Blum W., Riegler W., Rolandi L. Particle Detection with Drift Chambers
Подождите немного. Документ загружается.
412 11 Existing Drift Chambers – An Overview
[RIE 00] W. Riegler et al., Resolution limits of drift tubes, Nucl. Instr. and Meth. in Phys.
Res. A 443, 156-163 (2000), W. Riegler et al., Front-end electronics for drift tubes in
high-rate environment, Nucl. Instr. and Meth. in Phys. Res. A 446, 555-559 (2000),
M. Aleksa et al., Rate effects in high-resolution drift chambers, Nucl. Instr. and Meth.
in Phys. Res. A 446, 435-443 (2000), M. Deile et al., Dependence of drift tube per-
formance on the anode wire diameter, Nucl. Instr. and Meth. in Phys. Res. A 449,
528-535 (2000)
[SHI 88] A. Shirahashi et al., Performance of the TOPAZ time projection chamber, IEEE
Trans. NS-35, 414 (1988)
[SIE 90] P. Siegrist, CERN, private communication, September 1990
[SLD 85] SLD Design Report, SLAC Report 273 (revised 1985)
[SNI 88] F. Snider et al., The CDF Time Projection Chamber system, Nucl. Instrum. Methods
Phys. Res. A 268, 75 (1988)
[TOK 90] W.H. Toki, Review of straw chambers, in Proceedings of the 5th Int. Conf. for Col-
liding Beam Physics, Novosibirsk (USSR), March 1990; also as: SLAC preprint
SLAC-PUB-5232 (1990)
[VIL 86] F. Villa (ed.), Vertex Detectors, Proceedings of a Workshop for the INFN Eloisatron
Project, held September 1986 (Plenum, New York London 1988)
[WAG 81] A. Wagner, Central detectors, Phys. Ser. 23, 446 (1981)
[WAL 71] A.H. Walenta, J. Heintze and B. Sch
¨
urlein, The multiwire drift chamber, a new type
of multiwire proportional chamber, Nucl. Instrum. Methods 92, 373 (1971) (received
27 November 1970)
[WAL 78] A.H. Walenta, Left–right assignment in drift chambers and MWPC’s using induced
signals. Nucl. Instrum. Methods 151, 461 (1978)
[WAL 79] A.H. Walenta, The time expansion chamber and single ionization cluster measure-
ment, IEEE Trans. Nucl. Sc. NS 26, 73 (1979)
[WIL 86] H.H. Williams, Design principles of detectors at colliding beams, Annu. Rev. Nucl.
Part. Sc. 36, 361 (1986)
[YOU 86] C.C. Young et al., Performance of the SLD central driftchamber prototype, IEEE
Transactions Nucl. Sc. 33, 176 (1986)
Chapter 12
Drift-Chamber Gases
Gas is the medium in which the processes of ionization, drift and amplification
develop. We have discussed these processes in previous parts of this book. Here we
want to present some additional practical information relevant to the choice of gas
mixtures and their pressure, and relevant to the impurities that can be tolerated in
the gas.
12.1 General Considerations Concerning the Choice
of Drift-Chamber Gases
There are many gas mixtures that are successfully used in drift chambers. The tables
in Chap. 11 contain a representative collection. Often we find the noble gas argon
mixed with some quenching organic molecular gas. But purely organic mixtures
are also in use. These gases fulfil the basic requirement that the electron lifetime is
sufficiently long and that a stable amplification process exists.
An important choice has to be made with regard to the drift velocity and its de-
pendence on the electric field. A highly preferable situation arises if the drift velocity
varies little with the field, for then the coordinate measurement is less dependent on
the unavoidable field gradients, and on the pressure and temperature variations of
the gas. This happens at the maximum of the velocity–field relation (‘saturated drift
velocity’). Figure 12.1 shows the drift velocity of argon–isobutane mixtures, which
have this desirable property over a relatively long range of E field.
At these high velocities – typically 5 cm/μs – a fast time measurement is a ne-
cessity. If one wants to relax this requirement, one can reduce the drift field, and the
drift velocity is no longer saturated, with the consequence that it depends on space
charge and other changes of the drift field, thus requiring very careful calibrations.
The ‘dilemma of the Lorentz angle’ comes into sight here (see Sect. 11.3.2). If
the electric and magnetic fields are orthogonal one may wish to achieve a small drift
velocity at a high E field in order to keep the Lorentz angle small, cf. (2.7), (2.11).
W. Blum et al., Particle Detection with Drift Chambers, 413
doi: 10.1007/978-3-540-76684-1
12,
c
Springer-Verlag Berlin Heidelberg 2008
414 12 Drift-Chamber Gases
Fig. 12.1. Drift velocity as a function of the drift field, for various argon–isobutane mixtures
[BRE 74]
AhighE field at small electron energy must be the goal in the interest of a
low diffusion width; cf. the discussion following (2.63). The last two requirements
combine well in the ‘cold’ gas mixtures based on carbon dioxide or dimethylether.
In these, the random electron velocities and the drift velocity are much lower than
in the ‘hot’ gases at the same field strength. The cold gases have the disadvantage
that an operating point must be found in the unsaturated situation. High E fields are
more difficult to handle, since spurious discharges are a threat to the reliability of a
chamber. High E fields also cause electrostatic wire displacements; these are larger
for greater wire lengths.
In order to give an overview of achievable diffusion limits we reproduce in
Fig. 12.2 the characteristic electron energies calculated by Paladino and Sadoulet
[PAL 75] for the pure gases Ar, CH
4
and CO
2
, and for various argon–isobutane mix-
tures. The width of the diffusion after 1 cm of drift, calculated according to (2.63),
is shown in Fig. 12.3.
In TPCs, where the E and B fields are parallel, there is an advantage in having a
high value of
ωτ
, or a high product of the B field and the mobility (cf. Sect. 2.1).
One reason is the enormous suppression factor for the transverse diffusion constant
(2.72), another has to do with the track distortions caused by imperfections in the
uniformity of the magnetic field. Since in a good solenoid the inhomogeneities are
predominantly radial (not azimuthal), the distortions of the curvature of radial tracks
are suppressed by one power of
ωτ
; see (2.6).
Argon–methane mixtures with small CH
4
concentrations can be used in the sat-
urated situation of the drift velocity, and there they exhibit large values of
τ
. It can
be seen in Table 11.4 that such gas mixtures have actually been chosen for all the
large TPCs.
12.1 General Considerations Concerning the Choice of Drift-Chamber Gases 415
Fig. 12.2. Characteristic electron energies calculated and reported by [PAL 75]. (a) for various
mixtures of argon–isobutane; (b) argon, carbon dioxide and methane
416 12 Drift-Chamber Gases
Fig. 12.3 Diffusion width after 1 cm of drift, according to (2.63), using calculated electron energies
[PAL 75]
12.2 Inflammable Gas Mixtures
Most of the quenching gases used in drift chambers can burn in air. Thus in large
systems they may represent a security risk. We would like to know under what
circumstances and in what concentrations they are dangerous.
A gas mixture that contains some oxygen is ‘inflammable’ if a flame can exist and
propagate in it. (Confusingly enough, this same mixture is also called ‘flammable’
in the field.) The fact is established in a standardized vertical tube with ignition
from below. Whether a flame develops or not depends on the nature of the gas com-
ponents, their temperature and pressure, and in particular on their concentrations.
By varying the concentrations of a gas mixture, one finds non-inflammable com-
binations whose content of combustible or oxygen is not enough to make a flame.
Therefore there are lower and upper limits of concentration which mark the region
of inflammability.
The simplest mixture is a two-component gas, e.g., methane in air. At normal
temperature and pressure it is inflammable between approx. 5 and 15% CH
4
by
volume in the total, whilst the stoichiometric ratio for complete combustion is 9.5%.
There is no simple principle from which to derive such limits: they depend on details
of the flame dynamics. Let us note that these limits are markedly different when the
12.2 Inflammable Gas Mixtures 417
flame proceeds with the ignition on top rather than at the bottom. The monograph
by Lewis and von Elbe [LEW 61] may be consulted concerning this subject. More
recent determinations of the limits of inflammability such as the ones conducted at
CERN [CER 96] resulted in more conservative limits. This ‘progress’ is achieved
by a new judgement about the development of the test flame, which may appear in a
variety of forms – from a flame that propagates clearly to the top end of the test tube
to flames that do not clearly detach from the spark or do so but burn out halfway up.
Also, one is now more careful about gas purity as even small contaminations may
influence the result. In Table 12.1 we present the classical inflammability limits as
well as the new limits from [CER 96] for comparison.
When an inert gas is added to the previous mixture, we have a three-component
mixture, e.g., methane, argon, and air. Now the inflammability limits also depend
on the argon concentration. If there is too much argon, then there will be no flame,
irrespective of the methane/air ratio. The three-component gas mixtures are con-
veniently described by the diagrams in Fig. 12.4: a particular mixture is given by
the three volume concentrations c
1
,c
2
, and c
3
in the total mixture, which add up to
unity; they are represented by a point in the unilateral triangle whose three coordi-
nates are measured parallel to its sides (the sum of the three coordinates is always
equal to the length of one side).
Results of flame tests are now recorded in a diagram that delineates the region
of inflammability. Examples from measurements by the Physikalisch-Technische
Bundesanstalt, made at the request of CERN, are shown in Fig. 12.4.
Chamber gas does not contain oxygen, and inflammable mixtures will come
about by unintentional contact with air. A safe chamber gas is one that produces
Table 12.1 Limits of inflammability of some gases and vapours in air that are of interest for particle
detectors, from two different sources (vol. % of total mixture)
Compound Formula Limits of inflammability of gas or vapour
[LEW 61] [CER 96]
Lower Upper Lower Upper
Methane CH
4
5.3 15.0 4.4 16.9
Ethane C
2
H
6
3.0 12.5 2.4 14.6
Propane C
3
H
8
2.2 9.5 1.8 10.4
Butane C
4
H
10
1.9 8.5 1.4 8.9
Isobutane i-C
4
H
10
1.8 8.4 1.55 8.4
Pentane C
5
H
12
1.4 7.8
Isopentane i-C
5
H
12
1.4 7.6
2,2-dimethylpropane C
5
H
12
1.4 7.5
Hexane C
6
H
14
1.2 7.5
Ethylene C
2
H
4
3.1 32.0
Propylene C
3
H
6
2.4 10.3
Methylalcohol CH
3
OH 7.3 36.0
Isopropylalcohol C
3
H
7
OH 2.0 12.0
Diethylether C
4
H
10
O 1.9 48.0
418 12 Drift-Chamber Gases
Fig. 12.4a–d Diagrams to describe three-component gas mixtures by their volume concentrations
c
1
,c
2
and c
3
(see text). Limits of inflammability for four different mixtures containing air
a non-inflammable mixture with any amount of air. Looking, for example, at
Fig. 12.4a, we see that all the combinations of methane–argon–air concentrations
that have this property lie below the tangent that connects the region of inflamma-
bility with the point P (c
1
= 0,c
2
= 0, and c
3
= 1). All mixtures with a fixed
concentration ratio c
1
/c
2
lie on a straight line through P, which intersects the op-
posite side at the value c
2
– the methane-argon mixture is safe in this sense if the
methane concentration is not greater than 9% (or 6%, see later). Some safe concen-
trations of combustible gases are listed in Table 12.2.
In Table 12.3 we also list the new limiting safe mixtures recognized by the CERN
safety code, which are based on the more recent measurements from [CER 96]. A
comparison of the last two columns indicates a very small difference between the
behaviour of butane with different impurities, surely not very important in practical
applications.
A special point of interest is the tip of the region of inflammability that belongs
to the smallest air concentration. One might expect that it always represents the
stoichiometric concentrations. But this is not so – it is often displaced from the stoi-
chiometric ratio towards lower concentrations of the component that has the higher
diffusion [LEW 61]. Here is one more hint that no simple theory will predict the
safe concentrations.
12.2 Inflammable Gas Mixtures 419
Table 12.2 Standard limiting safe mixtures in air. A mixture of a combustible gas (component 1)
and an inert gas (component 2) is considered inflammable in air if, with a suitably chosen admixture
of air (component 3), the three-component mixture can be ignited. The table quotes concentrations
c
1
of component 1 at and below which no admixture of air can make a flame. The corresponding
concentrations c
2
are also given
Combustible c
1
(%) Inert gas c
2
(%) Reference
a
gas
CH
4
13/9 He 87/91 [BUR 48, ZAB 65]
i-C
4
H
10
2.7 Ne 97.3 [SCH 92]
CH
4
9 Ar 91 [BUR 48]
C
2
H
6
2 Ar 98 [SCH 92]
n-C
5
H
12
2 Ar 98 [SCH 92]
CH
4
23 CO
2
77 [BUR 48, ZAB 65]
C
2
H
6
12 CO
2
88 [BUR 48, ZAB 65]
C
3
H
8
11 CO
2
89 [BUR 48, ZAB 65]
n-C
4
H
10
9.7 CO
2
90.3 [BUR 48, ZAB 65]
i-C
4
H
10
9.3 CO
2
90.7 [BUR 48]
n-C
5
H
12
7.5 CO
2
92.5 [BUR 48, ZAB 65]
n-C
6
H
14
6.6 CO
2
93.4 [BUR 48, ZAB 65]
a
Where two references are given they are in agreement, except for CH
4
/He.
Inflammable gas mixtures have been used in numerous particle experiments. The
required safety precautions may be a burden on the operation of the experiment,
and it is advantageous to avoid such mixtures. However, for streamer chambers in
particular this does not seem to be possible, because most mixtures with the required
absorption power for the ultraviolet light are also inflammable. Special studies of
this problem have been undertaken by Benvenuti et al. [BEN 89] and several other
groups cited there.
Table 12.3 Official limiting safe mixtures between inflammable and inert gases in air, according to
[CER 96]. The table shows volume per cent of the inflammable gas in the total, below which the
two component mixture is considered non-inflammable
Inflammable gas (per cent purity)
Inert gas CH
4
C
2
H
6
C
3
H
8
iC
4
H
10
nC
4
H
10
nC
4
H
10
(% purity) (99.95) (99.4) (99.95) (99.5) (99.90) (99.5)
N
2
(99.995) 9.9 4.94 4.25 4.1 3.55 3.26
CO
2
(99.9) 22.45 9.09 7.95 7.95 6.52 6.45
He (99.996) 11.86 5.45 4.41 4.15 3.58 3.55
Ne (99.990) 9.2 4.37 3.45 3.26 2.74 2.70
Ar (99.996) 6.15 3.05 2.76 2.4 2.28 2.16
SF
6
(99.90) 50.4 20.4 20.4 19.43 16.97
CF
4
(99.7) 33.4 13.0 11.76 9.5 9.28 9.1
C
2
H
2
F
4
(99.7) 11.98 7.14 6.7 5.75 5.8
420 12 Drift-Chamber Gases
12.3 Gas Purity, and Some Practical Measurements
of Electron Attachment
Together with the gases intended to serve in a chamber, there will always be some
contamination, by impurities in the bottles delivered from industry, from outgassing
construction materials in the gas stream or produced in operation, or arriving from
the outside world through membranes that are not entirely tight.
Even minor contamination may have adverse effects on the chamber operation. It
can change the drift velocity (a drastic example is reported in Sect. 12.3.3), or they
may remove drifting electrons by attachment (see Sects. 12.3.1 and 12.3.2). They are
a cause of chamber ageing (Sect. 12.6). But they can also have a beneficial effect:
ionization tracks produced by pulsed UV lasers often depend on some unknown
molecular compound that happens to be in the chamber gas (cf. Sect. 12.4).
We have described in Chap. 2 various mechanisms of negative-ion formation
which may lead to a loss of electrons drifting through a chamber. It is chiefly by
varying the partial pressures of two clean gases that the two-body and three-body
processes can be experimentally distinguished. The values reported in Figs. 2.11 and
2.12a-d were obtained in this way, mostly by workers engaged in atomic-physics
research.
For the purposes of drift chambers it is not always necessary to isolate the exact
process or to know the electron energy. On the other hand one wants to have electron
attachment rates for the multi-component gas mixtures and for drift fields typical
for applications in particle experiments. We collect here some measurements of this
kind, keeping in mind that the detailed interpretation of the attachment process is
not always required.
12.3.1 Three-Body Attachment to O
2
, Mediated
by CH
4
,i-C
4
H
10
and H
2
O
Electron loss rates were measured in various drift-chamber gases with additions of
oxygen, water or methanol, by Huk, Igo-Kemenes and Wagner [HUK 88].
Working initially at 4 bar with Ar (88%) + CH
4
(10%) + i-C
4
H
10
(2%) and with
O
2
concentrations up to 440 ppm, they varied the total gas pressure and the hydro-
carbon and O
2
concentrations. They were able to show that the measured electron
attenuation rate R
tot
associated with oxygen could be understood as a superposition
of rates according to (2.77):
R
tot
=
∑
R
i
;
R
i
= k
(i)
1
N(O
2
)N(X
i
) ,
where N(O
2
)istheO
2
concentration and the N(X
i
) are the concentrations of the
other gas components. (The rate proportional to the square of N(O
2
) is negligible at
12.3 Gas Purity, and Some Practical Measurements of Electron Attachment 421
Table 12.4 Three-body electron attachment coefficients for O
2
and third body CH
4
determined
by [HUK 88] at fixed reduced drift fields E/P in drift-chamber gas composed of Ar(90%) +
CH
4
(10%) + O
2
(200 ppm) and Ar(80%) + CH
4
(20%) + O
2
(200 ppm)
E/P C(O
2
,CH
4
)
(V/cm bar) (μs
−1
bar
−2
)
100 167 ± 32
138 191 ± 30
163 176 ± 20
200 149 ± 27
250 146 ± 26
these low O
2
contaminations.) The various contributions could be separated, and the
three-body rate coefficients of CH
4
,i-C
4
H
10
and Ar were determined for a practical
range of drift fields. The contribution of Ar was zero inside the errors, as expected.
The results are quoted in Tables 12.4 and 12.5. The coefficient C(O
2
, X), measured
in units of (μs
−1
bar
−2
), is related to k
1
of (2.77), measured in units of (cm
6
s
−1
),
by the relation
C(O
2
,X)=
6.02×10
20
22.4
2
bar
−2
cm
6
10
−6
k
1
at N.T.P.
It should be noted that the three-body rate coefficients C(O
2
,CH
4
) and C(O
2
,
i-C
4
H
10
) given in Tables 12.4 and 12.5 for specific values of the reduced electric
field represent averages over the electron energies in the range of the gas mixtures
used. For details the reader is referred to the article by Huk et al. [HUK 88], which
is also an excellent entry point to the literature.
The authors then investigated the influence of water and methanol. In the absence
of oxygen, neither 0.3% H
2
O nor 0.1% CH
3
OH in Ar(90%) + CH
4
(10%) caused
any electron absorption outside the errors. But the situation changed when a small
amount (200 ppm) of O
2
was added to the argon–methane mixture. It turned out
that 0.1% of H
2
O is as efficient as 10% of CH
4
to mediate electron attachment to
oxygen under the operating conditions investigated. 500 ppm of methanol did not
change the attachment rate to oxygen inside the errors.
Table 12.5 Three-body electron attachment coefficients as in Table 12.4, but for O
2
and i-C
4
H
10
in gas composed of Ar(90%) + CH
4
(10%) + O
2
(200 ppm) + i-C
4
H
10
(0 to 4%)
E/P C(O
2
,i-C
4
H
10
)
(V/cm bar) (μs
−1
bar
−2
)
100 3240 ± 350
138 2819 ± 300
163 2490 ± 300
200 1970 ± 300
250 1570 ± 250