Photoluminescence, radioluminescence, and thermoluminescence in NaMgF3 activated with Ni2+ and Er3+

Author’s Accepted Manuscript Photoluminescence, radioluminescence, and thermoluminescence in NaMgF3 activated with Ni2+ and Er3+ Jethro Donaldson, Grant V.M. Williams www.elsevier.com/locate/jlumin

PII: DOI: Reference:

S0022-2313(15)30303-3 http://dx.doi.org/10.1016/j.jlumin.2016.01.004 LUMIN13801

To appear in: Journal of Luminescence Received date: 27 July 2015 Accepted date: 5 January 2016 Cite this article as: Jethro Donaldson and Grant V.M. Williams, Photoluminescence, radioluminescence, and thermoluminescence in NaMgF activated with Ni2+ and Er3+, Journal of Luminescence, http://dx.doi.org/10.1016/j.jlumin.2016.01.004 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Photoluminescence, radioluminescence, and thermoluminescence in NaMgF3 activated with Ni2+ and Er3+ Jethro Donaldson* and Grant V. M. Williams MacDiarmid Institute, School of Physical and Chemical Sciences, PO Box 600, Victoria University of Wellington, Wellington 6012, New Zealand ABSTRACT Photoluminescence (PL) and radioluminescence (RL) has been observed out to 1670 nm in NaMgF3 doped with Er3+ or Ni2+. The RL contains more emissions than seen in the PL, which is likely to be due to radiation induced excitations to higher energy excited states of the luminescent ion than are accessible during PL. Both luminescent ions have relatively long PL lifetimes and the long wavelength RL emission intensities are independent of absorbed dose history for high doses. This, coupled with the approximate radiological equivalence of NaMgF3 to water are desirable properties for a dosimeter for radiotherapy applications. The Čerenkov component seen for high energy radiation sources can be reduced by monitoring the longer wavelength RL emissions, or by temporal discrimination for pulsed sources. Room temperature glow peaks are seen in the NaMgF3:Er3+ thermally stimulated luminescence emission that imply the RL will be temperature dependent. This is not the case for NaMgF3:Ni2+ where no thermoluminescence is observed, hence Ni2+ is preferred over Er3+ for radiotherapy applications. KEYWORDS Photoluminescence; radioluminescence; thermally stimulated luminescence; fluoroperovskite. * Corresponding author ([email protected])

1

INTRODUCTION Radiation induced luminescence in inorganic solids has been the subject of investigation for many decades as there exist numerous applications in radiation dosimetry devices [1,2]. Of particular interest are the optically and thermally stimulated luminescence (OSL and TSL), and radioluminescence (RL) of such materials. In addition to facilitating some of these phenomena [1,3], the fluoroperovskite NaMgF3 has an effective atomic number and mass attenuation coefficient similar to that of liquid water for photon energies above 0.1 MeV. This renders it potentially useful for radiation dosimetry applications that demand radiological equivalence to soft tissues, such as dosimetry for radiation protection or radiation therapy. One particular application of NaMgF3 is as a RL point dosimeter where NaMgF3 is attached to the distal end of an optical fibre for dose verification of radiotherapy treatments delivered using medical linear accelerators. The use of luminescent solids coupled to optical fibres for remote readout is gaining acceptance in this area of application, with plastic scintillators particularly favoured for their unmatched soft tissue equivalence [4]. Due to their solid state nature, fibre coupled luminescence detectors are miniaturisable to a degree that existing dosimeter technologies are not, realizing increased spatial resolution and a reduction in radiation perturbation effects [4]. A stem signal due to capture of auto-fluorescence and a broad spectrum of Čerenkov photons in the irradiated fibre optics of such systems remains a complicating factor in the implementation of such devices. This occurs when using fibre coupled detectors with high energy radiotherapy radiation sources. The Čerenkov component of this background is typically dominant. Primary solutions to this challenge employed with plastic scintillators include chromatic discrimination

2

[5], the use of air filled waveguides [6], and temporal gating in pulsed radiations [7]. In using inorganic materials a wider range of emission energies and lifetimes can be accommodated, permitting dosimeters better suited to further possible solutions: more efficient temporal gating in pulsed radiations [8] and operation in the near infrared (NIR) where Čerenkov yield is significantly reduced [9]. Temporal discrimination exploits the prompt nature of the stem signal, which decays on the order of nanoseconds or less, while an infrared emission leverages the theoretical λ−2 dependence of spectral intensity expected of Cerenkov emissions, where λ is the wavelength [10]. We have previously investigated the inorganic crystal NaMgF3 activated with Eu3+ and Mn2+ in both bulk crystals [1] and nanoparticles [3], for applications in radiotherapy dosimetry. These dopants emit in the near ultraviolet and visible wavelength regions. In this report we present the results from photoluminescence (PL), RL, and TSL measurements on NaMgF3 doped with two infrared emitting species, Er3+ or Ni2+. This data supplements the limited coverage of these two dopants in NaMgF3 appearing in existing literature [11,12]. We show that NIR emissions are observed in the PL and RL for both luminescent ions and PL lifetimes are long enough for temporal discrimination. Ni2+ is the preferred dopant for radiotherapy application to minimize the Čerenkov background owing largely to the absence of detectable TSL. MATERIALS AND METHODS Polycrystalline samples were prepared by a slow cooling method. Fluorides were ground to powder and mixed in stoichiometric quantities in a vitreous carbon crucible. The crucible was then heated in an RF furnace to 1100 °C, which is above the melting point of NaMgF3, and

3

cooled slowly over a period of one day to 1010 °C. This process resulted in the formation of transparent polycrystalline material. To minimise the oxygen and water content, samples were prepared in a dry argon atmosphere. The molar fraction of both erbium and nickel dopants was 0.5%. A nickel doped sample at 5% was also fabricated and used for optical absorption measurements. Powder x-ray diffraction measurements on all samples showed that they were phase pure to within the limit of detection. Based on ionic radii [13] it is expected that Er3+ substitutes for Na+ on the perovskite A site, and Ni2+ for Mg2+ on the B site. In NaMgF3, the A site has a somewhat distorted cubic symmetry, while the B site remains octahedral [14]. PL spectra were measured using a Horiba Fluorolog spectrofluorometer with a xenon discharge source. This instrument was corrected for the spectral intensity of the light source and detector sensitivity, with upper and lower energy limits of the excitation source at approximately 40000 cm−1 and 12000 cm−1. Visible wavelength emissions were detected using a multi-alkali photomultiplier (Hamamatsu R928P) and infrared emissions using an InGaAs photodiode cooled with liquid nitrogen. Sensitivity at infrared wavelengths was cut off by the band gap of InGaAs at approximately 1550 nm. PL lifetimes were measured using the photomultiplier in conjunction with pulsed LED sources. It was not possible to measure lifetime data for emissions of wavelength longer than 850 nm. Absorption measurements were made using a PerkinElmer Lambda 1050 spectrophotometer equipped with deuterium and tungsten-halogen light sources. Visible wavelength emissions were detected with a multi-alkali photo-multiplier and infrared emissions with a Peltier cooled InGaAs photodiode, together covering from 200 nm to 1800 nm. The NaMgF3 sample containing 5%

4

Ni2+ was prepared for this measurement by cutting sections to 1 mm thickness using a diamond saw and polishing the exposed surfaces on alumina lapping discs down to 3 μm. RL was recorded using an Ocean Optics USB4000 fibre optic spectrometer. This instrument was fitted with a silicon CCD linear array and fixed grating of 300 nm blaze. The effective resolution was 2 nm. Irradiations were made using a tungsten x-ray tube operated at 40 kV and 40 mA, with 0.7 mm of aluminium filtration. Samples were located at a nominal distance of 6.5 cm for an approximate surface dose rate of 2 Gy s−1. X-ray luminescence was monitored during 3 hour exposures to assess the response to high cumulative radiation doses. TSL was measured using a custom apparatus detailed elsewhere [15]. Crystals were ground to powder and mixed with silicone oil to establish close thermal contact with the heating element, which was then cooled below room temperature in a dry nitrogen flow prior to heating. Luminescence was detected using an Ocean Optics USB2000+ fibre optic spectrometer. This instrument is similar that used for RL measurements, but with fixed grating of 400 nm blaze and a limiting resolution of 7.5 nm due to a much larger entrance slit. To accommodate the relatively poor sensitivity of this CCD spectrometer and the necessarily small sample quantities, radiation doses on the order of 100 Gy were made. Glow peaks were isolated by curve fitting in MATLAB, using an analytical expression for glow peaks with second order kinetics [16]. With the exception of TSL, all spectroscopic data were measured at room temperature. RESULTS AND DISCUSSION The PL from NaMgF3:Er3+ is plotted in figure 1(a) after fitting and subtracting a broad Eu2+ 4f65d(Eg) → 4f7 background due to a small amount of the Eu2+ contamination. The PL excitation peaks are well correlated with the known energy levels of Er3+ [17]. PL emissions at visible and

5

near infrared wavelengths for excitation at 27780 cm-1 (360 nm) can be seen in figure 1(a) with structured groups at 19050 cm−1, 18350 cm−1, 15270 cm−1, 11830 cm−1, 10200 cm−1, 8200 cm−1, and near 6500 cm−1. These groupings consist of narrow line emissions consistent with transitions within the 4f11 configuration of Er3+.The emissions in the near infrared were assigned to transitions from the 4S3/2 level to 4I11/2 and 4I13/2 (8200 cm−1 and 11830 cm−1), then from each to the ground state, 4I15/2 (10200 cm−1 and 6500 cm−1). With reference to the literature [11,18] the visible emissions are likely due to transitions from the 2H11/2 (19050 cm−1), 4S3/2 (18350 cm−1), and 4F9/2 (15270 cm−1) levels to the ground state. The energy level diagram and the excitation and emission transitions can be seen in figure 1(b). The PL decay from the NaMgF3:Er3+ 18350 cm−1 emission after 26670 cm−1 (375 nm) pulsed excitation can be seen in figure 2, decay of the 19050 cm−1 emission being identical. The PL decay could not be fitted to a single exponential. For this reason it was fitted to a stretched exponential that can be written as [19], () =  exp(−⁄)

(1)

where t is the time, τ is the PL lifetime, and β is the stretched exponential exponent. A stretched exponential PL decay indicates that there is a distribution of radiative and possibly also nonradiative emission lifetimes that can arise from structural disorder in the host crystal [20].In the case of NaMgF3:Er3+ such disorder may be due to a range of charge compensation defects. Both adjacent A site vacancies (Vc‒centre) and H‒centre defects are likely compensation mechanisms and have been shown to occur in a similar material, KMgF3:Eu3+ [21]. Vc‒Er3+ pairs or clusters thereof would perturb the local crystal field and so be expected to result in more than one distinct emission lifetime.

6

The fitted lifetime was  = 0.76 ± 0.04 ms and stretched exponential factor was ! = 0.85 ± 0.01. A double exponential function would also provide a satisfactory fit, and could perhaps be related to isolated Vc‒Er3+ pairs and clustering of Vc‒Er3+ pairs. It should be noted that while the A site is slightly distorted by a 17° rotation of the MgF6 octahedra [14] to accommodate size of the Na+ ion, there is only one structurally distinct A site in the absence of defects. It is also possible to estimate an average lifetime, 〈〉 [22], 〈〉 =

'

∫* %&(%)+% '

∫* &(%)+%

(2)

where I(t) is the PL intensity. This expression reduces to 〈τ〉 = τ for single exponential decay of the luminescence intensity. For the decay of both visible wavelength emissions in NaMgF3:Er3+, 〈τ〉 was 0.89 ± 0.05 ms, which is close to the stretched value. This is sufficiently long lived for temporal discrimination of prompt background emissions, such as Čerenkov photons when pulsed radiations sources are used. Optical absorption in the NaMgF3:Ni2+ sample containing 5% Ni2+ sample can be seen in figure 3(a) where the absorption coefficient is plotted. Three distinct peaks are observed in the range of measured energies that correspond to the Ni2+ spin allowed transitions from the ground state to the 3T2g(3F) (7650 cm−1), 3T1g(3F) (13100 cm−1), and 3T1g(3P) (24500 cm−1) energy levels. These findings are in good agreement with previously reported data on NaMgF3:Ni2+ at a similar dopant concentration [12]. Using the energy of these spin allowed absorptions, an estimation of the crystal field splitting energy 10Dq at 7850 cm−1 and the Racah parameter B of 925 cm−1 was made with reference to the Tanabe-Sugano diagram for the d8 configuration in octahedral complexes. Weaker peaks can be seen at 15550 cm−1, 21600 cm−1, and 32000 cm−1 that are consistent with the expected energies of the 1Eg(1D), 1T2g(1D), and 1T1g(1G) levels, as calculated

7

with the above values of 10Dq and B. An additional weaker and broad absorption appears at approximately 18000 cm−1 which is not consistent with the known energies of the Ni2+ ion. The PL excitation spectra from the NaMgF3:Ni2+ sample is also shown in figure 3(a), and it consists of two overlapping bands at 24510 cm−1 and 21510 cm−1. These bands appear closely correlated with the absorption data and so are assigned similarly. Identical PL excitation spectra were measured for the emission bands at 20200 cm−1, 13790 cm−1, and 6500 cm−1, as is expected because all these emissions derive ultimately from the same excited state. Principal PL emissions from NaMgF3:Ni2+ were observed at 20160 cm−1 and near 6000 cm−1, with additional emissions of much lower intensity at 13790 cm−1 and 9010 cm−1, when excited at 24630 cm−1 (406 nm) as can be seen in figure 3(a). The 20160 cm−1, 13800 cm−1, and 6000 cm−1 bands can be assigned to 1T2g(1D) → 3A2g(3F), 1T2g(1D) → 3T2g(3F), and 3T2g(3F) → 3A2g(3F) transitions within the 3d configuration of the NiF6 complex. These assignments reflect those made in other reports on Ni2+ in hosts crystals with octahedral dopant sites [12,23-25]. While it was not possible to measure the entirety of the emission from the 3T2g(3F) level, the absorption spectroscopy reported herein is otherwise consistent with that reported for NaMgF3:Ni2+ in the past [12], for which this emission was centred at 5800 cm−1. The emission at 9020 cm−1 was assigned to the transition 1T2g(1D) → 3T1g(3F) [24]. The subsequent spin allowed relaxation from 3

T1g(3F) to the ground state was notably absent. Possibly this state relaxes first to the 3T2g(3F)

level (~4800 cm−1), which would not be detectable with the cooled InGaAs detector employed in this study. The energy level diagram and the excitation and emission transitions can be seen in Fig. 3(b).

8

PL decay from the emissions at 20160 cm−1 and 13800 cm−1 after 26670 cm−1 (375 nm) pulsed excitation were made on NaMgF3:Ni2+ where the PL decay from the 20200 cm−1 emission is plotted in figure 2. It was not possible to fit the PL decays to a single exponential. For this reason the PL decays were fitted to the stretched exponential function given in equation 1. Both emissions had the same fitted lifetime of  = 0.124 ± 0.002 ms with an exponential stretch factor of ! = 0.885 ± 0.001 . While a double exponential function could also fit the PL decay, it is not clear why there would be two distinct sites as Ni2+ is expected to isoelectronically substitute for Mg2+. The average lifetime calculated using equation 2 was 〈〉 = 0.129 ± 0.002 ms for both the 20200 cm−1 and 13790 cm−1 emissions. The PL lifetimes are identical which is consistent with the emissions arising from the same excited state of 1T2g(1D). They are also on the same order of magnitude as reported for Ni2+ in similar perovskite hosts [23], and are also sufficiently long lived for temporal gating of a prompt Čerenkov background signal under pulsed irradiation. The low energy NIR emission from the 3T2g(3F) energy level of NaMgF3:Ni2+ (see figure 3) occurs in a region of the optical spectrum which is not only potentially advantageous in the mitigation of Čerenkov stem effects in radiotherapy applications, but is also relatively well aligned with the erbium window of optical transmission in silica fibre optics at 1550 nm. Attenuation of silica fibre at these wavelengths can theoretically be as low as 0.3 dB/km [26], albeit higher for polymer clad silica optics of large diameter and numerical aperture. This can be exploited to reduce signal attenuation and hence reduce measurement uncertainty in radiotherapy applications, where several tens of metres of fibre optics are necessary to traverse shielding mazes in typical treatment vaults.

9

Quenching of the 1T2g(1D) → 3A2g(3F) emission at molar concentrations above those of the crystals employed in this study has been reported for Ni2+ in other fluoroperovskite hosts [23], where it is attributed to cross relaxation between dopant ions. Quenching via this mechanism is not possible for emission from the 3T2g(3F) state due to the lack of any intermediate energy levels. This would suggest that the luminescence efficiency at NIR wavelengths may scale more favourably with dopant concentration. Radioluminescence measurements were made on NaMgF3:Er3+ during x-ray irradiation and the RL spectra can be seen in figure 4(a). Four peaks at 11830 cm‒1, 15270 cm‒1, 18350 cm‒1, and 19050 cm‒1 are also seen in the PL emissions (see figure 1) and are similarly assigned. However, there are a number of RL emissions that are not seen in the PL spectra. These peaks originate from levels above the PL excitation energy of 27780 cm−1. Specifically, from the 2P3/2 level relaxing to 2H11/2 (12400 cm−1), 4S3/2 (13110 cm−1), 4F9/2 (16220 cm−1), 4I11/2 (21200 cm−1), 4I13/2 (24900 cm−1) and the ground state (31400 cm−1). A broad emission of low intensity near 20000 cm‒1 could not be explained in terms of transitions within the 4f configuration. A similar band emission in the RL of NaMgF3:Er3+ has been previously attributed to exciton luminescence, but was not reported at room temperature [11]. In the Ni2+ doped sample a number of additional peaks were observed when excited with x-rays, which could not be explained in terms spin allowed transitions within the Ni2+ levels. Furthermore, these transition energies could not be reconciled with the d7 configuration of Ni3+, or with any plausible contaminant species such as Er3+ or Eu2+. They are indicated in figure 5(a). Possibly these emissions derive from the 1A1g(1S) level of Ni2+, which at 61900 cm−1 is well beyond the range of optical excitation sources available. The 20200 cm−1 and 13790 cm−1 emissions of Ni2+ observed in the PL emission spectra are also seen in the RL emission.

10

Optical absorption measurements in an un-doped NaMgF3 sample revealed intense radiation induced absorption at 300 nm, with less pronounced bands at 410 nm, 500 nm, and 700 nm. This is apparent in figure 5 where the transmittance is plotted for a 1 mm thick sample of un-doped NaMgF3 irradiated to several kGy. These findings closely reflect those reported by other investigators for NaMgF3 [27], and are similarly attributed to F‒centres and F‒centre amalgamations produced by radiolysis. Significant spectral changes in RL emission intensity were observed during x-ray irradiation for both samples. This can be seen in figures 4(b) and 5(b) where the RL peak intensities from NaMgF3:Er3+ and NaMgF3:Ni2+ are plotted for selected RL emissions. The time, and therefore dose, dependent changes are the product of two distinct processes: competition to radiative recombination from trapping defects, and optical absorption from trap filling and the aforementioned radiation induced defects. As radiation dose is accumulated the population of charge localized at trapping defects increases and eventually saturates, reducing the competition for delocalized charge carriers that may otherwise recombine and produce RL. Carrier trapping and an initial rise is also seen in the x-ray luminescence of other inorganic materials, such as α−Al2O3:C [8,28] and SiO2:Yb3+ [9]. This process generates the asymptotic growth in RL intensity observed in longer wavelength emissions where absorption from trap filling and radiation induced colour centres is minimal. The initial increase in the RL can be largely eliminated by deliberately saturating the carrier traps with suitably large priming doses. For any RL based radiation dosimetry system the filling of traps and the accumulation of optically active defect centres is detrimental, as the apparent change in RL sensitivity that occurs prevents such a dosimeter from retaining a calibration in the long term. As can be seen from the high dose response of the 845 nm emission from NaMgF3:Er3+ in figure 4(b), and the 730 nm

11

emission from NaMgF3:Ni2+ in figure 5(b), this radiation damage effect is significantly reduced if not eliminated in the near infrared. The concentration of colour centre defects absorbing in this wavelength region appears minimal or reaches an equilibrium after a relatively low radiation dose. Thus, infrared emitters are favourable not only for their potential to improve the ratio of detected RL to Čerenkov photons, but also for relative independence from the influence of F‒centre type absorptions. The NaMgF3:Er3+ crystal exhibited strong room temperature TSL, with distinct second order glow peaks at 272 K, 308 K, 327 K, and considerably weaker peaks at 367 K, 420 K, 540 K, 590 K, and 670 K. Both fits and experimental data are plotted in figure 6. These were fitted using the analytical approximation for second order thermoluminescence [16],

: <>
(/) = 4m exp 3;


? @< A 31 − m

B;< :a

: <>
? exp 3;

?+

B;
+ 1D

>B

(3)

where Im is the peak luminescence intensity, Tm the temperature in Kelvin at which Im was observed, E the Boltzmann constant in eV K−1, and Fa the trap activation energy in eV. / is temperature in Kelvin for which the observed luminescence intensity, /, is calculated. First order fitting using a similar expression [16] did not produce satisfactory results, particularly for the higher temperature peaks which seen in isolation appear symmetric in T . It can be seen in figure 6 that the TSL can be fitted with 8 distinct glow peaks. In order of increasing /m , the fitted values of Fa for each glow peak are 0.5 eV, 0.9 eV, 1.5 eV, 1.3 eV, 1.9 eV, 1.9 eV, 3.4 eV, and 2.4 eV.

12

With the exception of weaker emission wavelengths being reduced below measurement noise, no significant spectral changes exist between glow peaks. All peaks observed in the RL emission spectrum are seen in the most intense TL glow peak, shown in the inset of figure 6. Additional weak TL was observed at 700 nm, with faint glow peaks at 270 K and 420 K. This is known to arise from the platinum heating element of the measurement apparatus [15]. The intense room temperature TSL observed with the NaMgF3:Er3+ sample is likely to result in a strong temperature dependence of the RL intensity. This is potentially a barrier to applications in radiotherapy dosimetry, where measurements are performed at room or body temperature. Significant reduction in low temperature glow peak intensity following sintering and rapid quenching has been demonstrated for a similar material, NaMgF3:Eu2+. The crystals to which this technique was applied were fabricated using the same slow cooling method [1], and so similar technique might be leveraged to reduce this difficulty with NaMgF3:Er3+. It was observed that during repeated cycles of x-ray exposure, TSL readout, and thermal annealing, that the intensity of TSL per unit time of exposure to x-rays increased for the NaMgF3:Er3+ sample. This is indicative of the presence of deep or nearly disconnected traps [29], that are also seen in other fluoroperovskites [30], with trap energies not accessible at the maximum attainable annealing temperature of 720 K. No TSL was observed when the x-ray irradiated NaMgF3:Ni2+ sample was heated to the highest temperature attainable temperature of 720 K. This is consistent with the absence of room temperature TSL afterglow immediately after irradiation for RL spectroscopy. As the initial time dependence of RL intensity seen with this sample, shown in figure 5(b), is indicative of the competition between immediate charge carrier recombination via the luminescent ion and charge

13

carrier trapping, the lack of TSL is attributed a dominance of deep traps rather than a complete absence of trapping defects. The absence of low temperature TSL in NaMgF3:Ni2+ is a desirable property. If trapped charge carriers are not readily delocalised by heating to 720 K it follows that the TSL properties should not significantly influence the temperature dependence of RL intensity near room temperature. CONCLUSIONS In conclusion, we have observed photoluminescence and radioluminescence from NaMgF3 doped with Er3+ or Ni2+ where the emissions extend beyond 1600 nm. All of the PL and RL emissions from NaMgF3:Er3+ can be attributed to Er3+ 4f‒4f transitions. The average PL lifetime is 0.89 ms, which is significantly longer than the medical linear accelerator x-ray pulse duration that is typically in the μs range. Thus, NaMgF3:Er3+ would appear to a suitable candidate for radiotherapy applications where the influence of the Čerenkov background signal can be significantly reduced by temporal discrimination or by using long wavelength emissions. However, there are low temperature glow peaks with trap depths ranging from 0.5 eV to 2.4 eV that mean that the RL signal will be strongly temperature dependent, which is not desirable. The PL emissions from NaMgF3:Ni2+ can be attributed to Ni2+ 3d‒3d transitions. There are additional peaks in the RL that may be due to radiation induced Ni2+ optical excitation to higher energy levels that are not accessible during PL. The average PL lifetime is 0.13 ms, which is shorter than that found for Er3+, but it is still longer than the typical medical linear accelerator x-ray pulse duration. The NaMgF3:Ni2+ sample also did not display any TSL, which may be due to the presence of deep traps. The peak RL intensity at long wavelengths was independent of the dose history after an initial rise due to trap filling. These properties, combined with the approximate

14

radiological equivalence to water of NaMgF3, mean that NaMgF3:Ni2+ has potential as a dosimeter for radiotherapy applications. An appropriate priming dose would be required to remove the initial dose history dependence. It should be possible to significantly reduce the Čerenkov background signal by detecting the long wavelength emissions and using temporal discrimination. ACKNOWLEDGEMENTS We acknowledge funding from the MacDiarmid Institute for Advanced Materials and Nanotechnology, and the Ministry of Business, Innovation, and Employment (RTVU1405). REFERENCES [1]

C. Dotzler, G. V. M. Williams, U. Rieser, A. Edgar. Applied Physics Letters 91 (2007)

121910:1−3. [2]

G. V. M. Williams, S. G. Raymond. Radiation Measurements 46 (2011) 1099−1102.

[3]

G. V. M. Williams, S. Janssens, C. Gaedtke, S. G. Raymond, D. Clarke. Journal of

Luminescence 143 (2013) 219−225. [4]

A. R. Beierholm, C. F. Behrens, L. Hoffmann, C. E. Andersen. Radiation Measurements

56 (2013) 290−293. [5]

A-M. Frenlin, J-M. Fontbonne, G. Ban, J. Colin, M. Labalme, A. Batalla, A. Isambert, A.

Vela, T. Leroux. Medical Physics 32 (2005) 3000−3006. [6]

J. Lambert, Y. Yin, D. R. McKenzie, S. Law, N. Suchowerska. Physics in Medicine and

Biology 53 (2008) 3071−3080.

15

[7]

M. A. Clift, P. N. Johnston, D. V. Webb. Physics in Medicine and Biology 47 (2002)

1421−1433. [8]

C. E. Andersen, S. M. S, Damkjær, G. Kertzscher, S. Greilich, M. C. Aznar. Radiation

Measurements 46 (2011) 1090−1099. [9]

I. Veronese, C. Mattia, M. Fasoli, N. Chiodini, E. Mones, M. C. Cantone, A. Vedda.

Applied Physics Letters 105 (2014) 061103:1−4. [10]

J. V. Jelley. British Journal of Applied Physics 6 (1955) 277−232.

[11]

N. Kristianpoller, B. Trieman. Journal of Luminescence 24/25 (1981) 285−288.

[12]

M. L. Reynolds, G. F. J. Garlick. Infrared Physics 7 (1967) 151−167.

[13]

R. D. Shannon. Acta Crystallographica A 32 (1976) 761−767.

[14]

R. H. Mitchell, M. Alexander, L. M. D. Cranswick, I. P. Swainson. Physics and

Chemistry of Minerals 23 (2007) 705−715. [15]

J. W. Quilty, J. Robinson, G. A. Appleby, A. Edgar. Review of Scientific Instruments 78

(2007) 083905:1−6. [16]

G. Kitis, J. M. Gomez-Ros, J. W. N. Tuyn. Journal of Physics D: Applied Physics 31

(1998) 2636−2641. [17]

B. Henderson, G. F. Imbusch. Optical Spectroscopy of Inorganic Solids. Oxford

University Press, Oxford, 2006. [18]

D. K. Sardar, M. D. Shinn, W. A. Sibley. Physical Review B 26 (1982) 2382−2389.

16

[19]

J. Linnros, N. Lalic, A. Galeckas, V. Grivickas. Journal of Applied Physics 86 (1999)

6128−6134. [20]

R. Chen, P. L. Leung. Radiation Measurements 37 (2003) 519−526.

[21]

H. J. Seo, T. Tsuboi, K. Jang. Physical Review B 70 (2004) 205113:1‒8.

[22]

J. R. Lacowicz. Principles of Fluorescence Spectroscopy. Springer Science, New York,

2006. [23]

R. Alcala, C. B. Gonzalez, B. Villacampa, P. J. Alonso. Journal of Luminescence 48/49

(1991) 569−573. [24]

P. S. May, H. U. Güdel. Chemical Physics Letters 175 (1990) 488−492.

[25]

W. E. Vehse, K. H. Lee, S. I. Sun, W. A. Sibley. Journal of Luminescence 10 (1975)

149−162. [26]

M. Azadeh. Fiber Optics Engineering. Springer Science, New York, 2009.

[27]

J. R. Seretlo, J. J. Martin, E. Sonder. Physical Review B 14 (1976) 5404−5412.

[28]

A. R. Beierholm, C. E. Andersen, R. L. Lindvold, F. Kjaer-Kristoffersen, J. A. Medin.

Radiation Measurements 43 (2008) 898−903. [29]

E. G. Yukihara, V. H. Whitley, J. C. Polf, D. M. Klein, S. W. S. McKeever, A. E.

Akselrod, M. S. Akselrod. Radiation Measurements 37 (2003) 627‒638. [30]

G. V. M. Williams, C. Dotzler, A. Edgar. Journal of Materials Science: Materials in

Electronics 20 (2009) S268‒S271.

17

FIGURE CAPTIONS FIGURE 1: (a) Photoluminescence spectra of NaMgF3:Er3+ (0.5%), showing emission excited at 27780 cm‒1 (360 nm, red) and excitation detected at (1530 nm, blue). (b) Energy level diagram for Er3+ indicating with vertical lines the transitions observed in the emission spectrum depicted in (a). FIGURE 2: Photoluminescence emission decay for the emission at 18350 cm−1 (545 nm) from NaMgF3:Er3+ (0.5%) and the emission at 20160 cm−1 (496 nm) from NaMgF3:Ni2+ (0.5%). Both were excited by a 1.0 ms pulse at 26670 cm−1 (375 nm). Solid lines are stretched exponential fits to the decay of Er3+ (red) and Ni2+ (blue). FIGURE 3: (a) Photoluminescence spectra of NaMgF3:Ni2+ (0.5%) showing emission excited at 24630 cm−1 (406 nm, red), and excitation detected at 9010 cm−1 (1110 nm, blue). Absorption coefficient for NaMgF3:Ni2+ (5.0%) and the spin allowed transitions (black) from the Ni2+ ground state are plotted against the right hand vertical axis. (b) Energy level diagram for Ni2+ indicating with vertical lines the transitions observed in the emission spectrum depicted in (a). FIGURE 4: (a) Radioluminescence emission of NaMgF3:Er3+ (0.5%) with emission not observed in the PL indicated by *. (b) Integrated radioluminescence intensity as a function of time for selected emission peaks. FIGURE 5: (a) Radioluminescence emission of NaMgF3:Ni2+ (0.5%) with emissions no observed in the PL indicated by *. Also shown is the transmittance of un-doped NaMgF3 after a similar period of irradiation, illustrating radiation colour centre absorptions. (b) Integrated radioluminescence intensity as a function of time for selected emissions.

18

FIGURE 6: Glow curve for NaMgF3:Er3+ (0.5%) measured at 1 K/s (open circles). The plotted intensity is integrated under the 18350 cm−1 and 19050 cm−1 emissions. Also shown is a second order kinetics fit (red) and the individual fitted peaks (black). The inset shows the emission spectrum of the highest intensity glow peak near 310 K, which is representative of all glow peaks.

19

3

100

101

10

2

10

(a)

Intensity (arb. units)

5

850

10

λexc = 360 nm

1250

15

650

20 −1

25

Wavelength (nm) 450

Wavenumber (cm )

550

350

30

35

λems = 1530 nm

40 ×103

250

I11/2

I15/2

I13/2

4

4

4

F9/2

4

S3/2

4

H11/2

2

4 (b) G7/2 1

19050 cm

27780 cm

1

18350 cm

1

11830 cm

1

8200 cm

1

1

1

6500 cm 1 (approx.)

10200 cm

15270 cm

Non-radiative or below 6500 cm 1

Intensity (arb. units)

0

10 0.0

-3

10

-2

10

-1

10

0.5

1.0

2+

Ni

1.5

2.0

2.5 3.0 Time (ms)

3.5

3+

Er

4.0

4.5

5.0

Intensity (arb. units)

3

100

101

10

2

10

(a)

5

3

10

T2g( F)

3

850

650

3

15

T1g( F)

3

λexc = 405 nm

2000 1250

−1

25

T1g(3P)

Wavenumber (cm )

20

3

Wavelength (nm) 550 450

30

35

λems = 1100 nm

350

40

−1

×103

0

100

200

300

400

500

600

250 700

Absorption Coefficient (m )

A2g (3F)

3

3

T2g ( F)

3

T1g (3F)

3

Eg (1D)

1

T2g (1D)

1

(b) 3T1g (3P)

20160 cm

1

13800 cm

1

9020 cm

1

6170 cm

1

A1g (1G)

1

Intensity (arb. units)

(a)

10

0

10

1

10

850

*

15

650

*

550

*

20 25 −1 Wavenumber (cm )

*

Wavelength (nm) 450 350

30

*

3

35 ×10

(b)

Intensity (arb. units)

0.20 0.15 0.10 0.05 0.00 1.20 0.90 0.60 0.30 0.00 0.20 0.15 0.10 0.05 0.00 0

−1

2000

6000 Time (sec)

4000

8000

10000

λ = 400 nm k = 25000 cm−1

λ = 525, 545 nm k = 19050, 18350 cm−1

λ = 845 nm k = 11830 cm

Intensity (arb. units)

(a)

100

101

10

2

10

850

15

*

650 550

20

450

25

250

30 35 40 −1 Wavenumber (cm )

*

Wavelength (nm) 350

45

50 ×10

3

55

0

0.2

0.4

0.6

0.8

1

(b)

Intensity (arb. units)

Transmitance

0.24 0.18 0.12 0.06 0.00 1.20 0.90 0.60 0.30 0.00 0.40 0.30 0.20 0.10 0.00 0

2000

6000

8000

10000

λ = 290 nm k = 34480 cm−1

Time (sec)

4000

−1

λ = 500 nm k = 20000 cm−1

λ = 730 nm k = 13400 cm

Intensity (arb. units) 10

200

-2

-1

10

100

250

0

300

50

350

100

400

15

300

350

600

650

20 25 30 −1 Wavenumber (cm )

450 500 550 Temperature (K)

10

Temperature (°C) 150 200 250

Intensity (arb. units)

700

35

400

750