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

Подождите немного. Документ загружается.

372 11 Existing Drift Chambers – An Overview

Fig. 11.6a,b A drift cell of the CDHS chambers. (a) sense wires (small circles)andﬁeldwires

(large circles) with their potentials in kV; some distances are also indicated in mm; (b) equipoten-

tial lines, delineating sensitive areas for one quarter of the drift cell, as indicated in (a)

Fig. 11.7 Cross section of a drift-chamber module of the DELPHI Muon Identiﬁer; the delay line

is incorporated in the cathode

11.4 Planar Drift Chambers of Type 1 373

Fig. 11.8 A drift cell of the DC1 chambers in the HELIOS experiment; dimensions in mm

double-track separation. The solution adopted involves a slow gas which in turn

needed small drift cells in the high-multiplicity, high-rate environment. We see the

drift cell in Fig. 11.8. It must provide a very homogeneous ﬁeld because the gas is

not saturated; therefore in each drift cell the eight sense wires are surrounded by a

large number of ﬁeld wires. In an attempt to reduce the inﬂuence of the drift-time

variations, some of the ﬁeld wires were given such potentials that only a fraction

of the ionization along the track was collected on the sense wires. The whole stack

consists of six chambers with three different wire orientations. The performance

ﬁgures reveal that the outstanding measuring accuracy of 60 μm per wire, reached

with the prototype, worsened and became 120 μm in the real experiment where the

presence of many tracks deteriorated the critical ﬁeld conditions.

The muon chambers of the L3 experiment are the result of a colossal effort

to measure the momenta of fast muons with an accuracy of at least

4 ×10

−4

GeV/c

−1

, in cylindrical geometry. In the experiment, the radial space be-

tween 2.5 and 5.4 m was allocated to this task, inside a magnetic ﬁeld of 0.5 T. The

principle of the solution is sketched in Fig. 11.9; one observes that for the optimal

measurement of momenta in the magnetic ﬁeld, the chambers are concentrated at

the inner and outer edges, and in the middle of the radial space (cf. the discussion in

Sect. 8.2.6). Their local measuring accuracy is increased by sampling each track N

times, thus gaining a factor of 1

√

N over the accuracy of a single wire – provided,

the relative wire positions are sufﬁciently well known. For this purpose, the sense

wires have been pulled over optically ﬂat edges and are supported in the middle in

a controlled fashion. We show in Fig. 11.10 the drift cell of the chamber type in the

middle; it has N = 24 sense wires, 0.9 cm apart and 5.5 m long. Under operating

conditions the Lorentz angle is 19

◦

374 11 Existing Drift Chambers – An Overview

Fig. 11.9a,b Arrangement of

the muon chambers in the L3

experiment. (a) view along

the axis; (b) side view

In order to make full use of the local accuracy, the three chambers traversed by

a fast muon must be in well-known and constant relative positions. This has been

achieved to an accuracy of 30μm, or 10

−5

of the size of the mechanical structure.

This combination of size and positioning accuracy was unprecedented at the time in

high-energy physics experiments.

The disk-shaped forward tracking chambers for the H1 experiment are in our

collection on account of their circular symmetry and their radial sense wires. But

they are also interesting because, in addition to measuring tracks, they serve for the

detection of single photons created by transition radiators immediately upstream.

For this double function the gas was chosen to contain a sizeable fraction of xenon

with its short absorption length for soft X-rays.

The three-dimensional schematic drawing of Fig. 11.11 makes the 48 wedge-

shaped drift cells visible. Each one has at its centre the radial plane of the eight

radial sense wires, which are interspersed with ﬁeld wires, and on both sides the

cathode planes. So at the outer radius the drift space is larger than at the inner one

in the ratio of these radii; and in order to maintain a uniform drift ﬁeld in the cell,

the cathode is divided into axial strips on such radially increasing potentials that

the electric ﬁeld is everywhere the same and orthogonal to the sense-wire plane.

Suitable ﬁeld degraders close the volume of every cell. The drift-time coordinate

Fig. 11.10 One drift cell of a

middle muon chamber of the

L3 experiment

11.4 Planar Drift Chambers of Type 1 375

Fig. 11.11 Schematic blow-up of the radial-wire drift chamber of the H1 experiment

is in the azimuthal direction, thus appropriate for a measurement of the particle

momentum in the axial magnet. (The inﬂuence of the magnetic ﬁeld on the drift

directions has been neglected here, cf. Sect. 11.3.2.) The radial track coordinate is

obtained by charge division on pairs of wires. Similar chambers had previously been

constructed for the CDF experiment at Fermilab.

11.4.3 Type 1 Chambers without Field-Shaping Electrodes

The drift ﬁeld of a conventional drift chamber is established by conducting sur-

faces (ﬁeld wires, metal strips etc.) at suitable potentials. But according to an idea

of Allison et al. [ALL 82a], the drift ﬁeld can also be established by deposition

of electrostatic charge on the insulating surfaces that delimit the gas volume. They

showed that in a large ﬂat plastic box one can get a very homogeneous ﬁeld just with

one proportional wire in the middle and two small cathode strips along the parallel

376 11 Existing Drift Chambers – An Overview

Fig. 11.12 A chamber without ﬁeld-shaping electrodes

edges (Fig. 11.12). The deposition of charge is straightforward if the insulating sur-

faces are backed (on the outside) by grounded plates parallel to the surface. Looking

at Fig. 11.13a, the ﬁeld lines are as indicated at the moment that HV is applied to

the sense wire. Any positive charge from the wire ampliﬁcation will travel along the

ﬁeld lines and will fall onto the insulator and stay there to subsequently change the

ﬁeld conﬁguration.

This process continues until there are no more ﬁeld lines pointing at the insu-

lator, i.e. the ideal ﬁeld conﬁguration is reached (Fig. 11.13b). As long as there

is any charge ampliﬁcation going on at the wire, the ideal ﬁeld conﬁguration will

automatically be reached. It should be noted that for a thin chamber, the charge

on the wire must ﬁnally be equal to the charge on the insulating surface. This im-

plies that the gas ampliﬁcation factor, which is determined by the amount of charge

on the wire, cannot be altered without changing the drift ﬁeld in the volume. The

problematic side of this method is in the stability of the ﬁeld. Although the positive

charges are easily brought into place, they are not so easily removed following small

ﬂuctuations – these may be due to local discharges, bursts of particles, deposits on

the wire, etc., and are probably unavoidable. Allison et al. consider insulators with

a non-zero conductivity; these would create a dynamic equilibrium, which would

become a function of the rate of charge deposition.

On the other hand the new principle makes it easy to construct very simple drift

chambers. Franz and Grupen [FRA 82] have described the properties of a ﬂat, round

chamber (1.4 cm × 50 cm diam) with only one short (1.4 cm) wire in the centre.

Fig. 11.13a,b Field

conﬁguration (a) during the

process of charge deposition;

(b) ideal ﬁeld conﬁguration

11.5 Large Cylindrical Drift Chambers of Type 2 377

Becker et al. [BEC 82] have reported measurements with long and thin (20 cm ×

1.5 cm) plastic tubes, which could be bent. They have also discussed other applica-

tions of the new method. D

orr et al. [DOE 85] analyzed the behaviour of a circular

chamber operating in the limited streamer-mode.

11.5 Large Cylindrical Drift Chambers of Type 2

Spanning a large cylindrical volume with wires parallel (or approximately parallel)

to the axis requires strong end plates, which hold the wires, carry the combined

pulling force, and keep them in accurate positions. Usually the end plates, together

with the outer and inner cylinder mantles, also form the envelope for the gas.

The wire feedthroughs that transmit the high voltage and the signals are essential

elements of the construction.

11.5.1 Coordinate Measurement along the Axis – Stereo Chambers

Where charge division and time-difference methods were found insufﬁcient, the

idea of varying the wire directions throughout the volume has been widely used.

The equivalent of a stack of type 1 chambers with differing wire orientations is

a (type 2) ‘axial stereo chamber’. In a cylindrical chamber with many concentric

layers of drift cells, a stereo layer can be formed by rotating the end points on

one side of the wire that belong to this layer by the small ‘stereo angle’

and

by keeping the end points ﬁxed on the other side. The cylindrical surface of this

wire layer turns from a cylinder mantle into a hyperbolic surface. The azimuthal

wire position becomes a linear function and the radial wire position a quadratic

function of z to the lowest order of

z counting from the middle of the chamber. This

inhomogeneity distorts the electric ﬁeld to some extent. The achievable precision

z is equal to the measurement accuracy dr

achieved in the azimuthal direction

divided by the stereo angle. Hence there is interest in making

as large as the

increasing inhomogeneities permit. In practice, stereo angles are built to plus and

minus a few degrees, alternating with zero degrees in successive layers.

Obviously a z measurement can also be provided outside the cylindrical wire

chamber with a type 1 chamber. In some experiments special ‘z chambers’ have

been built for this purpose.

11.5.2 Five Representative Chambers

with (Approximately) Axial Wires

Table 11.2 shows ﬁve chambers selected in the spirit described at the beginning of

this chapter. In quoting performance ﬁgures we have indicated (see footnote 1 of the

table) whether they are from early or later stages of the development of the detector.

378 11 Existing Drift Chambers – An Overview

Table 11.2 Some large cylindrical drift chambers of type 2

Name of experiment Mark II OPAL ARGUS SLD CDF

Name of chamber New CDC Jet-chamber Drift-chamber CDC CTC

Method of internal z-measurement Stereo Charge division Stereo Charge division Stereo

Reference [ABR 89] [FIS 89] [ALB 89] [YOU 86] [BED 88]

[OPA 91] [SLD 85] [BER 91]

Geometry

Outer/inner radius (cm) 152/19 183/25 172/30 96/24 132/31

Length (cm) 230 320–400 200 180 321

No. of wires 36936 30528 37760 36504

No. of sense wires 5832 3826 5940 5120 6156

No. of sense wires per cell 6 159 1 8 12; 6

Sensewirediameter(μm) 30 50 30 25 40

Max. no. of measured points per track 72 159 36 80 84

Radial distance between sense wires (mm) 8.3 10 18 5 7

Max. drift distance (cm) 3 3–25 1 3 4

Stereo angle (

◦

) 4 0 2–5 3 3

Radial tilt of sense-wire plane (

◦

) 00 – 0 45

11.5 Large Cylindrical Drift Chambers of Type 2 379

Table 11.2 (continued)

Gas and ﬁelds

Gas (percentage concentration) Ar(89) Ar(88) C

(97) CO

(92) Ar(49.6)

+CO

(10)+CH

OH(3)+i-C

(8)+C

(49.6)

+CH

(1)+i-C

(2)+H

O(0.2)+C

OH(0.8)

Gas pressure (bar) 1 4 1 1 1

Electric drift ﬁeld (kV/cm) 0.7 0.9 1–2 (70%) 1.0 1.35

Magnetic ﬁeld (T) 0.47 0.4 0.8 0.6 1.5

Drift velocity (cm/μs) 5 0.3 4–5 0.9

ωτ

(approx.) 0.4 0.35 < 10.051

Performance (1)

Point-measuring accuracy

(mm) 0.17 (c) 0.12 (b) 0.19 (c) 0.055 (b) 0.2 (c)

0.135 (c)

Charge division

(mm) – 30–40 (b) – 9 (b) –

60 (c)

Double-track resolution:

(mm) 3.8 (c) 2.5 (c) 9 (a) 1 (a) 5 (c)

Accuracy of ionization measurement (%) 7.2 (c) 3.5 (b) 4.5–5.5 (c) –

3.8 (c)

Accuracy of momentum measurement (2) 0.0046 (c) 0.0012 (b) 0.009 (c) 0.0015 (a) 0.001 (c) (3)

(GeV/c)

−1

0.0022 (c) 0.002 (c) (2)

1. The performance ﬁgures are based on: (a) calculation and laboratory tests; (b) prototype and test beam measurement; (c) measurements in the running

experiment.

2. High p

, full track length, chamber alone.

3. Using the constraint of the primary vertex.

380 11 Existing Drift Chambers – An Overview

It is not unusual that the prototype performed better than the full chamber eventually

did in the experiment.

When the momentum-measurement accuracy actually achieved in an experiment

is compared to that expected from the point-measuring accuracy, the number of

points, and the magnetic ﬁeld strength (Sect. 8.5.1), it is often not as good as an-

ticipated. The overall mechanical accuracy, local and global ﬁeld distortions, and

imperfect calibrations all make their contribution to the momentum-measuring ac-

curacy, which sometimes improves only slowly as the apparatus is better and better

understood.

11.5.3 Drift Cells

The minimal wire conﬁguration needed to establish the drift and ampliﬁcation ﬁeld

around the proportional wire (which is at a positive potential) is a group of ﬁeld

wires at a negative potential around it. For a drift chamber of type 2, this elementary

‘cell’ is repeated to ﬁll the whole volume. Figure 11.14 shows, as an example, drift

cells of the ARGUS chamber; the full chamber consists of 5940 cells of this type.

Advantages of this arrangement are the high degree of homogeneity over the full

volume, the ﬁne granularity, which allows the use of single-hit electronics, and the

full track ionization that is employed. On the other hand, the local ﬁeld around the

wire in one cell is not very homogeneous and thus requires a very careful calibration

of the critical relation between drift time and wire–track distance. In Fig. 11.15 we

see the electron drift paths calculated for such a cell type [HAR 84]. They are in the

form of spirals, created by the B ﬁeld parallel to the wires. The contours of equal

arrival times are circles near the sense wire, but they are shaped by the ﬁeld-wire

geometry far from the sense wire. This means in practice that the time–distance

relation for tracks is a function of the orientation of the track with respect to the

drift cell.

The design of the OPAL drift cell goes in the opposite direction (Fig. 11.16).

The basic unit is a wedge-shaped sector that ﬁlls 15

◦

of azimuth over the entire

radial space so that the chamber consists of 24 of them. Each sector contains one

radial plane of sense wires which are interspersed by ﬁeld wires. The sense wires

are actually staggered to resolve the left–right ambiguity. Each sector is limited and

Fig. 11.14 Drift cells of the

ARGUS chamber. The sense

wire is, typically, at 2.9 kV

and the cathode wires are

grounded

11.5 Large Cylindrical Drift Chambers of Type 2 381

Fig. 11.15 Drift paths and

lines of equal arrival times for

a drift cell of the ARGUS

chamber. The time lines are

15 ns apart

separated from its two neighbours by two common radial planes of cathode wires at

such radially increasing potentials as to create a homogeneous drift ﬁeld orthogonal

to the central sense/ﬁeld-wire plane. The inner surface of the enclosing cylinder

carries suitable ﬁeld degraders. This geometry has also been called ‘jet chamber’

geometry. In this design the homogeneity is constructed into each sector, which is

relatively large. Close tracks are resolved electronically by using short pulses and

digital pulse-shape analysis. The dilemma of the Lorentz angle is diminished by a

high gas pressure and a low magnetic ﬁeld.

The cell design of the new MARK II chamber follows a similar line: radial

sense/ﬁeld-wire planes with radial cathode planes in between, but organized in 132

layers, 7.5 cm in radius. This is done in order to be able to create stereo angles in

consecutive layers. There are 972 cells in the chamber. Figure 11.17 shows a cell

with its electron drift trajectories.

Fig. 11.16 A section of the OPAL drift cell with electron drift lines. Upper plane: cathode wires;

lower plane: sense/ﬁeld wires