A. XRD phase

The powder XRD patterns of LiBaPO₄:0.01Eu²⁺ are shown in Fig. 1. All of the peaks can be indexed based on hexagonal structure and match well with the PDF#14-0270, indicating that the obtained samples are of single phase and the doped Eu²⁺ ions do not cause any significant change in the host crystal structure. The Eu²⁺ ions are supposed to reside on Ba²⁺ sites, because they have similar ionic radii and valences.

FIG. 1.

The XRD pattern of LiBaPO₄:0.01Eu²⁺ phosphor (upper) and that of PDF#14-0270 is given for comparison (bottom).

B. Reflectance

Figure 2(a) presents the diffuse reflectance spectra of LiBaPO₄ and LiBaPO₄:Eu³⁺. The spectrum of LiBaPO₄ shows a platform of high reflectance in the wave length range of 380–800 nm and then starts to decrease dramatically from 380 to 200 nm. This band is due to the host absorption. The body color of LiBaPO₄ is white, which agrees well with the spectrum. The near edge relation between absorption coefficient (α) and band gap E _g for a direct band gap material can be described by the following equation ^{Reference Omar27–Reference Amirtharaj, Seiler and Bass31} :

(1) $${\left( {{\rm{\alpha }}h{\rm{\nu }}} \right)^2} = A\left( {h{\rm{\nu }} - {E_{\rm{g}}}} \right)\quad ,$$

where A is a constant involving the properties of the bands and hν is the photon energy. According to Kubelka–Munk theory, absorption coefficient α in Eq. (1) can be substituted with ^{Reference Simmons32–Reference Nobbs34}

(2) $$F\left( {{R_\infty }} \right) = {{{{\left( {1 - {R_\infty }} \right)}^2}} \over {2{R_\infty }}}\quad ,$$

in which R _∞ is the reflectance of an infinitely thick material. It can be substituted with experimentally observed reflectance, because the transmittance of our samples is effectively zero. Figure 2(b) depicts the Tauc plot of LiBaPO₄ that is widely used for the determination of band gap. A line is drawn in the near edge region tangent to the point of inflection on the curve, and the hν value at the point of intersection of the tangent line and the horizontal axis gives the band gap E _g = 5.29 eV.

FIG. 2.

(a) Diffuse reflectance spectra of LiBaPO₄ and LiBaPO₄:Eu³⁺; (b) determination of the optical band gap (E _g) using Tauc plot for LiBaPO₄.

Dorenbos argued that the charge transfer (CT) energy of Eu³⁺ is a direct measure for the energy difference between the top of the VB and the 4f ground state of the Eu²⁺. ^{Reference Dorenbos35,Reference Dorenbos36} The energy of Eu²⁺ excited states as well as the ground state with respect to the band states of the host is crucial to understand the trapping and detrapping processes in a persistent phosphor. Therefore, we synthesized LiBaPO₄:Eu³⁺ and measured its diffuse reflectance spectrum. When Eu³⁺ ions are introduced into the host, new lines and band occur. The lines are ascribed to the intraconfigurational 4f–4f transitions of Eu³⁺ in the host: from ⁷F₀ to ⁵L₆ and ⁵D₂ at wave lengths 393 and 465 nm, respectively. ^{Reference Carnall, Fields and Rajnak37} The broad absorption band centered at ∼300 nm can be attributed to CT band arising from the top of VB to Eu³⁺ ion. The CT band overlaps the host fundamental absorption band; therefore, the absorption band in the 200–350 nm area changed. Our UV-visible spectrophotometer (Shimadzu UV-2450) was equipped with only one monochromator, and the emission of Eu³⁺ under CT excitation results in a false reflectance band at ∼320 nm. The 4f–4f transitions of Eu³⁺ are both parity and spin forbidden; therefore, these absorption/emission lines are usually weak. The diffuse reflectance spectrum of LiBaPO₄:Eu²⁺ is not achievable due to its efficient luminescence.

C. Photoluminescence

The steady-state PL and PLE spectra for the Eu²⁺-activated LiBaPO₄ are shown in Fig. 3(a). An intense broad band centered at ∼470 nm and can be attributed to the electric-dipole-allowed 4f ⁶5d ¹ → 4f ⁷ (⁸S_7/2) transition of Eu²⁺ was observed at ambient temperature. The emission band has a full width at half maximum (FWHM) of about 3400 cm⁻¹ that is larger than the characteristic value for Eu²⁺ emission in this wavelength region. This suggests that there is more than one site for the Eu²⁺ ions in the host lattice, which agrees with the crystal data. Monitored by the emission at 470 nm, the PLE spectrum exhibits an unresolved broad band in the range from 250 to 450 nm, which can be assigned to the transitions from the ground state ⁸S_7/2(4f ⁷) to the crystal-field-split components of the 4f ⁶5d ¹ configuration. The Stokes shift of LiBaPO₄:Eu²⁺ was estimated to be 6000 cm⁻¹. This is much larger than the normal value of 1000–4000 cm⁻¹ reported by Dorenbos. ^{Reference Dorenbos38} The cyan emission of Eu²⁺ in the host is mainly due to the large relaxation after the absorption of pump light.

FIG. 3.

(a) The PLE and PL spectra of LiBaPO₄:Eu²⁺, (b) PL spectra of LiBa_1−x PO₄:xEu²⁺, (c) the PLE and PL spectra of LiBaPO₄:Eu³⁺, and (d) persistent luminescence spectra of LiBaPO₄:Eu²⁺.

The variation of emission intensities of LiBa_1−x PO₄:xEu²⁺ phosphors with Eu²⁺ concentration (x) is shown in Fig. 3(b). It is obviously seen that the PL features of Eu²⁺ on the peak shape and positions remain unchanged as the concentration of Eu²⁺ ions increases. The concentration quenching of activators requires considerable energy overlap between PLE and PL spectra as well as the concentration of activators. We did not observe concentration quenching in the experimental range.

The PLE and PL spectra of LiBaPO₄:Eu³⁺ are shown in Fig. 3(c). The PLE spectrum by monitoring the ⁵D₀ → ⁷F₂ emission (617 nm) of Eu³⁺ ions consists of a broad band and some sharp lines in the UV and visible region. In the visible range, one observes well-separated excitation lines due to transitions from the ground state ⁷F₀ to the levels ⁵D_J, ⁵L_J, etc. The broad excitation band centered at 240 nm corresponds to a broad 2p → 4f CT band, which gives rise to allowed electric dipole transitions. The CT band is weaker than the intraconfigurational 4f–4f transitions, even though the 4f–4f transitions are both spin and parity forbidden, which indicates that most of the CT energy lost in the host lattice is due to nonradiative transitions. Upon the excitation with 394 nm, the PL spectra of LiBaPO₄:Eu³⁺ display ⁵D₀ → ⁷F_J (J = 0–2) emission lines of the Eu³⁺ ions. Because of efficient multiphonon relaxation in the phosphors, no emission from higher Eu³⁺ excited levels was detected. The bold lines imply that the spectrum has complex structure at high resolution because of the variety of Eu³⁺ sites and low site symmetry.

D. Persistent luminescence

After removal of the excitation light of 254 nm (low-pressure mercury discharge lamp) at room temperature, bright persistent luminescence was observed in LiBaPO₄:Eu²⁺. The spectral shape [Fig. 3(d)] of the persistent luminescence is in agreement with that of the PL spectrum of LiBaPO₄:Eu²⁺. Upon 365 nm (high-pressure mercury discharge lamp) excitation, LiBaPO₄:Eu²⁺ fluoresce efficiently; however, after ceasing the excitation, only weak persistent luminescence was observed.

Figure 4 shows the afterglow intensity as a function of time for the different concentrations of activator after 1 min excitation (254 nm). A concentration of x = 0.01 doped Eu²⁺ ions gives the brightest afterglow, although we did not observe concentration quenching for fluorescence even for x = 0.07 [Fig. 3(b)]. For most persistent phosphors, the concentration dependence of persistent luminescence is different from photoluminescence in view of their different mechanisms. The fluorescence quenching is due to the energy transfer between activators, which requires considerable interaction and energy resonance between activators. ^{Reference Blasse and Grabmaier39} On the other hand, persistent luminescence quenching results from electron transfer between traps/activators by tunneling, which requires electron cloud overlap and also energy resonance. ^{Reference Ju, Hu, Chen, Wang and Mu40} The width of potential barrier will be reduced at high Eu²⁺ concentrations. Therefore, the probability of tunneling increases and the trapped electrons can easily reach activators even at room temperature. The situation could be worse for shallow traps.

FIG. 4.

(a) The persistent luminescence isothermal decay curves of LiBa_1−x PO₄:xEu²⁺ and (b) fitting results of LiBaPO₄:0.01Eu²⁺ by I(t) = I ₀ + A exp(−t/τ₁) + B exp(−t/τ₂).

For many Eu²⁺-activated persistent luminescent materials, the afterglow intensity and duration can be improved by codoping certain rare earth elements. In particular, Dy³⁺ and Nd³⁺ are often chosen for this motive to enhance the afterglow of aluminates and silicates. This enhancement is attributed to creation or modification of traps in the material by the codopant. ^{Reference Ju, Hu, Chen, Wang and Mu26} For LiBaPO₄:Eu²⁺, different rare earth elements (Y³⁺, La³⁺, Ce³⁺, Gd³⁺, Tb³⁺, Dy³⁺, and Lu³⁺) were chosen as codopant but they hardly improve the afterglow. Instead, the La³⁺or Ce³⁺ codoped samples quenched the afterglow of LiBaPO₄:Eu²⁺.

In general, a decay process of a persistent phosphor contains a short-exponential decaying process and a long-exponential decaying one. ^{Reference Sakai, Katsumata, Komuro and Morikawa41} The long-decaying process determines the duration of the afterglow. The decay curves of LiBaPO₄:Eu²⁺ phosphor are well fitted with the double simple exponential function:

(3) $$I(t) = {I_0} + A\exp \left( { - t/{{\rm{\tau }}_1}} \right) + B\exp \left( { - t/{{\rm{\tau }}_2}} \right)\quad ,$$

where I(t) and I ₀ are the persistent luminescence intensities at times t and background, respectively; A and B are constants; and τ₁ and τ₂ are the decay times for the short and long decay components, respectively. The decay time values were determined to be 19.5 and 124.5 s, respectively.

E. Thermoluminescence

The persistent luminescence phenomenon is a special case of thermally stimulated luminescence at room temperature. ^{Reference Rodrigues, Brito, Hölsä and Lastusaari7} Thermoluminescence measurements provide important information on the nature of the traps in the persistent phosphors. ^{Reference McKeever4} The glow curve of LiBaPO₄:Eu²⁺ is shown in Fig. 5. A strong thermoluminescence band with maximum at 65 °C characterizes the glow curve, along with a much weaker band at high temperature. The trap depth E _T of the strong glow band can be calculated by the Urbach method, ^{Reference McKeever4,Reference Shalgaonkar and Narlikar42}

(4) $${E_{\rm{T}}} = {T_{\rm{m}}}{\rm{/500}}\quad {\rm{,}}$$

where T _m is the peak temperature (in Kelvin). The estimated trap depth for the glow peak is 0.68 eV. According to Ref. Reference Matsuzawa, Aoki, Takeuchi and Murayama12, the optimum trap depth is in the range of 0.4–0.7 eV, which is suitable for the thermal release rate at room temperature. Apparently, the strong glow peak at ∼65 °C is mainly responsible for the persistent luminescence at room temperature.

FIG. 5.

Thermoluminescence curve of LiBaPO₄:Eu^{2 +}.

F. The energy level scheme of LiBaPO₄:Eu²⁺

The energy of Eu²⁺ excited/ground states with respect to the band states of the host is crucial to understand the trapping and detrapping processes in a Eu²⁺-activated persistent phosphor. The ground level energy for the Eu²⁺ ion in the host band structure of LiBaPO₄ is difficult to obtain. The CT method ^{Reference Dorenbos35,Reference Dorenbos36} is not straightforward as it appears. Actually, if Eu³⁺ ion replaces a divalent ion (e.g., Ba²⁺ in the present case), Eu²⁺ and Eu³⁺ reside in different coordination polyhedra due to the charge-compensating defects. ^{Reference Ju, Hu, Chen, Wang and Hung24} The energy of the end state of CT is higher than the ground state of Eu²⁺ resulting from the repulsion of the charge-compensating defects. Indeed, the CT energy is usually close to the bottom of CB when Eu³⁺ replaces a divalent ion. The excited level energy for the Eu²⁺ ion in the host band structure of LiBaPO₄ is also difficult to obtain. Dorenbos suggested that the thermal quenching in Eu²⁺-activated compounds is due to the ionization of Eu²⁺, which is caused by the relatively small energy separation between the 5d levels of Eu²⁺and the bottom of the CB. Although some authors agree with this assumption, ^{Reference Smet, Eeckhout, Bos, Kolk and Dorenbos44} thermally assisted cross-over in the configuration coordinate diagram cannot be expelled yet, especially in cases in which large Stokes shift were observed, as in LiBaPO₄:Eu²⁺.

The response of our phosphors to 254 and 365 nm UV implies that pumping to higher excited states of Eu²⁺ redound to trap filling, so the lower 4f ⁶5d ¹ states are placed below the bottom of CB while the higher 4f ⁶5¹ states are placed in the CB. Once the lowest excited state is determined, the ground level can be estimated by analyzing the PLE and PL spectra. The interaction between the 5d electron and the 4f ⁶ core leads to a staircase energy structure on the low-energy side with intervals close to those of the multiplet ⁷F_J (J = 0–6) of Eu³⁺. ^{Reference Haecke, Smet and Poelman45} Using the intersection point (423 nm) of the PLE and PL curves, the lowest excited state of the ⁷F_J multiplet of the 4f ⁶ configuration (zero-phonon line) was determined to be 2.93 eV. Based on the considerations presented above, a rudimentary energy level scheme was constructed in Fig. 6.

FIG. 6.

Energy level scheme for luminescence in LiBaPO₄:Eu²⁺, T and E _F refer to traps and Fermi level, respectively. Arrows indicate electron transfer or transition.

With the scheme in Fig. 6 and the mechanism proposed by Hölsä et al., ^{Reference Brito, Hölsä, Laamanen, Lastusaari, Malkamäki and Rodrigues2,Reference Rodrigues, Brito, Hölsä and Lastusaari7,Reference Rodrigues, Brito, Hölsä, Stefani, Felinto, Lastusaari, Laamanen and Nunes15,Reference Aitasalo, Hölsä, Jungner, Lastusaari and Niittykoski46} the trapping and detrapping processes in the material can be depicted. Under the irradiation by UV light, Eu²⁺ is pumped to its 4f ⁶5d ¹ states, which are partly inside the CB. LiBaPO₄:Eu²⁺ is an efficient phosphor, so that the excited Eu²⁺ will mostly relax to its lowest 4f ⁶5d ¹ state and then radiatively back to its ground level. In addition to photoluminescence, small amounts of Eu²⁺ ions are ionized. These create free electrons in the CB and Eu³⁺ ions are left behind. Different research groups have recently confirmed this kind of valence-state change of europium. ^{Reference Carlson, Hölsä, Laamanen, Lastusaari, Malkamäki, Niittykoski and Valtonen47–Reference Korthout, Eeckhout, Botterman, Nikitenko, Poelman and Smet50} The delocalized electrons are trapped by positive lattice defects from the CB. The thermal energy at ambient temperature causes the detrapping of the trapped electrons to the CB. The delocalized electrons can go back to a Eu³⁺ core and subsequently emission occurs. Tunneling recombination is also possible if defects are situated near the Eu³⁺ cores.