-generation DOPs to zero, PqH = Pq̄H = 0,
where H = K ± , D± and q = u, d. The strange/kaon DOPs are PsK + = Pfav2 , Ps̄K − = Pfav , and PsK − = Ps̄K + = Pdis .
(3)
For the heavy/kaon DOPs and sea asymmetries involving Q = c, b we again use Eq. (S4) with ΛgK ± = 0.3 GeV as
its central value. For signal charm mesons D+ , estimates for all BMFrFs are essentially perturbative [73]. At leading
power in ΛQCD /mc they only involve the twist-2 gluon PDF fg ∼ fu,d and an O(1) coefficient χD± encoding the
probability for a free c (anti)quark to fragment into the experimentally selected set of D meson states, where we take

αs (mc ) αs2 (mc )

Pc̄D+ = PcD− = , PcD+ = Pc̄D− = PsD+ = PsD− = ,
π π2
αs2 (mb ) αs3 (mb )
P̄bD± = (1 ± 3) , ∆PbD± = (0 ± 5) . (S5)
π2 π3

Here the leading contribution to the valence BMFrF starts at αs2 /π 2 (or αs /π, relative to fc ) and arises from absorptive
contributions involving an additional Wilson-line attachment [73].
 10

C. Combined χ2 statistics

For a given pair of flavors (q, q ′ ), specifically (u, d) for the left panel and (c, s) for the right panel of Fig. 3, we define
the shorthand ∆κ ≡ (∆κq , ∆κq′ )T . Then, since all our constraints are linear, the combined χ2 distributions shown
in Fig. 3 take the form χ2 = ∆κ · A · ∆κ, with a vanishing χ2 = 0 for the best fit at ∆κ = 0 by construction. The
parameter covariance matrices A for the final combined limits are numerically given by
! !
(u,d) 0.00020336 −0.00029289 (u,d) 0.00015997 −0.00021552
Astat = , Astat+syst = (S6)
−0.00029289 0.00065974 −0.00021552 0.00045865

for the light-quark case, while for the heavy-quark case we have
! !
(c,s) 3.70891595 −0.1277616 (c,s) 3.23604087 −0.0528595
Astat = , Astat+syst = . (S7)
−0.1277616 0.0307346 −0.0528595 0.01616964