Recovery rate and homogeneity of doping europium into luminescent metal hydrides by chemical analysis †

During the investigation of concentration dependent properties of phosphors, such as emission intensities, the knowledge of the activator ion concentration is of great importance. Herein we present a study on recovery rate and homogeneity of the activator ion concentration by chemical analysis in luminescent europium doped metal hydrides. The analysis method was established on the model system Eu x Sr 1 (cid:1) x H 2 and applied to brightly emitting hydridic perovskites LiMH 3 :Eu 2+ and LiMD 3 :Eu 2+ (M ¼ Sr, Ba). The nominal activator ion concentrations calculated from initial weights are in good agreement with those determined via ICP-MS for 10 (cid:1) 3 % # x both for the hydrides (deuterides) and the parent alloys. The synthesis protocol thus allows a reliable method for reproducibly doping europium into metal hydrides. Luminescence spectra of LiEu x Sr 1 (cid:1) x H 3 with x ¼ 0.0037% show a broad band emission at room temperature typical for Eu 2+ 4f 6 5d 1 – 4f 7 transitions. Below 80 K, a vibronic ﬁ ne structure is observed with vibrational coupling frequencies of approximately 100, 370 and 970 cm (cid:1) 1 .


Introduction
Inorganic luminescent materials play an important role in efficient energy conversion and contribute to saving energy for lighting purposes.2][3] The electric-dipole and parity-allowed 4f 7 -4f 6 5d transition in Eu(II) produces high quantum yields and thus high emission intensities.In contrast to trivalent europium with an f-f transition Eu(II) further exhibits a strong dependence of the emission wavelength on the coordination of the activator ion.Eu(II) in oxide and halide host lattices oen emits blue or green light, 2 while more polarisable ions like suldes and nitrides induce emission in the green to red colour region. 2,4Luminescent materials usually contain only small amounts of the activator ion and for efficient host lattices luminescence can oen be observed even for ppm amounts of the activator ion.Many investigations have been carried out to study concentration dependencies of the luminescence intensity.Exact control of the activator ion concentration is of great importance, since luminescence properties may be strongly inuenced.For instance, too high concentrations oen lead to energy transfer and consequently to luminescence quenching. 5owever, analyses of the exact content of the activator ion or the local coordination environment via a method different from luminescence spectroscopy are rather scarce and concentrations are usually calculated according to initial weights.Lately, an NMR and ESR study regarding the Nd 3+ and Gd 3+ content in LaPO 4 was carried out, studying doping rates down to approximately 0.1 at%. 6hemical analyses of europium species in solution via inductively-coupled plasma-mass spectrometry (ICP-MS) are commonly used, e. g. in rock samples, 7 in water, 8,9 in ne chemicals, 10 in biological samples 11,12 or in non-radioactive model systems using europium as chemical homologue for the highly toxic americium (as one of the main components in radioactive waste) to analyse the complexation behavior of europium with natural organic matter 13,14 as well as the sorption/desorption properties in pore water/clay systems. 15nother feature oen neglected concerns the homogeneity of the activator ions within the host materials.In a recent study of Eu(II) in a metal-organic framework it was shown that at least for higher doping rates (20%) on a nanometer to low micrometer scale the europium content varied between 12-28%, proving a considerable heterogeneity. 16ecently, we have studied Eu 2+ luminescence in alkaline earth metal hydrides 17 and perovskite type hydrides. 18,19In order to check whether the europium concentrations of our luminescent materials calculated based on the initial weights are in accordance with the real europium content and whether europium is distributed homogeneously within the host lattice, we present a method of analysis of the phosphors using the model system Eu x Sr 1Àx H 2 within this work.EuH 2 and SrH 2 are isotypic and the existence of a solid solution series was shown before. 17n strontium-rich samples a broad band emission at 728 nm was observed and assigned to the Eu 2+ 4f 7 -4f 6 5d transition.The system was chosen due to its relative simplicity and we will also show that the methods can be applied for ternary systems such as the yellow emitting LiSrH 3 :Eu 2+ . 18pectra of SrH 2 , LiSrH 3 and LiBaH 3 with relatively high doping rates can be found in. 17,18SrH 2 samples with relatively low doping rates do not show emission, because emission intensities of alkaline earth metal hydrides are rather low.However, LiSrH 3 :Eu 2+ shows a bright emission even at very low doping rates and herein we also present emission spectra of a low doped sample whose europium concentration was determined during our study.

Sample preparation
Due to their air sensitivity, all solids were handled in an argon lled glove box.The alloys Eu x Sr 1Àx were synthesised by melting reactions of the elements (Eu ingot Alfa Aesar 99.9%, Sr pieces, distilled, Alfa Aesar 99.9% and Kristallhandel Kelpin, all mechanically surface cleaned before use) in niobium ampoules in a high-frequency furnace and hydrogenated in an autoclave at 650 K and approx.100 bar H 2 pressure (H 2 Praxair 99.9%) for about 2-3 days.The autoclave and crucibles consist of the hydrogen resistant Nicrofer® 5219 alloy (Inconel 718).Preparation of samples Eu x Sr 1Àx with x < 0.05 was performed by melting alloys with higher europium concentration with corresponding amounts of strontium metal.A list of the initial weights is given in the ESI.† The ternary hydrides were prepared as described in. 18,19Note that therein the europium concentrations are given with respect to the substitutional site M in AMH 3 , therefore 0.5% corresponds to an overall value of 0.25% in our present study.For ICP-MS measurements, three to four samples were taken from each intermetallic phase in order to assure representative sampling and to probe homogeneity, but only one sample from each hydride.

X-ray powder diffraction
X-ray powder diffraction data were collected on a PANalytical Philips X'Pert Pro with focusing Bragg-Brentano geometry and a ne focus X-ray tube (CuKa 1,2 radiation).The ratio of CuKa 1 and Ka 2 was determined using a silicon standard (NIST640d).Finely ground samples were enclosed between kapton foils with apiezon grease.The data collection time was 60 min and the diffraction range 10 # 2q # 110 .Crystal structures were rened using the program TOPAS 4.2 (Bruker AXS, Karlsruhe, Germany) 20 together with the fundamental parameter approach. 21The instrumental function was determined empirically by means of a reference scan of LaB 6 .Typically, complete crystal structures including hydrogen positions were taking as starting structures from literature and scale factors, lattice parameters, zero point shi, thermal displacement and microstructural parameters were rened.

Chemicals and standards for ICP-MS
The quality of all chemicals used was analytical reagent grade or better (e.g.suprapure) from Merck.All solutions were prepared with MilliQ deionised water (18.2MU, MilliQ-System, Millipore, Schwalbach, Germany).Single element standards of europium and holmium Certipur® from Merck were used and Argon 5.0 (99.999%,Praxair) as plasma gas.

ICP-MS measurements
For calibration a 1000 ppm europium stock solution was prepared by dissolving europium metal in HNO 3 (3.25%)and diluting to a concentration of 1 ppm europium.10 mL of a 10 ppm holmium solution as internal standard was added to correct time dependent sensitivity deviation of the ICP-MS system to the calibration samples as well as all other samples.Additionally, 300 mL of HNO 3 (65%, suprapure) were added to prohibit metal sorption onto the sample container surface.MilliQ water was then added to obtain the different europium concentrations (5-500 ppb).
All samples were rst dissolved in 1 mL HNO 3 (3.25%)and stocked up to a concentration of 1000 ppm (according to initial weights of the samples).Samples with europium concentration of 7 Â 10 À5 -7 Â 10 À7 % were diluted with a total sample dilution of 1 : 3 using an ICP-MS solution (300 mL HNO 3 65%, 1 mL internal standard, 636 mL MilliQ water).To all other samples 10 mL of internal standard and 300 mL of HNO 3 (65%) were added and stocked up with MilliQ water to reach a volume of 10 mL.9.5 mg of the pure strontium samples were dissolved in 5 mL HNO 3 (3.25%).10 mL of internal standard, 300 mL of HNO 3 (65%) and MilliQ water were added to reach a volume of 10 mL.
An Agilent 7500cx ICP-MS (Agilent Technologies, Santa Clara, USA) was used for the isotopic measurements; detailed analytical conditions are given in Table 1.A correction of the isotope intensities via normalisation by the intensity of 165 Ho as internal standard was carried out and samples were measured in triplicate. 15RSD values are in the range of 0.1 and 2.2%.

Luminescence measurements
Temperature dependent photoluminescence emission spectra and lifetime measurements were carried out on an Edingburgh FLS-920 spectrouorometer with a pulsed 376.8 nm Edingburgh laser diode for sample excitation, a monochromator according to Czerny-Turner (500 nm focal length) for emission, a R928P (Hamamatsu) photomultiplier tube for detection of the emission spectra and a H7422-02 photomultiplier tube (Hamamatsu) for detection of the decay curves.Low temperature measurements were carried out using an Oxford liquid helium cold nger.Due to moisture and air sensitivity, samples were enclosed in sealed silica tubes of 0.5-1 cm diameter and they were attached to the cold nger using high purity silver paint in order to facilitate heat transfer and minimise the lagging of real temperatures behind the nominal ones assigned to the spectra.Spectra were corrected for lamp intensity and photomultiplier sensitivity.Luminescence decay curves were recorded under pulsed excitation with the Edinburgh 376.8 nm pulsed diode laser.Neglecting a slight deviation from single exponential behavior in the higher temperature range (faster initial decay), a single exponential function was used for the analysis of the luminescence decay curves.

Results and discussion
Melting of the elements produced alloys Eu x Sr 1Àx with a silver luster, which were hydrogenated to yield metal hydrides Eu x Sr 1Àx H 2 as ne powders.For samples with 0.5% # x # 2% yellow to orange colors were observed, whereas samples with lower europium concentrations were colorless.Samples were single-phase according to X-ray powder diffraction and Rietveld renement (Fig. 1).Due to the low doping concentrations and like ionic radii of Sr 2+ and Eu 2+ rened lattice parameters vary very little and cannot reect the concentration of Eu 2+ properly given the accuracy of a routine X-ray powder diffraction measurement (Table 2).
Chemical analysis of the dissolved starting materials by ICP-MS gave 0.225 ppm europium in strontium from Alfa Aesar (99.9%) and 0.01 ppm europium in the strontium rod from Kristallhandel Kelpin.Thus samples Eu x Sr 1Àx and Eu x Sr 1Àx H 2 with europium contents x # 5 Â 10 À2 % were prepared from the latter material.As a consequence of the 0.01 ppm europium in the strontium metal europium in Eu x Sr 1Àx and Eu x Sr 1Àx H 2 samples the lowest reasonably achievable doping content was 10 À6 %.
Analyses of alloys Eu x Sr 1Àx show differences in europium content up to a factor of 2.8 between samples taken from different spots on the same alloy (Table 3).This hints at some inhomogeneity in the distribution of europium in strontium, which is not unexpected.Such inhomogeneity in the alloys most probably will cause inhomogeneity in the hydrides as well, which could not be measured, however, due to decrepitation of the alloys to a ne powder upon hydrogenation.
The relative deviation of europium content as determined by ICP-MS from the nominal contents in the metal hydrides Eu x Sr 1Àx H 2 is only 6% for x ¼ 0.5% and increases with decreasing europium content up to 70% for x ¼ 5 Â 10 À4 % (Table 3).Samples with lower europium content were analysed as well; however, the concentrations found for those samples lie within the same range as the blank samples and are therefore not reliable.At least for the higher contents shown in Tables 2  and 3, which are in the range of conventional activator contents in phosphors, the correspondence between nominal and experimentally determined europium contents is satisfactory (Fig. 2 and 3).Nominal europium contents are thus a reasonable measure of the actual content at least for 10 À3 % # x, where the deviations were found to be between 6% and 32%.
The analysis of the ternary luminescent materials yielded the following europium concentrations (see Table 4).
Again, the agreement between nominal and experimentally observed values is satisfactory.Nominal values based on the initial weights may therefore be regarded as a reasonable measure for the true activator content in europium doped metal hydrides prepared by melting of the metals and subsequent hydrogenation, even for very low dopant concentrations; however, europium distributions in powder samples may be inhomogeneous.This might lead to varying local environments and thus broadening of absorption and emission bands.In contrast to higher doped LiSrH 3 samples, which are of a light yellow colour, LiSrH 3 :Eu 2+ 0.0037% is colourless.However, in spite of the low europium content it still shows a bright yellow emission as described for samples with higher doping rates. 18In Fig. 4 temperature dependent spectra of LiSrH 3 :Eu 2+ 0.0037% are given.In good agreement with, 18 the spectra show a broad band emission at room temperature which is typically observed for Eu 2+ 4f 6 5d 1 -4f 7 emission and below approximately 80 K, a vibronic ne structure is observed.Due to the low europium content, this ne structure shows a slightly better resolution than that of the higher concentrated samples and higher modes are still clearly visible and can be assigned.Thus it may be concluded that the spectral resolution may be improved by the use of lower doping rates; if the emission intensities allow for it.The vibrational energies with respect to the zero phonon line 0-0-0 are listed in the ESI, † Table 1 and the assignment of the bands was done according to. 18A vibrational coupling with three different is observed (approx.100, 370 and 970 cm À1 ).
The temperature dependence of the lifetimes is shown in Fig. 5.In good agreement with earlier results, 18 the quenching temperature is approximately 150 K.

View Article Online
With about 530 ns, the lifetime at 10 K is slightly larger than for a higher europium concentration (0.25 mol%, approximately 420 ns).This indicates, together with the slightly better resolved vibronics, that in higher doped samples concentration effects might already play a role.It is well known that higher doping rates may lead to concentration quenching and a shortening of the observed lifetimes. 5

Conclusion
The exact knowledge of the doping concentration and homogeneity of the activator ion in luminescent samples is of great importance for concentration dependent effects such as emission intensities.Chemical analysis on europium doped alloys and their metal hydrides was performed by ICP-MS.Nominal activator ion concentrations calculated from initial weights in Eu x Sr 1Àx and Eu x Sr 1Àx H 2 are in good agreement with experimentally determined concentrations for 10 À3 % # x.The synthesis protocol thus allows a reliable method for reproducibly doping europium into metal hydrides.A good agreement between nominal and experimental values was also achieved for compounds LiMH 3 :Eu 2+ and LiMD 3 :Eu 2+ (M ¼ Sr, Ba) at dopant levels around 0.25% and at 0.0037% for LiSrH 3 :Eu 2+ .For the latter luminescence spectra show a broad band emission at room temperature which is typically observed for Eu 2+ 4f 6 5d 1 -4f 7 transitions.Below 80 K, a vibronic ne structure is observed with vibrational coupling frequencies of approximately 100, 370 and 970 cm À1 .

Fig. 2
Fig. 2 Log-log plot showing the relation between the nominal and real europium contents in the intermetallic phases Eu x Sr 1Àx .

Fig. 3
Fig. 3 Log-log plot showing the relation between nominal and real europium contents in the hydrides Eu x Sr 1Àx H 2 .

Table 1
Operating parameters and analytical conditions of the ICP-MS À1 ) This journal is © The Royal Society of Chemistry 2015

Table 2
Lattice parameters in Eu x Sr 1Àx H 2 as determined by Rietveld refinement on X-ray powder diffraction data

Table 3
Nominal and experimentally determined europium contents x of alloys Eu x Sr 1Àx and their hydrides Eu x Sr 1Àx H 2 in %