1 Introduction

Temperature measurements with high sensitivity and accuracy have become vital in many fields including daily life, industrial production, and scientific research [1,2,3]. In recent years, optical thermometry has attracted considerable interest owing to its noncontact characteristics, better accuracy, higher reliability, and broader applications as compared with traditional contact thermometers [4,5,6,7]. It can be used in harsh environments, such as those with high temperatures and pressure, and corrosive areas [8]. Among several optical thermometry methods, the fluorescence intensity ratio (FIR) based on the thermal coupled level transitions of rare-earth ions is believed to be the most promising and reliable for temperature measurements [9,10,11,12,13].

Upconversion luminescence materials codoped with Er3+ and Yb3+ ions have been studied [14,15,16]. As one of the most widely used activators, Er3+ ions have typical thermal coupled levels (TCLs) of 2H11/2 and 4S3/2, which relative emission intensity presents strong temperature dependence, and were applied to temperature sensing [17, 18]. By codoping Yb3+ ions as sensitizers, Er3+/Yb3+ doped materials can be effectively pumped by a commercial 980 nm laser diode and thus strengthen the intensity of upconversion luminescence [15]. This type of photoluminescence also has some other attractive advantages. In the previous published literatures [19], the absorption and scattering coefficients of human skin, subcutaneous adipose tissue and mucous in the wavelength range from 400 to 2000 nm have been studied. 980 nm have pretty low absorption and scattering coefficient, relative long tissue penetrate distance and non-biological toxicity, which make it possible to be used in biomedicine field. Some literature have reported the UC thermometer using FIR method for biomedical application [20,21,22]. For those broader application and attracted features, searching different optical temperature sensor materials can provide a more favorable opportunity to select different sensing functions according to the practical requirement.

The luminescence intensity and sensor sensitivity are essential for the practical application. Searching TCLs with larger energy gaps have been proposed for developing efficient methods with greater sensitivity [23]. In our previous work, the 2H11/2, 4S3/2, and 4F9/2 levels were shown to be thermal coupled multi-levels (TCMLs). The temperature sensitivity of FIR of 2H11/2/4F9/2 is approximately 4 times larger than that of 2H11/2/4S3/2 in Er3+/Yb3+ codoped PLZT ceramics [24]. To investigate the possibility of further enhancing the temperature sensitivity, we studied the thermal luminescence characteristics of Stark sublevels in the same sample. Stark sublevels can also be considered as TCLs for the small level gap. The emissions of small level gaps were studied as another valid strategy for improving sensitivity [25,26,27]. For instance, Liu reported the maximum absolute sensitivity of 2H11/2 to 4S3/2(1) levels as about onefold higher than that of traditionally used TCLs 2H11/2/4S3/2 in Er3+/Yb3+ codoped Ba3Y4O9 [28]. In addition, the sensitivity based on FIRs of 2H11/2 to 4S3/2(1) or 2H11/2 to 4S3/2(2) is almost twice enhanced in Er3+/Yb3+ codoped Gd2(WO4)3 [14]. As a traditional and widely studied material, PLZT transparent ceramic was selected as the host. The lower phonon energy (750 cm−1) and higher refraction index (2.401) of PLZT can result in higher transition probabilities than other rare-earth-doped host like LiNbO3 [29], and ceramics usually have better mechanical, thermal, and chemical resistances [29, 30].

In this study, the luminescence properties of the Stark sublevels of the 2H11/2, 4S3/2, and 4F9/2 levels in Er3+/Yb3+ codoped PLZT transparent ceramic were investigated based on the upconversion spectra under the 980 nm excitation, and the temperature-dependent emission intensities of the Stark sublevels were studied in the temperature range from 160 to 320 K. The FIRs of the Stark sublevels’ emissions were studied for the enhancement of temperature sensing sensitivity. Moreover, the results of this study are helpful for better understanding the temperature characteristics of the Stark sublevels’ relative intensities.

2 Experimental

The Er3+/Yb3+ codoped PLZT transparent ceramic was fabricated by the mixed-oxide method at Boston Applied Technologies Inc., USA. Then, a 3 mm × 3 mm × 2 mm sample was cut from the bulk ceramic and polished. This sample consists of 65 mol% lead zirconate plus 35 mol% lead titanate and 8 mol% lanthanum in the form of La2O3, i.e., PLZT(8/65/35), to which 0.5 mol% Er3+ ions in the form of Er2O3 and 2.5 mol% Yb3+ ions in the form of Yb2O3 had been added. The origins of components were PbO, ZrO2, La2O3, TiO2, Er2O3 and Yb2O3, respectively. PbO was purchased from Hammond Lead Products, with purity of 99.98%. All other oxides were purchased from Fisher Scientific have purity 99.85% or higher. A typical sintering process of the ceramic was published elsewhere [31]. By the inductively coupled plasma-atomic emission spectrometry (ICP-AES), the quantities of Er3+ and Yb3+ ions in the sample were determined to be 1.39 × 1020 ions/cm3 and 6.70 × 1020 ions/cm3, respectively. The upconversion spectrum and temperature fluorescence spectra were measured by a steady-state/lifetime spectrofluorometer (FLS 920, Edinburgh Instruments) excited by a 980 nm laser diode with emission slit size 0.5 nm, power 1000 mw, detection sensitivity (≤ 50 cps), integration time of 0.1 s and step of 0.5 nm, in here, the temperature of sample was real-time control by temperature control components of spectrofluorometer FLS 920 with 0.1 °C accuracy.

3 Results and discussion

The temperature-dependent upconversion spectra of Er3+/Yb3+: PLZT ceramic in the temperature range from 160 to 320 K were measured under 980 nm excitation (Fig. 1). The 2H11/2 and 4S3/2 levels are split into two sublevels 2H11/2(1) and 2H11/2(2) and 4S3/2(1) and 4S3/2(2), and the 4F9/2 level is split into three sublevels: 4F9/2(1), 4F9/2(2), and 4F9/2(3). Seven luminescence band peaks at 527 nm, 539 nm, 551 nm, 564 nm, 655 nm, 665 nm, and 680 nm were found according to the transitions from the 2H11/2(1), 2H11/2(2),4S3/2(1), 4S3/2(2), 4F9/2(1), 4F9/2(2), and 4F9/2(3) levels, respectively, to the ground level 4I15/2.

Fig. 1
figure 1

Temperature-dependent upconversion spectra of the Er3+/Yb3+: PLZT ceramic excited by 980 nm in the 160–320 K temperature range

Figure 2 shows the energy level scheme and relevant radiative transitions in this material. PLZT(Pb1-xLax(ZryTi1–y)1–xO3) have the perovskite structure with the formation of ABO3, in which La substitutes for Pb2+ in the A-site and the B-site vacancies are created for electric balance. When rare earth ion doped in PLZT, Re3+ ion (Er3+,Yb3+) will occupy the position of La3+ for their similar ionic radii and charge characters. Since in our sample, the concentration of Yb3+ is around five times larger than Er3+ ion and Yb3+ has a much larger absorption cross-section around 980 nm than that of Er3+ ion, which we have discussed in our previous work for the same sample, the 980 nm pumping is mainly absorbed by the Yb3+ ion [31]. Thus, the mechanism for the up-conversion emission can be explained as follows.

Fig. 2
figure 2

Schematic energy level diagram and transitions mechanisms of Er3+/Yb3+: PLZT ceramic

After absorbing the energy of incident light, the Yb3+ ions can be excited from the ground level 2F7/2 to upper levels 2F5/2. Due to the close energy level of 2F5/2 and 4I11/2, the energy transfer process effectively occur between Yb3+ and Er3+ ion, that is 2F5/2(Yb3+) + 4I15/2 (Er3+) → 2F7/2(Yb3+) + 4I11/2 (Er3+). Meantime, the Er3+ ion in its ground level can also absorb the pumping energy and excited to4I11/2 level. Then, the ions accumulate in Er3+:4I11/2 level absorb other photon energy and excited to upper 4F7/2 level. And Yb3+ ion can also absorb a second photon and transfer its energy to Er3+ ion by 2F5/2(Yb3+) + 4I11/2(Er3+) → 2F7/2(Yb3+) + 4F7/2(Er3+),which also contribute to the accumulation of 4F7/2 level. Ions in 4F7/2 level rapidly relax to the sublevels of 2H11/2 and 4S3/2 level, and then some of them radiatively relax to the ground level and generate green emission. The other ions relax to 4F9/2 level. Besides, the accumulation of 4F9/2 level also comes from the excited level 4I13/2 photon absorption and the energy transfer process: 2F5/2(Yb3+) + 4I13/2(Er3+) → 2F7/2 (Yb3+) + 4F9/2(Er3+). Finally, the ions in 4F9/2 sublevels decay to the ground level and generate red emission. The cross relaxation between Er3+ ions (4I11/2 + 4I11/2 → 4I15/2 + 4F7/2, 4I11/2 + 4I13/2 → 4I15/2 + 4F9/2) also help for the population accumulation of 4F7/2 and 4F9/2 level, which can’t be neglected in the sample with Er3+ concentration larger than 0.5 mol% [32]. But since those cross relaxation often obviously take place in the more heavily doped materials [33], and the Er3+ ion concentration is pretty low in our sample, those cross relaxation may be less due to the competitive mechanism.

Figure 1 also provides the temperature-dependent properties of the upconversion emissions from 160 to 320 K. The positions of the emission peaks barely changed with the increase of temperature, while the intensities changed remarkably. To further investigate the influence of temperature on the luminescence intensity of each transition, the relative integrated intensities normalized to the intensities at 160 K were calculated (Fig. 3). Here, the intensities of 2H11/2 (1) and 2H11/2 (2) were summed to be 539 nm to reduce calculation errors due to weaker luminescence. Figure 3 shows that all the transitions show strong temperature dependence. With the increase of temperature, the fluorescence integrated intensity of 539 nm evidently increased, while that of Is (the sum of intensities of 551 nm and 564 nm, 4S3/2(1, 2) → 4I15/2) and IF (the sum of intensities of 655 nm, 665 nm, and 680 nm, 4F9/2(1, 2, 3) → 4I15/2) uniformly decreased from 160 to 320 K, as shown in the inset of Fig. 3. This was because of thermal excitation between the TCLs [24].

Fig. 3
figure 3

Relative luminescence integrated intensities of the emissions of sublevels at different temperatures ranging from 160 to 320 K

In the green emission of Is, the intensities of 551 nm and 564 nm have the same decreasing trend with temperature, and the ions in the 4S3/2(1) and 4S3/2(2) levels were thermally excited to the 2H11/2 level simultaneously because of their similar small energy gap. However, in the red emission of IF, the intensity of 665 nm abnormally increases at 240 K. The abnormal increase can be attributed to the competitive effects of thermal excitation and nonradiative relaxation. The 4F9/2(2) level is laying in the gap of the 4F9/2(1) and 4F9/2(3) levels. With the increase of temperature, the population of the 4F9/2(2) is benefited by the nonradiative relaxation of the 4F9/2(1) level and thermal excitation of the 4F9/2(3) level as well as expended by the nonradiative relaxation to 4F9/2(3) and thermal excitation to 4F9/2(1) level. The 4F9/2(1) level with higher energy provides a thermal excitation barrier for transitions from the 4F9/2(2) to the 4S3/2 level. Therefore, the intensity of 665 nm increases from 160 to 240 K. With a further increase in temperature, the efficiency of thermal excitation increases, and the intensity of 665 nm decreases. Comparing the variation characteristics of intensities of 655, 665, and 680 nm from 240 to 320 K, thermal excitation was more dominant above 240 K, and their intensities show a similar decreasing trend with increasing temperature.

To elucidate the temperature sensing properties of TCMLs, the ratios between the emissions of the 4S3/2(1, 2) and 4F9/2(1, 2, 3) levels were examined (Fig. 4). The values of different ratios in Fig. 4 do not change above 240 K, indicating that the ions in the 4S3/2(1, 2) and 4F9/2(1, 2, 3) levels are thermally excited to the 2H11/2 level simultaneously. For the intensities of 655 nm or 665 nm, the ratios change dramatically under 240 K, which is because of thermal excitation and nonradiative relaxation for lower temperatures. However, the ratios of 551/564, 551/680, and 564/680 remain for the entire experimental temperature range. The most important deduction is that for the lowest level and intermediate ladder level, the relative populations of 4S3/2(1, 2) to 4F9/2(1, 2, 3) should not follow the Boltzmann distribution, and the FIRs of the levels cannot be used for temperature sensing in this sample.

Fig. 4
figure 4

a FIRs of the emissions of the sublevels of the 4S3/2 level or 4F9/2 levels, b FIRs of emissions the of sublevels of the 4S3/2 level and 4F9/2 levels

The experimental data of 539/551, 539/564, 539/655, 539/665, and 539/680 were obtained and depicted in Fig. 5. The FIRs of five pairs were fitted with \(A{\text{exp}}\left( { - B/T} \right) + C\) [24]. Here, A is determined by the degeneracy, the emission cross-section, and the angular frequency of two levels, B is determined by \(\Delta E/k_{{\text{B}}}\), and C is introduced to simplify the equation [34]. The fitting formula are listed in Table 1. The theoretical curves match the experimental results well. The values of B in Table 1 are similar and between the actual \(\Delta E/k_{{\text{B}}}\) of 2H11/2/4S3/2 and 2H11/2/4F9/2. The intermediate ladder level 4S3/2 can effectively reduce the transition interval limit of TCLs [24]. Meanwhile, the values of B are different owing to laser power fluctuations, the host matrix’s absorption, and the thermal effect of nonradiative relaxation [35].

Fig. 5
figure 5

Temperature-dependent luminescent behaviors of 539/551, 539/564, 539/655, 539/665, and 539/680 as a function of temperature

Table 1 FIRs of five pairs fitted with the Boltzmann distribution

The absolute sensitivity Sa and relative sensitivity Sr of FIRs were obtained, as shown in Fig. 6. The Sa of the five pairs increase nonlinearly with increasing temperature, among which the Sa of 539/655 increases for the entire temperature range and are maximized at 320 K as 142. 4 × 10–4 K−1. Meanwhile, Sr of 539/655 is superior to those of other pairs, which is greater than 1.5% K−1 in the temperature range from 170 to 290 K with maximum value of 2.21% K−1 at 203 K. The maximum temperature sensitivities of five pairs are listed in Table 2. The FIRs of Stark sublevels significantly enhanced the absolute sensitivity, and the maximum Sa is approximately 26 times larger than that of traditionally used 2H11/2/4S3/2 levels, which is also larger than that of other Er3+ doped hosts such as Al2O3, NaYF4, YAG, Gd2O3, and CaMoO4 [36,37,38,39,40,41,42,43]. The enhanced sensitivity is attributed to the energy gaps of corresponding thermal levels and the relative radiative transition probabilities [14]. The absolute sensitivity Sa of the Er3+/Yb3+: PLZT ceramic was compared with other Er3+/Yb3+ codoped materials based on the FIRs of Stark sublevels, as listed in Table 3. The maximum Sa of the Er3+/Yb3+: PLZT ceramics is higher than that of most Er3+/Yb3+ codoped phosphors. Thus, these Er3+/Yb3+: PLZT ceramics are promising materials for use as optical thermometers for temperatures higher than 160 K because of their enhanced temperature sensitivity.

Fig. 6
figure 6

a Absolute temperature sensitivity Sa and b relative temperature sensitivity Sr of the FIRs of five pairs

Table 2 Temperature sensitivities of FIRs of the emissions of Er3+/Yb3+ codoped PLZT ceramic
Table 3 Maximum absolute temperature sensitivities of some Er3+/Yb3+ codoped hosts based on FIRs of Stark sublevels

4 Conclusion

The temperature-dependent emissions of the Stark sublevels of Er3+/Yb3+ codoped PLZT transparent ceramic were studied within 160–320 K based on upconversion fluorescence spectra under 980 nm excitation. Notably, 2H11/2 and 4S3/2 level are split into two sublevels as 2H11/2(1, 2) and 4S3/2(1, 2), and 4F9/2 level is split into three sublevels as 4F9/2(1, 2, 3). All the level transitions have strong temperature dependence. The temperature characteristics of the emissions of the Stark sublevels were studied, and we found that the relative intensities of the lower level 4F9/2 and intermediate step level 4S3/2 in TCMLs cannot be used for temperature sensing. Hence, the FIRs of 539/551, 539/564, 539/655, 539/665, and 539/680 were studied and theoretically fitted for temperature sensing. The maximum values of absolute sensitivity Sa and relative sensitivity Sr were obtained as 142.4 × 10–4 K−1 at 320 K and 2.21% K−1 at 203 K, respectively. The Sa was enhanced 26 times, while the Sr is greater than 1.5% K−1 in the temperature range from 170 to 290 K. A comparison with other Er3+/Yb3+ codoped materials shows that Er3+/Yb3+: PLZT ceramics are a promising thermometer material for temperatures greater than 160 K. Meanwhile, the application of FIRs based on the emissions of Stark sublevels might enhance temperature sensing performance.