Huge enhancement of upconversion photoluminescence in In³⁺ doped Ba_0.85−xCa_0.15TiO₃:0.75%Er³⁺/xIn³⁺ lead-free ferroelectric ceramics

Abstract

Enhanced green and red upconversion emissions under 980 nm excitation are observed in Ba_0.85−xCa_0.15TiO₃:0.75%Er³⁺/xIn³⁺ lead-free ferroelectric ceramics which were synthesized by solid state reaction method. With increasing In³⁺ ions doping level from x = 0% mol the green and red emission intensities increase and reach the maximum at x = 1.75% mol. The maximum intensity of green emission at 550 nm is 20 times larger than In³⁺ ions undoped samples, which is unprecedented for Er³⁺-doped perovskite ferroelectric materials. Based on the ultraviolet-visible-near infrared absorption spectra, the enhancement of the luminescence intensity of the A-site In³⁺ doped Ba_0.85−xCa_0.15TiO₃:0.75%Er/xIn ceramics is attributed to In³⁺-doping-induced relative increase in isolate Er³⁺ and decrease in the clustered-Er³⁺. The effects of In³⁺ doping on the crystal structure, dielectric, ferroelectric, piezoelectric properties of the ceramics were also studied and discussed.

1 Introduction

Upconversion luminescence is an anti-Stokes type emission, in which the wavelength of emission light is shorter than the excitation wavelength, due to the sequential absorption of two or more photons [1]. Rare-earth ions doped perovskite ferroelectric materials, owing excellent ferroelectric, piezoelectric and luminescence properties, are currently attracting extensive interest and have potential applications in the field of sensors, optoelectronic and display devices etc. [2,3,4]. Improving luminescence efficiency of rare-earth doped materials is an important precondition for practical device applications, but there are still few studies on upconversion luminescence efficiency of perovskite ferroelectrics until now.

Er³⁺ ions are commonly used as emission center in upconversion luminescent materials because of its special energy level which has many fine splitting and metastable energy levels. The 4f electrons of Er³⁺ ions enable easy excitation in a wide photoluminescence emission range from visible to mid-infrared. Er³⁺ doped ferroelectrics can be excited by both 980 and 1550 nm infrared and show green emission at 530, 550 nm and red emission at 670 nm in visible light range [5,6,7]. Many studies have been carried out on the upconversion luminescence of different Er³⁺-doped materials, especially perovskite-type ferroelectrics, such as Er³⁺-doped PbZr_0.52Ti_0.48O₃, Pb(Mg_1/3Nb_2/3)O₃–PbTiO₃–Pb(Yb_1/2Nb_1/2)O₃, (K_0.48Na_0.48Li_0.04)(Nb_0.96Bi_0.04)O₃, LiNbO₃, BaTiO₃, Ba_0.95Sr_0.05Zr_0.1Ti_0.9O₃, and Na_0.5Bi_0.5TiO₃ etc. [8,9,10,11,12,13,14]. And the enhancement effect of Er³⁺/Yb³⁺ co-doped and Tm³⁺/Yb³⁺ co-doped in transparent phosphate glass–ceramics have been studied [15, 16]. Although there are many studies on electrical properties and luminescence properties of Er³⁺ doped ferroelectrics, only a low luminescence efficiency could be obtained. It is necessary to further improve luminescence efficiency of Er³⁺-doped perovskite ferroelectrics.

Recently, Li et al., Noto et al. and Qian et al. reported that the In³⁺ doping can improve the emission intensities of Ga₂O₃:Cr³⁺, CaTiO₃:Pr³⁺ and LiNbO₃:Er³⁺/Yb³⁺, respectively [5, 17, 18]. Zhang et al. and Han et al. found that the emission intensities of Y_2.94(Al_1−x−yGa_xIn_y)₅O₁₂:Er³⁺ (x + y = 0.15) and NaY_0.95−xYb_0.03Er_0.02F₄:In³⁺ increase by In³⁺ doping [19, 20]. Our previous studies also demonstrated that In³⁺-ions-doped perovskite-type Ba_0.85Ca_0.15TiO₃:0.75%Er/xIn achieved 10 times increase in green band emission intensity prominently [6, 7]. To further improve the upconversion photoluminescent properties of Ba_0.85Ca_0.15TiO₃:0.75%Er, we prepared a series of A-site In³⁺-ions-doped Ba_0.85−xCa_0.15TiO₃:0.75%Er³⁺ (B_1−xCT:0.75%Er) ceramics and studied the ferroelectric, piezoelectric, dielectric, structural and upconversion photoluminescent properties. We found that the green upconversion emission of B_1−xCT:0.75%Er/xIn could be dramatically enhanced by 20 times through nominal A-site In³⁺ doping. Ultraviolet-visible-near infrared absorption spectra reveals that the In³⁺ doping increases the relative volume fraction of isolate In³⁺ ions, which favors the enhancement of green emission.

2 Experiment

The Ba_0.85−xCa_0.15TiO₃:0.75%Er³⁺/xIn³⁺ (0 ≤ x ≤ 2%) ceramics were prepared by the conventional solid-state method using the high purity raw materials of BaCO₃ (99.8%), CaCO₃ (99.5%), TiO₂ (99.9%), Er₂O₃ (99.9%) and In₂O₃ (99.99%). These weighed powders were ball mixed in the alcohol medium for 24 h, then dried at 100 °C and calcined at 1200 °C for 4 h. The weight of the raw material and the weight of the pre-sintered precursor are showed in Table 1. The powders were ball milled again for another 24 h, and pressed into pellets with 10 mm in diameter and 1 mm in thickness. Samples were sintered at 1350 °C in the air for 3 h. For electrical measurements, top and bottom surfaces of the samples were coated with sliver paste and sintered at 600 °C for 2 h.

Table 1 The weight of the raw material and the weight of the pre-sintered precursor

X-ray diffraction with a Cu K_α radiation (Bruker D8 Advance) was used to characterize the phase structure of the samples, at a 40 kV tube voltage and a 40 mA tube current. Microstructures of the samples were characterized by a field emission scanning electronics microscope (FE-SEM, S-4800, HITACHI). Dielectric properties were measured between − 200 °C-150 °C using an Agilent E4980A Precision LCR analyzer at 100 kHz. Ferroelectric hysteresis loops were measured at room temperature using a Precision Multiferroic test system (Radiant Technologies, Inc. USA). The piezoelectric coefficient d₃₃ was measured using a quasi-static d₃₃ meter (ZJ-4AN). A spectrofluorometer (Edinburgh Instruments FS5, UK) was used to measure the upconversion emission spectra of the ceramics samples, and a laser diode operating at a fixed pump power of 49 mW and a beam waist of 1 mm was used to provide 980 nm laser source. The absorption spectra were measured by an ultraviolet-vis-near infrared (UV-Vis-NIR) absorption spectrophotometer (Cary 5000).

3 Results and discussion

The phase structures of the samples were studied by analyzing their XRD patterns and temperature-dependent dielectric constant (ε_r versus T) curves. Figure 1a displays XRD patterns of B_1−xCT:0.75%Er/xIn samples. All diffraction peaks can be indexed according to those of BaTiO₃. PDF#79-2264 and PDF#81-2200 are standard XRD patterns of tetragonal phase (T) and orthorhombic phase (O) of BaTiO₃, respectively. All samples with various In³⁺ content show perovskite structure and no impurity phase can be detected. The phases of two diffraction peaks were unchanged with different In³⁺ content in the 2θ range from 45°–46° (Fig. 1b). All samples are pure tetragonal phase at room temperature. Characteristic peaks strength was changed slightly, which represents the degradation of crystallization under high doped. With the doping of In³⁺ ions, the grain size of the samples did not change significantly, according to the SEM images (Fig. 1c). Figure 2a shows the curves of dielectric properties (ε_r and tan δ) against temperature. In the measured temperature region from − 200 to 150 °C a strong dielectric peak and three loss tan δ peaks appear. The dielectric peak corresponds to the cubic-to-tetragonal phase transition, and three loss tan δ peaks correspond to the cubic-to-tetragonal, tetragonal-to-orthorhombic, and orthorhombic-to-rhombohedral phase transitions, respectively [6, 7, 21]. With increasing x from 0 to 2%, the Curie Temperature T_C decreases linearly from 138 to 118 °C, similar to previously reported results [6, 7]. The reason for the decrease of T_C may probably arise from compositional and structural disorder due to the difference in the valences and ionic radii between Ba²⁺ and In³⁺. Moreover, the possible replacement of Ba²⁺/Ca²⁺ ions by Er³⁺ ions would lead to oxygen and titanium vacancies, which could also contribute to the decrease of T_C [22,23,24,25]. According to the dielectric and loss tanδ against the temperature data, we establish the phase diagram of the B_1−xCT:0.75%Er/xIn, as shown in Fig. 2b. It is noted that the tetragonal-to-orthorhombic transition temperature (T_O−T) is relativly insensitive to the In³⁺ content for 0 to 2%. All samples exhibit tetragonal phase at room temperature, which is consistent with the XRD results.

Fig. 1

a XRD pattern of the B_1−xCT:0.75%Er/xIn (0 ≤ x ≤ 2%) ceramics. b Enlarge XRD patterns from 45° to 46°. c SEM images of the x = 0%, 1%, and 2% samples

Fig. 2

a Polarization–electric field hysteresis loops of the B_1−xCT:Er/xIn (0 ≤ x ≤ 2%) ceramics. b, c 2P_r and d₃₃ as a function of In³⁺ content for the B_1−xCT:Er/xIn (0 ≤ x ≤ 2%) ceramics. d Temperature dependence of the dielectric constant for the B_1−xCT:Er/xIn (0 ≤ x ≤ 2%) ceramics. Inset is the dielectric loss against temperature for the samples. e Phase diagram of the B_1−xCT:Er/xIn (0 ≤ x ≤ 2%) ceramics

Figure 2c shows the polarization–electric (P-E) hysteresis loops of the B_1−xCT:0.75%Er/xIn (0 ≤ x ≤ 2%) ceramics whose remanent polarization (2P_r) are shown in Fig. 2d. With increasing In³⁺ content from x = 0%, 2P_r increases slightly and reaches a maximum of 23.5 µC/cm² at x = 0.5%. With further increase in the In³⁺ content 2P_r decreases rapidly to a minimum of 11.9 µC/cm² at x = 2.0%. As mentioned above, the replacement of A-site Ba²⁺ by In³⁺ would result in charge balance broken, formation of titanium and oxygen vacancies [22, 23]. It has been reported that oxygen vacancies induce defect dipoles [26, 27] which would suppress ferroelectric domain switching due to the pinned effect, resulting in an enhancement of the remnant polarization (2P_r) in the low x range 0%≤x≤0.5% [6, 28, 29]. However, a further increase in the In³⁺ content leads to a decrease in the remnant polarization. Excessive In³⁺ doping induced more oxygen vacancies cause the unit cell serious distortion and dissymmetry, thus ferroelectric domains breaking and leakage current take place [28, 30]. Additionally, the piezoelectric coefficient d₃₃ continuous decreases from ~ 130 to ~ 50 pC/N (Fig. 2e), with increasing In³⁺ content from 0 to 2%. All previously discussed effects including increasing In³⁺ ions, oxygen vacancies, pinning effect, leakage current and ferroelectric domains destruction, are negative factors of piezoelectric properties [31, 32]. These reasonable explanations are consistent with the experimentally observed variation trend of 2P_r and d₃₃.

Figure 3a shows the upconversion luminescence spectra of the B_1−xCT:Er/xIn using 980 nm laser excitation at room temperature. Each spectrum shows three emission bands, two strong green emission bands at 530 and 550 nm, and one weak red emission band at 670 nm, corresponding to the ²H_11/2/⁴S_3/2 → ⁴I_15/2 (green) and ⁴F_9/2 → ⁴I_15/2 (red) transitions of Er³⁺ ions, respectively. The location and intensity of the peaks are consistent with previous reports [6, 7, 33]. The emission intensity of all the three peaks are plotted against the In³⁺ content in the inset of Fig. 3a, where one can find that the maximum emission intensity of each peak appears at x = 1.75%. Notably, the maximum emission intensity at 550 nm is approximately 20 times larger than that of In³⁺ undoped samples. For the 530 and 670 nm emission peaks, the In³⁺ doping enhances the emission intensity by 16 and 20 times, respectively, as compared with those of In³⁺ undoped one. The emission intensities of all peaks (530, 550, 670 nm) reach the maximum at x = 1.75% and then decrease with further increases in the In³⁺ content. In³⁺-doping-induced changes in the energy level could improve the luminence intensity within a certain In³⁺ content, heavy In³⁺ doping would induce quenching effect and result in quick degeneration of emission intensity. As discussed in the above section, the In³⁺ doping induces oxygen vacancies which would induce defect levels, which act as traps for electrons populated in the conduction band and suppress energy transition to the excited levels of Er³⁺ ions [19]. In this work, 20 times enhancement in the emission intensity was unprecedented, comparing with previous reports [6, 7].

Fig. 3

a The upconversion emission spectra of the B_1−xCT:Er/xIn (0 ≤ x ≤ 2%) ceramics. Inset is the emission intensity at 530, 550 and 670 nm as a function of the In³⁺ content. b The emission spectra of the B_1−0.0175CT:Er/1.75%In ceramics under the excitation of different pump power. Inset is the Ln–Ln plot of the emission intensity versus pump power. c Simplified energy level diagram for the Er³⁺ ions upconversion mechanisms. d The absorption spectra of the B_1−xCT:Er/xIn (0 ≤ x ≤ 2%) ceramics. Inset is the emission intensity at 550 nm and the absorption cross section of the ²H_11/2 level as a function of the In³⁺ content

To further investigate the effect of In³⁺ doping on the luminescence mechanism of B_1−xCT:Er/xIn ceramics, we plot the emission spectra under different excitation pump power in Fig. 3b and measure the emission intensity (I) as a function of pump power (P) for the sample of x = 1.75%. As is reported in previous research works, [34, 35] the number of photons (n) required to populate the upper emitting state can be described by the relationship I ∝ Pⁿ. The estimated slopes (n) from the linear fits of ln(I) versus ln(P) are 2.03, 1.97 and 1.86, corresponding to the emission bands for 530, 550 and 670 nm, respectively. It indicates that the two-photon excitation mechanism dominates the upconversion luminescence of Er³⁺ ions in In³⁺ doped BCT:0.75Er% ceramics.

Figure 3c shows the simplified energy level diagram of the upconversion luminescence under 980 nm excitation [6, 7, 30, 33, 36]. According to the schematic diagram, 4f electrons of Er³⁺ ions are excited from the ground state ⁴I_15/2 to long-lived excited state ⁴I_11/2 through the ground state absorption (GSA) by the 980 nm excitation. Under sustaining excitation, the electrons at ⁴I_11/2 level absorb the second photon energy and transition to higher energy levels of ⁴F_7/2 through excited state absorption (ESA). Consequently, the electrons at ⁴F_7/2 level relax to ²H_11/2, ⁴S_3/2 and ⁴F_9/2 levels without radiation, and then radiative transition to the ground state, ²H_11/2/⁴S_3/2 → ⁴I_15/2 (green) and ⁴F_9/2 → ⁴I_15/2 (red), occurs. In addition, the multiphonon relaxation (MPR) during which electrons from excited levels jump into low level without radiation results in the decrease of photons absorption efficiency. Furthermore, the cross relaxation (CR) between two adjacent Er³⁺ ions is associated with the In³⁺ content and will be discussed in following sections.

The UV-VIS-NIR absorption spectra for selected samples in the wavelength ranging from 400 to 1200 nm are plotted in Fig. 3d. The absorption spectra have five absorption peaks at 488, 522, 654, 792, and 975 nm, corresponding to the transitions of Er³⁺ ions from the ground state ⁴I_15/2 to the ⁴F_7/2, ²H_11/2, ⁴F_9/2, ⁴I_9/2 and ⁴I_11/2 excited states, respectively. Upon the introduction of In³⁺ ions the intensities of these absorption peaks are remarkably enhanced. The absorption cross-section (σ_abs) which is an indicator to evaluate the distribution of Er³⁺ ions are calculated using the equation $\int_{{{\text{band}}}} {\sigma (\lambda )d\lambda = \sum\nolimits_{{{\text{band}}}} {{{\alpha (\lambda )} \mathord{\left/ {\vphantom {{\alpha (\lambda )} {Cd\lambda }}} \right. \kern-\nulldelimiterspace} {Cd\lambda }}} }$ [7, 37], where λ, α, and C are the absorption wavelength, absorption coefficient, and concentration of Er³⁺ ions, respectively. To reveal the intrinsic correlation between σ_abs and emission intensity, we plot σ_abs of ²H_11/2 and the emission intensity of 550 nm band against the content In³⁺ content in the inset of Fig. 3d. The variation of the emission intensity with the In³⁺ content shows similar trend as that of σ_abs. Both σ_abs and emission intensity reach the maximum values at x = 1.75%. According to previous reports [5, 7, 38], the doped Er³⁺ ions have two different distribution sites, i.e., Er³⁺ isolated sites and cluster sites and In³⁺ can modify the distribution of Er³⁺ sites according to the Judd–Ofelt (J–O) theory [39, 40]. Since σ_abs of Er³⁺ isolated sites is larger than that of Er³⁺ cluster sites [41], it thus can be concluded that the isolated sites are minimum at x = 0 and reaches the maximum at x = 1.75%. Furthermore, considering the effect between the adjacent Er³⁺ ions, the CR processes more possible occur among Er³⁺ cluster sites [5, 42], $({\text{i.e., CR1: }}{}^{4}{\text{S}}_{3/2} \to {}^{4}{\text{I}}_{9/2} + {}^{4}{\text{I}}_{15/2} \to {}^{4}{\text{I}}_{13/2} {\text{ and CR2: }}{}^{4}{\text{S}}_{3/2} \to {}^{4}{\text{I}}_{13/2} + {}^{4}{\text{I}}_{15/2} \to {}^{4}{\text{I}}_{11/2} )$ due to more small distance of Er³⁺ ions. The CR processes is no radiation relaxation process, the excited electrons decay to the low level without emission photons, influencing the efficiency of electrons radiative transition. In addition, the inherent ESA process happens among Er³⁺ isolated sites, which is the necessary condition for enhancing emission. Therefore, more Er³⁺ isolated sites generate, leading to more efficient ESA process and scarce CR process and the increase of both σ_abs and emission intensity, with the content of 0 ≤ x ≤ 1.75%. The x = 1.75% sample with the largest σ_abs and emission intensity should possess the maximal isolated Er³⁺ and the minimal cluster Er³⁺ sites. Notably, for x > 1.75%, the reason for the decrease of σ_abs and emission intensity may be due to the excess In³⁺ doping-induced formation of impurity level which have significant impact on the Er³⁺ energy level, reducing the efficiency of energy levels transition. According to the above analysis, all these interpretations consistent with the experimental results that In³⁺ can modify the Er³⁺ -sites distribution and Er³⁺ isolated sites, which is beneficial to the enhancement of upconversion luminescence intensity of B_1−xCT:0.75%Er ceramics.

4 Conclusions

In³⁺-doped Ba_0.85−xCa_0.15TiO₃:0.75%Er³⁺/xIn³⁺ upconversion photoluminescent ferroelectric ceramics have been prepared by the conventional solid-state method and the effects of In³⁺ content on the upconversion photoluminescent properties have been studied. Upon 980 nm excitation the 550 nm band upconversion emission intensity increases dramatically with increasing In³⁺ content and reaches the maximum at x = 1.75% mol, where the green emission intensity is unprecedentedly (~ 2000%) larger than that of the In³⁺ undoped Ba_0.85Ca_0.15TiO₃:0.75%Er³⁺ ones. UV-VIS-NIR absorption spectra reveal that the In³⁺ doping increases Er³⁺ isolated sites and reduces Er³⁺ cluster sites, thereby favoring the excited state absorption process and reduces cross relaxation. As a result, the upconversion efficiency is greatly enhanced. Our results demonstrate that the nominal A-site In³⁺ doping is an effective approach to enhance the upconversion green emission efficiency of Er³⁺-doped perovskite ferroelectrics.