1. Introduction

It is crucially important to precisely measure the temperature in many fields [1,2]. Generally, the conventional contact thermometers, such as thermocouple and resistance thermometers, are sensitive to the electromagnetic field and are therefore restricted in many occasions, although they can give an accurate reading [1]. Radiation thermometry using infrared photodetectors offers a truly contactless technique and does not require multiple optical components, enabling the convenient detection of temperature. However, it cannot be used in glass, liquids, materials with no emissivity and the cost of this technique is high [3–7]. Accordingly, the fluorescence based technique has been the focus of attention in recent decades for its feasibility in the above mentioned occasions, and it only needs the as-prepared functional materials to be contacted with or be injected into the object of interest [8–11]. More importantly, the fluorescence intensity ratio (FIR) method based on the thermally coupled energy levels (TCL) is especially promising and attractive because of its anti-interferences ability, repeatability and predictability.

Lately, Zhou et al. reported a new optical method for temperature measurement, i.e., the valley to peak intensity ratio (VPR) method [12]. The realization of this method depends on three conditions. Firstly, there must be a double-peak spectrum. Secondly, the profile of this spectrum must accord well with the Lorentzian profile throughout the experiment. Thirdly, the line widths of two fluorescence peaks of this double-peak spectrum must broaden homogeneously with the rise of temperature. Only when those conditions are met, the VPR of this double-peak spectrum would increase linearly, which could thus be used to monitor the temperature. Compared with the TCL-based FIR method, the VPR method has better anti-interference ability and less temperature measurement uncertainty [12]. However, it is hard to fulfill the above mentioned conditions simultaneously, which impedes the application of this method in reality, thus making it urgent to explore a simple and convenient VPR method. Furthermore, Zhou et al. utilized the downconversion (DC) emission to detect the temperature, which means that the temperature measurement might suffer from the autofluorescence irradiated by short-wavelength laser. It is obvious that utilizing the upconversion (UC) emission can solve this problem. But until now, no results about the VPR method based on the UC emission have been reported. Consequently, more efforts should be made to study and popularize the UC emission-based VPR method. It should be mentioned that the green UC emission of Er³⁺ was found to be sensitive to the pump power in our experiment, which greatly limits its scope although a higher thermal sensitivity could be obtained utilizing the green UC emission of Er³⁺. And the mechanism involved is still under investigation. In contrast, the red UC emission of Er³⁺ was found to possess the simple spectral structure and had a relative high intensity. Inspiringly, the red UC emission intensity could be further enhanced and easily adjusted by changing the Yb³⁺ concentration. Accordingly, the red UC emission of Er³⁺ was selected as the object of study.

In this work, we successfully developed a more advanced and attractive VPR method based on the UC emission, which just needs a double-peak spectrum constituted by TCL. Upon 976 nm diode laser excitation, the red UC emission of Er³⁺ in CaWO₄:Yb³⁺/Er³⁺ phosphor was investigated as a function of the temperature from 298 to 478 K. This spectrum was found to consist of two Stark components originating from two Stark sublevels of ⁴F_9/2 excited state to ⁴I_15/2 ground state of Er³⁺. These two Stark sublevels of ⁴F_9/2 state were confirmed to be TCL and their populations followed the Boltzmann distribution. The VPR of this double-peak spectrum was investigated and found to increase linearly with the rise of temperature, which demonstrated the feasibility of the proposed VPR method. In addition, the power dependence of the VPR method was investigated over a wide range of pump power.

2. Experimental

The CaWO₄:Yb³⁺/Er³⁺ phosphor was synthesized by a high-temperature solid-state reaction method [12]. The sintered sample had a diameter of 13 mm and a thickness of 1mm. The as-synthesized sample was fixed on a home-made heating chamber with an accuracy of ± 0.1 K in the experimental temperature range. A power-adjustable 976nm near-infrared (NIR) diode laser was used as the excitation source. The irradiated spot size of the sample was about 10 mm². In the range of 298 to 478 K, all UC emission spectra were collected by a monochromator (Zolix Instrument SBP 300) coupled with a photomultiplier (Zolix Instrument PMTH-S1-CR131) at intervals of 20 K. The resolution of the spectrometer was set at a moderate value of around 2 nm. Except for the temperature of the sample, other factors were consistent throughout the experiment. Powder X-ray diffraction (XRD) patterns were obtained using Rigaku D/MAX-2600/PC with Cu Kα radiation (λ = 1.5406 Å). The results revealed that the sample was well crystallized, and all diffraction peaks could be indexed by the data from the standard PDF card of CaWO₄ (No. 77-2236).

3. Results and discussion

Figure 1 shows the room-temperature UC emission spectrum of Er³⁺ in CaWO₄:Yb³⁺/Er³⁺ phosphor in the range of 620 to 700 nm, upon 976 nm diode laser excitation with the pump power of 40 mW. It can be seen from Fig. 1 that the spectrum consists of two fluorescence peaks centered at 655 and 669 nm, respectively. The left and right fluorescence peaks are attributed to the transitions from the higher and lower sublevels of ⁴F_9/2 state to ⁴I_15/2 state of Er³⁺. The inset of Fig. 1 presents the log-log plots of two fluorescence peak intensities I as a function of the pump power P. It is known that for an unsaturated UC process, I ∝ Pⁿ, where n is the number of excitation photons absorbed by the corresponding exciting states [13]. It can be seen from the inset of Fig. 1 that upon 976 nm NIR excitation, both n values are around 2 for two fluorescence peak emissions, suggesting that two-photon UC processes should be responsible for the red UC emission of Er³⁺. The corresponding UC mechanism is depicted in Fig. 2 [5]. Pumped by a 976 nm diode laser, the NIR photons are absorbed by Yb³⁺, causing the ²F_7/2→²F_5/2 transition of Yb³⁺. Subsequently, Er³⁺ is excited to its ⁴I_7/2 state from ⁴I_15/2 state via two sequential energy transfer (ET1 and ET2) processes. The Er³⁺ at ⁴F_7/2 state is then relaxed rapidly to the neighboring lower ⁴F_9/2 state. Furthermore, Er³⁺ at ⁴F_9/2 state can also be populated via another energy transfer (ET3) process. Finally, the ⁴F_9/2→⁴I_15/2 transition of Er³⁺ occurs and produces the red UC emission.

 

Fig. 1 Room-temperature UC spectrum of Er³⁺ in CaWO₄:Yb³⁺/Er³⁺ phosphor in the range of 620 to 700 nm; the inset shows the log-log plots of two fluorescence peak intensities as a function of the pump power.

Download Full Size | PDF

 

Fig. 2 The energy levels diagram of Yb³⁺ and Er³⁺ in CaWO₄:Yb³⁺/Er³⁺ phosphor and the corresponding UC processes.

Download Full Size | PDF

Figure 3 shows the temperature-dependent UC emission spectra of Er³⁺ in CaWO₄:Yb³⁺/Er³⁺ phosphor. They were measured at 298, 398 and 478 K, respectively. All spectra were normalized to the emission intensity of right fluorescence peak at 669 nm. As presented in Fig. 3, the left fluorescence peak increases markedly when the temperature is varied from 298 to 398 and then to 478 K. In other words, the FIR between left and right fluorescence peaks increases gradually with increasing the temperature, suggesting undoubtedly that the left and right fluorescence peaks respectively come from the higher and lower Stark sublevels of ⁴F_9/2 state to the same terminal ⁴I_15/2 ground state of Er³⁺. Besides, these two Stark sublevels of ⁴F_9/2 state were confirmed to be TCL, the population between which followed the Boltzmann distribution. In brief, due to the small energy gap between TCL, the lower emitting state would be thermally coupled to higher emitting state of TCL, causing the FIR between TCL to increase gradually. The FIR between TCL can be written as [8]:

 

Fig. 3 Temperature-dependent red UC emission spectra of Er³⁺ in CaWO₄:Yb³⁺/Er³⁺ phosphor. All spectra were normalized to the emission intensity of the right fluorescence peak.

Download Full Size | PDF

Meanwhile, it should be noted that with the rise of temperature, the population between two fluorescence peaks, which was controlled by the Boltzmann distribution, caused the emission intensity of valley at 661 nm to increase gradually. In addition, the valley and peak positions of the spectrum hardly changed throughout the experiment, which is advantageous to the stability of the VPR method for the detection of temperature in reality. The detailed changes of the VPR, as well as the FIR of the double-peak spectrum as a function of the temperature are given in the following part.

Figure 4(a) shows the data points of FIR and VPR as a function of temperature in the range of 298 to 478 K. Several test cycles were carried out and a good repeatability was achieved for both two methods. As shown in Fig. 4(a), the data points of FIR can be fitted well with Eq. (1), and the data points of VPR increase linearly with temperature. What is noteworthy is that there is slight difference between two methods. In the same temperature range, FIR is varied from 1.08 to 1.28 with the growth rate of 18.5%; while VPR is varied from 0.43 to 0.59 with the growth rate of 37.2%. As a consequence, the VPR method is more sensitive to the change of temperature under the same conditions. This is closely associated with the parameter of the relative sensitivity (Sr). For two different methods for the detection of temperature, it is reasonable to make a comparison of Sr [14,15], which can be expressed as:

 

Fig. 4 (a) FIR and VPR of double-peak spectrum as a function of temperature in the range of 298 to 478 K; (b) the comparison of Sr between the FIR and VPR two methods.

Download Full Size | PDF

Furthermore, we also investigated the anti-jamming capability of the VPR method in detail, and the results are shown in Fig. 5. It can be seen from Fig. 5 that in three different pump powers of 60, 90 and 120 mW, each group of three data points of the VPR keep consistent at every given temperature, thus making three fitting lines maintain consistent in the temperature range of 298 to 478 K. In addition, the study on the power dependence over a wide range was also performed. Taking the signal to noise ratio and the laser-induced heating effect into account, the appropriate pump power should greater than 45 mW but less than 180 mW. These experimental facts suggest that the VPR method is immune to the pump power over a wide range and thus appropriate for monitoring the temperature.

 

Fig. 5 The VPR of the double-peak spectrum as a function of the temperature in the range of 298 to 478 K. They were recorded in three different pump powers of 60, 90 and 120 mW, respectively.

Download Full Size | PDF

4. Conclusion

In conclusion, we successfully developed a more advanced and convenient VPR method based on the UC emission. Upon 976 nm diode laser excitation, the red UC emission of Er³⁺ in CaWO₄:Yb³⁺/Er³⁺ phosphor was studied as a function of the temperature from 298 to 478 K. This double-peak spectrum was found to consist of two Stark components originating from two Stark sublevels of ⁴F_9/2 excited state to ⁴I_15/2 ground state of Er³⁺. These two Stark sublevels of ⁴F_9/2 state were confirmed to be TCL and their populations accorded well with the Boltzmann distribution. The VPR of this double-peak spectrum was verified to increase linearly with the rise of temperature, which could thus be used to measure the temperature. The maximum relative sensitivity of the VPR method was obtained to be about 0.20% K⁻¹ at 298 K. The VPR method was proven to be independent of the pump power over a wide range. This excellent feature suggests that besides the FIR method, the VPR method is also a promising candidate for temperature measurement in practice.