- Email: [email protected]
Growth and scintillation properties of pure CsI crystals grown by micro-pulling-down method
Optical Materials 34 (2012) 1087–1091
Contents lists available at SciVerse ScienceDirect
Optical Materials journal homepage: www.elsevier.com/locate/optmat
Growth and scintillation properties of pure CsI crystals grown by micro-pulling-down method Daisuke Totsuka a,b,⇑, Takayuki Yanagida d, Yutaka Fujimoto a, Jan Pejchal c, Yuui Yokota a, Akira Yoshikawa a,d a
Institute for Materials Research, Tohoku University, 2-1-1 Katahira, Aoba-ku, Sendai, Miyagi 980-8577, Japan Nihon Kessho Kogaku Co., Ltd., 810-5 Nobe-cho Tatebayashi, Gunma 374-0047, Japan Institute of Physics AS CR, Cukrovarnicka 10, 162-53 Prague 6, Czech Republic d New Industry Creation Hatchery Center (NICHe) 6-6-10 Aoba, Aramaki, Aoba-ku, Sendai, Miyagi 980-8579, Japan b c
a r t i c l e
i n f o
Article history: Received 11 July 2011 Received in revised form 26 December 2011 Accepted 4 January 2012 Available online 28 January 2012 Keywords: Growth from melt Micro-pulling-down Pure CsI Scintillator
a b s t r a c t Single crystals of pure cesium iodide (CsI) have been grown from the melt using micro-pulling-down (l-PD) method. Two kinds of crucible (graphite one and quartz one) were used for the growth and the grown crystals were investigated by X-ray diffraction (XRD) and X-ray rocking curve (XRC) analysis. The XRD analysis did not confirm any impurity phases and a sub-grain structure was observed for each sample in the rocking curve measurement. Under X-ray irradiation, strong STE emission peaks around 300 nm were observed together with some luminescence related to unintentionally present impurities. The STE emission peaks are characterized by fast decay times of several ns and about 20 ns which are interpreted as the on-center-type STE (VK + e) and off-center type STE (H + F) recombinations, respectively. The light yield of the STE-related emissions has been estimated to be 3000 ph/MeV. Other emission peaks were observed at 410 nm and 515 nm. The former one can be related to Br-contamination and it is characterized by a relatively slow decay time of 6 ls. Concerning the latter one at 515 nm, similar luminescence was observed for the water-doped CsI grown by Bridgman method. Ó 2012 Elsevier B.V. All rights reserved.
1. Introduction The l-PD method is a relatively novel method for crystal growth from the melt [1] and it has several advantages compared to some conventional methods such as Czochralski (Cz) and Bridgman (BS) methods. In the l-PD, the shape of crucible plays an important role, because it is not only container of the melt but also the shaper for the grown crystal. The shape of the crystal is strongly depended on the configuration of the die part of the crucible positioned at the bottom and the die has a hole to allow liquid transport from the crucible. Crystals in the form of fiber, column, plate, and tube can be grown with a special shaped crucible [2]. Furthermore, from the aspect of material research, the fast growth speed of typically 0.05–1 mm/min is considerably attractive for a material screening. Many functional crystals including for example Pr-doped Lu3Al5O12 [3] for gamma-ray detection and rare-earthdoped LiCaAlF6 [4] for neutron measurements were developed using this convenient method so far. One of the most important ⇑ Corresponding author. Present address: Institute for Materials Research, Tohoku University, 2-1-1 Katahira, Aoba-ku, Sendai, Miyagi 980-0812, Japan. Tel.: +81 22 215 2217; fax: +81 22 215 2215. E-mail address: [email protected] (D. Totsuka). 0925-3467/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.optmat.2012.01.003
factors is the wettability between the crucible and the melt. When the wettability is good, the solid/liquid interface is formed under the crucible and the shape of grown crystal is formed by the shape of the die. Otherwise, solidification occurs inside the crucible and the diameter of grown crystal can be determined by the shape and dimensions of the hole. From the point of view of pulling the melt, namely the difficulty of crystal growth, it would be defined that good conditions for the l-PD crystal growth method are ‘‘good’’ or ‘‘poor’’ wettability between the melt of materials and crucibles. Generally, metal crucibles made of Pt, Ir or Re are used for an oxide crystal growth and graphite crucibles are used for a fluoride crystal growth. One unique feature of the crystals grown by the he l-PD method is uniform distribution of dopants or impurities along the growth direction [5], thus superior properties are expected (e.g. energy resolution in scintillator materials). However, most halide crystals except the fluoride ones can be grown only by the B.S. method with a sealed quartz ampoule [6] or the Cz method in a dry environment due to the increased hygroscopicity. Therefore, we developed a modified l-PD method with a removable chamber system for the growth of hygroscopic halide crystals. CsI is a well known fast scintillator with an emission peaking at 300 and 340 nm with a fast time response characterized by a decay time of 10 ns. It has been suggested by Nishimura et al. [7], that the
1088
D. Totsuka et al. / Optical Materials 34 (2012) 1087–1091
300 and 340 nm peaks are related to the intrinsic luminescence decay of the on-center-type STE (VK + e) and off-center type STE (H + F), respectively. It also has been observed that the crystals with poor quality exhibit the increase of slow component in longer wavelength range and decrease of scintillation light yield [8]. In the recent years, much effort was devoted to obtain better scintillation characteristics of pure and Tl-doped CsI single crystals. In this study, we grew pure CsI crystals by the modified l-PD method using two kinds of crucibles made of graphite and quartz material, then the crystallinity and scintillation properties were investigated. Actually, when trying to use platinum crucible the crystal could not be grown due to the corrosive nature of CsI and its unsuitable wettability. Then the crystalline and the scintillation properties were investigated. Water-doped CsI was also prepared by a vertical BS method to clarify the origin of impurity luminescence. Fig. 2. Photograph of two kinds of crucible, carbon graphite (left) and quartz (right).
2. Experiment 2.1. Crystal growth and crystalline properties Schematic drawing of l-PD crucible and the growth procedure for the system with radio-frequency (RF) heating is given in Fig. 1. The crucibles are placed on an alumina pedestal in a vertical quartz tube and are heated using RF generator. When the quartz crucible is used, carbon pedestal is inserted as a heater. The modified l-PD crystal growth furnace for halides has a removable chamber system, which can be moved in a glove box filled with Ar gas [9]. Preparation of a starting material and setup of crucible, seed and insulators were carried out inside the glove box with an atmosphere control to keep the oxygen and water moisture concentrations below 1 ppm. High purity (99.999%) CsI powder from Chemetall Company was used as a starting material. This starting material was weighed and charged into the crucibles. Fig. 2 displays these crucibles made of graphite and quartz. These crucibles had a cylindrical shape and were equipped with a hole of 1 mm in diameter. Pt wire with 0.5 mm in diameter was used as a seed instead of a seed crystal. The removable chamber was installed on vacuum and gas system with a stage movable in vertical direction. When the installation of the chamber was finished, the chamber was heated up and evacuated to less than 10 3 Pa by a rotary pump and turbo molecular pump to remove moisture in the raw material. Then the chamber was filled with the high purity Ar gas. After these processes, the crucible was heated up to the melting temperature of CsI. The
pulling rate was controlled between 0.01 and 0.1 mm/min. The grown crystals were cooled down to room temperature in 1 h after the pulling was finished. To examine the contaminations from the atmosphere inside the furnace, glow-discharge mass spectrometry analysis was carried out. The phase of the grown crystals was confirmed by the X-ray diffraction (XRD) analysis (RINT-2000, Rigaku corporation) in the 2h of 10°–90° with step of 0.02°. The X-ray was generated by Cu target using the tube voltage of 40 kV and current of 40 mA. The crystallinity was investigated by X-ray rocking curve (XRC) analysis for the (1 1 0) plane of grown crystals using a high-resolution diffractometer (Rigaku ATX) with Cu Ka1 radiation diffracted by a twobounce Ge (2 2 0) channel monochromator. Crystallinity of the grown CsI crystals was measured by x-scan and evaluated using the width of the peak. All the X-ray experiments were carried out at room temperature. 2.2. Scintillation evaluation The grown crystals were cut and polished to evaluate the scintillation properties. Pulse height spectrum measurements were carried out under gamma-ray (137Cs) excitation with a photomultiplier tube R7600U (Hamamatsu Photonics) connected to an ORTEC 113 preamplifier, an ORTEC 572 shaping amplifier and an Amptek Pocket MCA 8000 A multichannel analyzer for digital signal conversion. The bias voltage of the PMT was supplied at +600 V (ORTEC 556). The samples were mounted on the PMT with an optical grease (OKEN, 6262A) and covered with several layers of Teflon tape to collect scintillation photons. At the same time, decay time measurement was done using a digital oscilloscope (Tektronix TDS3052B). To evaluate the light yield, BGO crystal was used as a Ref. [10]. The radio-luminescence spectra measurement was performed at room temperature under X-ray irradiation. The excitation source was the RINT-2000 and the X-ray tube was supplied 40 kV and 40 mA. The emission spectra were measured using a Andor DU420-OE CCD cooled down to 213 K by a Peltier module. This CCD was coupled with ORIEL INSTRUMENTS monochromator with 285 grooves/mm and 280 nm blaze wavelength. The scintillation light was sent to CCD through a 2 m optical fiber to avoid direct irradiation of CCD by the X-ray. 3. Results and discussion 3.1. Crystal growth and crystallinity
Fig. 1. Schematic diagram of the modified l-PD method with a removable chamber system.
CsI crystals were successfully grown by the l-PD method with two kinds of crucibles and they are shown in Fig. 3. Both are trans-
D. Totsuka et al. / Optical Materials 34 (2012) 1087–1091
Fig. 3. Photographs of pure CsI single crystals, quartz-crucible-used (top) and graphite-crucible-used (bottom).
1089
Fig. 4. Powder X-ray diffraction patterns of CsI crystals. Black and red lines represent graphite-crucible and quartz-crucible-used crystals, respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
parent with a different shape. The physical dimensions of each crystal were 1 mm diameter and 15 mm in length for graphite crucible and 3 mm in diameter and 12 mm in length for quartz crucible, respectively. This difference of the diameter of the grown crystals could be due to the different wettability between CsI melt and the crucibles. In the case of graphite, the wettability seemed to be low because the solid/liquid interface was in the die and the diameter of grown crystal corresponded to that of the hole. On the other hand, the quartz material was better wetted by the CsI melt than the graphite material since the melt was hold on the bottom of the crucible and spread than the diameter of hole. The GDMS analysis showed the concentration of cationic impurities of the order of several ppm in both crystals. However, Br which could substitute the anion site was detected in large quantities around 0.8 mol% in the crystal grown in the graphite crucible and 0.1 mol% in the crystal grown in the quartz crucible. Many attempts to grow halide crystals were made using the furnace demonstrated in this study so far and the boiling point of bromides was proved to be lower than that of iodides, therefore Br contamination from the previous experiments could be observed. The powder XRD results are shown in Fig. 4. The diffraction patterns of both samples corresponded to the JCPDS chart #06-0311 and no impurity peaks were observed. The calculated lattice constants of CsI grown in quartz and graphite crucibles were 4.575 Å and 4.584 Å, respectively. This slight difference of the lattice constants could be affected by an interstitial-type lattice defect mainly due to Br ions. The crystal quality of the CsI samples grown by the l-PD method was characterized by (XRC) measurement and the results are shown in Fig. 5. It was difficult to perform accurate fitting of both XRC curves by a simple Gaussian. A bilateral symmetry was not observed and some peaks were clearly shown in the midslopes. Since the crystals were grown without the seed crystal, the grown samples would consist of sub-grain structure [11]. This became pronounced in the crystal grown in the quartz crucible. 3.2. Scintillation characteristics Fig. 6 shows the radio-luminescence spectra under the X-ray excitation. The spectra were normalized for the amplitude of the dominant emission peak at 300 nm. Water-doped CsI crystal prepared by the vertical BS method using a sealed ampoule was measured to identify the impurity emission peak. In all the CsI crystals, the emission peaks at 300 nm, 410 nm and 515 nm were observed.
Fig. 5. X-ray rocking curve of CsI crystals, graphite-crucible-used (top) and quartzcrucible-used (bottom).
It has been reported that the prominent emission band of 300 nm with a fast decay time is caused by STE [12]. The 410 nm luminescence band was observed in CsI:Br crystal [13] and it is related to a slow migration process with ls decay time. The ratio of this emission intensity with respect to the other emission peaks seems to be
1090
D. Totsuka et al. / Optical Materials 34 (2012) 1087–1091
Fig. 6. X-ray excited luminescence spectra of CsI crystals. Black line and red line show graphite-crucible-used and quartz-crucible-used, respectively. Blue line is the result of water-doped CsI crystal made by the vertical BS technique using sealed ampoule. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
different for each crystal. The crystal grown from the graphite crucible contained Br in a relatively high concentration, therefore we conclude that this peak is related to the Br impurity. The broader emission peak around 515 nm became the largest in the waterdoped CsI crystal. Thus, the broad peak around 515 nm can be ascribed to the water contamination. Fig. 7 shows the decay time profiles of both samples under 137Cs gamma-ray excitation. The decay times extracted from the fitting curves were determined to be 6 and 23 ns for the crystal grown in the graphite crucible and 7, 16 and 6033 ns for that grown in the quartz crucible, respectively. The component with several decay time is most probably related to the on-center-type STE (VK + e) recombination and the components with around 20 ns could be as ascribed to the off-center type STE (H + F) recombination [7]. The latter component was difficult to distinguish from the dominant peak around 300 nm in the shown radio-luminescence spectra because of poor resolution of spectrometer and CCD cam-
I(t) = 0.99 exp (-t/5.3ns) + 0.24 exp (-t/15.7ns)
Intensity (a.u.)
1
+ 0.04 exp (-t/6033ns) +0.03
0.1
0.01
0
100
200
300
400
500
Time (ns) I(t) = 1.04 exp (-t/5.0ns) + 0.32 exp (-t/20.1ns) + 0.03
Intensity (a.u.)
1
0.1
0.01
0
50
100
Time (ns) Fig. 7. Decay time profiles of graphite-crucible-used (top) and quartz-crucible-used (bottom), excited by 137Cs gamma-ray.
Fig. 8. Pulse height spectra under 137Cs irradiation. Black, red, and blue lines represent graphite-crucible-used, quartz-crucible-used, and BGO, respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
era. The slowest component with ls-order could be related to Br impurity. Fig. 8 shows 137Cs gamma-ray pulse height spectra of the pure CsI crystals compared with a BGO single crystal. The shaping time constant was set to be 2 ls for quartz-crucible crystal and 10 ls for graphite-crucible crystal. In all the samples, 662 keV photoelectron absorption peak was clearly observed. Light yields were determined by a comparison with BGO, which has the light output of 8000 ph/MeV. A correction factor taking into account the quantum efficiency of the photocathode was calculated for the corresponding positions of the emission peaks. Extracting the peak channel number from the single Gaussian function fitting of the photo-absorption peak and considering the quantum efficiency of the PMT of 30% at 300 nm (Fig. 6), a light yield of 3000 ph/MeV was deduced for the crystal grown from the quartz crucible. This value is well consistent with the previously reported results [14]. The MCA peak channel of crystal grown from graphite crucible was 1.8 times higher but it is difficult to evaluate exact light yield because this photo-absorption peak includes significant amount of slow components. 4. Conclusions Pure CsI single crystals were grown by the modified l-PD method using graphite and quartz crucibles. Transparent crystals were obtained. The higher wettability was observed for the quartz crucible. Many sub-grains were confirmed by XRC measurements, especially in the crystal grown from the quartz crucible. The strong STE emission consisting of two bands associated with the on-center-type STE (VK + e) and off-center type STE (H + F) were observed at 300 nm and their decay times were of several ns and around twenty ns, respectively. The light yield of these fast components was estimated to be 3000 ph/MeV for the crystal grown from the quartz crucible. Extrinsic luminescence was observed at 410 nm and 515 nm. The former band was related to Br and the latter to the water impurities. The scintillation decay time of the luminescence related to Br impurity was determined to be 6 ls. Acknowledgments This work was mainly supported by JST Sentan and partially by a Grant in Aid for Young Scientists (B)-15686001 and (A)-23686135 from the Ministry of Education, Culture, Sports, Science and Tech-
D. Totsuka et al. / Optical Materials 34 (2012) 1087–1091
nology of the Japanese government (MEXT). Partial assistance from the Yazaki Memorial Foundation for Science and Technology, Japan Science Society, Sumitomo Foundation, and Iketani Science and Technology Foundation are also gratefully acknowledged.
References [1] A. Yoshikawa, M. Nikl, G. Boulon, T. Fukuda, Opt. Mater. 30 (2007) 6–10. [2] Y. Yokota, V. Chani, M. Sato, K. Tota, K. Onodera, T. Yanagida, A. Yoshikawa, J. Cryst. Growth 318 (2011) 983–986. [3] M. Nikl, H. Ogino, A. Krasnikov, A. Beitlerova, A. Yoshikawa, T. Fukuda, Phys. Stat. Sol. (a) 202 (1) (2005) 4–6. [4] A. Yoshikawa, T. Yanagida, Y. Yokota, N. Kawaguchi, S. Ishizu, K. Fukuda, T. Suyama, K.J. Kim, J. Pejchal, M. Nikl, K. Watanabe, M. Miyake, M. Baba, K. Kamada, IEEE Trans. Nucl. Sci. 56 (2009) 3796–3799. [5] R. Simura, A. Yoshikawa, S. Uda, J. Cryst. Growth 311 (2009) 4763–4769.
1091
[6] W.M. Higgins, J. Glodo, E. Van Loef, M. Klugerman, T. Gupta, L. Cirignano, P. Wong, K.S. Shah, J. Cryst. Growth 287 (2006) 239–242. [7] H. Nishimura, M. Sakata, T. Tsujimoto, M. Nakayama, Phys. Rev. B 51 (1994) 2167. [8] B.K. Utts, S.E. Spagno, IEEE Trans. Nucl. Sci. NS-37 (1990) 134. [9] Y. Yokota, N. Kawaguchi, K. Fukuda, T. Yanagida, A. Yoshikawa, M. Nikl, J. Cryst. Growth 318 (2011) 908–911. [10] I. Holl, E. Lorenz, G. Mageras, IEEE Trans. Nucl. Sci. 35 (1998) 105–109. [11] N. Balamurugan, A. Arulchakkaravarthi, S. Selvakumar, M. Lenina, R. Kumard, S. Muralithard, K. Sivaji, P. Ramasamya, J. Cryst. Growth 286 (2006) 294–299. [12] Z. Wu, B. Yang, P.D. Townsend, J. Lumin. 128 (2008) 1191–1196. [13] G.I. Britvich, I.G. Britvich, V.G. Vasil’chenko, V.A. Lishin, V.F. Obraztsov, V.A. Polyakov, A.S. Solovjev, V.D. Ryzhikov, Nucl. Instrum. Meth. Phys. A 469 (2001) 77–88. [14] C. Amsler, D. Grogler, W. Joffrain, D. Lindelof, M. Marchesotti, P. Niederberger, H. Pruysa, C. Regenfus, P. Riedler, A. Rotondi, Nucl. Instrum. Meth. Phys. A 480 (2002) 494–500.
Contents lists available at SciVerse ScienceDirect
Optical Materials journal homepage: www.elsevier.com/locate/optmat
Growth and scintillation properties of pure CsI crystals grown by micro-pulling-down method Daisuke Totsuka a,b,⇑, Takayuki Yanagida d, Yutaka Fujimoto a, Jan Pejchal c, Yuui Yokota a, Akira Yoshikawa a,d a
Institute for Materials Research, Tohoku University, 2-1-1 Katahira, Aoba-ku, Sendai, Miyagi 980-8577, Japan Nihon Kessho Kogaku Co., Ltd., 810-5 Nobe-cho Tatebayashi, Gunma 374-0047, Japan Institute of Physics AS CR, Cukrovarnicka 10, 162-53 Prague 6, Czech Republic d New Industry Creation Hatchery Center (NICHe) 6-6-10 Aoba, Aramaki, Aoba-ku, Sendai, Miyagi 980-8579, Japan b c
a r t i c l e
i n f o
Article history: Received 11 July 2011 Received in revised form 26 December 2011 Accepted 4 January 2012 Available online 28 January 2012 Keywords: Growth from melt Micro-pulling-down Pure CsI Scintillator
a b s t r a c t Single crystals of pure cesium iodide (CsI) have been grown from the melt using micro-pulling-down (l-PD) method. Two kinds of crucible (graphite one and quartz one) were used for the growth and the grown crystals were investigated by X-ray diffraction (XRD) and X-ray rocking curve (XRC) analysis. The XRD analysis did not confirm any impurity phases and a sub-grain structure was observed for each sample in the rocking curve measurement. Under X-ray irradiation, strong STE emission peaks around 300 nm were observed together with some luminescence related to unintentionally present impurities. The STE emission peaks are characterized by fast decay times of several ns and about 20 ns which are interpreted as the on-center-type STE (VK + e) and off-center type STE (H + F) recombinations, respectively. The light yield of the STE-related emissions has been estimated to be 3000 ph/MeV. Other emission peaks were observed at 410 nm and 515 nm. The former one can be related to Br-contamination and it is characterized by a relatively slow decay time of 6 ls. Concerning the latter one at 515 nm, similar luminescence was observed for the water-doped CsI grown by Bridgman method. Ó 2012 Elsevier B.V. All rights reserved.
1. Introduction The l-PD method is a relatively novel method for crystal growth from the melt [1] and it has several advantages compared to some conventional methods such as Czochralski (Cz) and Bridgman (BS) methods. In the l-PD, the shape of crucible plays an important role, because it is not only container of the melt but also the shaper for the grown crystal. The shape of the crystal is strongly depended on the configuration of the die part of the crucible positioned at the bottom and the die has a hole to allow liquid transport from the crucible. Crystals in the form of fiber, column, plate, and tube can be grown with a special shaped crucible [2]. Furthermore, from the aspect of material research, the fast growth speed of typically 0.05–1 mm/min is considerably attractive for a material screening. Many functional crystals including for example Pr-doped Lu3Al5O12 [3] for gamma-ray detection and rare-earthdoped LiCaAlF6 [4] for neutron measurements were developed using this convenient method so far. One of the most important ⇑ Corresponding author. Present address: Institute for Materials Research, Tohoku University, 2-1-1 Katahira, Aoba-ku, Sendai, Miyagi 980-0812, Japan. Tel.: +81 22 215 2217; fax: +81 22 215 2215. E-mail address: [email protected] (D. Totsuka). 0925-3467/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.optmat.2012.01.003
factors is the wettability between the crucible and the melt. When the wettability is good, the solid/liquid interface is formed under the crucible and the shape of grown crystal is formed by the shape of the die. Otherwise, solidification occurs inside the crucible and the diameter of grown crystal can be determined by the shape and dimensions of the hole. From the point of view of pulling the melt, namely the difficulty of crystal growth, it would be defined that good conditions for the l-PD crystal growth method are ‘‘good’’ or ‘‘poor’’ wettability between the melt of materials and crucibles. Generally, metal crucibles made of Pt, Ir or Re are used for an oxide crystal growth and graphite crucibles are used for a fluoride crystal growth. One unique feature of the crystals grown by the he l-PD method is uniform distribution of dopants or impurities along the growth direction [5], thus superior properties are expected (e.g. energy resolution in scintillator materials). However, most halide crystals except the fluoride ones can be grown only by the B.S. method with a sealed quartz ampoule [6] or the Cz method in a dry environment due to the increased hygroscopicity. Therefore, we developed a modified l-PD method with a removable chamber system for the growth of hygroscopic halide crystals. CsI is a well known fast scintillator with an emission peaking at 300 and 340 nm with a fast time response characterized by a decay time of 10 ns. It has been suggested by Nishimura et al. [7], that the
1088
D. Totsuka et al. / Optical Materials 34 (2012) 1087–1091
300 and 340 nm peaks are related to the intrinsic luminescence decay of the on-center-type STE (VK + e) and off-center type STE (H + F), respectively. It also has been observed that the crystals with poor quality exhibit the increase of slow component in longer wavelength range and decrease of scintillation light yield [8]. In the recent years, much effort was devoted to obtain better scintillation characteristics of pure and Tl-doped CsI single crystals. In this study, we grew pure CsI crystals by the modified l-PD method using two kinds of crucibles made of graphite and quartz material, then the crystallinity and scintillation properties were investigated. Actually, when trying to use platinum crucible the crystal could not be grown due to the corrosive nature of CsI and its unsuitable wettability. Then the crystalline and the scintillation properties were investigated. Water-doped CsI was also prepared by a vertical BS method to clarify the origin of impurity luminescence. Fig. 2. Photograph of two kinds of crucible, carbon graphite (left) and quartz (right).
2. Experiment 2.1. Crystal growth and crystalline properties Schematic drawing of l-PD crucible and the growth procedure for the system with radio-frequency (RF) heating is given in Fig. 1. The crucibles are placed on an alumina pedestal in a vertical quartz tube and are heated using RF generator. When the quartz crucible is used, carbon pedestal is inserted as a heater. The modified l-PD crystal growth furnace for halides has a removable chamber system, which can be moved in a glove box filled with Ar gas [9]. Preparation of a starting material and setup of crucible, seed and insulators were carried out inside the glove box with an atmosphere control to keep the oxygen and water moisture concentrations below 1 ppm. High purity (99.999%) CsI powder from Chemetall Company was used as a starting material. This starting material was weighed and charged into the crucibles. Fig. 2 displays these crucibles made of graphite and quartz. These crucibles had a cylindrical shape and were equipped with a hole of 1 mm in diameter. Pt wire with 0.5 mm in diameter was used as a seed instead of a seed crystal. The removable chamber was installed on vacuum and gas system with a stage movable in vertical direction. When the installation of the chamber was finished, the chamber was heated up and evacuated to less than 10 3 Pa by a rotary pump and turbo molecular pump to remove moisture in the raw material. Then the chamber was filled with the high purity Ar gas. After these processes, the crucible was heated up to the melting temperature of CsI. The
pulling rate was controlled between 0.01 and 0.1 mm/min. The grown crystals were cooled down to room temperature in 1 h after the pulling was finished. To examine the contaminations from the atmosphere inside the furnace, glow-discharge mass spectrometry analysis was carried out. The phase of the grown crystals was confirmed by the X-ray diffraction (XRD) analysis (RINT-2000, Rigaku corporation) in the 2h of 10°–90° with step of 0.02°. The X-ray was generated by Cu target using the tube voltage of 40 kV and current of 40 mA. The crystallinity was investigated by X-ray rocking curve (XRC) analysis for the (1 1 0) plane of grown crystals using a high-resolution diffractometer (Rigaku ATX) with Cu Ka1 radiation diffracted by a twobounce Ge (2 2 0) channel monochromator. Crystallinity of the grown CsI crystals was measured by x-scan and evaluated using the width of the peak. All the X-ray experiments were carried out at room temperature. 2.2. Scintillation evaluation The grown crystals were cut and polished to evaluate the scintillation properties. Pulse height spectrum measurements were carried out under gamma-ray (137Cs) excitation with a photomultiplier tube R7600U (Hamamatsu Photonics) connected to an ORTEC 113 preamplifier, an ORTEC 572 shaping amplifier and an Amptek Pocket MCA 8000 A multichannel analyzer for digital signal conversion. The bias voltage of the PMT was supplied at +600 V (ORTEC 556). The samples were mounted on the PMT with an optical grease (OKEN, 6262A) and covered with several layers of Teflon tape to collect scintillation photons. At the same time, decay time measurement was done using a digital oscilloscope (Tektronix TDS3052B). To evaluate the light yield, BGO crystal was used as a Ref. [10]. The radio-luminescence spectra measurement was performed at room temperature under X-ray irradiation. The excitation source was the RINT-2000 and the X-ray tube was supplied 40 kV and 40 mA. The emission spectra were measured using a Andor DU420-OE CCD cooled down to 213 K by a Peltier module. This CCD was coupled with ORIEL INSTRUMENTS monochromator with 285 grooves/mm and 280 nm blaze wavelength. The scintillation light was sent to CCD through a 2 m optical fiber to avoid direct irradiation of CCD by the X-ray. 3. Results and discussion 3.1. Crystal growth and crystallinity
Fig. 1. Schematic diagram of the modified l-PD method with a removable chamber system.
CsI crystals were successfully grown by the l-PD method with two kinds of crucibles and they are shown in Fig. 3. Both are trans-
D. Totsuka et al. / Optical Materials 34 (2012) 1087–1091
Fig. 3. Photographs of pure CsI single crystals, quartz-crucible-used (top) and graphite-crucible-used (bottom).
1089
Fig. 4. Powder X-ray diffraction patterns of CsI crystals. Black and red lines represent graphite-crucible and quartz-crucible-used crystals, respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
parent with a different shape. The physical dimensions of each crystal were 1 mm diameter and 15 mm in length for graphite crucible and 3 mm in diameter and 12 mm in length for quartz crucible, respectively. This difference of the diameter of the grown crystals could be due to the different wettability between CsI melt and the crucibles. In the case of graphite, the wettability seemed to be low because the solid/liquid interface was in the die and the diameter of grown crystal corresponded to that of the hole. On the other hand, the quartz material was better wetted by the CsI melt than the graphite material since the melt was hold on the bottom of the crucible and spread than the diameter of hole. The GDMS analysis showed the concentration of cationic impurities of the order of several ppm in both crystals. However, Br which could substitute the anion site was detected in large quantities around 0.8 mol% in the crystal grown in the graphite crucible and 0.1 mol% in the crystal grown in the quartz crucible. Many attempts to grow halide crystals were made using the furnace demonstrated in this study so far and the boiling point of bromides was proved to be lower than that of iodides, therefore Br contamination from the previous experiments could be observed. The powder XRD results are shown in Fig. 4. The diffraction patterns of both samples corresponded to the JCPDS chart #06-0311 and no impurity peaks were observed. The calculated lattice constants of CsI grown in quartz and graphite crucibles were 4.575 Å and 4.584 Å, respectively. This slight difference of the lattice constants could be affected by an interstitial-type lattice defect mainly due to Br ions. The crystal quality of the CsI samples grown by the l-PD method was characterized by (XRC) measurement and the results are shown in Fig. 5. It was difficult to perform accurate fitting of both XRC curves by a simple Gaussian. A bilateral symmetry was not observed and some peaks were clearly shown in the midslopes. Since the crystals were grown without the seed crystal, the grown samples would consist of sub-grain structure [11]. This became pronounced in the crystal grown in the quartz crucible. 3.2. Scintillation characteristics Fig. 6 shows the radio-luminescence spectra under the X-ray excitation. The spectra were normalized for the amplitude of the dominant emission peak at 300 nm. Water-doped CsI crystal prepared by the vertical BS method using a sealed ampoule was measured to identify the impurity emission peak. In all the CsI crystals, the emission peaks at 300 nm, 410 nm and 515 nm were observed.
Fig. 5. X-ray rocking curve of CsI crystals, graphite-crucible-used (top) and quartzcrucible-used (bottom).
It has been reported that the prominent emission band of 300 nm with a fast decay time is caused by STE [12]. The 410 nm luminescence band was observed in CsI:Br crystal [13] and it is related to a slow migration process with ls decay time. The ratio of this emission intensity with respect to the other emission peaks seems to be
1090
D. Totsuka et al. / Optical Materials 34 (2012) 1087–1091
Fig. 6. X-ray excited luminescence spectra of CsI crystals. Black line and red line show graphite-crucible-used and quartz-crucible-used, respectively. Blue line is the result of water-doped CsI crystal made by the vertical BS technique using sealed ampoule. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
different for each crystal. The crystal grown from the graphite crucible contained Br in a relatively high concentration, therefore we conclude that this peak is related to the Br impurity. The broader emission peak around 515 nm became the largest in the waterdoped CsI crystal. Thus, the broad peak around 515 nm can be ascribed to the water contamination. Fig. 7 shows the decay time profiles of both samples under 137Cs gamma-ray excitation. The decay times extracted from the fitting curves were determined to be 6 and 23 ns for the crystal grown in the graphite crucible and 7, 16 and 6033 ns for that grown in the quartz crucible, respectively. The component with several decay time is most probably related to the on-center-type STE (VK + e) recombination and the components with around 20 ns could be as ascribed to the off-center type STE (H + F) recombination [7]. The latter component was difficult to distinguish from the dominant peak around 300 nm in the shown radio-luminescence spectra because of poor resolution of spectrometer and CCD cam-
I(t) = 0.99 exp (-t/5.3ns) + 0.24 exp (-t/15.7ns)
Intensity (a.u.)
1
+ 0.04 exp (-t/6033ns) +0.03
0.1
0.01
0
100
200
300
400
500
Time (ns) I(t) = 1.04 exp (-t/5.0ns) + 0.32 exp (-t/20.1ns) + 0.03
Intensity (a.u.)
1
0.1
0.01
0
50
100
Time (ns) Fig. 7. Decay time profiles of graphite-crucible-used (top) and quartz-crucible-used (bottom), excited by 137Cs gamma-ray.
Fig. 8. Pulse height spectra under 137Cs irradiation. Black, red, and blue lines represent graphite-crucible-used, quartz-crucible-used, and BGO, respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
era. The slowest component with ls-order could be related to Br impurity. Fig. 8 shows 137Cs gamma-ray pulse height spectra of the pure CsI crystals compared with a BGO single crystal. The shaping time constant was set to be 2 ls for quartz-crucible crystal and 10 ls for graphite-crucible crystal. In all the samples, 662 keV photoelectron absorption peak was clearly observed. Light yields were determined by a comparison with BGO, which has the light output of 8000 ph/MeV. A correction factor taking into account the quantum efficiency of the photocathode was calculated for the corresponding positions of the emission peaks. Extracting the peak channel number from the single Gaussian function fitting of the photo-absorption peak and considering the quantum efficiency of the PMT of 30% at 300 nm (Fig. 6), a light yield of 3000 ph/MeV was deduced for the crystal grown from the quartz crucible. This value is well consistent with the previously reported results [14]. The MCA peak channel of crystal grown from graphite crucible was 1.8 times higher but it is difficult to evaluate exact light yield because this photo-absorption peak includes significant amount of slow components. 4. Conclusions Pure CsI single crystals were grown by the modified l-PD method using graphite and quartz crucibles. Transparent crystals were obtained. The higher wettability was observed for the quartz crucible. Many sub-grains were confirmed by XRC measurements, especially in the crystal grown from the quartz crucible. The strong STE emission consisting of two bands associated with the on-center-type STE (VK + e) and off-center type STE (H + F) were observed at 300 nm and their decay times were of several ns and around twenty ns, respectively. The light yield of these fast components was estimated to be 3000 ph/MeV for the crystal grown from the quartz crucible. Extrinsic luminescence was observed at 410 nm and 515 nm. The former band was related to Br and the latter to the water impurities. The scintillation decay time of the luminescence related to Br impurity was determined to be 6 ls. Acknowledgments This work was mainly supported by JST Sentan and partially by a Grant in Aid for Young Scientists (B)-15686001 and (A)-23686135 from the Ministry of Education, Culture, Sports, Science and Tech-
D. Totsuka et al. / Optical Materials 34 (2012) 1087–1091
nology of the Japanese government (MEXT). Partial assistance from the Yazaki Memorial Foundation for Science and Technology, Japan Science Society, Sumitomo Foundation, and Iketani Science and Technology Foundation are also gratefully acknowledged.
References [1] A. Yoshikawa, M. Nikl, G. Boulon, T. Fukuda, Opt. Mater. 30 (2007) 6–10. [2] Y. Yokota, V. Chani, M. Sato, K. Tota, K. Onodera, T. Yanagida, A. Yoshikawa, J. Cryst. Growth 318 (2011) 983–986. [3] M. Nikl, H. Ogino, A. Krasnikov, A. Beitlerova, A. Yoshikawa, T. Fukuda, Phys. Stat. Sol. (a) 202 (1) (2005) 4–6. [4] A. Yoshikawa, T. Yanagida, Y. Yokota, N. Kawaguchi, S. Ishizu, K. Fukuda, T. Suyama, K.J. Kim, J. Pejchal, M. Nikl, K. Watanabe, M. Miyake, M. Baba, K. Kamada, IEEE Trans. Nucl. Sci. 56 (2009) 3796–3799. [5] R. Simura, A. Yoshikawa, S. Uda, J. Cryst. Growth 311 (2009) 4763–4769.
1091
[6] W.M. Higgins, J. Glodo, E. Van Loef, M. Klugerman, T. Gupta, L. Cirignano, P. Wong, K.S. Shah, J. Cryst. Growth 287 (2006) 239–242. [7] H. Nishimura, M. Sakata, T. Tsujimoto, M. Nakayama, Phys. Rev. B 51 (1994) 2167. [8] B.K. Utts, S.E. Spagno, IEEE Trans. Nucl. Sci. NS-37 (1990) 134. [9] Y. Yokota, N. Kawaguchi, K. Fukuda, T. Yanagida, A. Yoshikawa, M. Nikl, J. Cryst. Growth 318 (2011) 908–911. [10] I. Holl, E. Lorenz, G. Mageras, IEEE Trans. Nucl. Sci. 35 (1998) 105–109. [11] N. Balamurugan, A. Arulchakkaravarthi, S. Selvakumar, M. Lenina, R. Kumard, S. Muralithard, K. Sivaji, P. Ramasamya, J. Cryst. Growth 286 (2006) 294–299. [12] Z. Wu, B. Yang, P.D. Townsend, J. Lumin. 128 (2008) 1191–1196. [13] G.I. Britvich, I.G. Britvich, V.G. Vasil’chenko, V.A. Lishin, V.F. Obraztsov, V.A. Polyakov, A.S. Solovjev, V.D. Ryzhikov, Nucl. Instrum. Meth. Phys. A 469 (2001) 77–88. [14] C. Amsler, D. Grogler, W. Joffrain, D. Lindelof, M. Marchesotti, P. Niederberger, H. Pruysa, C. Regenfus, P. Riedler, A. Rotondi, Nucl. Instrum. Meth. Phys. A 480 (2002) 494–500.