Blum W., Riegler W., Rolandi L. Particle Detection with Drift Chambers

Подождите немного. Документ загружается.

382 11 Existing Drift Chambers – An Overview

Fig. 11.17 Mark II drift cell

with electron trajectories

The small cell of the SLD drift chamber (Fig. 11.18) is designed to create a better

separation between the ampliﬁcation region and the drift region by displacing the

ﬁeld wires from their position between the sense wires to a position surrounding

them. This is done to create a uniform drift region for the slow gas employed. Here

the good diffusion properties of carbon dioxide have been combined with its low

drift velocity at high electric ﬁelds. Thus the Lorentz angle is kept near 6

◦

at the

ﬁeld of 0.6 T. On the other hand, this imposes severe requirements on the electric

ﬁeld uniformity, since the drift velocity under these conditions is proportional to the

electric ﬁeld (‘unsaturated gas’). There are 640 drift cells in the detector. A very

favourable point-measuring accuracy of 55 μm was obtained with the prototype.

This requires that the wire displacements caused by mechanical inaccuracy, gravity,

and electrostatic forces as well as by the stereo arrangement are fully controlled to

this precision.

Fig. 11.18 SLD drift cell with

electron drift lines

11.5 Large Cylindrical Drift Chambers of Type 2 383

Fig. 11.19 A drift cell of the CDF central tracking chamber, with electron drift lines

The drift cells of the central tracking chamber of the CDF detector (Fig. 11.19)

are inclined by 45

◦

with respect to the radial direction in order to take the Lorentz

angle at the employed magnetic ﬁeld of 1.5 T into account. This solution also offers

advantages for triggering and for pattern recognition. The arrangement of the drift

cells, of which there are 660, in the end plate can be seen in Fig. 11.20. The cells

are divided radially into nine superlayers which alternate between wires in the axial

direction and wires at a ±3

◦

stereo angle. Tilted drift cells are also employed for

the drift chambers of the H1 and ZEUS experiments at the HERA e-p collider.

Fig. 11.20 The end plate of

the CDF central tracking

chamber with its drift cells,

which are inclined with

respect to the radial direction

384 11 Existing Drift Chambers – An Overview

11.5.4 The UA1 Central Drift Chamber

The detector ﬁlls a cylinder, 6 m long and 2.2 m in diameter, which contains the

collider-beam pipe on its axis. Equally horizontal, but at right angles to the beam, is

the magnetic ﬁeld of 0.7 T (see Fig. 11.21). Along its length, the cylinder is divided

into three cylindrical sections, each 2 m long: the central section and the two forward

ones. The wires are all parallel to the magnetic ﬁeld and organized in planes accord-

ing to the following scheme: the anode planes have alternating sense (35 μmdiam)

and ﬁeld (100 μm diam) wires every 5 mm; the cathode planes with 120−μmwires,

also every 5 mm, are parallel to the anode planes, leaving drift spaces of 18 cm be-

tween them. A peculiarity of the UA1 detector is the orientation of these planes –

they are vertical in the central cylinder part and horizontal in the forward parts. This

arrangement is a consequence of the horizontal magnetic ﬁeld; the solution opti-

mizes the momentum-measuring accuracy, the track curvature being measured by

the drift time. The charge-division technique is used for the coordinate along the

wire direction. Filled with a gas mixture of 40% Ar and 60% C

at ambient

pressure, the drifting electrons reach a velocity of 5.3 cm/μs under a Lorentz an-

gle of 23

◦

in their drift ﬁeld of 1500 V/cm. The total number of wires is 22 800, of

which 6110 are sense wires. A stiff vertical track is measured on 100 points and a

stiff track near the forward directions on up to 180 points.

Fig. 11.21 Geometry of the

UA1 central detector

11.5 Large Cylindrical Drift Chambers of Type 2 385

Performance ﬁgures were measured in the experiment as follows [NOR 90]: The

average point-measuring accuracy in the direction of the drift amounted to 350 μm.

The relative accuracy of momentum measurement, which depends very much on

the track orientation, was typically

p/p =(0.01 GeV/c

−1

)p for isolated tracks in

a favourable direction (obviously, the curvature of tracks that emanate from the in-

teraction region under a polar angle of 90

◦

and parallel to the magnetic ﬁeld cannot

be measured.) The electronic two-track resolution limit on a wire was about 5 mm.

The coordinate measurement accuracy along the wires was 3.5% r.m.s. of the wire

length after the gain reduction by a factor of three that was necessary when the de-

tector had to work in the high-luminosity environment; before this change, 2% had

been reached. These numbers illustrate the conﬂicting requirements of having the

smallest possible space charge in the drift region (small wire ampliﬁcation) and the

best possible charge-division accuracy (large ampliﬁcation) at the same time.

Ionization measurements of tracks in this complex detector were limited in accu-

racy and were not employed for particle identiﬁcation.

In Fig. 11.22 we reproduce an event as reconstructed by the computer; the high

number of measured points along each track make a clear picture even for the com-

plicated events created in high-energy hadron collisions – in fact this picture is

famous because it contains the ﬁrst Z

particle ever observed. The two arrows mark

the electron and the positron into which it decays.

11.5.5 The ATLAS Muon Drift Chambers (MDT)

The ATLAS experiment at the Large Hadron Collider (LHC) will investigate

pp-collisions with a total energy of 14 TeV; muons with transverse momenta up

to 1 TeV/c will be detected and measured in the muon spectrometer [ATL 97]. The

purpose of the muon drift chambers is to determine the momenta of the muons as

accurately as possible.

The ATLAS magnet produces a toroidal ﬁeld in air (ﬁeld lines going around

the beam) which offers a high bending power at much reduced Coulomb scattering

Fig. 11.22 Computer-reconstructed event in the UA1 central detector – the ﬁrst Z

particle ever

observed is seen to decay into an e

and an e

−

. Only one in two wire signals is displayed

386 11 Existing Drift Chambers – An Overview

Fig. 11.23 ATLAS muon drift chamber, consisting of a pair of threefold tube layers, separated by

a spacer frame

because there is no iron yoke. The chambers consist of simple aluminium drift tubes

with a diameter of 3 cm, carrying a central wire. (The theoretical work as well as the

practical measurements that were done in order to ﬁnd the best operating conditions

for these tubes were published in a series of papers in 2000 [RIE 00]). They are

mounted to form chambers as shown in Fig. 11.23; almost all chambers consist of

a pair of threefold plane layers of parallel tubes, the triple layers being separated

by a spacer. This is the basic unit, mostly 1 to 2 m wide and up to 6 m long in the

direction of the tubes. There are 1194 of them, covering a total surface of 5500 m

As the coordinate measurement of the drift tubes is perpendicular to the wires, the

chambers are oriented with the wires essentially following the magnetic ﬁeld lines.

In the ‘barrel’, the chambers are arranged in three cylindrical concentric layers so

that a stiff particle track from the interaction point is measured in all three layers for

optimum momentum resolution (see Fig. 11.24). The vertical detectors in the ‘end-

caps’ measure tracks in the more forward or backward directions in a similar way.

It should be noted that the muon drift chambers can only determine one coordinate,

the one that is useful for the momentum measurement. The azimuthal coordinate of

a track is measured in other detectors of the spectrometer. The overall dimensions

are characterized by a cylindrical volume into which the muon spectrometer can be

ﬁtted; it is 22 m in diameter and 45 m long. The barrel chambers occupy radial space

between 4.5 and 11 m, and the longitudinal space for the vertical end plates extends

from 7 to 11 m on either side of the interaction point. At the time of writing the

second edition of this book, the spectrometer is being assembled to start operation

in 2008.

The momentum measuring accuracy in the given magnetic ﬁeld depends essen-

tially on three factors: (1) the track measuring accuracy of the individual tube,

(2) the accuracy with which the relative positions of the tubes are known, and

11.5 Large Cylindrical Drift Chambers of Type 2 387

Fig. 11.24 Cross-sectional view of the barrel part of the ATLAS muon spectrometer. The cross

sections of the eight coils of the toroid magnet are visible as white circles; the chambers are located

inside and between them. The diameter of the barrel is approx. 22 m

(3) the multiple scattering in the chambers and the structural material traversed by

the particle.

1. The accuracy achieved for an individual drift tube in the absence of a magnetic

ﬁeld is depicted in Fig. 11.25 as a function of the distance between the track and

the wire. The working gas mixture for these measurements was 91%Ar + 4%N

5%CH

at 3 bar abs., which allows an r.m.s. measuring accuracy of approx. 75

mto

be reached for drift distances above 3 mm. For smaller drift distances the accuracy

quickly becomes worse as the primary ionization ﬂuctuation dominates. In view

of the problems of ageing in the high ﬂux regime (cf. Sect. 12.6.4), it was later

decided to go to 93%Ar + 7%CO

at 3 bar abs. The unfavourable dependence of

drift velocity on the E-ﬁeld of this gas had to be accepted. The resolution achieved

with the new mixture is shown in Fig. 7.19.

2. The geometrical integrity of each chamber is guaranteed by glueing the tubes

onto the frames. During the experiment, a laser calibration system will follow

the internal elastic deformations caused by changes of temperature or stress. For

the measurement of momentum it is important to know the relative positions of

the three chambers that together determine the sagitta of the particle track. For

388 11 Existing Drift Chambers – An Overview

Fig. 11.25 Measuring

accuracy of an individual drift

tube, without a magnetic

ﬁeld. Full circles:

measurements; open circles:

simulation [RIE 00]

100

110

120

130

02468101214

this purpose, the chambers of this ‘tower’ are interconnected by laser rays seen

by position-calibrated detectors on the chambers.

3. The mechanical structures for the support of the magnet, the chambers, and

all the other detectors of the ATLAS experiment are almost all made of aluminium.

They represent a mass of material which causes a lot of Coulomb scattering in ad-

dition to that of the chambers proper. It may be quite different from one track to

another; on average it causes relative transverse momentum measurement errors

typically between 2 and 3%. The aluminium of the chambers themselves

contributes approx. 1%.

The overall momentum-measuring accuracy of the muon spectrometer is calcu-

lated to be roughly

= 0.1atp

= 1 TeV/c, dominated by (1) and (2), and

0.03 at p

= 10 GeV/c, dominated by (3).

11.5.6 A Large TPC System for High Track Densities

The NA49 experiment at the CERN SPS accelerator is dedicated to the study of

hadron production on ﬁxed targets in proton-proton, proton-nucleus, and nucleus-

nucleus collisions [AFA 99]. A system of 4 large TPCs with a total sensitive volume

of 33 m

is the main instrument for tracking and dE/dx measurement. The incom-

ing beam passes between them, compare Fig. 11.26. Central lead-lead collisions at

the beam momentum of 158 GeV/c per nucleon can produce more than a thousand

tracks in one event.

Of the four TPC modules, the two smaller ones are operated inside strong and

not very homogeneous vertical magnetic ﬁelds, the two larger ones outside. The

drift is vertical, the sensitive wire chamber end plates horizontal. The wire grids are

11.5 Large Cylindrical Drift Chambers of Type 2 389

Fig. 11.26 Top view of the four TPCs; the two ‘vertex chambers’ inside the magnetic ﬁeld are on

the left, the two ‘main chambers’ to both sides of the beam outside the magnetic ﬁeld are on the

right. A central Pb + Pb event is depicted in a slice only 7 mm high around the beam plane, in order

to be able to distinguish tracks in the projection

designed as in the ALEPH TPC but have a reduced distance between pads and sense

wires. Their total area amounts to 30 m

. Signals are read out only from the cathode

pads of which there are 182 000. A maximum of 234 pad rows is available for the

measurement of fully recorded tracks, but the actual number of signal clusters lies

approximately between 20 and 160, depending on track overlap and kinematics in

the given acceptance region. The authors estimate that an over-all position precision

of about ±0.20 mm has been achieved for isolated tracks, varying with the drift

length between ±0.15 and ±0.28 mm.

The separation of neighbouring tracks is a major challenge in this experiment.

The diffusion properties of the gas and the other detector parameters have been

designed with the aim to reach a charge distribution within typically 5 mm FWHM.

The authors show that track pairs with average distances above 1.6 cm are fully

detected, those with distances of 1.0 cm are detected with 50% efﬁciency. Further

improvements seem to be in reach. It is remarkable that the TPC technique can be

extended to record tracks in such high particle densities.

This was made possible by an increased density of cathode pads and the associ-

ated electronic channels. The entire front-end electronics up to the digital outputs is

mounted on the TPC readout chambers, the pad widths being not more than 3.5 or

5.5 mm.

The high particle density is a challenging environment also for particle identiﬁ-

cation using dE/dx measurements. It is difﬁcult to control the pulse height variation

with drift time which is partly caused by the baseline shift of the electronic signals,

partly by the widening of the measured charge clusters at the required threshold,

but only to a smaller degree by electron attachment. It is also not easy to correct for

electronic cross-talk at these high particle densities. Despite of these severe condi-

tions, the authors are able to report an r.m.s. resolution for the dE/dx measurement

390 11 Existing Drift Chambers – An Overview

= 38%/

√

where the number N

of signal clusters lies between 20 and 160.

The method of charge estimation applied is the one of the truncated mean at the

50% level. The relativistic plateau is IR = 56% above minimum ionization.

Using the ratio

/IR as a measure for the ability of particle identiﬁcation one

can say that the NA49 TPC system reaches a similar level of identiﬁcation ability as

the detectors listed in Tables 10.3 and 10.4, once the losses of signal clusters from

track overlap and acceptance limitation are taken into account.

11.6 Small Cylindrical Drift Chambers of Type 2

for Colliders (Vertex Chambers)

Vertex detectors are needed for the measurement of the (primary) interaction vertex

or any secondary vertices that may occur in an event from the decay of short-lived

particles. The charm, bottom and

particles have lifetimes of the order of 10

−13

−12

s, so the tracks from the secondary vertices will miss the primary vertex by

correspondingly small distances, which are of the order of 30 to 300 μm. Ideally,

vertex chambers should have even better measurement accuracies.

This is the ﬁeld where the ﬁne-grained solid-state particle detectors with their su-

perb spatial resolution are often a better choice than gas drift chambers. These have

in their favour the simplicity of a proven technique and a somewhat better resis-

tance against high-radiation background. Also they are more easily extended over

large sensitive areas. Being closer to the interaction point they can often be built

smaller and with shorter wires than the large axial chambers, so that the wire posi-

tion can be extremely well deﬁned. This is a prerequisite for the high measurement

accuracy achieved in the vertex chambers. When measuring the position of a vertex

inside the vacuum tube, multiple scattering in the wall of the tube is often a limiting

factor, which depends on the momentum of the extrapolated particle as well as on

the tube radius and the thickness of the wall.

In order to reach the best possible point-measuring accuracy and double-track

resolution, the sense wires have to be as close as possible to each other (cf. the

discussion in Sect. 11.8). The limit imposed by electrostatic stability depends on

the wire length and the detailed electrostatic pattern, and it becomes more severe at

higher gas pressures because of the corresponding higher voltages.

To ﬁnd the correct gas, one has to balance the consequences of a slow ‘cool’

gas (see Sects. 2.2.4, 12.1) – slower electronics, high drift ﬁeld with low diffu-

sion, sensitivity to ﬁeld inhomogeneity through unsaturated drift velocity – with

the consequences of a fast ‘hot’ gas – faster electronics, lower sensitivity to ﬁeld,

temperature and pressure variations.

These questions have been studied in detail for the more recent vertex detec-

tors. The chambers listed in Table 11.3 are again intended to represent the different

choices in the ﬁeld. They are all axial drift chambers of type 2. The ones that reach

the highest precision are operated with slow cool gas at an elevated pressure. Hayes

has compared vertex drift chambers for the LEP and SLC detectors [HAY 88]. The

11.6 Small Cylindrical Drift Chambers of Type 2 for Colliders 391

Table 11.3 Some small drift chambers of type 2 (vertex chambers)

Name of experiment UA2 MARK J MARK II OPAL MAC ALEPH

Name of chamber JVD TEC DCVD VCH straw-VC ITC

Method of z measurement Charge division – – Time difference – Time difference

+stereo

Reference [BOS 89] [AND 86] [ALE 89] [CAR 90] [ASH 87] [DEL 90]

[AND 88a] [DUR 90] [OPA 91] [NEL 87] [BAR 89]

[AND 88b]

Geometry

Outer/inner radius (cm) 13/4 12/5 17/5 22/10 9/3.5 28/13

Length (cm) 100 58 55 100 43 200

No. of wires 1232 2750 1800 324 5136

No. of sense wires 208 168 190 648 324 960

No. of sense wires per cell (6) 13 14 19 12(6) 1 1

Sense-wire diameter (μm) 25 20 20 20 30 30

No. of measured points per track 13 14 19 18 6 8

Radial distance between sense wires (mm) (6) 6.4 2.4 2.9 5.0 (5.3) 6–10 11–19

Max. drift distance (cm) 2.5 1.3–2.6 3 0.8–1.9 0.35 0.47–0.65

Stereo angle (

◦

) 000400

Radial tilt of sense wire plane (

◦

) 0 ± 4150 – –