Source-linked AI summary
HiggsTools: BSM scalar phenomenology with new versions of HiggsBounds and HiggsSignals
Henning Bahl, Thomas Biekötter, Sven Heinemeyer, Cheng Li, Steven Paasch, Georg Weiglein, Jonas Wittbrodt
TL;DR
Existing tools were needed to test extended scalar-sector models against numerous collider searches and Higgs measurements. The paper presents HiggsTools, a modern C++ unification and extension of HiggsBounds and HiggsSignals with shared prediction inputs and expanded capabilities. The resulting framework adds broader search coverage, improved handling of scalar clusters and mass uncertainties, and unified measurement treatment.
Problem
Testing BSM scalar-sector models against the large set of available searches and Higgs measurements requires dedicated tools.
Method
The paper rewrites HiggsBounds and HiggsSignals in modern C++, unifies them in HiggsTools, and adds HiggsPredictions with interactive language interfaces.
Results
HiggsBounds gains improved mass-uncertainty and particle-clustering treatment plus expanded di-Higgs and doubly-charged-Higgs searches, while HiggsSignals unifies measurement handling.
Takeaways & Limitations
HiggsTools provides a common suite for defining BSM scalar predictions and confronting them with direct searches and 125 GeV Higgs data.
Takeaways & Limitations
HiggsTools does not internally distinguish resonant from non-resonant pair production, leaving that classification and separate inputs to the user.
Abstract
from arXiv · showhide
The codes HiggsBounds and HiggsSignals compare model predictions of BSM models with extended scalar sectors to searches for additional scalars and to measurements of the detected Higgs boson at 125 GeV. We present a unification and extension of the functionalities provided by both codes into the new HiggsTools framework. The codes have been re-written in modern C++ with native Python and Mathematica interfaces for easy interactive use. We discuss the user interface for providing model predictions, now part of the new sub-library HiggsPredictions, which also provides access to many cross sections and branching ratios for reference models such as the SM. HiggsBounds now implements experimental limits purely through json data files, can better handle clusters of BSM particles of similar masses (even for complicated search topologies), and features an improved handling of mass uncertainties. Moreover, it now contains an extended list of Higgs-boson pair production searches and doubly-charged Higgs boson searches. In HiggsSignals, the treatment of different types of measurements has been unified, both in the $χ^2$ computation and in the data file format used to implement experimental results.
1 Introduction
The paper motivates dedicated tools for testing extended scalar-sector models against the growing body of collider searches and Higgs measurements. It introduces HiggsTools as a modernized, unified framework extending HiggsBounds and HiggsSignals.
- Extended scalar-sector models predict additional scalars or modified 125 GeV Higgs couplings, which collider data constrain.
- Testing each BSM parameter point against all available collider searches and measurements requires dedicated computer tools because the dataset is large.
- HiggsBounds tests exclusion limits from new-scalar searches, while HiggsSignals tests compatibility with 125 GeV Higgs rate measurements.
- HiggsTools rewrites and unifies these codes in modern C++, adds HiggsPredictions, and provides C++, Python, and Mathematica interfaces.
- The framework adds support for non-resonant di-Higgs searches, doubly-charged Higgs bosons, and CP-sensitive coupling measurements.
2 The HiggsTools framework
HiggsTools unifies HiggsBounds and HiggsSignals and adds HiggsPredictions as a shared sub-library for model inputs and reference predictions.
- HiggsTools combines the functionality of HiggsBounds and HiggsSignals in one framework.
- The framework is organized around three subpackages and focuses its documentation on program flow and new features relative to older versions.
- HiggsPredictions defines physical models and supplies theory predictions for scalar production and decay rates.
- HiggsBounds evaluates direct-search bounds on scalar particles, while HiggsSignals evaluates compatibility with measurements of the approximately 125 GeV Higgs boson.
2.1 HiggsPredictions
HiggsPredictions provides a process-based interface for defining scalar models, deriving relevant production and decay predictions, and supplying tabulated reference rates. It supports several collider process topologies while leaving resonant-versus-non-resonant classification to the user.
- Model definition: Users define each BSM scalar through its mass, total width, charge, CP character, and relevant production and decay rates.
- Model definition: Effective couplings, SLHA files, and HiggsBounds data files provide alternative ways to generate or supply model predictions.
- Process definitions: The framework represents channel, chain-decay, pair-decay, and pair-production topologies for collider searches involving BSM scalars.
- Process definitions: HiggsTools does not distinguish resonant from non-resonant pair production internally, so users must provide separate inputs and define the classification criterion.
- Process definitions: HiggsPredictions automatically derives predictions for implemented processes from the supplied production cross sections and branching ratios.
- Process definitions: The branching-ratio machinery accounts for symmetry factors and currently restricts pair processes to overall neutral final states, including opposite charged BSM scalars.
- Reference predictions: HiggsPredictions includes tabulated reference cross sections and branching ratios for SM-like and selected non-SM-like scalar coupling structures.
- Reference predictions: Its non-SM-like cross-section predictions encode dependence on relevant couplings but omit effects from other BSM particles in production processes.
2.2 HiggsBounds
HiggsBounds evaluates model process rates against experimental limits using limit-specific assignments, particle clustering, and observed-to-limit ratios. HiggsBounds-6 broadens the implemented searches and improves limit-file integration, clustering, and mass-uncertainty treatment.
- Limit evaluation: HiggsBounds assigns model scalars and process rates to experimental limits, computes observed and expected ratios, and selects the most sensitive limit using the highest expected ratio.A parameter point is allowed when the observed ratio is below one for each particle’s most sensitive limit.
- Limit database: 258 experimental limits from LEP and the LHC are currently included in the HiggsBounds database.
- Limit types: HiggsBounds-6 encodes each experimental limit in a JSON file and distinguishes six limit types, simplifying the incorporation of new experimental data.The types cover channel, width-dependent, longer-chain, pair-decay, di-Higgs, and likelihood limits.
- Particle clustering: The clustering algorithm groups scalars with compatible masses and evaluates limits at rate-weighted masses and widths, including processes involving multiple BSM scalar types.A cluster is valid when its mass spread satisfies max(m_i) − min(m_i) ≤ r_abs + r_rel · mean(m_i).
- Mass uncertainties: HiggsBounds-6 varies each uncertain mass across its user-given range, selects the mass with the lowest observed ratio, and then evaluates the expected ratio there.For m_h = 130 GeV, the observed ratio decreases from ∼1.55 to ∼0.75 as the uncertainty increases from zero to 7 GeV, whereas for m_h = 125 GeV it stays essentially constant.
- New search coverage: The new version extends resonant and non-resonant Higgs pair-production searches and adds searches for doubly-charged Higgs bosons.The doubly-charged searches include leptonic final states and a recent bosonic-final-state search.
2.3 HiggsSignals
HiggsSignals unifies the treatment of peak-centered, mass-centered, and STXS measurements while computing correlated χ2 values from model predictions and measurement data. HiggsSignals-3 also incorporates measurements whose dependence extends beyond simple rates, exemplified by a CMS CP analysis.
- Unified measurement treatment: The modern C++ reimplementation unifies peak-centered observables, mass-centered observables, and STXS measurements.
- Measurement data and χ2: HiggsSignals replaces individual ATLAS and CMS Higgs mass measurements with one PDG-combination file and currently implements 129 individual measurements.Its χ2 computation accounts for correlations among the measurements.
- Parameter-dependent measurements: HiggsSignals-3 implements measurements that depend on model parameters beyond simple rates, including the CMS H → τ+τ− CP analysis.The analysis depends on gluon-fusion and vector-boson-fusion signal strengths and the CP-violating phase ϕτ.
- Parameter-dependent measurements: The new version directly incorporates coupling-structure dependencies into χ2 calculations that previously required external evaluation.For the CMS analysis, ϕτ is defined through the CP-even and CP-odd tau-Yukawa coefficients cτ and ˜cτ.
3 C++, Python, and Mathematica interfaces
HiggsTools provides C++, Python, and Mathematica interfaces for defining scalar predictions and running HiggsBounds and HiggsSignals analyses. Users initialize prediction and analysis objects, load datasets, and retrieve exclusion or χ2 results through parallel workflows.
- Interfaces: HiggsTools offers C++, Python, and Mathematica interfaces, with similar C++/Python syntax but a distinct Mathematica structure.The C++ libraries use Higgs/Bounds.hpp, Higgs/Predictions.hpp, and Higgs/Signals.hpp, while Python imports corresponding Higgs modules.
- Providing predictions: Users initialize HiggsPredictions and define scalar particles, production cross sections, decay widths, branching ratios, or effective couplings.Reference-model predictions can be used instead of supplying every cross section and decay quantity explicitly.
- Running HiggsBounds: HiggsBounds loads limit files, applies them to model predictions, and returns whether the selected parameter point is allowed or excluded.For each particle, the limit with the highest expected sensitivity is selected; an observed ratio below one is required for allowance.
- Running HiggsBounds: HiggsBounds can also report selected-limit details, including the excluded particle, observed and expected ratios, and the experimental search description.The example output identifies an h exclusion with an observed ratio of 1.676 and an expected ratio of 0.774.
- Mathematica: Mathematica users initialize the package and datasets, add particles, set their properties, and inspect applied limits or measurement information through dedicated functions.Installing the Mathematica executable automatically initializes HiggsPredictions, HiggsBounds, and HiggsSignals objects.
4 Examples
The examples demonstrate HiggsTools applications to charm-Yukawa constraints, resonant Higgs-pair searches, Higgs-width modifications, and 2HDM parameter constraints. They also show how reference-model choices and unified HiggsSignals/HiggsBounds analyses affect the resulting interpretations.
- Charm Yukawa coupling: HiggsSignals scans the CP-even and CP-odd charm-Yukawa couplings, with the lowest ∆χ2 region including the SM prediction.Using SMHiggs instead produces a vanishing-coupling best fit and disfavors the SM at about 1σ.
- Charm Yukawa coupling: The choice of reference model changes the gluon-fusion prediction from σ(ggH) = 41.93 pb to σ(ggH) = 48.52 pb and can alter the best-fit interpretation.SMHiggs uses QCD NNLO cross sections, whereas SMHiggsEW uses N3LO heavy-top predictions; the example emphasizes that the correct reference model is crucial for reliable bounds.
- Resonant h125-pair production: HiggsBounds-6 implements an extended set of resonant h125-pair searches and can yield substantially stronger bounds when resonant h125-pair production is relevant.The illustrated limits cover σ(pp →H →h125h125) across searches and combine results from different final states.
- Higgs-boson width: For the Higgs-width scenario, the low-∆χ2 band includes the SM point at ceff = 1 and BR(h125 →NP) = 0, while the best fit is ceff = 1.21 and BR(h125 →NP) = 0.32.The SM point has ∆χ2 = 0.06 relative to the best fit, and the plot marks total widths of 1, 10, and 100 times the SM prediction.
- 2HDM constraints: The 2HDM effective-coupling analysis assumes heavy-state effects on loop-induced couplings are negligible when mH = mA = mH± ≫mh.Effective couplings as functions of α and β provide the model information used to obtain the relevant predictions.
5 Conclusions
HiggsTools-1 unifies HiggsPredictions, HiggsBounds, and HiggsSignals through a common prediction interface while extending their functionality. The updated suite supports broader search and measurement implementations and demonstrates these capabilities in physics examples.
- Conclusions: HiggsTools-1 combines HiggsPredictions, HiggsBounds, and HiggsSignals into one suite with a common interface for Higgs production and decay predictions.HiggsBounds tests exclusion limits, while HiggsSignals confronts models with the measured Higgs mass and rates near 125 GeV.
- Conclusions: HiggsPredictions facilitates defining physical models and providing the scalar production and decay predictions required by the other programs.
- Conclusions: HiggsBounds classifies experimental limits into six types and improves particle clustering for cases where several scalars contribute to one limit.The update also adds di-Higgs search channels and doubly charged Higgs-boson limits.
- Conclusions: HiggsSignals now supports measurements depending on model parameters beyond simple rates, including CMS’s dedicated CP analysis for Higgs decay to τ +τ −.
- Conclusions: Applying the correct Higgs-boson cross-section prediction was demonstrated to be crucial for obtaining reliable bounds on the charm Yukawa coupling.