Luminescent Nanothermometer Operating at Very High Temperature—Sensing up to 1000 K with Upconverting Nanoparticles (Yb³⁺/Tm³⁺)

Abstract

Lanthanide-based luminescent nanothermometers play a crucial role in optical temperature determination. However, because of the strong thermal quenching of the luminescence, as well as the deterioration of their sensitivity and resolution with temperature elevation, they can operate in a relatively low-temperature range, usually from cryogenic to ≈800 K. In this work, we show how to overcome these limitations and monitor very high-temperature values, with high sensitivity (≈2.1% K^–1) and good thermal resolution (≈1.4 K) at around 1000 K. As an optical probe of temperature, we chose upconverting Yb³⁺–Tm³⁺ codoped YVO₄ nanoparticles. For ratiometric sensing in the low-temperature range, we used the relative intensities of the Tm³⁺ emissions associated with the ³F_2,3 and ³H₄ thermally coupled levels, that is, ³F_2,3 → ³H₆/³H₄ → ³H₆ (700/800 nm) band intensity ratio. In order to improve sensitivity and resolution in the high-temperature range, we used the 940/800 nm band intensity ratio of the nonthermally coupled levels of Yb³⁺ (²F_5/2 → ²F_7/2) and Tm³⁺ (³H₄ → ³H₆). These NIR bands are very intense, even at extreme temperature values, and their intensity ratio changes significantly, allowing accurate temperature sensing with high thermal and spatial resolutions. The results presented in this work may be particularly important for industrial applications, such as metallurgy, catalysis, high-temperature synthesis, materials processing and engineering, and so forth, which require rapid, contactless temperature monitoring at extreme conditions.

1. Introduction

Temperature is a crucial parameter for most of industrial processes, like sintering, formation of metal alloys, catalytic reactions, formation of new materials under extreme conditions, and so forth. Hence, its rapid, accurate, and online monitoring is a challenging task for many specialists working in various fields of science, industrial researchers, and material engineers.¹⁻⁶ For these purposes, various luminescence thermometry techniques have been proposed, developed, and applied.⁷⁻¹⁸ However, in general, because the luminescence of materials is significantly quenched at increasing temperature, these optical methods are usually limited to low- (cryogenic to biological; below ≈350 K) and mild-temperature (≈400–800 K) ranges.¹⁹⁻²²

The most commonly applied materials in luminescence thermometry are lanthanide-based inorganic compounds, such as fluorides, oxides, vanadates, and phosphates.²⁰⁻²⁴ These matrices are thermally stable and they have relatively low phonon energies, allowing development of highly efficient phosphors and exhibiting a strong photoluminescence signal.²⁵⁻²⁷ On the other hand, lanthanide ions (Ln³⁺) are optically active species, that is, dopant ions acting as sensitizers (predominantly Ce³⁺ and Yb³⁺) and luminescence activators (e.g., Eu³⁺, Tb³⁺, Nd³⁺, Er³⁺, Tm³⁺).^13,24−29 This is because of their beneficial and unique spectroscopic features, such as narrow emission bands, long luminescence lifetimes, variety of emitting states, large Stokes shift, and the possibility of converting lower energy photons to higher energy ones, known as upconversion (UC) process, that is, anti-Stokes emission.^{20−27,30−33}

Because of the abovementioned benefits, luminescence thermometry is dominated by Ln³⁺-based optical sensors, in which sensing is realized predominantly by the use of ratiometric techniques (band intensity ratio) and to a lesser extent by luminescence lifetimes.²⁰⁻²³ For ratiometric sensing, the most commonly utilized ions are Er³⁺, Tm³⁺, and Nd^{3+14,17,20,26,34−36} because their thermally coupled levels (TCLs) are located in the visible and NIR spectral ranges.^22,34−37 Such excited TCLs are separated by an energy gap ΔE ≈ 50–2000 cm^–1 to provide efficient thermalization in the measured temperature range.^20−22,38 In such case, the principle of sensing is associated with the thermal relocation of the excited electron to another excited state (with a slightly higher energy), and temperature readout is correlated with changes in the band intensity ratio observed in the spectra measured as a function of temperature.²⁷ On the other hand, the use of upconverting nanomaterials is commonly proposed for biological applications (subcutaneous sensing) because of their small size and the use of NIR laser as an excitation source, eliminating harmful UV radiation.³⁹⁻⁴⁷ However, nanomaterials exhibiting UC phenomena, and excited in the NIR range, are also beneficial in industrial applications. This is because they allow high-spatial resolution of temperature sensing, as well as they do not need a high-energy excitation source, which can cause other undesired effects, such as uncontrolled polymerization during catalytic processes or additional heating.^1,27,48−50 It is worth noting about novel application of lanthanide-based luminescent thermometers in optical vacuum sensing, that is, conversion of luminescent thermometers into low-pressure sensors, utilizing the effect of laser-induced heating of the materials, which is enhanced under vacuum conditions.^27,51

Here, for the first time, we present a very sensitive and accurate way of monitoring very high-temperature values, up to ≈1000 K, using the Yb³⁺–Tm³⁺ codoped YVO₄ upconverting nanomaterial. Temperature sensing is realized by the use of both TCLs and non-TCLs of Tm³⁺ and Yb³⁺ ions, located in the red and NIR spectral ranges. By the use of the band intensity ratio associated with non-TCLs of Yb³⁺ and Tm³⁺ ions (940/800 nm), we have achieved unprecedentedly high sensitivity (2.1% K^–1) and temperature resolution (1.4 K) at extreme temperature values.

2. Experimental Section

2.1. Synthesis of YVO₄:Yb³⁺ 20 mol %–Tm³⁺ 0.5 mol % NPs

A typical synthesis was performed per 1 g of the final product. The starting materials Y₂O₃, Yb₂O₃, and Tm₂O₃ (Stanford Materials, USA, 99.99%) were dissolved in an excess of hydrochloric acid (Sigma-Aldrich, ACS 37%). The solutions were evaporated three times to remove the excessive HCl. The obtained YCl₃, YbCl₃, and TmCl₃ solutions were diluted with deionized water to prepare 0.5 M solutions. Then, the stoichiometric amounts (calculated based on the formula Y_0.795Yb_0.2Tm_0.005VO₄) of appropriate YCl₃, YbCl₃, and TmCl₃ solutions were mixed with a proper amount (1:1 molar ratio, with regard to Y_0.795Yb_0.2Tm_0.005VO₄) of trisodium citrate dihydrate (Chempur, >99%). Afterward, stoichiometric amount of NH₄VO₃ (Sigma-Aldrich, ACS >99%) was dissolved in water, and then, the appropriate amount (1:1 molar ratio) of NaOH (POCH, pure p.a. 98.8%) was added to the system to obtain a transparent yellow solution. The as-prepared NH₄VO₃ solution was added dropwise (at room temperature) to the mixed lanthanide chloride solution, maintaining vigorous stirring. Finally, deionized water was added to the above mixture to get the total volume of 40 mL, and pH = 10 of the system was adjusted with 1.5 M solution of NaOH, upon stirring. The obtained yellow mixture was transferred to the hydrothermal autoclave (Berghof DAB2) and heated at 473 K for 20 h, under a pressure of 30 bar. The synthesized nanoparticles (NPs) were collected and purified by centrifugation and washing five times with water/ethanol mixture. The obtained material was dried at 353 K for 24 h in the air atmosphere and then ground in an agate mortar. Finally, the sample was annealed at 1123 K for 2 h to improve crystallinity and luminescence signal of the material and, after this, again ground in a mortar. The elemental composition of the final product was determined using inductively coupled plasma optical emission spectroscopy, resulting in the formula YVO₄:Yb³⁺ 19.61 mol %–Tm³⁺ 0.52 mol %.

2.2. Characterization

Powder diffractogram was measured using a Bruker AXS D8 ADVANCE diffractometer, with Cu Kα radiation (λ = 0.15406 nm). Transmission electron microscopy (TEM) image was taken with a Hitachi HT7700 transmission electron microscope, operating at 100 kV accelerating voltage. Emission spectra were measured with an Andor Shamrock 500 spectrometer coupled with the silicon CCD camera and corrected for the apparatus response. The excitation source was a tunable continuous wave Ti:Sapphire laser (Spectra Physics 3900-S, pumped with a 15 W 532 nm Spectra Physics Millennia), which was set to 975 nm. The beam was focused to a spherical spot of about ≈0.5 mm, and the output power was adjusted to ≈200 mW.

3. Results and Discussion

3.1. Properties at Ambient Conditions

All reflexes in the diffractogram of the measured powder for the synthesized material, YVO₄:Yb³⁺ 20 mol %–Tm³⁺ 0.5 mol % (Figure S1), fit well the reference pattern from the Inorganic Crystal Structure Database (ICDD) of tetragonal YVO₄, crystallizing in a space group I4₁/amd (card no. #2504), confirming the successful synthesis of the vanadate structure, and indicating on the substitution of Y³⁺ ions by Yb³⁺ and Tm³⁺ ions in the crystal lattice, where the local point symmetry for the metal ions is D_2h. The broadening of the observed reflexes is associated with the small (nano) size of the synthesized particles. The TEM image of the obtained product (Figure 1a) reveals that it is composed of nearly spherical, nonuniform, and agglomerated NPs, with sizes in the range from 50 to 100 nm.

Figure 1.

(a) TEM image, (b) UC emission spectrum, and (c) a simplified energy-level diagram for the synthesized YVO₄:Yb³⁺–Tm³⁺ NPs; inset (b) presents a photograph of the sample under 975 nm laser irradiation at ambient conditions.

Figure 1b presents the UC emission spectrum for the prepared nanomaterial, exciting with a laser radiation at 975 nm. The spectrum consists of several sharp bands, characteristic of Tm³⁺ ions, located in the spectral range of 450–850 nm, and one band above 900 nm typical of Yb³⁺ ions, which is partially cutoff by the short-pass 950 nm filter applied. All observed bands are associated with 4f–4f intraconfigurational radiative transitions, that is, anti-Stokes UC emission of Tm³⁺: ¹G₄ → ³H₆ (≈470 nm), ¹G₄ → ³F₄ (≈640 nm), ³F_2,3 → ³H₆ (≈700 nm), ³H₄ → ³H₆ (≈800 nm), and phonon-assisted emission of Yb³⁺: ²F_5/2 → ²F_7/2 (>900 nm).^26,34,52,53 In this system, Yb³⁺ acts as a sensitizer, which transfers absorbed NIR light to the neighboring Tm³⁺ (UC emitter), pumping its excited states mainly via energy transfer UC processes.^26,27 A schematic, simplified representation of the main radiative and nonradiative processes occurring in the studied system is presented in the energy level diagram in Figure 1c.

3.2. Properties at High-Temperature Conditions

First, it is worth noting that we have chosen YVO₄ as a host for the optically active lanthanide ions (Yb³⁺ and Tm³⁺), mainly because of its resistance to high-temperature treatment; possibility of codoping with different lanthanide ions; facile synthesis under high-temperature conditions, resulting in a well-crystallized, nanosized product, which exhibits intense and bright luminescence. Temperature has a significant impact on the absolute and relative intensities of the observed UC emission bands of Tm³⁺ ions and the NIR emission band of Yb³⁺ ions (see Figure 2a). Because of the thermal quenching processes, as the temperature increases, the intensity of Tm³⁺ bands gradually decreases, except of the thermalized band located around 700 nm, whose intensity increases as a function of temperature. A similar effect of decreasing intensity is observed for the Yb³⁺ band. However, the rate of this decrease is much lower compared to the Tm³⁺ UC emission bands. Hence, this evident difference in the quenching rates will be further used for temperature-sensing purposes. The initial slight increase in intensity of the Yb³⁺ band is plausibly caused by its phonon-assisted character (non-UC emission, below excitation wavelength).⁵³ Please note that, despite the fact, phonon-assisted transitions are favored by temperature, their intensity may vary as they compete with other thermal quenching processes.⁵³ This is why, in general, the use of phonon-assisted bands together with UC emission bands for ratiometric temperature sensing should result in high sensitivity of the LIR-based luminescent thermometers.⁵³

Figure 2.

(a) UC emission spectra for the synthesized YVO₄:Yb³⁺–Tm³⁺ NPs recorded at different temperature values; λ_ex = 975 nm; (b,c) determined luminescence intensity ratios—LIRs for (b) TCLs of Tm³⁺ (700/800 nm) and (c) non-TCLs of Yb³⁺ and Tm³⁺ (940/800 nm) as a function of temperature; inset (a) presents a photograph of a sample under 975 nm laser irradiation, taken at ≈1000 K, clearly showing the dominant black body emission from the sample and the environment.

From the point of view of temperature sensing, an important conclusion can be drawn analyzing the spectra, namely, the use of Tm³⁺ bands centered at ≈470 and ≈640 nm becomes pointless above around 800 K, as the intensity of the corresponding transitions (¹G₄ → ³H₆ and ¹G₄ → ³F₄, respectively) decreases to the noise level (see Figures 2a, S2 and S3 in the Supporting Information). This is why, as this work is focused on sensing in a very high-temperature range, we have focused on two diverse band intensity ratios, corresponding to the thermally coupled levels (TCLs) of Tm³⁺ (700/800 nm; ³F_2,3 → ³H₆/³H₄ → ³H₆) and the non-TCLs of Yb³⁺ and Tm³⁺ (940/800; ²F_5/2 → ²F_7/2/³H₄ → ³H₆).

However, as expected, the spectra measured above ≈800 K have a significant contribution of black body radiation (see Figure S3), that is, thermal emission from the sample and the environment (furnace tube, measuring holder, etc.). This is why, before starting thermometric evaluation of the luminescent nanomaterial, the raw data were processed in order to remove the background thermal emission. In order to do this, we recorded also the spectra of pure black body emission (Figure S4) at adjusted temperature values (laser turned off), just before the luminescence measurement, and then, we simply subtracted the undesired background emission. Finally, to correctly determine the band intensity ratios, the spectra were converted to the energy scale (eV) via the Jacobian transformation (Figure S5), ensuring good reliability of data and the applied fittings.⁵⁴⁻⁵⁶

The thermalization processes between the ³F_2,3 and ³H₄ excited levels of Tm³⁺ are well-established in the literature.^26,34 These TCLs are separated by relatively large energy (usually ΔE ≈ 1500–2000 cm^–1), resulting in effective thermalization in a high-temperature range and good relative sensitivity.^26,34 As the temperature increases, the higher energy I₂ band (700 nm) increases, and the intensity of the lower energy I₁ band (800 nm) decreases, according to the Boltzmann-type distribution

1

where LIR is the luminescence intensity ratio of these UC emission bands; I₁ and I₂ are the integrated intensities of the bands; ΔE is the energy difference between the barycenters of the I₂ and I₁ bands; k_B is the Boltzmann constant; T is the absolute temperature; and B is a constant dependent on rates of total spontaneous emission, branching ratio of the transitions in respect to the ground state, state degeneracies, and transition angular frequencies.^20,35,57 Applying eq 1, we correlated the determined 700/800 nm band intensity ratio with temperature, resulting in ΔE = 1777 cm^–1 (which is similar to the value derived from the spectra, i.e., 1825 cm^–1), B = 1.491, and R² = 0.999 (Figure 2b). The calculated LIR parameter increases more than 100-times in the measured temperature range, that is, from ≈0.001 up to 0.125.

On the other hand, the second LIR parameter, associated with non-TCLs of Yb³⁺ and Tm³⁺, corresponding to the 940/800 nm band intensity ratio, was correlated with temperature applying the following exponential function

2

where A₁, A₂, A₃, C₁, C₂, and C₃ are fitting constants. The presented empirical function was applied because of the absence of an appropriate physical model conforming the observed, temperature-dependent changes in the LIR parameter, associated with the non-TCLs used. Please note that a similar empirical, exponential function was applied by Brites et al.,¹² in order to correlate LIR of the non-TCLs of Pr³⁺ with temperature in the cryogenic range. We have obtained very good fitting results (see Figure 2c), with R² = 0.999, using the following values of the fitting parameters: LIR = −2.339 × 10^–5 exp(T/55.05) + 4.110 × 10^–3 exp(T/104.3) + 1.644 × 10^–5 exp(T/54.03) + 0.1732. In this case, the determined LIR parameter increases approximately 200-times, that is, from ≈0.25 up to 50. Both determined LIR parameters change reversibly with the temperature, confirming the reliability of the sensing methods used (see Figure S6).

In the next step, in order to quantitatively compare the sensing performance of the developed nanothermometer, the relative temperature sensitivities, S_r (% K^–1), were calculated. The S_r value shows how the spectroscopic parameter is measured, here the LIR changes per 1 K of the absolute temperature, and is calculated as follows

3

Figure 3 shows how the S_r values change with temperature, both for TCLs (a) and non-TCLs (b). It is clear that for sensing in the low-temperature range, it is more beneficial to use TCLs of Tm³⁺ (700/800 nm) because of high S_r values, for example, ≈2.9% K^–1 at 300 K, compared to relatively low S_r values for sensing with non-TCLs of Yb³⁺/Tm³⁺ (940/800 nm), that is, ≈0.27% K^–1 at 300 K. However, as the temperature increases, the S_r values significantly decrease for the 700/800 nm band intensity ratio (S_r ≈ 0.25% K^–1 at 1000 K), as commonly observed for TCLs of various Ln³⁺ ions.^{15,20,26,27,35,36,58} Fortunately, by using the 940/800 nm band intensity ratio, it is possible to preserve high sensitivity of temperature sensing because for these non-TCLs, S_r increases with the temperature, being around 2.1% K^–1 above 1000 K.

Figure 3.

(a,b) Determined relative temperature sensitivities (S_r) and (c,d) temperature resolutions (δT) for (a,c) TCLs of Tm³⁺ (700/800 nm) and (b,d) non-TCLs of Yb³⁺ and Tm³⁺ (940/800 nm) as a function of temperature.

To further examine the performance of the developed high-temperature sensor, we have calculated the temperature resolution (uncertainty; δT) associated with both band intensity ratios used, applying eq 4

4

where δLIR is the uncertainty of the determined LIR value, calculated according to the formula

5

where δI₁ and δI₂ are the uncertainties of determination of the corresponding band intensities (noise-to-signal ratio, i.e., integrated intensity of the noise divided by the integrated intensity of the band). Initially, at room temperature (300 K), the determined temperature resolutions are very good, that is, δT ≈ 0.43 and 0.15 K for TCLs and non-TCLs used, respectively. Typically, as a result of signal-to-noise deterioration, caused by thermal quenching of luminescence, the uncertainties of luminescence sensing increase with temperature elevation. As expected, with increasing temperature, the calculated uncertainties for TCLs and non-TCLs increase, for example, δT ≈ 2.4 and 0.31 K at 900 K. With further temperature elevation, δT rapidly increases, up to above 20–25 and 1.3–1.4 K at extreme temperature values (above 1000 K) for TCLs and non-TCLs, respectively. It is worth noting that the good temperature resolution for sensing with non-TCLs (940/800 nm) preserved at extreme temperature values is thanks to their high S_r values, as well as decent signal-to-noise levels (small δLIR) associated with Yb³⁺ (²F_5/2 → ²F_7/2) and Tm³⁺ (³H₄ → ³H₆) transitions located in the NIR spectral region.

In addition to the theoretical thermal resolution calculated based on the relative sensitivities of the thermometer (using eq 4), we have determined also the experimental temperature resolution of the sensors, according to the procedure reported by Savchuk et al.(58) This method takes into account the experimental errors, which may affect the sensing readouts and bias the determined temperature values. In order to obtain the mentioned experimental resolution, we have collected a series of fifty spectra recorded at the highest measured temperature value (≈1009 K). Afterward, both LIR parameters were determined, and using the calibrations curves, as shown in Figure 2b,c, the corresponding temperature values were plotted, as shown in Figure S7, as histograms. The resulting data were fitted with Gauss function, and the obtained full-width at half maximum of the peaks is equal to the resolution of the sensor. The obtained experimental temperature resolutions are similar (slightly worse) compared to the theoretical values, that is, 32 versus 28 K (TCLs) and 1.7 versus 1.4 K (non-TCLs) at ≈1009 K.

In order to compare the high-temperature performance of various lanthanide-based, luminescent thermometers, we summarized the maximum relative temperature sensitivities (S_r max) and S_r at the highest recorded/detectable temperature values for the luminescent thermometers operating in the high-temperature range, as shown in Table 1. We agreed the criterion of the temperature threshold, starting from 600 K, which seems to be rationale for such comparison, as our studies concern sensing in a very high-temperature range. In the moderate temperature range, around room temperature, it is beneficial to use TCLs of Tm³⁺ to obtain very good S_r value (2.86% K^–1 at 300 K) compared to available literature data. In the presented comparison, there are only few reports showing significantly higher S_r max values (up to 9% K^–1), however, these values were obtained at very low (cryogenic) temperature values. Obviously, there are some other reports showing luminescent thermometers with similar or higher S_r values but they operate in a low- or mild-temperature range.^17,53 On the other hand, there is no doubt that in a very high-temperature range, we achieved the highest S_r (2.13% K^–1 at 1009 K), compared to other reports, where sensitivity significantly decreases with temperature, and the S_r values are very low in the range of 800–1000 K, that is, usually around 0.3–0.1% K^–1 (see Table 1). Please note that the temperature resolution (δT) is not given in the presented comparison because the vast majority of researchers do not report on this parameter, especially in the high-temperature region. However, as δT depends on S_r and signal-to-noise ratio, we may assume that at present, temperature resolution (at ≈800–1000 K) of our developed sensor is plausibly the best among the available luminescent thermometer, operating in a high-temperature range.

Table 1. Maximum Relative Temperature Sensitivities (S_r max) and S_r at the Highest Measured Temperature (T_max) for Different Lanthanide-Doped Inorganic Materials.

dopant ions	host	Sr max (% K–1)	T (K)	T-range (K)	Sr at Tmax (% K–1)	transitions	λ (nm)	refs
Yb3+–Tm3+	YVO4	2.13	1009	300–1009	2.13	2F5/2 → 2F7/2/3H4 → 3H6	940/800	this work
		2.86	300		0.25	3F2,3 → 3H6/3H4 → 3H6	700/800
Yb3+–Nd3+	La2O3	1.45	290	290–1230	0.10	4F7/2 → 4I9/2/4F5/2 → 4I9/2	762/825	(59)
Yb3+–Er3+	Gd2O3-AuNPs	0.72	423	423–1050	0.14	2H11/2 → 4I15/2/4S3/2 → 4I15/2	523/560	(14)
Yb3+–Er3+	Al2O3	0.37	500	295–973	0.10	2H11/2 → 4I15/2/4S3/2 → 4I15/2	523/545	(60)
Mo–Er3+	Yb3Al5O12	0.43	467	300–973	0.08	2H11/2 → 4I15/2/4S3/2 → 4I15/2	522/546	(61)
Yb3+–Ho3+	CaWO4	0.50	923	303–923	0.50	5F1,5G6 → 5I8/5F2,3,3K8 → 5I8	460/487	(62)
Yb3+–Er3+	Gd2O3	0.83	300	300–900	0.20	2H11/2 → 4I15/2/4S3/2 → 4I15/2	523/548	(63)
Yb3+–Er3+	NaYF4@SiO2	1.02	300	300–900	0.13	2H11/2 → 4I15/2/4S3/2 → 4I15/2	520/550	(50)
Nd3+	CaWO4	0.27	730	303–873	0.19	4F5/2 → 4I9/2/4F3/2 → 4I9/2	805/872	(64)
		0.37	873	303–873	0.37	4F7/2 → 4I9/2/4F5/2 → 4I9/2	755/805
Nd3+	YVO4	9.00	123	123–873	0.18	4F5/2 → 4I9/2/4F3/2 → 4I9/2	808/880	(65)
Cr3+–Nd3+	Y3Al5O12	1.80	195	100–850	0.05	4F5/2, 2H9/2 → 4I9/2/4F3/2 → 4I9/2	810/895	(66)
Mn3+/4+–Nd3+	Y3Al5O12	1.80	190	123–823	0.30	5T2 → 5E/4F3/2 → 4I9/2	575/885	(67)
Dy3+	BaYF5	2.00	270	270–800	0.25	4I15/2 → 6H15/2/4F9/2 → 6H15/2	410/510	(68)
		1.80	270	270–800	0.24	4I15/2 → 6H15/2/4F9/2 → 6H11/2	410/460
Eu3+	GdAlO3	2.96	293	293–793	0.48	5D1 → 7F1/5D0 → 7F2	555/614	(69)
Yb3+–Er3+	Y2O3	2.05	303	303–783	0.32	4G11/2 → 4I15/2/2H9/2 → 4I15/2	387/410	(70)
Yb3+–Ho3+	β-NaLuF4	1.60	390	390–780	0.38	5F1/5G6 → 5I8/5F2,3/3K8 → 5I8	443/482	(71)
Yb3+–Tm3+	LaPO4	3.00	293	293–773	0.50	3F2,3 → 3H6/3H4 → 3H6	700/800	(26)
Yb3+–Tm3+	YPO4	2.40	293	293–773	0.48	3F2,3 → 3H6/3H4 → 3H6	700/800	(26)
Eu3+	YNbO4	2.70	303	303–773	0.40	5D1 → 7F1/5D0 → 7F1	538/595	(72)
Sm3+	YNbO4	0.43	500	303–773	0.32	4F3/2 → 3H5/2/4G5/2 → 6H5/2	531/566	(72)
Yb3+–Tm3+	YF3	0.43	750	300–750	0.43	3F2,3 → 3H6/3H4 → 3H6	700/776	(34)
Yb3+–Nd3+	LiLaP4O12	0.40	268	100–700	0.05	4F3/2 → 4I9/2/2F5/2 → 2F7/2	870/1000	(73)
Yb3+–Er3+	NaYF4	1.57	260	298–693	0.20	2H11/2 → 4I15/2/4S3/2 → 4I15/2	523/548	(74)
Yb3+–Er3+	YNbO4	0.9	390	298–673	0.30	2H11/2 → 4I15/2/4S3/2 → 4I15/2	525/545	(75)
Dy3+	YVO4	1.80	298	298–673	0.38	4I15/2 → 6H15/2/4F9/2 → 6H15/2	455/480	(76)
Yb3+–Ho3+	KLu(WO4)2	0.54	297	297–673	0.02	5F5 → 5I8/5S2, 5F4 → 5I8	540/650	(77)
Nd3+	NaYF4	1.20	323	323–673	0.40	4F5/2 → 4I9/2/4F3/2 → 4I9/2	803/864	(78)
		1.38	323	323–673	0.40	4F7/2 → 4I9/2/4F5/2 → 4I9/2	740/803
		2.70	323	323–673	0.70	4F7/2 → 4I9/2/4F3/2 → 4 I9/2	740/864
Yb3+–Tm3+	GdVO4	0.20	673	297–673	0.20	1G4 → 3H6, 3F3 → 3H6	RGB ratio (red/blue)	(79)
Mn2+–Eu3+	Zn2SiO4:Mn2+/	3.05	303	303–623	0.50	4T1 → 6A1/5D0 → 7F2	520/612	(80)
	Gd2O3:Eu3+
Nd3+	YAlO3 perovskite	1.83	293	293–611	0.41	2H9/2, 4F5/2 → 4I9/2/4F3/2 → 4I9/2	820/890	(15)
Pr3+	Sr2GeO4	9.00	22	17–600	0.50	(<300 K) 4f15d1 → 3H4–6, 3F2–4/3P0 → 3H6, 1D2 → 3H4	300/620	(12)
						(>300 K) 3P0 → 3H4/3P0 → 3H6, 1D2 → 3H4	490/620

It is worth to discuss more about two other studies (refs (14) and (59)) showing temperature sensing above 1000 K, reported by Debasu et al. and Gao et al.(14,59) In the first report, the authors synthesized the Gd₂O₃:Yb³⁺–Er³⁺ nanorods decorated with Au NPs (luminescent-plasmonic nanomaterials), working as optical heaters/thermometers.¹⁴ The authors used two different thermometric parameters, that is, the commonly used LIR of Er³⁺ (525/550 nm—TCLs) for the T-range 300–1050 K and the black body emission of the nanomaterial (shift/change of the spectral profile according to Planck’s law) for the T-range 1200–2000 K, observed because of the laser heating of the sample, which was significantly enhanced with plasmonic gold NPs. The relative sensitivity around 1000 K of this sensor was one order of magnitude lower (see Table 1), and the reported temperature resolution was similar (≈2 K), compared to our sensor (non-TCLs). However, the authors used varied pump power during the measurements (excitation power increased with temperature because temperature elevation was induced with laser heating of the sample), leading to the improved signal at extreme temperature. Please note that typical temperature sensing experiments require measurements at fixed pump power. Moreover, the reported nanosensor is composed of two different phases (heterostructures), hence its synthesis is much more complex and difficult, compared to our nanothermometer. In the second report, the authors synthesized La₂O₃:Yb³⁺–Nd³⁺ micron-sized crystals (lower spatial resolution, compared to nano-sized sensors), working as a luminescent thermometer in the T-range 290–1230 K.⁵⁹ For temperature sensing, the authors used also ratiometric approach, that is, TCLs of Nd³⁺ (LIR 762/825 nm), resulting in resolution of about 3 K at the high-T range and very low sensitivity (≈0.1% K^–1).

4. Conclusions

We have shown the possibility of highly sensitive and accurate determination of very high temperature (up to above 1000 K) via luminescence thermometry. This was possible applying a ratiometric technique and non-TCLs of lanthanide ions (Yb³⁺ and Tm³⁺). As a luminescent probe, we have used the synthesized inorganic nanomaterial, composed of YVO₄:Yb³⁺–Tm³⁺ NPs (<100 nm), exhibiting simultaneous UC luminescence of Tm³⁺ and intense emission of Yb³⁺ located in the NIR spectral region. We have tested two approaches, using TCLs of Tm³⁺ (700/800 nm) and non-TCLs of Yb³⁺/Tm³⁺ (940/800 nm). Both determined LIR parameters revealed good temperature resolution up to ≈800 K. However, at extreme temperature values, the temperature resolution for TCLs significantly decreases, that is, δT ≈ 25–30 K at around 1000 K, whereas for non-TCLs, the resolution was one order of magnitude higher, namely, δT ≈ 1.3–1.4 K. Such a good temperature resolution was achieved because of the high sensitivity (S_r ≈ 2.1% K^–1) and a reasonable signal-to-noise ratio, preserved up to about 1000 K for the non-TCLs used, that is, 940/800 nm band intensity ratio, corresponding to the radiative transitions ²F_5/2 → ²F_7/2 and ³H₄ → ³H₆ of Yb³⁺ and Tm³⁺. Such a highly sensitive method for accurate monitoring of high temperature and the developed optical thermometer could be used, for example, in industrial applications, for processes requiring very high-temperature conditions, as well as submicron-sized resolution.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsami.0c13011.

• XRD pattern; UC emission spectra; thermal emission; cycling of the LIR parameter; and experimental temperature resolution histograms (PDF)

The authors declare no competing financial interest.