LHC Dark Showers Workshop

27 Oct 2025, 14:00 → 30 Oct 2025, 17:00 Europe/Zurich

73/1-031 - 73-1-031 (CERN)

Annapaola De Cosa (ETH Zurich (CH)), Deepak Kar (University of the Witwatersrand (ZA)), Felix Kahlhoefer (Karlsruhe Institute of Technology), Jackson Carl Burzynski (Simon Fraser University (CA)), Jannicke Pearkes (University of Colorado Boulder (US)), Kevin Pedro (Fermi National Accelerator Lab. (US)), Suchita Kulkarni (University of Graz), Sukanya Sinha (The University of Manchester (GB))

Following up on the initial discussions during the Virtual Dark-showers workshop earlier this year, we are organising the first hybrid LHC Darkshowers workshop (at CERN & Virtual) on 27-30th October, 2025. The effort is now part of the LHC BSM working group, as a cross-subgroup task force.

This 3/4-day meeting will promote further discussion about our 2 objectives:

1. Establishing common benchmarks for LHC Run 3 experimental results 
2. Further developing dark shower models and simulations.

The workshop will include: followup presentations for ongoing efforts from previous workshops; open discussions; a lightning round for new ideas; and potentially a post-workshop hackathon to jumpstart technical efforts. If you would like to have input into the agenda topics, please fill out the registration form by September 8, 2025. Later registrations will still be accepted, but the agenda will be finalized in mid-September.

Following the workshop, contributors and steering committee members will prepare a white-paper based on the studies and from there plan other short workshops. We can use as an example the discussions we had at the Semi-Visible jet workshop in Zurich, the January dark-showers workshop and follow the recommendations highlighted in the report from the MITP dark sector workshop.

If you are interested in contributing to the benchmarking effort, please join the LHC Darkshowers taskforce channel within LHC BSM WG mattermost server. The taskforce mattermost channel will be used for all benchmarking related communications before and after the October workshop. To join the taskforce channel, you first need to join the main LHC BSM WG server and then join the relevant taskforce channel here: https://mattermost.web.cern.ch/lhc-bsm-wg/channels/lhc-dark-showers-taskforce

Monday 27 October

14:00 → 15:55 Overviews & new ideas

Conveners: Annapaola De Cosa (ETH Zurich (CH)), Jackson Carl Burzynski (Simon Fraser University (CA))

14:00 Workshop introduction

Speaker: Sukanya Sinha (The University of Manchester (GB))

LHC Darkshowers Oct 2025 intro.pdf

14:15 CMS overview

Speaker: Cesare Tiziano Cazzaniga (ETH Zurich (CH))

presentation_DarkShowers2025.pdf

14:40 ATLAS overview

Speaker: Ian Alejandro Ramirez-Berend (Carleton University (CA))

LHC_DarkShowersWorkshop_IRB_27Oct25.pdf

15:05 Holography for darkshowers

Speaker: Nick Evans

DarkShowers_NE.pdf

15:30 Glueballs and Lattice for darkshowers

Speaker: Graham Douglas Kribs (University of Oregon (US))

kribs_dark_showers_2025_10.pdf

15:55 → 16:30 Break

16:30 → 16:55 Progress reports

Status updates from working groups established at January workshop

Conveners: Jannicke Pearkes (University of Colorado Boulder (US)), Suchita Kulkarni (University of Graz)

16:30 Semi-visible Emerging Jet

The dark shower collider phenomenology splits the signatures into the “semi-visible” and “emerging-jets”, corresponding to seemingly orthogonal simplified models (s-channel vs t-channel). However, those are not exclusive signatures, and hence I will discuss how to hunt for dark showers that give visible and invisible displaced signatures. These Semi-Visible Emerging Jets (SVEJ) have striking signatures at the LHC, and I will show a proof-of-concept on the LHC sensitivity to these signatures, and then non-trivial interplay with the CalRatio ATLAS search.

Speaker: Dr José Francisco Zurita (IFIC, CSIC‐UV)

Zurita_SVEJ_CERN.pdf

16:55 → 19:10 Lightning talks

Conveners: Jannicke Pearkes (University of Colorado Boulder (US)), Suchita Kulkarni (University of Graz)

16:55 A Novel Beam Dump Experiment for Dark Sector Searches Using the CMS Detector

We propose a novel scheme for performing a beam-dump-like experiment with the general-purpose detectors (ATLAS and CMS) at the LHC. Collisions of high-energy protons result in jets containing a number of energetic hadrons and electromagnetic objects that are essentially “dumped” to hadronic and electromagnetic calorimeters, respectively, and induce the production of secondary hadrons, electrons, and photons in calorimetric showers. We envision a situation where new physics particles are produced by the interactions of these secondary particles inside the calorimeters. For proof of principles, we consider the axion-like particles (ALPs) produced via the Primakoff process in the presence of their interaction with photons at CMS. We argue that the drift tube chambers and the ME0 module of the muon system can serve as detectors to record the photons from the ALP decay, demonstrating that assuming the background level can be controlled as discussed in this work, the resulting sensitivity reach is competitive due to their close proximity to the signal source points. We further show that the LHC does not suffer from a barrier, dubbed beam-dump “ceiling”, that typical beam-dump experiments hardly surpass. This gives the LHC great potential to explore a wide range of parameter space. This analysis can be extended to investigate various types of light mediators with couplings to the Standard Model leptons and quarks.

Speaker: Hyunyong Kim (Seoul National University (KR))

LHC_Dark_Showers_Workshop_v1.pdf

17:10 Reinterpretation of the early Run-3 displaced-jets search at CMS for emerging-jets signatures

In this talk, I will present the reinterpretation of the early Run-3 displaced-jets search at CMS [Rep. Prog. Phys. 88 (2025) 037801] for emerging-jets signatures. The search analyzes a data sample of proton-proton collisions at a center-of-mass energy of 13.6 $\mathrm{TeV}$, corresponding to an integrated luminosity of 34.7 $\mathrm{fb}^{-1}$, collected with the CMS detector at the CERN LHC in 2022. Novel techniques in trigger, reconstruction, and machine learning have been developed for and employed in this search, leading to significant improvements in sensitivities to displaced-jets signatures arising from long-lived particle decays.

We further interpret this search for emerging-jets signatures. Models with a dark sector having a new confining gauge group are considered. New dark quarks are charged under the gauge group, and can undergo parton shower and hadronization within the dark sector, producing long-lived dark hadrons that can subsequently decay into standard model particles. We examine the cases where the dark quarks are pair produced from the decays of a new heavy vector boson $\mathrm{Z}'$ or the Higgs boson. For the model where a $\mathrm{Z}'$ boson decays to a pair of dark quarks, production cross sections larger than 6.5 $\mathrm{fb}$ can be excluded for $\mathrm{Z}'$ boson masses larger than 600 $\mathrm{GeV}$. For the models where the Higgs boson decay to a pair of dark quarks, the lowest limits on the branching fraction of the Higgs boson to dark quarks reach below $1\%$ for dark hadron proper decay lengths smaller than $1~\mathrm{m}$. These are the most stringent limits to date on these models.

Speaker: Jingyu Luo (Brown University (US))

LHC_DS_DisplacedJets.pdf

17:25 How invisible is it?

Semi-visible jet signatures have rich phenomenology for experiments to explore. While the details of the dark shower parameters are being studied carefully, the community has identified a critical parameter, $r_\mathrm{inv}$, to capture the main characteristics of such signatures. It is the fraction of invisible dark hadron decays. $r_\mathrm{inv}$ is correlated with key observables in the analyses, such as the transverse missing energy. It is of great interest to reconstruct this quantity at the analysis level, to help with designing more powerful search strategies and directly probe the nature of the excess in case of a discovery. In a previous work[1], we investigated the possibility of reconstructing $r_\mathrm{inv}$ in the boosted regime using a simple decomposition method, which gave us promising results. In this ongoing work, we attempt to apply a ML-based technique to regress $r_\mathrm{inv}$, achieving a much better resolution.

Speaker: Yin Li (Sun Yat-Sen University (CN))

How invisible is it.pdf

17:40 Electroweak Portal Dark Shower

We present the phenomenology of the electroweak portal dark shower. The talk will include a brief introduction to the UV theory. We will also present model-independent experimental bounds from current LHC data and future projections.

Speaker: LINGFENG LI (Brown U.)

EW_Dark_Shower_DSWS.pdf

17:55 Probing the Higgs Portal to a Hidden Valley at FCCee

Within the Hidden Valley (HV) scenario, for a QCD-like dark sector, novel experimental signatures emerge, characterized by sprays of particles resembling hadronic jets containing stable dark bound states. The resulting signature is characterised by missing momentum aligned with one of the jets, and is known as semi-visible jets (SVJs). We investigate the potential role of the Higgs boson as a portal to HV, specifically leading to SVJ signatures at the FCC-ee. We present new simplified models allowing the decays of the unstable dark bound states to photons, leptons and quarks. We investigate also scenarios with different fractions of invisible states and introduce efficient analysis strategies to probe them. The presented model and signature can be extended also in the context of the LHC, and provide a further benchmark to probe new SVJ signatures also at hadron colliders via the Higgs boson.

Speaker: Cesare Tiziano Cazzaniga (ETH Zurich (CH))

dark_showers_fccee.pdf

18:10 Are SVJs hidden from the Lund Plane?

In the context of Dark Matter detection at colliders, several new signatures have emerged that bridge experimental data with the theoretical models. Among these, Semi-Visible Jets (SVJs) represent a novel promising signature, particularly those mediated by a resonant Z′ boson that enables the production of heavy dark hadrons. In such a scenario, jets consist of both visible SM particles and invisible dark matter components, resulting in missing transverse energy that is aligned with the jet axis. This feature disrupts conventional LHC search strategies, which typically rely on large angular separations between jets and MET. Consequently, SVJ events closely mimic QCD backgrounds, complicating their discrimination. The complexity is further heightened since the poorly constrained nature of the dark sector; factors such as the dark confinement scale, dark meson masses and the fraction of the invisible DM hadrons. Our study consider different topologies, including heavy gauge boson Z′, resulting in different kinematic features for the produced SVJ from the QCD ones. Traditional SVJ searches have relied either on global kinematic observables or on jet substructure features such as energy correlation functions that highlight differences between SVJ and QCD jets. In this work, we propose a multimodal Transformer-based network with cross-attention, designed to jointly analyse both sources of information. By combining kinematic observables with local jet-level features (encoded in the jet constituents), the model achieves a more effective separation between signal and background. In addition, the adoption of a cross-attention mechanism serves to establish a conditional correlation between the global kinematic variables of the event and the corresponding features within the SVJ that arise from dark hadron dynamics. To reduce the network dependence on specific theoretical assumptions, we train it on mixed datasets generated with different values of the theoretical parameters. Furthermore, we analyze the latent space representations learned by the network to quantify the impact of each theoretical parameter
on the embedded data distribution.

Speaker: Mr Miguel Angel Avendano Bernal (University of Southampton)

Dark_Showers_Conference_Miguel_Talk_v5.pdf

18:25 Surveying the Hidden Valley landscape

The confining and asymptotically free nature of confining Hidden Valleys (HVs) naturally give rise to long-lived mesons and to “dark showers”. Due to the vast theory space inherent to HV theories, we seek to reinterpret existing collider searches by keeping the mediator properties fixed while varying parameters of the HV theory itself. Such an endeavour necessitates accurate modelling of the simulation properties dependence on their fundamental theory parameters.

We specifically focus on theories with very long-lived mesons where we need to reinterpret existing CMS searches targeting long-lived particle decays in the muon system. By using this model-agnostic approach, we show that this existing CMS search is sensitive to the fundamental parameters of the confining DS itself and present model dependent exclusions on its parameter space.

Speaker: Joshua Lockyer

Lockyer_DSWorkshop_final.pdf

18:40 Search for Strongly-Coupled Dark Matter via t-channel Production of Semi-Visible Jets​

We present a search for strongly coupled dark matter in the form of semi-visible jets (SVJs) using the full Run~2 CMS dataset. SVJs arise in Hidden Valley models with a confining dark QCD sector, where dark quarks produced via \textit{t}-channel scalar exchange hadronize into dark hadrons---some decaying to Standard Model particles and others escaping detection. This results in jets with both visible and invisible components, and missing transverse momentum ($p_T^{\text{miss}}$) that is aligned with a jet rather than isotropic. After a set of event preselection requirements, an event-level deep neural network (DNN) trained with a Lagrange multiplier to ensure independence from $p_T^{\text{miss}}$ is applied to select candidate events. This DNN output is used to define the ABCD plane for data-driven background estimation. Within each selected event, individual jets are tagged using both supervised and unsupervised SVJ taggers. The ABCD method is used to estimate Standard Model backgrounds and validate closure directly in data using low-signal control regions, ensuring reliable background predictions in the presence of semi-visible signals.

Speaker: Akshat Shrivastava (Texas Tech University (US))

Dark_showers_Lightning_Talk_Shrivastava_Akshat_final.pdf

Tuesday 28 October

14:00 → 16:00 Progress reports

Status updates from working groups established at January workshop

Conveners: Felix Kahlhoefer (Karlsruhe Institute of Technology), Kevin Pedro (Fermi National Accelerator Lab. (US))

14:00 Flavor Changing Dark Currents

We investigate classes of models that can produce a wide range of r_inv values. The model parameter ranges are chosen to occupy phase space regions that are well-understood from perturbative and lattice QCD results. Different parameter variations can change r_inv while preserving similar showering behavior or produce different showers with degenerate or near-degenerate r_inv values. We present the first exploration of these variations to quantify their impact on observables such as jet substructure, in order to ensure both the breadth and physical reality of proposed simplified benchmark models.

Speaker: Emily Ann Smith (Fermi National Accelerator Lab. (US))

Dark Showers-EmilySmith.pdf

14:25 Studying CKKWL merging/matching scheme for t-channel generation

Comparison of currently used MLM matching scheme and more well motivated CKKWL scheme for t-channel SVJ production will be presented.

Speaker: Deepak Kar (University of the Witwatersrand (ZA))

SVJ-matching-dkar.pdf

14:40 Physics and PYTHIA; What's hidden in the weeds

Speaker: Prof. Matthew James Strassler (Harvard University)

PYTHIAWeeds.pdf

15:10 Hidden Valleys in Herwig

Speaker: Dominic William Stafford (Deutsches Elektronen-Synchrotron (DE))

DS_WS_Herwig_281025.pdf

16:00 → 16:20 Break

16:20 → 18:00 Benchmarking proposals

Conveners: Deepak Kar (University of the Witwatersrand (ZA)), Felix Kahlhoefer (Karlsruhe Institute of Technology)

16:20 Comments on Benchmarking

Speaker: Tim Cohen (CERN)

Cohen_Comments on Benchmarking.pdf

16:50 $t$-channel models

Speakers: Pedro Klaus Schwaller (Mainz University), Pedro Schwaller

T-channel_models_Pedro_2025.pdf

17:20 Dark matter and dark showers in strongly interacting dark sectors at the sub-GeV scale

Speaker: Elias Bernreuther (UC San Diego)

stronglyinteracting_dm.pdf

18:30 → 20:30 Social Dinner: Dinner at Luigia Academy Meyrin

Wednesday 29 October

14:00 → 16:00 Benchmarking panel discussion

Conveners: Suchita Kulkarni (University of Graz), Sukanya Sinha (The University of Manchester (GB))

SSinha_LHC-DS-WS-Oct2025-benchmarking discussion.pdf

Panellists: Kevin Pedro (CMS), Deepak Kar (ATLAS), Graham Kribs (Theory), Matthew Strassler (Theory)

16:00 → 16:20 Break

16:20 → 18:00 Formation of working groups toward white paper

Thursday 30 October

09:00 → 17:00 Hackathon

Conveners: Deepak Kar (University of the Witwatersrand (ZA)), Kevin Pedro (Fermi National Accelerator Lab. (US))

09:00 Kickoff

List of hackathon topics

Speaker: Kevin Pedro (Fermi National Accelerator Lab. (US))

List of hackathon topics

09:30 Work session

10:30 Coffee break

11:00 Work session

12:00 Lunch

13:00 Work session

14:30 Coffee break

15:00 Work session

16:30 Closeout