Lepton Flavor Violating Higgs decays at the Compact Linear Collider
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: High Energy Physics: Phenomenology. The day in particle phenomenology: models, signatures and constraints.
Kai: We're looking at the paper: "Lepton Flavor Violating Higgs decays at the Compact Linear Collider".
Mira: Here's the gist of the paper. Lepton flavor violating Higgs decays could appear in scenarios beyond the Standard Model of particle physics The observation of any Higgs boson decays into two leptons with different flavors would be a clear indication of new physics Sensitivity to Lepton Flavor Violating Decays This study investigates the potential sensitivity of the Compact Linear Collider CLIC facility to LFV Higgs decays at both √s = 1.4 TeV and √s = 3 TeV energy stages The dominant production process for a SM Higgs boson is e+e- → hνν¯ The expected sensitivities are of order 10−5 for the h → eµ decay and of order 10−4–10−5 for inclusive τ channels Analysis of the h → eµ Decay Channel The final state h → eµ is described as the cleanest, as both final state particles are stable Event selection criteria for 1.4 TeV collisions included ∆θeµ −0.8 and Evis < 800 GeV The analysis uses a Boosted Decision Tree (BDT) classifier
49: to select kinematic variables for the h → eµ channel at 1.4 TeV The expected 95% CL limit for BR(h → eµ) at √s = 1.4 TeV is BR(h → eµ) < 3.86×10−5 without electron beam polarization Probing Inclusive Tau Leptonic Modes The study also explores two inclusive τ leptonic modes, h → τµ and h → eτ For the h → τµ decay channel at 1.4 TeV, the dominant requirement is on ∆θτµ < 2.43 rad The expected 95% CL limit for BR(h → τµ) at √s = 1.4 TeV is BR(h → τµ) < 1.50×10−4 Comparison with Existing Experimental Limits The results demonstrate that CLIC operation at 1.4 TeV has the potential to reach a sensitivity 12 times better than that of the CMS and ATLAS experiments
26, 27: This sensitivity increases to a factor of around 33 times better at 3 TeV operation The projected limits appear significantly more restrictive than current LHC constraints
26–28: These results demonstrate that CLIC operation at 1.4 TeV has the potential to reach a sensitivity 12 times greater than that of the CMS and ATLAS experiments (see Table??
Title and authors: Kai: So, we're looking at this paper titled "Lepton Flavor Violating Higgs decays at the Compact Linear Collider." It's about how sensitive a future collider, CLIC, can be to something that shouldn't happen in the Standard Model.
Mira: Exactly. The core idea is that if we see a Higgs boson decay into two leptons with different flavors, like an electron and a muon, it would mean there’s new physics happening somewhere beyond our current particle physics models.
Lev: And the paper focuses on testing those very limits using CLIC at two different energy levels, one at one point four teravolt and another at three teravolt.
Kai: Right, so what does the paper actually do to find this potential sign of new physics? It sets specific upper limits on branching fractions for decays like h to e mu, h to tau mu, and h to e tau.
Mira: The study looks at the dominant production process for a Higgs boson at an electron electron collider, which is e plus minus colliding with neutrinos producing a Higgs > Page one of that work reads: "The discovery of a scalar Higgs boson at the LHC one two has offered significant insights into the mechanism of electroweak symmetry breaking
three–five: , thereby opening new avenues for investigating BSM physics within the Higgs sector <ref:2605.08501#pg2,The discovery of a scalar Higgs boson at the LHC 1, 2>."
Lev: And those limits they calculate are quite tight, especially when you consider how much integrated luminosity CLIC can deliver at those different energies.
Kai: The paper gives a specific sensitivity range, saying that in the absence of finding anything, ninety-five percent confidence level upper limits in the region of ten to ten minus five could be placed on the branching fractions of these processes assuming integrated luminosities of four ab minus one at point four teravolt and five ab minus one at three teravolt <ref:2605.08501#pg1,could be placed on the branching fractions of these processes assuming integrated>.
Mira: That range is what makes it interesting. The Standard Model predicts these decays are extremely rare because they have to go through loop contributions involving neutrinos, which are very light, resulting in very small branching ratios > Page two of that work reads: "The Standard Model (SM) of particle physics forbids lepton flavor violation (LFV) at tree level <ref:2605.08501#pg2,The Standard Model (SM) of particle physics forbids lepton flavor violation (LFV>."
Lev: So the paper is basically saying that because the Standard Model suppresses these decays so much, if we don't see them, it means any new physics causing lepton flavor violation must be quite subtle.
Kai: And the paper describes how they set limits on h to e mu specifically at one point four teravolt, noting that without electron beam polarization, the upper limit for the branching fraction of h to e mu is less than three point eight six times ten minus five > Page one of that work reads: "In this article, we study the sensitivity of a future Compact Linear Collider (CLIC) to such processes, namely, h to e mu, h to tau mu, and h to e tau <ref:2605.08501#pg1,In this article, we study the sensitivity of a future Compact Linear>."
Title and authors: Mira: And they also look at those inclusive tau leptonic modes like h to tau mu and h to e tau. For the decay into tau mu at point four teravolt, the upper limit is less than one point five times ten minus four > Page one of that work reads: "In the absence of any observation, ninety-five percent CL upper limits in the region of ten to ten minus five could be placed on the branching fractions of these processes assuming integrated luminosities of four ab minus one at point four teravolt and five ab minus one at three teravolt <ref:2605.08501#pg1,could be placed on the branching fractions of these processes assuming integrated>."
Lev: The paper also discusses how they plan to improve this search using various techniques, which we'll talk about next.
Kai: Right, so the researchers are suggesting some ways to make this search even more powerful than just setting a simple upper limit on the branching fraction. They mention several improvements for their analysis pipeline.
Mira: They suggest using Boosted Decision Trees, or BDTs, to select kinematic variables like the mass of the leptons and the angle between them, or cos theta star and transverse energy > Page three of that work reads: "for tracking, enclosing an ultra-thin silicon-pixel vertex detector <ref:2605.08501#pg1>."
Lev: That makes sense for efficiency. They also talk about modeling how tau leptons are reconstructed using tools like TauFinder, and they incorporate things like the minimum momentum to enter reconstruction and isolation cone settings derived from their parameter scan > Page three of that work reads: "The resulting jet-energy resolution, for isolated central light-quarks jets with energy in the range from one hundred GeV to one TeV, is sigma E over E less than three point five percent <ref:2605.08501#pg1>."
Kai: That sounds like they are trying to maximize the signal while minimizing background noise by using those sophisticated cuts and reconstructions > Page three of that work reads: "A strong solenoidal magnet located outside the HCAL provides a four T magnetic field <ref:2605.08501#pg1>."
Mira: They also mention a specific background suppression strategy, saying they can apply cuts based on kinematic correlations, for instance, at three teravolt they might use an angle like less than zero point nine radians to suppress background processes > Page one of that work reads: "Figure one: Cross section as a function of center-of-mass energy for the main Higgs production processes at an e plus minus collider for a Higgs mass of mh = one hundred twenty-six GeV thirty-two <ref:2605.08501#pg3,Figure 1: Cross section as a function of center-of-mass energy>."
Lev: So, it's not just about the raw sensitivity numbers from CLIC, but about designing an intelligent analysis system that can actually sift through all that data effectively. That’s where the real engineering comes in.
Title and authors: Kai: Exactly. And when we look at these results, what they show is that CLIC operation at point four teravolt has the potential to reach a sensitivity twelve times better than what the CMS and ATLAS experiments are currently achieving > Page one of that work reads: "This work was carried out in the framework of the CLICdp Collaboration © two thousand twenty-six CERN for the benefit of the CLICdp Collaboration <ref:2605.08501#pg1,This work was carried out in the framework of the CLICdp Collaboration>."
Mira: And they say this sensitivity increases to about thirty three times better if you operate at three teravolt, which is quite a big jump > Page one of that work reads: "In the absence of any observation, ninety-five percent CL upper limits in the region of ten to ten minus five could be placed on the branching fractions of these processes assuming integrated luminosities of four ab minus one at point four teravolt and five ab minus one at three teravolt <ref:2605.08501#pg1,could be placed on the branching fractions of these processes assuming integrated>."
Lev: Those projected limits appear much tighter than what we expect from the current LHC constraints, which is a big thing for pushing our search limits > Page one of that work reads: "Reproduction of this article or parts of it is allowed as specified in the CC-BY-four point zero license <ref:2605.08501#pg1,Reproduction of this article or parts of it is allowed as specified>."
Kai: So, to wrap up on this paper, "Lepton Flavor Violating Higgs decays at the Compact Linear Collider," it shows that CLIC has a pathway to probe these rare decays with significant precision > Page one of that work reads: "In this article, we study the sensitivity of a future Compact Linear Collider (CLIC) to such processes, namely, h to e mu, h to tau mu, and h to e tau <ref:2605.08501#pg1,In this article, we study the sensitivity of a future Compact Linear>."
Mira: It really highlights that observing any Higgs decay into two different flavored leptons would be a very clear sign that the Standard Model needs an extension > Page two of that work reads: "Consequently, the observation of any Higgs boson decays into two leptons with different flavors would be a clear indication of new physics <ref:2605.08501#pg2,Consequently, the observation of any Higgs boson decays into two leptons with>."
Lev: For me, it shows what's actually feasible for real hardware. It tells us how much overhead we need to build into the error correction systems to actually pull these limits when you run at those high energies > Page three of that work reads: "The vertex and tracking systems are surrounded by a highly-granular sampling calorimetry system, composed of an electromagnetic and a hadronic calorimeter (ECAL and HCAL), optimized for particle flow reconstruction <ref:2605.08501#pg1>."
Kai: We're done with this one. Next up, we'll be looking at how other experiments are trying to probe similar BSM physics in different ways.
The paper's summary: Kai: So, basically, this paper is about how sensitive we can make our Compact Linear Collider to these rare decays where a Higgs turns into two leptons with different flavors, like an electron and a muon.
Mira: Right. The core idea is that if we don't see those decays at the machine, it tells us something specific about physics beyond the Standard Model that's messing with how we see lepton flavors.
Lev: From an error correction standpoint, it means they’re calculating how much data you need to collect at different energies—one point four teravolt and another three teravolt—to actually pull these limits off.
Kai: They set specific upper limits on the branching fractions for decays like h to e mu, h to tau mu, and h to e tau.
Mira: And they show how tight those limits get when you factor in the integrated luminosity of that machine at those different energies. It’s not just a simple number anymore; it’s tied directly to how much time you spend running the experiment.
Lev: The authors mention using advanced AI systems to do this precision limit setting, which means they can calculate those ninety-five percent confidence level upper limits based on the actual luminosity you plan to use.
Kai: And they don't just say "if we don't see it, then new physics must be small"; they give us the numbers for what that small is. For instance, at one point four teravolt, the limit on h to e mu is less than three point eight six times ten minus five without using any electron beam polarization.
Mira: That's what makes it compelling; it’s a concrete constraint on new physics that we can test with the machine they’re designing.
Lev: If you were trying to run this on actual hardware, you'd need to worry about those background processes and how the AI system handles all that complex kinematic information, like the Boosted Decision Trees they use.
Kai: They also talk about modeling tau lepton reconstruction very carefully using tools like TauFinder and figuring out exactly what kind of cuts you need to apply based on your energy settings.
Mira: It really hammers home that this isn't just theory; it’s a roadmap for how to build the experiment and analyze the data to actually find these tiny deviations.
Lev: So, they show us that CLIC at one point four teravolt could be twelve times better than what current experiments like CMS and ATLAS are doing right now.
Kai: And if you push it up to three teravolt, that sensitivity jumps by a factor of around thirty three times better. That’s a big improvement in the potential search power.
Mira: It suggests that these projected limits appear much tighter than what the current Large Hadron Collider constraints are putting out on similar things.
Lev: That's what matters because it means we could be looking for new physics that is currently just below our detection threshold, but this machine would push us right into the observable region.
Kai: So, observing any Higgs decay into two different flavored leptons would be a clear sign that the Standard Model needs an extension.
Mira: Exactly. It’s a direct path to finding evidence for new physics in the Higgs sector.
Lev: What this means for you on the ground is that we need to build error correction systems and analysis pipelines capable of handling these kinds of extremely precise measurements when they run at those high energies.
The paper's improvements: Kai: So, we’re looking at how they plan to make this search even better than just setting a simple upper limit on the branching fraction.
Mira: They suggest using Boosted Decision Trees, or BDTs, to select kinematic variables like the mass of the leptons and the angle between them, or cos theta star and transverse energy.
Lev: That makes sense for efficiency because it lets you build a smarter filter that maximizes signal while trying to minimize background contamination from other processes.
Kai: They also talk about modeling how tau leptons are reconstructed using tools like TauFinder, and they incorporate specific steering parameters, like the minimum momentum to enter reconstruction and isolation cone settings derived from their parameter scan.
Mira: That level of detail in the tau reconstruction modeling is what you need to accurately predict the efficiency and fake rates for those decays into tau mu, tau e, or electron tau.
Lev: And on the background suppression side, they propose channel-specific cuts based on kinematic correlations; for example, at three teravolt they might use an angle less than zero point nine radians to suppress background processes.
Kai: So what this means for us is that the analysis pipeline isn't just brute force; it needs to be highly optimized with these kinds of tailored cuts and intelligent variable selection.
Mira: It really shows that the search power comes from designing a system that understands the physics, not just collecting more raw data.
Lev: When I think about running this on real hardware, all those complex AI systems need to run fast enough on the processing units so they don't become a bottleneck in the overall error correction scheme.
Kai: The paper also touches on how to improve the jet-energy resolution, saying that for isolated central light quarks jets with energy in that range from one hundred gigaelectronvolts to one terahigh-energy electronvolt, it falls with the square of the sampling interval.
Mira: That’s a concrete number about how well you can measure particle energy in this environment, which is crucial for those precise kinematic cuts they mentioned earlier.
Lev: It points toward needing very granular sampling calorimetry systems to get that kind of resolution when you're dealing with high energy jets.
Kai: So, the implication here is that the success of this search depends heavily on how well we engineer the detector and how smart our AI system can be at filtering out noise.
Mira: It’s a very practical look at turning a theoretical possibility into a measurable experimental constraint.
Conclusion: Kai: So, to wrap up, this paper on "Lepton Flavor Violating Higgs decays at the Compact Linear Collider" is about setting very specific upper limits on how rare these lepton flavor violating Higgs decays could be.
Mira: Basically, they show us that if we don't see these decays at a future collider, it constrains new physics in the Standard Model very tightly.
Lev: It’s interesting because they are using the CLICdp Collaboration framework to project what we can actually measure at those different energy stages of one point four teravolt and three teravolt.
Kai: The implication is that any observation of a Higgs decaying into two different flavored leptons would be a very clear sign that the Standard Model needs an extension.
Mira: Right. It’s a direct path to finding evidence for new physics in the Higgs sector, which we haven't seen yet.
Lev: From my side, it means we need to build error correction systems and analysis pipelines capable of handling these kinds of extremely precise measurements when they run at those high energies.
Kai: It really shows how much overhead we have to put into the hardware and the software to actually pull these limits off in a real experiment.
Mira: And it highlights that the search power comes from designing a system that understands the physics, not just collecting more raw data points for a simple count.
Lev: The next thing we look at is how other experiments are trying to probe similar BSM physics in different ways, because this is one piece of the puzzle.
Francisca Garayb, Gabriel Vegab, Philipp Roloffa
CERN · Pontificia Universidad Catolica de Chile · Millenium Institute for Subatomic Physics at the High Energy Frontier - SAPHIR
hep-ph, hep-ex
Submitted: 2026-05-08
Updated: 2026-10-29
License: http://creativecommons.org/licenses/by/4.0/
Importance score: 83/100
The gist: Lepton flavor violating Higgs decays could appear in scenarios beyond the Standard Model of particle physics The observation of any Higgs boson decays into two leptons with different flavors would be…
Key concepts
- Lepton Flavor Violating (LFV) Decays
- These are hypothetical Higgs decays where the Higgs boson turns into two leptons of different flavors, such as an electron and a muon (h -> eµ). Observing this decay would be a clear signal that new physics beyond the Standard Model is at play, as it is not predicted by the Standard Model.
- Compact Linear Collider (CLIC)
- CLIC is a proposed high-energy particle collider designed to operate at energies of 1.4 TeV and 3 TeV. The study investigates its potential to precisely measure Higgs decays into leptons, using specific energy stages and production processes like e+e- -> hνν¯ as the baseline for sensitivity calculations.
- Boosted Decision Tree (BDT)
- A BDT is a machine learning tool used in particle physics analysis to select the most promising events from a large dataset. It uses multiple kinematic variables, such as angles and energy, to distinguish rare signal events (like h -> eµ) from unwanted background noise with high efficiency.
- Branching Ratio (BR)
- The branching ratio is the probability that a specific particle decay will occur. The study sets upper limits on the BR of Higgs decays, such as h -> eµ, by comparing expected signal rates at CLIC to the experimental sensitivity achieved at those energy levels.
Terminology
Summary
Lepton flavor violating Higgs decays could appear in scenarios beyond the Standard Model of particle physics The observation of any Higgs boson decays into two leptons with different flavors would be a clear indication of new physics
Sensitivity to Lepton Flavor Violating Decays
This study investigates the potential sensitivity of the Compact Linear Collider CLIC facility to LFV Higgs decays at both √s = 1.4 TeV and √s = 3 TeV energy stages The dominant production process for a SM Higgs boson is e+e- → hνν¯ The expected sensitivities are of order 10−5 for the h → eµ decay and of order 10−4–10−5 for inclusive τ channels
Analysis of the h → eµ Decay Channel
The final state h → eµ is described as the cleanest, as both final state particles are stable Event selection criteria for 1.4 TeV collisions included ∆θeµ −0. p. 4 Additionally, an electron energy threshold of Ee > 8 GeV was applied to avoid poorly reconstructed electrons, as those below this energy fall outside the detector’s acceptance. 4 TeV collisions,
the cuts |∆θeµ
| < 2. The final event selection employs a Boosted Decision Tree (BDT) classifier [49] using a subset of the
kinematic variables defined previously. 4 TeV, those used are: meµ
, |∆θeµ
|, |∆φeµ
|, βeµ
, cosθ
∗
and
Evis. 994 at 3 TeV. The BDT score distribution is compared with the background-only BDT score distribution to derive
the expected limit on BR(h → eµ). All results are based on
√
s = 13 TeV pp collisions corresponding to an integrated luminosity of about 139 fb−1
. The most stringent indirect constraint on the h → eµ decay arises from searches for µ → eγ decays [29], yielding a bound of BR(h → eµ) ≤ O(10−8
) [30, 31].Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Probing Inclusive Tau Leptonic Modes
The study also explores two inclusive τ leptonic modes, h → τµ and h → eτ For the h → τµ decay channel at 1.4 TeV, the dominant requirement is on ∆θτµ < 2. p. 9 The same criteria as before were used to maintain high signal efficiency. 4 TeV, the dominant requirement is on ∆θτµ , which removes about 12% of the background
while leaving the signal essentially unaffected. 994 at 3 TeV. The BDT score distribution is compared with the background-only BDT score distribution to derive
the expected limit on BR(h → eµ). 4 TeV, the dominant requirement is on ∆θτµ , which removes about 12% of the background
while leaving the signal essentially unaffected. This exploits the fact that background processes tend to
produce leptons in nearly back-to-back topologies, whereas signal events from h → τµ do not.Where the paper says it
Where the paper says it
Where the paper says it
Comparison with Existing Experimental Limits
The results demonstrate that CLIC operation at 1.4 TeV has the potential to reach a sensitivity 12 times better than that of the CMS and ATLAS experiments [26, 27] This sensitivity increases to a factor of around 33 times better at 3 TeV operation The projected limits appear significantly more restrictive than current LHC constraints [26–28]
Conclusion on CLIC Potential
These results demonstrate that CLIC operation at 1.4 TeV has the potential to reach a sensitivity 12 times greater than that of the CMS and ATLAS experiments (see Table? p. 1 While recent LHCb measurements [24,
25] show consistency with SM expectations, the theoretical interest in LFV signatures remains strong,
particularly given the enhanced sensitivity of future lepton colliders such as CLIC. The ATLAS and CMS experiments at the Large Hadron Collider (LHC) have searched for LFV Higgs
decays.Where the paper says it
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Precision limit setting for new physics searches: The improved AI system can calculate 95% confidence level upper limits on lepton flavor violating Higgs branching fractions, such as h → eµ, based on CLIC integrated luminosities and specific center-of-mass energies (1.4 TeV and 3 TeV).
-
Event selection optimization: The system can apply complex multivariate analysis like Boosted Decision Trees (BDTs) using kinematic variables such as mll', ∆θll', ∆φll', cosθ∗, Evis, Erest, and θmiss to maximize signal efficiency while minimizing background contamination.
-
Tau lepton reconstruction modeling: The AI can simulate tau lepton reconstruction using tools like TauFinder and incorporate specific steering parameters (e.g., minimum pT to enter reconstruction) and isolation cone settings derived from the study's parameter scan to accurately predict reconstruction efficiencies and fake rates for h → τe, h → τµ, and h → eτ channels.
-
Background suppression strategy: The system can implement channel-specific cuts based on kinematic correlations; for example, applying ∆θτµ-0.9 at 3 TeV to suppress background processes with high effectiveness against signal contamination.
Abstract
Lepton flavor violating Higgs decays could appear in scenarios beyond the Standard Model of particle physics. In this article, we study the sensitivity of a future Compact Linear Collider (CLIC) to such processes, namely, h to eμ, h toτμ, and h to eτ. In the absence of any observation, 95% CL upper limits in the region of 10-4 to 10-5 could be placed on the branching fractions of these processes, assuming integrated luminosities of 4,ab-1 at sqrt s =1.4 TeV and 5,ab-1 at sqrt s =3 TeV.
Sources
- Updated CLIC luminosity staging baseline and Higgs coupling prospects
- The International Linear Collider Technical Design Report - Volume 4: Detectors
- Physics and Detectors at CLIC: CLIC Conceptual Design Report
- A detector for CLIC: main parameters and performance
- TMVA - Toolkit for Multivariate Data Analysis
- CLIC Higgs coupling prospects with 100 Hz operation
Related papers
- MadNIS at NLO
- The NANOGrav 15-year data set: A Limit on Continuous Spin Gravitons
- DIS-secting the universe with high-energy neutrinos
- Crowdsourcing Gravitational Waves from Superradiant Axions
- Classification of g-modes for neutron stars with a strong transition: Novel universal relation including slow stable hybrid stars
- Scalar axion field of toroidal electromagnetic pulses