Abstract
We have prepared NaYF4:Er3+/Yb3+/Ca2+ phosphors using the solvothermal method and discussed the influence of annealing on the crystal structure, morphology, and photoluminescence characteristics in this work. The results show that, with the increase of annealing temperature, the crystallization phase changes from a single cubic structure to a mixed phase of cubic and hexagonal phases and the morphology is also condensed from spherical particles to sheets. In particular, the PL intensity of the annealed phosphor becomes much stronger than that of the unannealed phosphor. An annealed NaYF4:Yb3+, Er3+/Ca2+ phosphor is selected as an example to investigate the performance of high temperature sensor based on the fluorescence intensity ratio between two green emissions. The measurement range and the maximum absolute sensitivity SA of the annealed sensor material are superior to the corresponding parameters of the same phosphor without annealing. The curve of relative sensitivity curve SR decreases as the ambient temperature increases, but that of the revolution δT presents an opposite trend. All the results show that the NaYF4:Er3+/Yb3+/Ca2+ phosphor is more suitable for application on the high temperature sensor after annealing.
1. Introduction
The investigation of the method and materials for temperature measurement with excellent performance is still a hot topic because of its comprehensive application in various fields [1–7]. Especially in the last two decades, the optical sensing technique of measuring high temperature based on the florescence intensity ratio (FIR) of the photoluminescence (PL) intensities resulting from the emission of material-doped rare earth ion (REI) has attracted increasing attention due to its wide dynamic range, high sensitivity, almost perfect anti-interference ability, and so on [8–16]. Recently, it has been found that NaYF4 phosphor is a favorable host for the doping of trivalent REIs due to its lower phonon energy, higher luminescence efficiency, and stable structure [17–20], and a lot of works have been reported on the design and preparation of NaYF4 materials-doped REIs, as well as its application in optical detection, biological labeling, and temperature sensing. [21–24]. For example, Zeng et al. [25] investigated the synthesis and upconversion luminescence of hexagonal phase NaYF4:Yb, Er3+ phosphors with controlled size and morphology; Schietinger et al. [26] discussed the plasmon-enhanced upconversion of single NaYF4:Yb3+/Er3+ codoped nanocrystals; Liu et al. [27] designed the chemical sensors based on periodic mesoporous organosilica@NaYF4: Ln3+ nanocomposites; Shan and Ju [28] synthesized controllable lanthanide-doped NaYF4 upconversion nanocrystals via ligand-induced crystal phase transition and silica coating. Of cause, the chemical and physical stability of the NaYF4 host is still deficient compared to the oxide matrix, which causes the significant influence of annealing temperature on the PL intensity and crystal structure of NaYF4:Yb3+, Er3+/Ca2+ materials [29, 30].
Unfortunately, the PL intensity of NaYF4 material doped with REIs is not very satisfactory and there is still room for improvement. Therefore, there have been some reports on the incorporation of Ca2+ ions into NaYF4 in order to improve the PL characteristics of the luminescent material [31, 32]. For example, Li et al. [33] prepared a series of hexagonal NaYF4:Yb/Er microcrystals codoped with different Ca2+ contents by a facile trisodium citrate-assisted hydrothermal method, and they found that the upconversion photoluminescence from ultraviolet to visible is significantly enhanced with the increase of Ca2+ concentration. The reasons for the enhancement are that the crystal structure exhibits the asymmetry and the crystallinity of NaYF4:Yb/Er is further improved after Ca2+ codoping.
On the basis of the discussions above, we also try to codope Ca2+ ions into NaYF4:Yb3+, Er3+ phosphor in our work in order to improve and enhance PL characteristics. Therefore, NaYF4:Yb3+, Er3+ phosphors have been prepared by solvothermal method in this work, and Ca2+ ions are simultaneously codoped into the phosphor in the preparation process to improve the PL characteristics. In particular, the synthesized NaYF4:Yb3+, Er3+/Ca2+ phosphors have been annealed at different temperatures in order to further enhance the PL intensity. The X-ray diffraction (XRD) patterns show that the crystallization phase gradually changes from the cubic phase to the mixed phase of cubic and hexagonal phase with the increase of annealing temperature. At the same time, the images of the scanning electron microscope (SEM) reveal that the morphology is also condensed to the layer of spherical particles after the phosphor is annealed at higher temperature. In addition, the PL intensity becomes much stronger than that of the phosphor without annealing. The NaYF4:Yb3+, Er3+/Ca2+ phosphor annealed at 1,000°C is chosen as an example to demonstrate the performance of optical high-temperature sensors (OHTS) based on the FIR between two green emissions of Er3+ ions. All parameters of the annealed temperature sensor material, including the measurable range and absolute/relative sensitivities, as well as the temperature resolution, are better than those of the nonannealed phosphor, which indicates that the temperature sensing behavior of the NaYF4:Yb3+, Er3+/Ca2+ phosphor has been significantly improved after annealing.
2. Preparation and Characterization of NaYF4:Er3+, Yb3+/Ca2+ Phosphor
In this section, the preparation process of NaYF4:Er3+, Yb3+/xCa2+ phosphors by solvothermal method is introduced in detail because this technique has many excellent advantages for synthesis of the powder, such as high purity, good crystallinity and dispersion, narrow particle size distribution, controllable morphology and phase formation, and so on [34, 35].
It should be noted beforehand that all chemical reagents, including NaCl, YCl3·6H2O, YbCl3·6H2O, ErCl3·6H2O, NH4F, polyetherimide (PEI), CaF2, and (CH2OH)2, are analytical grade and can be used directly without further purification.
Figure 1 illustrates the entire preparation process of NaYF4:Er3+/Yb3+/xCa2+ phosphors. First, the raw materials of NaCl (1.2 mmol, the same unit as below except ml for (CH2OH)2)), YCl3·6H2O (0.48), YbCl3·6H2O (0.108), and ErCl3·6H2O (0.012) are weighed, respectively, and dissolved together in (CH2OH)2 (9 mL). The mixed liquid is stirred with a magnetic stirrer at room temperature until the solution becomes clear and marked as A. Similarly, both PEI (3.0) and NH4F (0.006) are also dissolved in (CH2OH)2 (9 mL), stirred and marked as B. Subsequently, the solutions A and B are mixed together with xCaF2 (x = 0–0.7 and the interval is 0.1 mmol). The mixed solution is stirred again for 30 min until it appears a uniform state, and then it is poured into a reaction vessel and placed in a drying oven that has been preheated to 200°C. Third, all the semifinished samples are put into different centrifuge tubes, and after the reaction above is completed, they are put into a centrifuge and cooled to room temperature about 150 min later. The speed and time of the centrifuge are set, respectively, to 9,000 rpm and 15 min, respectively. It is observed that the transparent colloid settles to the bottom of the tube after the supernatant has been removed. An appropriate amount of anhydrous ethanol is added to the centrifuge tube containing the synthetized sample and shaken manually. Then, the tubes are further vibrated in an ultrasonic cleaning device to evenly disperse the agglomerated nanoparticles and to make the organic molecules on the surface of the nanoparticles soluble in anhydrous ethanol. Centrifugalization is then repeated at the same speed and time. Finally, the colloid at the bottom of the centrifuge tube will become a white precipitate after repeating the steps of the washing process three to four times, and all the powders are dried in a drying oven working at 80°C for 60 min. Now, each sample is divided into two groups: one group is annealed for 2 hr at 600, 700, 800, 900, and 1,000°C, respectively; another group keeps its original state without annealing and is used for the experimental control group.

The diffraction patterns obtained with an X-ray diffractometer (Shimadzu Company, Japan) of representative NaYF4:Er3+, Yb3+/Ca2+ phosphors are demonstrated, as shown in Figure 2. It can be seen that the red spectrum belonging to the phosphor without annealing is agreement with the standard chart (JCPDS 77-2,042), which means that this type of phosphor has a single α phase, i.e., a stable cubic structure (a = b = c = 5.47 Å). It can also be found from the blue line that, however, only a few diffraction peaks coincide with those in JCPDS 77-2,042, whereas the remaining peaks are in accordance with JCPDS 28-1,192, meaning that the same phosphor, after annealing at 1,000°C, has a mixed phase, including α phase and β phase (hexagonal crystal, a = b = 5.96 Å, c = 3.51 Å). The XRD characterization results of the other samples annealed at different temperatures also prove that there exists a significant effect of annealing on the crystal structure of NaYF4:Er3+, Yb3+/Ca2+ phosphor and the specific reasons for this trend of change need to be further investigated.

The SU8000 scanning electron microscope (Hitachi, Japan) is adopted to characterize representative samples and the images are shown in Figure 3. The morphology of the NaYF4:Er3+, Yb3+/Ca2+ nanophosphor without annealing is a collection of spherical particles, as shown in Figure 3(a), where the inset exhibits the statistical distribution of particle size and the maximum, minimum, as well as average sizes are approximately 89, 15, and 31 nm, respectively. Most of the particles, however, are concentrated around the 20–40 nm range; in other words, the distribution range is narrow. The particles are condensed into sheets when the sample is annealed at high temperature. Figure 3(b) demonstrates the representative morphology of the phosphor annealed at 1,000°C and it consists of some sheets with various sizes ranging from 0.2 to 0.6 μm [36, 37]. In short, the annealing process can change the morphology of the same phosphor from a granular shape to a crystal flake.

(a)

(b)
3. Results and Discussion
In the following discussions, the doping amounts of trivalent ytterbium and erbium ions are fixed at 0.108 and 0.012 mmol, respectively, based on the existing reports and previous works. The next step is to optimize the Ca2+ ion concentration and the annealing temperature.
Figure 4 displays the variation of PL intensity as a function of the Ca2+ doping amount under the excitation of a 980-nm semiconductor laser. It can be seen that there are three luminescence bands, including the intense green emissions centered at 524 and 544 nm, red emissions at 651 and 668 nm, and a faint peak at 407 nm, corresponding, respectively, to the transitions of Er3+ ions from the excited state levels 2H11/2, 4S3/2; 4I9/2; 2H9/2 to the ground state level 4I15/2. Within each band, there exist a few peaks because the energy level of the Er3+ ion undergoes a splitting phenomenon under the influence of the crystal field. It can also be known, as shown in Figure 4, that the center wavelength of the NaYF4:Yb3+, Er3+/Ca2+ phosphors is almost identical to that of the NaYF4:Yb3+, Er3+ phosphors with undoped Ca2+ ions, suggesting that there exists almost no effect of Ca2+ content on the wavelength of the NaYF4:Yb3+, Er3+/Ca2+ phosphors. The inset, as shown in Figure 4, reveals that the optimum doping level of Ca2+ ions is 0.4 mmol for green emission at 544 nm, 0.3 mmol for green/red emission at 524, 651, and 670 nm, as well as 0.2 mmol for blue emission at 407 nm. Therefore, in the following experiments, the Ca2+ doping level is selected to be 0.4 mmol in order to gain a stronger green emission for application in a temperature sensor using FIR technology. Figure 5 demonstrates the change in photoluminescence intensity of NaYF4:Er3+, Yb3+/0.4 Ca2+ phosphor as a function of annealing temperature and it can be known from the figure that the PL intensity becomes increasingly stronger, and especially the increased amplitude of the intensity becomes more obvious with the increase in annealing temperature. The main reasons for this are: (i) the morphologies of the phosphor change with increase in annealing temperature from particles to sheet; in other words, the latter resembles a bulk material in which the particle density becomes higher, resulting in stronger PL intensity; (ii) the crystal structure of NaYF4:Er3+, Yb3+/Ca2+ phosphor with a single α phase also changes to α–β mixed phase when the same sample is annealed at 1,000°C, which indicates that α–β mixed phase may be more compatible to the doping environment of Er3+, Yb3+, and Ca2+ ions, making the PL intensity intenser.


The upconversion photoluminescence spectra of the green emission band of NaYF4:Yb3+, Er3+/0.4 Ca2+ phosphors excited by a 980-nm semiconductor laser are shown in Figure 6. It can be known that (i) the PL intensities of the samples, whether being annealed at 1,000°C (blue line) or not (red line), become weaker significantly at 544 nm, as well as intenser slightly at 524 nm with the increase of ambient temperature from 303 (solid line) to 543 K (dotted line); (ii) the PL intensities of annealed sample is far stronger than these of the same phosphor without annealing, even though the intensities of the latter are multiplied by 10 and 80, respectively. The inset, as shown in Figure 6, refers to the photos of emitting green light from the two kinds of phosphors shotted at room temperature. Another is the sketch map of Er3+ levels corresponding to its green emission. Interestingly, the peak wavelength of the phosphor without annealing appears “redshift” at about 544 nm, whereas “blueshift” at 524 nm with the increase of ambient temperature from 303 to 543 K, and the reason lies in the energy levels of Er3+ ions occur the movement slightly because of the additional energy originating from thermal energy in its crystal Hamiltonian item. The peak wavelengths of the sample with annealing, however, do not appear red or blueshift, indicating that the structure of the energy level becomes more stable after the phosphor is annealed. Figure 7 reveals the change trends of PL intensity with increasing ambient temperature. It can be found that the integral intensities at both 524 and 544 nm (red dotted and star lines) of the unannealed phosphor weaken gradually until they are difficult to measure when the temperature exceeds 543 K, meaning that the measurable range of this sort of phosphor will be lower than 240 K once it is applied to optical temperature sensor. The inset, as shown in Figure 7, is the enlarged drawing belonging to the intensity variations of the sample without annealing in order to be examined intuitively and conveniently.


According to the Boltzmann distribution formula, i.e.,where Ni denotes the population at the energy level Ei, k depicts the Boltzmann constant, T declares the absolute temperature, and c is a constant, respectively; the population Ni in each energy level will descend with the increase in temperature T, which is the main reason for the weakening of the PL intensities at 524 and 544 nm. On the other hand, the photon energy is also strengthened with increase of the ambient temperature, which drives Er3+ ions at 4S3/2 level to 2H11/2, causing the intensity of the green emission at 544 nm to be further weakened but that at 524 nm to be enhanced. Therefore, the integral intensity of the green emission at 524 nm varies slightly on a small scale.
Although the integral intensity of green emission at 544 nm (blue star) from the annealed phosphor also abates clearly and that at 524 nm (blue dotted) fluctuates slightly at the small scale, they are far stronger than these emitting from the unannealed sample. In other words, the intensity ratio between I544 and I524 can present change prominently, and the feature is beneficial to the application in high temperature sensors because the upper limit of measurable temperature can extend to about 750 K.
The fluorescence intensity ratio between two green upconversion emissions at 524 and 544 nm can be defined as [38, 39]:where Ij (j = 524 and 544 nm) denotes the integral PL intensity of the green emissions at 524 and 544 nm. N represents the population number in the energy levels 2H11/2 and 4S3/2. g, σ, and ω refer to the degeneracy, the emission cross-section, and the angular frequency of fluorescence transitions. ΔE exhibits the energy gap between the coupling levels 2H11/2 and 4S3/2. K and T mean the Boltzmann constant and absolute temperature, respectively, and the preexponential constant C = gHσHωH/gSσSωS can be determined by the fitting process of experimental data.
Figure 8 demonstrates the FIR change as a function of temperature. One can be learnt from the figure that the curve of the unannealed phosphor (red line) shows a monotonous upward trend first with the increase of ambient temperature from 303 to 453 K and then closes to saturation within 453–543 K, implying that the measurable range of the unannealed phosphor is only about 150 K if the sample is used for temperature sensing material. In addition, there exist a piecewise fitting functions R = 4.19 exp (−927.8/T) and R = 0.84 exp (−208.1/T), respectively. FIR change of the annealed phosphor (blue line), however, appears the increase trend throughout from room temperature to 723 K, that is, its measuring scope is 2.8 times larger than that of the unannealed sample. Furthermore, the preexponential coefficient in its fitting function R = 6.27 exp (−999.4/T) is also bigger than that in the red line, indicating that the sensitivity of the former applied to temperature measuring is more excellent than that of the latter.

Moreover, the absolute sensitivity SA is defined to further unveil the performance of an optical high temperature sensor, and its expression is given as follows:
The absolute sensitivity changes of both kinds of NaYF4:Yb3+, Er3+/0.4 Ca2+ phosphors (annealed and unannealed) as a function of temperature are shown in Figure 9. As discussed above, the sensitivity SA of the phosphor annealed at 1,000°C (blue line) is better than that of the unannealed phosphor (red line), and their maximum values are 33.942 × 10−4 K−1 at 483 K and 24.399 × 10−4 K−1 at 453 K, respectively. Especially, the absolute sensitivity of the latter descends sharply from 24.399 × 10−4 to 4.007 × 10−4 K−1 in the interval of 453–473 K, further proving that the unannealed phosphor is not suitable for the application of high temperature sensors.

In addition, according to the work reported in Ref. [40], the relative sensitivity SR and temperature resolution δT are also key parameters for evaluating the optical temperature sensor and can be computed using the following formulae:andwhere δR denotes the standard deviation of FIR and δR/R is the relative error to determine the temperature parameters. Here, we take δR/R = 0.1% in the computing process, which is superior to 0.5% used often in the existing works [12, 41, 42].
Based on Equations (4) and (5), the relative sensitivity SR and temperature resolution δT of the NaYF4:Yb3+, Er3+/0.4 Ca2+ phosphor annealed at 1,000°C are calculated and shown in Figure 10. It can be observed that the SR curve presents a trend of gradual decline from about 1.1% to 0.16% when the ambient temperature rises, which implies that the higher the temperature is, the better the performance of the sensing material becomes. Instead, the δT curve goes up with the increase of temperature from 303 to 783 K. The minimum value of the temperature resolution is about 0.16 K at 303 K and its maximum value is around 1.1 K even at the highest temperature 783 K. In a word, NaYF4:Yb3+, Er3+/0.4 Ca2+ phosphor has become a promising optical material for high temperature measurement after it is annealed due to its excellent sensing performance, including intense PL intensity, wide measurement range, high absolute sensitivity, low relative sensitivity, and favorable resolution.

The performance comparison of key indicators reported by the existing representative optical high temperature sensors is listed in Table 1 on the basis of the Refs. [43–46]. It can be known that the temperature measurement range of the annealed sensing material prepared in our work is wider compared to that reported in other works. The performance of our absolute sensitivity SA is in a middle level, but the relative sensitivity SR is more superior due to its intense PL characteristics and good anti-interference ability.
4. Conclusion
In summary, a series of NaYF4 phosphors codoped with Er3+,Yb3+, and Ca2+ ions have been prepared by the solvothermal method, and the doping amount of Ca2+ ions is optimized when the ones of Er3+ and Yb3+ ions are fixed. In addition, the synthesized NaYF4:Yb3+, Er3+/0.4 Ca2+ phosphors have been annealed at different temperatures and the results show that the photoluminescence intensity enhances significantly with the increase of annealing temperature. XRD patterns reveal that the crystal structure of the unannealed phosphor is a single α phase, but it will become α–β mixing phase once the sample is annealed at 1,000°C. SEM images prove that the morphology has been condensed to a sheet from spherical particles belonging to an unannealed phosphor after the phosphor is annealed at high temperature. Subsequently, we choose NaYF4:Yb3+, Er3+/0.4 Ca2+ phosphor annealed at 1,000°C as an example to investigate the performance of optical high temperature sensor according to the technique of florescence intensity ratio between two green emissions originated from Er3+ ion. The measurable temperature range of the annealed sensing material is 303–723 K and its maximum absolute is 34 × 10−4 K−1 at 493 K, which are superior to the corresponding parameters of the phosphor without annealing. The relative sensitivity descends with the increase of ambient temperature, but the revolution presents an opposite trend. All the results prove that the optical sensing material of NaYF4:Er3+, Yb3+/0.4 Ca2+ phosphor is suitable for application to the high temperature measurement after it is annealed.
Data Availability
All data included in this manuscript are available from the corresponding author on reasonable request; email: lnnulicr@aliyun.com.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
Acknowledgments
The authors extend their appreciation to the Guiding Project of Heilongjiang Key Research and Development Program for funding this work through a large group research project under grant number GZ20220059.