The Compact Muon Solenoid (CMS) on the Large Hadron Collider
\mbox(LHC) at CERN is a large, general--purpose detector
for the measure-ment of particles resulting from 14~TeV
proton--proton (pp) collisions. The hadronic calorimeter
(HCAL), now under construction, has been developed for the
containment and measurement of jets with particular emphasis
on sensitivity for the determination of missing transverse
energy
(\mbox\hboxE\kern-0.6em\lower-.1ex\hbox/_T).
HCAL is configured as a central barrel and two end-cap
calorimeters contained within a 4 Tesla magnetic field, and
an outer barrel residing outside the magnet. Energy sampling
in HCAL is done using alternating layers of brass absorber
and scintillating tiles readout over optical fibers. We
describe the hadronic calorimeter, with particular emphasis
on the instrumentation developed for the readout system.
[Q14.002] Test Beam Results of the first production prototype (PPP1) of the forward calorimeter of the CMS Experiment
Ahmet Sedat Ayan (Department of Physics and Astronomy, University of Iowa), CMS HF Calorimeter Collaboration
We present the test beam results of the first production
prototype (PPP1) of the hadronic forward calorimeter (HF) of
the CMS experiment. The HF is based on the detection of the
Cherenkov light produced by shower particles and consists of
300 micron quartz fibers embedded in an iron matrix. The
results presented are about the light yield of the detector,
energy resolution for electron and hadron detection and the
signal uniformity and linearity. The signal generation
mechanism gives this type of detector unique properties:
narrow, shallow shower profiles and extremely fast signals.
These properties are presented in detatil.
[Q14.003] Optimization of Calorimeter Responses for the DZero Detector
Kwok Chan (University of Rochester), DZero Collaboration
Using events produced by the DZero Run II event simulation
program, the responses of the calorimeter are optimized by
reweighting the energy measurements in longitudinal layers
of electromagnetic and hadronic sections. The results show
that both the energy resolution and uniformity of the
calorimeter response for electrons and pions are
significantly improved by the optimization procedure.
[Q14.004] Large scale test of MSGC + GEM detectors in a high intensity hadron beam
Pascal Vanlaer (IIHE/ULB, Brussels), CMS forward MSGC Team
A large scale test of microstrip gas counters (MSGC) equipped with gas electron multipliers (GEM) was conducted at the Paul Scherrer Institute, Switzerland, in order to investigate the robustness of that technology in an LHC-like environment. The setup comprised 16 modules, each consisting of 4 MSGC's with a common GEM foil and a common drift plane. The total sensitive area was 1.3~m^2 and counted 16384 readout channels. The modules were exposed to a beam of 360~MeV/c pions at an intensity of 4.10^3~Hz/mm^2 during 376 hours.
During that period, all detectors were operated at voltage settings corresponding to 98% detection efficiency for MIP's at the LHC. The Signal-to-Noise ratio was about 33. Discharges induced by heavily ionizing particles were recorded on MSGC substrates and GEM's at a tolerable rate, and lead to the loss of 24 anode strips. This damage, when extrapolated to 10 years of operation at the LHC, corresponds to 8% of the total number of channels, and is considered acceptable for tracking. At the end of the test, the detectors were brought to S/N \simeq 100 without significant increase of sparking rate.
The CMS MSGC forward group has mastered large-scale production of MSGC + GEM detectors, and this test has demonstrated the capability of this technology, of coping with the hostile irradiation environment expected at the LHC.
[Q14.005] Track Segment Finding with the CDFII Online Track Processor
Christopher Neu (Ohio State University), CDF Collaboration
With increased accelerator luminosity and detector upgrades,
Run II at the Tevatron offers not only unprecedented physics
opportunities, but also exciting new technical challenges.
At CDF, the new Central Outer Tracker (COT) coupled with the
decreased bunch spacing requires the design of a new track
processor to identify tracks in the central detector. This
critical component of the triggering system must be
efficient, fast and accurate. The eXtremely Fast Tracker
(XFT) meets these criteria. The XFT is divided into two
major subsystems, the segment finder and the segment linker.
We report on the XFT's role in the Level 1 triggering system
at CDF and the Finder subsytem. The Finder identifies track
segments within a 12-wire layer of the COT. The device is
highly parallel and makes use of field programmable gate
arrays. The design, testing and commissioning of the Finder
are detailed.
[Q14.006] CDFII's Track Segment Linking in the Online Track Processor
Carlos Sanchez (Ohio State University), CDF Collaboration
At the CDF experiment, the eXtremely Fast Tracker (XFT)
addresses the challenge of track trigger processing at high
beam crossing rates. The LINKER subsystem of the XFT has the
specific role of reconstructing tracks by linking track
segements across four layers of the new Central Outer
Tracker. Every 132 ns, the Linker subsystem reports the
position and transverse momentum of up to 288 tracks. The
system is designed to achieve a momentum resolution of
\delta p_T/p_T^2 < 1% and identification
efficiency greater than 96% for tracks with p_T > 1.5
GeV/c. The XFT exploits a highly parallel, pipelined design
in order to accurately perform under tight timing
requirements. Here we report on the design and testing of
the LINKER subsystem, and the integration of the LINKER into
the XFT system.
[Q14.007] The Promise of Computational Grids in the LHC Era
Paul Avery (University of Florida), CMS Experiment Collaboration
The LHC era brings unprecedented challenges in information technology:(1) providing rapid access to massive data stores of 100 PB or more and (2) providing transparent access to heterogeneous computing resources throughout the world across an ensemble of networks of varying capability and reliability.
I discuss here how these challenges can be met by a hierarchical computational Grid of data analysis centers linked by Gbps networks. Such a Grid will allow physicists to play key roles in all stages of the data analysis, from development of the reconstruction programs and software infrastructure, to the extraction of first physics results. This capability will be extended to physicists at their home institutions through the use of an integrated distributed Grid system architecture designed for efficient petaByte-scale data access and analysis.
Several Grid projects are already underway. The
network-monitoring and control software, and the modeling
and optimization systems and methods, will be widely
applicable functional components that should also drive the
design and implementation of future distributed systems in
HEP and many other fields of science, engineering and
industry.
[Q14.008] Designing Polymer Blends using Neural Networks, Genetic Algorithms, and Markov Chains
Nilay Roy (Center for Simulational Physics, University of Georgia, Athens, Georgia 30602), Walter Potter (Department of Computer Science, University of Georgia, Athens, Georgia 30602), David Landau (Center for Simulational Physics, University of Georgia, Athens, Georgia 30602)
Obtaining miscibility in blends involving engineering
polymers has always been a challenge in the field of polymer
physics. However the fruits of succeeding has almost always
resulted in the development of a cost effective new material
with properties reflecting those of the parent polymers. In
order to predict miscibility a large number of theoritical
techniques are available, but none can make use of the
enormous existing information of these materials. A novel
method is presented here that uses Neural Networks to select
polymers (or simple modifications of these polymers) as
candidates for the blend. Then the technique of Genetic
Algorithms and Markov Chains is used to predict the
miscibility of the blend by finding for example the gibbs
free energy of mixing or a dimensionless interaction
parameter. After equilibriation by running the genetic
algorithm a tranformation is made from the populations in
the genetic algorithm to obtain a transition matrix. Using
this and a remapping back to generate a new population
extremely high convergence not seen using conventional
genetic algorithm techniques is obtained. Several examples
of this method is also presented.
[Q14.009] Kernel Estimation in High-Energy Physics
Kyle Cranmer (University of Wisconsin-Madison)
The theory of Kernel Estimation offers a non-parametric method for estimation of the parent distribution from which data are drawn. These estimates can either be used for parametrizing the distribution of discriminating variables for use in confidence level calculations or as a powerful multivariate analysis tool. The method is presented and several software packages are introduced.