Heavy Stable Charged Particles search by Novel ...

4 downloads 607 Views 278KB Size Report
Patterns are software generated on the basis of the simulation results. .... The CMS Trigger Supervisor: Control and Hardware Monitoring System of the CMS ...
Heavy Stable Charged Particles search by Novel Pattern Comparator Procesor A.Zagoździńska, K.T.Poźniak, R.Romaniuk ABSTRACT This paper describes the idea of the particle recognition at the Pattern Comparator Boards at the RPC detector at the CMS experiment at the LHC at CERN. This solution enables the muons and the Heavy Stable Charged Particles recognition and distinguishing. The described algorithms are implemented in the FPGA structures. They are able to realise the fast analysis of the data from the whole system and work on the reliable triggering decision with the same delay. The implementation is compared with the other solutions. The further development of the RPC system for the HSCPs search is described. Those solutions should provide the reliable data about the hipotetical existence of the HSCPs. Key words: LHC FPGA CMS Resistive Plate Chamber System, RPC Heavy Stable Charged Particles HSCP pattern comparator algorithms, PAC muon

1. INTRODUCTION Most of the High Energy Physics (HEP) experiments are focused on discovering particles and dependencies described by the Standard Model Theory (SM) [1]. Most of electroweak and electrostrong iteractions have been experimentally confirmed. One of the last missing elements of the SM is Higgs Particle [2]. Its existence can be confirmed or denied with a high precision at Large Hadron Collider (LHC) at CERN. There are also theories that enhance the SM: technicolors models, supersymmetry (SUSY) [3] theories and additional dimensions. A further exploration is to confirm fenomenas described by those theories and fill the SM drawbacks. The SUSY theory predict Heavy pseudo-Stable Charged Particles (HSCP) existence. A HCSPs mass is over 200GeV/c2 and occur in three configurations: – lepton similar – can be recognised by a detector as muons, their lifetime is long enough to leave the detector and hit surrounding rocks, – R-hadrons – contain hadron particle but interact like quarks and gluons, – particles charged more than 1 or less than -1. Search of such particles is possible with the use of particle accelerators, detectors and computer network. Detetecting systems are multilevel and have a different precision. For example the CMS detector at the LHC experiment contains such layers: Silicon Tracker, Electromagnetic and Hadron Calorimeter and Muon Chambers. The Tracker and Calorimeters are surrounded by Superconducting Solenoid. There is also a return yoke between layers of the Muon Chambers. Most of particles produced in the collision such as hadronsor electrons can not achieve the outside Muon Chamber layers. Only muons and HSCPs are able to exceed the magnetic field of the Superconducting Solenoid and reach Muon Chambers. Some subsystems at the CMS experiment at the LHC are developed to recognize HSCPs. The RPC system in which the Warsaw CMS Group is involved is working on search of a stau (τ) particle. The stau is a supersymetric partner of a τ lepton. It is also called NLSP (Next to Lightest Symetric Particle). It can be registered as the typical muon so there are special algorithms required to distinguish between those two particles. For the technical point of view the most important feature of stau is its low speed. It is not able to leave the detector during the one bunch crossing (BX) time. For this reason one trajectory of the hypothetical stau can be registered as a muon more than once in a few BXs. Research of such low scale fundamental forces and particles requires a very high energy of particles collisions. For that reason accelerators are built. There biggest accelerator complex over the world is situated at CERN in Switzerland. There are 8 accelerators: AD, CLIC, CNGS, ISOLDE, nTOF, PS, SPS and LHC. The Large Hadron Collider (LHC) enables the particles collisions with the energy of 7TeV. It is closer than 10-10s to the beginning of the Universe. The aim is to reach the energy of 14TeV in this accelerator. Particles are formed in bunches that are collided every 25ns. Results of the collisions are constantly registered by the detectors: ALICE, ATLAS, CMS, LHCb, TOTEM, LHCf. The detectors

consist on many small subsystems that's cooperation results with the set of data interesting from the physical point of view. One of such subsystem at the CMS experiment is RPC (Resistive Plate Chamber) system.

2. RPC SYSTEM OVERVIEW The RPC system consist on the RPC detector and electronic system. This system implements on the first level triggering algorithms. All the incoming data from the detector is analysed to reject the unnecessary information and register only the most interesting data for the further processing. The RPC system is divided into the functional blocks that consist on dedicated electronics. The detector chambers are block filled with the gas. The flying particle ionise the gas. Electrons are multiplied that cause the pulse generation. The pulse propagates through the wire and is registered by the Frond-End Boards (FEB). The analog signal is converted into the digital and transmitted further. This part of electronics works in the high magnetic field in the detector cavern. For this reason it must be resistant to the high radiation level. The data is transmitted to the counting room to be analysed for the presence of the muon tracks. The general data flow from the FEBs to the last functional block of the RPC system is shown on the fig.1. The FEBs convert the analog data to digital and send it to the Link Boards (LB). The data frame is marked with the information about its source. The 96 bit data frame is compressed into 24 bits and serialised. Than it is transmitted through the optical links to the Trigger Boards (TB). The TBs receive and deserialise the data. This idea is shown on the fig. 2. Each TB consist on 12 Pattern Comparator boards (PAC) and FPGAs that realise the Ghost Busting algorithms. The PAC structure and functionality is described in the next chapter. The Ghost Busting is rejection process of the artificially generated double information about the same particle. The information about the muons candidates is transmitted from the TBs to the Sorter Boards. The effect of sorting algorithms is identification of the four highest quality muon candidates. The similar information is generated by the other CMS subsystems. The further analysis and triggering is realised by the High Level Trigger (HLT). The HLT algorithms are implemented in computer farms and the processing is offline. The data that were not rejected by triggers are saved to the hard drives and tapes and propagated through the LHC Computing Grid.

Figure 1: The direction of the data flow through the RPC system

Figure 2: The data compression and transmission between the Link Boards and Trigger Boards

3. PATTERN COMPARATOR ALGORITHMS The pattern recognition is realised by the Pattern Comparator (PAC) boards, that are part of the Trigger Boards (TB). The TB boards receive compressed data through the optical links. The input signal is multiplied by a splitter and sent to all the TB. On the board signal is decompressed and appropriate divided between PAC boards. The single TB contains 12 PACs and analyse the data from 30º of the one detector tower. This area is logically divided into smaller parts – cones. Each cone consist on 6 detector layers with multiple strips. The narrowest part of the cone contains 8 strips and is called “reference layer”. PAC must find the most energetic track of the muon that crossed the reference layer. After sorting and ghostbusting the best 4 cases are sent to the further system parts. A pattern comparator algorithm is implemented in VHDL with the use of parameters so that it can be used in the different FPGA chips. The current version of PAC boards consist on Altera's Stratix III chips. 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. Patterns are software generated on the basis of the simulation results. Single pattern must be included into the one detector cone. The single pattern with a fired strip at the reference layer is shown on the fig 3a. The worse quality pattern without a fired strip at the reference layer is shown on the fig 3b. Main parameters are pT code and quality. The quality describes the number of fired strips of each layer in the following cone. There is also a time parameter that is described in a next chapter. 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, - 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. Generated patterns are converted into the VHDL code and compiled with a whole firmware for PAC boards. Each PAC gets the separate patterns to analyse the incoming data.

a)

b)

Figure 3: The logic cones a) with a fired strip at the reference layer and b) without a fired strip at the reference layer

4. SEARCH OF THE SLOW PARTICLES The slow particles such as the hypothetical HSCPs can fire strips in more than 1BX time. For this reason the patterns are enhanced with a time parameter for each layer. The patterns with the time information are shown on fig 4. In first case the whole pattern is active during the time defined in BXs. The trigger decision latency grows of additional BXs. The solution allows both muons and HSCPs registration. Distinguish between them is not possible in the real time. On basis of this solution exclusive trigger can be implemented fig 5. The 2BXs pattern and two 1BX patterns are active. If the particle fires the first 3 strips during the first BX and the rest of them in the second BX, the 2BXs pattern will cause the trigger. None of the single BXs can not cause the trigger at this time. If one of the single patterns fires the trigger, the 2BXs pattern will be negated. This solution enables distinguishing between muons and HSCPs but requires the triple comparison. There is not enough resources in the current chips to implement such solution.

Figure 4: Time flexed patterns for 1BX, 2BXs and 3BXs, this is also an additional latency value for the trigger decision

Figure 5: The triple pattern comparison with the distinguishing between muons and HSCPs

The last configuration is shown on fig 6. In this solution each layer is defined with the separate time parameters. For example layers number 0, 1, 2 are active in the BX0, 3rd layer is active in BX0 and BX1, and 4th, 5th layers are active in BX1 and BX2. The time parameter equal 0 means the pattern for the muon particle. This solution is currently tested for the PAC boards at the RPC system.

Figure 6: The pattern with a time distinguishing between layers

5. CONCLUSIONS Modification of the RPC system at the CMS experiment is independent to reach better results for search of the new physics described by the SM theory and enhancing theories. The newly implemented solution enables clear distinguish between the muons and HCSPs. The further development of the system is aimed to reach the greater time resolution and particle flight direction recognition. The better time resolution should allows to recognise the particle parameters with a greater precision. The data size will be predicted on basis of the simulation and the results of current implementation. It will allow to develop the better compression and decompression algorithms. The flight direction recognition will be interesting area of research. The interesting question is how to distinguish between the cosmic muons and the “boomerang” particle that hit rocks surrounding the detector and fly back into the detector area. The RPC system development sets the border for the physical research. It is important to prepare efficient tools for further search before the Higgs particle discovery or rejection.

REFERENCES [1] Chris Quigg, The Standard Model (Electroweak Theory), European School of High-Energy Physics, 2002 [2] Martin Schmaltz, Physics beyond the standard model (Theory): Introducing the Little Higgs, Nuclear Physics B - Proceedings Supplements Vol.117, 2003 [3] P. Binetruy, Supersymmetry: Theory, Experiment, and Cosmology, Oxford University Press, ISBN: 978-0199652730 , 2006 [4] Website http://lhc.web.cern.ch/lhc/ [5] Website http://atlas.web.cern.ch/Atlas/Collaboration/ [6] Website http://cms.web.cern.ch/ [7] Website http://lhcb.web.cern.ch/lhcb/ [8] The LHCb Collaboration, The LHCb Detector at the LHC, J. Instrum. 3, 2008 [9] D.Froidevaux, P.Sphicas, General-purpose detectors for the Large Hadron Collider, Annu. Rev. Nucl. Part. Sci. 56:375–440, 2006 [10] [11] The CMS Collaboration, The data-acquisition system of the CMS experiment at the LHC, Journal of Physics: Conference Series 331 (2011) 022021, 2010 [12] H.Spieler, Front-End Electronics and Signal Processing, Instrumentation in Elementary Particle Physics, AIP Conf. Proc. 674, 2002 [13] Website http://aliceinfo.cern.ch/Public/en/Chapter2/Chap2_TOF.html [14] J. Christiansen, HPTDC High Performance Time to Digital Converter, HPTDC v. 1.3, 2004 [15] Website http://hepwww.rl.ac.uk/Vertex03/Talks/JornGrosseKnetter.pdf [16] Website http://project-slhc.web.cern.ch/project-slhc/about/ [17] Website http://clic-study.org/ [18] Website http://public.web.cern.ch/public/en/Research/LEPExp-en.html [19] Website http://www-bdnew.fnal.gov/tevatron/ [20] The CMS Collaboration, CMS The TriDAS ProjectTechnical Design Report, Vol.1:The Trigger Systems,

CERN/LHCC 2000 – 38 CMS TDR 6.1, 2000 [21] Website http://www-cdf.fnal.gov/ [22] The CDF Collaboration, Using Feedback to Control Deadtime in the CDF Trigger System, 2010 [23] T. Yoshioka et al., Upgrade of the Level-0 Trigger System for BNL-E949, IEEE Transactions on Nuclear Science, vol. 51, issue 3, pp. 334-339, 2004 [24] I.Magrans de Abril, The CMS Trigger Supervisor: Control and Hardware Monitoring System of the CMS Level1 Trigger at CERN, 2008 [25] M.Mozer, Triggers for New Physics at the LHC, J. Phys.: Conf. Ser. 171 012101, 2009 [26] R.Achenbach et al., The ATLAS Level-1 Calorimeter Trigger, JINST 3 P03001, 2008 [27] The ATLAS Collaboration, Tag and Probe method for Electron ID efciencies, 2009 [28] M.Konecki, Performance of the RPC trigger, CMS Warsaw Group, 2012 [29] N.Neufeld, Status and Prospects of LHC Experiments Data Acquisition, CERN/PH, 2009 [30] J.R. Andersen1 et al., Discovering Technicolor, Eur. Phys. J. Plus (2011) 126: 81, 2011 [31] The CMS Collaboration, CMS The TriDAS Project Technical Design Report, Volume 2: Data Acquisition and High-Level Trigger, CERN/LHCC 02-26 CMS TDR 6, 2002