Heavy Stable Charged Particles search by RPC system at CMS
detector at LHC accelerator at CERN
A.Zagoździńskaa , K.T.Poźniaka , R.Romaniuka , P.Zalewskib
a
Dept. of Electronics and Information Technology, Warsaw University of Technology, Poland;
b
National Centre for Nuclear Research, Poland;
ABSTRACT
A paper describes a possibility of upgrading trigger RPC system at CMS detector for research of Heavy Stable Charged
Particles (HSCPs). An experiment for the High Energy Physics (HEP) is located at CERN at the world's biggest
accelerator complex mostly aimed to the research of particles and forces described in Standard Model (SM) theory. The
SM explains the structure of matter but omits some important factors as the Gravity. There are three main branches of
the New Physics that expand the SM: supersymmetry (SUSY), technicolors and additional dimentions. The SUSY
provides the hipotetic HSCPs existence. Algorithms for search of those particles are implemented as a FPGA firmware
in the RPC trigger system. Proposed methods expand the functionality of the existing trigger firmware. They are
supposed to implement the simultaneous HSCPs and muons identification. The proposed solutions increase the trigger
time resolution that is the important element of the system competitiveness to the other systems of the CMS detector.
They also enable distinction between the particles produced inside the detector as the result of the collisions and the
particles generated outside the detector that fly back through it. Such generation can be the result of the HSCP the
reaction with rocks surrounding the detector cavern. The described methods are supposed to provide the selective trigger
for the data most wanted for the HEP experiments research that could introduce the experimental science to the Great
Unification.
Keywords: FPGA, High Energy Physics, Heavy Stable Charged Particles, trigger, Electronics, CMS,
1. INTRODUCTION
A latest research in the High Energy Physics (HEP) field was mainly focused on the search of Higgs particle. Scientists
of Compact Muon Solenoid (CMS) and A Toroidal LHC ApparatuS (ATLAS) experiments at the Large Hadron Collider
(LHC) presented their preliminary results at the particle physics conference ICHEP2012 in Melbourne. The discovery of
Higgs particle have been introduced. A new particle was observed in the mass region 125-126 GeV in both experiments.
This fully confirms the Standard Model (SM)[1] theory and gives a field to further considerations. The SM describes the
electromagnetic, strong and weak forces. All predicted particles have been discovered. However, the SM description
excludes the Gravity. There are three main branches of the theoretical New Physics that try to fulfill the gap[2]:
supersymmetry (SUSY)[3], technicolors and additional dimensions. The SUSY assume the existence of symmetric
particles which spin values differs of ½ to the basic particle. The technicolor models describe the Higgs boson as a
particle with its internal structure that is visible at the energy levels over 1TeV. The problem is that scientists do not
know how to model it. The last theory contains additional dimensions to explain phenomena that are over the SM. All
those theories try to bring us closer to the Grand Unification Theory(GUT). The GUT is supposed to merge all those
considerations into the one experimentally verified model. This paper describes a search of the particles described in
various extensions of the SM theory at the CMS experiment at CERN. Especially the possibility of the Heavy Stable
Charged Particles (HCPCs) in the Resistive Plate Chamber (RPC) trigger system is presented.
2. RESEARCH OF THE HSCP AT THE CMS DETECTOR
Accelerator complexes are built for the research and search of fundamental particles, hypothetical particles and
phenomena. The particles or ions are accelerated in linear or circular accelerators. They are formed into bunches and
collided with a shield or with another bunches. Results of the collisions are registered by detectors. There are also nonPhotonics Applications in Astronomy, Communications, Industry, and High-Energy Physics
Experiments 2013, edited by Ryszard S. Romaniuk, Proc. of SPIE Vol. 8903, 89031K
© 2013 SPIE · CCC code: 0277-786X/13/$18 · doi: 10.1117/12.2032247
Proc. of SPIE Vol. 8903 89031K-1
accelerator experiments where detectors collect cosmic muons (Auger Project, DAMA) but they are not in the area of
interest of this paper. The world's biggest accelerator complex is located at CERN in Switzerland. There are 8 main
accelerators AD, CLIC, CNGS, ISOLDE, nTOF, PS, SPS and LHC[5][6]. Some of them are supposed only to accelerate
the particles and inject them into the other accelerators. This is how the LHC accelerator works. The particles are
accelerated in LINAC, BOOSTER, PS, SPS and finally injected in two points of the LHC. Two counter beams are
accelerated by the magnetic field of 9300 magnets at the way of 27km. There are protons or heavy ions runs. The current
collision energy is 3.5TeV per beam. The aim is to reach 7TeV per beam (14TeV of the total energy). There are 4 bunch
crossing points where the main experiments are located: ALICE[7], ATLAS[8], CMS[9] and LHCb[10]. There are also
smaller LHC experiments that investigate a small angles of the collisions: TOTEM and LHCf.
The result of the collision is generation of the new particles. Some of the particles e.g. pions have the short life time.
They are not able to reach outside layers of the detector or leave it. The muons and some hypothetical particles have the
long life time and can be detected by the external muon detectors. The Heavy pseudo-Stable Charged Particles (HSCP)
are provided by some supersymmetric models that expand the SM. These are leptoquarks which carry both baryon
number and lepton number and also supersymmetric couple to quarks and leptons. They are color triplets. They are
charged with the fractional electric charge. Their mass could be over 150GeV/c2 . The slowest HSCP live long enough
to stop in the detector layers or to leave the detector (75ns to 106ns are searched). They can also hit the rocks
surrounding the detector and fly back through the detector as the “boomerang” particle. Gluinos with mass < 370GeV/c2
for lifetimes 10us-1000s have been already excluded. Some subsystems at the CMS experiment are supposed to search
the stau (τ) particle. This is the supersymmetric partner of the tau (τ) lepton.
A detailed description of the CMS detector can be found elsewhere[12]. The DT, CSC and RPC muon detectors are
installed in the outside detector layers in a barrel and endcaps. Each covers the different region. The CMS trigger
consists of a two-stage system. The first level (L1) of the CMS trigger system is composed of custom hardware
processors. It uses the data from the calorimeters and muon detectors to select a subset of the events. The data rejection
result is compared between the subsystems in the Global Trigger. The final data triggering is realized by the High Level
Trigger (HLT). The HLT processor farm decreases the event rate from ~100 kHz to ~300 Hz, before data storage. The
data storage[13] is realized with the parallel data acquisition system. The processing and acquisition is synchronized to
the global clock.
3. STRUCTURE OF THE RPC SYSTEM
The RPC system is one of the first level trigger subsystems at the CMS detector. It covers barrel and endcap regions of
the detector[14]. It is based on signals coming from Resistive Plate Chambers. It provides a fast recognition of a muon
with specified transverse momentum (pT). Algorithms of searching and identification of the muons are mostly
implemented in FPGA structures. The result of the processing is selection of 8 best muon candidates: 4 from the barrel
region and 4 from endcaps. The pT value and the track quality are two main criteria of the choice. Resistive Plate
Chamber consists of two parallel plates made out of Bakelite with a bulk resistivity of 1010- 1011 Ωcm, forming a gas
gap of a few millimeters. The particle crossing the chamber induces a charge that is detected by Front End Board
electronic system[15]. It is situated in a magnetic field up to 3.5T. Signal from the detector is transmitted to Front-End
Boards (FEB). The FEBs are installed in the detector cavern. The system structure is shown on figure 1.
Fig 1: Data flow in the RPC system
Proc. of SPIE Vol. 8903 89031K-2
The signal processing is performed by the Front End Chips (FEC). These are custom ASIC chips designed to amplify,
discriminate, and do analog to digital conversion of the signal. Each FEC gets an information from 8 detector strips.
After a conversion into LVDS standard, pulses are transmitted to the link system. The system contains 1232 Link Boards
(LB) and Control Boards (CB). All of them placed in racks in the detector cavern. LB and CB process the RPC data,
compress [16] and transmit it through optical links to the Trigger Crate. CB also handle the communication to the control
software and to LB. They also provide automatic loading of the firmware to LB and control of FEB. Every two LB are
slave to the CB. LB contains two FPGA structures (Xilinx Spartan-3 XC3S1000FG456). One of them process the data
received from 96 detector strips. The second FPGA provides the compression and transmission of the data to CBs. The
CB master boards have the similar structure enhanced with GOL (Gigabit OpticaL) transmitter chips. The data are sent
through the optical links to Trigger Boards (TB) where the Pattern Recognition algorithm is realized. Signals from the
optical links are splitted and deserialized. The data are transmitted to Pattern Comparator (PAC) Mezzazine boards on
TB. There are 4 PAC on each TB, each contains single FPGA chip Altera Stratix EP2S60 (barrel data analysis) or Altera
Stratix EP2S30 (endcap data analysis). The data are demultiplexed and decoded. For each optical link three 96 bit
vectors are created. Those bits correspond to detector strips in a single bunch crossing (BX). They are the input data to
Pattern Comparator algorithm. The algorithm is described in the next chapter. The output of the algorithm are muon
candidates with a specified pT, quality and BX number. Results are transmitted in a LVDS standard to the Sorter Crate
(SC). The crate contains two Half Sorter Boards and the Final Sorter Board. The final result is 8 muon candidates: 4 for
the barrel and 4 for endcaps. It is transmitted with the 40MHz speed (LVDS standard) to the Global Muon Trigger
(GMT) system. There results are compared with the other subsystem response to elaborate a final decision. The RPC
trigger system contains also data acquisition part (DAQ), expanded diagnostics for firmware and some other
functionality to reject false muon candidates. That want be described in this paper.
An interesting part of the system is the Component Internal Interface (CII). The CII java software generates the mirror of
the registers and memory elements implemented in firmware. The communication between the corresponding
components is provided. The user has an access to the internal registers through the website. The CII implementations
are described elsewhere [17].
4. THE RPC SYSTEM IN A SEARCH OF MUONS AND HSCP
The RPC chambers are supposed to register muon candidates. Indeed they are sensitive to each charged particle that
crosses the detecting layers. As the HSCP particles are highly ionizing particle, they can also be detected. The time of
flight (ToF) and a energy deposit are two features that enables distinguish between those particles. The HSCP produced
in one BX can moves through the detector during the next few BXs. If such particle moves with velocity β>0.8 it can be
misidentified as a muon. It can be registered in a time window that is defined for each detectors layer. Although the
analysis presented in [18] shows that the large fraction of particles registered at the RPC has β<0.8. Such particles can be
recognized during the offline analysis on basis of merged results of the tracker and muon (CSC, DT) subsystems. As
shown on figure 2, there is no possibility to register HSCP of a particular velocity in one time gap (between two
horizontal lines) for all layers. This Monte Carlo simulation consider the region for eta = 0.46. For this reason there was
a need of a new RPC PAC algorithm implementation.
Fig 2: HSCPs time of flight versus inversed velocity for 6 layers of the
detector in Barrel Region eta=0.46
Proc. of SPIE Vol. 8903 89031K-3
Most of the muons hit with the small angle into the detectors endcaps, only the small part of them hit into the barrel
detector. If HSCP exist, the probability of the hit is much higher in the endcap region. As the particle cross the detector
layers it can fire strips of the detector layers. It can be 1 to 4 or 6 layers dependably on the detector region. The number
of fired layers determines a muon candidate quality. In this way the particle track is sampled in the time window. The
sample is marked with the information about the appropriate BX number during the processing. Figure 3 shows a
simulation of the basic PAC algorithm implemented in FPGA. Input data was extracted out of the data set registered in
RPC detector. Layer 0 to 5 correspond with the detector layers. Width of each layer is defined as a number of strips in a
cone. Fired strips are ones, the rest are zeros. EpacsData array contains patterns definition. The result is code (1F), sign
(0) and quality (1).
Fig 3: Results given by PAC algorithm for a data triggered in the detector
Patterns for a comparison are software generated by the algorithm implemented in C++ on the basis of the simulation
results. It generates file with pattern arrays. It is compiled with PAC firmware. The firmware is implemented in VHDL
as a universal parametrized block. It can be used in the different FPGA chips. Each PAC gets separate patterns to
analyze the incoming data. The basic algorithm of the particle track recognition is based on a path and mask comparison.
A particle must hit at least three detector layers to be registered. The result of the comparison is quality and pT code
marking of each registered particle. This enables fast and reliable exclusion of the less interesting data. The quality is
dependable on a number of fired layers. Third layer is called reference layer. A single pattern with a fired strip at the
reference layer is shown on figure 4a. The worse quality pattern without a fired strip at the reference layer is shown on
figure 4b.
Fig 4: The logic cones a) with a fired strip at the reference layer and b) without a fired strip at the reference layer
Main parameters are pT code and quality. The quality describes the number of fired strips of each layer in the following
cone. The usage of a separate pattern for each particle is impossible because of limited logic resources. For this reason
patterns are grouped into energetic patterns. A few patterns are merged or included in one in the following steps:
– the pattern with active strip of a reference layer is the strongest one, so the cone containing such pattern is
chosen,
Proc. of SPIE Vol. 8903 89031K-4
–
cones are shifted so that the first strip of the reference layer was “0” strip, multiplied cones are rejected or the
list of cones is updated with the new one,
– “pattern inclusion” - for the highest quality patterns, the similar patterns with the lower number of active strips
are rejected,
– “pattern merge” - for the lowest pT, patterns are merged. There are two requirements: the same sign and the
same active strip at the reference layer.
The latest version of PAC algorithm was enhanced with a time parameter. The example of the time distinguished pattern
is shown on figure 5. Each layer of is defined with the use of separate time parameters. For example layers number 0, 1
are active in the BX0, 2,3 are active in BX1 and 4th, 5th layers are active in BX3. The time parameter equal 0 means the
flat pattern that is fired by a muon particle. Results produced by the upgraded firmware will be available after the
technical stop of LHC.
Fig 5: Simulation of a late particle hitting into the detector's layers
5. CONCLUSIONS
The modified method of charged particles detection implemented in the RPC system enables the reliable detection and
determination of the flight direction of the HSCPs. The distinguish between the muon and the stau is possible in the
existing hardware. The search can be mostly realized with the use of the existing tools. The usage of the reprogrammable
chips enables modification of the firmware to follow the growing luminosity and changes of the HEP requirements. The
flexibility of the system firmware gives the possibility of working in the different conditions. This feature is one of the
most important for the system competitiveness to the other detectors. The RPC system is actually able to recognize the
HSCPs. The influence on the basic muons recognition efficiency is examined by the experimental physicists from the
Warsaw CMS Group. The results will be submitted in the next years.
To follow the requirements the future upgrade of the system is planned. The search of the “boomerang particles” or
distinguish between “fast” HSCP and muon will require the additional changes in the system as strips division and
improvement of a time resolution. Quart higher time resolution is planed. Additional markers for the particle's flight
direction will be presumably added. The data size will be greater. Then the revised compression algorithm will be crucial
not to cause the dead time. The basis for the algorithm implementation is the simulation and current results analysis. The
result of simulation for muons is shown elsewhere [19]. Most of the muons hit with the small angle into the detectors
endcaps, and only the small part of them hit into the barrel detector. If the HSCPs exists the probability of their hit into
the endcaps is much higher. That will also affect on the data compression algorithm.
This work was supported in part by Polish National Science Centre grant N N202 167440.
Proc. of SPIE Vol. 8903 89031K-5
REFERENCES
[1] Website, https://twiki.cern.ch/twiki/bin/view/CMSPublic/PhysicsResultsEXO
[2] Quigg, C. “The Standard Model (Electroweak Theory)”, European School of High-Energy
Physics, 2002
[3] Schmaltz, M.,“Physics beyond the standard model (Theory): Introducing the Little Higgs”, Nuclear Physics B Proceedings Supplements Vol.117, 2003
[4] Binetruy, P., “Supersymmetry: Theory, Experiment, and Cosmology”, Oxford University Press,
ISBN: 978-0199652730 , 2006
[5] Froidevaux, D., Sphicas, P.,“General-purpose detectors for the Large Hadron Collider”, Annu. Rev. Nucl. Part.
Sci. 56:375–440, 2006
[6] Website, http://lhc.web.cern.ch/lhc/
[7] Website, http://aliweb.cern.ch/
[8] Website, http://atlas.web.cern.ch/Atlas/Collaboration/
[9] Website, http://cms.web.cern.ch/
[10] Website, http://lhcb.web.cern.ch/lhcb/
[11] The LHCb Collaboration, “The LHCb Detector at the LHC”, J. Instrum. 3, 2008
[12] The CMS Collaboration, “The CMS experiment at the CERN LHC”, JINST 03 (2008) S08004
[13] The CMS Collaboration, “The data-acquisition system of the CMS experiment at the LHC”, Journal of
Physics: Conference Series 331 (2011) 022021, 2010
[14] Bunkowski, K., Krolikowski, J.,“Optimization, Synchronization, Calibration and Diagnostic of the RPC PAC
Muon Trigger System for the CMS detector”, CERN-THESIS-2009-160 ; CMS-TS-2010-009
[15] Spieler, H., “Front-End Electronics and Signal Processing”, Instrumentation in Elementary Particle Physics,
AIP Conf. Proc. 674, 2002
[16] Gorski, M., Kudla, I.M., Pozniak, K.T., “Resistive Plate Chamber (RPC) based muon trigger system for the
CMS experiment data compression/decompression system”, Nuclear Instruments and Methods in Physics
Research A 419 (1998) 701Ð706
[17] Pozniak, K.T, “Internal Interface I/O communication with FPGA circuits and hardware description standard
for applications in HEP and FEL electronics ver. 1.0”, TESLA Report 2005/22
[18] Montoya, C. et al., “Search for Heavy Stable Charged Particles in the CMS Experiment using the RPC
Detectors”, CERN-THESIS-2011-032 ; CMS-TS-2011-018
[19] Website, https://twiki.cern.ch/twiki/bin/view/CMSPublic/
Proc. of SPIE Vol. 8903 89031K-6