Subscriber access provided by UNIV OF LOUISIANA
Functional Inorganic Materials and Devices
Unravelling the critical role of site occupancy of lithium codopants in Lu2SiO5:Ce3+ single-crystalline scintillators Yuntao Wu, Jing Peng, Daniel Rustrom, Merry Koschan, Camera Foster, and Charles L. Melcher ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.8b19040 • Publication Date (Web): 22 Jan 2019 Downloaded from http://pubs.acs.org on January 26, 2019
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Applied Materials & Interfaces
Unravelling the critical role of site occupancy of lithium codopants in Lu2SiO5:Ce3+ single-crystalline scintillators Yuntao Wu,*,a,b Jing Peng,c Daniel Rustrom,a,b Merry Koschan,b Camera Foster,a,b and Charles L. Melcher a,b,d,e a.
Scintillation Materials Research Center, University of Tennessee, Knoxville, TN 37996, USA Department of Materials Science and Engineering, University of Tennessee, Knoxville, TN 37996, USA c. Department of Chemical & Biomolecular Engineering, University of Tennessee, Knoxville, TN 37996, USA d. Bredesen Center for Interdisciplinary Research and Graduate Education, University of Tennessee, Knoxville, TN 37996, USA e. Department of Nuclear Engineering, University of Tennessee, Knoxville, TN 37996, USA b.
ABSTRACT: Lithium codoping has emerged as an effective strategy to enhance the light yield of oxide scintillators for radiation detection applications, but the understanding of the actual role played by Li+ remains unclear. In this work, we comprehensively study the effects of Li codoping on optical and scintillation properties of Lu2SiO5:Ce (LSO:Ce) single crystals, and reveal the critical role of site occupancy of Li. High quality LSO:Ce single crystals codoped with 0.05, 0.1 and 0.3 at% Li ions were grown by the Czochralski method. The optical absorption spectra confirm non-conversion of stable Ce3+ to Ce4+ in Li-codoped LSO:Ce regardless of Li codoping concentration. The photoluminescence decay kinetics suggest an enhanced ionization of the excited 5d1 state of Ce3+ centers in highly codoped samples. A simultaneous improvement of scintillation light yield, decay time, and afterglow is achieved in LSO:Ce codoped with low concentrations of Li. The preferential occupation of Li at interstitial spaces and lutetium sites is proven to rely on its codoping concentration by using 7Li nuclear magnetic resonance technique. The concentration-dependent site occupancy of Li alters the defect structures of LSO:Ce, in particular resulting in a distinct change in the number of cerium-spatially-correlated oxygen vacancies confirmed by thermoluminescence and afterglow measurements. KEYWORDS: scintillator, single crystal, codoping, site occupancy, defect structure
INTRODUCTION Radiation detection, as an important technology for detecting and identifying high-energy particles and rays, has broad and growing use for diagnostic nuclear medicine, nuclear security, particle physics, and astrophysics. A critical objective in developing radiation detectors is production of high-performance radiation detection materials. Semiconductor materials, such as High-Purity Germanium (HPGe) and the room temperature semiconductor Cd0.9Zn0.1Te (CZT), are of great interest because of their high sensitivity and excellent energy resolution. However, the HPGe detectors must be operated under cryogenic conditions to eliminate thermal noise, and they require highly accurate supporting electronics.1 The development of CZT is mainly limited by its problematic crystal growth. Its solid-solution nature leads to inhomogeneity in composition and the precipitation of tellurium inclusions during crystal growth.2 So far, these high-performance detection materials are not widely deployed due to their nonportability and high cost. Inorganic scintillators, another important type of radiation detection materials, have been intensively developed and widely applied due to their relatively low cost and good performance, such as the well-known NaI:Tl 3 and Lu2SiO5:Ce.4 To meet the increasing performance requirements of detectors, over the past twenty years several notable oxides and halides were discovered and developed,5 including Gd3Ga3Al2O12:Ce,6 (Lu0.75Y0.25)3Al5O12:Pr,7 LaBr3:Ce,8 SrI2:Eu,9 KSr2I5:Eu,10 CsSrI3:Eu,11 KCaI3:Eu,12 and KCa0.8Sr0.2I3:Eu.13-15 Ion codoping is also known as a useful method for engineering properties of existing inorganic scintillators for targeted applications:16
i) In positron emission tomography (PET), the timing resolution and light yield of oxyorthosilicates 17,18 and aluminate garnets 19-23 were improved by divalent ion codoping (e.g. Ca2+ and Mg2+). These improvements were attributed to both the introduction of stable Ce4+ ions with a much faster emission, which bypasses the initial holetrapping step of the stable Ce3+ scintillation emission under irradiation,18,19,22 and to the suppression of charge carrier trapping at defects.19,22,24 ii) In computed tomography (CT), the persistent afterglow of commercialized thallium activated cesium iodide can be diminished by Eu2+, Sm3+, or Yb2+ codoping.25-27 These codopants can introduce shallow and/or deep energy levels in the forbidden-gap of the host. The trapping processes occurred at these energy levels will interference the migration and relaxation processes of free charge carriers.25-27 The afterglow of the recently developed KCaI3:Eu was reduced by orders of magnitude via Sc3+ codoping.28 This positive effect is a consequence of the preferentially formed Sc3+ interstitial, which can suppress the formation of iodine vacancies that act as deep electron traps without introducing any additional deep electron traps.28 iii) For gamma camera and single photon emission computed tomography (SPECT) applications the energy resolution of NaI:Tl, the workhorse scintillator widely employed in these imaging systems, can be improved by Sr2+ or Ca2+ codoping which suppresses undesirable slow components of the scintillation process.29 For gamma-ray spectroscopy applications, the energy resolutions of cutting-edge halide scintillators CeBr3, LaBr3:Ce, and KCaI3:Eu were improved to 2%, 3%, and 2.7% respectively at 662 keV by Sr2+ and Zr4+ codoping.30-32 These improvements were attributed to the reduction of the Auger quenching.32-34 The charge
ACS Paragon Plus Environment
ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
carrier population in halide scintillators can be significantly reduced by non-linear quenching due to the Auger recombination, and as a consequence, the scintillation yield measured at different gammaray energies will show an non-proportional response. The introduction of shallow traps by codopants can temporally capture the electrons/holes to enhance the carrier separation, and lead to a reduction of non-linear quenching.32-34 In recent years, lithium codoping has emerged as a useful approach to enhance the scintillation light yield of many high temperature oxides, such as aluminate garnets and oxyorthosilicates.35-38 For example, a 1.6 fold increase of light yield in Y3Al5O12:Ce (YAG:Ce)35 and a 30% increase in Lu3Al5O12:Ce (LuAG:Ce)36 were achieved with Li codoping. We recently reported that the light yield of (Lu0.75Y0.25)3Al5O12:Pr was enhanced by around 60% upon Li codoping; an additional effect of the high light yield is a significant improvement in energy resolution from 4.8 to 4.1% at 662 keV, the best value ever reported for single-crystalline oxide scintillators.37 Motivated by the enhanced performance achieved in garnet scintillators, a codoping study using low levels of Li was conducted on Lu2SiO5:Ce single crystals in our lab. In a recent research letter,38 we reported a simultaneous improvement of afterglow, light yield, and decay time of LSO:Ce by 0.05 at% Li codoping. There have been many studies that have found different effects of monovalent Li codoping on the oxidation state of activators and defect structures compared to divalent codoping,35,36,39-41 but the understanding of the role played by Li ions in oxide scintillators remains vague. In Ref. 38, we presented a preliminary study on the effect of Li codoping on LSO:Ce, and found that Li ions prefer to occupy interstitial spaces (Lii) rather than substitutional sites (LiLu) in LSO:Ce codoped with low levels of Li. In this current work, we aim to fully unravel the critical role of Li codopants in the scintillation mechanism of LSO:Ce by clarifying the internal correlation between the oxidation state of cerium, relevant defect structures, and the site occupancy of Li in LSO:Ce single crystals codoped with different Li concentrations, such as 0.05, 0.1 and 0.3 at%. We will first discuss the correlation between codopant concentration and scintillation performance; second, the influence of Li codoping concentration on the oxidation state of cerium and the luminescence characteristics of the Ce1 and Ce2 centers are investigated by optical absorption spectroscopy, photoluminescence spectra and decay time, and X-ray excited radio-luminescence (RL) spectra; third, the effect of Li codoping concentration on the defect structures is studied by thermoluminescence (TL) and afterglow measurements; finally, the Li site occupancy and its critical role were revealed by the 7Li nuclear magnetic resonance technique.
EXPERIMENTAL METHODS Crystal growth LSO:0.1 at% Ce crystals codoped with 0, 0.05, 0.1 and 0.3 at% Li were grown by the Czochralski method using Lu2O3, SiO2, CeO2, and Li2CO3 starting materials with at least 99.99% purity. The atomic percentage of the dopant and codopant concentrations is relative to that of lutetium in the melt. As is typical in Czochralski growth with dopants that have low segregation coefficients, the dopant concentrations in the finished crystal will be lower than the nominal concentration in
Page 2 of 21
the melt. The molten raw material was contained in iridium crucibles that were inductively heated by a 7 kHz Hüttinger power supply. The growth atmosphere was a mixture of nitrogen and a small fraction of oxygen. All as-grown boules are 32 mm in diameter and 110 mm long except for the 0.3 at% Li-codoped boule which was 50 mm long (Fig. 1a-c). As observed in Fig. 1(d), all 5 mm3 polished samples luminesce strongly under a hand-held ultraviolet lamp (em=360 nm) except for the 0.3 at% Li-codoped sample. To ensure that samples have the same Ce concentration and proportional Li concentrations, all samples were cut from their respective boules at the same distance below the end of the seed.
Figure 1. As-grown boules of Li codoped LSO:Ce: (a) 0.05 at% Li, (b) 0.1 at% Li, and (c) 0.3 at% Li. (d) 5 mm3 samples under UV irradiation.
Optical property measurements Optical absorption measurements as a function of wavelength were made with a Varian Cary 5000 UV-VIS-IR spectrophotometer in the same manner as reported in Refs. 24 and 38. Photoluminescence emission (PL) and excitation (PLE) measurements were carried out with a HORIBA Jobin Yvon Fluorolog-3 spectrofluorometer. A 450W continuous xenon lamp was utilized as the excitation source. Photoluminescence decay profiles were conducted on the same spectrofluorometer using a time-correlated-single-photon counting module. The excitation sources are HORIBA Jobin Yvon NanoLEDs (pulsed light-emitting diodes). The duration of the light pulse was shorter than 2 ns. Scintillation property measurements The scintillation decay profile was acquired using a timecorrelated single-photon counting equipment excited by a 22Na source.42 The absolute light yield (the number of photons per energy unit created by gamma-rays in the scintillator) was evaluated using a pulse processing chain consisting of a Hamamatsu R2059 photomultiplier tube (PMT), an Ortec 672 Amp, a Canberra model 2005 pre-Amp and a Tukan 8k multi-channel analyzer. The shaping time was 10 s to provide full light integration. Each sample was measured under irradiation with
ACS Paragon Plus Environment
Page 3 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Applied Materials & Interfaces
a 15 Ci 137Cs source. The sample was directly coupled to the PMT. A few drops of mineral oil were used as the couplant between the sample and the PMT. To maximize the light collection, a 50 mm diameter PTFE-lined dome-shaped reflector was used as a top reflector. The reproducibility of the LY measurements is ±5%. A Hamamatsu R6231-100 high quantum efficiency PMT was used to evaluate the energy resolution at 662 keV. The energy resolution is expressed as a percentage (E/E) of the full width at half maximum of a given energy peak divided by the location of the peak centroid. Radioluminescence (RL) emission spectra were acquired in transmission mode under excitation of a copper X-ray source (X-ray Model; CMX003) operated at 35 kV and 0.1 mA. The emission spectra were recorded with a 150 mm focal length monochromator and a broadband PMT. The emission intensities were corrected for the spectral sensitivity of the PMT. The afterglow profiles of the samples were acquired with a Hamamatsu R2059 PMT operated at -1500 Vbias. A Tetratex TX3104 PTFE membrane was used as the reflector. Prior to the afterglow measurements, the samples were held in the dark for 24 hours, and then irradiated at room temperature with x-rays for 15 min while being held in the dark. The luminescence of each sample was recorded as a function of time beginning immediately after irradiation. Nuclear magnetic resonance measurements 7Li NMR spectra were recorded with a Bruker Avance III magnet at 155.5 MHz using a 3.2 mm MAS broad band probe (zirconia rotors). The experiments were performed with the samples spinning at 15 kHz at 20˚C. The 7Li chemical shift was referenced to 1M LiCl in distilled water at 0 ppm. A single pulse sequence was employed to record this data with a recycle delay (d1) of 10 s. About 100000, 50000 and 25000 scans were accumulated for LSO:Ce single crystals codoped with 0.05, 0.1, and 0.3 at% Li, respectively, in order to achieve a reasonable signal-to-noise ratio. Thermoluminescence measurements Thermoluminescence (TL) measurements were performed in a temperature range from 10 to 550 K. After X-ray irradiation at 10 K with an X-ray tube operated at 35 kV and 0.1 mA for 15 min, the TL glow curve was recorded by heating the sample at a rate of 3 K/min. Prior to the TL measurements, the samples were individually heated to 550 K to ensure that all traps in the temperature range were empty. A Hamamatsu R2059 PMT optically coupled to a cryostat’s borosilicate window was used to record the spectrally unresolved emission from the sample. Standard NIM electronics were used to convert the PMT current signal into a voltage signal. The voltage signal was digitized by a National Instruments 6002-E data acquisition card. The sample temperature and the signal intensity were correlated by software that was developed in-house.
RESULTS AND DISCUSSION Li concentration dependence of scintillation performance
Pulse height spectra of 5 mm3 non-codoped and Li-codoped LSO:Ce samples acquired by a Hamamatsu R2059 PMT under 137Cs gamma-ray source irradiation are plotted in Fig. 2. The absolute light yield was estimated by using the single photoelectron method 43 while considering the emission-weighted quantum efficiency of the PMT. The light yield of the non-codoped LSO:Ce sample is about 32,000 photons/MeV, close to the published value for LSO:Ce.44 A remarkable enhancement of the light yield to 39,000 photons/MeV was achieved by 0.05 at% Li codoping.38 As the Li concentration was further increased, the light yield slightly decreased to 37,000 photons/MeV in the 0.1 at% Li codoped sample, and substantially dropped to a few hundred photons/MeV in the 0.3 at% Li-codoped sample. The pulse height spectra acquired for energy resolution evaluation are shown in Fig. 3. The energy resolution defines the capability of a scintillator to distinguish between ionizing radiations of slightly different energies. The energy resolution of LSO:Ce at 662 keV slightly improves from 10% for non-codoped sample to 9.7% for both 0.05 at% and 0.1 at% codoped samples.
Figure 2. 137Cs Pulse height spectra acquired by a Hamamatsu R2059 PMT for non-codoped, 0.05, 0.1, and 0.3 at% Li-codoped LSO:Ce single crystals. Note: in Fig. 2, no photopeak can be identified for the 0.3 at% Li-codoped sample. Its absolute light yield was evaluated by acquiring a pulse height spectrum using a 40 times higher gain in amplifier.
Figure 3. 137Cs Pulse height spectra acquired with a Hamamatsu R6231-100 PMT for non-codoped, 0.05, and 0.1 at% Li-codoped LSO:Ce single crystals. The energy resolution of each sample evaluated at 662 keV is also plotted.
The scintillation decay profiles of non-codoped, 0.05, 0.1, and 0.3 at% Li-codoped LSO:Ce samples are plotted in Fig. 4. All curves can be well fit by a single exponential function. The scintillation
ACS Paragon Plus Environment
ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
decay time is continuously shortened as the Li concentration increases. The decay decreases from 45.4 ns for non-codoped, 42.1 ns for 0.05 at% Li, 41.4 ns for 0.1 at% Li, to 25 ns for 0.3 at% Licodoped sample. Orders of magnitude higher background in 0.3 at% Li codoped sample suggests its stronger afterglow level.
Figure 4. Scintillation decay profiles of non-codopd, 0.05, 0.1, and 0.3 at% Li-codoped LSO:Ce single crystals under 22Na irradiation.
Evidence of non-conversion of stable Ce3+ to Ce4+ after Li codoping Aliovalent codoping tends to change the valence of activation ions in luminescence materials. Optical absorption spectroscopy was utilized to study the influence of monovalent Li codoping on the oxidation state of cerium ions in LSO:Ce. The optical absorption spectra of non-codoped and Li-codoped samples are shown in Fig. 5. As confirmed by X-ray absorption spectroscopy 45 and electron energy loss spectroscopy 38, the cerium oxidation state in as-grown LSO:Ce single crystals is purely Ce3+. As expected, in the noncodoped LSO:Ce sample only absorption bands corresponding to the transitions from the 4f ground state to the 5d sublevels of stable Ce3+ ions are observed. In contrast to the effect of Ca2+ or Mg2+ codoping on the conversion of stable Ce3+ to Ce4+ by,17,18 all of the Li-codoped LSO:Ce samples, regardless of Li concentration, show solely the Ce3+ 4f-5d absorptions, indicating that there was no conversion of stable Ce3+ to Ce4+.
Figure 5. Optical absorption spectra of 5 mm3 non-codoped, 0.05, 0.1, and 0.3 at% Li-codoped LSO:Ce single crystals.
Page 4 of 21
non-codoped LSO:Ce peak at 3.15 and 2.9 eV for Ce1 centers, and 2.64 eV for Ce2 centers. 48-51 The typical PL lifetimes of the Ce1 and Ce2 centers are 33 and 46 ns, respectively.50,51 To clarify the effects of Li codoping on the luminescence characteristics of the Ce1 and Ce2 centers, the PL, PLE and RL spectraand the decay kinetics of all Li-codoped samples were investigated. The PL and PLE spectra of Ce1 and Ce2 centers in noncodoped and Li-codoped LSO:Ce are normalized to the peak maximum for a better comparison (see Fig. 6(a-d)). The shape of the excitation spectra of the Ce1 and Ce2 centers in the 0.3 at% Licodoped sample is different from that of non-codoped, 0.05 at%, and 0.1 at% Li-codoped samples. For the 0.3 at% Li-codoped sample, the intensity of the excitation bands from the lowest 4f level to the higher-lying 5d levels is lower than that of the excitation band from the lowest 4f level to the 5d1 level. The emission spectra of the noncodoped and Li-codopd samples can be well fitted with three Gaussian components with peak positions at 3.15 eV (393 nm), 2.91 eV (426 nm), and 2.64 eV (470 nm).50,51 The relative contribution of the Ce2 emission significantly decreased in the 0.1 at% Li-codoped sample, and is almost undetectable in the 0.3 at% Li-codoped sample. Because no stable Ce4+ was induced by Li codoping, a similar trend in reduction of the Ce2 emission was observed in the RL spectra (see Fig. 7). The ratio of the integral emissions of Ce2 and Ce1 decreases from 0.91 for non-codoped LSO, 0.45 for 0.05 at% Li, 0.43 for 0.1 at%, to 0.03 for the 0.3 at% Li codoped sample. The photoluminescence decay profiles of Ce1 and Ce2 emissions in non-codoped and Li-codoped LSO:Ce are shown in Fig. 8(a) and (b). All decay profiles can be fitted by a single exponential function. The decay constant of the Ce1 emission first increases from 33.2 ns for non-codoped to 35.0 ns for 0.05 at% Li, then gradually decreases to 32.3 ns for 0.1 at% Li, and 24.8 ns for 0.3 at% Li. The Ce2 emission shows a continuous shortening of the decay constant from 46.1 ns for non-codoped, 41.5 ns for 0.05 at% Li, 41 ns for 0.1 at% Li, to 20.5 ns for 0.3 at% Li. The noticeable increase in the background value relative to the decay amplitude for highly codoped samples suggests an enhanced ionization of the excited 5d1 level of the Ce3+ centers, since the resulting delayed recombination processes are spread to much longer time scales than the delay between two subsequent excitation pulses. For LSO:Ce (or LYSO:Ce), the thermal ionization process from the Ce3+ lowest energy 5d levels to the conduction band at room temperature has been proven by a photoconductivity method 52 and a purely optical method 53. The accelerated PL decay times of the Ce1 and Ce2 centers also suggests an increased probability of non-radiative recombination of Ce1 and Ce2 through the ionization process, because such a process leads to a decay time shortening but leaves an electron in the conduction band and a temporary Ce4+ behind. That electron can then subsequently radiatively recombine with the other temporary Ce4+ centers, resulting in delayed emissions. However, at this time, the cause of the prolonged decay time of Ce1 emission for the 0.05 at% Li-codoped sample is not clear.
Effect of Li codoping on luminescence characteristics of Ce1 and Ce2 centers In the LSO host there are two distinct occupation sites for Ce ions, namely the seven-oxygen coordinated site (Ce1) and the sixoxygen coordinated site (Ce2).46,47 The Ce3+ 5d-4f emissions in
ACS Paragon Plus Environment
Page 5 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Applied Materials & Interfaces
Figure 6. Normalized PL and PLE spectra of Ce1 and Ce2 centers in (a) non-codoped, (b) 0.05 at% Li, (c) 0.1 at% Li, and (d) 0.3 at% Li-codoped LSO:Ce samples.
Figure 7. X-ray excited RL spectra of (a) non-codoped, (b) 0.05 at% Li, (c) 0.1 at% Li, (d) 0.3 at% Li-codoped LSO:Ce samples. All RL spectra can be well fitted with three Gaussian peaks at 3.15 eV (393 nm), 2.91 eV (426 nm), and 2.64 eV (470 nm). The two green peaks are Ce1 emission, and the blue peak is Ce2 emission.
Figure 8. PL decay profiles of (a) Ce1 and (b) Ce2 centers in non-codoped, 0.05, 0.1, and 0.3 at% Li-codoped LSO:Ce samples. The excitation and emission wavelengths are 360 nm and 392 nm for Ce1, and 333 nm and 500 nm for Ce2, respectively. The error bar of the decay constant is 0.1 ns.
state of the activator in order to achieve electrical neutrality in the lattice, due to the charge mismatch of the Li ions with the host cations. It is proven that Li codoping does not convert the stable Ce3+ to Ce4+ in the LSO host. Thus, the introduction of Li codopants will result in a change in the type and/or concentration of defects as a consequence of charge compensation. Thermoluminescence is an effective tool for studying the defect structure of luminescent materials. The TL glow curves of noncodoped and Li-codoped LSO:Ce samples are shown in Fig. 9(a). The dominant TL peaks at 351 and 408 K in non-codoped LSO:Ce have been ascribed to Ce-spatially-correlated oxygen vacancies (VO).54 Specifically, these TL peaks result from the recombination of electrons trapped in VO with holes localized at Ce3+ centers through thermally-assisted tunnelling processes.54 The integral intensity of the VO-related TL peaks is reduced by 80% in 0.05 at% Li-codoped LSO:Ce compared to that of non-codoped LSO:Ce. This suggests that 0.05 at% Li codoping effectively suppresses the formation of VO and/or dissociates the spatially-correlated Ce and VO. The fact that the intensity of the VO-related TL peaks increases with further increases in Li concentration clearly indicates that high codoping levels favor the formation of Ce-spatially-correlated VO. Nonetheless, the intensity of the VO-related TL peaks in 0.1 at% Li codoped samples is still lower than that of non-codoped sample. In LSO:Ce, the VO defects are deep electron traps that can delay the recombination of electron-hole pairs at the Ce centers, leading to a strong afterglow at room temperature.55 The room temperature afterglow profiles of non-codoped and Li-codoped LSO:Ce samples are presented in Fig. 9(b). A positive correlation between the intensity of VO-related TL peaks and the afterglow level is observed. The afterglow level of 0.05 at% Li-codoped LSO:Ce is one order of magnitude lower than that of non-codoped LSO:Ce. Doubling the Li codoping to 0.1 at% Li results in an increase in afterglow to a level comparable with that of non-codoped LSO:Ce. Further increasing the Li-codoping concentration to 0.3 at% results in an extremely low light yield of a few hundreds of photons/MeV, and the afterglow level significantly increases to one order of magnitude higher than that of non-codoped LSO:Ce. This confirms the reciprocal relationship between light yield and afterglow. Therefore, the Liconcentration-dependent light yield is mainly determined by the variation in concentration of cerium-spatially-correlated VO.
Effect of Li codoping on the defect structure Codoping of monovalent Li ions into the LSO lattice will inevitably lead to changes in the defect structure and/or the oxidation
ACS Paragon Plus Environment
ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Figure 9. (a) TL glow curves and (b) X-ray induced afterglow profiles of non-codoped and Li-codoped LSO:Ce single crystals.
Critical roles of Li site occupancy To clarify the critical role of Li on defect structures, the 7Li nuclear magnetic resonance technique was used to investigate the site occupancy of Li and its local environment in all samples. All of the 7Li NMR spectra shown in Fig. 10 can be well fitted by two Lorentzian functions. All samples have two 7Li peaks with an isotropic chemical shift of -1.6 and 0.1 ppm. However, the percentage and linewidth of these two 7Li peaks in all samples are quite different. The percentage and linewidth of the 7Li NMR peaks at -1.6 and 0.1 ppm as a function of Li codoping concentration are plotted in Fig. 11. There are two Li occupation sites in the LSO host lattice. Theoretical calculation results suggest that under O-poor synthesis conditions the Li ions prefer to occupy interstitial spaces and/or Lu sites depending on the Fermi level.56,57 In our previous study of 0.05 at% Li-codoped LSO:Ce, the 7Li peaks at 0.1 ppm and -1.6 ppm were assigned to the positively charged Lii interstitials (six-coordinated) and the negatively charged LiLu substitutional sites (seven-coordinated), respectively.38 The Lii interstitials are dominant in the 0.05 at% Li-codoped LSO sample. However, with increasing Li concentration the percentage of LiLu substitutional sites becomes dominant compared to that of Lii interstitials. Specifically, 72.7% and 84.5% of Li ions occupy the seven-coordinated Lu sites in 0.1 at% and 0.3 at% Li-codoped samples, respectively. Because of that, the formation of VO becomes energetically favorable when the Fermi level moves downward in the gap in the presence of LiLu as acceptors. This explains the enhanced intensity of VO-related TL peaks in 0.1 and 0.3 at% Li codoped samples (see Fig. 9a) Static NMR lineshape analysis can provide useful information on the ion mobility.58,59 Since linewidth is inversely proportional to the spin-spin relaxation time, larger linewidth means smaller relaxation time, indicating less mobility. For samples at all Li-codoping concentrations the linewidth of the 7Li peak at -1.6 ppm corresponding to LiLu is essentially the same, suggesting almost unchanged mobility of Li ion at that site. In contrast, the linewidth of the 7Li peak at 0.1 ppm associated with Lii decreases from 996.5 Hz for 0.05
Page 6 of 21
at% Li, to 335.3 Hz for 0.1 at% Li, and to 146.2 Hz for 0.3 at% Li. The decreasing linewidth indicates the presence of more mobile Li-containing species in the samples codoped with higher Li concentrations. Since the Lii interstitials are coordinated with six oxygens,57 the mobile Li-containing species indicate the existence of VO adjacent to Lii to form {Lii+VO} complex defects. According to Ref. 56, under O-poor conditions the {Lii+VO} complex defect has the lowest formation. Therefore, the significant reduction of VO-related peaks in the 0.05 at% Li sample is believed to be mainly caused by a reduction in VO formation rather than the dissociation of spatially-correlated Ce and VO by the formation of {Lii+VO} complex defects. Similarly, in Y3Al5O12 codoped with a low Li concentration (
