An Update to the Letter of Intent for MATHUSLA: Search for Long-Lived Particles at the HL-LHC
Cristiano Alpigiani, Juan Carlos Arteaga-Velázquez, Austin Ball, Liron Barak, Jared Barron, Brian Batell, James Beacham, Yan Benhammo, Karen Salomé Caballero-Mora, Paolo Camarri, Roberto Cardarelli, John Paul Chou, Wentao Cui, David Curtin, Miriam Diamond, Keith R. Dienes, Liam Andrew Dougherty, Giuseppe Di Sciascio, Marco Drewes, Erez Etzion, Rouven Essig, Jared Evans, Arturo Fernández Téllez, Oliver Fischer, Jim Freeman, Jonathan Gall, Ali Garabaglu, Stefano Giagu, Stephen Elliott Greenberg, Bhawna Gomber, Roberto Guida, Andy Haas, Yuekun Heng, Shih-Chieh Hsu, Giuseppe Iaselli, Ken Johns, Audrey Kvam, Dragoslav Lazic, Liang Li, Barbara Liberti, Zhen Liu, Henry Lubatti, Lillian Luo, Giovanni Marsella, Mario Iván Martínez Hernández, Matthew McCullough, David McKeen, Patrick Meade, Gilad Mizrachi, O.G. Morales-Olivares, David Morrissey, Meny Raviv Moshe, Antonio Policicchio, Mason Proffitt, Dennis Cazar Ramirez, Matthew Reece, Steven H. Robertson, Mario Rodríguez-Cahuantzi, Albert de Roeck, Amber Roepe, Joe Rothberg, James John Russell, Heather Russell, Rinaldo Santonico, Marco Schioppa, Jessie Shelton, Brian Shuve, Yiftah Silver, Luigi Di Stante, Daniel Stolarski, Mike Strauss, David Strom, John Stupak, Martin A. Subieta Vasquez, Sanjay Kumar Swain, Guillermo Tejeda Muñoz, Steffie Ann Thayil, Brooks Thomas, Yuhsin Tsai, Emma Torro, Gordon Watts, Charles Young, Jose Zurita
CCERN-LHCC-2020-014LHCC-I-031-ADD-1
An Update to the Letter of Intent for MATHUSLA:Search for Long-Lived Particles at the HL-LHC
MATHUSLA mathusla-experiment.web.cern.ch
Cristiano Alpigiani, Juan Carlos Arteaga-Vel ´azquez, Austin Ball, Liron Barak, Jared Barron, Brian Batell, James Beacham, Yan Benhammo, Karen Salom ´eCaballero-Mora, Paolo Camarri, Roberto Cardarelli, John Paul Chou, WentaoCui, David Curtin, Miriam Diamond, Keith R. Dienes, , Liam Andrew Dougherty, Giuseppe Di Sciascio, Marco Drewes, Erez Etzion, Rouven Essig, JaredEvans, Arturo Fern ´andez T ´ellez, Oliver Fischer, Jim Freeman, Jonathan Gall, Ali Garabaglu, Stefano Giagu, Stephen Elliott Greenberg, Bhawna Gomber, Roberto Guida, Andy Haas, Yuekun Heng, Shih-Chieh Hsu, Giuseppe Iaselli, Ken Johns, Audrey Kvam, Dragoslav Lazic, Liang Li, Barbara Liberti, ZhenLiu, Henry Lubatti, Lillian Luo, Giovanni Marsella, Mario Iv ´an Mart´ınezHern ´andez, Matthew McCullough, David McKeen, Patrick Meade, GiladMizrachi, O.G. Morales-Olivares, David Morrissey, Meny Raviv Moshe, AntonioPolicicchio, Mason Proffitt, Dennis Cazar Ramirez, Matthew Reece, Steven H.Robertson, Mario Rodr´ıguez-Cahuantzi, Albert de Roeck, Amber Roepe, JoeRothberg, James John Russell, Heather Russell, Rinaldo Santonico, MarcoSchioppa, Jessie Shelton, Brian Shuve, Yiftah Silver, Luigi Di Stante, DanielStolarski, Mike Strauss, David Strom, John Stupak, Martin A. SubietaVasquez, Sanjay Kumar Swain, Guillermo Tejeda Mu ˜noz, Steffie Ann Thayil, Brooks Thomas, Yuhsin Tsai, Emma Torro, Gordon Watts, Charles Young, Jose Zurita University of Washington, Seattle Universidad Michoacana de San Nicol´as de Hidalgo, Mexico (UMSNH) CERN Tel Aviv University University of Toronto University of Pittsburgh a r X i v : . [ phy s i c s . i n s - d e t ] S e p Ohio State University Universidad Aut´onoma de Chiapas, Mexico (UNACH) Sezione di Roma Tor Vergata, Roma, Italy Rutgers, the State University of New Jersey University of Arizona University of Maryland Universit´e catholique de Louvain YITP Stony Brook University of Cincinnati Benem´erita Universidad Aut´onoma de Puebla, Mexico (BUAP) Karlsruhe Institute of Technology Fermi National Accelerator Laboratory (FNAL) Universit`a degli Studi di Roma La Sapienza, Roma, Italy Hyderabad University New York University Institute of High Energy Physics, Beijing Politecnico di Bari, Italy Boston University Shanghai Jiao Tong University Universit`a del Salento, Lecce, Italy TRIUMF Universidad San Francisco de Quito (USFQ) Harvard University McGill University University of Oklahoma SLAC INFN and University of Calabria University of Illinois Urbana-Champaign Harvey Mudd College Carleton Unversity University of Oregon Instituto de Investigaciones F´ısicas (IIF), Observatorio de F´ısica C´osmica de ˆa Chacaltayaˆa, UniversidadMayor de San Andr´es (UMSA) National Institute of Science Education and Research, HBNI, Bhubaneswar Lafayette College University of Notre Dame Instituto de F´ısica Corpuscular (CSIC-UV), Valencia, Spain SLAC National Accelerator Laboratory Karlsruhe Institute of Technology Institute for Theoretical Physics
E-mail: [email protected]
BSTRACT : We report on recent progress in the design of the proposed MATHUSLA Long LivedParticle (LLP) detector for the HL-LHC, updating the information in the original Letter of Intent(LoI), see CDS:LHCC-I-031, arXiv:1811.00927. A suitable site has been identified at LHC Point 5that is closer to the CMS Interaction Point (IP) than assumed in the LoI. The decay volume has beenincreased from 20 m to 25 m in height. Engineering studies have been made in order to locate much ofthe decay volume below ground, bringing the detector even closer to the IP. With these changes, a 100m x 100 m detector has the same physics reach for large c τ as the 200 m x 200 m detector described inthe LoI and other studies. The performance for small c τ is improved because of the proximity to the IP.Detector technology has also evolved while retaining the strip-like sensor geometry in Resistive PlateChambers (RPC) described in the LoI. The present design uses extruded scintillator bars read out usingwavelength shifting fibers and silicon photomultipliers (SiPM). Operations will be simpler and morerobust with much lower operating voltages and without the use of greenhouse gases. Manufacturing isstraightforward and should result in cost savings. Understanding of backgrounds has also significantlyadvanced, thanks to new simulation studies and measurements taken at the MATHUSLA test standoperating above ATLAS in 2018. We discuss next steps for the MATHUSLA collaboration, andidentify areas where new members can make particularly important contributions. ontents MATHUSLA (Massive Timing Hodoscope for Ultra-Stable neutraL pArticles) [1, 2] is a proposedlarge-scale dedicated Long-Lived Particle (LLP) detector to be situated at CERN near one of the maindetectors. It will be able to reconstruct the decay of neutral LLPs, produced in HL-LHC collisions, asdisplaced vertices (DV) in a near-zero-background environment. The physics case for LLP searchesat the HL-LHC in general and MATHUSLA in particular was explored in detail in [3]. MATHUSLAwould be able to extend the sensitivity in long lifetime and LLP cross section by several orders ofmagnitude compared to the main detectors alone, depending on the production and decay mode. Inparticular, it would allow searches for LLPs with lifetimes near the upper Big Bang Nucleosynthesisbound set by cosmology, and also play a vital role in the search for Dark Matter (DM).We report several updates on the design on the MATHUSLA detector in Section 2. Site-specificengineering studies have been carried out by CERN engineers to identify a suitable location for the– 1 –ATHUSLA detector on CERN-owned land adjacent to CMS. This informs the updated “MATH-USLA @ CMS” geometry, with a 100 m ×
100 m area and a 25 m high decay volume that is exca-vated 20m below grade. This is only the area of the earlier MATHUUSLA200 benchmark [1, 3, 4],a major factor in reducing costs. The updated location is significantly closer to the collision point thanearlier benchmarks, resulting in near-identical LLP sensitivity compared to MATHUSLA200. Wealso identify extruded scintillators as the leading technology choice for MATHUSLA’s tracker systemand are currently conducting detailed design studies.In Section 3 we present new LLP sensitivity estimates for the updated MATHUSLA@CMS ge-ometry, which confirm that the new geometry has near-identical sensitivity to MATHUSLA200. Im-portantly, this means that the myriad of projections in the physics case white paper [3] can be appliedalmost verbatim to the updated design. We also draw upon recent studies from the literature to em-phasize the vital role MATHUSLA plays in the hunt for Dark Matter (DM), especially in scenarioswhere the DM abundance is controlled by the properties of an LLP. Finally, we summarize a recentstudy [5] demonstrating that MATHUSLA can characterize any discovered LLP in great detail, pro-vided it is integrated into the CMS L1 trigger system to allow a combined analysis with main detectorinformation.There have been significant advances in our understanding of backgrounds to LLP searches atMATHUSLA, thanks to several detailed simulation studies as well as measurements conducted bythe MATHUSLA test stand [6] in 2018. As a result, we can confirm earlier estimates that downwardtraveling cosmic rays, muons from the LHC and atmospheric neutrinos can be vetoed and are unlikelyto constitute a background to LLP searches at MATHUSLA. On the other hand, these new resultsalso focus attention on very rare backgrounds, namely the production of long-lived pions, muons andneutral kaons due to cosmic ray inelastic back-scattering in the detector floor. Several veto strategiesare available, but reliable understanding of these ultra rare events necessitates careful study, which arecurrently in progress.We comment briefly on a possible upgrade to MATHUSLA in order to study cosmic rays inSection 5, and then sum up the current status, as well as next steps for the collaboration, in Section 6.It is our aim to complete a technical design report and a robust cost estimate by early 2021, andthere are ample opportunities for new collaboration members to contribute, especially in the areas ofhardware design, geometry optimization, simulation studies, and cosmic ray physics. We now summarize recent studies in choosing a location for MATHUSLA near CMS, the updatedMATHUSLA @ CMS geometry and modular design, and details of the tracking system based onextruded plastic scintillators.
A site has been identified a LHC Point 5 close to the CMS Interaction Point (IP) that allows for a100 m x 100 m detector footprint to be located on the CERN site at P5. The detector building andits relation to the CMS IP is shown in Figure 1; the dashed red line indicates the CERN non-fenced– 2 – igure 1 . Location of proposed MATHUSLA detector at the CMS site. domain boundry. The near side of this site is approximately 70 m from the CMS IP, considerablycloser than the 100 m assumed in the Letter of Intent.Site-specific engineering studies have been carried out by CERN engineers to locate 20 m ofdecay volume below grade, which together with 5 m of decay volume above ground level gives a totaldecay height of 25 m. This is an increase of 5 m over the design described in the LoI [2]. Figure 2shows the excavated volume with retaining walls and other structures. There are two rows of supportthat divide the volume into three regions, each of which is served by a bridge crane with 33 m span.The other columns support the detectors. At the end of the excavated volume is a surface assemblyarea measuring 100 m x 30 m.Putting the detector longitudinally closer to the IP and placing most of the decay volume belowground have increased the solid angle coverage significantly. Thus a smaller detector with 100 mx 100 m footprint in this configuration has the same physics reach as the 200 m x 200 m detectordescribed in the LoI. We demonstrate this in Section 3.
Further simulations studies and information from the test stand we operated at P1 [6] have informedour revised MATHUSLA layout at CMS shown in Figure 3. The current detector concept shownin Figure 4 is a 100 m detector consisting of 100 m × m units. Each detector unit comprises 9-layers of scintillating-detector planes that provide position and timing coordinates of charged particlesresulting from the decay of long-lives particles in the MATHUSLA detector decay volume. Thereare five detector planes, separated by m at the top, two additional planes, also separated by mlocated 5 m below (to enhance the particle position measurement precision close to the floor) andtwo additional planes at the floor for rejecting charged particles from the LHC and cosmic muon– 3 – igure 2 . Engineering details of the partially excavated structure that would house MATHUSLA. Figure 3 . MATHUSLA@CMS geometry relative to the CMS collision point. backscattering. The total height of ∼ m includes a ∼ m LLP decay volume, 21 m of whichwould be excavated, and 12 m above the surface that hosts the tracker and the crane system used forassembly and maintenance.The layout of the the building that houses the 100 m ×
100 m experimental area and the adjacent30 m ×
100 m adjacent area for the detector assembly is shown in Figure 1. The structure, which islocated on the surface near the CMS IP fits well on CERN owned land. Having a large part of thedecay volume underground brings it closer to the IP, which increases the solid angle in the acceptance– 4 – igure 4 . Schematic view of the MATHUSLA detector modular concept: left the ( m × m) units; rightthe detector planes in each module for LLPs generated in the collisions. To adjust to the available land, this proposal has a 7.5 m offsetto the centre of the beams. The site allows for the detector to be as close as 68 m from the IP, whichis show in red in Figure 4. Each of the m × m detector planes consists of an assembly of extruded scintillating bars whoselength, width and thickness is . m, . cm, cm, respectively. Each bar is extruded with a holeat the centre into which a wave-length shifting (WLS) fibre is inserted and connected to an SiPM.To facilitate installation the scintillating bars are assembled into sub-units that comprise barsresulting in . m × . m sub-units that allow for overlapping the sub-units by cm longitudinallyand . cm transversely in order to avoid gaps in coverage. In this arrangement the sub-units providehermetic coverage over a m × m area. We now present updated (and slightly improved) LLP sensitivity estimates for the new MATH-USLA@CMS benchmark geometry presented above, and demonstrate that the reach is essentiallyidentical to the old MATHUSLA200 benchmark geometry from the original LOI [2]. We also em-phasize that LLP searches are instrumental in the hunt for Dark Matter and are often the only wayof observing the DM directly, as demonstrated by several recent studies [7–10]. We also summarizerecent work [5], which shows that analysis of MATHUSLA and CMS data together can characterizethe LLP in great detail, including determining the production mode, decay mode, and (if applicable)parent particle mass. This serves as powerful motivation to integrate MATHUSLA with CMS so it cansupply a L1 trigger signal, ensuring that the necessary information in the main detector is recorded.– 5 – .001 0.100 10 1000 10 - - - - c τ X ( m ) B r ( h -> XX ) σ pp → h → XX ( f b ) s =
14 TeV, 3ab - h → XX, X → jj m X =
20 GeV
BBN L i m i t HL - LHC h → invisiblesensitivity A TL A S M S D V M A T H U S L A a t C M S D ecays do m i n a n t l y i n M a i n D e t ec t o r Figure 5 . Red curve: MATHUSLA@CMS sensitivity (4 observed events) for LLPs of mass m X = 20 GeV produced in exotic Higgs decays. Black curve: reach of ATLAS search for a single hadronic LLP decay in theMuon System at the HL-LHC [11].
200 400 600 800 1000 120010 μ [ GeV ] c τ [ m ] Number of observed higgsino → gravitino events4840 200 400 600 800 1000 120010 μ [ GeV ] F [ G e V ] Number of observed higgsino → gravitino events m [ k e V ] Figure 6 . Number of ˜ H → ˜ G + ( Z, h ) events that MATHUSLA@CMS could observe from electroweakproduction of higgsinos at the HL-LHC with an integrated luminosity of 3 ab − . Left: higgsino mass µ versuslifetime cτ in meters. Right: higgsino mass µ versus the SUSY breaking scale as parametrized by √ F in GeV(label on left axis) or gravitino mass m / in keV (label on right axis). In a wide swath of parameter spacewith higgsino lifetimes ranging from smaller than m to larger than m, MATHUSLA could provide adiscovery of new physics with electroweak cross-sections for which the HL-LHC would fail to discover newphysics. – 6 – .1 LLP Benchmark Models The physics case for LLP searches in general and at MATHUSLA in particular was discussed ingreat detail in the physics case white paper [3] and subsequent studies [5, 7, 10, 12–33]. All thesensitivity estimates in [3] and [5, 7, 10, 12–33] assumed the MATHUSLA200 benchmark geometryfrom the original Letter of Intent [2], which is identical to the original proposal [1]: a 200 m ×
200 m ×
20 m decay volume, with 100 m displacement both horizontally and vertically from the LHCinteraction point. The much more realistic MATHUSLA@CMS benchmark geometry introduced inthis note is significantly smaller, with a partially excavated decay volume of a 100 m ×
100 m ×
25 m, displaced 70 m horizontally and 60 m vertically from the CMS interaction point. We now showupdated sensitivity curves of MATHUSLA@CMS for several LLP benchmark models at the LHC:exotic Higgs decays, long-lived Higgsinos, as well SM+S (light scalar mixing with the Higgs) andRHN (Right-Handed Neutrinos). This demonstrates that the sensitivity is essentially identical to theold MATHUSLA200 benchmark, since the reduced distance from the IP makes up for the smallerdecay volume. Therefore, all of the MATHUSLA200 sensitivity estimates apply for the updatedMATHULSA@CMS benchmark almost verbatim. Note that the sensitivity estimates we present here, like those of [3], assume perfect detectionefficiency as long as the LLP decays in the decay volume, and assume zero backgrounds after therigorous geometric DV reconstruction cuts have been applied. Both the zero-background assumption(see new studies in Section 4) and perfect reconstruction assumption (see preliminary reconstructionstudies in LOI [2]) are good approximations for LLPs decaying into a high multiplicity of final states,which is the case for MATHUSLA’s most important physics target, hadronically decaying LLPs withmasses in the O (10 GeV) − O (100 GeV) range. The extent to which these ideal assumptions holdfor leptonically decaying or very light LLPs with masses < ∼ GeV is the subject of ongoing study bythe MATHUSLA collaboration, and also depends on details of the final detector design.We first consider examples of weak- or TeV-scale LLPs produced at the LHC. Figure 5 showsthe sensitivity to hadronically decaying LLPs produced in exotic Higgs decays, which arises in alarge variety of new physics scenarios [35], including solutions to the Hierarchy Problem like NeutralNaturalness [36–39]. The LLP cross section sensitivity on the right axis approximately applies tomost weak-scale LLP production processes [3]. Figure 6 shows the reach for meta-stable Higgsinoswithin Supersymmetry with gauge-mediated SUSY breaking [40, 41]. In both cases, the observableLLP production rate is nearly identical to the sensitivity of the old MATHUSLA200 benchmark [1, 2],with slightly improved sensitivity at shorter lifetimes.We now consider two of the benchmark models for low-mass LLP that were also studied by thePhysics Beyond Colliders (PBC) working group [19]. Figure 7 shows the sensitivity for a singletscalar LLP that has a tiny mixing angle θ with the Higgs boson (for details see [19, 42]). The sensi-tivity is again nearly identical to MATHUSLA200. Part (b), where an exotic Higgs decay branchingfraction of Br( h → ss ) = 0 . is assumed as an additional LLP production process, demonstrates This small offset of the detector center line from the beam is neglected in our simulation studies, since its preciseparameters may change and it is unlikely to have a significant effect on LLP sensitivity. The only significant difference is that MATHUSLA@CMS has an improved sensitivity for shorter LLP lifetimes com-pared to MATHUSLA200, again due to the smaller distance to the IP. – 7 –
ATHUSLA @ CMS to redo these, just output plot in the same way, then give new plot pdf the same dimensions & coordinates as the aligned one - - - - - - mS / GeV S i n θ (a) MATHUSLA @ CMS, 3 ab -1 - - - - - - mS / GeV S i n θ - - - - - - m S ( GeV ) S i n θ (b) (c) Figure 7 . Purple curves: sensitivity of MATHUSLA@CMS for a singlet scalar LLP s mixing with Higgsmixing angle θ . (a) Assuming production in exotic B , D , K meson decays only. (b) Assuming additionalproduction in exotic Higgs decays with Br( h → ss ) = 0 . . Figures (a) and (b) are reproduced from the PBCBSM Working Group report [19] with the purple MATHUSLA@CMS curves added. This shows sensitivityof various other existing and proposed experiments, as well as the old MATHUSLA200 benchmark estimates(yellow curves). (c) Same scenario as (b) but showing the entire MATHUSLA sensitivity due to h → ss decays. the advantage gained by the high LHC energy even when searching for very light LLPs, since theycan be produced in high-scale processes that are kinematically suppressed at intensity frontier ex-periments. Also shown are the expected sensitivities of other proposed experiments like FASER [4],CODEX-b [9], and SHiP [43].Figure 8 (a) - (c) shows MATHUSLA’s reach for Heavy Neutral Leptons (HNL) that dominantlymix with only electron, muon or tau active neutrinos. These estimates are slightly improved comparedto the previous results in [3, 19], since the production and decay rate calculations have been updated toagree with the latest results in [34], also adopted by e.g. [44]. The effect of this improved calculationfor the old MATHUSLA200 benchmark geometry can be seen in Figure 8 (d) by noting the smalldifferences between the dark yellow and green curves. The effect of the new MATHUSLA@CMS– 8 – ATHUSLA @ CMS, 3 ab -1 solid: B, D mesonsdashed: Kaons �� - � � �� �� � �� - �� �� - �� �� - �� �� - � �� - � �� - � �� - � �� - � �� - � �� - � �� - � � � [ ��� ] | � � � MATHUSLA @ CMS �� - � � �� �� � �� - �� �� - �� �� - �� �� - � �� - � �� - � �� - � �� - � �� - � �� - � �� - � � � [ ��� ] | � μ � MATHUSLA @ CMS, 3 ab -1 solid: B, D mesonsdashed: Kaons (a) (b) MATHUSLA @ CMS �� - � � �� �� � �� - �� �� - � �� - � �� - � �� - � �� - � �� - � �� - � �� - � � � [ ��� ] | � τ � MATHUSLA @ CMS, 3 ab -1 (from B, D mesons) �� - � � �� �� � �� - �� �� - �� �� - �� �� - � �� - � �� - � �� - � �� - � �� - � �� - � � � [ ��� ] | � μ � MATHUSLA200 ( PBC, B,D only ) MATHUSLA200 ( updated, B,D only ) MATHUSLA200 ( updated, Kaons ) MATHUSLA @ CMS ( B,D only ) MATHUSLA @ CMS ( Kaons ) (c) (d) Figure 8 . Purple curves: sensitivity of MATHUSLA@CMS for a Heavy Neutral Lepton LLPs produced in
B, D (solid) and K decays (dashed) for HNLs that mix predominantly with electron (a), muon (b) or tau (c)active neutrinos. These figures are reproduced from the PBC BSM Working Group report [19] with the purpleMATHUSLA@CMS curves added. This shows sensitivity of various other existing and proposed experiments,as well as the old MATHUSLA200 benchmark estimates (yellow curves). (d) For the muon-dominated mixingscenario, comparison of the old MATHUSLA200 sensitivity estimate in the PBC and MATHUSLA physics casewhite papers [3, 34] (dark yellow) to the MATHUSLA200 (green) and MATHUSLA@CMS (purple) sensitivitycomputed using the latest results in [34] (blue). The dark yellow and purple curves are the same as in plot (b).The small difference between the green and purple curves demonstrates that the new MATHUSLA geometryhas very little effect on sensitivity, and the difference between the yellow and purple curves in (a)-(c) is mainlydue to the improved HNL production and decay calculation. geometry is then extremely minor, same as for the other LLP production modes, as can be seen bycomparing the green to the purple curve. Additionally, the production of HNLs in Kaon decays isnow taken into account (dashed curves). This is denoted separately, since the resulting HNLs arerelatively soft and MATHUSLA’s ability to reconstruct such light, soft LLPs depends on the finaldetector design. – 9 – .2 Dark Matter It is perhaps not commonly appreciated that LLP searches are crucial to complete the search programfor DM. In models like Freeze-In DM (FIDM) [7, 45], inelastic DM (iDM) [46, 47], co-annihilatingDM [8] or co-scattering DM [9], the relic abundance of the stable DM candidate is determined bythe properties of an LLP in the thermal plasma of the Big Bang. This LLP carries the same quantumnumber which stabilizes DM, and decays into DM + SM final states. The DM particle itself could bealmost completely sterile, precluding a direct detection signal, and production of the parent LLP atcolliders could then be the only way to produce and observe DM.The mass of the LLP can easily exceed the GeV scale, meaning they can only be produced inthe high-energy collisions of the LHC, while the connection between the LLP properties and the DMrelic abundance often pushes the necessary LLP searches into the long-lifetime regime. All of thismeans that MATHUSLA has a unique and invaluable role to play in the search for Dark Matter, whichwe demonstrate in this section by briefly reviewing several recent studies from the literature. (Allof these studies used the original MATHUSLA200 benchmark, but as we demonstrated in the lastsection, those results also apply to the updated MATHUSLA@CMS geometry.)We first consider Freeze-In DM [7, 45]. In FIDM, the DM candidate X is essentially sterile, withtiny or non-existent direct coupling to SM particles. X therefore never reaches equilibrium with theSM bath during the Big Bang, and its abundance, assumed to be negligibly small after reheating atthe end of inflation, cannot be set by thermal freeze-out. However, if there is a parent particle A inthermal equilibrium with the plasma that has a long-lived decay A → B SM X , (3.1)where B SM is some SM state, then the X abundance can be gradually populated while A is relativisticduring the radiation-dominated era, . Most of the X particles are produced around T ∼ m A / , justbefore A disappears from the plasma due to its efficient annihilation into SM particles. This is referredto as DM “freeze-in”. The lifetime and mass of the parent particle A directly set the DM abundance, and since it hassizable coupling to the SM, A is an LLP that can be produced at colliders. A simplified modelof FIDM, considered in [3] and further analyzed in [7], serves to illustrate the DM reach of LLPsearches. It consists of one electroweak doublet Dirac fermion ψ and one singlet Dirac fermion χ with a higgs couping y χ ¯ ψHχ . In the regime m ψ > m χ and y χ (cid:28) , the two fermions acquire a smallmixing after electroweak symmetry breaking. The ψ -dominated mass eigenstate χ with mass m then acts as the parent LLP with sizable SM couplings, decaying to the χ -dominated mass eigenstate χ with mass m , which is the FIDM candidate. This model is similar to the singlet-doublet DM While we focus on the DM reach from LLP searches, there are also much more exotic possibilities. For example,MATHUSLA could at as a direct detection experiment for strongly-interacting very heavy DM, showing up as a very slowdownward track with multiple hits in each tracker element [48]. Assuming standard cosmology, see [49, 50] for the impact of modified thermal histories Freeze-in scenarios can also utilize suppressed annihilations of particles in the SM bath to produce X , for a recentreview see e.g. [51]. – 10 –
00 300 500 700 900 1100 1300 1500 1700 1900 m (GeV) − c τ χ ( m )
200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 m (GeV) − − − − − m ( G e V ) MATHUSLA200 95% CL (3000 fb ) ⌦ h = 0.12 ( m = 1 GeV, T EW = 50 GeV) ⌦ h = 0.12 ( m = 1 GeV, T EW = 160 GeV) ⌦ h = 0.12 ( m = 10 MeV, T EW = 50 GeV) ⌦ h = 0.12 ( m = 10 MeV, T EW = 160 GeV) ⌦ h = 0.12 ( m = 100 KeV, T EW = 50 GeV) ⌦ h = 0.12 ( m = 100 KeV, T EW = 160 GeV) Lyman- α exclusion DV + MET 95% CL (3000 fb -1 ) Disappearing Tracks 95% CL (3000 fb -1 ) MATHUSLA200 (4 observed events, 3000 fb -1 ) Figure 9 . Left:
HL-LHC reach of MATHUSLA (red, 4 observed events) and ATLAS/CMS searches for DV +MET (purple) and disappearing tracks (cyan) in the mass-lifetime plane of the parent LLP χ in our simplifiedFIDM model. The blue region excluded from Large-Scale Structure constraints derived from Lyman- α forestobservations. The solid and dotted near-horizontal lines denote mass-lifetime relationship required for a 1 GeV,10 MeV or 100 KeV DM particle χ to have the observed relic abundance, for different temperatures of theelectroweak phase transition. Figure adapted from [7] and [3]. Right: reach of MATHUSLA (4 observed events)in the LLP-DM mass plane. This figure is adapted from [3] and assumes the mass-lifetime relationship of theearlier analysis which differs slightly from the newer analysis [7], but still illustrates the reach of MATHUSLAin the DM mass plane. model considered in [52, 53], and serves as a close analogue of the Higgsino-Axino system [49, 54–56] or a feebly interacting Higgsino-Bino system. Figure 9 shows the reach of MATHUSLA in eitherthe mass-lifetime plane of the LLP (with near-horizontal curves showing where DM of varying masshas the correct relic abundance), or the LLP-DM mass plane. Also shown is the projected reach ofdisappearing track and DV+MET searches at ATLAS/CMS for the HL-LHC. DM masses above theMeV scale imply the long-lifetime regime for the parent LLP, making MATHUSLA the most sensitiveexperiment to probe large regions of FIDM parameter spaceAnother example of this very direct LLP-DM connection is inelastic DM [46, 47], where thestable DM state χ can only interact with the SM by turning into a slightly heavier and meta-stabledark state χ . This means direct detection experiments can only detect iDM with a mass splittingsmaller than the nuclear recoil energy < ∼
100 keV , and larger mass splittings have to be probed atcolliders.The recent study Ref. [10] examined a simple iDM model with a Dirac pair of two dark Weylfermions oppositely charged under a dark broken U (1) D that has mass m A (cid:48) and kinematic mixing (cid:15) with the SM photon. In addition to the Dirac mass between the two fermions, symmetry breakinggenerates small Majorana mass terms. This results in two mass eigenstates χ , with small masssplitting ∆ = ( m − m ) /m < ∼ . and an off-diagonal coupling ie D A (cid:48) µ ¯ χ γ µ χ to the dark photon.For m A (cid:48) > ∆ · m , χ is an LLP that decays to a DM particle χ and two SM fermions via anoff-shell A (cid:48) with a lifetime given dominantly by the mass splitting ∆ and the photon portal coupling– 11 –
10 10 - - - - - m @ GeV D e Fermionic iDM, m A ¢ = m , D= a D = W c > W D M BaBar L H C H D M J L L H C H t i m i n g L LEPBelle II F A S E R C O D E X - b M A T HU S L A L H C b - - - - - m @ GeV D e Fermionic iDM, m A ¢ = m , D= a D = W c > W D M BaBar LEPBelle II C O D E X - b M A T HU S L A Figure 10 . Reach of MATHUSLA and other LHC experiments and searches for iDM with a dark photon ofmass m A (cid:48) that has kinetic mixing (cid:15) with the SM photon, and mass splittings ∆ in the percent range. The blackcurve indicates where thermal o-annihilations χ χ → A (cid:48) → f ¯ f to SM fermions give the observed DM relicdensity. Figure taken from [10]. (cid:15) . χ χ pairs can be produced via an s -channel dark photon in colliders and fixed target experiments,allowing LLP searches to probe this class of models. Figure 10 shows the reach of MATHUSLA,FASER, CODEX-b and various LHC searches for dark matter of mass m with dark photon kineticmixing (cid:15) for mass splittings in the percent range ∆ ∼ . . The black line shows where χ has theright relic density from thermal freeze-out via χ χ → A (cid:48) → ¯ f f co-annihilations to SM fermions.Note that only the LHC can produce χ χ pairs at high rate for dark photon masses above ∼
10 GeV ,demonstrating the importance of the energy frontier for probing even relatively light DM models, anddirect detection experiments have no reach due mass splittings around . − GeV. MATHUSLAprovides the best reach and is the only way to probe large parts of motivated iDM parameter spacewhere the DM candidate has the observed relic density, in particular for mass splittings below apercent.iDM and FIDM demonstrate how the properties of an LLP can be directly tied to the observedrelic DM abundance. It is also possible for LLPs to be motivated or even necessary to realize agiven DM mechanism, but without the properties of the LLP directly setting the precise abundance.This is the case for Asymmetric DM (ADM) [57], where the baryon asymmetry of the universe hasa common origin with an asymmetric DM relic abundance. The primordial asymmetry is sharedbetween the visible and dark sectors by higher dimensional operators, which only have to be largeenough to establish chemical equilibrium between the dark and visible sectors. Realizing this ideawithin supersymmetric models, for instance, de-stabilizes the Lightest Supersymmetric Particle (LSP)in the visible sector and makes it an LLP with a decay to both visible and DM states. Crucially, sincethe relic abundance in ADM is not tied to the significant equilibrium interactions of the DM with SMparticles, standard probes of DM, such as direct detection, may be insensitive, so that DM productionin LLP decays is the only viable discovery channel. The reach of MATHUSLA for SUSY with– 12 – - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Figure 11 . In the co-annihilation scenario studied by [8], with DM mass m χ , relative mass splitting to nextlightest dark state δ , and light scalar mediator mass m φ , the left plot shows existing cosmological and colliderconstraints on the plane of DM mass and φ − Higgs mixing angle. The right plot shows the reach of MATH-USLA, CODEX-b, FASER (analogous to Fig. 7), LHCb as well as direct detection experiments. LLP searchesprovide the only probe of large regions of DM parameter space. Figure from [8] unstable LSPs is demonstrated for unstable Higgsinos in Fig. 6 and was studied for RPV SUSY in [3].TeV-scale LLPs with a wide range of lifetimes can be probed, and depending on the decay channel,the lifetime sensitivity can be several orders of magnitude better than main detector searches.Conversely to the ADM example, some DM models predict and rely on the existence of LLPs ina very instrumental fashion, but the LLP signature itself does not involve the DM particle directly. Forexample, the co-annihilation scenario studied by [8] features a light scalar φ that mixes with the Higgsand provides the final state for χ i χ j → φφ annihilations that set the abundance of the DM candidate χ . The properties of φ set the DM abundance, but the search for φ corresponds exactly to the searchfor the SM+S simplified model, see Fig. 7, and involves no DM production in itself. Even so, the LLPsearch can provide the most sensitive probe of the model, at far smaller couplings than is possible indirect detection experiments. This is demonstrated in Fig. 11 from [8].Another well-known example in this class is the ν MSM [58, 59]. In this scenario, DM is a keVsterile neutrino produced in the early universe through the resonant Shi-Fuller mechanism [60]. Thechemical potentials that are required for this resonance to occur are produced in the decays of thetwo heavier siblings of the DM sterile neutrino [61]. Those heavier right-handed neutrinos are LLPsand can be searched for at MATHUSLA and other LLP experiments, see Fig. 8. MATHUSLA wouldtherefore indirectly probe the mechanism of DM generation [62, 63].Both of these last scenarios vividly demonstrate that simplified LLP scenarios like the HNL andSM+S model show up as components in more complete hidden sectors in general and dark matter– 13 –odels in particular, additionally motivating these searches. At the same time it is important to keepin mind that even in above scenarios like co-annihilation, other LLP signatures can easily be generated.The iDM scenario [10] discussed above is one example, another is the related scenario of freeze-outthrough co-scattering [64], which can result in long-lived states that could only be produced in exoticZ decays [9] at the LHC, where MATHUSLA could provide the best sensitivity.Finally, there are many theories where DM is embedded in a larger hidden sector that, by itsnature, is motivated to contain observable LLPs, even if other aspects of the new physics determinethe DM properties and abundance. For example, any composite hidden sector that contains a stableDM candidate will generically include a tower of slightly heavier states with possibly small masssplitting, meaning hidden sector production at colliders gives rise to LLP signals at collider-observablelifetimes [65] with DM in the final states of the unstable particle decays. This is highly motivated inSIMP and ELDER models [66, 67], where → annihilations set the relic abundance, and whichare most easily realized in composite theories (though perturbative UV completions that give rise toLLPs also exist [68]). Dynamical Dark Matter [15, 69, 70] is another example in this broad category,where a tower of unstable states of widely varying lifetimes make up the observed DM relic density,and shorter-lived states in the spectrum could be observable as LLPs at colliders.The above discussion makes clear that LLP searches are a vital component of the DM searchprogram. Depending on the DM scenario, it is possible that the properties or existence of LLPsdirectly determines the abundance of DM, or that the nature of the hidden sector which gives riseto DM also has to give rise to LLPs. In many cases, LLP decays to DM + SM are the only DMproduction mechanism that is available. This raises the question of how MATHUSLA could detectDM in LLP decays. We briefly discuss this below Here we summarize several recent studies [5, 13, 71] that demonstrate MATHUSLA, working intandem with the CMS main detector, can characterize the newly discovered physics in great detail, in-cluding determination of LLP decay mode, mass, production mode, and underlying parameters of thenew physics theory like the parent particle mass. Unlocking this surprisingly broad analysis capabil-ity, given that MATHUSLA does not collect energy or momentum information, requires MATHUSLAto act as a Level-1 Trigger for CMS, allowing for the correlation of information from both detectors.Let us start with the basic properties of the LLP in MATHUSLA. It has been shown [71] thatMATHUSLA can use geometric information from LLP decays to extract the likely velocity, or boost b , of the LLP, thereby determining the LHC bunch crossing that produced the LLP to be identified.The multiplicity of final states in the DV strongly constrains the likely LLP decay modes, with lep-tonic/hadronic decays giving rise to low- or high-multiplicity final states. Refinements based purelyon geometrical information as well as measurement of non-relativistic tracks are possible, includingestimation of the hadronic decay’s jet flavor [71] or the detection of an invisible component, possiblysignifying the production of DM in an LLP decay [13].Determining the LLP mass, as well as other parameters of the underlying theory, is possible bycorrelating information from MATHUSLA with the main detector, and by observing that for a known– 14 – ample of LLP events. Each eventis an LLP decay in MATHUSLApaired with a hard vertex in CMS.No jets with p T > GeV in90% of events? B -meson DecayOne lepton with p T > GeV in20% of events? Charged Current n jet ≥ in85% of events? n jet ≥ in30% of events? Heavy Parent 1-jetHeavy Parent 2-jetVBF-like jets in & of events? Exotic Higgs DecayLLP boostdistribution spread σ b > h b i + . ? Direct Pair ProductionHeavy ResonanceYesNo YesNo Yes Yes NoNo YesNo YesNo Figure 12 . Summary of a prototype hierarchical LLP production mode classification algorithm. A sampleof LLP observations is considered, and at each step a cut on a sample-level variable is used to classify theproduction mode. Extension to include additional LLP production modes, like parent paraticles decaying toLLP + leptons, or more general higgs-like scalars, is straightforward [5]. – 15 –LP production mode, the measured boost of the LLP is highly correlated with its mass. This wasrecently studied in detail [5], and we summarize the main results here.If MATHUSLA can supply a L1 trigger signal to the main detector, each LLP decay reconstructedin MATHUSLA can be associated with an event recorded at CMS. Given a sample of observedLLPs, a simple classifier using cuts on the collective kinematic properties of the MATHUSLA-CMSevents is sufficient to classify samples of LLP observations into one of the simplified LLP productionmodels recently defined in [72]. The production modes considered include exotic B meson decay toone LLP plus SM mesons and/or leptons (not included in the simplified models of [72] but addedhere as a stand-in for low-scale LLP production mechanisms), “Exotic Higgs decay” to two LLPs, aheavy vector-like resonance decaying to two LLPs (“Heavy Resonance”), a charged vector resonancedecaying to one LLP and one SM charged lepton (“Charged Current”), a pair-produced heavy parentdecay to an LLP plus SM jets (“Heavy Parent”), and “Direct Pair Production” of LLPs through aneffective 4-point interaction. The framework can also be easily included to include other productionmodes modes like heavy parent decay to leptons and LLPs or more general Higgs-like scalar decay toLLPs. Fig. 12 shows how the cuts are defined and sequentially applied in the classifier to discriminatebetween each production model. Assuming only 100 observed LLP events originating from a singleproduction mode, samples can be accurately classified with probability from − , dependingon the model, for a wide range of BSM particle masses and widths.Once the LLP production topology is identified, information from MATHUSLA and CMS can beused to determine the underlying parameters, most importantly LLP mass and, if applicable, the massof the LLP parent particle. For Exotic B -decay, Exotic Higgs Decay and Direct Pair Production, theLLP boost distribution reveals the LLP mass at < ∼ precision with only 100 observed events. Forthe other production modes, an additional variable is needed in addition to LLP boost to determineboth LLP mass and parent particle mass. A maximum likelihood fit in lepton p T (Charged Current) orjet H T (Heavy Parent) measured in the main detector together with the measured LLP boost distribu-tion in MATHUSLA allows both parent and LLP masses to be determined with percent-level precisionassuming only 100 observed events. The Heavy Resonance model is the most challenging due to theabsence of additional visible objects in the main detector at lowest order. The LLP boost very accu-rately determines m parent /m LLP , but absolute determination of the parent mass scale requires eithervery high statistics ( ∼ While uncertainty in the LLP velocity and pileup can both lead to a list of possible LLP production vertices, this is notprohibitive for the tasks of classifying the production mode and measuring model parameters [5]. – 16 –
Backgrounds to LLP Searches
The new physics reach of MATHUSLA relies on its ability to distinguish LLP decays from the otherparticles that interact with or inside the detector. The vast majority of tracks observed in MATHUSLAwill originate from cosmic rays and, at a much lower rate, energetic muons from the LHC. It is notobvious how any of these processes could fake an LLP decay, given the stringent geometric andtiming requirements on the DV signal, but the high integrated rate of these backgrounds over the runof the HL-LHC necessitates a careful analysis. Atmospheric neutrinos can also scatter inelasticallywith atoms in the air-filled decay volume, a process which occurs at much lower rate than otherbackgrounds but which has the potential to be dangerous since it can give rise to a genuine DV. In thissection, we provide a major update on the background estimates in [1, 2]. Data from the MATHUSLATest Stand [6] and additional simulations provide a much clearer picture of the cosmic ray backgroundin Section 4.1. Detailed simulations of muon production at the LHC and propagation in the rocktowards MATHUSLA for the updated geometry presented in this note are presented in Section 4.2.Finally, Section 4.3 discusses full simulation studies of atmospheric neutrinos scattering off air inthe MATHUSLA decay volume, updating earlier analytical calculations. These new results largelyconfirm earlier estimates that the background-free regime for LLP searches should be achievable,but also identify important areas of future detailed study that are currently being undertaken by theMATHUSLA collaboration.
The discussion of cosmic rays at MATHUSLA@CMS is calibrated by measurements taken with theMATHUSLA Test Stand [6]. The Test Stand had an active area of 2.5 m × The measured downward track rate was about 30 Hz [6]. Simulations were used to estimate thatthe geometric acceptance of the Test Stand for downward traveling cosmic rays was The trackreconstruction efficiency was estimated from unbiased trigger events in data to be 40%. Taking thisinto account, the observed rate of downward tracks corresponds to a measurement of the local cosmicray muon flux: Φ CR µ ≈ . − min − , (4.1)in good agreement with expectation [77]. This can be applied to the 100 m ×
100 m area of MATH-USLA@CMS. MATHUSLA has much greater geometrical acceptance than the test stand, and is alsolikely to have much better track reconstruction efficiency. We therefore assume that every cosmic ray A reconstructed track has to pass through all layers of the test stand detector, and its narrow vertical shape results inthis low geometric acceptance. – 17 –itting the top tracking layer of MATHUSLA can be reconstructed. This results in a downward CRrate of 2 MHz over the entire detector, or N down ∼ × (4.2)CR tracks over the run-time of the HL-LHC, in agreement with earlier estimates. Downward-travelingCRs by themselves are extremely unlikely to fake an LLP: with a ∼ ns time resolution, imposing therequirement that a two-pronged DV is made up of only upwards going tracks rejects all cosmic raysat MATHUSLA, even before taking into account geometrical information [1, 2]. However, the CRscan give rise to other backgrounds that need closer examination. CRs that hit the floor of the decay volume can produce upwards traveling SM particles through in-elastic backscatter, or through the decay of stopped muons in the floor. The measured ratio of upwardtracks to downward tracks in the test stand without beam was N TSup N TSdown = (7 . ± . · − . (4.3)Different CR particle species produce upwards traveling tracks at different rates, necessitating detailedsimulations to estimate the rate of upwards tracks.Our simulations reproduced the measured ratio of upward tracks to downward tracks with veryreasonable agreement. The composition of downward cosmic ray tracks was estimated in our simula-tions to be 49% positive muons, 43% negative muons, 4% electrons, 3% positrons, and 1% protons.The upward tracks induced by cosmic rays can be attributed to (by incident downward particle) 39%neutrons, 20% protons, 19% positive muons, 10% negative muons, 5% positrons, 4% photons, and 3%electrons. The upward tracks themselves are composed primarily of electrons, positrons, and chargedpions. Other particles such as protons, muons, and charged kaons collectively represent less than 5%of the upward tracks.There are two distinct populations of these upward tracks that can be distinguished by timing.About 76% of the tracks are consistent with the timing of an upward particle immediately createdby the incident downward cosmic ray interacting with material in the test stand or the floor. Theremaining 24% of tracks are delayed by tens to thousands of nanoseconds relative to the incidentparticle. These are the result of low-energy muons stopping in or near the test stand and decaying,leading to a late upward electron or positron.The measured ratio of upwards tracks to downward tracks in Eqn. (4.3) can be readily appliedto MATHUSLA@CMS, since geometrical acceptances and efficiencies largely drop out of the ratio.This allows us to estimate that MATHUSLA would see roughly N up ∼ × (4.4)upwards tracks over the HL-LHC run, mostly electrons, protons, and charged pions.This flux of upward tracks could fake a DV by random crossings. The rate of this backgroundcan be readily estimated, assuming the tracks occur randomly over the are of the detector and are– 18 –ncorrelated. Assuming two tracks have to pass within distance ∆ d of each other in a time interval ∆ t in order to pass geometrical and timing DV reconstruction cuts, the number of “fake” DVs fromrandom upward tracks is roughly N fake DV ∼ . × (cid:18) N up (cid:19) (cid:18) ∆ d (cid:19) (cid:18) ∆ t (cid:19) (4.5)Given that MATHUSLA’s trackers will have < ∼ Br( π + → e + e − e + ν e ) ≈ × − [77], that could fake a multi-pronged DV from BSM LLP decays O (1) times throughout the entire MATHUSLA run. Muons,which constitute a much smaller fraction of upwards traveling particles, also have a rare decay Br( µ − → e − e + e − ν µ ¯ ν e ) ≈ × − that could contribute to LLP backgrounds at a similar rate.Finally, neutral kaons make up a tiny but as yet undetermined fraction of produced particles. Theyhave a dominant decay to three charged particles, making them another possible LLP background.The above discussion must be regarded as strictly preliminary. Reliably estimating the rates ofthese extremely rare processes requires careful study and high-statistics simulations to estimate themuon and especially K L production rates from cosmic rays incident on the detector floor. This canthen be combined with the known decays of these particles to arrive at a more complete picture ofthese ultra-rare backgrounds to LLP searches. Several veto-strategies are available, which depend onthe spectrum of the produced upwards-traveling particle species as well as the precise design of thefloor detector, which could be optimized to help reject these backgrounds by detecting the originatingcosmic ray hit. The MATHUSLA collaboration is currently in the process of conducting the requiredstudies. In this section, we provide an estimate for the rate of upwards traveling Muons produced at the HL-LHC and detected by MATHUSLA. Compared to earlier estimates, these simulations use a moreaccurate material description of the rock and CMS cavern. Furthermore, we now only count muonsthat pass through scintillator layers, which gives us a better estimate of the geometric acceptance.These simulations, run with the appropriate geometrical differences, closely reproduce the number ofupwards tracks observed in the MATHUSLA test stand while with the beam on [6].High energy muons with enough energy to reach MATHUSLA are dominantly produced in W -production and ¯ bb production at the HL-LHC, which is simulated in PYTHIA8 [78]. The b ¯ b productioncross-section is obtained directly from PYTHIA8, while the W production cross-section at 14 TeV isextrapolated from the measured 13 TeV W -production cross-section [79] using MadGraph [80] toextract the LO dependence on √ s . The new GEANT4 simulation [76] propagates muons through– 19 –hree distinct layers of rock from the IP to the MATHUSLA detector, and includes details of theexperimental cavern UXC 55, the PX56 shaft, and the modular MATHUSLA@CMS detector design.The number of muons observed by MATHUSLA is defined as the number of events with Muonsthat hit the detector’s top seven layers. Assuming an instantaneous luminosity of cm − s − atthe HL-LHC, we obtain a rate of about . × ( . × ) muons per hour from W -production( ¯ bb production with p bT >
40 GeV ). Scaling up the total number slightly to account for small contri-butions from Z -bosons, tau- and charm-decays, we obtain a total rate of approximately N µ ≈ (cid:18) . × hour (cid:19) × (cid:18) L cm − s − (cid:19) (4.6)at MATHUSLA. The total number of muons from HL-LHC collisions passing through MATHUSLA@ CMS, assuming integrated luminosity of 3000fb − , is therefore N µ ≈ × . (4.7)This is about 50 × more than the rate estimated in earlier background estimates for the MATH-USLA200 geometry [2], which has similar solid angle coverage (consistent with the near-identicalacceptance for LLP decays). This can be understood as arising from the smaller amount of rock sepa-rating MATHUSLA @ CMS from the LHC collision compared to MATHUSLA200. It happens to bethe case that a muon must have kinetic energy close to m W / to penetrate about 100m of rock. ForMATHUSLA200, this means that W decay dominates the upward muon rate. Reducing the amount ofmaterial drastically increases the muon contribution from W and Z decay due to the kinematic edgein the distribution of muon momenta. It also allows softer muons from ¯ bb production to contribute incomparable numbers.Upwards LHC muons do not constitute a background to LLP searches. However, the muonbackground studies in the LOI [2] include predictions for rare events arising from the muons flyingthrough the decay volume, like highly subdominant rare muon decay modes. Rescaling those ratesby 50 suggests that MATHUSLA @ CMS might see O (10) µ → eeeνν decays, which would have tobe vetoed using the floor detector or main detector information. Material interactions can be rejectedusing a fiducial veto. Future studies will refine on these estimates and explicitly demonstrate theassociated rejection strategies. Neutrinos can scatter off nuclei in the air-filled MATHUSLA decay volume, giving rise to chargedparticles that go through the tracking layers and mimic an LLP decay inside the detector. This back-ground category can be further divided into neutrinos from the LHC, and those from cosmic rayinteractions in the atmosphere. Previous calculations [1] estimated that the rate of these backgroundprocesses is very low. We have now conducted detailed calculations using GENIE [81] to simulateneutrino interactions confirm this to be the case.Neutrinos were simulated using GENIE v2.12.10 to generate neutrino events in the energy range0-10 GeV incident on an air target (76.8% N, 23.2%O), using GENIE’s pre-calculated DefaultPlus-– 20 –ECWithNC set of interaction cross-sections. Samples were generated for muon and electron neutri-nos and anti-neutrinos, and reweighted according to the total neutrino flux dN ν /dE ν incident on theMATHUSLA detector, as we discuss below.The atmospheric neutrino background was estimated using flux measurements from the Frjusexperiment [82]. Based on the simulations, the expected number of neutrino interactions occurringwithin the volume of MATHUSLA100 over one year is ≈
30, assuming the detector runs for 50%of the year. A fairly modest cut asking for each event to have two tracks with β > < − over the complete HL-LHC run,the total number of neutrino events expected is < β > ≈ MATHUSLA’s large detection area and highly granular tracking system makes it an excellent cosmic-ray (CR) telescope to study extended air showers (EAS) and performing CR measurements up to thePeV scale. Highly detailed measurements of EAS arrival time, particle multiplicity, and spatial dis-tributions allow reconstruction of EAS core and direction and would be greatly aided by installing ahybrid digital-analogue tracking layer of Resistive Plate Chambers (RPC) similar to the ARGO-YBJexperiment. Such a possible upgrade is now being closely investigated, since CR measurements con-stitute a guaranteed physics return on the investment of building MATHUSLA. This will be discussedin an upcoming MATHUSLA publication [83], and is an area where new members with cosmic rayexpertise could make significant contributions.
The MATHUSLA collaboration plans to finalize a Technical Design Report for the design of MATH-USLA, including a robust cost estimate, in 2021. The plan for the following years includes the con-struction of a prototype detector module that can already set relevant limits on certain LLP scenariosand conduct cosmic ray measurements during the extended LHC shutdown to further calibrate back-ground predictions. Subsequent construction and commissioning of the remaining detector modulesshould then allow the full detector to come online in time for first beam at the HL-LHC.– 21 –o achieve these goals, significant R&D is needed to design and optimize the detector hardwareitself, including the scintillators, wavelength shifting fibers, SIPMs, DAQ and trigger systems. De-tailed simulation studies including full material effects are needed to demonstrate LLP decay recon-struction and explicit background rejection in the unique environment of MATHUSLA. Similarly, itis a high priority to conduct careful analyses of rare backgrounds like production of upward travelingmuons, pions and particularly neutral kaons, as well as studying their associated veto strategies andpossible implications for the detector design and geometry. Finally, studies are underway to under-stand the cosmic ray physics case, and the implications of a possible hybrid digital-analog detectionlayer to reconstruct dense air shower cores.The collaboration is accepting new members, and new contributions are particularly welcome inall of the above areas, from detector hardware to simulation and cosmic ray physics.
References [1] J. P. Chou, D. Curtin, and H. J. Lubatti,
New Detectors to Explore the Lifetime Frontier , Phys. Lett.
B767 (2017) 29–36, [ arXiv:1606.06298 ].[2]
MATHUSLA
Collaboration, C. Alpigiani et al.,
A Letter of Intent for MATHUSLA: A DedicatedDisplaced Vertex Detector above ATLAS or CMS. , arXiv:1811.00927 .[3] D. Curtin et al., Long-Lived Particles at the Energy Frontier: The MATHUSLA Physics Case , Rept.Prog. Phys. (2019), no. 11 116201, [ arXiv:1806.07396 ].[4] FASER
Collaboration, A. Ariga et al.,
Letter of Intent for FASER: ForwArd Search ExpeRiment at theLHC , arXiv:1811.10243 .[5] J. Barron and D. Curtin, On the Origin of Long-Lived Particles , arXiv:2007.05538 .[6] M. Alidra et al., The MATHUSLA Test Stand , arXiv:2005.02018 .[7] J. M. No, P. Tunney, and B. Zaldivar, Probing Dark Matter freeze-in with long-lived particle signatures:MATHUSLA, HL-LHC and FCC-hh , JHEP (2020) 022, [ arXiv:1908.11387 ].[8] R. T. D’Agnolo, C. Mondino, J. T. Ruderman, and P.-J. Wang, Exponentially Light Dark Matter fromCoannihilation , JHEP (2018) 079, [ arXiv:1803.02901 ].[9] G. Aielli et al., Expression of Interest for the CODEX-b Detector , arXiv:1911.00481 .[10] A. Berlin and F. Kling, Inelastic Dark Matter at the LHC Lifetime Frontier: ATLAS, CMS, LHCb,CODEX-b, FASER, and MATHUSLA , Phys. Rev. D (2019), no. 1 015021, [ arXiv:1810.01879 ].[11] A. Coccaro, D. Curtin, H. J. Lubatti, H. Russell, and J. Shelton, Data-driven Model-independentSearches for Long-lived Particles at the LHC , Phys. Rev.
D94 (2016), no. 11 113003,[ arXiv:1605.02742 ].[12] M. Bauer, M. Heiles, M. Neubert, and A. Thamm,
Axion-Like Particles at Future Colliders , Eur. Phys. J.C (2019), no. 1 74, [ arXiv:1808.10323 ].[13] A. Ibarra, E. Molinaro, and S. Vogl, Potential for probing three-body decays of Long-Lived Particleswith MATHUSLA , Phys. Lett. B (2019) 127–131, [ arXiv:1809.01683 ].[14] A. E. Nelson, M. Park, and D. G. Walker,
Composite Higgs Models with a Hidden Sector , Phys. Rev. D (2019), no. 7 076015, [ arXiv:1809.09667 ]. – 22 –
15] D. Curtin, K. R. Dienes, and B. Thomas,
Dynamical Dark Matter, MATHUSLA, and the LifetimeFrontier , Phys. Rev. D (2018), no. 11 115005, [ arXiv:1809.11021 ].[16] D. Dercks, J. De Vries, H. K. Dreiner, and Z. S. Wang, R-parity Violation and Light Neutralinos atCODEX-b, FASER, and MATHUSLA , Phys. Rev. D (2019), no. 5 055039, [ arXiv:1810.03617 ].[17] FASER
Collaboration, A. Ariga et al.,
FASER’s physics reach for long-lived particles , Phys. Rev. D (2019), no. 9 095011, [ arXiv:1811.12522 ].[18] S. Demidov, S. Gninenko, and D. Gorbunov, Light hidden photon production in high energy collisions , JHEP (2019) 162, [ arXiv:1812.02719 ].[19] J. Beacham et al., Physics Beyond Colliders at CERN: Beyond the Standard Model Working GroupReport , J. Phys. G (2020), no. 1 010501, [ arXiv:1901.09966 ].[20] Z. S. Wang and K. Wang, Long-lived light neutralinos at future Z − factories , Phys. Rev. D (2020),no. 11 115018, [ arXiv:1904.10661 ].[21] P. deNiverville and H.-S. Lee,
Implications of the dark axion portal for SHiP and FASER and theadvantages of monophoton signals , Phys. Rev. D (2019), no. 5 055017, [ arXiv:1904.13061 ].[22] J. Serra, S. Stelzl, R. Torre, and A. Weiler,
Hypercharged Naturalness , JHEP (2019) 060,[ arXiv:1905.02203 ].[23] A. Das, P. B. Dev, and N. Okada, Long-lived TeV-scale right-handed neutrino production at the LHC ingauged U (1) X model , Phys. Lett. B (2019) 135052, [ arXiv:1906.04132 ].[24] I. Boiarska, K. Bondarenko, A. Boyarsky, M. Ovchynnikov, O. Ruchayskiy, and A. Sokolenko,
Lightscalar production from Higgs bosons and FASER 2 , JHEP (2020) 049, [ arXiv:1908.04635 ].[25] E. J. Chun, A. Das, S. Mandal, M. Mitra, and N. Sinha, Sensitivity of Lepton Number Violating MesonDecays in Different Experiments , Phys. Rev. D (2019), no. 9 095022, [ arXiv:1908.09562 ].[26] A. Krovi, I. Low, and Y. Zhang,
Higgs Portal to Dark QED , arXiv:1909.07987 .[27] M. Bauer, O. Brandt, L. Lee, and C. Ohm, ANUBIS: Proposal to search for long-lived neutral particlesin CERN service shafts , arXiv:1909.13022 .[28] Z. S. Wang and K. Wang, Physics with far detectors at future lepton colliders , Phys. Rev. D (2020),no. 7 075046, [ arXiv:1911.06576 ].[29] S. Jana, N. Okada, and D. Raut,
Displaced Vertex and Disappearing Track Signatures in type-III Seesaw , arXiv:1911.09037 .[30] M. Hirsch and Z. S. Wang, Heavy neutral leptons at ANUBIS , Phys. Rev. D (2020), no. 5 055034,[ arXiv:2001.04750 ].[31] A. Banerjee, H. Kim, O. Matsedonskyi, G. Perez, and M. S. Safronova,
Probing the Relaxed Relaxion atthe Luminosity and Precision Frontiers , JHEP (2020) 153, [ arXiv:2004.02899 ].[32] F. Kling and S. Trojanowski, Looking forward to test the KOTO anomaly with FASER , Phys. Rev. D (2020), no. 1 015032, [ arXiv:2006.10630 ].[33] H. K. Dreiner, J. Y. Gnther, and Z. S. Wang,
R-parity Violation and Light Neutralinos at ANUBIS andMAPP , arXiv:2008.07539 .[34] K. Bondarenko, A. Boyarsky, D. Gorbunov, and O. Ruchayskiy, Phenomenology of GeV-scale HeavyNeutral Leptons , arXiv:1805.08567 . – 23 –
35] D. Curtin et al.,
Exotic decays of the 125 GeV Higgs boson , Phys. Rev.
D90 (2014), no. 7 075004,[ arXiv:1312.4992 ].[36] Z. Chacko, H.-S. Goh, and R. Harnik,
The Twin Higgs: Natural electroweak breaking from mirrorsymmetry , Phys. Rev. Lett. (2006) 231802, [ hep-ph/0506256 ].[37] G. Burdman, Z. Chacko, H.-S. Goh, and R. Harnik, Folded supersymmetry and the LEP paradox , JHEP (2007) 009, [ hep-ph/0609152 ].[38] N. Craig, A. Katz, M. Strassler, and R. Sundrum, Naturalness in the Dark at the LHC , JHEP (2015)105, [ arXiv:1501.05310 ].[39] D. Curtin and C. B. Verhaaren, Discovering Uncolored Naturalness in Exotic Higgs Decays , JHEP (2015) 072, [ arXiv:1506.06141 ].[40] G. Giudice and R. Rattazzi, Theories with gauge mediated supersymmetry breaking , Phys. Rept. (1999) 419–499, [ hep-ph/9801271 ].[41] S. Knapen and D. Redigolo,
Gauge mediation at the LHC: status and prospects , JHEP (2017) 135,[ arXiv:1606.07501 ].[42] J. A. Evans, Detecting Hidden Particles with MATHUSLA , Phys. Rev. D (2018), no. 5 055046,[ arXiv:1708.08503 ].[43] S. Alekhin et al., A facility to Search for Hidden Particles at the CERN SPS: the SHiP physics case , Rept. Prog. Phys. (2016), no. 12 124201, [ arXiv:1504.04855 ].[44] P. Coloma, E. Fernndez-Martnez, M. Gonzlez-Lpez, J. Hernndez-Garca, and Z. Pavlovic, GeV-scaleneutrinos: interactions with mesons and DUNE sensitivity , arXiv:2007.03701 .[45] L. J. Hall, K. Jedamzik, J. March-Russell, and S. M. West, Freeze-In Production of FIMP Dark Matter , JHEP (2010) 080, [ arXiv:0911.1120 ].[46] D. Tucker-Smith and N. Weiner, Inelastic dark matter , Phys. Rev. D (2001) 043502,[ hep-ph/0101138 ].[47] E. Izaguirre, G. Krnjaic, and B. Shuve, Discovering Inelastic Thermal-Relic Dark Matter at Colliders , Phys. Rev. D (2016), no. 6 063523, [ arXiv:1508.03050 ].[48] J. Bramante, B. Broerman, R. F. Lang, and N. Raj, Saturated Overburden Scattering and theMultiscatter Frontier: Discovering Dark Matter at the Planck Mass and Beyond , Phys. Rev. D (2018), no. 8 083516, [ arXiv:1803.08044 ].[49] R. T. Co, F. D’Eramo, L. J. Hall, and D. Pappadopulo, Freeze-In Dark Matter with Displaced Signaturesat Colliders , JCAP (2015) 024, [ arXiv:1506.07532 ].[50] F. D’Eramo, N. Fernandez, and S. Profumo, Dark Matter Freeze-in Production in Fast-ExpandingUniverses , JCAP (2018) 046, [ arXiv:1712.07453 ].[51] J. Alexander et al., Dark Sectors 2016 Workshop: Community Report , 8, 2016. arXiv:1608.08632 .[52] L. Calibbi, L. Lopez-Honorez, S. Lowette, and A. Mariotti,
Singlet-Doublet Dark Matter Freeze-in:LHC displaced signatures versus cosmology , JHEP (2018) 037, [ arXiv:1805.04423 ].[53] L. Calibbi, A. Mariotti, and P. Tziveloglou, Singlet-Doublet Model: Dark matter searches and LHCconstraints , JHEP (2015) 116, [ arXiv:1505.03867 ]. – 24 –
54] L. Covi, J. E. Kim, and L. Roszkowski,
Axinos as cold dark matter , Phys. Rev. Lett. (1999)4180–4183, [ hep-ph/9905212 ].[55] L. Covi, H.-B. Kim, J. E. Kim, and L. Roszkowski, Axinos as dark matter , JHEP (2001) 033,[ hep-ph/0101009 ].[56] R. T. Co, F. D’Eramo, and L. J. Hall, Gravitino or Axino Dark Matter with Reheat Temperature as highas GeV , JHEP (2017) 005, [ arXiv:1611.05028 ].[57] D. E. Kaplan, M. A. Luty, and K. M. Zurek, Asymmetric Dark Matter , Phys. Rev. D (2009) 115016,[ arXiv:0901.4117 ].[58] T. Asaka and M. Shaposhnikov, The nuMSM, dark matter and baryon asymmetry of the universe , Phys.Lett.
B620 (2005) 17–26, [ hep-ph/0505013 ].[59] T. Asaka, S. Blanchet, and M. Shaposhnikov,
The nuMSM, dark matter and neutrino masses , Phys. Lett.
B631 (2005) 151–156, [ hep-ph/0503065 ].[60] X.-D. Shi and G. M. Fuller,
A New dark matter candidate: Nonthermal sterile neutrinos , Phys. Rev. Lett. (1999) 2832–2835, [ astro-ph/9810076 ].[61] M. Laine and M. Shaposhnikov, Sterile neutrino dark matter as a consequence of nuMSM-inducedlepton asymmetry , JCAP (2008) 031, [ arXiv:0804.4543 ].[62] L. Canetti, M. Drewes, and M. Shaposhnikov,
Sterile Neutrinos as the Origin of Dark and BaryonicMatter , Phys. Rev. Lett. (2013), no. 6 061801, [ arXiv:1204.3902 ].[63] L. Canetti, M. Drewes, T. Frossard, and M. Shaposhnikov,
Dark Matter, Baryogenesis and NeutrinoOscillations from Right Handed Neutrinos , Phys. Rev.
D87 (2013) 093006, [ arXiv:1208.4607 ].[64] R. T. D’Agnolo, D. Pappadopulo, and J. T. Ruderman,
Fourth Exception in the Calculation of RelicAbundances , Phys. Rev. Lett. (2017), no. 6 061102, [ arXiv:1705.08450 ].[65] L. Li and Y. Tsai,
Detector-size Upper Bounds on Dark Hadron Lifetime from Cosmology , JHEP (2019) 072, [ arXiv:1901.09936 ].[66] Y. Hochberg, E. Kuflik, T. Volansky, and J. G. Wacker, Mechanism for Thermal Relic Dark Matter ofStrongly Interacting Massive Particles , Phys. Rev. Lett. (2014) 171301, [ arXiv:1402.5143 ].[67] E. Kuflik, M. Perelstein, N. R.-L. Lorier, and Y.-D. Tsai,
Elastically Decoupling Dark Matter , Phys. Rev.Lett. (2016), no. 22 221302, [ arXiv:1512.04545 ].[68] A. Dery, J. A. Dror, L. Stephenson Haskins, Y. Hochberg, and E. Kuflik,
Dark Matter in VerySupersymmetric Dark Sectors , Phys. Rev. D (2019), no. 9 095023, [ arXiv:1901.02018 ].[69] K. R. Dienes and B. Thomas, Dynamical Dark Matter: I. Theoretical Overview , Phys. Rev. D (2012)083523, [ arXiv:1106.4546 ].[70] K. R. Dienes and B. Thomas, Dynamical Dark Matter: II. An Explicit Model , Phys. Rev. D (2012)083524, [ arXiv:1107.0721 ].[71] D. Curtin and M. E. Peskin, Analysis of Long Lived Particle Decays with the MATHUSLA Detector , Phys. Rev.
D97 (2018), no. 1 015006, [ arXiv:1705.06327 ].[72] J. Alimena et al.,
Searching for Long-Lived Particles beyond the Standard Model at the Large HadronCollider , arXiv:1903.04497 . – 25 –
73] T. Sato,
Analytical Model for Estimating Terrestrial Cosmic Ray Fluxes Nearly Anytime and Anywherein the World: Extension of PARMA/EXPACS , PLOS ONE (Dec, 2015) 1–33.[74] T. Sato, Analytical Model for Estimating the Zenith Angle Dependence of Terrestrial Cosmic Ray Fluxes , PLOS ONE (Aug, 2016) 1–22.[75] “EXPACS Homepage in English.” https://phits.jaea.go.jp/expacs/ , May, 2019.[76] GEANT4
Collaboration, S. Agostinelli et al.,
GEANT4: A Simulation toolkit , Nucl. Instrum. Meth.
A506 (2003) 250–303.[77]
Particle Data Group
Collaboration, M. Tanabashi et al.,
Review of Particle Physics , Phys. Rev. D (2018), no. 3 030001.[78] T. Sjostrand, S. Mrenna, and P. Z. Skands, A Brief Introduction to PYTHIA 8.1 , Comput. Phys. Commun. (2008) 852–867, [ arXiv:0710.3820 ].[79]
ATLAS
Collaboration, G. Aad et al.,
Measurement of W ± and Z -boson production cross sections in pp collisions at √ s = 13 TeV with the ATLAS detector , Phys. Lett. B (2016) 601–621,[ arXiv:1603.09222 ].[80] J. Alwall, M. Herquet, F. Maltoni, O. Mattelaer, and T. Stelzer,
MadGraph 5 : Going Beyond , JHEP (2011) 128, [ arXiv:1106.0522 ].[81] C. Andreopoulos et al., The GENIE Neutrino Monte Carlo Generator , Nucl. Instrum. Meth.
A614 (2010) 87–104, [ arXiv:0905.2517 ].[82] K. Daum, W. Rhode, P. Bareyre, et al.,
Determination of the atmospheric neutrino spectra with the Frjusdetector , Z. Phys. C - Particles and Fields (1995) 417428.[83] MATHUSLA cosmic ray whitepaper , In preparation ..