L009 $\mathcal{B}(\tau^-\to\mu^-\mu^+\mu^-)$

Tau to three muons charged-lepton-flavor-violating four-lepton decay
Status REVIEWED VERIFIED Medium Code: NO Priority Medium

PDG / equivalent values

Observable Value Year Experiment / source Provenance
Branching fraction upper limit for $\tau^- \to \mu^- \mu^+ \mu^-$ 1.9e-8 2025 Particle Data Group 2025 tau listing source ↑
Branching fraction upper limit for $\tau^- \to \mu^- \mu^+ \mu^-$ 1.9e-8 2024 Belle II Collaboration source ↑
Branching fraction upper limit for $\tau^- \to \mu^- \mu^+ \mu^-$ ? 2026 LHCb Collaboration Run 2 preprint source ↑

Why this constrains the RS scan

This mode tests lepton-sector flavor violation that need not be dipole dominated. In warped or composite-flavor extensions, flavor-nonuniversal lepton profiles, KK gauge exchange, flavor-changing \(Z\) couplings, or scalar exchange can generate effective \(\tau\mu\mu\mu\) operators. The observable is therefore complementary to \(\tau\to\mu\gamma\) and \(\mu\to e\gamma\): dipoles can feed the three-muon final state through an off-shell photon, but the four-lepton final state is also directly sensitive to contact LFV.

What's changed since the original paper

Relative to the arXiv:0804.1954 RS-flavor baseline, the experimental frontier moved to \(10^{-8}\)-level limits (L009.yaml:pdg\_or\_equivalent.rs\_context; L009.yaml:pdg\_or\_equivalent.historical\_experimental\_limits). Belle and BaBar published full three-lepton tau searches in 2010, reaching \(2.1\times 10^{-8}\) and \(3.3\times 10^{-8}\), respectively (L009.yaml:pdg\_or\_equivalent.historical\_experimental\_limits[0] and L009.yaml:pdg\_or\_equivalent.historical\_experimental\_limits[1]). LHCb demonstrated a hadron-collider tau-to-three-muons program, setting \(4.6\times 10^{-8}\) with Run 1 data and \(1.9\times 10^{-8}\) with Run 2 (L009.yaml:pdg\_or\_equivalent.historical\_experimental\_limits[2]; L009.yaml:pdg\_or\_equivalent.post\_pdg\_update). Belle II now supplies the PDG benchmark, \(1.9\times 10^{-8}\) (L009.yaml:pdg\_or\_equivalent.primary\_current\_limit).

Validity and model dependence

Experimentally this is a clean upper limit on a forbidden charged-LFV decay. As a constraint on the current quark-scan code it is lepton-extension-only: the present quark $\Delta F = 2$ pipeline has no direct tau-lepton matching. As a future RS constraint, the translation is model dependent because a dipole-only assumption, a contact-operator assumption, and a full lepton-sector EFT can give different correlations with \(\tau\to\mu\gamma\), \(\mu\to e\gamma\), and \(\mu\to 3e\).

Code coverage in this repo

NO. Targeted greps over quarkConstraints/, qcd/, flavorConstraints/, neutrinos/, yukawa/, warpConfig/, solvers/, scanParams/, and tests/ found no implementation of \(\tau\to3\mu\), tau four-lepton LFV, or contact LFV. The adjacent LFV code is only \(\mu\to e\gamma\): flavorConstraints/muToEGamma.py:1, flavorConstraints/muToEGamma.py:75, and scanParams/scan.py:524. The modern phenomenology surface is neutral-meson mixing only, see quarkConstraints/modern/phenomenology.py:23.

Implementation difficulty

MEDIUM. A first implementation needs a new four-lepton LFV observable and a convention for lepton-flavor Wilson coefficients or RS matching. It does not require lattice inputs or hadronic long-distance treatment, but a production-grade implementation should decide whether to include dipole/contact interference and correlations with the existing \(\mu\to e\gamma\) lane.

Reason: A new four-lepton/contact LFV observable and lepton-sector Wilson-coefficient convention are needed. A first bound does not require lattice inputs or hadronic long-distance calculations, but correlations with dipoles and other LFV modes must be specified.

Key references

Process-local raw keys: PDG2025\_Tau3Mu, BelleII2024\_Tau3Mu, LHCb2026\_Tau3Mu, LHCb2015\_Tau3Mu, Belle2010\_Tau3Leptons, BaBar2010\_Tau3Leptons, and CFW2008\_RSWarpedFlavor.
Source SHAs
7 snapshot(s) tracked in flavor_catalog/processes/charged_lepton/L009.yaml
Access dates
2026-05-16
Worklog
flavor_catalog/worklogs/L009/
LaTeX source
flavor_catalog/processes/charged_lepton/L009.tex
Anchors generated
2026-05-18T01:53:04Z

Branching fraction upper limit for tau^- -> mu^- mu+ mu^-

https://pdg.lbl.gov/2025/listings/rpp2025-list-tau.pdf
accessed 2026-05-16 sha ad13c510e28a... snapshot L009/pdg_2025_tau_listing.txt RESOLVED
value 1.9e-08
RESOLVED
Match snapshot line 20
L17: Test of lepton family number conservation.
L18: 
L19: PDG best listed limit:
L20: BR(tau^- -> mu^- mu+ mu-) < 1.9 x 10^-8, CL = 90%.
L21: Document ID: ADACHI 24R, Belle II, 424 fb^-1, e+e- at sqrt(s) = 10.6 GeV.
L22: 
L23: Other listed measurements in the same PDG block:

Snapshots live under flavor_catalog/references/<process_id>/. Source-line anchoring is automated; the status pill reflects match confidence, not editorial review.

Branching fraction upper limit for tau^- -> mu^- mu+ mu^-

https://arxiv.org/abs/2405.07386
accessed 2026-05-16 sha 190f6aa5e4a0... snapshot L009/belleii_2024_arxiv_2405_07386.txt RESOLVED
value 1.9e-08
RESOLVED
Match snapshot line 27
L24: candidate is observed, compatible with the expected background.
L25: 
L26: Canonical result:
L27: BR(tau^- -> mu^- mu+ mu-) < 1.9 x 10^-8 at 90% confidence level.
L28: The arXiv abstract states this was the most stringent bound to date at
L29: the time of the paper.

Snapshots live under flavor_catalog/references/<process_id>/. Source-line anchoring is automated; the status pill reflects match confidence, not editorial review.

Branching fraction upper limit for tau^- -> mu^- mu+ mu^-

https://arxiv.org/abs/2601.20785
accessed 2026-05-16 sha 5dbcfd57d132... snapshot L009/lhcb_2026_arxiv_2601_20785.txt UNRESOLVED
value
UNRESOLVED

Automated anchoring could not pinpoint this value inside the snapshot. Manual verification is needed: open the live source and confirm against the context surrounding the value claim.

Open source manually
Snapshots live under flavor_catalog/references/<process_id>/. Source-line anchoring is automated; the status pill reflects match confidence, not editorial review.