Spectral Characterization of CeF₃-YF₃-TbF₃ Nanoparticles for Temperature Sensing in 80–320 K Temperature Range

Abstract

The studied Ce_0.5Y_0.5−XTb_XF₃ (X = 0, 0.001, 0.002, 0.005, 0.01, and 0.05) nanoparticles were synthesized via the water-based co-precipitation method. All the samples demonstrated diameters in the 17–20 nm range and a hexagonal phase corresponding to the phase of CeF₃. Under 266 nm excitation (4f–5d absorption band of Ce³⁺), the luminescence spectrum shape was notably dependent on temperature. The integrated luminescence intensity ratio (LIR) of Ce³⁺ and Tb³⁺ (⁵D₄–⁷F₃) peaks was chosen as a temperature-dependent parameter. It was shown that the LIR functions linearly decay. The rate of decay decreases with the increase in Tb³⁺ concentration. This was explained by the fact that in the case of low Tb³⁺ concentrations, the spectral temperature dependence is mostly based on effective thermal quenching of Ce³⁺ luminescence. At higher Tb³⁺ concentrations, there is a higher probability of Ce³⁺ to Tb³⁺ energy transfer. Here, the efficiency of the temperature dependence of this process is lower, and the rate of LIR decay is lower as well.

1. Introduction

Luminescence temperature sensing is a very powerful tool to measure the temperature of small objects (micro-devices, parts of microcircuits, living cells, etc.) with high spatial resolution. In this method, temperature-dependent parameters of the luminescence of special phosphors are analyzed [1,2,3,4]. Among a huge variety of organic and inorganic phosphors, rare earth-doped fluoride nanoparticles play a very important role due to a lack of photobleaching, bright luminescence, high mechanical and chemical stability, and low cytotoxicity [5,6,7,8]. In particular, Pr³⁺ [9], Dy³⁺ [10], and Er³⁺ [11] phosphors have been successfully utilized as luminescence temperature sensors due to the presence of thermally coupled electron levels. Here, the temperature sensitivity depends on an energy gap between these thermally coupled levels. However, it is almost impossible to affect the value of the energy gap because of the shielded nature of the rare earth ion f-electron shell. An approach allowing this limitation to be overcome is the use of double-doped phosphors. In this approach, the temperature-dependent energy transfer efficiency between donor and acceptor ions plays a crucial role in the temperature sensitivity of the luminescence (for f-f luminescence). For example, there are Tm³⁺/Yb³⁺ [12], Nd³⁺/Yb³⁺ [13], Pr³⁺/Yb³⁺, and Er³⁺/Yb³⁺ [14] down-conversion systems. However, this approach also has some limitations, including the spectral overlap of donor and acceptor and back energy transfer from acceptor to donor, which smooths the sensitivity [15].

It seems that the more efficient method of increasing the sensitivity is the use of double-doped phosphors, where the donor is excited via an allowed 4f–5d transition and the acceptor demonstrates f-f luminescence [16,17]. Here, an interesting example is Ce³⁺/Tb³⁺-doped phosphors. Indeed, the Ce³⁺ ion has a 4f¹ configuration and demonstrates UV or blue luminescence due to the 5d–4f allowed transition. In turn, Tb³⁺ demonstrates bright f-f luminescence in the visible part of the spectrum due to transitions from the excited ⁵D₃ and ⁵D₄ states to the lower ⁷F_J (J = 0–6) ones. The Ce³⁺–Tb³⁺ energy transfer processes are described in [18,19]. Here, the luminescence intensity ratio (LIR) of Ce³⁺ and Tb³⁺ luminescence peaks is a temperature-dependent parameter due to the different thermal quenching efficiencies of Ce³⁺ and Tb³⁺. However, it seems that the physical mechanism of the luminescence of Ce³⁺/Tb³⁺-doped phosphors is more complicated due to the presence of Ce³⁺–Tb³⁺ energy transfer, Tb³⁺ multiphonon relaxation (⁵D₃–⁵D₄), and the cross-relaxation of Tb³⁺ [20]. It should also be noted that Ce³⁺/Tb³⁺ based phosphors are highly necessary in imaging technologies and the light-emitting diode industry [21,22,23,24].

In our previous works, we studied the spectral kinetic characteristics of CeF₃-TbF₃-YF₃ nanoparticles, as considered here, in the ~300–500 K temperature range. Here, we perform a spectral characterization of the samples in the lower 80–320 K temperature range [20].

One of the main novelties of this work is that here we consider a double-doped system (Ce³⁺/Tb³⁺), where the donor ion (Ce³⁺) is optically excited via 4f–5d transition. This class of materials is significantly less studied compared to double-doped materials excited via f-f transitions (Nd³⁺/Yb³⁺, Tm³⁺/Yb³⁺, etc.). Due to the high sensitivity of 5d electrons to the surroundings, this class of materials has potentially higher temperature sensitivities. No less important is that we consider several samples having different Tb³⁺ concentrations in order to manipulate and maximize the thermometer performances.

The objective of this paper was to study the physical mechanism of the luminescence temperature sensitivity of the synthesized CeF₃-TbF₃-YF₃ nanoparticles in the ~80–320 K temperature range.

2. Results

2.1. Physical Characterization of the Samples

The TEM image of Ce_0.5Y_0.49Tb_0.01F₃ nanoparticles is presented in Figure 1. It can be seen that the synthesized particles are not perfectly spherical. In addition, the Tb³⁺ concentration does not affect the nanoparticle morphology.

The size distribution histogram is presented in Figure 2.

The average diameter of the samples is in the 17−19 nm range. The XRD patterns of the samples are presented in Figure 3.

It can be seen that all the samples demonstrate a hexagonal phase corresponding to the CeF₃ host (COD ID—1011350, Space group P63/mcm). The impurity peaks and amorphous phase were not detected. We calculated the sizes of the coherent stuttering domain (CSD) according to the well-known Scherrer equation [25]:

$$CSD=K{\displaystyle \frac{\lambda }{\beta ·Cos\theta }}$$

(1)

where K is a shape-factor (we took 0.94 as the value); λ—X-ray wavelength (Kα = 0.154 nm); β—full width at half maximum (FWHM); and θ is an angle extracted from the 2θ peak position. In addition, we calculated the lattice parameters. All the calculated values are listed in

Table 1. Structural parameters of the samples depending on the Tb³⁺ concentration (X). Here. 2θ is an XRD peak position; FWHM—full width at half maximum; CSD—coherent stuttering domain; d is a physical diameter obtained from the TEM image.

Sample Ce0.5Y0.5−XTbXF3	2θ, Degree	FWHM, β, Degree	CSD, nm	d, nm	Lattice Parameters a, Å	Lattice Parameters c, Å
X = 0.001	28.50	0.7011 ± 0.0091	12	18 ± 1	6.97 ± 0.07	7.08 ± 0.06
X = 0.002	28.52	0.6868 ± 0.0122	12	20 ± 1	6.98 ± 0.08	7.10 ± 0.06
X = 0.005	28.50	0.5326 ± 0.0075	15	17 ± 1	6.98 ± 0.07	7.08 ± 0.05
X = 0.01	28.30	0.5944 ± 0.0093	14	20 ± 1	7.01 ± 0.08	7.13 ± 0.11
X = 0.05	28.48	0.6302 ± 0.0078	14	21 ± 1	6.99 ± 0.07	7.09 ± 0.06
X = 0	28.52	0.6277 ± 0.0097	14	19 ± 1	6.98 ± 0.07	7.09 ± 0.06

.

It can be seen that the peak position weakly depends on Tb³⁺ content, probably because Tb³⁺ and Y³⁺ have similar ionic radii. The CSD sizes are slightly smaller than the physical diameters (d). This can be explained by the fact that the Scherrer equation takes into consideration size effects only. However, in the obtained samples, Ce³⁺ and Y³⁺ have different ionic radii, and the samples probably contain tensions. These tensions, additionally, broaden the XRD peaks. The room temperature luminescence spectra of the Ce_0.5Y_0.5−XTb_XF₃ (X = 0.001; 0.002; 0.005; 0.01; 0.05) samples are presented in Figure 4. The spectra are normalized at the 544 nm peak of Tb³⁺. The 266 nm excitation wavelength corresponds to the 4f–5d absorption band of Ce³⁺.

All the luminescence peaks were interpreted as radiative transitions from the excited states of both Ce³⁺ and Tb³⁺. In particular, four intense ⁵D₄–⁷F_J (J = 6, 5, 4, and 3) Tb³⁺ peaks are clearly observed. The intensity of Tb³⁺ peaks increases with the increase in its concentration. Finally, it can be concluded that the synthesized samples are 17–21 nm nanoparticles in diameter. The nanoparticles demonstrated the expected hexagonal phase. Impurity peaks and the amorphous phase were not detected. The luminescence spectra show the presence of both Ce³⁺ and Tb³⁺ peaks.

2.2. Temperature-Dependent Spectral Characterization of the Samples

The normalized at 266 nm luminescence spectra of Ce_0.5Y_0.5−XTb_XF₃ (X = 0) (without Tb³⁺) under 266 nm excitation (4f–5d absorption band of Ce³⁺ ions) are presented in Figure 5. The spectra were recorded in the 80–320 K temperature range.

It can be seen that the Ce³⁺ luminescence significantly decreases with the increase in temperature compared to excitation. There is also a broadband luminescence in the approximately 450–600 nm spectral range. This broadband luminescence can be ascribed to defects. It should also be noted that the intensity of this broadband luminescence sharply decreases with the temperature increase. The normalized integrated intensity values of the Ce³⁺ peak of Ce_0.5Y_0.5−XTb_XF₃ (X = 0) (without Tb³⁺) in the 80–320 K temperature range are presented in Figure 6. Note that integrated intensities were calculated according to Figure 5.

As Figure 7 shows, the Ce³⁺ intensity weakly decreases in the 80–150 K range, and then sharply decreases from ~150 to 320 K (note that the LIR does not have units). The mechanism of Ce³⁺ luminescence thermal quenching is complex and described in the literature [26,27]. It can be explained by the conventional configurational coordinate diagram (Figure 7).

There is an activation energy for the Ce³⁺ luminescence quenching process. Suggestively, this energy corresponds to a temperature around 150 K, after which the efficiency of Ce³⁺ quenching increases sharply. In particular, the lift part of the 5d curve of Ce³⁺ only has one crossing point D (Figure 7), which is the crossing point of the excited and ground state of Ce³⁺. The increase in the environmental temperature gives excited electrons the extra energy to move from C to D. The quantity of electrons in D increases, which allows more electrons to return to the ground state non-radiatively through the path ABCDA. In the case of Ce³⁺/Tb³⁺ ion pairs, the temperature dependence of the spectral characteristics is more complicated. The normalized at 544 nm (Tb³⁺ peak) luminescence spectra of Ce_0.5Y_0.5−XTb_XF₃ (X = 0.001, 0.002, 0.005, 0.01, and 0.05) nanoparticles detected in the 80–320 K temperature range are presented in Figure 8a–e, respectively.

It can be seen that the shape of the spectra notably depends on temperature. The relative change in Ce³⁺ and Tb³⁺ intensities takes place. In particular, the Ce³⁺ intensity gradually decreases with the increase in Tb³⁺ concentration (Figure 8a–e). On the other hand, the above-mentioned broadband luminescence is also observed for Ce_0.5Y_0.5−XTb_XF₃ (X = 0.001 and 0.002) samples. The same broadband we observed in [25]. The difference in broadband intensities for different samples can be explained by the intense Tb³⁺ luminescence compared to the broadband intensity. In order to realize the ratiometric luminescence thermometry approach, the luminescence intensity ratio (LIR) of Ce³⁺ and Tb³⁺ peaks should be calculated. However, this broadband can be a notable drawback for the LIR calculation. In turn, the broadband does not affect the Tb³⁺ peak at 621 nm corresponding to the ⁵D₄–⁷F₃ transition. Thus, we calculated the LIR of Ce³⁺ (250–350 nm) and Tb³⁺ (615–625 nm). The LIR functions are presented in Figure 9.

As Figure 9 shows, all the LIR functions decrease with the increase in temperature. In addition, the decay rate of the LIR functions decreases with the increase in temperature. It can be seen that the decay rates of two samples, Ce_0.5Y_0.5−XTb_XF₃ (X = 0 and 0.001) without Tb³⁺ and with the lowest concentration, demonstrated almost the same decay rate. Particularly, Δ = LIR (80 K) − LIR (320 K) = ~30. Indeed, the Ce³⁺ thermal quenching is highly efficient due to efficient electron–phonon coupling. Hence, the Ce³⁺ intensity in Ce_0.5Y_0.5−XTb_XF₃ (X = 0, without Tb³⁺) can operate as a luminescent thermometer based on the analysis of intensity. However, this approach is not reliable because the luminescence intensity also depends on the fluctuation of excitation power. Here, the double-doped ratiometric Ce³⁺/Tb³⁺ systems are independent of this disadvantage.

The decreasing character of the LIR functions can be explained by the different efficiencies of Ce³⁺ and Tb³⁺ thermal luminescence quenching. Indeed, as was mentioned above, due to the 5d–4f nature of Ce³⁺ luminescence, it has stronger electron–phonon coupling compared to Tb³⁺ f–f luminescence. However, the decay rates of the LIR functions notably depend on the Tb³⁺ concentration. Specifically, the rate decreases with the increase of Tb³⁺ concentration. Discussion of the obtained results necessitates an analysis of the chemical composition. Indeed, the chemical composition of the studied samples is complex. As was mentioned above, since the phase of the sample corresponds to CeF₃, both Y³⁺ and Tb³⁺ substitute Ce³⁺ positions. In the case of the samples with lower Tb³⁺ concentrations (Ce_0.5Y_0.5−XTb_XF₃ (X = 0.001, 0.002, 0.005), it can be suggested that the majority of Ce³⁺ ions do not interact with Tb³⁺ ones. Hence, we observe the luminescence of “pure” Ce³⁺ without interaction with Tb³⁺, and the luminescence of Ce³⁺ that is in contact with Tb³⁺. In the case of Tb³⁺ luminescence, the Tb³⁺ excitation goes through ABEFG path (Figure 7). Finally, the dependence of LIR functions on Tb³⁺ content can be explained by the competition between the more temperature-dependent ABDA (Ce³⁺ thermal quenching) and the less temperature-dependent ABEFG paths. In order to obtain the quantity characteristic of the obtained samples, we calculated relative temperature sensitivity plots according to the convenient equation below:

$$S_r={\displaystyle \frac{1}{LIR}}\left|{\displaystyle \frac{\mathit{\partial }\left(LIR\right)}{\mathit{\partial }T}}\right|·100\%$$

(2)

The S_r plots are presented in Figure 10.

One of the most important characteristics is temperature uncertainty. This performance is determined as follows:

$$\delta T={\displaystyle \frac{1}{S_r}}·{\displaystyle \frac{\sigma \left(LIR\right)}{LIR\left(T_0\right)}}$$

(3)

where σ(LIR) is the standard deviation, T₀ is the temperature at which uncertainty was calculated by repetitive measurements (8 times), and S_r is the relative temperature sensitivity [%·K⁻¹]. Since the LIR functions are linear, the temperature uncertainty values are constant in the whole temperature range. In particular, δT (X = 0.001) ≈ 0.05 K, δT (X = 0.002) ≈ 0.07 K, δT (X = 0.005) = 0.07766 ≈ 0.08 K, δT (X = 0.01) ≈ 0.35 K, and δT (X = 0.05) ≈ 0.90 K. It can be seen that we achieved competitive values. In particular, the maximum S_r was around 0.7%/K at 300 K. In turn, the LaOBr:Ce³⁺/Tb³⁺ exhibited S_r = 0.4 at 300 K [27]. In turn, LiScSiO₄: Ce³⁺/Tb³⁺ showed S_r around 0.5 at 300 K [28]. The interesting material YBO3: Ce³⁺/Tb³⁺ showed a linear increase in the Sr from 0.41 (at 290 K) to 0.5%/K (at 330 K) [29]. Finally, Ce³⁺/Tb³⁺ in the fluoride host of Na_1.5Y_2.5F₉ demonstrated S_r around 0.5%/K ar 300 K. The majority of the papers do not deal with concentration series.

3. Materials and Methods

The studied Ce_0.5Y_0.5−XTb_XF₃ (X = 0, 0.001, 0.002, 0.005, 0.01, and 0.05) nanoparticles were synthesized via the water-based co-precipitation method. The synthesis procedure is described in our previous work [20]. The morphology of the samples (shape and average diameter) was studied by means of transmission electron microscopy (TEM) using a Hitachi HT7700 Exalens transmission electron microscope (Tokyo, Japan). The phase composition of Ce_0.5Y_0.5−XTb_XF₃ was investigated via the X-ray diffraction (XRD) method (Bruker D8 X-ray diffractometer (conventional Cu K_α radiation at 0.154 nm)) (Billerica, MA, USA). The XRD simulation was carried out using VESTA software, version 3 (Ibaraki, Japan). The schematic diagram of the experimental setup for measuring the luminescent properties of the nanoparticles is shown in Figure 11.

The luminescence excitation was carried out via a 266 nm pulse laser (the 4th harmonic of the YAG:Nd laser from Lotis TII LS-2147, Lotis Tii, Minsk, Belarus). Pulse duration and pulse repetition were 10 ns and 10 Hz, respectively). The photoinduced luminescence spectra of Ce_0.5Y_0.5−XTb_XF₃ were recorded using the EPP2000, StellarNet spectrometer (Tampa, FL, USA) (~0.5–1.0 nm spectral resolution). We kept the same integration time of 3 s. The spectral characterization was carried out in the 80–320 K temperature range via the so-called “cold finger” method. The control of temperature was performed via “CRYO industries” thermostatic cooler equipped with the LakeShore Model 325 temperature controller (Lake Shore Cryotronics, Inc., Westerville, OH, USA). The temperature setting accuracy was 0.2 K according to the description. Liquid nitrogen was used as a cooling agent. The samples were deposited as a thin layer onto a copper substrate and subsequently loaded into a cryostat. All the calculations were carried out using the Origin Pro 9.0 software.

4. Conclusions

The studied Ce_0.5Y_0.5−XTb_XF₃ (X = 0, 0.001, 0.002, 0.005, 0.01, and 0.05) nanoparticles were synthesized via the water-based co-precipitation method. All the samples demonstrated diameters in the 17–20 nm range. The nanoparticles exhibited the expected hexagonal phase corresponding to the phase of CeF₃. Under 266 nm excitation (4f–5d absorption band of Ce³⁺), the luminescence peaks of both Ce³⁺ (5d–4f) and Tb³⁺ (⁵D₄–⁷F_J, J = 3, 4, 5, and 6) were clearly observed. Spectral characterization of the samples showed that the samples’ luminescence spectral shape was notably dependent on temperature. The integrated luminescence intensity ratio (LIR) of the Ce³⁺ and Tb³⁺ (⁵D₄–⁷F₃) peaks was chosen as a temperature-dependent parameter. It was shown that the LIR functions linearly decay. However, the rate of decay decreases with the increase of Tb³⁺ concentration. This was explained by the fact that in the case of low Tb³⁺ concentrations, the spectral temperature dependence is mostly based on the effective thermal quenching of Ce³⁺ luminescence. At higher Tb³⁺ concentrations, there is a higher probability of Ce³⁺ to Tb³⁺ energy transfer. Here, the efficiency of the temperature dependence of this process is lower, and the rate of LIR decay is lower as well.