An overview document with appendices of short summaries of R&D proposals will be put together to assist R&D committee in its decision process. Needs to be completed by late January.

preparation for simulation work required for Barcelona results

use the official "LCD" simulation framework
volunteers need to come forward

coordinate work with international colleagues
converge on limited number of well conceived tracking designs

A working group web page will be set up as a central information source. Reminder that the working group organizers are expected to be rotated frequently.

Discussion

important to consider the effect of machine backgrounds in the tracking detector simulation
R&D projects need to focus on difficult issues, not on essentially solved ones.

Linear Collider tracking R&D in Europe (Ron Settles)

Motivation for choices of reference detector

Recoil mass resolution for Higgs production through "Higgstrahlung" process drives the need for excellent dp/p. Hadronic WW and tt channels require good energy flow measurements, therefore good granularity. Although some consideration was made of an all Si tracker, the group decided to concentrate on a TPC design. Beats down background using high granularity (about 10 times the granularity of the ALEPH TPC is proposed). Minimize material to improve matching to vertex detector and reduce the number of photon conversions (about 100-600 1 MeV photons/bunch in barrel are expected from machine backgrounds). Overall backgrounds (photons, charged particles and neutrons) not expected to pose a difficult challenge for TPC operation.

Detector R&D

Past R&D efforts are described, most recently, in DESY 97-123E. Current tracking R&D include:

M. Tonutti, etal.	Honeycomb Intermediate Tracker	background studies, improved design
R. Nahnhauer, etal.	Fiber trigger detector	more design work
R. Orava, etal.	GEM chamber	GEM R&D, chamber design
M. Gruwe, etal.	GEM TPC readout	design and prototype work
P. Colas, etal.	Micromegas TPC readout	design
R. Settles, etal.	TPC	gas, prototype work
E. Rulikowska	Gas studies	general gas studies

Experience with the SLD drift chamber (Tracy Usher)

The SLD drift chamber uses a slow gas and achieves 100mm resolution per point. The chamber has worked very well. Tracy summarized some of the problems encountered in operating the drift chamber at a linear collider.

Low momentum tracks displaced from z = 0. These are caused by off momentum beam electrons hitting masks that protect the CCD vertex detector, located at z = ± 20 cm. This causes difficulties in the track trigger, in that the rate for single tracks is too high to be used as a stand alone trigger. Having z information available in the track trigger would have helped.
Synchrotron radiation, from tails of beam passing through focusing elements, is the dominant source of background hits. Individual hits saturate the ADCs for 4-5 neighboring channels. Static currents up to 1-2 mA per layer observed. Occupancy often is of order 10%, sometimes reaching 20%. At such high occupancies track finding efficiency starts to suffer. Background bursts, typically due to klystron trips, result in very high occupancy and HV trips.
One wire broke in September 1997. A quick patch requires three days. To fix properly takes a week or more.

Experience with the ALEPH TPC (Ron Settles)

The ALEPH TPC provides momentum resolution of s_p/p² = 0.0012 GeV^-1 and when combined with the ITC and vertex detector, the resolution improves to 0.0006 GeV^-1 . It provides up to 21 space points per track in a 1.5 T field. The pad segmentation was limited by the readout cable space. Multiplexed solutions would allow much higher density. Greater care should have been taken in the initial magnetic field mapping. A laser system is used to monitor the drift velocity. Space point resolutions (high p_t) are 173 mm in r-f and, after a tracking software upgrade, 450 mm in r-z. Two track separation in r-f is about 1.5 cm and much worse in r-z. This is limited by the electronics, not diffusion. Expect these could be much better in a LC TPC design.

Alignment has been a continuing battle, especially with the vertex detector and tracking software upgrades. More challenging now that 1 million Z's are not available each year. The TPC geometry has advantages is that it can be quite long, with the wire chambers easy to service, but for LC better tools will be needed to measure distortions.

Question: How fast can you recover from a background burst?
Answer: When wire chambers trip, it requires about 5 minutes to bring them back to operating voltage.

The CDF Upgrade tracking system (David Gerdes)

The CDF II tracking system consists of the SVX II, an intermediate silicon layer (ISL), and an drift chamber (COT). Two years ago the leading contenders for outer tracking was a combination of scintillating fibres and straw tubes. The SVX II consists of 5 layers from 2.5 cm to 10 cm in radius. Five layers are used instead of four, in the previous tracking system, in order to allow standalone tracking in the high background environment. The ISL is a single layer for h < 1 and two layers for 1< h < 2, which provides a good space point in a low occupancy detector to match tracks between the COT and SVX II. The COT occupies the space from 44 to 132 cm in radius, is 1.7 % radiation lengths, uses a fast gas (90 mm/ns), and has a 2 track resolution of about 4 mm. The drift field in the cells are provided by metalized mylar panels, not wire planes. This helps isolate the effects that a broken wire might cause. The momentum resolution is s_p/p² = 0.0007 GeV^-1 .

The CMS tracking system (Muzaffer Atac)

The CMS tracking system consists of pixels, silicon strips and MSGCs. A total of 75 m² of Si is to be used, with 5.4 million channels. In the initial phase of operation at LHC (low luminosity) the Si pixels will be a 5 cm radius and some of the Si strip and MSGC layers will be absent. Later the Si pixels will be moved back to 7 cm.

MSGC R&D has been a long term project (some 10 years). There have been successes and failures along the way. The CMS design uses a 300 mm glass substrate, a 3 mm drift gap with 3500 V applied for the drift field and 520 V between the anode and cathode strips. Stringent tests still need to be made before the device is to be successful. The group is contemplating a combined MSGC - GEM design.

The DØ upgrade tracking system (Norman Graf)

The DØ environment required a rad hard, fast, triggerable tracking system to fit within its new coil (radius 50 cm). The technology chosen was a 4 layer silicon microstrip tracker in the barrel up to radius of 12 cm and in disks perpendicular to the beam for large h tracking. Outside that, there are 8 layers of scintillating fibers (20 cm < r < 50 cm). Each layer consists of an axial doublet layer and a stereo doublet layer. The 77,000 fibers (830 mm Ø and 1.6 to 2.5 m in length) are readout with VLPCs (at 9° K). Altogether, the fibers add 6% of a radiation length. The momentum resolution of the tracking system is expected to be s_p/p² = 0.002 GeV^-1 .

The ATLAS tracking system (Hartmut Sadrozinski)

The ATLAS inner detector (ID) consists of Si pixels, Si strips, and transition radiation detectors (TRT) made of straw tubes. In the barrel the systems are coaxial to the beam, and for large h the Si strips and TRTs are aligned in planes perpendicular to the beam. Good momentum resolution is required to determine the sign of 1 TeV leptons. The design resolution is s_p/p² = 0.3 TeV^-1 . Tracks are measured in about 11-13 layers of silicon and 35 layers in the TRT. The total material, including services, is a strong function of h, ranging from 0.3 to over 1 radiation length. The expected occupancy in pixels is < 10^-5, in silicon strips < 10^-5, and up to 40% in the TRT. Such high occupancies will affect tracking and pion identification.

Lessons learned:

difficult to deal with high particle flux. Radiation damage forces detectors to large radii.
difficult to deal with 40 MHz BX. Increased power demands, more mass.

Recommendations:

keep front end electronics and services outside of tracking volume
use disks to close off acceptance in endcap
put one additional silicon layer right inside the magnet

Robotic drift chamber stringing (S. Csorna)

CLEO III chamber geometry made human stringing difficult or impossible. A robotic procedure was developed over a period of 8 months. The chamber was strung vertically with the robot feeding the wires from the top to a human at the lower endcap. The robot uses CCD video cameras and pattern recognition software to align itself. The system worked very well. Using such systems, one could consider building chambers with much smaller diameter wires and/or much smaller cells than have been built in the past.

Monday Session

Traking group planning discussion (Keith Riles)

Plans for the short term (over next few months)

prepare to submit R&D proposals

try to coordinate submissions, joint submission would be ideal

set up tracker simulation infrastructure

Medium term goals (over next 12 months)

carry out simulation studies
initial winnowing of tracker options

It is expected that the R&D proposals for the first year will concentrate on simulation. Any detector R&D proposals submitted for FY99 need to be specific to LC, to have a good chance of receiving funding. The working group organizers are asking for proposal abstracts from interested groups within the next month. The abstracts will be used to decide whether joint submission of a single proposal or individual submissions of proposals makes the most sense. In the latter case, the working group coordinators will write a general recommendations document. In any case the organizers will try to minimize overlap of R&D proposals.

In order to make progress on simulation infrastructure, technology advocates are requested to provide realistic hit simulation over next 2 months. These will then be incorporated into LCD framework, in order to start generating large MC samples. The fallback solution is track parameterization, but this would leave a credibility gap. For pattern recognition and fitting, TRF++ (from DØ) will be brought into the LCD framework.

The following is a list of problems that need further study:

Mike Peskin?s homework problems
forward tracking for luminosity as a function of Ecm

David Miller suggests that acollinearity distribution of Bhabha scattering can be used to determine the luminosity as a function of energy. This requires angular resolution of 1 mrad for scattering angle above about 100 mrad.

is an intermediate tracking system needed for the ?L? detector
two track separation issues (q and f)
vertex-mass reconstruction (b/c discrimination)
requirements of dE/dx
?L? option: is a special tracking layer at tracker/ ECAL boundary a good idea?

Flat panel detectors (John Kadyk)

Active Matrix Arrays are commercially available, developed for flat panel displays, x-ray imagers and page readers. Typical pixel size is 100 ´ 100 mm² pixels over an area of 50 ´ 50 cm² . Pixels of size 50 ´ 50 mm²are also possible. Commercial applications do not demand fast readout. Readout speed needs to be improved for particle detector applications. John expects that 10 ns/pixel will be easy to achieve. A fast clear of all channels (10 ns) is possible with these devices. Costs are dropping rapidly, with increasing production.

Gas microdetectors R&D ( Madhu Dixit)

MicroStrip Gas Chambers (MSGCs) were born 11 years ago. Since then much progress has been made, and a number of variants on the original idea now exist. HERA-B and CMS have MSGCs as part of their tracking systems, indicating that the technology has reached a certain level of maturity.

MSGCs suffered a problem discharges and shorts in test beams that were not seen in lab tests. This is now understood as being due to non-quenched streamer mode processes induced from electron emission from the edges of the cathode. CMS has treats this problem by coating the edge of their cathodes (edge passivation).

Gas Electron Multipliers (GEM) consist of a 50 mm thick kapton metalized with 5 mm gold and perforated with an array of micro-holes. It can be used alone or as a multiplying stage for a MSGC. Microdot detectors are essentially a pixilated form of a MSGC. Good gain and aging properties have been obtained with such detectors. Micromegas detectors have a micro mesh just above strips (supported by quartz fibres). They offer fast signal rise times and very high rate capability.

Micro gas avalanche detectors have two remaining problems:

heavily ionizing particles will limit gains in such detectors (except micro-dot chambers) to 1500-2000. This does not leave much room for error
aspect ratio gives poor performance for tracks incident > 10 degrees from perpendicular

possible solution using CsI coatings (see talk by John Kadyk)
not a problem for use in TPC readout

Silicon readout electronics R&D (Bruce Schumm)

Bruce identified many issues needing further study

is it feasible to turn off electronics between NLC bunch trains, in order to reduce cooling requirements
investigate ideas to maximize S/N - allowing thinner detectors
optimize timing resolution - it would then be possible to remove background hits from other bunch crossings than the one containing the physics event
improve spatial resolution - investigate ways to reduce charge spreading from Lorentz angle

An estimated cost for the first year study would be between $60K and $100K.

Diamond detector R&D (Hans Ziock)

Diamond is a unique semiconducting material in that the bandgap is very large, 5.5 V, about 5 times that of Si.
This results in a very low dark current. The detectors can provide fast signals, and are radiation hard. The disadvantage with diamond is that because it takes more energy to produce an electron-hole pair, 13 eV / e-h pair (c.f. 3.6 for Si), fewer e-h pairs are produced per radiation length: 4500 e-h / 0.1 % X₀ (c.f. 8900 for Si). There is not an obvious match to diamond technology and central tracking at a new LC. It could be useful for far forward tracking or beam monitoring.

Silicon drift detector R&D (Sanjeev Pandey)

Silicon drift detectors give pixel information, but with far fewer channels than pixel detectors. The drift time gives the z coordinate and the segmented anode gives the transverse coordinate. Resolutions of order 20 mm are easily achieved and two track separations of 500 mm are easily resolved. Silicon drift chambers are being used for the STAR experiment at RHIC. The experiment will have 216 SDD wafers for a total of 104 K channels. These are in production at this time, with a total cost of the detector being $7M, including salaries involved in development. The detectors were prototyped in experiment E896 (Au+Au at AGS), where good performance was observed even in the high B field of 6.4 T. The detectors operate well at room temperature.

There are a number of R&D issues that would be appropriate for improving these detectors for LC applications:

improve position resolution to 5 mm (by decreasing the anode pitch)
reduce thickness from 300 mm to 150 mm
build SDDs on larger wafers (6") for longer drift distance, smaller number of channels

The technology for SDDs is relatively mature. The cost to build a new detector would be about $15/channel (including R&D, but not readout electronics). The expertise from the STAR group is available after STAR construction is complete, in summer of 2000.

TPC R&D (Mike Ronan)

The STAR experiment at RHIC has chosen to perform the central tracking with a TPC. This is a very challenging environment where the average charged multiplicity is 2500. The front end electronics are mounted on the endplate, with signals transmitted by fiber, allowing a high granularity readout. Several R&D studies should be performed for a TPC at the linear collider, including

detector simulations
software development
chamber design

gas, field cage, readout device and geometry

Columnar CsI drift electrodes for Gas microdetectors (John Kadyk)

One way to solve the aspect ratio problem in MSGCs is to replace the 3 mm gas gap with a 400 mm thick columnar CsI layer. This provides a high efficiency for secondary electron emission. When a field is applied to collect the electrons, actual gain in CsI itself is observed. This method works in that hit efficiency no longer depends on incident angle when used with an MSGC. In fact, enough electrons can be generated from the CsI itself, that a MSGC is not required! The material is not small, however, about 1-2 % X0 per layer.

Back to tracking home page