Hatano Y., Katsumura Y., Mozumder A. (Eds.) Charged Particle and Photon Interactions with Matter - Recent Advances, Applications, and Interfaces

Подождите немного. Документ загружается.

890 Charged Particle and Photon Interactions with Matter

l v

att d e

= τ

(31.13)

An alternative way of expressing the above equation is

l K

att M

[ ]

(31.14)

where

is the electric eld

[M] is the impurity concentration in ppm

is the trapping constant, which has been measured to be (0.15 ± 0.03) ppm cm

/kV for LAr

(Buckley

etal., 1989)

–1

0 10

I/I

20 30 40

100 ppb

200 ppb

1 ppm

5 ppm

10 ppm

50 60

Distance (cm)(b)

–2

Absorption in H

O-doped Xe

I/I

200 ppb

100 ppb

5 ppm

1 ppm

10 ppm

0 10 20 30 40 50

–1

–2

Absorption in O

-doped Xe

0.1

Oxygen

Water vapor

Xe scintillation spectrum

0.01

Absorption coeﬃcient (m

–1

)

130

(a)

140 150

Wavelength (nm)

160 170 180 190 200

Figure 31.9 (a) Absorption coefcients for 1 ppm of water vapor and 1 ppm of oxygen (Watanabe and

Zelikoff, 1953; Watanabe etal., 1953) The xenon scintillation spectrum is superimposed. (b) Scintillation

light intensity as a function of the distance from the light source for various concentrations of water in LXe.

2005a. With permission.)

Applications of Rare Gas Liquids to Radiation Detectors 891

An RGL should in principle be transparent to its own scintillation light due to the scintilla-

tion mechanism through the excimer state,

, as shown in Figure 31.1. However, possible con-

taminants in an RGL, such as water and oxygen at ppm level, considerably absorb scintillation

photons (Watanabe and Zelikoff, 1953; Watanabe etal., 1953). In Figure 31.9a, the absorption

coefcients for vuv light are shown for 1 ppm contamination of water vapor and oxygen. The

absorption spectra of water and oxygen largely overlap with the xenon scintillation spectrum.

Given these absorption coefcients and neglecting the scattering (l

abs

< l

Ray

), the light intensity

as a function of the distance from the light source for various concentrations of the contaminant

is shown in Figure 31.9b and c. Apparently, water is the worst contaminant. Since water tends to

absorb light with shorter wavelengths, only a component with longer wavelengths survives for a

long distance.

Furthermore, removal of radioactive impurities is very important for rare event searches, such

as neutrinoless double-beta decay and dark matter search. A detailed example is described in

Section 31.4.2.

It should be remembered that the energy loss distribution by a relativistic particle (∼GeV/n) with

Z ≤ 2 does not follow a Maxwell distribution but a Landou–Vavilov distribution (Badwar, 1973;

Ahlen, 1980; Shibamura etal., 1987).

31.3 appliCations in aCCelerator physiCs

31.3.1 ionization caloriMeter

The rst use of RGLs in particle physics experiments was in the form of a calorimeter,

which is a

device that measures the energy of high-energy particles or γ rays. A calorimeter must fully absorb

incoming particles and all of secondary particles within its sensitive volume. A RGL ionization

calorimeter was developed as a sampling calorimeter consisting of a multi-parallel-plate electrode

system (passive medium) immersed in the RGL (active medium), as shown in Figure 31.10. LAr,

among RGLs, is the one most often used for this type of detector because it is available in large

volumes due to its cheaper cost. The advantage of the RGL ionization calorimeter is that any con-

guration of the detector can be formed depending on the experiment at will. Furthermore, it is easy

to determine the positional distribution of energy deposition given by incident particles in a detector

by using strip electrodes of different orientations coated on the insulated plate, such as a printed

* Calorimeters are usually classied as electromagnetic or hadron calorimeters. An electromagnetic calorimeter is one

specically designed to measure the energy of particles that interact primarily via electromagnetic interaction, while

a hadronic calorimeter is one designed to measure particles that interact via a strong nuclear force. The same detector

media

and similar readout techniques can be used, but the responses are very different in both cases.

–HV

Figure 31.10 Multilayer conguration for the RGL sampling calorimeter.

892 Charged Particle and Photon Interactions with Matter

circuit board. Then, a very accurate determination of the incident direction, the conversion point of

the

particle, and particle identication can be provided.

The

energy resolution in a sampling calorimeter is determined mainly by sampling uctuation.

Aformula for sampling uctuation was derived from an analogy with Fano’s formalism for the

straggling of the total ionization produced by ionization radiation with a xed energy, and it is

expressed as follows (Hoshi etal., 1982):

d cell

= −













2 1

∆

(31.15)

where

is the total energy deposited in the active medium

is its standard deviation

ΔE

cell

is the average energy loss per unit cell in the calorimeter

is the energy of the incident particles

It is obvious from Equation 31.15 that the energy resolution improves by increasing E

or decreasing

ΔE

cell

. There is a technical limit up to which ΔE

cell

can be decreased. On the other hand, it is pos-

sible to increase E

, namely, by thinner electrodes and thicker gaps of the RGL. This point of view

ultimately leads to the idea of a (quasi-) homogeneous calorimeter. Homogeneous calorimeters are

much larger devices compared with sampling calorimeters. For reducing the volume of the device,

LKr and LXe are more attractive media because of their shorter radiation lengths*; in addition, a

better energy resolution is achievable owing to their lower W-values. The LKr calorimeter is begin-

ning to be used in practical experiments (Fanti etal., 2007; Peleganchuk, 2009), and the study of the

LXe

calorimeter has also started as test devices (Baranov etal., 1990; Okada etal., 2000).

The

disadvantage of RGL ionization calorimeters is their slow time response, that is, the integra-

tion time to collect full charge, which originates from the electrons’ drifting time (t

) in the liquid

gap, is quite long (Figure 31.2). The liquid gap of ordinary RGL ionization calorimeters is a few

mm; therefore, the electron drifting time is usually of the order of μsec. The so-called initial current

measurement, which uses much a shorter integration time than t

, was introduced to rene this weak

point (see Figure 31.11 (Aubert etal., 1991)). There is not a great difference in the energy resolution

between the methods of full integration and the initial current measurement. Hence, the (quasi-)

homogeneous

RGL calorimeter was established.

31.3.2 na48 lkr caloriMeter

A good example of the working of a quasi-homogeneous LKr calorimeter can be seen from the

NA48 experiment performed at CERN

†

using its super proton synchrotron (SPS) accelerator. The

aim of the NA48 experiment was the precise measurement of direct CP violation in the system of

neutral kaons as well as the measurement of the very rare K

and charged kaons decays (Fanti etal.,

1999; Lai etal., 2003; Batley etal., 2009). Neutral kaons are produced at two targets located at

different distances from the detector, thus creating simultaneously a K

and a K

beam for the CP

violation experiment. The experiment is performed to measure the decays of kaons from these two

beams into two neutral or two charged pions. The detection of photons from a neutral pion decay is

achieved by means of the LKr calorimeter. In order to suppress the huge background from

0 0

3→ π

Bremsstrahlung and electron pair production are the dominant processes for high-energy electrons and photons. Radiation

length, X

, is dened as the length that the high-energy electron travels in matter losing its energy by bremsstrahlung to

1/e. A value for 1GeV is usually taken, since the radiation length, X

, becomes almost independent of energy above 1 GeV.

†

European Organization for Nuclear Research.

Applications of Rare Gas Liquids to Radiation Detectors 893

decays, very good spatial, energy, and time resolutions are required. The LKr volume is 125cm deep

(∼27radiation lengths) along the direction of the beam and forms an octagon of about a 5.5m

cross

section. The anodes and cathodes are 40μm thick copper–beryllium ribbons running almost parallel

to the beam (Figure 31.12). They form 13,500 readout cells having a cross section of 2 × 2cm

. Cold

preampliers are directly connected to one end of the anode ribbons (downstream side). In order to

stand the high rate of particles (∼1MHz) from the K

beam and to still provide the good time reso-

lution required for K

tagging, an initial current readout with fast shaping by means of a linear

amplier is used. The shaped pulse has a width of only 70ns, while the drift time in the calorimeter

cell is over 3μs. The performance of the LKr calorimeter was tested and the results were as follows

(in units of GeV, cm, and ns) (Barr etal., 1996; Costantini, 1998; Jeitler, 2002):

= ⊕ ⊕

0 032 0 09

0 0042

. .

X Y

.= ⊕

0 42

0 06 (31.16)

2 5.

I , Q

(a)

200100 300 400 t (ns)

100

(b)

200 t (ns)

(δ)

100 400

(c)

t (ns)

≈t

Figure 31.11 (a) Drift current, I, and integrated charge, Q, versus drift time for ionized electrons.

The maximum drift time is 400 ns. (b) Response of a shaping amplier to a very short current pulse

(δ). (c) Response of a shaping amplier to the current form shown in (a). (From Aubert, B. etal., Nucl.

Instrum. Methods Phys. Res. A., 309, 438, 1991. With permission.)

894 Charged Particle and Photon Interactions with Matter

31.3.3 atlaS lar caloriMeter

The LHC (Large Hadron Collider) experiments at CERN began in 2008. The optimization of the

detector is driven by the requirement to detect new particles, in particular, the Higgs boson* decay-

ing to two photons or to four electrons, in the mass range from 90 to 180GeV. The ATLAS experi-

ment should be capable of measuring the Higgs mass with 1% precision using the calorimeter system

alone. This translates into the requirements of a sampling term of less than

10%

/ E

, associated with

a global constant term of 0.7% and an angular resolution better than 50mrad/ E. The technology

chosen for the ATLAS electromagnetic calorimeter is Pb/LAr sampling with accordion-shaped

electrodes and absorbers. LAr has been chosen because of its intrinsic linear behavior, stability

of the response over time, and radiation tolerance. In addition, the accordion geometry minimizes

inductances in the signal path, allowing the use of the fast shaping (40ns) needed for operation with

25ns bunch intervals between collisions at the LHC.

The ATLAS LAr calorimeter system consists of an electromagnetic barrel calorimeter and two

end-cap cryostats with electromagnetic (EMEC), hadronic (HEC), and forward (FCal) calorime-

ters (Figure 31.13). The barrel, covering a pseudo-rapidity

†

range of │η│ < 1.475, shares its cryostat

with the superconducting solenoid, the calorimeter being behind the solenoid. Each wheel consists

of 16 modules of 1024 lead absorbers interleaved with readout electrodes. Each end-cap, cover-

ing the pseudo-rapidity range of 1.4 < │η│ < 3.2, consists of eight modules (see Figure 31.14b).

To correct for the energy lost in the dead material in front of the calorimeter, both end-caps and

barrel are completed with pre-sampler detectors, covering the

│η│ < 1.8 range. These pre-samplers

are thin layers of argon equipped with readout electrodes, but without an absorber. Figure 31.14a

shows the segmentation in the (η, ϕ) plane and along the longitudinal electromagnetic shower

development in the barrel. The HEC is structured into two wheels: the front, HEC1, and the rear,

* The Higgs boson is the only Standard Model particle that has not yet been observed. Experimental detection of the Higgs

boson

would help explain the origin of mass in the universe.

†

η = −ln tan(θ/2), where θ is the polar angle of the produced particles against beam axis. The azimuthal angle ϕ is mea-

sured

around the beam axis.

2 cm×2 cm cell

Beam

Cathodes

Anodes

Figure 31.12 Cell geometry of NA48 LKr calorimeter. (From Costantini, F., Nucl. Instrum. Methods

Phys. Res. A.,

409, 570, 1998. With permission.)

Applications of Rare Gas Liquids to Radiation Detectors 895

HEC2, with 32 modules each, covering the pseudo-rapidity range of 1.5 < │η│ < 3.2. It consists

of copper-plate absorbers of 25mm (50 mm) thickness for HEC1 (HEC2) interleaved with elec-

trodes. The readout technique (electrostatic transformer) of the HEC splits the gap between two

absorbers into four sub-gaps, each with a width of 1.85mm. The readout granularity is Δη × Δϕ =

0.1 × 0.1 for 1.5 < │η│ < 2.5 and 0.2 × 0.2 for 2.5 < │η│ < 3.2. The HEC is designed to measure

the

jets in the forward direction. The FCal shares the same cryostat with the EMEC and the HEC

Figure 31.13 The ATLAS LAr calorimeter: (1) the electromagnetic barrel; (2) the EMEC end-cap; (3) the

HEC end-cap; (4) the FCal end-cap. (From Aharrouche, M., Nucl. Instrum. Methods Phys. Res. Sect. A., 581,

373,

2007. With permission.)

Towers in sampling 3

Δφ× Δη = 0.0245 ×0.05

Trigger

tower

Δη= 0.0982

Square towers in

sampling 2

Strip towers in sampling 1

(a) (b)

Δφ= 0.0245 × 4

36.8 mm× 4

=147.3 mm

37.5 mm/8= 4.69 mm

Δη= 0.0031

470 mm

4.3X

1.7X

1500

Trigger tower

Δη= 0.1

Δη= 0.025

Δφ= 0.0245

Figure 31.14 (a) Projection of the electromagnetic module showing the longitudinal and lateral seg-

mentation. (b) Schematic view of the electromagnetic end-cap wheel (Aubert etal., 2005; Aharrouche,

2007). (From Aharrouche, M., Nucl. Instrum. Methods Phys. Res. Sect. A., 581, 373, 2007. With permission.)

896 Charged Particle and Photon Interactions with Matter

and covers the pseudo-rapidity range of 3.2 < │η│ < 4.9. It consists of three consecutive modules

along the beam line. The modules consist of a tungsten matrix in FCal2 and FCal3 (copper matrix

in FCal1) housing cylindrical electrodes, consisting of copper rods. The LAr gaps are very small:

0.250, 0.375, and 0.500mm, for FCal1, FCal2, and FCal3, respectively. The readout granularity

is Δη × Δϕ = 0.2 × 0.2. The design of the FCal was constrained by the high radiation level in the

most forward region. The detailed studies of the construction and performance are described in

Aubert etal. (2005).

31.3.4 lxe Scintillation caloriMeter (Meg experiMent)

The MEG experiment searches for a rare muon decay, μ

→ e

γ, which is forbidden in the standard

model. Several new theories beyond the standard model predict the branching ratio of the decay just

below the current experimental limit, that is, 1.2 × 10

−12

(Brooks etal., 1999). The MEG experi-

ment aims a sensitivity of 10

−14

(Mori, 1999), where most predictions are covered. The discovery of

the muon decay is a probe for new physics beyond the standard model. A μ

→ e

γ decay event is

characterized by the clear two-body nal state where the decay positron and the γ ray are emitted in

opposite directions with energies equal to half the muon mass (E = 52.8MeV). While positrons of

this energy are abundant from the standard Michel decay of muons, γ rays with such high energies

are very rare. Therefore, the key requirement for the MEG detector is a high energy resolution

for γ rays, since the accidental background rate decreases, at least, with the square of the energy

resolution

(Mori, 1999).

The

excellent properties of LXe as a scintillator and the developments of new PMTs with high

quantum efciency, fast timing, and their possible operation in LXe motivated the choice of the

detector for the MEG experiment. With the Xe scintillation light detected by a large number of

PMTs immersed in the liquid volume, the detector’s energy resolution is proportional to

where N

is the number of photoelectrons, and was estimated to be better than 1% (1σ) for 52.8MeV

γ rays (Doke and Masuda, 1999). In addition, the γ ray interaction points can also be determined

from the light distribution on each PMT (Mori, 1999). Following the R&D with a small proto-

type (Mihara etal., 2002) and with further improvement of the metal channel-type PMT, a large

LXe prototype detector with 228 PMTs of 2in. was constructed and tested with 10–83MeV γ rays

(Mihara etal., 2004). The detector and associated cooling and purication apparatus, shown sche-

matically in Figure 31.15, were tested at the Paul Scherrer Institute in Switzerland where the nal

MEG experiment is located. A pulse tube refrigerator (PTR) was used for the rst time to condense

Xe gas and maintain the liquid temperature stable over a long period. The PTR, optimized at the

LXe temperature, had a cooling power of 189W (at 165K) (Haruyama, 2002). The prototype was

large enough to test both the reliability and stability of the cryogenic system based on the PTR, and

the

efciency of the purication system required for the calorimeter’s performance.

While

the transmission of the scintillation light is affected by Rayleigh scattering, this process

does not cause any loss of light, and hence does not affect the energy resolution. Two types of cir-

culation systems for Xe purication were developed to improve the light attenuation length. One is

gaseous purication using a heated metal getter, in which the evaporated Xe is brought to the hot

getter system, and then the Xe is recondensed in LXe in the detector (Baldini etal., 2005a). This

method is effective in removing all types of impurities, but is not efcient for water, which was iden-

tied as the major contributor to light absorption, because its vapor pressure at the LXe temperature

is too low. Furthermore, the purication speed is limited by the cooling power. The other method

was developed specically to remove water at a much higher rate (100L/h) by circulating LXe with

a cryogenic centrifugal pump. Water is efciently removed by a lter containing molecular sieves.

With this method, the total impurity concentration was reduced in 5h from 250 to 40 ppb in the total

100 L

LXe volume of the MEG prototype detector (Mihara et al., 2006).

novel calibration technique was developed and applied to the prototype detector. A lattice of

α point sources deposited on thin (100 μm diameter) gold-plated tungsten wires was permanently

Applications of Rare Gas Liquids to Radiation Detectors 897

suspended in the volume and xed at the surfaces of the large vessel containing the LXe (Baldini

etal., 2006). This method was successfully implemented and used to determine the relative quan-

tum efciencies of all the PMTs and monitor the stability of the detector during the experiment.

In the nal MEG detector, the PMT conguration is similar to that in the prototype; hence, the

methods of reconstructing the γ ray energy and interaction points were proven with the prototype

detector. The reconstructed energy distribution for 55MeV γ rays from π

decay is shown in Figure

31.16 (Baldini etal., 2005b). The spectrum is asymmetric with a low-energy tail, caused mainly

by γ ray interactions with materials in front of the sensitive region and by the leakage of shower

Refrigerator

Outer vessel

Inner vessel

LXe

Blue LEDs

241

PMTs

Signal

1 m

Unit: mm

Front view Side view

Figure 31.15 LXe γ ray prototype detector. PMTs are installed on six faces of a rectangular solid holder.

(From

Mihara, S. etal., Cryogenics, 44, 223, 2004. With permission.)

120

100

0 200 400 600 800 1000 1200 1400

×10

Number of photons

Counts

Figure 31.16 Energy spectrum for 54.9MeV γ rays measured by the large prototype detector. (From

Baldini, A. etal., Transparency of a 100 liter liquid xenon scintillation calorimeter prototype and measure-

ment of its energy resolution for 55 MeV photons. In 2005 IEEE International Conference on Dielectric liq-

uids,

June 26–July 1, 2005 (ICDL 2005), Coimbra, Portugal, pp. 337–340. With permission.)

898 Charged Particle and Photon Interactions with Matter

components. The estimated energy resolution at the upper edge is 1.2 ± 0.1% (1σ) (Baldini etal.,

2005b), as shown in Figure 31.16, which is roughly equal to the value obtained by extrapolation of

the small prototype data. The events that remain above 55MeV are known to be due to the mistag-

ging of π

events. The position and time resolutions are evaluated to be 5mm and ∼120ps (FWHM),

respectively (Baldini etal., 2006). The MEG cryostat, consisting of cold (inner) and warm (outer)

vessels, is designed to reduce passive materials at the γ ray entrance. The same refrigerator tested on

the prototype will be used for cooling, with an emergency cooling system based on liquid nitrogen

owing

through pipes installed both inside the cold vessel and on its outer warm vessel.

Gaseous

and liquid purication systems are located on site for removing impurities possibly

retained after liquefaction. The construction of the calorimeter was completed in the autumn of

2007. Its calibration started in the same year and physical data taking continues to date. The MEG

ray scintillation calorimeter, with a total LXe mass of almost 2.7

and 846 PMTs, is currently the

largest LXe detector in operation worldwide (Figure 31.17). The many new technologies that were

brought by the MEG LXe detector have a great inuence on the detector, such as those developed

for

dark matter searches discussed in the following text.

31.4 appliCations in non-aCCelerator physiCs

31.4.1 liQuid argon tiMe projection chaMber (icaruS)

The technology of the LAr TPC was rst proposed by C. Rubbia in 1977 (Rubbia, 1977). It com-

bines the characteristics of a bubble chamber with the advantages of the electronic readout: it is

fully and continuously sensitive, self-triggering, and able to provide three-dimensional views of

ionizing events with particle identication from dE/dx and range measurement, and it can be oper-

ated over large active volumes. This detector is also a superb calorimeter of very ne granularity

and high accuracy. The ICARUS project envisages the usage of LAr TPC detectors for studies of

neutrinos from different sources as well as for searches for proton decay. This detection technique

offers 3D imaging with a granularity for tracking and calorimetric measurements, with a fully elec-

tronic detector. Owing to this, it gives good measurements in a wide range of neutrino energies or

background-free

and high-efciency measurements, for example, for proton decay.

conguration of the electrodes suitable for the 3D readout of the events was proposed by Gatti

etal. (1979). A voltage is applied to each wire plane in order to ensure the transparency of the grids

(a) (b)

Figure 31.17 (a) Picture of the LXe scintillation calorimeter. (b) Inside the detector.

Applications of Rare Gas Liquids to Radiation Detectors 899

to the drifting electrons (Bunemann etal., 1949). The position and charge of the track elements are

measured by means of the current induced in each wire. The drifting electrons reach and cross in

sequence following the wire planes (Figure 31.18): (1) A plane of wires running in the y-direction

(x and y are the two orthogonal directions along which the wires are placed), functioning as a

screening grid. (2) A plane of wires running again in the y-direction, located below the screening

grid; they measure, by induction, the x-coordinate. (3) A plane of wires running in the x-direction,

located below the previous plane; they measure the y-coordinate. Since this is the last sensitive

plane, the electric eld is such that this plane collects the drifting electrons. An electron drifting

in the sensitive volume above the screening grid does not produce any current in the coordinate

planes (in the case of perfect shielding). A positive induced current starts in the coordinate plane

when the electron crosses the screening grid, becomes negative when the electron crosses the coor-

dinate plane, and ends when the electron crosses the y-coordinate plane. The charge signal from the

y-coordinate plane has similar behavior, with the exception that only the positive current is pres-

ent. Thus, the segmentation of the readout system provides x–y coordinates. The coordinate along

the drift eld (z-direction) can be inferred from the drift time, measured as the time delay of the

arrival of electrons to the readout system with respect to a trigger signal (e.g., by direct scintillation).

From this information, the three-dimensional reconstruction of the ionization particle track can be

obtained.

Figure 31.19 shows a cosmic ray–induced shower as observed in the ICARUS LAr TPC.

fundamental requirement of the LAr TPC is that electrons produced by the ionizing par-

ticles can travel unperturbed from the point of production to the collecting wires. To this effect,

the system has to be able to purify and maintain pure LAr with a concentration of electronegative

impurities lower than 1 ppb of O

equivalent for the longest period of time possible. By using stan-

dard procedures for high vacuum (suitable materials, cleaning, design of internal components) and

vacuum conditioning of the internal surfaces, a high LAr purity after lling can be achieved. The

pollution of LAr is mainly due to the outgassing of inner surfaces in contact with gaseous Ar, while

the contamination from objects immersed in the liquid is negligible. In the absence of large convec-

tive motions, recirculation of the gas is sufcient to maintain the LAr purity. Commercial lters

provide

a purity level that is well above the experimental requirements.

purity monitor chamber for the measurement and control of the electron lifetime is developed

(Bettini etal., 1991). This chamber is a doubly girded ionization chamber, as shown schemati-

cally in Figure 31.20. The anode (A) and the cathode (K) are almost completely screened by two

grids (GA and GK) located a short distance from each electrode. Field intensities (E1, E2, and E3)

Induction signal

Collection signal

Drift time

Screen grid

Ionization track

Drift direction

Charge

Screen grid

1st sense wire

2nd sense wire

2nd sense wire grid

(y view)

1st sense wire grid

(x view)

Figure 31.18 Conguration of the anode wire plane and expected signals.