- Email: [email protected]
Lu2S3:Ce3+, A new red luminescing scintillator
JL*H NIMMI
B
Beam interactions with Materials&Atoms
EiSEVIER
Nuclear Instruments and Methods in Physics Research B 134 (1998) 304-309
Letter to the Editor
Lu2S3:Ce3+, A new red luminescing scintillator J.C. van? Spijker ‘.*, P. Dorenbos
’
‘, C.P. Allier ‘, C.W.E. van Eijk ‘, A.R.H.F. G. Huber ’
Ettema
b,
Abstract Scintillation
properties
of Lu&:Ce’+
crystals are presented.
The studied crystals of -1 x I x 1 mm3 have the corun-
dum structure (density 6.25 g/cm”, effective atomic number 66.8). The linear attenuation length of a 51 I keV photon is 1.35cm. The light yield is 2.5 000-30 000 photons per MeV and the decay time is rather short, i.e. 32 ns. The emission spectrum is between 550 and 700 nm, peaking at 592 nm. Consequently by means of silicon diodes. 0 1998 Elsevier Science B.V. PACS: 29.40.Mc: 29.30.K~; 78.55.H~ &ZJWW&: Scintillation; y-ray detection;
light can be efficiently detected
Rare earth sulfide
In recent years several interesting Lu based scintillators were found, with Ce3+ as the luminescent center. Examples are LulSiOs:Ce3’ [l] and LuAIOs:Ce’+ 121. These kind of scintillators attenuate y-rays efficiently, because of the high atomic number of Lu and the high density of the compounds. Electrons and holes created in the crystal, upon X-ray or y-ray interaction can transfer their energy
*Corresponding author: Tel.: 15 12781954;fax: 15 2786422: e-mail: spijker@!iri.tudelft.nl. ’ These investigations have been supported bv the Nether_ lands Technolo~y~oundation (STW). . _ 0168-583X/98/$19.00 0 1998 Elsevier Science B.V. All rights resewed. PIISO168-583X(98)00667-8
the scintillation
fast and efficiently to CeZ’ centers and because of the allowed 5dAf dipole transition of the Ce)+ luminescence by means of this dopant, scintillation decay times in the order of 30 ns can be achieved. So far the search for new scintillators has been focused mainly on fluoride and oxide host materials. Sulfides did not gain much attention, despite the fact that they are well known in other luminescence applications. For example, doped sulfides are used in cathode ray tubes and in electroluminescent devices [3,4]. Sulfides have also been studied to find new laser materials [5]. In this letter we present the scintillation properties of Ce”’ doped LulSi crystals, which were grown by the method of chemical vapor transport
in Hamburg, Germany [6]. The crystals studied here are irregularly shaped with dimensions in the order of 1 x 1 x 1 mm3. The crystal structure, determined by diffraction analysis, is of the socalled z-phase [7], with the corundum structure and space group R 3 c. The coordination of the rare earth atoms is trigonal anti-prismatic with coordination number 6 and point symmetry C3”. The effective atomic number is 66.8 and the density 6.25 g/cm3. LulS3 is not hygroscopic. The emission spectrum under X-ray excitation was measured with an X-ray tube with a Cu anode operated at 35 kV and 25 mA. The spectrum, shown in Fig. 1, is measured with a red-sensitive Philips XP2254fB photomultiplier tube (PMT) and a HlO Jobin Yvon monochromator (grating 1200 grooves/mm, blazed at 250 nm). The optical absorption spectrum is measured using a light source with tungsten filament, an ARC VM504 monochromator (grating 1200 grooves/mm) and a Thorn EM1 type 9426 PMT. The optical emission measurements were done in reflection using the system Quantamaster of Photon Technology international. In Fig. 1 the optical absorption and the at ;lexc= 450 nm excited emission spectrum are also shown. The absorption spectrum consists of a strong absorption below
Fig. 1. (a) X-ray induced emission spectrum of Lu&:Ce?' . (b) optical absorption spectrum of Lu&:C$* (c) the i.,,, =450 nm excited emission spectrum. The latter is corrected for the transmission of the optical system and the QE of the PMT. The X-ray induced emission spectrum is not corrected for the transmission of the measurement system and the QE of the XP?254/B PMT. The peak at 650 nm indicated with an arrow is an artifact.
330 nm, a tail between 330 and 400 nm and a broad absorption band peaking at 460 nm. This spectrum strongly resembles the absorption spectrum of P-La&:Ce3’, measured by Scharmer et al. [S]. By comparison we attribute the 460 nm absorption band in Lu2S3:Ce”+ to 4f-5d absorption of Ce3+. Assuming that the absorption peak cross section for Ce”+ in LuzS3 is equal to that in pLa?&:Ce’+ we deduced the Ce3* concentration in Lu,_z,Cez,,Sj as .~=0.005. The absorption at 330 nm is attributed to the onset of the fundamental absorption band. The corresponding estimate of the bandgap energy of 3.8 eV is -0.6 eV higher than that reported by Schevciw et al. 181. The ieyc = 450 nm excited emission consists of a broad emission between 500 and 750 nm due to 5d-4f Cei’ luminescence. When we transform wavelength to energy we can reasonably well fit the ‘F~,z and ‘F7i? levels by two Gaussian shaped peaks, located at 16 770 + 80 cm-’ (2.08 + 0.01 eV) and 14 760 + 80 cm-’ ( 1.83 + 0.01 eV) below the lowest 5d absorption level. The spin-orbit splitting of the ‘F ground state is 2010 t 160 cm-‘. which is typical for Ce3+ [9]. The Stokes shift amounts 4900 If: 160 cm-‘. No other optically active impurities could be detected. Light yields were derived from pulse height spectra, using three different light detectors: Philips XP2020Q and XP2254/B PMTs, and a Hamamatsu S5345 Avalanche Photodiode (APD). The crystals were mounted on a light detector with a coupling liquid and were wrapped in 5 layers of 0.1 mm thick Teflon tape. The crystals were excited by y-quanta of a ‘“‘Am and a 13’Cs source. Standard spectroscopic techniques were used to record the pulse height spectra. In pulse height measurements with the PMTs the peak position of the photopeak was compared with that of the single electron spectrum. Thus the photoelectron yield per absorbed MeV of radiation (phe/MeV) was deduced. Using the quantum efficiencies (QEs) these yields were converted into photonslMeV (ph/MeV). The QEs of the XP202OQ and the XP2254lB PMT were determined with the method as described in [lo]. In order to determine the light yield out of a pulse height spectrum measured with the APD, first we calibrated the preamplifier and spectro-
306
J. C. can’t Spijhw
et al. I Nwl.
Insrr. and Meth. in Phys. Res
scopic amplifier with pulses of known charges. The number of generated e-h pairs can then be calculated from the photopeak position knowing that the gain for a detected photon in the visible is about 12.8 (at a bias voltage of -260 V). For the QE of the APD we used the data presented by the manufacturer. The photoelectron yields of 10 different Lu2S3:Ce3+ crystals were measured with the XP2254/B PMT using the 59.5 keV y-rays from a “‘Am source. The mean yield at 0.5 ps shaping time amounts 3 100 phe/MeV with a standard deviation of 340 phe/MeV. Since the crystals are small and irregularly shaped, there is a possible loss in light yield due to a non optimal coupling of the crystal to the detector. Indeed. after sanding and polishing an irregularly shaped crystal to one with plan parallel faces an increase of -20% in light yield is observed. The crystal, with dimensions of 0.2 x 0.4 x 2 mm’, showing the highest yield, was mounted on all three different scintillation-light sensors. Fig. 2 shows the pulse height spectrum of y-quanta of a “‘Cs source, measured with the crystal coupled onto an XP22541B PMT. In this spectrum, the pand y-ray background due to the natural abundance of radioactive ‘76Lu in the Lu&:Ce”’ crystal is negligible. In the pulse height spectrum, the characteristic escape peaks of the K, and I$ Lu X-rays arise 54 and 62 keV below the 662 keV photopeak. Considering the almost 100% X-ray fluo-
600
h
I
700
800
B 134 (1998) 304-309
rescence yield of Lu, and that for a 54 keV X-ray the linear attenuation coefficient is -2 mm-‘, we expect escape from the 0.2 mm thick crystal in -65% of the cases when we have absorption by the photoelectric effect. In Fig. 2 the ratio in peak intensity of the photopeak and the escape peak is about $ thus consistent with the expectations. In the inset of Fig. 2 the complete pulse height spectrum is presented, showing also the photopeak of the lJ7Ba X-ray absorption at -32 keV. The pulse height spectrum measured with the APD is shown in Fig. 3. The crystal and the APD were excited with X-rays and y-rays of a “‘Am source. In this spectrum the 59.5 keV scintillation peak in LuzS3:Ce3+ is located at channel 560. The peaks at channel 1660, 2060 and 2370 are due to the direct detection in the APD of Xrays of 137Np. The light yield of the Lu&:Ce3’ crystal measured with the calibrated setup amounts 21 100 e-h pairs/MeV. A simple method to obtain the number of e-h pairs per MeV of absorbed radiation in the LuzSi:Ce3+ crystals is by comparison of position of the photopeak of the scintillation in the Lu&:Ce” crystal with that of the direct detection of X-rays in the APD, considering that the e-h production requires 3.6 eV. We then find a light yield of 13 90013.6 x 560/ 1660 x 110.0595= 21 900 e-h pairs/MeV. In this case the yield is &80/ too high, since the APD-amplification for visible light is somewhat higher than for X-rays.
energy [keV]
Fig. 2. Pulse height spectrum of “‘Cs y-rays recorded with LuzSi:Ce’+ coupled onto an XP22541B tube. The spectrum is measured with a shaping time of 0.5 vs. In the inset the total pulse height spectrum is shown. Note the logarithmic .v-axis.
Fig. 3. The pulse height spectrum of “‘Am X-rays and y-rays recorded with Lu&:Ce’+ coupled to the APD. The spectrum is measured with a shaping time is 0.5 ps.
307
J.C. cun’t Spijker et al. I Nucl. Instr. and Meth. in Phys. Res. B 134 (1998) 304-309
The measured light yields of the 0.2 x 0.4 x 2 mm3 crystal are compiled in Table 1. In this table the error in light yields in column 5, is due to the uncertainty in peak position of the photopeak and single electron peak (column 3). the error in the mean QE at the scintillation emission (column 4) and the inaccuracy in light collection due to mounting, which is estimated to be 10%. The light yield, measured with the XP2254/B PMT and the APD is -28 000 ph/MeV. In the XP22541B error contributions for QE and mounting inaccuracy are of the same order of magnitude. In the APD the main error is the inaccuracy in mounting. The light yield measured with the XP2020Q PMT seems to be somewhat lower, i.e. -22 000 ph/MeV. However, this value is consistent with the other results, since the uncertainty in the mean QE of the XP202OQ at the LulS3:Ce3t emission, is quite large. In all cases with increasing shaping time the measured yields increase. In going from 0.5 to 10 us an increase of approximately 15% is observed. This implies the presence of a decay time component in the 3 us range. The decay time spectrum was measured using the multi-hit method, employing XP2020Q PMTs as start and stop PMTs [l 11. For excitation we used y-rays of a “‘Am source. The spectrum is shown in Fig. 4. It shows a main decay time of 32 ns, persisting over almost two decades of intensity. The presence of a decay time component in the 3 us range, as was deduced from the pulse Table 1 The light yields of the 0.2 x 0.4 x 2 mm3 Lu&:Ce’+ represents the mean QE at the scintillation emission in the light detector Light detector
S5345 APD
Shaping
0.5 3
time (us)
102
..
7
f .
s
‘..
5
I
!,.
8
.1
.;’
10'
._
‘.
‘I = 2.2 ns
..:;.i .
.:. '.... .I _.,. _ ... ._
ii._
@a
___. .
.
. . ..,
I
504
time [ns]
Fig. 4. The scintillation decay time spectrum of Lu2Sj:Ce’- induced by y-rays of a “‘Am source, measured with XP2020Q start and stop PMTs. The accumulation time is 28 hours. The peak at the start of the scintillation decay curve at t = 0 is an experimental artifact.
cannot be observed in height measurements, Fig. 4. This is due to the low number of counts in the scintillation decay tail which is comparable with the number of background counts. Let us consider the efficiency of energy transfer from the generated e-h pairs to the luminescence centers. The energy needed for the creation of an e-h pair amounts 2-3 times the bandgap [ 121. The bandgap for Lu2S3 is 3.8 eV. Then the energy needed for the creation for an e-h pairs is at most -11 eV. This means that per MeV absorbed radiation energy 90 000 e-h pairs are created. The maximum light yield we measured is -28 000 phi MeV, which means that only 30% of the e-h pairs
crystal. measured with three different light detectors. The given QE in column 4 of Lu&:Cez+. It includes the collection efficiency of the charge carriers generated
Measured light yield (phe/MeV) or (e-h pairs/MeV)
QE at LuzSz:CeZ+ emission (‘%I)
Light yield (phe/MeV)
21 100 + 400 23 700 It 400
79 2 2
26 700 f 2600 30 000 ? 3000
X P2020Q
0.5 3 10
291 f 10 316 + 10 354 + 10
1.8 ? 1
12800 20200f 21 900 ? 13 900 15600 24600&
XP2254lB
0.5 3 10
3922 * 60 4068 + 60 4454 f 60
13 * 2
30 100 + 5500 31 300 f 5700 34 300 & 6300
308
J.C. oan’t Sp(jker et ul. I Nud.
Table 2 A comparison
of the scintillation
properties
of Lu&:Ce”
Instr. and Meth. in Phys. Rrs. B 134 (1998) 3044309
with other Lu based scintiifators
Properties
LuzSi05:Ce”+
Density (g/cm’) Etfective atomic number Attenuation length for a 511 keV photon (cm) Photoelectric fraction at 511 keV Emission maximum (nm) Light yield (ph/MeV) Decay time (ns) Refractive index Number of background counts due to the presence 17’Lu (count/s cmi)
1.4 64.4 1.14 0.34 420 17 000 42 1.81 318
of
[I,181
LuAlOq:Ce’”
[I81
8.34 65 1.05 0.32 365 11 400 a 17.5 1.95 323
LuzSq:Ce’+ 6.25 66.7 1.35 0.35 592 -28 000 32 25-2.8 ’ 270
B The crystal growth technique of LuAIOl:Ce’+ is not yet optimized. It is well possible that the light yield increases for better crystal quality. b The refractive index for Lu2SI:Ce” is not known. we assume that it comparable with other rare earth sulfides. i.e. between 2.5 and 2.8
gives rise to radiative
recombination
on Ce3+. Yet,
In Table 2 some properties of Lu&:Ce3+ are compared with those of Lu2SiOs:Ce’+ and LuA103:Ce3+. The light yield of the new scintillator is approximately equal to that of Lu&O&e’+. It is somewhat higher than that of the fast 17 ns scintillation component of LuA103:Ce3+. An advantage of the Lu&:C$+ scintillator is its emission in the red. Thus, the scintillation light can be efficiently detected by silicon photodiodes. The results presented were obtained on small crystals. We do not know, whether efficient Lu&:Ce3’ scintillation crystals with sizes of e.g. 1 x 1 x 1 cm3 can be grown. So far only small Lu& crystals have been grown [7,13]. The crystal growth of other rare earth sesquisulfides with sizes larger than I x I x 1 mm3 is rarely reported [14,15]. The melting point of Lu& is around 1750°C [16,17]. The refractive index of LuzS3 is probably in the range of 2.5-2.8, observed for other rare earth sesquisul~des [14]. This is slightly higher than that of other high-density scintillators. such as Bi4Ge3017-and CdWO+ which have a refractive index of 2.15 and 2.3, respectively. Since standard photodetectors and coupling liquids only have a refractive index around 1.5, a large part of the generated scintillation light might be reflected on the crystal-photodetector interface. If clear crystals the light
yield
is relatively
high.
can be obtained, possible light losses can be minimized using highly reflective coatings. From transmission measurements on the small crystals we observed that self-absorption of the scintillation light in the crystal does not pose a problem.
References Nucl. Ins&. and Meth. A 314 (1992) ?I?. B.1. Minkov, in: P.A. Rodnyi, C.W.E. van Eijk (Eds.). Record of the International Workshop PHYSCIW Physical Processes in Fast Scintillators. St. Petersburg. Russian Federation. 3 September-3 October. 1994, p, 182 (Stratech report TUD-SCIR-94-04) Y.A. Ono, M. Fuyama. 0. Ken-ichi, K. Tamura, M. Ando. J. Appl. Phys. 66 (I I) (1989) 5564. A.N. Belskv, 0. Krachni. V.V. Mikhailin. J. Phys.: Condens. Matter 5 (1993) 9417. E.G. Scharmer. M. Leiss. G. Huber, J. Phys. C 15 (1982) 1071. M. Leiss. J. Phys. C 13 (1980) 151. T. Schleid, F. Lissner. Z. Anorg. Allg. Chem. 615 (1992) 19. 0. Schevciw. W.B. White, Mater. Res. Bull. IS (1983) 1059. G. Blasse. W. Schipper. J.J. Hamelink, Inorg. Chem. Acta 189 (1991) 77. P. Dorenbos. J.T.M. de Haas. R. Visser, C.W.E. van Eijk, R.W. Hollander, Nucl. Instr. and Meth. A 325 (1993) 367. W.W. Moses. Nucl. Instr. and Meth. A 336 (1993) 253. P.A. Rodnyi, P. Dorenbos. C.W.E. van Eijk. Phys. Status Solidi B 187 (1995) 15.
[II C.L. Melcher. J.S. Schweitzer. PI
[31 I41 PI [61 [71 PI [91 VOI PiI PI
J. C. can’t Spijker et al. I Nucl. Instr. and Meth. in Phys. Rex B 134 (ISUS) 304-309 [13] A.W. Sleight. C.T. Prewitt. Inorg. Chem. 7 (1968) 2282. [14] V.P. Zhuze. A.I. Shelykh, Sov. Phys. Semicond. 23 (3) (1989) 245. [l.S] J.R. Henderson, M. Muramoto, E. Loh. J.B. Gruber. J. Chem. Phys. 47 (1967) 3347. [lh] H.F. Franzen, A.V. Hariharan. J. Chem. Phys. 70 (1979) 4907.
309
[17] H.F. Franzen. A.V. Hariharan, Advan. Chem. 186 (1980) 309. [18] M. Moszynski, D. Wolski. T. Ludziejewski. M. Kapusta, A. Lempicki, C. Brecher. D. Wisniewski, A.J. Wojtowicz. Nucl. Instr. and Meth. A 385 (1997) 123.
B
Beam interactions with Materials&Atoms
EiSEVIER
Nuclear Instruments and Methods in Physics Research B 134 (1998) 304-309
Letter to the Editor
Lu2S3:Ce3+, A new red luminescing scintillator J.C. van? Spijker ‘.*, P. Dorenbos
’
‘, C.P. Allier ‘, C.W.E. van Eijk ‘, A.R.H.F. G. Huber ’
Ettema
b,
Abstract Scintillation
properties
of Lu&:Ce’+
crystals are presented.
The studied crystals of -1 x I x 1 mm3 have the corun-
dum structure (density 6.25 g/cm”, effective atomic number 66.8). The linear attenuation length of a 51 I keV photon is 1.35cm. The light yield is 2.5 000-30 000 photons per MeV and the decay time is rather short, i.e. 32 ns. The emission spectrum is between 550 and 700 nm, peaking at 592 nm. Consequently by means of silicon diodes. 0 1998 Elsevier Science B.V. PACS: 29.40.Mc: 29.30.K~; 78.55.H~ &ZJWW&: Scintillation; y-ray detection;
light can be efficiently detected
Rare earth sulfide
In recent years several interesting Lu based scintillators were found, with Ce3+ as the luminescent center. Examples are LulSiOs:Ce3’ [l] and LuAIOs:Ce’+ 121. These kind of scintillators attenuate y-rays efficiently, because of the high atomic number of Lu and the high density of the compounds. Electrons and holes created in the crystal, upon X-ray or y-ray interaction can transfer their energy
*Corresponding author: Tel.: 15 12781954;fax: 15 2786422: e-mail: spijker@!iri.tudelft.nl. ’ These investigations have been supported bv the Nether_ lands Technolo~y~oundation (STW). . _ 0168-583X/98/$19.00 0 1998 Elsevier Science B.V. All rights resewed. PIISO168-583X(98)00667-8
the scintillation
fast and efficiently to CeZ’ centers and because of the allowed 5dAf dipole transition of the Ce)+ luminescence by means of this dopant, scintillation decay times in the order of 30 ns can be achieved. So far the search for new scintillators has been focused mainly on fluoride and oxide host materials. Sulfides did not gain much attention, despite the fact that they are well known in other luminescence applications. For example, doped sulfides are used in cathode ray tubes and in electroluminescent devices [3,4]. Sulfides have also been studied to find new laser materials [5]. In this letter we present the scintillation properties of Ce”’ doped LulSi crystals, which were grown by the method of chemical vapor transport
in Hamburg, Germany [6]. The crystals studied here are irregularly shaped with dimensions in the order of 1 x 1 x 1 mm3. The crystal structure, determined by diffraction analysis, is of the socalled z-phase [7], with the corundum structure and space group R 3 c. The coordination of the rare earth atoms is trigonal anti-prismatic with coordination number 6 and point symmetry C3”. The effective atomic number is 66.8 and the density 6.25 g/cm3. LulS3 is not hygroscopic. The emission spectrum under X-ray excitation was measured with an X-ray tube with a Cu anode operated at 35 kV and 25 mA. The spectrum, shown in Fig. 1, is measured with a red-sensitive Philips XP2254fB photomultiplier tube (PMT) and a HlO Jobin Yvon monochromator (grating 1200 grooves/mm, blazed at 250 nm). The optical absorption spectrum is measured using a light source with tungsten filament, an ARC VM504 monochromator (grating 1200 grooves/mm) and a Thorn EM1 type 9426 PMT. The optical emission measurements were done in reflection using the system Quantamaster of Photon Technology international. In Fig. 1 the optical absorption and the at ;lexc= 450 nm excited emission spectrum are also shown. The absorption spectrum consists of a strong absorption below
Fig. 1. (a) X-ray induced emission spectrum of Lu&:Ce?' . (b) optical absorption spectrum of Lu&:C$* (c) the i.,,, =450 nm excited emission spectrum. The latter is corrected for the transmission of the optical system and the QE of the PMT. The X-ray induced emission spectrum is not corrected for the transmission of the measurement system and the QE of the XP?254/B PMT. The peak at 650 nm indicated with an arrow is an artifact.
330 nm, a tail between 330 and 400 nm and a broad absorption band peaking at 460 nm. This spectrum strongly resembles the absorption spectrum of P-La&:Ce3’, measured by Scharmer et al. [S]. By comparison we attribute the 460 nm absorption band in Lu2S3:Ce”+ to 4f-5d absorption of Ce3+. Assuming that the absorption peak cross section for Ce”+ in LuzS3 is equal to that in pLa?&:Ce’+ we deduced the Ce3* concentration in Lu,_z,Cez,,Sj as .~=0.005. The absorption at 330 nm is attributed to the onset of the fundamental absorption band. The corresponding estimate of the bandgap energy of 3.8 eV is -0.6 eV higher than that reported by Schevciw et al. 181. The ieyc = 450 nm excited emission consists of a broad emission between 500 and 750 nm due to 5d-4f Cei’ luminescence. When we transform wavelength to energy we can reasonably well fit the ‘F~,z and ‘F7i? levels by two Gaussian shaped peaks, located at 16 770 + 80 cm-’ (2.08 + 0.01 eV) and 14 760 + 80 cm-’ ( 1.83 + 0.01 eV) below the lowest 5d absorption level. The spin-orbit splitting of the ‘F ground state is 2010 t 160 cm-‘. which is typical for Ce3+ [9]. The Stokes shift amounts 4900 If: 160 cm-‘. No other optically active impurities could be detected. Light yields were derived from pulse height spectra, using three different light detectors: Philips XP2020Q and XP2254/B PMTs, and a Hamamatsu S5345 Avalanche Photodiode (APD). The crystals were mounted on a light detector with a coupling liquid and were wrapped in 5 layers of 0.1 mm thick Teflon tape. The crystals were excited by y-quanta of a ‘“‘Am and a 13’Cs source. Standard spectroscopic techniques were used to record the pulse height spectra. In pulse height measurements with the PMTs the peak position of the photopeak was compared with that of the single electron spectrum. Thus the photoelectron yield per absorbed MeV of radiation (phe/MeV) was deduced. Using the quantum efficiencies (QEs) these yields were converted into photonslMeV (ph/MeV). The QEs of the XP202OQ and the XP2254lB PMT were determined with the method as described in [lo]. In order to determine the light yield out of a pulse height spectrum measured with the APD, first we calibrated the preamplifier and spectro-
306
J. C. can’t Spijhw
et al. I Nwl.
Insrr. and Meth. in Phys. Res
scopic amplifier with pulses of known charges. The number of generated e-h pairs can then be calculated from the photopeak position knowing that the gain for a detected photon in the visible is about 12.8 (at a bias voltage of -260 V). For the QE of the APD we used the data presented by the manufacturer. The photoelectron yields of 10 different Lu2S3:Ce3+ crystals were measured with the XP2254/B PMT using the 59.5 keV y-rays from a “‘Am source. The mean yield at 0.5 ps shaping time amounts 3 100 phe/MeV with a standard deviation of 340 phe/MeV. Since the crystals are small and irregularly shaped, there is a possible loss in light yield due to a non optimal coupling of the crystal to the detector. Indeed. after sanding and polishing an irregularly shaped crystal to one with plan parallel faces an increase of -20% in light yield is observed. The crystal, with dimensions of 0.2 x 0.4 x 2 mm’, showing the highest yield, was mounted on all three different scintillation-light sensors. Fig. 2 shows the pulse height spectrum of y-quanta of a “‘Cs source, measured with the crystal coupled onto an XP22541B PMT. In this spectrum, the pand y-ray background due to the natural abundance of radioactive ‘76Lu in the Lu&:Ce”’ crystal is negligible. In the pulse height spectrum, the characteristic escape peaks of the K, and I$ Lu X-rays arise 54 and 62 keV below the 662 keV photopeak. Considering the almost 100% X-ray fluo-
600
h
I
700
800
B 134 (1998) 304-309
rescence yield of Lu, and that for a 54 keV X-ray the linear attenuation coefficient is -2 mm-‘, we expect escape from the 0.2 mm thick crystal in -65% of the cases when we have absorption by the photoelectric effect. In Fig. 2 the ratio in peak intensity of the photopeak and the escape peak is about $ thus consistent with the expectations. In the inset of Fig. 2 the complete pulse height spectrum is presented, showing also the photopeak of the lJ7Ba X-ray absorption at -32 keV. The pulse height spectrum measured with the APD is shown in Fig. 3. The crystal and the APD were excited with X-rays and y-rays of a “‘Am source. In this spectrum the 59.5 keV scintillation peak in LuzS3:Ce3+ is located at channel 560. The peaks at channel 1660, 2060 and 2370 are due to the direct detection in the APD of Xrays of 137Np. The light yield of the Lu&:Ce3’ crystal measured with the calibrated setup amounts 21 100 e-h pairs/MeV. A simple method to obtain the number of e-h pairs per MeV of absorbed radiation in the LuzSi:Ce3+ crystals is by comparison of position of the photopeak of the scintillation in the Lu&:Ce” crystal with that of the direct detection of X-rays in the APD, considering that the e-h production requires 3.6 eV. We then find a light yield of 13 90013.6 x 560/ 1660 x 110.0595= 21 900 e-h pairs/MeV. In this case the yield is &80/ too high, since the APD-amplification for visible light is somewhat higher than for X-rays.
energy [keV]
Fig. 2. Pulse height spectrum of “‘Cs y-rays recorded with LuzSi:Ce’+ coupled onto an XP22541B tube. The spectrum is measured with a shaping time of 0.5 vs. In the inset the total pulse height spectrum is shown. Note the logarithmic .v-axis.
Fig. 3. The pulse height spectrum of “‘Am X-rays and y-rays recorded with Lu&:Ce’+ coupled to the APD. The spectrum is measured with a shaping time is 0.5 ps.
307
J.C. cun’t Spijker et al. I Nucl. Instr. and Meth. in Phys. Res. B 134 (1998) 304-309
The measured light yields of the 0.2 x 0.4 x 2 mm3 crystal are compiled in Table 1. In this table the error in light yields in column 5, is due to the uncertainty in peak position of the photopeak and single electron peak (column 3). the error in the mean QE at the scintillation emission (column 4) and the inaccuracy in light collection due to mounting, which is estimated to be 10%. The light yield, measured with the XP2254/B PMT and the APD is -28 000 ph/MeV. In the XP22541B error contributions for QE and mounting inaccuracy are of the same order of magnitude. In the APD the main error is the inaccuracy in mounting. The light yield measured with the XP2020Q PMT seems to be somewhat lower, i.e. -22 000 ph/MeV. However, this value is consistent with the other results, since the uncertainty in the mean QE of the XP202OQ at the LulS3:Ce3t emission, is quite large. In all cases with increasing shaping time the measured yields increase. In going from 0.5 to 10 us an increase of approximately 15% is observed. This implies the presence of a decay time component in the 3 us range. The decay time spectrum was measured using the multi-hit method, employing XP2020Q PMTs as start and stop PMTs [l 11. For excitation we used y-rays of a “‘Am source. The spectrum is shown in Fig. 4. It shows a main decay time of 32 ns, persisting over almost two decades of intensity. The presence of a decay time component in the 3 us range, as was deduced from the pulse Table 1 The light yields of the 0.2 x 0.4 x 2 mm3 Lu&:Ce’+ represents the mean QE at the scintillation emission in the light detector Light detector
S5345 APD
Shaping
0.5 3
time (us)
102
..
7
f .
s
‘..
5
I
!,.
8
.1
.;’
10'
._
‘.
‘I = 2.2 ns
..:;.i .
.:. '.... .I _.,. _ ... ._
ii._
@a
___. .
.
. . ..,
I
504
time [ns]
Fig. 4. The scintillation decay time spectrum of Lu2Sj:Ce’- induced by y-rays of a “‘Am source, measured with XP2020Q start and stop PMTs. The accumulation time is 28 hours. The peak at the start of the scintillation decay curve at t = 0 is an experimental artifact.
cannot be observed in height measurements, Fig. 4. This is due to the low number of counts in the scintillation decay tail which is comparable with the number of background counts. Let us consider the efficiency of energy transfer from the generated e-h pairs to the luminescence centers. The energy needed for the creation of an e-h pair amounts 2-3 times the bandgap [ 121. The bandgap for Lu2S3 is 3.8 eV. Then the energy needed for the creation for an e-h pairs is at most -11 eV. This means that per MeV absorbed radiation energy 90 000 e-h pairs are created. The maximum light yield we measured is -28 000 phi MeV, which means that only 30% of the e-h pairs
crystal. measured with three different light detectors. The given QE in column 4 of Lu&:Cez+. It includes the collection efficiency of the charge carriers generated
Measured light yield (phe/MeV) or (e-h pairs/MeV)
QE at LuzSz:CeZ+ emission (‘%I)
Light yield (phe/MeV)
21 100 + 400 23 700 It 400
79 2 2
26 700 f 2600 30 000 ? 3000
X P2020Q
0.5 3 10
291 f 10 316 + 10 354 + 10
1.8 ? 1
12800 20200f 21 900 ? 13 900 15600 24600&
XP2254lB
0.5 3 10
3922 * 60 4068 + 60 4454 f 60
13 * 2
30 100 + 5500 31 300 f 5700 34 300 & 6300
308
J.C. oan’t Sp(jker et ul. I Nud.
Table 2 A comparison
of the scintillation
properties
of Lu&:Ce”
Instr. and Meth. in Phys. Rrs. B 134 (1998) 3044309
with other Lu based scintiifators
Properties
LuzSi05:Ce”+
Density (g/cm’) Etfective atomic number Attenuation length for a 511 keV photon (cm) Photoelectric fraction at 511 keV Emission maximum (nm) Light yield (ph/MeV) Decay time (ns) Refractive index Number of background counts due to the presence 17’Lu (count/s cmi)
1.4 64.4 1.14 0.34 420 17 000 42 1.81 318
of
[I,181
LuAlOq:Ce’”
[I81
8.34 65 1.05 0.32 365 11 400 a 17.5 1.95 323
LuzSq:Ce’+ 6.25 66.7 1.35 0.35 592 -28 000 32 25-2.8 ’ 270
B The crystal growth technique of LuAIOl:Ce’+ is not yet optimized. It is well possible that the light yield increases for better crystal quality. b The refractive index for Lu2SI:Ce” is not known. we assume that it comparable with other rare earth sulfides. i.e. between 2.5 and 2.8
gives rise to radiative
recombination
on Ce3+. Yet,
In Table 2 some properties of Lu&:Ce3+ are compared with those of Lu2SiOs:Ce’+ and LuA103:Ce3+. The light yield of the new scintillator is approximately equal to that of Lu&O&e’+. It is somewhat higher than that of the fast 17 ns scintillation component of LuA103:Ce3+. An advantage of the Lu&:C$+ scintillator is its emission in the red. Thus, the scintillation light can be efficiently detected by silicon photodiodes. The results presented were obtained on small crystals. We do not know, whether efficient Lu&:Ce3’ scintillation crystals with sizes of e.g. 1 x 1 x 1 cm3 can be grown. So far only small Lu& crystals have been grown [7,13]. The crystal growth of other rare earth sesquisulfides with sizes larger than I x I x 1 mm3 is rarely reported [14,15]. The melting point of Lu& is around 1750°C [16,17]. The refractive index of LuzS3 is probably in the range of 2.5-2.8, observed for other rare earth sesquisul~des [14]. This is slightly higher than that of other high-density scintillators. such as Bi4Ge3017-and CdWO+ which have a refractive index of 2.15 and 2.3, respectively. Since standard photodetectors and coupling liquids only have a refractive index around 1.5, a large part of the generated scintillation light might be reflected on the crystal-photodetector interface. If clear crystals the light
yield
is relatively
high.
can be obtained, possible light losses can be minimized using highly reflective coatings. From transmission measurements on the small crystals we observed that self-absorption of the scintillation light in the crystal does not pose a problem.
References Nucl. Ins&. and Meth. A 314 (1992) ?I?. B.1. Minkov, in: P.A. Rodnyi, C.W.E. van Eijk (Eds.). Record of the International Workshop PHYSCIW Physical Processes in Fast Scintillators. St. Petersburg. Russian Federation. 3 September-3 October. 1994, p, 182 (Stratech report TUD-SCIR-94-04) Y.A. Ono, M. Fuyama. 0. Ken-ichi, K. Tamura, M. Ando. J. Appl. Phys. 66 (I I) (1989) 5564. A.N. Belskv, 0. Krachni. V.V. Mikhailin. J. Phys.: Condens. Matter 5 (1993) 9417. E.G. Scharmer. M. Leiss. G. Huber, J. Phys. C 15 (1982) 1071. M. Leiss. J. Phys. C 13 (1980) 151. T. Schleid, F. Lissner. Z. Anorg. Allg. Chem. 615 (1992) 19. 0. Schevciw. W.B. White, Mater. Res. Bull. IS (1983) 1059. G. Blasse. W. Schipper. J.J. Hamelink, Inorg. Chem. Acta 189 (1991) 77. P. Dorenbos. J.T.M. de Haas. R. Visser, C.W.E. van Eijk, R.W. Hollander, Nucl. Instr. and Meth. A 325 (1993) 367. W.W. Moses. Nucl. Instr. and Meth. A 336 (1993) 253. P.A. Rodnyi, P. Dorenbos. C.W.E. van Eijk. Phys. Status Solidi B 187 (1995) 15.
[II C.L. Melcher. J.S. Schweitzer. PI
[31 I41 PI [61 [71 PI [91 VOI PiI PI
J. C. can’t Spijker et al. I Nucl. Instr. and Meth. in Phys. Rex B 134 (ISUS) 304-309 [13] A.W. Sleight. C.T. Prewitt. Inorg. Chem. 7 (1968) 2282. [14] V.P. Zhuze. A.I. Shelykh, Sov. Phys. Semicond. 23 (3) (1989) 245. [l.S] J.R. Henderson, M. Muramoto, E. Loh. J.B. Gruber. J. Chem. Phys. 47 (1967) 3347. [lh] H.F. Franzen, A.V. Hariharan. J. Chem. Phys. 70 (1979) 4907.
309
[17] H.F. Franzen. A.V. Hariharan, Advan. Chem. 186 (1980) 309. [18] M. Moszynski, D. Wolski. T. Ludziejewski. M. Kapusta, A. Lempicki, C. Brecher. D. Wisniewski, A.J. Wojtowicz. Nucl. Instr. and Meth. A 385 (1997) 123.