Harvesting Light from BaHfO3/Eu3+ through Ultraviolet, X-ray, and Heat Stimulation: An Optically Multifunctional Perovskite.

Materials with optical multifunctionality such as photoluminescence (PL), radioluminescence, and thermoluminescence (TL) are a boon for a sustainable society. BaHfO3 (barium hafnium oxide [BHO]) under UV irradiation demonstrated visible PL endowed by oxygen vacancies (OVs). Eu3+ doping in BHO (BHOE) introduces f-state impurity levels just below the conduction band for both Eu@Ba and Eu@Hf sites, causing efficient host-to-dopant energy transfer, generating intense 5D0 → 7F1 magnetic dipole transitions (MDT) with internal quantum yield of ∼70%. X-ray photoelectron spectroscopy and electron paramagnetic resonance showed the formation of OVs in both BHO and BHOE samples with more vacancies in the doped sample. The positron lifetime measurements suggested that Eu3+ ions are distributed at both Ba2+ and Hf4+ sites. The association of OVs with Hf4+ and Eu3+ ions due to high charge/radius ratio is considered to be responsible for lowering the symmetry around Eu3+ ions to C 4v in BHOE. Density functional theory studies of defect formation energy justified the same. Time-resolved emission spectroscopy showed distinct spectra for Eu@Ba and Eu@Hf sites corresponding to symmetric and asymmetric environments, respectively. This could be highly relevant in designing color tunable phosphor by forcing dopant ions at one specific site because Eu@Ba displayed orange emission whereas Eu@Hf displayed red emission. We could further harness BHOE for X-ray scintillator application by designing a thin film, which showed efficient conversion of high-energy X-ray into visible light. Under beta irradiation; both BHO and BHOE showed distinct TL glow curves as shallow traps were formed in the former and deep traps in the latter, which could have long-term implications in harnessing this material for persistent luminescence. We believe that BHO/BHOE demonstrated an extraordinary credential as a perovskite for multifunctional applications in the area of defect-induced light emission, UV phosphor, X-ray scintillator, and TL crystals.

Introduction

Materials designed with an objective to achieve superior optical properties require deep understanding of the defects, traps, radiation/structural stability, doping efficiency, dopant’s local structure, efficacy of host-to-dopant energy transfer, and so forth.[1−5] The set of objectives in making high quantum yield phosphor, which can have multifunctional applications, are specific to the need. On top of that, a material that can serve as a phosphor, scintillator, and thermoluminescent dosimeter would be a boon to the optical material scientific community as it will drastically reduce load on material engineers and designers.[6−8]

Efficient phosphor is expected to have significant global dominance in the coming days for designing phosphor-converted white light-emitting diodes (pc-LEDs) as well as display panels.[2] Red phosphor in particular is in extremely high demand to resolve the current existing problem of highly correlated color temperature (more than 5000 K) and low color rendering index (CRI, lower than 80) in the commercial white phosphor (YAG/Ce3+ coupled with GaN blue LED).[9]

Scintillators are luminescent materials that convert high-energy radiations such as gamma, X-ray, and so forth into thousands of low-energy photons immediately and are found to be extremely useful in the area of security, photodynamic therapy, and radiation detection.[7] Because of its deeper penetration than near-infrared light, X-ray has several advantages in the medical science. In the literature, radioluminescence (RL) measurements are mostly focused on single crystals owing to their high transparency and superior energy conversion efficiency.[10] However, single crystal growth is very tedious, costly, and time taking and needs very specialized laboratories and experimental facilities. On the other hand, many scintillating materials can be useful in powder form as well, such as in X-ray intensifier screens.[7] Powder materials for RL are expected to give more reliable qualitative information in luminosity, time response, and emission spectra.[7] Thermoluminescence (TL) measurements on functional materials give very valuable information related to the nature and density of traps present. Moreover, the spectral positions, peak profiles, and intensities of the TL glow curve provide important information about the trapped states/charges and energy transfer processes in a crystalline lattice, resulting in the emission of light.[8]

Perovskite structures in general have demonstrated excellent ability to host rare earth ions for light-emitting applications.[11−13] Barium hafnium oxide (BHO) with the formula BaHfO3 represents a very interesting class of nonpolar perovskite materials with superlative properties such as a large band gap ∼6 eV, high dielectric constant, low phonon frequency <700 cm–1, high transparency in UV–visible region, which makes it an ideal host for rare earth-doped photoluminescence (PL).[14−16] BHO also has very high density (8.5 g/cc) and high effective atomic number (Zeff = 64.6), making it an ideal host for scintillators and X-ray phosphors.[17] In fact, the BHO sample itself demonstrated broad RL at around 410 nm, which disappears on doping, though.[18] Dobrowolska et al. explored the RL of Eu3+-doped BaHfO3 (BHOE) and proposed that europium-occupying Hf4+ site is amenable only to X-ray excitations, and hence, they bring out a strategy to force the dopant into either the Ba2+ or the Hf4+ site.[19] The same group pointed out that one can force europium at A and B sites by co-doping Y3+ and La3+, respectively.[20] There have been few reports on PL properties of BHOE as well. In one of the works, the authors have stabilized both Eu3+ and Eu2+ to produce white light emission.[18] Guzman–Olguin also discussed about PL of BHOE under UV excitation but did not give any information on the role of defects, local dopant structure and host–dopant energy transfer efficacy.[14] Drag–Jarzabek also explored PL of BHOE, but their work was more focused on designing the same using a single molecular precursor rather than photophysical aspects.[21] There have been few reports on PL of other RE, such as Pr3+- and Tb3+-doped BHO.[15,17]

Based on all the above discussions, it is quite clear that the local structure of the dopant is very important for achieving efficient PL or RL. However, reports pertaining to that are mostly based on experimental results and are without theoretical validations. Moreover, with such low doping, the prediction made by only experimental data may not be reliable/extrapolated, curtailing the advancement of BHOE as a reliable UV phosphor or X-ray scintillator. We took the initiative to resolve this issue by performing defect formation energy calculations to decipher information pertaining to the local site occupancy of europium ions in BHO. Moreover, defects are expected to have profound influence on PL and RL, and there is no report to the best of our knowledge wherein defects have been probed in BHO and BHOE. In this work, we have used positron annihilation lifetime spectroscopy and density functional theory (DFT) calculations to understand the defects in both the samples. Probably for the first time, we report the appearance of bright dual blue and green bands in undoped samples and unravel their origin by electronic structure calculations using DFT.

Experimental Section

Both BHO and BHOE were synthesized using a high-temperature solid-state diffusion method. Powder X-ray diffraction (XRD) measurements were taken using a Proto-AXRD benchtop system using a Cu Kα line of ∼1.5405 Å. An Edinburgh-made fluorescence spectrometer (model CD 920) was used for the PL measurements. RL measurements were taken on polytetrafluoroethylene-embedded BaHfO3 discs. Positron annihilation lifetime measurements were taken on powder samples using a lifetime spectrometer having a time resolution 265 ps. The TL was measured using the Lexsyg research imaging TL-OSL-RF system. The instrumentation and other experimental procedures were the same as those used in our earlier works. To avoid repetition, further details related to synthesis, sample preparation for RL measurements, instrumentation, and computational methodology are given in Sections S1–S4 of the Supporting Information, respectively.

Results and Discussion

This section deals with all the experimental and the theoretical results pertaining to BHO and BHOE.

X-ray Diffraction

Figure a,b shows the XRD patterns of BHO and BHOE. The patterns are completely identical, which shows that europium doping has not resulted in any unwanted impurity or did not lead to any phase segregation of Eu2O3. The pattern completely matches with that of standard BaHfO3 (JCPDS no. 24-0102) with the cubic perovskite structure having a space group of Pm3̅m. Though all major diffraction peaks match with the BHO phase, two impurity peaks corresponding to HfO2 (ICDD-741506) located at 2-theta value of 28 and 31° are also seen in the XRD pattern of BHO and BHOE, which is not expected to affect the PL properties.

Figure 1

(a) Powder XRD pattern of BHO and BHOE along with the standard JCPDS file no. 24-0102 and (b) crystal structure of BHO.

BHO shows an ideal cubic perovskite structure with Ba at the cube corners, Hf at the center of the cube, and oxygen atoms at the face-center of the cube, forming regular octahedrons presented in Figure b.

Probing Oxygen Vacancy: EPR and XPS Measurements

X-ray photoelectron spectroscopy (XPS) measurements were taken to understand the formation and evolution of oxygen vacancy in BHO and BHOE perovskites. XPS survey scans of BHO and BHOE are shown as Figure a. The scans do not show the presence of any impurity peak and show all the characteristic elemental peaks. Figure b shows the O 1s XPS spectra from BHO and BHOE. The main peak could be fitted into three major peaks at 529.3, 531.2, and 532.6 eV in both the samples. In addition, a small peak at 525.9 eV is observed in BHO, but its origin is not clear yet. The areas under each peak are also indicated in Figure b. The peak at 529.6 eV is due to lattice oxygen (OL), while the peak at 531.2 eV is due to oxygen vacancies (OV) or defects (OV), and the peak at 532.6 is due to adsorbed oxygen (OA) on the surface of the sample.[22−24] From the areas under the various deconvoluted peaks in O 1s XPS spectra, the fraction of OV area is 0.24 and 0.35 in BHO and BHOE, respectively. This suggests an increase in the OVs after Eu doping, as also suggested by positron annihilation lifetime data. Enhanced OVs might be caused by partial doping of Eu at Hf sites, as indicated by PL and DFT studies discussed later.

Figure 2

(a) XPS survey scans, (b) O 1s XPS spectra, and (c) room temperature X-band EPR spectra of BHO and BHOE.

Electron paramagnetic resonance (EPR) is considered a very sensitive technique for probing the electronic spin state. Because both BHO and BHOE are non-magnetic, they are not expected to give any PER signal on their own. Figure c shows the room temperature EPR spectrum of BHO and BHOE. Both the samples showed EPR signal around g ∼ 2.00 (H ∼ 333 mT), typical of an electron-trapped defect center and mainly associated with singly ionized oxygen vacancy (VȮ), which is also known as F+ center.[25] This singular and asymmetric EPR signal is associated with VȮ, and other vacancy-related defects are reported to vary between 1.9560 and 2.0030, endowed by a variation in synthesis, local structure, and thermal conditions.[25] Broadness of this particular peak further reflects disordered and heterogeneous environments surrounding the vacancy similar to the surface of nanostructure materials.

PL Spectroscopy of BHO

Figure a shows the emission spectra of non-activated BHO samples under several excitation wavelengths. The emission profile broadly shows similar spectral features irrespective of excitation photons. The only difference that could be seen is the emission intensity, and the one excited with 250 nm energy shows the maximum emission output. The Gaussian deconvoluted spectra shown in Figure b under 250 nm excitation showed a dual band peaking at 435 and 515 nm in the blue and green regions, respectively. The color coordinate diagram shows predominant green emission marked with the letter b (Figure d). ABO3 perovskite crystals are known to possess abundant defects in their band gap owing to their very high structural flexibility.[26,27] Depending on their localization, they can be classified further as shallow and deep defects. Under exposure to UV photons (∼250 nm), these shallow and deep defects lead to creation of localized states in the band gap of BHO, and further inhomogeneous charge distribution leads to the formation of electron traps. We could not find any report wherein PL properties of BHO have been discussed barring one. Ye et al.[28] have synthesized single crystalline hollow micro- and nanospheres of BHO and observed several PL peaks in the range of 400–650 nm under 214 nm excitations. They have ascribed the violet-blue band to the shallow and green-yellow band to deep defects present in the wide band gap of BHO. However, their analogy was based on other reports published on PL properties of CaZrO3, BaZrO3, SrZrO3, SrTiO3, and so forth.[29−32] Park et al.[33] have carried out temperature-dependent PL on SrHfO3 and proposed the absence of any substantial PL at room temperature but could observe dual bands at 1.75 and 2.5 eV at low temperatures, which they attributed to defect states originating from structural disorder or OVs. OV-induced visible PL was also observed in HfO2 thin films and ThO2 nanocrystals.[34,35] We also believe in the prominent role of defects in PL properties of BHO, but the nature/configuration of actual defects is difficult to be predicted unless the data are complemented by other experimental and theoretical studies. This is more so for identifying the nature of OVs, whether they are neutral, singly ionized or doubly ionized. We have carried out positron annihilation lifetime spectroscopy as well as DFT calculations to probe the origin of the blue and green bands in BHO (Figure ).

Figure 3

(a) Excitation-dependent emission spectra, (b) Gaussian deconvoluted emission spectra under 250 nm excitation, (c) PL excitation spectra under 435 and 515 nm emission, and (d) luminescence decay profile for blue and green bands of BHO.

Figure 5

(a) Excitation-dependent emission spectra, (b) PL excitation spectra under 598 and 616 nm emissions, (c) luminescence decay profile for the598 nm band, and (d) time-resolved emission spectra of BHOE.

Figure 8

Alignment of energy levels for BHO in the absence and presence of vacancy defects and europium ions.

The excitation spectra (Figure c) recorded under 435 and 515 nm emissions showed multiple peaks located approximately around 235, 250, 265, 280, 295, and 310 nm. Such multiple excitation bands again suggest the presence of several defect states in BHO. The 235 nm peak is ascribed to the band edge of BHO. Other peaks have contributions from defect states as well as O2– → Hf4+ charge transfer. Luminescence lifetime measurements (Figure d) on the two emission bands, viz. 435 and 515 nm, under 250 nm excitation show biexponential behavior. Lifetime values for BHO for the 435 nm blue band are 5.9 and 15.8 μs with their relative intensities as 54 and 46%, respectively. The green band at 515, on the other hand, has a lifetime of 7.3 and 20.8 μs with relative intensities as of 73 and 27%, respectively. These clearly suggested that different defects contribute to blue and green PL bands. The one having a lifetime value of ∼6–7 μs has a major contribution in both the emission bands, whereas the one having a longer lifetime of ∼16–21 μs contributes relatively less. The fractional contribution varies for both blue and green PL bands. Papernov et al.[35] have reported that the visible PL from OV states in HfO2 thin films has a microsecond-scale lifetime.

Theoretical Band Gap and Density of States for BHO

The calculated unit cell parameter (4.199 Å) is found to be close to the experimentally reported value (4.171 Å).[24] This justifies the reliability of the present computational methodology. The electronic structure of BHO has been described by analyzing total density of states (DOS) and projected DOS (PDOS) (Figure a). The calculated band gap is 3.55 eV, which is underestimated due to the limitation of the standard density functional theory. As can be seen from Figure a, the valence band maximum (VBM) is mostly dominated by O 2p states, while the conduction band minimum (CBM) is mainly contributed by Hf 5d and Ba 4d states. Critical analysis of PDOS indicates that there exists minor contribution of the Hf 5d states to the upper part of the valence bands, indicating covalent nature between Hf and O. Similarly, the bottom part of the conduction bands shows hybridized states of Hf 5d and Ba 4d states with a small contribution from O 2p states. This behavior is consistent with that reported in an earlier theoretical study.[36]

Figure 4

DOS of (a) pure BHO, (b) BHO in the presence of neutral oxygen vacancy (VO), (c) BHO in the presence of singly charged oxygen vacancy (VO1+), and (d) BHO in the presence of doubly charged oxygen vacancy (VO2+). The vertical dashed line indicates the Fermi level.

To model BHO with oxygen vacancy defect, we considered a 2 × 2 × 2 supercell, which contained a total number of 40 atoms (8 Ba, 8 Hf, and 24 O), and removed 1 oxygen atom. Electronic structure calculation was carried out with the optimized geometry. Figure b shows the DOS plot for BHO in the presence of one neutral oxygen vacancy, which indicates the presence of discrete occupied impurity states, 0.43 eV below the CBM. This is due to the presence of excess electrons in the system. Analysis of PDOS indicates that the defect state is the hybridized state of Hf (s, p, d), Ba (d), and O (p) orbitals. In the case of singly positively charged oxygen vacancy, there exists a partially occupied defect state in the mid-gap region (1.63 eV above the VBM) while a partially unoccupied state, adjacent to the CBM, which results in lowering of CBM energy (Figure c). In the case of doubly charged oxygen vacancy, the DOS (Figure d) looks very much similar to that of the neutral variety. However, all the energy levels are shifted by 0.20 eV in the upward direction in comparison to that of the neutral variety.

PL Spectroscopy of BHOE

Figure a shows the emission spectra of BHOE under 235, 248, and 254 nm excitations. The spectral features remain the same and show only a small variation in intensity. PL excitation spectra shown in Figure b show only a broad peak around 250 nm due to O2– → Eu3+ charge-transfer transitions, and the forbidden intra f–f transitions with weaker intensity corresponding to 7F0 → 5L6 (395 nm), 7F1 → 5L3 (414 nm), and 7F0 → 5D2 (467 nm) can be seen.

The emission spectra consist of typical europium peaks located at 575, 598, 616, 654, and 708 nm due to 5D0 → 7F0, 5D0 → 7F1, 5D0 → 7F2,5D0 → 7F3, and 5D0 → 7F4 transitions, respectively. Host emission almost completely disappears. This suggests that under 250 nm host excitation, efficient host-to-europium energy transfer takes place. The color coordinate diagram shows an orange-red emission marked with letter “p” in Figure d. The details of the color coordinate calculation methodology have been mentioned in Figure S5.

Figure 9

(a) RL emission spectra of BHOE. (b) Comparison of RL and PL emission intensities for BHOE. (c) Mechanism of RL and (d) color coordinate diagram for PL in BHO (indicated as “b”) and PL (“p”)/RL (“r”) of BHOE. The inset of (a) depicts the film used for measurements, and the inset of (b) shows the calculated asymmetry ratio value for RL and PL.

The color purity of the samples was calculated using eq where (x, y), (xd, yd), and (xi, yi) are the color coordinates of the phosphor sample, dominant red region wavelength (616 nm in this case), and white illumination, respectively. In this work, (xd, yd) for 616 nm was  (0.681, 0.323), and the standard (xi, yi) coordinate was (0.310, 0.316). The color purity of the present sample was found to be 96.57, which suggests its high potential as an efficient red-emitting phosphor for pc-LEDs.

The peak at 575 nm due to 5D0 → 7F0 is allowed neither by electric dipole transition (EDT) nor by magnetic dipole transition (MDT) and is observed only in the case when europium ions have a local site symmetry of C1, C, or ≤ C. Other important transitions are located at 598 and 616 nm, which are allowed, respectively, via MDT and EDT. In europium spectroscopy, 598 has pure MDT (ΔJ = ±1) origin and 616 has hypersensitive EDT (ΔJ = ±2) origin, and their ratio (616/598), known as asymmetry ratio (AR), is an important marker for local symmetry of europium ions. The fact that the intensity of the 598 nm peak is much stronger than that of the 616 nm peak clearly suggests that europium ions are preferentially occupying the symmetric site though their fractional distribution at the asymmetric site could not be completely ruled out as the 616 nm peak also has substantial intensity.

In the BHO perovskite lattice, Ba2+ ions are 12-coordinated (161 pm) in the form of a cube, whereas Hf4+ ions are 6-coordinated (71 pm) as octahedra. The size of 6-coordinated europium ions is about 101 pm, and its incorporation in small-sized Hf4+ is thermodynamically very unstable and will lead to very high lattice strain and distortion. Moreover, the charge difference between Eu3+ and Hf4+ will further invoke the need for OVs for charge compensation, which will create additional strain in the BHO lattice. The combined strain and distortion is expected to induce severe lowering of symmetry around europium ions in BHOE. Furthermore, every europium ion that occupies such a small-sized hafnium ion will create very high strain in the lattice, though some fraction of Eu3+ ions that occupy Hf4+ could not be completely ruled out. The large-sized Ba2+ ions, on the other hand, trigger large-fraction Eu3+ ions substituted at Ba2+ sites, though such charge mismatch invokes the generation of barium vacancies (BVs) for charge compensation. However, local symmetry around europium ions in BHOE will not have an inversion symmetry due to the generation of substantial amounts of BVs and distortion due to size mismatch. PALS data discussed in later sections also suggest that this charge compensation could also be achieved through creation of oxygen interstitials. The appearance of charge-compensating defects could be the reason for observation of MDT/EDT non-allowed 5D0 → 7F0 transition, which ultimately lowers the symmetry around europium ions. At the same time, Eu3+ occupying Hf4+ with adjacent OVs as charge compensating defects leads to 5D0 → 7F2 EDT.

We have also carried out Stark splitting analysis (Figure ) as induced by the host crystalline field. The 616 nm peak, owing to 5D0 → 7F2 transition, is hypersensitive in nature, and hence, its intensity is strongly affected by the local Eu3+symmetry and the host crystalline field. In this case, this peak is split into two bands (Figure c). On the other hand, the 598 nm MDT peak, owing to 5D0 → 7F1 transition, is not affected by external factors, and hence, its intensity is negligibly affected by the Eu3+ symmetry or the host crystalline field. In this case, this peak does not present any splitting (Figure b). The other EDT located at 708 nm, due to 5D0 → 7F4 transition, also undergoes quadruple splitting. Based on the number of stark components, it can be deciphered that local site symmetry around europium ions in BHO is C4,[30] even though MDT is more intense than EDT, suggesting high symmetry of europium ions in BHOE. However, the actual symmetry is lowered by lattice strain owing to size mismatch and the presence of BVs and OVs as charge-compensating defects endowed by the dopant and lattice site/charge mismatch.

Figure 6

Stark splitting analysis of (a) 5D0 → 7F0, (b) 5D0 → 7F1, (c) 5D0 → 7F2, and (d) 5D0 → 7F4 of BHOE.

The PL lifetime decay curve of BHOE for the 598 nm band under 250 nm excitation is shown in Figure c. The decay profile demonstrates biexponential behaviorwhere I(t) is the PL intensity, t is the time after excitation, and Ti is the decay lifetime for the ith component with intensity I. I0 is the background or detector zero offset.

Different photophysical phenomena can contribute to such multiexponential behavior, such as different emissive centers, host-to-dopant energy transfer, defects, and so forth.[37]

After biexponential fitting, we could deduce two different lifetime values: short-lived with T1 = 1.42 ms and intensity of 25%, and long-lived with T2 = 3.84 ms and intensity of 75%.

Considering the selection rule and photophysical process governing the f–f transition, a long-lived T2 component should be ascribed to symmetric europium ion, as in that case the 4f → 4f transition of Eu3+ ions would be more LaPorte forbidden. The shorter lived T1 component, on the other hand, should be ascribed to relatively low symmetry europium ions as the 4f → 4f transition of Eu3+ ions would be more relaxed. Based on our earlier discussion, we proposed that europium ions in BHOE mostly occupy the Ba2+ site, which results in the formation of BVs via

At the same time, fractions of europium ions occupy the Hf4+ site as well, which leads to the formation of OVs via

In BHO perovskite lattice, these defects are randomly distributed. There would be two different scenarios for europium occupancy: and . A large fraction of europium ions occupy the barium site, with BVs as charge-compensating defects (. However, there can be other less probable situations wherein europium ions would occupy the hafnium site, with OVs as charge-compensating defects (. The T1 component (25%) could be ascribed to ( and is responsible for both 5D0 → 7F0 and 5D0 → 7F2 transitions. The T2 (75%) component, on the other hand, could be attributed to ( and is responsible for the 5D0 → 7F1 MDT transition.

Based on PL emission and lifetime data analysis, we inferred that T1 (1.42 ms, 25%) arises because of Eu3+ ions occupying Hf4+ with configuration (, whereas the major component T2 (3.84 ms, 75%) could be ascribed to Eu3+ ions occupying Ba2+ with configuration (. In order to identify the emission spectra of these three components, we carried out time-resolved emission spectroscopy (TRES). Through TRES, by applying suitable delay times and selecting proper slicing window, we could elucidate the emission spectra of individual components, as shown in Figure d. The emission spectra of T1 have almost equally intense EDT/MDT, large spectral splitting, and significantly intense 5D0 → 7F0 transition, clearly suggesting that this particular europium ion has very low symmetry arising because of Eu3+ ions occupying the Hf4+ site near OVs (. The emission spectra of T2 have very intense MDT, low degree of spectral splitting, and complete absence of 5D0 → 7F0 transition, clearly suggesting that this particular europium ion has very high symmetry with the center of inversion and arises definitely because of Eu3+ ions occupying the Ba2+ site with configuration (. Thus, one can achieve efficient and tunable PL in BHOE by preferential doping of Eu@Ba2+ or Eu@Hf4+ sites, with the former resulting in orange emission and the latter displaying red emission.

Local Site of Europium Ions in BHO: a DFT Study

In the present study, we considered 2 × 2 × 2 supercells, which contain a total of 40 atoms (8 Ba, 8 Hf, and 24 O), and introduced 1 Eu at the cationic lattice site. We considered two different scenarios: substitution of one Ba site by Eu and substitution of one Hf lattice site by Eu, which are represented by EuBa and EuHf, respectively. We optimized both types of defect structures and calculated the defect formation energy to find out the preferred lattice site for Eu using the relationship[38,39]where Edoped and Eperfect represent the energy of the doped and defect-free BaHfO3, calculated with equal supercell size; μ represents the chemical potential of the element X, and n stands for the number of elements added (q = −1) or replaced (q = +1) to model the doped system. The value of formation energy of Eu at the Ba lattice site is 1.5 eV, which is lowered in comparison to that of Eu at the Hf site, with a value of 3.51 eV. The difference between the two cases is just ∼2.0 eV. This clearly suggests that europium has a larger preference to occupy Ba2+, but a smaller fraction can occupy Hf4+ sites as well.

Positron Annihilation Lifetime Spectroscopy

The positron lifetime spectra in both the samples could be fitted to a maximum sum of three lifetime components, with the third component of 2 ns accounting to only <1% of intensity. This component is observed in all powder samples due to positronium formation on the surface of the powder samples. The other two lifetime components, their intensities, and the intensity weighted average positron lifetimes in these samples are given in Table .

Table 1

Summary of Positron Lifetime Data in Undoped and Doped BaHfO3

sample	τ1 (ps)	I1 (%)	τ2 (ps)	I2 (%)	τave (ps)
BHO	171.1 ± 4.9	63.8 ± 4.8	308 ± 11	35.5 ± 4.7	220 ± 17
BHOE	186.0 ± 3.9	71.6 ± 3.7	345 ± 13	27.9 ± 3.6	231 ± 15

The first positron lifetime (τ1) in oxide powder samples is usually due to positron annihilations from the delocalized bulk, while the second positron lifetime is due to positron trapping in the defects. In undoped samples too, the two lifetime components are obtained, showing that the undoped sample also has significant defects. It is to be noted that anion vacancies are poor in trapping positrons, while the cation vacancies and associated defects trap positrons very efficiently. The ab initio calculated positron lifetime in BaTiO3 was reported to be 152 ps;[40] Panda et al., on the other hand, have obtained a lifetime of 159 ps in BaCeO3.[41] The first lifetime component of 170 ps is close to the expected value from defect-free bulk, with contribution from shallow positron traps like OVs. The presence of these vacancies is also evidenced in PL emission spectrum. The second positron lifetime of 308 ps has to be from either Ba vacancy or vacancy cluster associated with Hf vacancy such as VHf–Vo. The calculated positron lifetime for Ba vacancy in BaTiO3 is 293 ps.[40] Upon Eu doping, it is seen that both positron lifetimes increase with marginal reduction in the intensity corresponding to the second lifetime component. Marginal decrease in the intensity of the second component shows that the positrons are less efficiently trapped. Partial trapping of positrons in the defects is also expected to lower the bulk positron lifetime. The doping of Eu at Ba2+ sites can be charge compensated by BVs or oxygen interstitials, while the doping of Eu at the Hf site can be charge compensated by the formation of OVs. As discussed earlier, the formation of BVs is expected to cause very efficient positron trapping in these defects or saturation trapping of positrons, leading to single lifetime components corresponding to BVs. Alternatively, the Eu3+ can be distributed between Ba2+ and Hf4+ sites, which effectively charge compensates without the necessity for any additional charge-compensating defects.

The positron lifetimes suggest that the Eu is doped at both Ba2+ and Hf4+ sites and hence does not cause drastic changes in the positron lifetimes. The substitution of Eu@Ba also seems to be compensated by oxygen interstitials rather than BVs, as indicated by minimal change in the defect positron component (I2), which is expected to increase drastically if BVs were significant. However, the marginal increase in the individual as well as the positron lifetimes still suggests that the doping is causing the formation of defects. It is likely that even though Eu is being doped at both Hf and Ba sites, the OVs seem to be more associated with Hf and Eu due to high charge/radius ratio. The association of OVs might be responsible for the asymmetry observed in the PL emission of BHOE.

Efficacy of Host Sensitized Energy Transfer: a DFT Study

Figure a,b represents the DOS for BHOE for Eu occupying both Ba and Hf sites, respectively. The impurity state appears only at the adjacent CBM for the EuBa system, whereas the impurity state appears adjacent to both VBM and CBM for the EuHf system. The impurity state adjacent to the CBM is composed of the Eu (f) orbital. The impurity states close to the VBM are contributed by the O (p) and Eu (f) orbitals. It is interesting to note that in both the cases Eu introduces impurity states just below the CBM, which are contributed by mainly Eu (f) state. This is consistent with the experimental observation of disappearance of host emission and intense Eu-to-host emission due to doping with Eu.

Figure 7

DOS for Eu-doped BaHfO3 with (a) Eu at the Ba lattice site (EuBa) and (b) Eu at the Hf lattice site (EuHf). The vertical dashed line indicates the Fermi level.

To summarize the electronic structure of BHO in the presence of vacancy defects and europium ions, we aligned the energy levels for the defect-containing system with respect to that of defect-free BaHfO3 (Figure ). The effective gap is found to be reduced by 1.08, 1.92, and 1.17 eV due to discrete defect levels for VO, VO1+, and VO2+, respectively, with respect to the VBM–CBM gap of ideal BHO. The behavior for VO and VO2+ is consistent with the experimental optical observation for BHO. Thus, one can conclude that the observed optical property that is different from that of the ideal system is attributed to the presence of oxygen vacancy.

RL Spectroscopy of BHOE

Considering the problem associated with powder sample as far as commercial viability is considered owing to its poor adhesion, lower flexibility, and mechanical strength, we ensembled BHOE powder in the form of a film for RL measurement. The photograph of the film is shown in the inset of RL emission spectra shown in Figure a. The RL emission spectra could not be detected in the case of undoped BHO samples, but the BHOE sample displayed efficient visible light emission under X-ray excitation. It needs to be understood here that the mechanism of excitation is different with UV photons (PL) and highly energetic X-ray photons (RL). PL excitation with UV activates the luminescent ions, whereas X-ray interacts with holes and electrons of the host, and it is possible that X-ray excitation may induce change in dipole moment. The RL emission spectrum shown in Figure a shows typical europium features as was observed in PL (Figure a).

Based on earlier studies by Dobrowolska and group,[19,20] it was found that only the europium ions occupying the Hf4+ site were amenable to be efficiently excitable with X-ray, that is, the T1 component (. The integral emission intensity of MDT for PL was found to be more intense than that for RL (Figure b).This could be attributed to the fact that only 25% of europium ions occupy the Hf4+ site compared to 75% occupying the Ba2+ site. This was also well established using formation energy calculation using DFT. Moreover, the value of AR is slightly higher for RL compared to PL (inset of Figure b). This again could be correlated to the fact that , which is X-ray excitation active, has higher distortion and strain compared to . This is also reflected by the red color high purity, as can be seen from the color coordinate diagram (indicated by r) in Figure d. The mechanism of RL (Figure c) typically involves four steps: (1) generation of excitons via photoelectric effect on X-ray irradiation; (2) relaxation of electrons and holes to generate secondary electrons, holes, phonons, plasmons, photons, and so forth; (3) thermalization of electrons and holes leading to their transport near the band edge; and (4) their recombination with the emission center (europium ions in this case) to generate visible light.

TL of BHOE

The TL measurements were taken to find out the TL-related defects in the material. It is evident from Figure that pure BHO gives relatively weaker TL compared to BHOE. The TL measurements on BHO depicted two low-temperature TL peaks: one high-intensity peak at a temperature of 72 °C and another low-intensity peak at 121 °C. This suggests the presence of two types of shallow traps in BHO. Because the low-temperature TL peaks (72 °C) possess relatively higher intensity compared to high-temperature TL peaks (121 °C), this means that the samples possess increased numbers of shallow traps and are also likely to fade at faster rates and, thus, are not suitable for long-term measurements. We believe the dual-band BHO is associated with these two shallow traps. These shallow traps in BHO are probably associated with the presence of OVs and responsible for PL.

Figure 10

TL glow curves for BHO and BHOE for an absorbed dose of 20 Gy using 90Sr/90Y beta source.

For BHOE, only a singular TL peak at a high temperature of ∼180 °C (453 K) is observed, which suggests the presence of only deep traps in BHOE unlike BHO. The dosimetric characteristic of thermoluminescent phosphor depends mainly on the kinetic parameters associated with its glow peak as they provide very important information related to TL mechanism. The most important TL parameter is trap depth (Et) also known as activation energy. It is defined as thermal energy needed to liberate the trapped electrons and holes. Others are the order of the kinetics (b) and the frequency factor (s). Order of the kinetics (b) can be deciphered by calculating the geometrical factor (μg) of the glow curve from the known values of shape parameters using eq .[42]

The trap depth (Et) was estimated as the average of three energies (Eτ, Eδ, and Eω) using Chen’s method.[43,44]

Various parameters of TL curves determined from above expressions are given in Table .

Table 2

Peak and Kinetic Parameters from TL Glow Curves in Figure

sample	Tm (°C)	ω (°C)	Tl (°C)	Th (°C)	τ (°C)	δ (°C)	μg	Eτ (eV)	Eδ (eV)	Eω (eV)	Et (eV)	freq factor
BHO	72.4	24.3	60.0	84.3	12.3	12.0	0.493	1.330	1.293	1.320	1.313	3.7 × 1018
	121.3	45.9	91.9	137.8	29.3	16.5	0.360	0.517	0.437	4.890	0.481	1.1 × 105
BHOE	179.7	92.7	137.7	230.4	41.9	50.7	0.548	0.632	0.664	0.650	0.649	1.1 × 106

The first-order peaks are considered asymmetrical in nature with τ (= Tm – Tl) almost 50% larger than δ(= Th – Tm) and are characterized by a geometrical factor μg ∼ 0.423. Second-order peaks, on the other hand, are practically symmetrical (δ = τ) and are characterized by a geometrical factor μg ∼ 0.524.[42] BHOE displays the most intense glow peak located at 180 °C. The nature of the peak was found to be symmetrical, and the shape factor μg for the same was calculated as ∼0.548, which suggests that this peak obeys second-order kinetics. This further highlighted the fact about BHOE, which suggests that the probability of retrapping before recombination is non-zero. On the other hand, BHO displayed a dual band: 72 and 121 °C, both of which obey mixed-order kinetics. The dominant process in second-order kinetics is that electrons and holes are retrapped in their respective traps on heating. Moreover, based on the calculated activation energy values, it can be postulated that there are two different kinds of traps in BHO: one being shallow and the other deeper in nature, with trap depths of 0.48 and 1.31 eV. On the other hand, in a doped BHOE perovskite sample, there is only one kind of trap with a depth of ∼0.65 eV.

Judd–Ofelt Calculation: Photophysical Properties

We further calculated various radiative properties such as radiative transition rate (AR), non-radiative transition rate (ANR), branching ratios (β), Judd–Ofelt (J–O) parameters (Ω), and internal quantum yield (IQY) using the Judd–Ofelt analysis method. The details of various calculations used have been reported elsewhere.[37,45] The corrected emission spectrum was used for area calculations, and an average lifetime value of 2.87 was used.

Because MDT is not affected much by crystalline field and other external factors, its transition rate invariably remains constant with an approximate value of 50 s–1.[46] For BHO, the refractive index value is 2.16. Emission quantum efficiency of the 5D0 emitting level of BHOE was calculated usingwhere the AR rate was obtained by summing over the radiative rates for each 5D0 → 7FJ (J = 1–4) transition. The J–O parameter and other photophysical values are mentioned in Table .

Table 3

Photophysical Properties of Eu3+-Doped LaPO4

	ARAD (s–1)	ANRAD (s–1)	η (%)	Ω2 (10–20 cm2)	Ω4 (10–20 cm2)	β1 (%)	β2 (%)	β4 (%)
BHOE	269	112	∼70.1	1.25	3.72	48	32	22

It was observed that for BHOE, the Ω4 value is greater than the Ω2 value, indicating the existence of a symmetric environment around europium ions in BHO, which is highly probable because the majority of europium ions occupy larger sized Ba2+ and the environment is less distorted and more flexible. The Eu–O bond has lower covalence and is less polarizable in BHOE. This is also reflected in the intense MDT compared to EDT in our PL emission spectra shown in Figure a.

From Table , it can also be seen that radiative and non-transition rate values are, respectively, 269 and 112 s–1, which gives the IQY value of 70.1%.

This value was calculated based on radiative and non-radiative lifetimes theoretically based on J–O calculations by taking the reciprocal of AR and ANR. τR  in this case would be 3.7 ms and τNR would be 8.93 ms. The PLQY calculated based on τf and τR would be 68.23%. Moreover, it is slightly different from the one calculated (∼70%) based on radiative and non-radiative transition rates probably because excited state lifetime is more affected by non-radiative channels. The branching ratio values showed higher contribution of orange emissions as expected due to higher intensity of 598 nm MDT compared to 616 nm EDT.

Conclusions

The important role of OVs (neutral as well as ionized) was manifested in visible PL observed in BHO, and the results were corroborated using DFT and lifetime spectroscopy. XPS and EPR confirmed the presence of OVs in both BHO and BHOE perovskites. No drastic change in positron lifetime was observed, which suggested that the Eu3+ ion is distributed at both Ba2+ and Hf4+ sites with doping -induced formation of defects. This analogy was further justified on O 1s XPS spectra, which showed higher density of OVs in BHOE compared to undoped samples due to distribution of europium ions at both Ba2+ and Hf4+ sites. PALS also suggested that OVs are more closely associated with Hf4+ and Eu3+ due to a high charge/radius ratio.

Defect formation energy calculation along with PL lifetimes showed that a large fraction of europium ions stabilize at the Ba2+ site with excited state lifetime of 3.84 ms, whereas one-fourth occupies the Hf4+ site with lifetime value of 1.42 ms. TRES showed a very intense MDT for Eu@Ba with a predominant orange emission and relatively strong EDT with appearance of 5D0 → 7F0 peaks for the Eu@Hf site with a predominant red emission. Stark splitting analysis concluded that the overall symmetry of europium ions is lowered significantly to C4 due to strain, distortion, and formation of charge-compensating defects. PALS also suggested that OVs is one of the most probable defects responsible for the lowering of symmetry around europium despite a majority of it occupying the Ba2+ site. The efficient host–to-europium energy transfer was validated by the creation of an impurity state below the CBM, mainly by the Eu (f) state for both Eu@Ba and Eu@Hf sites. BHOE under X-ray excitation demonstrated an orange-red emission endowed by the Eu@Hf site, though the intensity was lower compared to PL endowed by lower density of X-ray excitable Eu@Hf centrers. TL measurements under beta irradiation show the appearance of shallow and deep traps, respectively, in BHO and BHOE. BHOE can thus have potential for further exploration in the area of afterglow phosphors.