`
`179
`
`melt
`
`1
`
`Fig. 9.6. Crystal growth ac;cording to the Bridgman-Stockbarger method
`
`[ 17]. Csl: Tl crystals up to several inches in diameter can be grown out of quartz
`ampoules, but today Pt crucibles are used for the growth of all alka li halides.
`Crystals of BaF2 as well as CeF3 can also be grown by the Bridgman-Stockbarger
`method .
`The elements of the Czochralski technique are shown schematically in Fig. 9.7.
`The melt is contained in the crucible which is heated either by radio-frequency(cid:173)
`induction heating or by resistance heating. The pull rod with a chuck containing the
`seed crystal at its lower end is positioned axially above the crucible. The seed crystal is
`dipped into the melt and the melt temperature adjusted until the meniscus is supported.
`The pull rod is rotated slowly and then lifted. By careful adjustment of the power
`supplied to the melt, the diameter of the crystal is controlled as it grows. Rotation rates
`are commonly in the range of 1-100 rpm. Pulling rates can vary from one millimeter
`per hour for certain oxide crystals to several tens of millimeters per hour for halide
`crystals. The whole assembly is enclosed within an envelope which permits control
`of the ambient gas. The principal advantages of the technique include the fact that the
`crystal remains unstrained when it cools, so that a high structural perfection can be
`obtained. To yield material of high and controllable purity, it is necessary to fabricate
`the crucible from a material which is not attacked by the molten charge.
`For oxide crystals like Bi 4 Ge3 0 12 (Tm: 1044°C) and CdW04 (T111 : 1272"'C) Pt
`crucibles are normally used. For crystals with a higher melting temperature, such as
`Gd 2 Si05 :Ce (T 111 : 1950"'C), Ir crucibles are applied [18].
`Oxide crystals are usually grown by the Czochralski technique, but alkali halide
`crystals can also be grown by this method. The perfection of these crystals is high,
`but the size is normally limited to 3-4 inches diameter. However, with a special
`
`Vizio EX1018 Page 0190
`
`
`
`180
`
`9. X-Ray Phosphors and Scintillators (Counting Techniques)
`
`1
`c:::: >
`
`Seed(cid:173)
`holder
`
`/
`
`Quartz
`tube
`
`Crucible
`
`~
`
`~
`
`~
`
`~
`
`<!1TIIT!l>
`<l!IIIIID>
`cmmiD>
`
`r.f. heater
`
`<Uiml1>
`<Uiml1>
`<l!illiD>
`<l!illiD>
`<!ITIID1>
`
`<l!illiD>
`<mmn>
`
`Fig. 9.7. Crystal growth according to the Czochralski method
`
`Thermo-couple
`
`technique, viz. simultaneous feeding of the starting material during the crystal growth
`process, perfect halide crystals with up to 20 inches ( 50 em) diameter have been
`grown with a length up to 75 em [19].
`Bi 4 Ge3 0 12 crystals will only be really colorless when the raw materials Bi 2 0 3 and
`Ge2 0 have a high purity (5-6N). The growth atmosphere has to be oxygen, otherwise
`the Pt crucible will be attacked.
`The tendency of Bi 4 Ge3 0 12 crystals to grow with a "core" can be partly prevented
`by application of a higher rotation rate up to 100 rpm. But this inconvenience can be
`avoided using other growth techniques. Indeed the horizontal Bridgman-Stockbarger
`method has quite recently become popular for Bi 4 Ge3 0 12 growth, especially in China
`[20]. The details as to how this technique is applied are not known, but it must be
`successful. High quality crystals of dimensions 30 x 30 x 240 mm3 have been obtained,
`and are, for example applied in electromagnetic calorimeters. In order to grow large
`stoichiometric single crystals of CdW04 with a diameter of 3 inches, the Czochralski
`technique has to be slightly modified to prevent the evaporation of cadmium at the
`melting point of CdW04 .
`
`Vizio EX1018 Page 0191
`
`
`
`9.4. Material Preparation (Crystal Growth)
`
`181
`
`Table 9.5. Some properties of scintillators based on alkali halides [ 4,9,26]
`
`Property
`density (g cm- 3 )
`emission maximum (nm)
`light yield (photons Me y-I)
`decay time (ns)
`afterglow (% after 6 ms)
`stability
`
`mechanical behavior
`
`Nal :TJ+
`
`Csl :TJ+
`
`3.67
`415
`40.000
`230
`0.3-5
`hygro-
`scopic
`brittle
`
`4.51
`560
`55.000
`1000
`0.5-5
`hygro-
`scopic
`deform-
`able
`
`Csl: Na
`4.51
`420
`42.000
`630
`0.5-5
`hygro-
`scopic
`deform-
`able
`
`Csl
`4.51
`315
`2.000
`16
`
`-
`
`hygro-
`scopic
`deform-
`able
`
`The large anisotropy of the thermal expansion along the [01 0] axis of Gd 2 Si05 : Ce
`makes the growth of single crystals rather difficult. Cracks due to cooling down
`from about 1950'"'C to room temperature occur because of large residual stresses in
`the grown crystal. But up to now crystals up to 2 inches in diameter have been
`obtained without cracks. The topics of crystal growth have been treated in many
`review papers and books [21-24]. As an illustration of what can be achieved today ,
`Fig. 9.8 shows several Bi 4 Ge3 0 12 crystals. Two freshly-grown ingots without cores
`are shown, together with several machined crystals.
`
`Fig. 9.8. Crystals of Bi4 Ge3 0 12 grown by the Czochralski method. Two freshly grown ingots as
`well as several machined crystals are shown. The authors are grateful to the Crismatex Company
`who made this photograph available
`
`Vizio EX1018 Page 0192
`
`
`
`182
`
`9. X-Ray Phosphors and Scintillators (Counting Techniques)
`
`9.5 Scintillator Materials
`
`9.5.1 Alkali Halides
`
`Two of the alkali halides have been used as a scintillator material, viz. Nal and Csl,
`both doped with Tl+. Table 9.5 summarizes some of their properties. Also included
`are Csl: Na and undoped Csl. The emission spectra of the Tl+ -doped crystals are
`given in Fig. 9.9.
`These materials have a very high light yield (except for undoped Csl). For
`Nal: Tl+, for example, the radiant efficiency calculated from the light yield is about
`~ of the maximum possible efficiency (see Table 9.4). The low light yield of Csi
`is, certainly for a part, due to thermal quenching [26]. For certain applications the
`l f-LS) is acceptable. Unfortunately the afterglow
`decay time of these scintillators ( <
`is considerable and the stability poor. It depends on the application whether these
`scintillators can be applied or not (see Table 9.3). Nai: TJ+ is probably the most
`extensively used scintillator.
`The emission of the Tl+-doped alkali halides is due to the 3 P 1- 1S 0 transition on
`the TJ+ ion (see Sect. 3.3.7). It is usually assumed that the afterglow is due to hole
`trapping in the host lattice (trapped exciton, see Sect. 3.3.1 ), whereas the electron is
`trapped by the activator. In Csi: Nathe emission is due to an exciton bound to aNa+
`ion, in Csi to self-trapped exciton emission.
`
`Fig. 9.9. The emission spectra of Nal: Tl (a) and Csl : Tl(b) at room temperature under X-ray
`excitation
`
`600 nm
`
`Vizio EX1018 Page 0193
`
`
`
`9.5. Scintillator Materials
`
`183
`
`Table 9.6 Some properties of the tungstate scintillators and Bi4 Ge30 12 [4,9,26]
`
`Property
`density (g cm- 3)
`emission maximum (nm)
`I ight yield (photons Mev- 1)
`decay time (ns)
`afterglow (% after 3 ms)
`stability
`mechanical behavior
`
`ZnW04
`7.87
`480
`10.000
`5000
`< 0.1
`good
`brittle
`
`CdW04
`7.99
`480
`14.000
`5000
`< 0.1
`good
`brittle
`
`Bi4Ge3012
`7.13
`480
`9.000
`300
`0.005
`good
`brittle
`
`9.5.2 Thngstates
`
`The scintillators ZnW04 and CdW04 have high densities (see Table 9.6). Their light
`yields are lower than for the alkali halides, but their afterglow is weak. The crystals
`cleave easily which makes machining difficult. CdW04 is toxic.
`Their maximum efficiencies can be expected to be below 10% (Sect. 4.4). From
`the data in Table 9.6, CdW04 is estimated to yield 3.5%.
`These tungstates show broad band emissions originating from the tungstate oc(cid:173)
`tahedron (in contradistinction with the scheelite CaW04 , they have the wolframite
`crystal structure). This type of luminescence was treated in Sect. 3.3.5.
`
`Some properties of BGO are summarized in Table 9.6. It is a very interesting ma(cid:173)
`terial from the fundamental as well as from the applied point of view. It is used in
`calorimeters and PET scanners. The decay time is rather short, the afterglow weak,
`and the density high. Weber [27] has reviewed the history and properties of this scin(cid:173)
`tillator. Figure 9.10 gives the luminescence spectra. The relevant optical transitions
`were discussed in Sects. 3.3.7 and 5.3.2.
`Due to the large Stokes shift of the emission, i.e. the large relaxation in the excited
`state, the <:;mission of Bi4Ge3 0 12 is for the greater part quenched at room temperature.
`The calculated TJmax is 6%, but should be reduced to 2% in view of this quenching.
`The actual value calculated from the data in Table 9.6 is 2%. This shows that, apart
`from the quenching, there are not many radiationless losses.
`The compound Bi 2Ge3 0 9 [28] shows similar luminescence properties [29,30]. The
`Stokes shift is even larger than in Bi 4 Ge3 0 12 (20 000 vs 17 500 em- •, respectiveiy).
`As a consequence the emission intensity is quenched at 150 K, making the material
`unsuitable for application.
`The compound Bi 4Si 30 12 shows luminescence properties which are very similar
`to those of the germanium analogue [31 ,32].
`
`Vizio EX1018 Page 0194
`
`
`
`184
`
`9 . X -Ray Phosphors and Scintillators (Counting Technique s)
`
`600
`
`500
`
`400
`
`300
`
`nm
`
`250
`
`Fig . 9 .10. The emiss ion (left) and excitation (right) spectra of the luminescence of Bi 4 Ge30 12
`
`Table 9.7. Some properties of Ce3+ -activated scintillators (9,26]
`
`Host
`lattice
`
`BaF2
`LaF3
`CeF3
`YAI03
`Gd2SiOs
`Lu2SiOsa
`glassc
`Y3Al 5 012
`
`Ce3+ cone.
`(mole%)
`
`emi SSIOn
`max. (nm)
`
`0.2
`10
`100
`0 . 1
`0 .5
`
`4
`0.4
`
`310, 325
`290, 305
`310, 340
`350, 380
`440
`420
`390
`550
`
`light yield
`(photons
`Mev- 1)
`7000
`900
`4000
`17.000
`9000
`25.000
`1500
`14.000
`
`I
`
`decay
`time (ns)
`
`density
`(g cm- 3 )
`
`60b
`27
`30
`30
`60b
`40
`70b
`65
`
`4.89
`5.89
`6.16
`5.55
`6 .71
`7.4
`2.5
`5
`
`a C.L. Melcher, p. 415 in Ref. I
`b
`and longer component
`c
`composition (Si02 )o.ss (Mg0)o.24 (AI203)o.oo(Li20)o.oo (Ce203)o.o4
`
`The compound Gd 2Si05 h as a complicated crystal structure with two crystallographic
`sites for the rare earth ions. In recent years it has become popular as a scinti1lator.
`Crystals can be grown with the Czochralski method [19,24] . Their properties are
`summarized in Table 9.7. It is not hygroscopic, but it cleaves easily which can be a
`problem for certain applications.
`The luminescence of Ce3+ was treated in Sects. 3.3.3 and 5 .3 .2 . The emis(cid:173)
`sion transition (Sd ~ 4f) is fully allowed, so that short decay times can be ex-
`
`Vizio EX1018 Page 0195
`
`
`
`9.5. Scintillator Materials
`
`185
`
`pected. This is actually observed (see Table 9.7). The variation of the decay time
`is mainly determined by the emission wavelength according to the relation r ~ A2
`(see Sect. 3.3.3.a). The light yield of Gd2 Si05 : Ce3+, although not low, is below that
`expected (7Jobservd ~ 2.5%, 77max ~ 8%).
`Suzuki et al. [33] have reported on the ultraviolet and y-ray excited luminescence
`of Gd 2 Si05 : Ce3+. At 11 K they were able to find luminescence from two different
`Ce3+ ions, one with an emission maximum at about 425 nm, the other with an
`emission maximum at about 500 nm. The respective lowest excitation bands have
`their maximum at 345 and 380 nm, and the respective decay times are 27 and 43 ns.
`The former luminescence is hardly quenched at room temperature, the intensity of
`the latter decreases above 200 K, and at room temperature only 20% is left. Under
`y-ray excitation at room temperature the luminescence is dominated by the 425 nm
`emission, since the other is quenched for the greater part. Peculiarly enough, the decay
`shows under these conditions a long component (r ~ 600 ns) which is not observed
`for Y 2 Si05 : Ce3+ and Lu 2 Si05 : Ce3+.
`The results of previous chapters of this book allow us to propose a simple expla(cid:173)
`nation for these experimental observations. First we note that the ratio of the decay
`times ( '"'-' 0.65) is about equal to the squared ratio of the emission band maxima
`(~ 0.75) as is to be expected from r ~ A2 (see Sect. 3.3.3a).
`The long decay component of Gd2 Si05 : Ce3+ can be ascribed to the fact that part
`of the electron-hole pairs formed by y-ray irradiation are captured by Gd3+ ions. The
`excitation energy migrates over the Gd3+ ions as described in Sect. 5.3.1. In this way
`the Ce3+ ions are populated in a delayed way, so that a long decay component is
`observed. This effect does not occur in Y 2 Si05 : Ce3+ or Lu 2 Si05 : Ce3+, since the
`host lattice ions do not have energy levels below the band gap energy.
`The crystal structure of Gd 2 Si05 shows that one Ce3+ ion is coordinated by 8
`oxygen ions belonging to silicate tetrahedra and 1 oxygen which is not bounded to
`silicon. The latter oxygen is coordinated tetrahedrally by four rare earth ions. The
`other Ce3+ ion is coordinated by 4 oxygen ions belonging to silicate tetrahedra and 3
`oxygen ions which are not. The latter Ce3+ ion is more strongly covalently bonded,
`because the oxygen ions without silicon neighbors do not have enough positive charge
`in their immediate surroundings to compensate their twofold negative charge [34].
`Consequently, this Ce3+ ion has its energy levels at lower energy (see Sect. 2.2), as
`has also been observed for Tb3+ in Gd2 Si05 [34].
`Finally the lower quenching temperature of this Ce3 + emission remains to be
`explained. It is important to note that the oxygen ions, coordinated by four rare
`earth ions only, form a two-dimensional network in the crystal structure of Gd2 Si05 .
`The longer-wavelength emitting Ce3+ ions are located in this network. There is a
`striking structural analogy with the structure of Y 2 0 3 where every oxygen is tetra(cid:173)
`hedrally coordinated by 4 yttrium ions, so that here the network is threedimensional.
`Actually Ce3+ in Y 2 0 3 does not luminesce due to photoionization (see Sect. 4.5).
`Also in CatGdO(B03 )3 the Ce3+
`ion does not luminesce [35]. In this host lat(cid:173)
`tice a similar structural network can be distinguished. Therefore we conclude that
`the low quenching temperature of the Ce3+ ion in Gd 2 Si05 which is coordinated
`by three oxygen ions which do not coordinate to silicon must be explained in the
`same way.
`
`Vizio EX1018 Page 0196
`
`
`
`186
`
`9. X-Ray Phosphors and Scintillators (Counting Techniques)
`
`The quenching of one of the two Ce3+ centres in Gd 2 Si05 is at least partly
`responsible for the discrepancy between the observed and theoretical values of the
`efficiency.
`The host lattice Lu 2 Si05 has a different crystal structure which is also the struc(cid:173)
`ture of Y 2 Si05 • Luminescence studies do not show a large amount of quenching
`of the Ce3+ emission at room temperature [33]. It is therefore not surprising that
`in this structure much higher yields can be obtained than in Gd2 Si05 . The exper(cid:173)
`imental values are close to the theoretical maximum (see Table 9.4). It has been
`found that Y 2 Si05 : Tb3+ yields, under X-ray excitation, also a higher efficiency than
`Gd 2 Si05 : Tb3+ [34]. This suggests that the Gd 2 Si05 structure contains in some way
`centers which compete with the activator ions for the capture of charge carriers. The
`scintillator Lu 2 Si05 : Ce3+ seems, therefore, to have many advantages. Unfortunately,
`the cost price of pure Lu 2 0 3 is extremely high.
`
`9.5.5 CeF3
`
`Some properties of CeF3 scintillators are gathered in Table 9.7. This material is
`one of the serious candidates for a new generation of high-precision electromagnetic
`calorimeters to be used for the new large proton collider to be built at CERN (Geneva).
`For that purpose one needs a total crystal volume as large as 60 m 3 [2], which is nearly
`two orders of magnitude more than at present (1.2 m 3 Bi 4 Ge3 0 12 ). As mentioned,
`the relatively low light yield of CeF3 is not detrimental for this specific application
`(see Table 9.3). In view of these plans, a large amount of research has already been
`performed on CeF3 [36-38]. In passing, it should be mentioned that the high costs of
`this proposal have forced the scientists involved to look for a cheaper solution based
`on a reasonable compromise between cost and performance.
`CeF3 is a material with 100% activator concentration. As argued in Sect. 5.3.2, the
`large Stokes shift of the Ce3+ emission localizes the excited state, so that concentration
`quenching by energy migration does not occur.
`In our opinion the paper by Moses et al. [36] on the scintillation mechanisms in
`CeF3 is a fine example of how scintillators should be studied from a fundamental point
`of view. A combination of techniques was used, viz. (time-resolved) luminescence
`spectroscopy, ultraviolet photoelectron spectroscopy, transmission spectroscopy, and
`the excitation region was extended up to tens of eV by using synchrotron radiation.
`Further, powders as well as crystals with composition La 1_xCexF3 were investigated.
`The emission depends strongly on the value of x and on the excitation energy
`(see also Fig. 9.11 ). The intrinsic Ce3+ emission consists of a narrow band with
`maxima at 284 and 300 nm. These are ascribed to transitions from the lowest level
`of the 5dconfiguration to the two levels of the 4Jground configuration eF512 , 2 F 712 )
`(Sect. 3.3.3). If x > 0.1, an additional emission band appears at longer wavelength
`(around 340 nm) which sometimes even dominates (see Fig. 9.11 ). This one is as-
`cribed to Ce3+ ions close to defects.
`.
`Instructive is the emission for x < 0.01: it consists of the intrinsic Ce3+ emission
`band, p~+ emission lines, and a broad band extending from 250 to 500 nm which
`is ascribed to self-trapped exciton (STE) emission from the host. The STE consists
`
`Vizio EX1018 Page 0197
`
`
`
`9.5 . Scintillator Materia ls
`
`187
`
`a
`
`1
`
`b
`
`/
`
`300
`
`400 nm
`
`Fig. 9.11. Emission spectra of CeF3 . Curve a is for a powder under X -ray excitation, curve b
`for a crystal under y-ray excitation. From data in Ref. [36)
`
`of an electron bound to a V K centre which is a hole trapped by two fluorine ions
`(Sect. 3.3.1 ). This shows that the lattice itself also
`forming a pseudo-molecule F 2 -
`traps the electron-hole pairs. At room temperature the STE migrates over the lattice,
`ending its life by radiative recombination, transfer to Ce3+ or to p~+ (the latter being
`present as an impurity), or by nonradiative recombination. This illustrates that for
`this composition the factor S in Eq. (4.6) is far from one if one considers the Ce3+
`emission.
`In CeF3 energy transfer (Chapter 5) from intrinsic to extrinsic Ce3+ ions takes
`place [36,37]. Extrinsic Ce3+ ions are Ce3+ ions near to imperfections in the crystal
`lattice. Radiative as well as nonradiative transfer contributes. Actually the 340 nm
`emission shows a build-up which is equal to the decay of the 290 nm emission. The
`decay times of these emissions are about 30 and 20 ns, respectively. This agrees well
`'"""' A2 relation (Sect. 3.3.3a).
`with the T
`Under high-energy excitation an initial much faster decay has been observed. This
`phenomenon was studied by Pedrini et al. [37]. This fast component is of greater
`importance if there are more defects (impurities) in the lattice. However, even in
`very pure crystals it is present. When a high-energy particle is absorbed, the region
`of its relaxation has a radius of 10-100 nm. Therefore the electronic excitation is
`correlated in space and time. Auger relaxation in excited pairs has been proposed as
`a loss mechanism. The importance of this process decreases with temperature, since
`the excited states become more and more mobile at higher energies, so that the pair
`of excited ions can dissociate.
`
`Vizio EX1018 Page 0198
`
`
`
`188
`
`9. X -Ray Phosphors and Scintillators (Counting Techniques)
`
`The light yield of CeF3 is low, viz. 4000 photons McV- 1 • This corresponds to
`'"'"' 1 %, whereas TJmax ""' 8%. Thi s shows that the greater part of the emission is
`TJ
`quenched. It is usually admitted that impurity rare earth ions cannot be responsible for
`such a loss. However, fluoride crystals will always contain a certain (low) amount of
`oxygen. If the presence of 0 2
`forces one of the neighboring cerium ions to become
`-
`Ce4 + for charge compensation, a bulky quenching center is created, since intervalence
`charge-transfer transitions (Sect. 2.3.7) are at low energy and yield seldom emission
`[39]. If we add to this loss the Auger process mentioned above, it is understandable
`that the light yield of CeF3 is considerably lower than is expected.
`Moses et at. [36] have determined the quantum efficiency (Sect. 4.3) of the CeF3
`luminescence. For direct Ce3+ excitation it is high. Lower quantum efficiencies are
`found if the excitation starts at the F- ion (2p). For 100 e V the total quantum efficiency
`is about 0.7. The energy efficiency is then 3%. Also this is relatively low, and the
`authors suggest nonradiative recombination on quenching centers in order to explain
`this.
`
`9.5.6 Other Ce3+ Scintillators and Related Materials
`
`The strong potential of scintillators like Gd 2 Si05 : Ce3+ , Lu 2 Si05 : Ce3+ and CeF3
`have prompted a search for other Ce3+ -activated scintillators. Recently many new
`ones have been proposed [1,9,11]. Some of these have been included in Table 9.7.
`New ones are still appearing.
`Here we mention BaF2 : Ce3+, YA10 3 : Ce3 + and Y 3 A1 5 0 12 : Ce3+ (see Table 9.7).
`""' 30 ns, light yield 4000 photons
`Further there are reports on CeP5 0 14
`(7
`MeV- 1 [II]), LuP0 4 :Ce3+ (25, 17200 [11]), CsGd2 F 7 : Ce3 + (30, 10000 [40]), and
`GdA103 : Ce3+ (7 ,...___
`I ns, which is very short; no light yield given [41]).
`A slightly different approach is the use of Nd3+ as suggested by Van Eijk's group
`[9]. The Nd3+ ion has 4j 3 configuration with a Sd ---+ 4f emission transition in the
`ultraviolet ('"""' 175 om). Since this is an allowed transition at very high energy, the
`radiative decay time is even shorter than for Ce3+, viz. 6 ns (in LaF3 : Nd3+). See
`also Sect. 3.3.3. The light yield is a few hundred photons per MeV. Several other
`host lattices have been tried, but light yields never surpass 1000 photons MeV- 1 [42].
`An important problem is formed by the absorption of Nd3+ emission by rare earth
`impurities.
`
`9.5.7 BaFz (Cross Luminescence; Particle Discrimination)
`
`The optical transitions in the luminescence of BaF2 are discussed in Sect. 3.3.1 0. Two
`emissions are observed, a very fast cross luminescence (7 = 0.8 ns, emission maxima
`195 and 220 nm) and a much slower self-trapped exciton emission (7 = 600 ns,
`maximum '"""' 310 nm) which is for the greater part quenched at room temperature.
`
`Vizio EX1018 Page 0199
`
`
`
`9.5. Scintillator Materials
`
`189
`
`Table 9.8. Some properties of BaF2
`
`density
`maximum emission}
`wavelengths
`decay times
`
`light yield
`
`4.88 g cm- 3
`310 nm (slow)
`220 and 195 nm (fast)
`630 ns (slow)
`0.8 ns (fast)
`6.500 photons Me y-t (slow)
`2.500 photons Mev- 1 (fast)
`
`I
`
`a
`
`I 0
`
`....J
`
`b
`
`Fig. 9.12. Some data on the scintillation of BaF2. (a): emission spectrum at room temperature
`under y-ray excitation; the slow component is indicated by s, the fast component by f. (b):
`temperature dependence of the light output (LO) of BaF2 under y-ray excitation in a wide
`region around room temperature; the fast component (f) is temperature independent, the slow
`component (s) decreases strongly with increasing temperature
`
`BaF2 is not hygroscopic and large crystals can be grown. The slow emission com(cid:173)
`ponent can be reduced by a factor of four by adding 1% LaF3 . Table 9.8 gives some
`of its properties, whereas Fig. 9.12 shows the emission spectrum and the temperature
`dependence of both components [26].
`
`Vizio EX1018 Page 0200
`
`
`
`190
`
`9 . X-Ray Phosphors and Scintillators (Counting Techniques)
`
`At the moment there is exceiient agreement between the observed emission spec(cid:173)
`trum of BaF2 and the spectrum obtained from an ab initio calculation on basis of
`a molecular cluster approach [43,44]. This confirms the spectral assignments. The
`relatively low light yield of the fast component is a drawback for the application.
`Unfortunately the slow emission component is also of an intrinsic nature and captures
`a greater share of the electron-hole pairs than the fast component. The whole topic
`has been reviewed by Van Eijk [9,45].
`The BaF2 scintillator can be nicely used for particle discrimination (see Sect. 9.3
`and Fig. 9.5), because the intensity ratio of the fast and the slow components of the
`emission depends on the nature of the excitation. Figure 9.5 shows clearly that the
`heavier the exciting particles are, the less fast emission there is. The literature does
`not seem to contain an explanation for this effect. It should be realized that heavy
`particles will not penetrate deeply into the scintillator. This means that the excitation
`density must be very high. Since at room temperature the exciton will be more mobile
`than the electron-hole pair responsible for the cross luminescence, Auger interactions
`(Sects. 4.6 and 9 .5.5) will effect the fast component more strongly than the slow
`component. Note an interesting wordplay here: whereas the excited state of the cross
`luminescence is fast in the emission process, it is slow as far as migration through
`the lattice is concerned (the reason for this is the strongly localized character of the
`hole in the Ba 5p core band; see Fig. 3 .27). The slowly emitting exciton is fast in the
`lattice.
`
`9.5.8 Other Materials with Cross Luminescence
`
`In view of the strong interest in very fas t scintillator emi ssion, 1t 1s not surpns mg
`that many other compounds have been investigated for cross luminescence. It is
`essential, of course, that the cross-luminescence emission energy is smaller than the
`bandgap energy, since otherwise the cross luminescence cannot be emitted (see also
`Fig. 3.27). This is illustrated in Table 9.9 [9]. The table shows excellent agreement
`between prediction and observation.
`Table 9.10, finally, shows some compounds for which cross luminescence has
`been definitely observed [9]. All decay times are of the same order of magnitude
`('"'"' 1 ns), whereas the light yields do not reach the level of 2000 photons MeV- 1 • It
`is, at this time, too early to predict whether cross luminescence will have an important
`application or not.
`
`9.6 Outlook
`
`For many years scintillator research was often performed in the conviction that the
`physical mechanisms were unknown, that predictions were impossible, and that new
`materials should be found by trial and error [46]. As shown in this book and elsewhere
`
`Vizio EX1018 Page 0201
`
`
`
`9.6. Outlook
`
`191
`
`Table 9.9 On the possibility of cross luminescence [9]
`
`compound
`
`BaF2
`SrF2
`CaF2
`CsCI
`CsBr
`Csl
`KF
`KCI
`
`Ec-Evs a (eY)
`4.4-7.8
`8.4-12.8
`12.5-17.3
`1-5
`4-ti
`0-7
`7.5-10.5
`10-13
`
`Eub (eY)
`10.5
`I 1.1
`12.6
`8.3
`7.3
`6.2
`10.7
`8.4
`
`predictedc
`+
`
`0
`-
`+
`+
`0
`+
`-
`
`observedd
`+
`-/STE
`-/STE
`+
`+
`-/STE
`+
`-/STE
`
`a energy differences between top of core band and bottom or top of the valence band
`b energy gap
`c +: cross luminescence (CL) possible, -: CL impossible, o: CL doubtful
`d +: CL observed, -: no CL observed, STE: exciton emission observed
`
`Table 9.10. Scintillators with cross luminescence at 300 K [9]
`
`compound
`
`BaF2
`CsF
`CsCI
`RbF
`KMgF3
`KCaF3
`KYF4
`LiBaF3
`CsCaCI3
`
`emiSSIOn
`maximum (nm)
`195, 220
`390
`240, 270
`203, 234
`140-190
`140-190
`140-190
`230
`250, 305
`
`light yield
`(photons Mev- 1 )
`1400
`1400
`900
`1700
`1400
`1400
`1000
`1400
`1400
`
`T (ns)
`
`0.8
`2.9
`0.9
`1.3
`1.5
`<2
`1.9
`I
`<I
`
`[ 10,11 ,47], such a conviction is not justified. The knowledge from other fields of
`luminescence can be very helpful. The light yield is predictable from Eq. (4.6); TJmax
`is a well-known quantity in cathodoluminescence and q in photoluminescence. The
`transfer factorS in Eq. (4.6) is hard to predict and depends probably on the perfection
`and purity of the crystals. Decay times are easily predictable (see Sect. 3.3).
`Modern instrumentation, in addition, offers large potential in studying fundamental
`processes in scintillators. In the first place the synchrotron should be mentioned which
`makes monochromatic excitation up to very high energies possible. Examples were
`mentioned above [36,37].
`These developments tend to promote scintillator research into becoming "big sci(cid:173)
`ence". Also, international cooperation is increasing (see, for example, Ref. [38]), and
`the multidisciplinarity of this type of research is growing. Actually, the short history
`of the CeF3 scintillator sketched above illustrates all these developments very well.
`
`Vizio EX1018 Page 0202
`
`
`
`192
`
`9. X-Ray Phosphors and Scintillators (Counting Techniques)
`
`CeF3
`BaF2
`Csl
`
`PbF2
`PbW04j];[~T]~JJJJ77??J??I7777??7?IIIT
`NaBi(W04)2
`PbA1204
`PbGa204
`Fluoride glasses
`
`Y Al03:Ce,Pr (YAP)
`GdAl03:Ce
`LaAI03:Ce
`
`Lu2Si05:Ce (LSO)
`LuP04:Ce
`
`Gd2Si05:Ce (GSO) j~~~;!~~~;prrysszzl
`
`CsG d2F7:Ce
`Li(Yb,Gd)F4:Ce
`BaThF6
`
`Sma ll samples S mall samples Sto l Scm long
`Not a ll
`Basic
`crys ta ls
`p r operti es
`properties
`knovvn
`known
`
`Large crysta ls
`> 15 em
`
`S tatus of some promising fas t
`scintillators s tudie d by the Crystal Clear
`collaboration
`
`Crys tal s for
`hi g h
`p erformance
`ca lorimete r s
`
`Lo w light yield
`but chea p er
`solutions
`
`Hi gh li g ht
`yield for low
`energy
`calorimetry
`
`Ot h ers
`
`Fig. 9.13. An overview of scintillator research in the Crystal Clear collaboration on sc intillator
`crystals for calorimeter application ( 1993). Reprinted with permission from their status report
`of 3rd September 1993
`
`This does not necessarily mean that many new scintillators can be expected. The
`reason for this view is in the transfer factor S. In order to bring it close to 1, it is
`necessary to work with simple systems, and many of these have already been checked.
`For the same reason we do not have high expectations of amorphous scintillators. They
`contain "intrinsically" too many centers which will contribute to quenching. The open
`areas will soon be filled in us ing the theor y available at the moment. However, to
`optimize a given composition to a scintillator which satisfies the requirements of the
`application is still a hell of a job. It requires cooperation between materials scientists of
`several backgrounds (crystal growth, defect chemistry, solid state physics, materials
`science, spectroscopy, and radiation damage). An example of such a group is the
`
`Vizio EX1018 Page 0203
`
`
`
`Crystal Clear Collaboration (see authors of Ref. [38]). Figure 9.13 illustrates results
`from their work as summarized in the Status Report of this group of 3 September
`1993. It will be interesting to see how this picture evolves in the coming years.
`
`References
`
`193
`
`References
`
`I. De Notaristefani F, Lecoq P, Schneegans M ( 1993) (eds), Heavy scintillators for scientific
`and industrial applications. Editions Frontieres, Gif-sur- Yvette
`2. Lecoq P ( 1994) J Luminescence 60/61 :948
`3. Knoll GF ( 1987) Radiation detection and measurement, Wiley, New York
`4. Grabmaier BC, p 65 in Ref. [I]; J Luminescence 60/61:967
`5. Schotanus P ( 1988) thesis, Technical University, Delft
`6 . Anderson OF ( 1982) Phys Letters B I 18:230
`7. Melcher CL, p 75 in Ref. [I]
`8. van Eijk CWE, pp 161 and 60 I in Ref. [I]
`9 . van Eijk CWE ( 1993) Nucl. Tracks RadiaL Meas. 21:5
`I 0. Blasse G ( 1994) J Luminescence 60/61 :930
`II. Lempicki A, Wojtowicz AJ (1994) J Luminescence 60/61:942; Lempicki A, Wojtowicz AJ,
`Berman E ( 1993) Nucl lnstr Methods A322:304
`12. McKeever SWS ( 1985) Thermoluminescence of solids, Cambridge University, Cambridge;
`McKeever SWS, Markey BG, Lewandowski AC ( 1993) Nucl. Tracks Radiat. Meas. 21 :57
`13. Azorin J, Furetta C, Scacco A ( 1993) Phys Stat Sol(a) 138:9
`14. Wisshak K, Guber K, Kappeler F, Krisch J, Muller H, Rupp G, Voss F ( 1990) Nucl. lnstr.
`Methods A292:595
`15 . Migneco E, Agodi C, Alba R, Bellin G, Coniglione R, Del Zoppo A, Finocchiaro P,
`Maiolino C, Piattelli P, Raia G, Sapienza P ( 1992) Nucl. Instr. Methods A 314:3 1
`16. Nestor OH (1983) Mat. Res. Soc. Symp. Proc. 16:77
`17. Grabmaier BC ( 1984) IEEE Trans. Nucl. Sci. NS-31 :372
`18. Takagi K, Fukazawa T ( 1983) Appl. Phys. Lett. 42:43
`19. Goriletsky VI, Nemenov VA, Protse nko VG, Radkevich AV , Eidelman LG (I 980) Proc. 6th
`conf. on crystal growth, Moscow , p Ill 20
`20. Chongfau H
`( 1987) (Shanghai Institute of Ceramics, Academia Sinica, unpublished
`manuscript), cited as ref. 59 by Gevay G, Progress Crystal Growth Charact 15:181
`21. Wilke KTh, Bohm J ( 1988) Kristallzi.ichtung. Verlag Harr