Visualizing the valence states of europium ions in Eu-doped BaAl2O4 using X-ray nanoprobe mapping

It is anticipated that X-ray nanoprobes will open new avenues with significant characterization ability for unravelling the emission mechanisms of phosphor materials.


Introduction
Phosphor-converted white-light-emitting diodes (wLEDs) have been studied widely to develop more stable and efficient wLEDs (Xia et al., 2019). Investigating the emission mechanism of long-afterglow phosphors will help realize various applications in lighting. Alkaline earth aluminates are good candidates for host matrices for luminescent materials, and are widely used as hosts for rare earths. Persistently luminescent phosphors composed of alkaline earth aluminates can be represented by the general formula XAl 2 O 4 (X = Mg 2+ , Ca 2+ , Sr 2+ or Ba 2+ ) (Lephoto et al., 2012). Recently, the widely used traditional host matrix BaAl 2 O 4 has received extensive attention because of its unique properties, such as low cost, easy synthesis, high physical stability and better luminescence efficiency Lephoto et al., 2012;Tian et al., 2021;Rezende et al., 2012Rezende et al., , 2016Rezende et al., , 2015.
Many researchers have extensively investigated methods to obtain perfect luminescent phosphors. For example, Tian et al. (2021) reported a new BaAl 2 O 4 -YAG:Ce composite ceramic phosphor for high-efficiency wLEDs. Yin et al. (2020) reported a BaAl 2 O 4 :Eu 2+ -Al 2 O 3 composite ceramic to enhance the luminescence output. Rafiaei et al. (2020) reported the synthesis, crystal structure and optical and adsorption properties of BaAl 2 O 4 :Eu 2+ , Eu 2+ /L 3+ (L = Dy, Er, Sm, Gd, Nd, and Pr) phosphors. As BaAl 2 O 4 doped with rare earths has significant application in luminescent phosphors, in this report ISSN 1600-5775 we propose a powerful analysis method to help researchers improve their manufacturing methods.
Rare earth ions play the most important roles as activators in these phosphors. In particular, the valence states of the rare earth ions significantly affect the emission properties of the phosphors. In the case of a europium (Eu)-doped BaAl 2 O 4 phosphor, a broad-band emission at $500 nm can be attributed to the Eu 2+ emissions (4f 5d ! 4f transitions), and the narrow emission peaks at around 560-750 nm are associated with the Eu 3+ emissions ( 5 D 0 ! 7 F i , i = 0 to 4). X-ray nanoprobe (Sham, 2014;Martínez-Criado et al., 2014; techniques using a synchrotron source can be applied to characterize these phosphors (Huang et al., 2021). By exploiting the advantages of X-ray nanoprobes, including the continuously tunable X-ray energy (4-15 keV) and excellent spatial resolution of the nano-focused X-ray beam (<100 nm), we can easily and quickly investigate the valence states of the rare earth ions in the selected local area to unravel the emission mechanisms of phosphor materials. An X-ray nanoprobe can be used to perform X-ray absorption spectroscopy (XAS) across the L-edges of Eu (6.977 keV), Dy (7.790 keV), Er (8.358 keV), Sm (6.716 keV), Gd (7.243 keV), Nd (6.208 keV) and Pr (5.964 keV) to obtain information on the valence states.
In this study, we developed visualization methods for the characterization of the Eu-doped BaAl 2 O 4 phosphor. Using an X-ray nanoprobe, X-ray fluorescence (XRF) and X-ray excited optical luminescence (XEOL) mapping can clearly reveal the distributions of the constituent elements, the valence states of the Eu 2+ and Eu 3+ ions, and the different emission wavelengths ( em ). The accuracy of the estimated valence state distributions was examined using XAS spectra. As the X-ray nanoprobe can provide excellent spatial resolution, we selected different local areas with different valence states of the Eu 2+ and Eu 3+ ions to study their emission properties. The XEOL spectra consist of one broad intense peak at $500 nm and narrow emission peaks at around 560-750 nm in the local areas richer in Eu 2+ and Eu 3+ , respectively. In addition, a weaker emission at $390 nm, which is related to the F colour centre of -Al 2 O 3 , is observed.

Experiment
The XAS, XRF and XEOL experiments were conducted on the Taiwan Photon Source (TPS) 23A X-ray nanoprobe beamline located at the National Synchrotron Radiation Research Center (NSRRC) in Taiwan. The capabilities of the TPS 23A X-ray nanoprobe beamline have previously been described in detail by Lin et al. (2019Lin et al. ( , 2020. This beamline can deliver an X-ray beam spot size of less than 60 nm. The test powder sample of Eu-doped BaAl 2 O 4 (Ba 0.97 Eu 0.03 Al 2 O 4 ) was doped with Eu 2+ and Eu 3+ ions and purchased from Dott Technology. The XAS and XRF spectra were measured using a silicon drift detector (SDD; Vortex-ME4, Hitachi). The XEOL spectra were collected using a multimode optical fibre (with a core diameter of 400 mm) attached to a spectrometer (iHR320, Horiba) with a deep thermoelectric cooling charge-coupled device (Syncerity BI UV-Vis) and a resolution of 2048 Â 70 pixels. The XEOL mapping images were acquired through a photomultiplier tube, which is installed in another spectrometer (iHR550, Horiba).

Results
The mapping images used to visualize the Eu-doped BaAl 2 O 4 phosphor, shown in Figs. 1-3, have the same measured areas. Through XRF and XEOL mapping, detailed information about the measured area can be obtained. Fig. 1 shows the XRF and XAS analyses of the Eu-doped BaAl 2 O 4 phosphor. As the X-ray energy was tuned at 6.985 keV, which is above the Ba L 3 -edge (5.247 keV), Al K-edge (1.560 keV) and Eu L 3 -edge (6.977 keV), the elemental distributions of Ba, Al and Eu were observed directly, as shown in Figs. 1(a), 1(b) and 1(c), respectively. The Eu L 3 -edge XAS spectra exhibit a strong white line caused by the electronic transitions from 2p 3/2 to 5d. In particular, the resonances of the Eu 2+ and Eu 3+ ions are located at 6.975 and 6.983 keV, respectively.
The core electrons in the Eu 3+ ions have a larger binding energy than those in the Eu 2+ ions, with the difference between the resonance energies equal to $8 eV (Korthout et al., 2013). On the basis of features of the Eu 2+ and Eu 3+ ions, XAS can reveal the corresponding amounts of the Eu 2+ and Eu 3+ ions. Fig. 1(d) shows the XAS spectrum of the area marked with the white dashed circle in Fig. 1(c). The XAS spectrum of this local area shows that the fluorescence yield of the Eu 2+ resonance (6.975 keV) is higher than that of the Eu 3+ resonance (6.983 keV), which indicates that the highlighted local area is richer in Eu 2+ .  Although the elemental distribution of Eu can be obtained from Fig. 1(c), it cannot provide information on the positions of the Eu 2+ and Eu 3+ ions. Therefore, we developed a visualization method that directly images the valence states of the Eu ions. As the X-ray energy of the synchrotron source is continuously tunable, we selected three X-ray energies at 6.970, 6.975 and 6.983 keV for XRF mapping; these correspond to the background, Eu 2+ resonance and Eu 3+ resonance, respectively. Using the following equations, we subsequently obtained the distributions of the valence states of the Eu ions, Assuming that the concentrations of the Eu 2+ and Eu 3+ ions are x% and y%, respectively, the fluorescence yields of I 6.975 keV and I 6.983 keV can be estimated using equation (1). The corresponding ratio parameters C 1 , C 2 and C 3 were determined from the XAS spectra of the reference materials for Eu 2+ (EuS) and Eu 3+ (Eu 2 O 3 ) (Korthout et al., 2013). In the report by Korthout et al., the Eu L 3 -edge XAS spectra of the EuS and Eu 2 O 3 reference materials show that the values of I Eu 2þ ð6:975 keVÞ , I Eu 2þ ð6:983 keVÞ , I Eu 3þ ð6:975 keVÞ and I Eu 3þ ð6:983 keVÞ were 1.9, 1, 0.36 and 2.1, respectively. So, the ratio parameters C 1 , C 2 and C 3 can be calculated to be equal to 1.9, 0.17 and 0.476, respectively.
After estimating the above parameters, we obtained the distributions of the Eu 2+ and Eu 3+ ions using equations (6) and (7), as shown in Figs. 2(a) and 2(b), respectively. To verify the accuracy of the estimated results, we obtained the XAS spectra of the selected local areas to measure the valence states of the Eu ions. Fig. 2(c) shows the XAS spectrum of the area marked with the red dashed circle in Fig. 2(a). The XAS spectrum in Fig. 2(c) shows that the fluorescence yield of the Eu 2+ resonance (6.975 keV) is higher than that of the Eu 3+ resonance (6.983 keV), indicating that this local area is richer in Eu 2+ . According to our estimation, another local area marked with the red dashed circle in Fig. 2(b) is clearly richer in Eu 3+ . Fig. 2(d) illustrates the XAS spectrum of the local area marked in Fig. 2(b), which is consistent with the expected results. In Fig. 2(d), it can be observed that the fluorescence yield of the Eu 2+ resonance (6.975 keV) is lower than that of the Eu 3+ resonance (6.983 keV), indicating that this local area is richer in Eu 3+ .
The XAS spectra demonstrate the accuracy of the estimated results, implying that the valence states of the Eu ions can be easily and quickly visualized using XRF mapping. This model relies on the approximation that EuS and Eu 2 O 3 are used as the reference for Eu 2+   Estimation results for the distributions of (a) Eu 2+ and (b) Eu 3+ ions. Panels (c) and (d) show XAS spectra of the locations/areas marked with the red dashed circles in panels (a) and (b), respectively. The XAS spectra demonstrate the accuracy of the estimated results, implying that the valence states of the Eu ions can be easily and quickly visualized using XRF mapping. increase the discrepancy while the difference between the resonance energies is smaller.
Although this model can provide the relative quantitative values of Eu 2+ and Eu 3+ , the validity of the quantitative values of the valence states can be measured by XAS spectra. In the quantitative analysis reported by Yamamoto & Yukumoto (2018), the oxidation state of Eu in phosphor samples was established using XAS to investigate the influence of the constituent elements, absorption edge and measurement mode on the evaluated oxidation states.
The variation in the valence states of the Eu ions tends to influence the emission properties. Because the Eu 2+ and Eu 3+ ions have different distributions, as shown in Fig. 2, we used XEOL mapping to study further the emission distribution of the main emission wavelength of the Eu-doped BaAl 2 O 4 phosphor. Figs. 3(a), 3(b), 3(c) and 3(d) show the XEOL mapping at emission wavelengths ( em ) of approximately 390, 500, 588 and 698 nm, respectively. The emission distributions of the Eu-doped BaAl 2 O 4 phosphor at different emission wavelengths ( em ) were clearly visualized. At em = 500 nm, the well documented emission of the Eu 2+ ions due to the 4f 5d ! 4f transitions was observed, along with emissions of the Eu 3+ ions at em = 588 and 698 nm due to transitions of the 5 D 0 ! 7 F i (i = 0 to 4) states (Rezende et al., 2016). However, we suggest that em = 390 nm is attributed to the F centre of -Al 2 O 3 , which will be discussed later. Thus, XEOL mapping can be used to visualize clearly the emission behaviour of the Eu-doped BaAl 2 O 4 phosphor. Compared with Figs. 3(a), 3(c) and 3(d), Fig. 3(b) shows a larger emission intensity and better emission uniformity. This result demonstrates that the main contribution to the luminescence intensity of Eu-doped BaAl 2 O 4 comes from the Eu 2+ activator.
To study further the emission properties of the Eu-doped BaAl 2 O 4 phosphor in different local areas, we selected four such local areas, indicated by the white dashed circles in Figs. 3(a)-3(d), and plotted the corresponding XEOL spectra. The room-temperature XEOL spectra of the four local areas (P1-P4) for excitation across the Eu L 3 -edge (6.977 keV) are shown in Figs. 4(a)-4(d).
The XEOL spectra corresponding to each of the four local areas were acquired at three X-ray energies: below the Eu L 3edge (6.960 keV), at the Eu 2+ resonance (6.975 keV) and at the Eu 3+ resonance (6.983 keV). The XEOL spectra of P1-P4 exhibit one common broad intense peak at 500 nm, which can be attributed to the Eu 2+ 4f5d ! 4f transitions.
The above results not only reinforce the conclusion that the Eu 2+ activator dominates the luminescence intensity in the Eu-doped BaAl 2 O 4 phosphor, but also illustrate that the Eu 2+ ions have unique local symmetry in the matrix. According to Rezende et al. (2011Rezende et al. ( , 2016, Eu 2+ ions may be incorporated on more than one non-symmetric site, as substitution might occur at either the Al 3+ or Ba 2+ site, which results in more than one band in the emission spectra associated with the 4f 5d ! 4f transition. Another shoulder emission due to the doping of Eu 2+ ions on the non-equivalent sites can be observed in the intense peak. In addition, the emission intensity of the single   Figs. 3(a)-3(d). The XEOL spectra corresponding to each of the four local areas were acquired at three X-ray energies: below the Eu L 3 -edge (6.960 keV), at the Eu 2+ resonance (6.975 keV) and at the Eu 3+ resonance (6.983 keV).
intense peak in P1-P4 decreased slightly as the X-ray energy was tuned across the Eu L 3 -edge. This behaviour suggests that the energy transfer to the Eu 2+ 4f5d ! 4f transitions is less efficient above than below the Eu L 3 -edge (Huang et al., 2021).
In addition to the single intense peak at $500 nm produced by the Eu 2+ 4f 5d ! 4f transitions, the XEOL spectra of P1-P4 exhibit other weaker peaks. The XEOL spectra of P1 shown in Fig. 4(a) consist of a special weaker emission line at $390 nm. This emission may be attributed to the F-centre emission in -Al 2 O 3 (den . The F-centre is a type of colour centre that is associated with an oxygen vacancy with two electrons (Wang et al., 2013;Ghamnia et al., 2003;Itou et al., 2009). den Engelsen, Fern, Ireland, Yang & Silver (2020) also observed an F-centre emission in the Eu-doped BaAl 2 O 4 phosphor via photoluminescence and cathodoluminescence.
The XEOL spectra of P2 shown in Fig. 4(b) consist of only one intense peak produced by the 4f5d ! 4f transitions of Eu 2+ and very weak peaks associated with the 5 D 0 ! 7 F i (i = 0 to 4) transitions of the Eu 3+ ions, showcasing the perfect luminescence intensity of the Eu-doped BaAl 2 O 4 phosphor. Compared with P1 and P2, the XEOL spectra of local areas P3 and P4 clearly consist of emission peaks associated with the 5 D 0 ! 7 F i (i = 0 to 4) transitions of the Eu 3+ ions, as shown in Figs. 4(c) and 4(d), respectively. Clearly, the emission lines caused by the Eu 3+ transitions in P3 and P4 are narrower than those in P1 and P2, indicating more crystal homogeneity in local areas P3 and P4 than in P1 and P2. The broad Eu 3+ emission is a result of crystal inhomogeneity causing small distortions around the Eu ions (Rezende et al., 2016;Gasparotto et al., 2008).
To study the valence states of the Eu ions, the same four local areas P1-P4 were used to measure the Eu L 3 -edge XAS spectra, as shown in Figs. 5(a)-5(d). The coexistence of the two common valence states of Eu 2+ and Eu 3+ can be seen in the XAS spectra of P1-P4, represented by the two well resolved edge resonances. Although the Eu 2+ and Eu 3+ ions coexist, the corresponding concentrations of the Eu 2+ and Eu 3+ ions in these local areas can be determined from the fluorescence yield of the XAS spectra. The XAS spectrum of P1 shows that the fluorescence yield of the Eu 2+ resonance is similar to that of the Eu 3+ resonance, indicating similar concentrations of the Eu 2+ and Eu 3+ ions in local area P1. However, the XAS spectrum of P2 shows that the fluorescence yield of the Eu 2+ resonance is higher than that of the Eu 3+ resonance, indicating that local area P2 is richer in Eu 2+ . Compared with P1 and P2, the XAS spectra of P3 and P4 show the opposite behaviour, that is, the Eu 3+ resonances have a higher fluorescence yield and therefore local areas P3 and P4 are richer in Eu 3+ .
The results of the XAS spectra are consistent with those of the XEOL spectra, as shown in Fig. 5. Because local area P2 is richer in Eu 2+ , the XEOL spectra of P2 consist of only one intense peak at $500 nm produced by the Eu 2+ 4f 5d ! 4f transitions. As P3 and P4 are richer in Eu 3+ , their XEOL spectra consist of narrow emission lines around 580-700 nm that are produced by the 5 D 0 ! 7 F i (i = 0 to 4) transitions of the Eu 3+ ions.
Since we can obtain the distributions of the Eu 2+ and Eu 3+ ions from the results of Figs. 2(a) and 2(b), a pixel-by-pixel analysis can be conducted to determine the overall correlation between em = 390, 500, 588 and 698 nm and Eu 2+ and Eu 3+ ions. For a given map, em emission intensity and Eu ions were plotted against each other for each pixel in the map. Figs. 6(a)-6(d) and 6(e)-6(h) show the emission intensity of em = 390, 500, 588 and 698 nm as a function of Eu 2+ and Eu 3+ ions, respectively. The correlations are consistent with the measured results of the XEOL and XAS spectra. The emission intensity of em = 390 nm shown in Figs. 6(a) and 6(e) has largest emission intensity at around 50% Eu 2+ or Eu 3+ ions, suggesting that P1 has similar concentrations of Eu 2+ and Eu 3+ ions. Figs. 6(b) and 6(f) show that the emission intensity of em = 500 nm is a zero correlation with Eu 2+ or Eu 3+ ions. This suggests that, regardless of whether local areas P1-P4 are richer or poorer in Eu 2+ or Eu 3+ ions, the single intense peak at $500 nm is still the main contribution to the luminescence intensity. The emission intensities of em = 588 and 698 nm show negative correlation with the Eu 2+ ions shown in Figs. 6(c) and 6(d), and positive correlation with the Eu 3+ ions shown in Figs. 6(g) and 6(h). This result is also corroborated by the fact that the local areas P3 and P4 are richer in Eu 3+ . Thus, the emission mechanisms of Eu-doped BaAl 2 O 4 phosphors can be further understood through such a correlation analysis.

Conclusions
In this paper, we report powerful characterization capabilities for investigating the features of Eu-doped BaAl 2 O 4 phosphor materials using an X-ray nanoprobe. XRF    phosphors, including the elements, the valence states of the Eu ions and the different emission wavelengths ( em ). The accuracy of the estimated valence state distributions was examined by performing XAS across the Eu L 3 -edge (6.977 keV), and the corresponding concentrations of the Eu 2+ and Eu 3+ ions were obtained from the XAS spectra.
Exploiting the excellent spatial resolution of the X-ray nanoprobe, we selected four local areas with different valence states of the Eu 2+ and Eu 3+ ions to study their emission properties. The XEOL spectra consisted of one broad intense peak at $500 nm and narrow weaker emission peaks at around 560-750 nm in the local areas richer in Eu 2+ and Eu 3+ , respectively. In addition, a weaker emission at $390 nm relating to the F colour centre of -Al 2 O 3 was also observed.
The XEOL spectra demonstrated that the main contribution to the luminescence intensity of Eu-doped BaAl 2 O 4 comes from the Eu 2+ activator and the emission intensity will not be influenced by the concentration of Eu 2+ or Eu 3+ ions.
We believe that X-ray nanoprobes will open new avenues with significant characterization ability for unravelling the emission mechanisms of phosphor materials.