WorksPhysics and Astronomy
Prospects for ditauonium discovery at colliders
Abstract
The feasibility of observing ditauonium, the bound state of two tau leptons, at colliders (BES III at GeV, Belle II at GeV, a future super tau-charm factory (STCF) at , and the FCC-ee at GeV) as well as in hadronic and photon-photon collisions at the LHC, is studied. Cross sections and expected yields for spin-0 para- () and spin-1 ortho- () ditauonium are presented for nine different production and decay processes. Para-ditauonium can be observed at the FCC-ee via photon fusion in its diphoton decay (). Ortho-ditauonium can be observed at STCF via , where a threshold scan with monochromatized beams can also provide a very precise extraction of the tau lepton mass with a keV) uncertainty or less. Observing pp at the LHC is possible by identifying its displaced vertex with a good control of the combinatorial dimuon background. In addition, we compute the rare decay branching fractions of ditauonium into quarkonium plus a photon.
Introduction
Leptons of opposite electric charge () can form short-lived “onium” bound states under their quantum electrodynamics (QED) interaction. Among the six possible leptonium states—(), (), (), (), (), and ()—only the first two, positronium [1] and muonium [2], have been observed to date. The lightest system, positronium, has been thoroughly studied for precision QED tests [3], and in searches for violations of the discrete space-time CPT symmetries (where C, P, and T are charge conjugation, parity, and time reversal, respectively) either singly or in various combinations [4, 5]. This work focuses on the heaviest leptonium state, ditauonium, proposed originally in [6, 7, 8], and whose detailed spectroscopic properties [9] and first feasibility studies for its production via two-photon fusion at and hadron colliders [10], have been recently presented. Its ground state (principal quantum number ) has two states with total angular momentum and , depending on the relative (opposite or parallel) orientation of its constituent tau leptons, known as para-() and ortho-() ditauonium, respectively. These states, labelled as and using the spectroscopic notation (for total spin and orbital angular momentum), have quantum numbers and , respectively.
The mass of the ditauonium ground state is MeV, with a binding energy of keV, to next-to-next-to-leading-order (NNLO) accuracy in the electromagnetic coupling . This value accounts for a eV downwards Lamb shift, and there is a minuscule hyperfine splitting between the and states [9]. The ditauonium mass uncertainty is fully dominated by the current experimental precision of the tau lepton mass, MeV [11]. The total decay widths of para- and ortho-ditauonium are and meV, corresponding to lifetimes of and fs, respectively [9]. Each of the tau leptons forming ditauonium can itself decay weakly with lifetimes fs [11], about 10 times longer than that of the combined system, and thus ditauonium can be really produced as a bound state. As a matter of fact, and despite being the shortest-lived leptonium, the ditauonium ground states have lifetimes that are 5 to 8 orders-of-magnitude longer than those of similar charmonium states, with masses also in the 3–4 GeV range, that decay much promptly through the strong interaction. The dominant partial decay widths of into diphotons and into dileptons () are known up to NNLO accuracy [9], and amount to
corresponding to branching fractions of and , respectively, with numerical values given for the states, and theoretical uncertainties commensurate with the last digit quoted.
Hadronic decays of the are also possible (), but the final states are less clean and subject to larger backgrounds than the leptonic ones, and not considered here.
The motivation for the production and study of ditauonium is threefold. First, ditauonium can provide new high-precision information about fundamental properties of the free tau lepton itself such as its mass, as explained below, and associated precision tests of the SM (e.g., lepton flavour universality) and of quantities (e.g., the CKM element from decays, or empirical relationships claimed among lepton masses [12]) that parametrically depend on [13]. Second, since the tau lepton is 3500 and 17 times more massive than the electron and muon, respectively, the ditauonium Bohr radius is the smallest of all leptonium systems, and its associated minimum “photon ionization” energy (Rydberg constant), , is the largest. This implies that is the most strongly bound of all leptonia, and that the investigation of its properties can thereby provide new tests of QED and CPT symmetries at much smaller distances than other exotic atoms. Third, ditauonium features enhanced sensitivity to any physics beyond the standard model (BSM) that is suppressed by powers of or affected by uncertainties from hadronic effects, as is the case for, e.g., positronium or muonic-hydrogen states, respectively. Since various hints of violation of lepton flavour universality have appeared in the last years in studies with free leptons [13, 14], it appears relevant to examine if any such effects shows up in the properties of the ditauonium bound states compared to its lighter siblings, positronium and dimuonium.
So far, the only ditauonium production mode considered in detail in the literature is that of exclusive para-ditauonium in two-photon fusion followed by its diphoton decay [10], shown in the (a) panels of Figure 1 and 2. In this work, we consider for the first time seven other ditauonium production and decay modes, including para- and ortho-states (rest of diagrams in both figures). Because of C symmetry and the Landau–Yang selection rule [15, 16], production via (quasireal) photon fusion, as well as decay into two photons, is only possible for the spin-0 para-ditauonium state (diagrams (a) and (b) of Figure 1 and Figure 2). Conversely, the -channel production followed by dilepton decays, which both proceed through spin-1 virtual photon or boson exchanges, are only accessible to ortho-ditauonium (diagrams (d), (e) and (f) of Figure 1, and (c) of Figure 2). Para-ditauonium can also be produced via fusion in annihilation provided the final state contains an extra photon (Figure 1 (c)).

Figure 1. Diagrams for para- (top) and ortho- (bottom) ditauonium production in collisions. Top: Two-photon fusion followed by (a) diphoton and (b) Dalitz decays, s-channel with (c) diphoton decays with final-state radiation (FSR). Bottom: s-channel production with dilepton decays at (d) , and at with (e) initial-state radiation (ISR) and (f) FSR.

Figure 2. Diagrams for ditauonium production at the LHC: exclusive production via two-photon fusion (with 's emitted from hadrons ) followed by (a) diphoton or (b) Dalitz decays, and (c) Drell–Yan-like production with a recoiling gluon.
Experimental considerations
The concrete choices of the ditauonium production and decay processes scrutinized here (Figure 1 and Figure 2) are driven by experimental considerations that can facilitate the observation of the resonance. Potential discovery at four different colliders, as well as at four CERN experiments during the high-luminosity LHC (HL-LHC) phase, will be examined. Among the machines, we consider (i) BES III running at center-of-mass (c.m.) energies GeV with total integrated luminosities up to fb [17], (ii) a future Super-Tau-Charm-Factory (STCF) at integrating up to ab of data per year [18, 19], (iii) the Belle II experiment at Super-KEKB with GeV and ab [20], (iv) the future circular collider (FCC-ee) running at the pole with GeV and ab [21, 22]. Future linear [23, 24], as well [25], colliders running at higher energies will feature integrated luminosities that are likely too low for the production of the rare ditauonium resonances, and will not be considered hereafter. At the HL-LHC, we consider p-p collisions at TeV at ATLAS [26] and CMS [27], covering central pseudorapidities ( for photons and charged leptons) and up to ab expected, and the LHCb experiment [28] that covers the forward acceptance () with fb. Production in the LHC fixed-target mode [29] with the 7 TeV proton beam on a hydrogen gas target, corresponding to p-p collisions at GeV with fb expected at the LHCb/ALICE interaction points, is also discussed. Photon-fusion processes in ultraperipheral collisions (UPCs) [30] of p-p, proton-lead (p-Pb), and lead-lead (Pb-Pb) at the LHC are also studied. The possibility to reconstruct the very soft decay products (charged leptons and/or photons) of ditauonium produced in UPCs appears only feasible for LHCb in p-p collisions, and for the ALICE experiment [31] in the heavy-ion mode with nb and 0.66 pb for Pb-Pb and p-Pb, respectively [32].
In the processes and where ditauonium is produced alone at rest, its decay vertex is indistinguishable from the primary collision vertex. This holds even in the presence of longitudinal boosts because displaced decays along the beam axis fall within the typical collision spot sizes in the direction (e.g., at Belle II and FCC-ee, the luminous region has a typical -width). For these cases, the decay products are swamped by much larger prompt diphoton and/or dilepton (plus photon) backgrounds, and its observation requires the largest possible integrated luminosities. The measurement through resonant -channel production, (Figure 1 (d)), relies on counting the number of dilepton events and observing an excess from the ortho-ditauonium decay above the expectations from background processes that have orders-of-magnitude larger cross sections. Maximizing the signal counts requires also to run collisions with c.m. energies as close as possible to the pole mass of the very narrow Breit–Wigner (BW) peak.
For the rest of the processes, the presence of decay charged leptons accompanied by the additional production of a photon or a jet (in the p-p case), may facilitate the measurement of ditauonium by exploiting its finite decay length (leading to an exponential tail of secondary vertices with proper decay lengths for and , respectively), and the boost given by the extra photon or jet emitted. Both considerations lead to a potentially visible decay away from the or p-p interaction point at an average transverse distance of , free from very large prompt SM physics backgrounds, such as processes. The experimental signature is similar to that of low-mass long-lived dark photons that decay via decays at and p-p colliders [20, 33, 34, 35], with the advantage that the mass, width, and lifetime of the particle are known. Transverse ditauonium boosts of the order of are needed in order to identify the tail of secondary dilepton-decays vertices in the standard silicon trackers with typical vertex position resolutions of the order of . We will therefore provide estimates of the number of ditauonium events with displaced vertices, as a proxy for the signal discovery possibilities. Although decays with take place sufficiently separated from the interaction point to expect negligible backgrounds from prompt SM processes, in particular in collisions, such events will nonetheless be potentially subject to instrumental effects that may difficult its observation. In a real analysis, vertex resolution tails, lepton track misreconstruction, and, in the case of dilepton final states, conversions occurring at larger radii but wrongly extrapolated to a vertex closer to the nominal interaction point, should be taken into account. Resolution effects will also slightly reduce the signal efficiency. All such subtle reconstruction effects must be evaluated with full detector simulations, for both signal efficiency and background rejection, that are beyond the scope of this work.
Results
In the following subsections, we present the expected ditauonium cross sections times branching fractions, and associated number of signal events at various facilities and at the LHC, for all processes of Figure 1 and Figure 2. The quoted ditauonium yields are for the expected integrated luminosities of each machine including simple, but realistic, detector acceptance and mass resolution effects (both effects only introduce small signal losses for modern detectors, as shown in the more detailed data analyses of the simulated data carried out in [10]). We will focus on the production of the para- and ortho-ditauonium ground states (), because excited states (densely spaced a few keV above the mass) have suppressed cross sections [9]. For various processes, if relevant, we provide realistic estimates of the number of background events () and of the expected statistical significance of the signal counts (), given by the ratio in terms of number of standard deviations () above the background-only expectation. For other channels, we provide the expected number of events with a displaced vertex beyond , and/or , as indicative of the feasibility of its observation.
Para-ditauonium via
The two-photon fusion production of para-ditauonium followed by its diphoton decay, , in and UPCs (panels (a) of Figure 1 and Figure 2) has been studied in detail in [10]. Simulated signal and background events have been generated with the HELAC-Onia 2.6.6 Monte Carlo (MC) code [36, 37], complemented with the gamma-UPC photon fluxes [38]. Para-ditauonium is implemented as a modified version of the meson at the mass with a width of MeV (and corresponding lifetime) accounted for by reshuffling the momentum of the resonance according to its associated BW distribution [39]. The measurement is very challenging as the final state is swamped by pairs of photons from decays of charmonium resonances (, , and mesons) plus the light-by-light (LbL) scattering continuum [40]. All such background contributions overlap with the state and have up to 100 times larger expected yields (Table I). Evidence (3) and discovery (5) of para-ditauonium appear only feasible at Belle II and FCC-ee, respectively, by exploiting their full integrated luminosities and carrying out in-situ high-precision measurements of the diphoton widths of the charmonium mesons. Taking advantage of the displaced decay vertex appears hopeless given its production at rest (mostly only longitudinal boosts are present in the fusion of two quasireal photons) and the coarse vertex pointing capabilities of reconstructed photons.
| Colliding system, , , detector | LbL | ||||||
| at 3.78 GeV, 20 fb, BES III | 3.6 ab | 15 ab | 13 ab | 30 ab | 0.25 ab | – | – |
| at 7 GeV, 1 ab, STCF | 0.15 fb | 0.23 fb | 0.33 fb | 0.73 fb | 6.0 ab | 6 | – |
| at 10.6 GeV, 50 ab, Belle II | 0.35 fb | 0.52 fb | 0.77 fb | 1.7 fb | 0.015 fb | 750 | 3 |
| at 91.2 GeV, 50 ab, FCC-ee | 2.8 fb | 3.9 fb | 6.0 fb | 12 fb | 0.11 fb | 5 600 | 5 |
| p-p at 14 TeV, 300 fb, LHCb | 2.0 fb | 2.8 fb | 4.3 fb | 6.3 fb | 80 ab | 24 | – |
| p-Pb at 8.8 TeV, 0.66 pb, ALICE | 6.3 pb | 8.7 pb | 13 pb | 21 pb | 0.25 pb | 0.13 | – |
| Pb-Pb at 5.5 TeV, 14 nb, ALICE | 15 nb | 21 nb | 31 nb | 62 nb | 0.59 nb | 6.4 | – |
Table I. Photon-fusion production cross sections times diphoton branching fraction, , for the para-ditauonium signal and backgrounds (overlapping C-even charmonium resonances, and LbL scattering over ) at various facilities, and in p-p, p-Pb, and Pb-Pb UPCs at the LHC. The last two columns list the total produced events for the integrated luminosities at each collider, and the expected statistical significance [10].
Para-ditauonium via
The two-fusion production cross sections of shown in panels (b) of Figure 1 and Figure 2, are computed with the same setup used in the previous section, but now the Dalitz decays are considered with widths [9],
where the numerical values are quoted for the ground state (), and lead to branching fractions of and for , respectively. Table II lists the expected para-ditauonium production cross sections times Dalitz decay branching fractions, as well as the corresponding signal yields for the colliding systems considered. In UPCs at the LHC, the cross sections are small for the comparatively low values in heavy-ion collisions, and less than one produced event is expected. In collisions, the cross sections are tiny, reaching 3.2 ab at most, but since the integrated luminosities are very large the number of expected events, combining dielectron and dimuon channels, are 20 and 160 at Belle II and FCC-ee, respectively. Unfortunately, the background cross sections from Dalitz decays of charmonium mesons and the continuum have expected cross sections many orders of magnitude larger than that of the signal. Although Dalitz decays of , , and have not yet been observed, since this decay, , shares the same underlying diagram as the diphoton case, one expects the same relative cross sections shown in Table I for the diphoton channel. Given the factor of 35 lower number of Dalitz, compared to diphoton, signal events, one does not expect any significant . Also, the transverse boost of produced via photon fusion at Belle II and FCC-ee is very small, , and one will not be able to separate its secondary vertex either. Observation of this exotic atom appears therefore unfeasible in this mode.
| Colliding system, , , detector | |||
| at 3.78 GeV, 20 fb, BES III | 7.2 zb | – | – |
| at 7 GeV, 1 ab, STCF | 0.18 ab | 0.14 | 0.04 |
| at 10.6 GeV, 50 ab, Belle II | 0.44 ab | 17 | 5 |
| at 91.2 GeV, 50 ab, FCC-ee | 3.22 ab | 125 | 36 |
| p-p at 14 TeV, 300 fb, LHCb | 2.3 ab | 0.5 | 0.2 |
| p-Pb at 8.8 TeV, 0.66 pb, ALICE | 7.5 fb | – | – |
| Pb-Pb at 5.5 TeV, 14 nb, ALICE | 17.5 pb | 0.25 | – |
Table II. Photon-fusion production cross sections times Dalitz decay branching fraction, with , and expected number of events for the para-ditauonium signal at various facilities and in ultraperipheral collisions at the LHC.
Para-ditauonium via e+e- to TT0(gammagamma)+gamma, with an FSR photon
The cross section to produce para-ditauonium accompanied with an FSR photon in annihilation via -channel virtual or fusion (Figure 1 (c)) is given by the following expression
where and are the Z boson mass and width, and and are sine and cosine of the Weinberg angle. In the infinite boson mass limit, the cross section reads
Figure 3 shows the corresponding cross section as a function of c.m. energy over –100 GeV, with the inset displaying a zoom in the –10 GeV peak region. Starting at the threshold, the cross section increases with energy up to a maximal value of 1.44 ab reached at GeV, followed by a steady decrease for higher , except for a clear BW structure at GeV corresponding to the production and ditauonium-mediated decay of the Z boson into three photons (albeit with negligible cross sections)3. In order to retain as much as possible of the cross section, we focus on the dominant diphoton decay. The corresponding cross sections and yields for collisions at various colliders are listed in Table III.

Figure 3. Cross section (in ab) for the production of para-ditauonium plus an FSR photon, , as a function of the c.m. energy over – GeV. The inset shows a zoom in the – GeV region.
| Colliding system, , , detector | ||
| at 4.3 GeV, 1 ab, STCF | 1.1 ab | 1 |
| at 7 GeV, 1 ab, STCF | 0.37 ab | 0.37 |
| at 3.78 GeV, 20 fb, BES III | 0.69 ab | 0.014 |
| at 10.6 GeV, 50 ab, Belle II | 0.085 ab | 4 |
| at 91.2 GeV, 50 ab, FCC-ee | ab | – |
Table 3. Para-ditauonium production cross sections in annihilation accompanied by an FSR photon, , and corresponding expected number of events at various facilities.
Finite yields are only expected at Belle II, , thanks to its much larger integrated luminosity compared to the rest of low-energy machines. Such low number of signal counts, and a triphoton final-state with poor secondary vertex resolution, precludes a search based on the para-ditauonium displaced vertex tail. A search based on identifying an excess above the continuum appears also hopeless as prompt backgrounds sharing the same final state have notably larger cross sections. Using HELAC-Onia, we find that the production of three photons with at least one pair having an invariant mass within MeV, and the three photons satisfying , yields pb ab events at Belle II. In addition, the production of a meson plus a photon shares the same final state and mass window, and results in ab ab counts. With such daunting number of background events, measuring the channel considered here appears experimentally unfeasible.
Ortho-ditauonium via
The cross section for resonant production in collisions at a c.m. energy is theoretically given by the relativistic BW expression:
For an run exactly at the pole mass, , this cross section amounts to . However, similar to the case of resonant Higgs boson production at the FCC-ee, [41], two effects reduce substantially this purely theoretical cross section. First, the actual beams are never perfectly monoenergetic but have a distribution of incoming energies that is many orders-of-magnitude wider than the width of the very narrow ditauonium resonance. Secondly, the occurrence of ISR photon emission from the beams will shift the actual collision energy below the BW peak. Both effects render the term much larger than in the denominator of Eq. (6), and strongly reduce the actual cross section. For a Gaussian c.m. energy spread of width (with ), the actual -channel cross section with both effects convolved can be written as
where the and structure functions describe the distribution of energy of each beam reduced by a fraction after collinear initial photon emission, and where the relativistic Voigtian function reads
which can be expressed in terms of the complementary error function [42]. In order to carry out the two integrations in Eq. (7), we use the leading-logarithmic expressions for , and the parametrization of the fractions of beam energies , of Ref. [43]. With those ingredients at hand, one can compute the cross sections for any and values. Figure 4 (left) shows the resonant cross section in collisions at and as a function of the c.m. spread . Assuming that one runs exactly at the mass, the cross section decreases very rapidly, following a dependence (top black curve). If the c.m. energy of the collision misses the mass of the resonance by (middle green curve) or by (bottom red curve), the cross section peaks at for a spread, and decreases to the left and right of it. The drop in the cross section in the region is much smaller for the -keV shift, than for the -keV case, because running above the resonance profits from possible ISR radiative returns that bring the c.m. energy onto the BW peak.

Figure 4. Resonant cross section in collisions at , as a function of c.m. spread (left), and as a function of minus the mass for fixed with monochromatized beams (right). In the right plot, the different symbols indicate the expected measured cross sections for scan points (with each) where the actual mass () coincides with the then-current value (black dots), or is shifted by (blue and magenta squares) from it (see text for details).
The experimental observation of resonant production is a “counting experiment” whereby one measures the number of dilepton events searching for an excess issuing from the ortho-ditauonium decays above the expectations from the background-only processes. We focus on the decay mode which has smaller continuum backgrounds than the channel. As discussed in the direct Higgs boson production in collisions at at the FCC-ee [44], in order to maximize the cross section one needs to: (i) run at a value as close as possible to the true ditauonium mass (known today with a 250-keV uncertainty, from twice the tau-lepton mass value precision), (ii) to be able to measure and calibrate the actual machine within the same high accuracy, and (iii) to reduce the c.m. energy spread down to values of the same order. Regarding the first point, the BES III collaboration performed a fine mass scan experiment in 2018 with five points around the tau pair production threshold with a spread of MeV, and a total integrated luminosity of 140 pb [45]. The final expected uncertainty in will be around 50 keV. Similarly, Belle II forecasts to reach tau mass uncertainties below 150 keV (systematic and statistical sources combined) with about 300 fb of data [46]. Combining these results, plus all others from the PDG world average [11], it is not inconceivable to reach a precision of keV) in the tau mass (i.e., twice this value, keV), in the mass) whenever a dedicated search for the ditauonium is carried out at a future STCF. Regarding the calibration, the BES-III experiment has already achieved a high accuracy, at the level of , by exploiting the measurement of monochromatic laser photons backscattered from the beams in its beam energy measurement system (BEMS) [47]. The application of the same technique at three different lepton colliders can reach accuracies of the actual c.m. energy not worse than 50 keV [48]. Last but not least, on the c.m. energy spread, the possibility of monochromatization of electron/positron beams has been considered several times in the literature [49, 50, 51], with values of keV at the tau-pair production threshold theoretically achievable.
Table IV lists the expected resonant cross sections and number of events at various facilities. We list first the two BES-III scan runs performed around the -pair threshold in 2011 and 2018 with pb and MeV, respectively [45, 52]. For each of these two cases, we expect about 2 and 60 events produced respectively, which are too small numbers to be observed on top of the orders-of-magnitude larger dimuon continuum background. On the other hand, the STCF is expected to integrate 1 ab around the threshold with a default MeV spread [18, 19]. One single run under these conditions will produce more than half-a-million particles decaying into dimuons, enough to observe its production with a significance around . For the STCF facility, we consider in addition the possibility to monochromatize the beams down to keV, albeit with a factor of ten loss in the integrated luminosity. Such assumptions should be considered as conservative given that spreads as low as keV are expected at these colliding energy (albeit with corresponding losses in the beam luminosities) [50, 51]. With such a monochromatization working point, we expect more than 2.2 million ditauonium events produced per 0.1-ab scan point that would allow a detailed study of the resonance. In particular, if the c.m. energy is calibrated to within 50-keV with the BEMS (or any similar) technique, one can then determine the ditauonium mass within this precision by carrying out a scan with e.g., five collision points around the production threshold even with a reduced 0.1 ab integrated luminosity per point. With such a setup, one would be able to determine the position of the mass peak within an accuracy just driven by the keV) calibration value. This is illustrated in Figure 4 (right) where the -channel cross section around the pole for a fixed keV spread is shown. The three sets of 5-scan points show the expected extracted cross sections (with error bars approximating the statistical uncertainties only) for three possible cases where the actual ditauonium mass corresponds to the assumed PDG mass at the moment of performing the experiment, or where it is actually shifted by +50 keV with respect to it. The plot shows clearly that a Gaussian fit (dashed curves) to whichever measured set of five cross section points should provide a peak value whose position will only be driven by the accuracy in the knowledge of the c.m. energy calibration. Such an extraction of the value with keV uncertainty would allow a direct determination of the lepton mass, (with keV), to within keV. Given that tau lepton decays into final states with invisible neutrinos that complicate the accurate measurement of its mass (and that any approach based on a high-luminosity threshold scan will inevitably produce ditauonium, even without monochromatized beams, as shown in Table V), no other -determination method can likely match the precision and accuracy of the ditauonium-mass presented here. We note that the cross section is about 2% (50% for monochromatized beams) of the total tau-pair production for the collisions at considered in Table IV. Namely, about 2% of the pairs produced in standard threshold scans will form an ortho bound state. It is obvious from this study that ditauonium should be an integral part of the experimental programme of precision studies of the tau lepton at any future STCF.
| Colliding system, ( spread), , experiment | |||||||
| at 3.5538 GeV (1.47 MeV), 5.57 pb, BES III | 1.9 pb | 117 pb | 6.88 nb | 10.4 | 2.1 | 38 300 | 0.01 |
| at (1.24 MeV), 140 pb, BES III | 2.2 pb | 103 pb | 6.88 nb | 310 | 63 | 0.06 | |
| at (1 MeV), 1 ab, STCF | 2.6 pb | 95 pb | 6.88 nb | ||||
| at (100 keV), 0.1 ab, STCF | 22 pb | 46 pb | 6.88 nb |
Table IV. Cross sections and expected number of events for the -channel production of ortho-ditauonium (), and for the and (background) continua, in at at various facilities. The last column lists the expected signal statistical significance.
| Colliding system, , , detector | |||||||
| at 3.78 GeV, 20 fb, BES III | 3.0 fb | 18.3 pb | 6.4 nb | 12 | 12 | ||
| at 3.78 GeV, 1 ab, STCF | 3.0 fb | 18.3 pb | 6.4 nb | 610 | 610 | ||
| at 10.6 GeV, 50 ab, Belle II | 36 ab | 223 fb | 10.5 pb | 370 | 370 | ||
| at 91.2 GeV, 50 ab, FCC-ee | 0.61 ab | 3.8 fb | 10 fb | 6 | 6 |
Table V. Production cross sections and expected yields for ortho-ditauonium and prompt dilepton backgrounds in annihilation with ISR at four different colliders. The backgrounds are computed within and .
Ortho-ditauonium via in ISR collisions
The cross section to produce ortho-ditauonium in annihilation at after ISR emission (Figure 1 (e), with a collinear photon) is given by
with the same parameters defined in Eq. (4), and where the last factor encodes the change in the ditauonium cross section due to the emission of ISR photon(s). This latter factor is defined as an effective luminosity in terms of the electron and positron structure functions, and ,
where is the fraction of the beam momentum carried out by the after collinear ISR emission. This effective luminosity can be evaluated either via MC methods [53] or in a parton-distribution-function (PDF)-like approach [54, 55, 56, 57]. Again, in the infinite boson mass limit, a more handy formula is available
Here, we use the effective luminosity at leading-logarithmic accuracy from Eq. (67) of Ref. [55] with the so-called ‘mixed’ prescription that partially accounts for a few extra non-leading terms in the cross section. Table V lists the corresponding ISR ortho-ditauonium cross sections () at four different colliders. The largest signal cross sections, in the few fb range, are for BES-III and STCF running at a above, but not too far from, the production threshold. About 25 and 1200 events are expected in the combined dilepton decays at each machine, respectively. At Belle II, despite two orders-of-magnitude smaller cross sections, one expects about 750 dilepton signal events thanks to its much larger integrated luminosities, whereas the FCC-ee energies are too large for ISR ditauonium production. In all cases, the number of prompt dilepton continuum events below the peak is, however, orders-of-magnitude larger, and the possibility to observe its displaced vertex decay is negligible given that the ISR photons are emitted collinear to the beam and do not impart any appreciable transverse boost to the resonance. In the next section, the more favorable case where the ISR/FSR is emitted at a large angle is discussed.
Ortho-ditauonium via with a large-angle photon
In order to transversely boost the ortho-ditauonium and facilitate the observation of its displaced decay vertex, we consider where the emitted photon has a large scattering angle (i.e., it is not collinear with the beams, ) in the rest frame of the and system (Figure 1 (e) and (Figure 1f) for a non-collinear photon). In the above frame, the differential cross section with respect to the cosine of the relative angle is
where and . For or , the cross section has the asymptotic limit
The scalar three-momentum in the same frame is fixed at , where is the invariant mass square of the final system. At Belle II collisions with GeV, the three-momentum in the rest frame is GeV. One can integrate Eq. (13) over from to , where are positive real numbers smaller than one, to avoid the collinear divergence of the term for . The derived cross sections and yields values for expected at different facilities can only be a fraction of the ISR ones quoted in Table V, because the underlying physical process is the same except that now we require a hard photon within the detector acceptance that boosts the in the opposite direction. Given the low number of events already expected for ISR collisions at BES-III and FCC-ee (Table V), there is no chance to see any event of this sort at both machines. Similarly, the STCF collision energies or integrated luminosities will not be adequate to produce a photon hard enough to boost the ditauonium resonance. One is therefore constrained to just consider the Belle II case, for which the cross sections, where the photon is emitted within the detector acceptance () are given in Table VI. The prompt dilepton-plus-photon backgrounds are many orders of magnitude larger so only displaced vertices will make the visible. Simulated signal events are generated with HELAC-Onia and passed through the Belle-II acceptance for the final-state particles. Combining dielectron and dimuon decays, the total expected number of events is with only one event expected in the tail of the displaced decays. The measurement at Belle-II is therefore certainly challenging, but given that the mass and width of the resonance are known, the search at values away from the interaction point would be worth a try.
| Colliding system, , , detector | with | ||||||
| at 10.6 GeV, 50 ab, Belle II | 7.1 ab | 51 fb | 1.9 pb | 73 | 73 | 0.5 | 0.5 |
Table VI. Production cross sections and expected yields for ortho-ditauonium and prompt dilepton backgrounds (), plus a hard large-angle photon (), in annihilation at Belle II. The prompt backgrounds are computed for (for , we also require the angle between and to be to avoid final-state collinear divergences). The last columns list the expected events with a displaced vertex beyond 30 m.
Ortho-ditauonium via pp to
Ortho-ditauonium can be produced in p-p collisions through the Drell–Yan-like process shown in Figure 2. At LO in pQCD, the cross section for at a given nucleon c.m. energy can be obtained from
where are PDFs evaluated at parton momentum fractions and factorization scale , is the squared partonic invariant mass, , and the partonic cross section of quarks of charge is
The NLO cross sections for inclusive production at the LHC are obtained using the expressions above with the CT14nlo PDFs [58] and an NLO/LO ratio computed with MadGraph5_aMC@NLO v2.6.6 [59] from the Drell–Yan pp processes with the invariant mass of the dimuon system set at , and the central scale set at the transverse mass of the pair. The corresponding ortho-ditauonium () cross sections are listed in the first column of Table VII for p-p collisions at TeV, and GeV (fixed target). The quoted uncertainties of 30–40% are derived from the usual independent variations of theoretical scales around within a factor of two. The PDF uncertainties are negligible compared to those.
| Colliding system, , , detector | with | |||||
| p-p at 14 TeV, 3 ab, ATLAS/CMS | fb | fb | 1100 | 1100 | 130 (10) | 130 (10) |
| p-p at 14 TeV, 300 fb, LHCb | fb | fb | 110 | 110 | 5 (–) | 5 (–) |
| p-p at 114.6 GeV, 10 fb, ALICE/LHCb | fb | fb | – | – |
Table VII. Ortho-ditauonium cross sections and yields for p-p collisions in collider and fixed-target modes at the LHC. The first columns list the NLO cross sections for inclusive p-p and p-p (with GeV) production. The second columns give the expected number of events within the LHC detectors acceptances. The last columns list the expected number of events with the additional requirement of a secondary vertex with .
The process of interest is that where is accompanied by an away-side (mini)jet, p-p , that provides the resonance with a non-zero and allows the use of the displaced decay vertex to identify it above the large Drell–Yan p-p background. The second column of Table VII details the cross section for ditauonium production accompanied by a (mini)jet, , with GeV. The NLO -factor for this process is , as derived also with MadGraph5_aMC@NLO. The produced number of boosted events as function of a given minimum are shown in Figure 5 for the ATLAS/CMS (left) and LHCb (right) integrated luminosities. For the full HL-LHC programme, one expects about 10 events produced with GeV (corresponding to boosts) in LHCb and ATLAS/CMS, respectively.

Figure 5. Expected number of p-p events at TeV above a given minimum within the ATLAS/CMS (left) and LHCb (right) acceptances.
The expected numbers of actually visible events, passing the typical HL-LHC detector acceptance cuts, are listed in the second columns of Table VII. The requirement of both decay leptons to be within the ATLAS/CMS (LHCb) acceptance of () and have GeV ( GeV), reduces the yields by about a factor of ten. Thus, a bit more than one thousand events are expected to be produced at ATLAS and CMS, one hundred at LHCb, and just a bunch of them in the fixed-target mode. The number of expected events with a secondary vertex displaced by at least 30 m with respect to the primary vertex, are listed in the last columns of Table VII. At ATLAS and CMS, about 130 (10) events with (100) m are expected, indicating that its observation is possible given that, in principle, there is no known long-lived particle that produces a dimuon peak at . A good control of the reducible combinatorial backgrounds is, however, required. A detailed experimental analysis goes beyond the scope of this paper, but we can outline a few key elements that can facilitate the measurement.
Experimental studies of long-lived dimuon resonance searches in ATLAS and CMS [60, 61, 62] indicate that displaced dimuons originate dominantly from different heavy quarks decays, and partially (mis)reconstructed or misidentified muons from secondary charged pion or kaon decays. The muon isolation requirement strongly suppresses copious backgrounds from dijet and multijet events that yield uncorrelated (either genuine or misidentified) pairs of muons from particles decays within different jets. The signed difference in azimuthal angles (transverse collinearity angle) between the of the dimuon system and provides an important discriminating variable between signal and backgrounds. By requiring a significant background reduction is achieved, and one is left with contributions from background sources that yield dimuons exclusively or predominantly at small including: dimuon decays of nonprompt low-mass resonances such as or mesons from B-hadron decays; cascade decays of B hadrons; and dimuons formed from a pair of unrelated nonprompt in the same jet. If well reconstructed, most such background events will not have a consistent with the mass and width requirements. Nonetheless, if needed to maximize the analysis sensitivity, multivariate analyses combining multiple discriminating observables in a single powerful discriminator, such as e.g., a boosted decision tree algorithm, can be further employed.
Ditauonium decays into quarkonia-plus-photon,
Ditauonium can decay into final states containing a light quark-antiquark pair (), either directly in the ortho-ditauonium case (), or via the Dalitz channel for para-ditauonium where one photon splits into the pair (). In principle, the decay is not kinematically accessible because the lightest charm D meson has a mass that is more than half the ditauonium mass. It is however possible for ditauonium to decay into light charmonium states plus a photon, via and (Figure 6). The corresponding

Figure 6. Diagrams for para- (left) and ortho- (right) ditauonium production in and p-p collisions followed by decays.
partial widths can be derived from the NNLO expressions for Dalitz decays and for diquark-plus-photon decays respectively, multiplied by the standard long-distance matrix elements (LDMEs), . LDMEs describe the formation of each quarkonium system in the non-relativistic QCD (NRQCD) framework, as an expansion in terms of the relatively low velocity of the bound charm quarks [63]. The partial widths for the quarkonium decay modes, suppressed by a factor and a small phase space compared to the leading decay channels, read
where the numerical results correspond to the ground states, the QCD coupling evaluated at the scale is taken as , and keV is the dielectron partial decay width [11]. The uncertainty of the branching fraction in Eq. (16) is small because it is expressed in terms of the experimentally well-known value. The decay width, Eq. (17), uses the LDME of for its numerical evaluation, with an uncertainty estimated by using the different values of the wavefunction at the origin given in [64]. In addition, relativistic corrections (i.e., higher order in ) and missing higher-order terms can both change this width by . The latter two uncertainties have not been included in the estimate quoted here. The corresponding decay branching fractions amount to:
Such values are so low that experimental searches for ditauonium in the quarkonium + photon decay channels appear hopeless. The decay is also possible, but it is expected to be further -suppressed compared to . In addition, is also kinematically allowed. Its computation requires the knowledge of the transition, and may have larger phase space suppression due to . We refrain ourselves from considering this hadronic channel here.
Summary
We have evaluated the possibility to experimentally produce and observe ditauonium, the bound state of two tau leptons, in high-energy and hadron collisions. Ditauonium is the heaviest and most compact purely leptonic “atomic” system, and remains experimentally unobserved to date. Its experimental study can provide a high-precision extraction of the tau mass, as well as novel QED tests sensitive to physics beyond the standard model, such as e.g., violations of lepton flavour universality, that may not impact its lighter siblings (positronium and dimuonium). Nine production and decay channels have been studied at four different facilities (BES III, Belle II, STCF, and FCC-ee) as well in hadronic and photon-fusion collisions at the LHC. The overall conclusion is that the observation of the para- and/or ortho-ditauonium states is challenging but feasible.
Para-ditauonium can be measured at Belle II and FCC-ee via photon fusion in its dominant diphoton decay with branching fraction, provided the large background from the overlapping decay can be precisely controlled in-situ. On the other hand, its Dalitz decay with , will likely remain unobserved. Ortho-ditauonium can be observed in the -channel in a normal run at a future STCF operating at the threshold. In addition, running with monochromatized beams, leading to a c.m. energy spread, the ditauonium cross sections are large, and its mass can be determined from a threshold scan within a precision just given by the c.m. energy calibration. For a typical c.m. energy calibration, the mass of the tau lepton can be obtained with an uncertainty half this value, . Such a ditauonium-based approach will likely become the most precise and accurate experimental means to determine the mass. Ditauonium should thus be incorporated as an integral part of the experimental programme of precision studies of the tau lepton at any future STCF. The ISR production of ortho-ditauonium has too low cross sections at all colliders to be observable over the huge prompt dilepton backgrounds, unless one requires the production of a large-angle photon that boost the resonance decay away from the interaction point. This latter possibility may lead to the observation of a few events with displaced transverse decays at Belle II. Finally, the associated production with a jet, , can be observed by ATLAS and CMS at the HL-LHC by measuring its displaced dimuon vertex beyond , with a good control of the large instrumental combinatorial dimuon backgrounds. Last but not least, the very rare decay branching fractions of ditauonium into quarkonium plus a photon, and , have been computed for the first time.
Acknowledgments.— H.-S.S. thanks S. Frixione for discussions about the electron structure function. Supports from the European Union’s Horizon 2020 research and innovation programme (grant agreements No. 824093, STRONG-2020, EU Virtual Access “NLOAccess”, and No. 951754 “FCC Innovation Study”), the ERC grant (grant agreement ID 101041109, “Boson”), the French ANR (grant ANR-20-CE31-0015, “PrecisOnium”), and the CNRS IEA (grant No. 205210, “GlueGraph”), are acknowledged.
References
- [1]M. Deutsch, “Evidence for the formation of positronium in gases,” Phys. Rev. 82 (1951) 455. [2] V. W. Hughes, D. W. McColm, K. Ziock, and R. Prepost, “Formation of muonium and observation of its Larmor precession,” Phys. Rev. Lett. 5 (1960) 63.
- [3]S. G. Karshenboim, “Precision physics of simple atoms: QED tests, nuclear structure and fundamental constant,” Phys. Rept. 422 (2005) 1, arXiv:hep-ph/0509010 [hep-ph].DOI
- [4]W. Bernreuther, U. Löw, J. P. Ma, and O. Nachtmann, “How to test CP, T and CPT invariance in the three photon decay of polarized s-wave triplet positronium,” Z. Phys. C 41 (1988) 143.
- [5]T. Yamazaki, T. Namba, S. Asai, and T. Kobayashi, “Search for CP violation in positronium decay,” Phys. Rev. Lett. 104 (2010) 083401, arXiv:0912.0843 [hep-ex]. [Erratum: Phys. Rev. Lett. 120, 239902 (2018)].DOI
- [6]J. W. M. Doeft, “Does a heavy positronium atom exist?,” Phys. Rev. Lett. 35 (1975) 1605.
- [7]C. Avilez, R. Montemayor, and M. Moreno, “Tauonium: τ+τ−, a bound state of heavy leptons,” Lett. Nuovo Cim. 21 (1978) 301.DOI
- [8]C. Avilez, E. Ley Koo, and M. Moreno, “Comments on the observability of the tauonium tau, the bound state of τ+τ−,” Phys. Rev. D 19 (1979) 2214.DOI
- [9]D. d’Enterria, R. Perez-Ramos, and H.-S. Shao, “Ditauonium spectroscopy,” Eur. Phys. J. C 82 (2022) 923, arXiv:2204.07269 [hep-ph].
- [10]D. d’Enterria and H.-S. Shao, “Observing true tauonium via two-photon fusion at e+e− and hadron colliders,” Phys. Rev. D 105 (2022) 093008, arXiv:2202.02316 [hep-ph].
- [11]PDG Collaboration, P. A. Zyla et al., “Review of Particle Physics,” PTEP 2020 (2020) 083C01.
- [12]Y. Koide, “A New View of Quark and Lepton Mass Hierarchy,” Phys. Rev. D 28 (1983) 252.DOI
- [13]HFLAV Collaboration, Y. Amhis et al., “Averages of b-hadron, c-hadron, and τ-lepton properties as of 2021,” arXiv:2206.07501 [hep-ex].
- [14]S. Bifani, S. Descotes-Genon, A. Romero Vidal, and M.-H. Schune, “Review of Lepton Universality tests in B decays,” J. Phys. G 46 (2019) 023001, arXiv:1809.06229 [hep-ex].DOI
- [15]L. D. Landau, “On the angular momentum of a system of two photons,” Dokl. Akad. Nauk SSSR 60 (1948) 207.
- [16]C.-N. Yang, “Selection rules for the dematerialization of a particle into two photons,” Phys. Rev. 77 (1950) 242.DOI
- [17]BESIII Collaboration, M. Ablikim et al., “Future Physics Programme of BESIII,” Chin. Phys. C 44 (2020) 040001, arXiv:1912.05983 [hep-ex].
- [18]Charm-Tau Factory Collaboration, A. E. Bondar et al., “Project of a Super Charm-Tau factory at the Budker Institute of Nuclear Physics in Novosibirsk,” Phys. Atom. Nucl. 76 (2013) 1072.
- [19]STCF Working Group Collaboration, X. Zhou, “Experimental Program at Super Tau-Charm Facility,” PoS CHARM2020 (2021) 007.
- [20]Belle II Collaboration, W. Altmannshofer et al., “The Belle II Physics Book,” PTEP 2019 (2019) 123C01, arXiv:1808.10567 [hep-ex]. [Erratum: PTEP 2020, 029201 (2020)].
- [21]FCC Collaboration, A. Abada et al., “FCC Physics Opportunities: Future Circular Collider Conceptual Design Report Volume 1,” Eur. Phys. J. C 79 (2019) 474.
- [22]FCC Collaboration, A. Abada et al., “FCC-ee: The Lepton Collider: Future Circular Collider Conceptual Design Report Volume 2,” Eur. Phys. J. ST 228 (2019) 261.
- [23]ILC International Development Team Collaboration, A. Aryshev et al., “The International Linear Collider: Report to Snowmass 2021,” arXiv:2203.07622 [physics.acc-ph].
- [24]O. Brunner et al., “The CLIC project,” arXiv:2203.09186 [physics.acc-ph].arxiv.org/abs/2203.09186
- [25]N. Bartosik et al., “Detector and Physics Performance at a Muon Collider,” JINST 15 no. 05, (2020) P05001, arXiv:2001.04431 [hep-ex].DOI
- [26]ATLAS Collaboration, G. Aad et al., “The ATLAS Experiment at the CERN Large Hadron Collider,” JINST 3 (2008) S08003.
- [27]CMS Collaboration, S. Chatrchyan et al., “The CMS Experiment at the CERN LHC,” JINST 3 (2008) S08004.
- [28]LHCb Collaboration, R. Aaij et al., “Physics case for an LHCb Upgrade II - Opportunities in flavour physics, and beyond, in the HL-LHC era,” arXiv:1808.08865 [hep-ex].
- [29]C. Hadjidakis et al., “A fixed-target programme at the LHC: Physics case and projected performances for heavy-ion, hadron, spin and astroparticle studies,” Phys. Rept. 911 (2021) 1, arXiv:1807.00603 [hep-ex].
- [30]A. J. Baltz, “The physics of ultraperipheral collisions at the LHC,” Phys. Rept. 458 (2008) 1, arXiv:0706.3356 [nucl-ex].DOI
- [31]ALICE Collaboration, K. Aamodt et al., “The ALICE experiment at the CERN LHC,” JINST 3 (2008) S08002.
- [32]D. d’Enterria et al., “Opportunities for new physics searches with heavy ions at colliders,” arXiv:2203.05939 [hep-ph].
- [33]BaBar Collaboration, J. P. Lees et al., “Search for a dark photon in e+e− collisions at BaBar,” Phys. Rev. Lett. 113 (2014) 201801, arXiv:1406.2980 [hep-ex].DOI
- [34]BESIII Collaboration, M. Ablikim et al., “Dark photon search in the mass range between 1.5 and 3.4 GeV/c2,” Phys. Lett. B 774 (2017) 252, arXiv:1705.04265 [hep-ex].
- [35]LHCb Collaboration, R. Aaij et al., “Search for A′ → μ+μ− decays,” Phys. Rev. Lett. 124 (2020) 041801, arXiv:1910.06926 [hep-ex].
- [36]H.-S. Shao, “HELAC-Onia: An automatic matrix element generator for heavy quarkonium physics,” Comput. Phys. Commun. 184 (2013) 2562, arXiv:1212.5293 [hep-ph].DOI
- [37]H.-S. Shao, “HELAC-Onia 2.0: an upgraded matrix-element and event generator for heavy quarkonium physics,” Comput. Phys. Commun. 198 (2016) 238, arXiv:1507.03435 [hep-ph].arxiv.org/abs/1507.03435
- [38]H.-S. Shao and D. d’Enterria, “gamma-UPC: Automated generation of exclusive photon-photon processes in ultraperipheral proton and nuclear collisions with varying form factors,” arXiv:2207.03012 [hep-ph].DOI
- [39]S. Frixione, B. Fuks, V. Hirschi, K. Mawatari, H.-S. Shao, P. A. Sunder, and M. Zaro, “Automated simulations beyond the Standard Model: Supersymmetry,” JHEP 12 (2019) 088, arXiv:1907.04988 [hep-ph].
- [40]D. d’Enterria and G. G. da Silva, “Observing light-by-light scattering at the Large Hadron Collider,” Phys. Rev. Lett. 111 (2013) 080405, arXiv:1305.7142 [hep-ph]. [Erratum: Phys. Rev. Lett. 116, 129901 (2016)].
- [41]D. d’Enterria, A. Poludniak, and G. Wojcik, “Measuring the electron Yukawa coupling via resonant s-channel Higgs production at FCC-ee,” Eur. Phys. J. Plus 137 (2022) 201, arXiv:2107.02686 [hep-ex].
- [42]R. A. Kycia and S. Jadach, “Relativistic Voigt profile for unstable particles in high energy physics,” J. Math. Anal. Appl. 463 (2018) 1040, arXiv:1711.09304 [math-ph].DOI
- [43]S. Frixione, O. Mattelaer, M. Zaro, and X. Zhao, “Lepton collisions in MadGraph5_aMC@NLO,” arXiv:2108.10261 [hep-ph].arxiv.org/abs/2108.10261
- [44]D. d’Enterria, “Collider constraints on axion-like particles,” in FIPS 2020, CERN. 2021. arXiv:2102.08971 [hep-ex].
- [45]BES-III Collaboration, M. N. Achasov, X. H. Mo, N. Y. Muchnoi, I. B. Nikolaev, S. A. Privalov, and J. Y. Zhang, “Tau mass measurement at BES-III,” EPJ Web Conf. 212 (2019) 08005.
- [46]Belle-II Collaboration, F. Abudinén et al., “τ lepton mass measurement at Belle II,” arXiv:2008.04665 [hep-ex].
- [47]E. V. Abakumova et al., “The Beam Energy Measurement System for the Beijing electron-positron collider,” Nucl. Instrum. Meth. A 659 (2011) 21, arXiv:1109.5771 [physics.acc-ph].
- [48]M. N. Achasov and N. Y. Muchnoi, “C.M.S. energy calibration in BES-III and VEPP-2000 experiments,” JINST 15 (2020) C07005.DOI
- [49]J. Kirby, “Physics prospects at the tau-charm factory,” Frascati Phys. Ser. 5 (1996) 747.
- [50]A. Bogomyagkov, V. Druzhinin, E. Levichev, A. Milstein, and S. Sinyatkin, “Low-energy electron-positron collider to search and study (μ+μ−) bound state,” EPJ Web Conf. 181 (2018) 01032, arXiv:1708.05819 [physics.acc-ph].DOI
- [51]V. I. Telnov, “Monochromatization of e+ e− colliders with a large crossing angle,” arXiv:2008.13668 [physics.acc-ph].
- [52]BESIII Collaboration, M. Ablikim et al., “Precision measurement of the mass of the τ lepton,” Phys. Rev. D 90 (2014) 012001, arXiv:1405.1076 [hep-ex].
- [53]H.-S. Shao, “Initial state radiation effects in inclusive J/ψ production at B factories,” JHEP 04 (2014) 182, arXiv:1402.5840 [hep-ph].DOI
- [54]E. A. Kuraev and V. S. Fadin, “Calculation of radiative corrections to the cross-section of one-photon annihilation by means of structure functions,” Jan., 1985. DESY L-Trans-297.
- [55]W. Beenakker et al., “WW cross-sections and distributions,” in CERN Workshops on LEP2 Physics (1995). 2, 1996. arXiv:hep-ph/9602351.DOI
- [56]S. Frixione, “Initial conditions for electron and photon structure and fragmentation functions,” JHEP 11 (2019) 158, arXiv:1909.03886 [hep-ph].DOI
- [57]V. Bertone, M. Cacciari, S. Frixione, and G. Stagnitto, “The partonic structure of the electron at the next-to-leading logarithmic accuracy in QED,” JHEP 03 (2020) 135, arXiv:1911.12040 [hep-ph].DOI
- [58]S. Dulat, T.-J. Hou, J. Gao, M. Guzzi, J. Huston, P. Nadolsky, J. Pumplin, C. Schmidt, D. Stump, and C. P. Yuan, “New parton distribution functions from a global analysis of quantum chromodynamics,” Phys. Rev. D 93 (2016) 033006, arXiv:1506.07443 [hep-ph].DOI
- [59]J. Alwall et al., “The automated computation of tree-level and next-to-leading order differential cross sections, and their matching to parton shower simulations,” JHEP 07 (2014) 079, arXiv:1405.0301 [hep-ph].DOI
- [60]CMS Collaboration, V. Khachatryan et al., “Search for long-lived particles that decay into final states containing two electrons or two muons in proton-proton collisions at √s = 8 TeV,” Phys. Rev. D 91 (2015) 052012, arXiv:1411.6977 [hep-ex].
- [61]ATLAS Collaboration, M. Aaboud et al., “Search for long-lived particles in final states with displaced dimuon vertices in pp collisions at √s = 13 TeV with the ATLAS detector,” Phys. Rev. D 99 (2019) 012001, arXiv:1808.03057 [hep-ex].DOI
- [62]CMS Collaboration, A. Tumasyan et al., “Search for long-lived particles decaying to a pair of muons in proton-proton collisions at √s = 13 TeV,” arXiv:2205.08582 [hep-ex].DOI
- [63]G. T. Bodwin, E. Braaten, and G. P. Lepage, “Rigorous QCD analysis of inclusive annihilation and production of heavy quarkonium,” Phys. Rev. D 51 (1995) 1125, arXiv:hep-ph/9407339. [Erratum: Phys.Rev.D 55, 5853 (1997)].arxiv.org/abs/hep-ph/9407339
- [64]E. J. Eichten and C. Quigg, “Quarkonium wave functions at the origin,” Phys. Rev. D 52 (1995) 1726, arXiv:hep-ph/9503356.DOI