The muon detector should fulfill three basic tasks: muon identification, trigger, and momentum measurement. The high field solenoid magnet and its flux-return iron, which also serves as the absorber for muon identification, ensure the performance of these tasks. The muon detector is designed to have the capability of reconstructing the momentum and charge of muons over the entire kinematic range of the LHC. The system is extraordinarily robust with muon momentum measured three times independently.

7.1 Performance requirements

Muons from pp collisions at the LHC provide clean signatures for a wide range of new physics processes. Some of these processes are expected to be rare and will require the highest luminosity. The task of the muon detector is to identify these muons and to provide a precision measurement of their momenta over a wide range from a few GeV up to a few TeV. Cross section of CMS muon system layout is shown in Fig.3.1

To address the anticipated new physics at the LHC, we have designed a muon detector to provide the following functionality and performance:

  • geometric coverage: pseudorapidity covered up to (polar angles from 9.4 to 170.6 degrees), with the minimum possible acceptance losses due to gaps and dead areas,
  • muon identification: at least of material is present at all angles; a track can be identified as a muon candidate if it has penetrated through the muon detector,
  • transverse momentum resolution muon, the muon detector alone for ; for pt=10 GeV, 7-20% for pt=100 GeV, and 15-35% for pt=1 TeV .
  • matching with the central tracker: momentum depend spatial position matching for 1 TeV muon has to be less than 1 mm in the bending plane and less than 10 mm in the non-bending plane,
  • beam crossing tag: efficiency > 99%.
  • muon trigger: precise muon chambers and fast dedicated detectors provide a trigger with sharp pt thresholds from a few GeV up to 100 GeV.

7.2 Design considerations

The muon system has been designed to satisfy the performance requirements within reasonable cost constraints. Because the solenoidal field of CMS bends tracks in the r- plane there is a precise vertex constraint (~15 µm) which may be used. Thus the muon momentum may be measured in three independent ways:

  • sagitta measurement in the inner tracker;
  • bending angle measurement right after the coil; and
  • sagitta measurement in the return yoke.

Obviously the best resolution is obtained when all three methods are combined, and the redundancy of the measurements makes the system robust against various kinds of backgrounds. Resolution at low pt is limited by multiple scattering; resolution at high pt is limited by the chamber resolution. With resolutions on the order of ~100 µm per station, in the standalone measurement multiple scattering dominates forpt below ~100 GeV.

With multiple chamber layers in each station providing many distributed measurement points, pattern recognition of track segments is facilitated. With many points per station, track segment identification is not disturbed by rejection of hits affected by noise or backgrounds such as soft deltas and electrons from neutrons and gammas.

7.3 Detector overview

The muon detector consists of four muon stations interleaved with the iron return yoke plates. The magnetic flux in the iron provides the possibility of an independent momentum measurement, and the number of muon stations ensures the reliability of the system.

In the endcap regions, the forward muon detector comprises four layers of muon stations (MF1 to MF4) interleaved with the three iron disks making up the magnet yoke endcaps. Each muon station contains Cathode Strip Chambers (CSC) and dedicated trigger detectors RPCs covering the area below Room is left above for a possible high rate trigger detector. The CSCs consist of six layers of azimuthal strips and wires perpendicular to the central radial line, except for the MF1 chamber, which experiences the full field of the solenoid and has a tilted wires.

The MF1 station should meet specific CMS requirements to provide a very good spatial resolution of 50-75 µm for muon pattern recognition and efficient matching with the inner tracker and efficient acceptance coverage over the full range of pseudorapidity for triggering. Because MF1 is located inside the solenoid in the slot between the endcap HCAL and RF1 it should operate in conditions of a very strong nonuniform axial magnetic field BZ up to 3.35 Tesla and highest charge particle background up to 1.5 kHz per cm² while the maximum rate for other forward stations is less than 0.25 kHz per cm² [7.1]. To decrease strip occupancy MF1 station is divided in the radial "virtually," the strips are divided but are contained within one enclosure. A perspective drawing of the muon chamber layout is shown in Fig. 7.1.

Each chamber subtends 10 degrees in azimuth and provides an additional overlap of several strip widths at each side to provide continuous azimuthal coverage. Each station is divided azimuthaly into 36 sectors. To avoid dead areas in the azimuthal coverage at the edges of the chambers, all sectors are deployed in overlapping fashion. A gap of 64 cm between the Endcap HCAL and the RF1 iron disk provides space for the overlapping CSCs, for the electronics boards and cabling, for the dedicated RPC trigger chambers.

7.4 MF1 station description

7.4.1 Cathode Strip Chambers

Cathode strip chambers are multiwire proportional chambers with cathode planes containing some configuration of strips. The ability to measure the induced charge distribution on several strips gives these chambers a high intrinsic precision, while the short drift paths gives a fast time response measured on the anode wires.

It has been demonstrated with a wide variety of chambers that a resolution of 50-70 µm can be achieved with strip widths around 5 mm [7.2]. Although the maximum drift times can exceed the 25 ns corresponding to the beam crossing interval, a measurement of the first pulse coming from six chamber planes provides a reliable bunch crossing tag. (A more complicated algorithm may be required to guarantee the integrity of this measurement in the presence of backgrounds.)

Radial strips provide a natural frame of coordinates for measuring muon momentum. They become narrower toward the inner edge of the chamber. Several options such as the use of stereo strips are under consideration to resolve "ghosts." For MF1 chambers( Fig. 7.2.) the maximum strip length is 170 cm and maximum wire length is 53 cm. Low-mass paper honeycomb panels provide rigidity for the chambers and determine the uniformity of the gas gap and consequently of the gas amplification [7.3]. These panels provide also small capacitance of cathode strips etched on a thin G10 skins. It is important to keep input electronics noise at low level to achieve high spatial resolution for r--coordinate.

Wire diameter will be 30 µm. The width of the groups of ganged wires ranges from 2.5 to 5.2 cm. The anode cathode distance will be 2.5 mm, and the gas mixture , although other mixtures are still under study. Parameters of MF1 chambers are shown in the Table 1. With the present baseline design, there are 69120 cathode strips and 27648 anode wire channels in the MF1 system.

7.4.2 Effects of the magnetic field

The effect of the magnetic field on drifting electrons can degrade the resolution of gaseous detectors, and this was an initial concern for CSC devices, which had never been tested at high fields. There are effects of radial BR and axial BZ components of magnetic field in the muon endcap CSCs. The effect of a radial component of magnetic field parallel to the strips is similar to the well known effect of track inclination.

Effect of the axial component of magnetic field, orthogonal to the anode plane dominates in MF1 CSCs. The systematic shift of the induced charge distribution across the strips is proportional to the distance between the track and the nearest anode wire and the tangent of the Lorentz drift direction angle. The compensation of this effect can be achieved by anode wire plane rotation with respect to the strips by the Lorentz angle when the Lorentz drift direction becomes parallel to the strips. In case of radial strips the axial field effect cannot be fully compensated but can be minimized to the reasonable value. Recent beam studies at RD5, the CMS test facility at the CERN SPS, have verified that this effect can be corrected by stringing anode wires in a direction which is not exactly perpendicular to the strips but is slightly rotated.

7.4.3 Readout Electronics

The front-end electronics for the endcap muon system handles two main tasks: 1) it acquires charge and timing information from the CSCs, stores it during the many bunch crossings (approximately 128, or 3.2 µs) necessary before a trigger decision can be made in the central trigger logic and returned to the front end, and transfers it to the data acquisition system when the trigger criteria are satisfied; and 2) it identifies and measures local track segments quickly to provide input to the muon trigger logic. The system must be suitable for chamber mounting and its components must be able to withstand the expected radiation level. Front-end cathode electronics

In chambers of the CSC type, the spatial coordinate across the direction of the strips is measured by charge interpolation of the signals read from consecutive strips. The resolution of this interpolation is linearly proportional to the percentage error of the charge measurement. For CSCs with strip width equal to twice the anode-to-cathode distance, this is approximately given by where w is the strip width and Q is the total cathode charge. To achieve the best possible spatial resolution for MF1 chambers, the cathode charge measurement must be accurate to less then 1% and strip capacitance should be small as it was mentioned above. The electrical specification for the MF1 cathode readout electronics is summarized in Table 2.

Notes to the Table 2:

1. A level of 2000 e- (0.32 fC) equivalent input noise produced an ADC "pedestal rms." 3.2 counts. Noise level applies to 100 pF input capacitance and 300 ns shaping time.

2. Peaking time for "semi-gaussian" shaping. Transfer response of system is given by:

H(s)=st(1=st)-(1+n) ,
where st=peaking time and n=number of integration's. The most likely choice is n=4, since n=2 leads to a long uncancelled tail.

3. Ratio of input charge to ADC least count. This value depends on such things as the value of the feedback capacitor in the charge-sensitive preamplifier, the gain of the shaping amplifier, the ballistic deficit factor and least-count sensitivity of the readout ADC. For preamplifier-only designs, the required conversion gain is 5 mV/pC (10 mV/pC desirable).

4. 12-bit range will keep a larger fraction of the Landau tail on scale and will allow operation over a wider range of chamber gas gains.

5. This is maximum average rate of Level 1 accepts.

6. Subject of R&D

7. Ambient temperature range over which all other system specifications must be met.

8. Total power for electronics divided by total number of channels. Includes readout and trigger sections.

9. The muon electronics are situated outside the calorimeter and therefore are expected to experience only low radiation levels. This remains to be verified.

10. Calculations are in progress. The stated value assumes a neutron fluence scenario of Fn=106 cm-2 s-1 at L=1034 cm-2 s-1 for ten LHC years.

The physical layout of the frontend electronics could be the following: sixteen cathode strip channels from each of the six layers in a station are connected to each front-end strip card mounted on the chamber. These cards carry preamplifiers, data acquisition and storage chips, and basic trigger pattern recognition circuitry. Wire cards discriminate anode signals and also incorporate bunch timing and trigger pattern recognition circuitry. All data from both strip and wire cards, whether for data acquisition or for triggering, pass through a single motherboard mounted on the face of each chamber. Front-end Trigger

The momentum resolution required for a trigger capable of limiting the single muon trigger rate to the few kHz level was mentioned earlier: a threshold of 20 - 40 GeV with better than 30% momentum resolution. Such azimuthal resolution is achievable at the trigger level up to by localizing a track to within a one strip width maximum deviation.

Patterns among a set of hits in anode wire groups are found in the same way as those for the strips, although the segmentation is much coarser and the patterns are not pt dependent. An algorithm to obtain the bunch crossing is to require a minimum number of anode wire groups out of six layers to contain hits, and to then take the timing from the first hit out of the six layers. The coincidence of tagged track signals in both coordinates forms the basis of the CSCbased muon trigger.

7.5 Assembly and Mechanical Structure

The design of the endcap iron toroids must balance many considerations. The large magnetic field and consequent large axial magnetic forces (greater than 10000 tons) imply that significant mechanical distortions must be accommodated. Because the MF1 is located in the 64 cm width slot between endcap HCAL and RF1 special design and access procedure are needed for installation, cabling and service. The iron disks must move to provide access to the inner components of the detector and allow chamber installation and maintenance.

7.6 Simulation

CSC simulation program which includes details of particle passage through the chamber, gaseous processes of ionization and electron drift in a magnetic field environment, cathode charge signal induction, electronics influence on charge measurements etc. has been developed for CSC study.

A Monte Carlo program MARS'94 was developed for inclusive simulation of threedimensional hadronic and electromagnetic cascades in matter and for hadrons, muons, lowenergy neutrons, photons and electrons transport in accelerator and detector components at energies up to 20 TeV. Energy thresholds are 10 MeV for muons and charged hadrons, 100 KeV for electrons and photons, 2 x 10-6 KeV for neutrons. This code provides fast cascade simulation and has user-friendly geometrical block for complex geometry's and composite materials description.

MARS'94 together with FLUKA and GCALOR codes was used for the CMS muon system radiation background shielding design described in the Technical Proposal of CMS [7.1]. This design takes into account the results of detailed studies of radiation background formation processes inside muon chambers [7.8]. The main forthcoming activity on the muon system simulations will be related with:

  • muon chambers background signal generation;
  • occupancy fluctuations simulations;
  • radiation shielding optimization;
  • chamber components induced radioactivity calculations;
  • background influence on the trigger efficiency studies.

7.7 Research and Development

Since MF1 should operate in an axial magnetic field up to 3.35 Tesla one of the major goal of R&D was to study the CSC spatial resolution in a strong magnetic field. To measure the Lorentz angle in 3 Tesla magnetic field the small size (0.5x0.5 m²) 4layer prototype [7.4] was constructed in Dubna and tested in RD5 facility in 1993. In high energy muons the spatial resolution of this prototype was measured as 52 µm with a good uniformity across the strips. Measurement of Lorentz angle (magnetic field parallel to cathode strips) was performed with cosmic rays.

The results shown in Fig. 7.3 give a clear determination of the Lorentz angle for several values of the field. Based on 1993 test results the real size engineering prototype of MF1 sector was constructed at Dubna [7.5]. This prototype was instrumented with new front end readout electronics.

The MF1 prototype comprises two 6-layer trapezoidal chambers overlapped in r with radial strips. Dimensions are shown in Fig. 7.4 Cathode strips readout electronics is based on a low noise analog signal processor GASPLEX designed at CERN [7.6]. Following the 1993 Lorentz angle measurements anode wires in the CSC 2 are tilted by an angle with respect to the strips for Lorentz effect compensation in a 3 Tesla magnetic field. Anode wires in CSC 1 are orthogonal to strips. CSC 1 was tested in a 200 GeV muon beam. Spatial resolution at normal incidence is shown in Fig. 7.5a and equals to 49.6 µm. CSC 2 was tested in 1994 in the 3 Tesla magnetic field perpendicular to the anode plane. An average spatial resolution over the full sensitive chamber area of 63.4 µm was achieved (Fig. 7.5b ) and it was demonstrated also that a properly compensated chamber has a good resolution in a strong magnetic field.


One can conclude that the CSC spatial resolution fully satisfies MF1 requirements in a strong magnetic field. The magnet M1 will be modified for 1995 so that it can be rotated by 90 degrees and put into "end cap configuration" to carry out an integrated test of the end cap calorimeters and the MF1 chamber prototype.

Muon-induced electromagnetic secondaries are a potential danger for pattern recognition, tracking and triggering was studied by analyzing the data vertical cosmic rays with 4-layer Dubna prototype [7.4]. The electron production probability in CSC of 11.8%± 1.2% per layer was obtained. The estimate of the fraction of electrons confined to one layer is amounts to 78.8%±12.3% of the events with electrons [7.7]. This estimation is in agreement with DTBX group result.

Further R&D effort is required in a number of areas to address the remaining design issues and to progress toward system fabrication, the overall goal being to construct a muon detector which meets our performance specifications at the lowest possible cost and on pace with the LHC schedule. To focus on the most urgent projects for the next two years, we can enumerate the following areas:

  • to carry out an integrated test of endcap calorimeters and MF1 muon chamber in the magnetic field;
  • to construct a full size MF1 prototype which has no radial acceptance losses;
  • to complete the conceptual design of the front-end data acquisition and trigger electronics, and to validate the design by producing enough channels for prototype chamber testing;
  • to design low-cost, low-power ASIC circuits for the readout of the chambers;
  • to design and test the local alignment system.

7.8 Schedule of the Project

By 1997:

  • To complete full design of MF1 CSCs ready for mass-production. It requires a chambers simulation, full understanding and testing of CSCs, layout and trigger design. Work is well in progress.
  • Layout design: continue particles background simulation, shielding design, MF1, HCAL and VF integration.
  • Construction and investigation of prototypes and test chambers. We are already at the level of full size devices.
  • Front-end readout, trigger and slow control electronics design with a standard endcap outputs for DAQ and CSCs muon trigger and its test.

In 1997:

  • To set up CSC massproduction facility in Dubna. It requires full technology design tooling design and fabrication
  • Facility construction. At present we have available two areas: ~300 m² and ~1000 m² which are being used for prototypes construction. We plan on these areas to set up CSC production facility oriented for MF1 chamber mass-production but which is capable to manufacture all the CSC for all MF stations if needed (backup).

Since 1997:

  • To start up mass production of CSCs fully instrumented with electronics and other features. It requires first chambers in series and preproduction series manufacturing and test.

7.9. The MF1 Cost estimate

The cost review is presented in the Table 3 and taken into account the following scheme:

  • wastage materials for chambers construction 5%
  • spare cathode, anode and trigger electronics 5%
  • spare HV and LV channels 5%
  • spare CSC chambers (materials + assembly) 3 pc

The following items were not taken into account:

  • R & D;
  • preproduction;
  • monitoring system;


[7.1] I.Azhgirey and A.Uzunian, "CMS muon system radiation background shielding", CMS TN/94-266,1 November 1994

[7.2] L.Barabash, A.Baranov, I.Golutvin et al., "A Study of the Accuracy of Proportional Chambers with Cathode Strip Readout", JINR, P9-82-724, Dubna, 1982 and Nucl. Instrum. Methods v. A236, p. 271-273, (1985), Bo Yu, "Gas Proportional Detectors with interpolating Cathode Pad Readout for High Track Multiplicities", BNL 47055, December, 1991, A.Chvyrov, Yu.Ershov, I.Golutvin et al., "The Spatial Resolution of 3x0.3 m2 Dubna Prototype CSC", GEM TN 93-466, 14 September 1993

[7.3] L.S.Barabash, A.V.Vishnevsky, I.A.Golutvin et al., "A study of the Spatial Resolution of a Large Proportional Chamber Using Cathode Information Readout", JINR, P13-81-554, Dubna, 1981

[7.4] Dubna 0.5x0.5 m2 RD5 CSC Prototype 1993, in: Milestones Report, CMS Collaboration, CERN/LHCC 93-48, p.70, 1993

[7.5] Dubna Full Size Prototype of an MF1 Sector, 1994, to be published

[7.6] J.C.Santiard et al., "GASPLEX a Low Noise Analog Signal Processor for Readout of Gaseous Detector", CERN/ECP 94-17, 25 October 1994

[7.7] C.Allbajar et al., "Electromagnetic Secondaries in Muon Detection at LHC CERN-PPE/94. To be submitted to Nucl. Instrum. and Methods.

[7.8] I.Azhgirey and A.Uzunian, "Muon Chambers Occupancy in the Low-Energy Background Fields", CERN CMS TN 94-265; "Background Electron Track Length Distribution in CSC", CERN CMS TN 94-267.