Scintillator Design Via Codoping

3 downloads 0 Views 787KB Size Report
Appl. 2 (2014) 044009. [22]. M. Kobayashi, Y. Usuki, M. Ishii, N. Senguttuvan, K. Tanji, M. Chiba, et al.: Nucl. Instr. Meth. Phys. Res. A 434 (1999) 412. □□□.
Proc. Int. Symp. on Radiation Detectors and Their Uses (ISRD2016) JPS Conf. Proc. 11, 020001 (2016) https://doi.org/10.7566/JPSCP.11.020001

Scintillator Design Via Codoping C. L. MELCHER1,2, M. KOSCHAN1, M. ZHURAVLEVA1,2, Y. WU1, H. ROTHFUSS1,2,4, F. MENG1,2, M. TYAGI1, S. DONNALD1,3, K. YANG1,2, J. P. HAYWARD1,3, and L. ERIKSSON1,2,4 1

Scintillation Materials Research Center, Materials Science and Engineering Department, 3 Nuclear Engineering Department, University of Tennessee, Knoxville, TN, USA 4 Siemens Healthcare Molecular Imaging 2

E-mail: [email protected] (Received April 28, 2016) Scintillation materials that lack intrinsic luminescence centers must be doped with optically active ions in order to provide luminescent centers that radiatively de-excite as the final step of the scintillation process. Codoping, on the other hand, can be defined as the incorporation of additional specific impurity species usually for the purpose of modifying the scintillation properties, mechanical properties, or the crystal growth behavior. In recent years codoping has become an increasingly popular approach for engineering scintillators with optimal performance for targeted applications. This report reviews several successful examples and its effect on specific properties. KEYWORDS:

1.

scintillator, codoping

Introduction

Scintillators are materials that convert ionizing radiation into pulses of visible light, and therefore may be used to detect electromagnetic radiation such as X-rays and gamma rays, as well as energetic charged particles. The traditional view of the scintillation process starts with the incident radiation ionizing some of the atoms making up the material. The energetic electrons and holes resulting from this process eventually thermalize via electron-electron and electron-phonon interactions and migrate through the crystal lattice until they become trapped either at luminescence centers or defects. The luminescence centers are thus elevated to excited states that subsequently decay via the emission of characteristic photons, often in the visible range of the spectrum. Efficient detection of gamma rays and X-rays by scintillators requires compositions that are optimized in several ways. Good detection efficiency requires high density and high effective atomic number to maximize the cross sections for photoelectric absorption, Compton scattering, and pair production. Luminescence centers should exhibit high quantum efficiency for good light output and electric dipole allowed radiative transitions for short decay times and consequently high count-rate capability in high flux scenarios such as active interrogation. The host material band gap should be small in order to

■■■ 020001-1

©2016 The Author(s)

This article is available under the terms of the Creative Commons Attribution 4.0 License. Any further distribution of this work must maintain attribution to the author(s) and the title of the article, journal citation, and DOI.

020001-2

JPS Conf. Proc. 11, 020001 (2016)

minimize the energy required for the production of electron-hole pairs but must be larger than the activator’s excited state to ground state transition energy. Low defect concentrations are needed for fast and complete energy transfer from the matrix to the luminescence centers. Minimal absorption of the emitted light is necessary for good light collection by the photo-sensor, and an index of refraction near 1.5 is preferred in order to minimize internal reflections. Chemical inertness and mechanical strength facilitate cost effective detector processing and high detector reliability and lifetime. Useful neutron detectors, on the other hand require high sensitivity for neutrons and at the same time minimal sensitivity for gamma rays or, when counting neutrons, the ability to reject gamma rays since false neutron counts are often unacceptable. For fast neutron sensing, this may be achieved by low density and low atomic number scintillator compositions that include elements with high neutron scattering cross sections. In slow neutron sensing, scintillator compositions include isotopes that have high absorption cross sections. In addition, the interaction of neutrons with the material must result in ionization of the matrix and ultimately the excitation of luminescence centers. The desirable properties of the luminescence centers include a short radiative lifetime and emission energy between 3 and 6 eV. 2. The first practical single crystal scintillator Sodium iodide activated with thallium was first used to detect gamma rays by Hofstadter in 1948 [1]. A short time later, the ability to measure the energy of gamma rays with NaI:Tl was demonstrated [2], and numerous applications in gamma spectroscopy followed [3]. For at least two decades NaI:Tl was the most commonly used inorganic scintillator crystal in a wide variety of nuclear physics endeavors ranging from basic science investigations to medical imaging to geophysical exploration. Due to its relatively high density (3.67 g/cm3) and atomic number compared to organic materials, NaI:Tl efficiently detects gamma rays and X-rays. Its scintillation decay time of 230 ns accommodates high counting rates, and its emission wavelength peaking at 410 nm is a good match to the spectral response of bialkali photomultiplier tubes. It does suffer from sensitivity to moisture in the atmosphere, but effective hermetic packaging technology has largely solved this problem. 3. The specialization of scintillators Despite the remarkable success of NaI:Tl, the increasing demands of various radiation detection applications eventually stimulated the discovery and development of new scintillators with properties that matched the requirements of the targeted applications. The search for new scintillators in the 1970’s and 1980’s resulted in the discovery of Bi4Ge3O12 by Weber and Monchamp [4] which was rapidly developed for use in high energy particle detection at the L3 detector at CERN as well as for positron emission tomography. In the 1990’s two new scintillators emerged, Lu2SiO5:Ce as a faster and higher light output scintillator for positron emission tomography [5-7] and PbWO4 as a faster and more efficient scintillator for the CMS detector at CERN . Oil well logging spurred the development of Gd2SiO5:Ce [8, 9] and more recently has implemented LaBr3:Ce [10] primarily due to detection efficiency, temperature response,

■■■

020001-3

JPS Conf. Proc. 11, 020001 (2016)

and energy resolution. The demanding energy resolution requirements in nuclear security applications have pushed the development of both SrI2:Eu [11] and LaBr3:Ce. Table I summarizes some of the major scintillators that have replaced or are beginning to replace NaI:Tl in specific applications. Table I.

Important scintillator properties for various applications.

Application

Important properties

Particle physics

High density, radiation hardness, availability in large size High density and atomic number, fast decay time, high light output Good energy resolution High density and atomic number, good temperature response, fast decay time

Positron Emission Tomography Nuclear security Oil well logging

Examples of successful scintillators Bi4Ge3O12, PbWO4 Bi4Ge3O12, Lu2SiO5:Ce LaBr3:Ce, SrI2:Eu Gd2SiO5:Ce, LaBr3:Ce

4. The implementation of codopants In the 1990’s scintillation materials researchers began to follow a strategy previously used for optimizing the growth and properties of laser crystals, i.e. the use of codopants. Table II lists examples of codopants that have been used in scintillators to optimize a number of different properties. More details are given in the sections following the table. Note that this is not meant to be an exhaustive list but rather a collection of examples to illustrate some advances that have been made in recent years. Table II.

Examples of codopants and their effects.

Scintillator Lu2SiO5:Ce Lu2SiO5:Ce Gd2SiO5:Ce LaBr3:Ce NaI:Tl Y2SiO5:Ce Gd3Al2Ga3O12:Ce PbWO4

Codopant Ca Zn Zr Sr, Ca Eu Ca Ca Ln3+

Effect Shorten decay time Stabilize growth behavior Increase light output Improve energy resolution Shift emission wavelength Enhance charge transport Manipulate charge state Improve radiation hardness

4.1 Increase light output The scintillation emission in Ce activated crystals such as Gd2SiO5:Ce arises primarily from 5d to 4f transitions in the trivalent species. High light output has therefore been associated with maximizing the amount of Ce3+ and minimizing the amount of Ce4+ which does not luminesce. Researchers at Hitachi Chemical Co. reported substantial improvement of light output in Gd2SiO5:Ce by codoping with

■■■

020001-4

JPS Conf. Proc. 11, 020001 (2016)

tetravalent ions that presumably occupy Gd3+ sites in the crystal lattice [12]. The apparent reduction in Ce4+ concentration is attributed to charge compensation by the tetravalent codopants. Positive results were obtained with Zr4+, Hf4+, and Ge4+ codopants, and the largest increase of ~20% was observed with Zr4+ codoping. Experiments were conducted to determine that the optimum Zr4+ concentration was ~200 ppm. In addition a noticeable improvement in the optical transmission was observed and also attributed to the reduction in Ce4+. 4.2 Shorten decay time The scintillation process is a complex one that comprises several phenomena that occur on a wide range of time scales. In the case of photoelectric absorption of the incoming gamma ray, typically an electron is ejected from the K shell of an atom with an energy equal to the incoming gamma ray energy minus the K shell energy. The atom with the K shell hole relaxes by either generating an Auger electron or through a series of radiative steps through the outer shells. Both the primary electron and the Auger electron relax via electron-electron interactions on a time scale of 1-100 fs which is similar to the time scale of radiative relaxation of a K shell hole. The result of this step is a large number of conduction band electrons, valence band holes, and core and valence band excitons. The next step, thermalization, involves electron-phonon interactions during which electrons move to the bottom of the conduction band and holes move to the top of the valence band, thus producing e-h pairs with energy equal to the band gap. This process occurs on a time scale of roughly 1-10 ps. The time for e-h migration to luminescence centers varies widely according to the distance between centers as well as the number and depth of traps created by crystalline defects and impurities. Times as long as 1-10 ns are not uncommon. The time scale of the final step, luminescence emission, ranges from < 1 ns to > 1 ms. Many useful scintillators utilize the 5d-4f transition with a characteristic time constant of 20-70 ns (e.g. Ce3+), or the 4f65d-4f transition with a characteristic time constant of approximately 1 µs (e.g. Eu2+). The time profile of the scintillation emission is important because it ultimately determines the count-rate capability and the coincidence timing capability of the scintillator. Spurrier et al. [13] discovered that codoping Lu2SiO5:Ce with Ca2+ reduced the decay time by as much as 25%. Figure 1 shows the relationship between Ca2+ concentration and scintillation decay.

Fig. 1. Decay time as a function of Ca2+ concentration in Lu2SiO5:Ce. Reused from [13] with permission.

4.3 Stabilize growth behavior In Czochralski crystal growth, control and stability of crystal diameter and therefore of the growth process as a whole depends on adequate surface tension of the melt to form a stable meniscus. Spurrier et al. [14] discovered that while Ca2+ codoping of

■■■

020001-5

JPS Conf. Proc. 11, 020001 (2016)

Lu2SiO5:Ce had desirable effects on scintillation properties, high levels of Ca2+ concentration resulted in low surface tension that prevented stable crystal growth. It was found that by adding a second codopant, Zn4+, the surface tension of the melt was restored and stable growth was enabled. 4.4 Improve energy resolution The energy resolution of a scintillator + PMT is determined by photon statistics, crystal inhomogeneity, and nonproportionality and can be expressed as follows [15]: 1 ∆𝐸 2 2 2 2 = [𝑅𝑖 + 𝑅𝑝 + 𝑅𝑀 ] 𝑅=( ) 𝐸 𝐹𝑊𝐻𝑀 2 2 + 𝑅𝐼𝑁𝐻 𝑅𝑖2 = 𝑅𝑁𝑃 is the intrinsic resolution. 𝑅𝐼𝑁𝐻 relates to inhomogeneity.

𝑅𝑁𝑃 relates to nonproportionality and

𝑅𝑝 is the “transfer efficiency” or the probability that a photon emitted by the crystal results in a photoelectron arriving at the first dynode and that is fully multiplied by the PMT. 1

𝑅𝑀 = 2.35 × [1 +

𝜐(𝑀) 2 𝑁𝑝𝑒

]

where 𝜐(𝑀) is the variance in electron multiplication

and 𝑁𝑝𝑒 is the number of photoelectrons. Typically 𝜐(𝑀) ∼ 0.1 − 0.2 In scintillators with very high light output, the nonproportionality of the response vs. energy can be a limiting factor for energy resolution, and codoping has been explored as a possible technique to improve this characteristic. Alekhin et al. [16] and Yang et al. [17] determined that the nonproportionality of LaBr3:Ce scintillators could be controlled by means of codoping with various divalent and monovalent ions. In particular Sr2+ and Ca2+ had a significant beneficial effect and enabled the energy resolution to approach 2.0% at 662 keV. 4.5 Shift emission spectrum The spectral match of a scintillator’s emission and the response of the photodetector is fundamental to the performance of a radiation detector. In the case NaI:Tl, the emission peak of 410 nm is well matched to the response of bialkali photomultipliers but is not so well matched to the response of silicon-based devices such as some avalanche photodiodes and silicon photomultipliers. Shiran et al. [18] were able to successfully use codoping with Eu2+ to shift the emission from 410 nm to 450 nm, which provides a significantly better match to the longer wavelength sensitivity of the silicon-based photosensors.

■■■

020001-6

JPS Conf. Proc. 11, 020001 (2016)

4.6 Enhance charge transport It is apparent that the observed emission time of a scintillator is dominated by the migration time of excitons and e-h pairs and the excited state lifetime of the luminescence center. In a perfect defect-free scintillation crystal with a single type of luminescence center with a single radiative transition, the excited state of the luminescence centers would be populated in