Boron codoping of Czochralski grown lutetium ...

Journal of Crystal Growth 486 (2018) 126–129

Contents lists available at ScienceDirect

Journal of Crystal Growth journal homepage: www.elsevier.com/locate/crys

Boron codoping of Czochralski grown lutetium aluminum garnet and the effect on scintillation properties Camera Foster a,b,⇑, Merry Koschan b, Yuntao Wu a,b, Charles L. Melcher a,b a b

Materials Science and Engineering, University of Tennessee, Knoxville, TN, USA Scintillation Materials Research Center, University of Tennessee, Knoxville, TN, USA

a r t i c l e

i n f o

Article history: Available online 1 February 2018 Communicated by T. Paskova Keywords: A1. Doping A2. Czochralski method B1. Rare earth compounds B1. Oxides B3. Scintillators

a b s t r a c t Many single crystal scintillators, such as Lu3Al5O12, have intrinsic defects that impede their performance. In addition to doping with activators such as cerium, codoping can be used to improve the scintillation properties of a variety of scintillators. In particular, boron has been shown to improve the light yield, energy resolution, and self-absorption of other garnet scintillators, such as GGAG, when incorporated into the lattice via codoping. In this study, single crystals of LuAG: 0.2 at.% Ce codoped with varying concentrations of boron were grown via the Czochralski method at a rate of 1.2 mm/h. Results will show the effect boron codoping has on the scintillation properties of LuAG: Ce, including light yield, decay time, and self-absorption. Ó 2018 Published by Elsevier B.V.

1. Introduction In many inorganic scintillators, efficient luminescence is dependent on dopants to create a radiative transition within the forbidden band [1,2]. For example, in medical imaging applications cerium is commonly used as an activator due to its 5d-4f transition which contributes to bright and fast scintillation with an emission wavelength that is suitable for photodetectors [3,4]. In addition to dopants, codopants have been used in order to mitigate the diminished scintillation properties that result from charge carrier traps and have been shown to improve light yield and decay time in some scintillators [4–7]. Codoping has also been reported to modify the growth behavior by altering the surface tension of the melt, as well as the mechanical strength of the material [8,9]. Donnald and Tyagi [10,11] have shown that boron codoping has an effect on single crystal Gd3Ga3Al2O12: Ce (GGAG: Ce) scintillators. Donnald discovered an absolute light output increase of about 10% due to improved energy migration that reduced charge carrier trapping at low temperature, which also improved the nonproportionality and subsequently the energy resolution [10]. Tyagi went on to show that the incorporation of B3+ as a codopant changes the defect structure in the lattice and reduces competing trap centers above

⇑ Corresponding author at: Department of Materials Science and Engineering, University of Tennessee, Knoxville, TN, 37996, USA. E-mail address: [email protected] (C. Foster). https://doi.org/10.1016/j.jcrysgro.2018.01.028 0022-0248/Ó 2018 Published by Elsevier B.V.

room temperature via charge compensation, ultimately improving the scintillation performance [11]. In this study we evaluate the impact that boron codoping has on another well-known garnet scintillator, cerium doped Lu3Al5O12 (LuAG). This compound has a desirable high density and relatively fast scintillation decay time, but its scintillation properties suffer due to the presence of electron traps arising from the existence of anti-site defects of LuAl in the lattice [12–15]. Divalent codoping with Ca2+, Mg2+, Sr2+, and Ba2+ has been carried out on LuAG: Ce and showed an increase in luminescence efficiency in samples grown via the micro-pulling down method[16]. Additionally, recent studies on LuAG: Ce, Li+ ceramics have been reported to exhibit promising scintillation characteristics by suppressing traps [17]. The motivation for this current work comes from the positive impact of boron codoping on the GGAG: Ce scintillator. We hypothesize that boron codoping in LuAG: Ce will have a similar effect on the light yield and defect structure. In this work we investigate a similar dopant concentration as in [10,11] with increasing concentrations of B3+ to determine whether incorporating boron in the LuAG: Ce matrix also affects the scintillation performance.

2. Experimental methods Four boules of cerium-doped LuAG were grown via the Czochralski method in a Cyberstar Oxypuller growth station using an automated system in which the derivative of the crystal weight was the process variable. The 4 N pure Lu2O3, a-Al2O3, CeO2, and H3BO3 raw materials were added directly to a 60 mm Ø iridium

127

C. Foster et al. / Journal of Crystal Growth 486 (2018) 126–129

crucible. Dopant and codopant were added in nominal concentrations of 0.2 at.% Ce and 0, 0.1, 0.2, and 0.4 at.% B with respect to the Lu atom. It should be noted that concentrations in the boule will differ due to segregation at the solid–liquid interface. Crystal growth was initiated on LuAG: Pr seed crystals oriented in the h1 1 1i direction. The atmosphere was primarily nitrogen with a small fraction of a percent oxygen. After growth the 32 mm diameter boules were cut into 555 mm pixels for measurements (Fig. 1). To mitigate any influence on the segregation of the dopants in the lattice, samples were gathered from a similar portion of the boule length. In order to calculate the light yield of each sample, pulse height spectra for each composition were collected using a pulse processing chain consisting of an R2059 photomultiplier tube (PMT), a Canberra model 2005 pre-amplifier, an Ortec 672 amplifier, and a Tukan 8k multi-channel analyzer [18]. Each sample was excited with a 10 lCi Cs-137 (662 keV) source and was coupled to the PMT with Corning optical grease. Light capture was enhanced by covering five sides of each sample with multiple layers of Teflon tape, and a reflective Spectralon dome was placed on top. An annealing study was also carried out by heating these samples in a CM Model 1730-12 horizontal tube furnace in an oxidizing (air) atmosphere for a soak time of 10 h at 1300 °C. After annealing samples were kept in containers to prevent light exposure before measurements, similar to the procedure reported in Tyagi et al. [11]. Scintillation decay times were measured using two R2059 Hamamatsu PMTs and a 137Cs gamma source in the configuration described by Bollinger and Thomas [19]. The radioluminescence was excited by x-rays from a CMX003 X-ray generator operated at 35 kV and 0.1 mA with an Acton SP-2155 monochromator within the spectral range of 200–800 nm. Absorbance measurements were completed on a Varian Cary 5000 UV–Vis-NIR Spectrophotometer on polished samples 1 mm thick. Photoluminescence (PL) emission and excitation spectra were acquired with a HORIBA Jobin Yvon Fluorolog-3 Spectrofluorometer using a 450 W continuous Xenon lamp as the excitation source.

3. Results and discussion Fig. 1 depicts a LuAG: Ce and LuAG: Ce, B crystal grown in this study under UV excitation, as well as the cut 555 mm pixels of all compositions used for measurements. The concentrations noted in Fig. 1 are the nominal concentrations of the dopant and codopant that were added directly to the crucible. While the concentration in the finished crystal has not been determined, we assume due to its 3+ oxidation state and small ionic radius that boron will readily incorporate into the crystal lattice, and is likely to have a larger distribution coefficient in LuAG than the larger Ce3+ ions. The artifacts near the bottom of the longer boule are cracks initiated during extraction at growth termination. Each crystal cracked in a similar way during extraction; however, the majority of the boule was crack-free and very transparent samples were able to be collected for analysis of each composition. Gamma-ray pulse height spectra were collected for each sample to determine the absolute light yields which are shown in Table 1. The sample containing the nominal boron concentration of 0.1 at.% had the largest increase in light yield, 10%, which is similar to the increase previously observed for boron codoped GGAG [10]. The other boron concentrations resulted in smaller increases in light yield, but in all cases the light yield of the codoped samples was higher than that of Ce-only LuAG. Overall, the light yields of these samples were lower than some previously reported values for LuAG: Ce. We attribute this to the fact that both the Ce concentration and the growth conditions in this study followed the earlier work on GGAG and were not optimized specifically for LuAG. It is typical for each scintillator compound to require optimization of both the activator concentration and the growth conditions in order to achieve the best performance. In this study, we were more interested in a direct comparison to a previously studied garnet scintillator, and we therefore followed the same procedures. Future work will pursue optimization of the synthesis parameters specifically for LuAG. A comparison of light yield before and after annealing can be seen in Fig. 2. The reduction of light yield after air annealing is similar to the reduction previously reported by Tyagi, et al. for boron codoped GGAG, in which charge carrier traps are emptied during annealing. However, after exposure to light for 24 h, the light yield did not return to the original value as was reported in reference [11]. The charge carrier traps found in GGAG: Ce, B are filled after light exposure, but the situation is quite different for LuAG: Ce, B. White light exposure can fill some traps spatially correlated with cerium ions via optical absorption of Ce3+ [20]. In LuAG, because of the dominant LuAl anti-site defect, located at 0.29 eV below the conduction band minimum is electrically neutral, it is hardly filled during light exposure [21]. Further investigation of the defect profiles of these samples is needed to form a definitive explanation on this topic. The decay times for LuAG single crystals exhibited both a primary slow component and a secondary fast one, as displayed in Fig. 3 with values in Table 1. The slow component has been attributed to the capture of free charge carriers in shallow traps, causing delayed recombination of the Ce3+ activation site [13,22]. Com-

Table 1 Absolute light yield and scintillation decay time profiles for each sample and their respective weights.a

Fig. 1. As grown boule of LuAG: Ce, B, 80 mm in length, with a cut LuAG: Ce crystal below it for comparison under 366 nm UV excitation (top). Pixels 555 mm in dimension used for measurements (bottom).

a

Boron concentration (%)

Light yield, ph/MeV

Fast s, ns

Slow s, ns

0 0.1 0.2 0.4

10,386 11,444 11,002 10,963

59 41 70 35

462 396 536 413

(12%) (11%) (16%) (10%)

All values were collected on as-grown samples of 555 mm size.

(88%) (89%) (84%) (90%)

128


Fig. 2. Chart depicting light yield as a function of boron concentration before (black) and after air annealing (red) for 10 h at 1300 °C, as well as after light exposure for 24 h (blue). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 3. Emission wavelengths of LuAG: Ce, B pixels under 30 kV, 40 mA X-ray excitation. The emission between 250 and 450 nm is due to host defects in LuAG and the emission at 530 nm is due to Ce3+.

pared to the non-codoped sample, there was a decrease of both the short and long decay time constants for 0.1 and 0.4 at.% B. The 0.2 at.% B sample had slower decay constants of 536 ns and 70 ns. The generally slower decay of this sample suggests a higher density of shallow charge carrier traps that can slow the migration of electrons and holes to the Ce activator centers. Possible origins of the shallow traps include antisite defects and oxygen vacancies, both of which have been reported previously in LuAG, and both of which are sensitive to small unintended variations in crystal growth parameters. Future thermoluminescence studies may improve our understanding of the nature of the traps and provide insight into the factors that affect scintillation time profiles in LuAG. Radioluminescence of LuAG: Ce, B was investigated via the collection of emission spectra under X-ray excitation (Fig. 4). All samples showed the characteristic Ce3+ emission at a wavelength of 530 nm. The emission peaks at 300 and 370 nm are attributed to host lattice defects, such as the LuAl anti-site defect [12]. As the boron concentration increased, the intensity of the defect peaks

Fig. 4. Emission wavelengths of LuAG: Ce, B pixels under 30 kV, 40 mA X-ray excitation. The emission between 250 and 450 nm is due to host defects in LuAG and the emission at 530 nm is due to Ce3+.

Fig. 5. Absorbance spectra of LuAG: Ce showing the effect of increasing boron concentration. The peaks at 340 nm and 450 nm correspond to the Ce 4f-5d1 and Ce 4f-5d2, transitions respectively.

changed only slightly, if at all, implying that B codoping does not suppress the existence of host lattice defects in LuAG:Ce. In order to look more closely at the impact of codoping on the efficiency of the activation center, absorption and PL excitation and emission measurements were completed. Fig. 5 depicts the absorption of these samples with increasing boron codopant concentration. The location of the peaks positioned at 340 and 450 nm, corresponding to the Ce 4f-5d1 and Ce 4f-5d2 transitions, respectively, did not shift with boron concentration. The small variations in intensity are not related to the concentration of boron but rather to the abundance of Ce3+ sites and the experimental uncertainty within each measurement [12]. In Fig. 6, the PL excitation and emission are shown. Similarly to the absorption spectra, the samples all were found to have the characteristic excitation from the 4f-5d transitions of cerium. In addition, the samples had a broad photoluminescent emission wavelength with two convoluted peaks centered at 505 and 520 nm that are due to the splitting of the Ce3+ 4f ground state [23].


129

codoping in other garnet scintillators; however, this codopant does not affect the dominating defect structure of LuAG, i.e. anti-site defects, and requires more research in order to discover its role in the scintillation mechanism. Acknowledgements This work was partially supported by Siemens Medical Solutions, Molecular Imaging, USA and the Tennessee Higher Education Commission. References

Fig. 6. Photoluminescence curves of LuAG: Ce with increasing boron concentration. The excitation was measured at 530 nm emission and the emission spectra were collected under a 340 nm excitation wavelength. The inset figure emphasizes the ‘‘shelf” residing at 350–450 nm due to F+ center defects.

The self-absorption of these peaks was not improved with the incorporation of boron into the matrix, unlike the results presented in Tyagi, et al. Additionally, when looking at the spectra, there appears to be a ‘‘shelf” residing around 350–450 nm. The intensity for this characteristic peak increased with boron concentration and is attributed to an F+ center defect, or electron trapped in an oxygen vacancy [23,24]. 4. Conclusion In this study, we found that boron codoping in LuAG: Ce produces a 10% increase in light yield for 0.1% B and an approximately 6% increase for 0.2% and 0.4% B. This is similar to previous reports by both Tyagi and Donnald that incorporating boron into GGAG: Ce, can improve the scintillation light yield by around 10% [10,11]. While the light yield was modestly increased with boron codoping, the defect behavior of LuAG: Ce, B was not found to be similar to that of GGAG: Ce, B. In particular, the boron codoping did not affect the anti-site defect of the LuAG host and did not reduce self-absorption in the PL spectra. The reduction in light yield after air annealing was similar to that reported for GGAG: Ce, B, although the light yield could not restored after light exposure as reported for GGAG: Ce, B. The scintillation decay time was accelerated for all samples except for LuAG: Ce, 0.2% B. The incorporation of boron into the crystal did not significantly affect the wavelength of either the host lattice defect emission or Ce3+ emission. The wavelengths of absorbance and PL are unchanged with increasing boron concentration, as expected; however, the PL emission exhibits a ‘‘shelf” at 400 nm, attributed to an F+ center that increases with boron concentration. Ultimately, it was found that boron codoping in the LuAG: Ce matrix has an effect on scintillation light yield similar to that of B

[1] P. Lecoq, Inorganic Scintillators for Detector Systems, Springer, 2006. [2] G.F. Knoll, Radiation Detection and Measurement, John Wiley & Sons, 2010. [3] M. Nikl, Scintillation detectors for x-rays, Measure. Sci. Technol. 17 (4) (2006) R37. [4] C. Ronda, Review – scintillators for medical imaging: a tutorial overview, ECS J. Solid State Sci. Technol. 5 (1) (2016) 5. [5] M.A. Spurrier, P. Szupryczynski, K. Yang, A.A. Carey, C.L. Melcher, Effects of Ca2 + co-doping on the scintillation properties of LSO:Ce, IEEE Trans. Nucl. Sci. 55 (3) (2008) 1178–1182. [6] Y. Wu, F. Meng, Q. Li, M. Koschan, C.L. Melcher, Role of Ce4+ in the scintillation mechanism of codoped Gd3Ga3Al2O12:Ce, Phys. Rev. Appl. 2 (4) (2014) 044009. [7] Y. Wu et al., Defect engineering by codoping in KCaI3: Eu2+ single-crystalline scintillators, Phys. Rev. Appl. 8 (3) (2017) 034011. [8] C.L. Melcher, Scintillator Design Via Codoping, vol. 11, ed: JPS Conf. Proc., 2016, p. 8. [9] M.A. Spurrier, P. Szupryczynski, H. Rothfuss, K. Yang, A.A. Carey, C.L. Melcher, The effect of co-doping on the growth stability and scintillation properties of lutetium oxyorthosilicate, J. Cryst. Growth 310 (7) (2008) 2110–2114. [10] S.B. Donnald, The effect of B3+ and Ca 2+ co-doping on factors which affect the energy resolution of Gd3Ga3Al2O12: Ce, IEEE Trans. Nucl. Sci. 60 (5) (2013) 5. [11] M. Tyagi, Effect of codoping on scintillation and optical properties of a Cedoped Gd3Ga3Al2O12 scintillator, J. Phys. D: Appl. Phys. 46 (475302) (2013) 12. [12] M. Nikl, The antisite LuAl defect-related trap in Lu3Al5O12: Ce single crystal, Phys. Stat. Sol. (B) 242 (14) (2005) 3. [13] K. Blazek, A. Krasnikov, K. Nejezchleb, M. Nikl, T. Savikhina, S. Zazubovich, Luminescence and defects creation in Ce3+-doped Lu3Al5O12 crystals, Physica status solidi (b) 241 (5) (2004) 1134–1140. [14] M. Nikl et al., Shallow traps and radiative recombination processes in Lu3Al5O12 single crystal scintillator, Phys. Rev. B 76 (19) (2007) 195121. [15] M. Nikl, Development of LuAG-based scintillator crystals – a review, Prog. Cryst. Growth Characteriz. Mater. 59 (2013) 26. [16] K. Kamada, Co-doping effects on luminescence and scintillation properties of Ce doped Lu3Al5O12 scintillator, Nucl. Instrum. Meth. Phys. Res. A (782) (2015) 4. [17] S. Liu et al., Effect of Li+ ions co-doping on luminescence, scintillation properties and defects characteristics of LuAG: Ce ceramics, Optic. Mater. 64 (Supplement C) (2017) 245–249. [18] Z. Guzik, S. Borsuk, K. Traczyk, M. Plominski, TUKAN-an 8K pulse height analyzer and multi-channel scaler with a PCI or a USB interface, IEEE Trans. Nucl. Sci. 53 (1) (2006) 231–235. [19] G.T.L. Bollinger, Measurement of the time dependence of scintillation intensity by a delayed coincidence method, Rev. Scientific Instrum. 32 (1961) 7. [20] P. Dorenbos, C.W.E.V. Eijk, A.J.J. Bos, C.L. Melcher, Afterglow and thermoluminescence properties of Lu2SiO5:Ce scintillation crystals, J. Phys.: Condensed Matter 6 (22) (1994) 4167. [21] M. Fasoli et al., Band-gap engineering for removing shallow traps in rare-earth Lu3Al5O12 garnet scintillators using Ga3+ doping, Phys. Rev. B 84 (8) (2011) 081102. [22] W. Chewpraditkul et al., Scintillation properties of LuAG:Ce, YAG: Ce and LYSO: Ce crystals for Gamma-ray detection, IEEE Trans. Nucl. Sci. 56 (6) (2009) 3800–3805. [23] M. Nikl et al., Traps and timing characteristics of LuAG:Ce3+ scintillator, Physica Status Solidi (a) 181 (1) (2000) R10–R12. [24] M. Nikl, V.V. Laguta, A. Vedda, Energy transfer and charge carrier capture processes in wide-band-gap scintillators, Physica status solidi (a) 204 (3) (2007) 683–689.