Lanthanide Upconverted Luminescence for Simultaneous Contactless Optical Thermometry and Manometry–Sensing under Extreme Conditions of Pressure and Temperature

Abstract

The growing interest in the miniaturization of various devices and conducting experiments under extreme conditions of pressure and temperature causes the need for the development of small, contactless, precise, and accurate optical sensors without any electrical connections. In this work, YF₃:Yb³⁺-Er³⁺ upconverting microparticles are used as a bifunctional luminescence sensor for simultaneous temperature and pressure measurements. Different changes in the properties of Er³⁺ green and red upconverted luminescence, after excitation of Yb³⁺ ions in the near-infrared at ∼975 nm, are used to calibrate pressure and/or temperature inside the hydrostatic chamber of a diamond anvil cell (DAC). For temperature sensing, changes in the relative intensities of the Er³⁺ green upconverted luminescence of ²H_11/2 and ⁴S_3/2 thermally coupled multiplets to the ⁴I_15/2 ground state, whose relative populations follow a Boltzmann distribution, are calibrated. For pressure sensing, the spectral shift of the Er³⁺ upconverted red emission peak at ∼665 nm, between the Stark sublevels of the ⁴F_9/2 → ⁴I_15/2 transition, is used. Experiments performed under simultaneous extreme conditions of pressure, up to ∼8 GPa, and temperature, up to ∼473 K, confirm the possibility of remote optical pressure and temperature sensing.

Introduction

In the past, nowadays, and probably in the future, experiments performed simultaneously as a function of both pressure and temperature had been, are, and will be crucial for a considerable number of scientists and material engineers.¹⁻⁶ This is mainly because of the fundamental importance of these physical quantities that affect the physical, chemical, and biological properties of materials and living organisms.² On the other hand, most of the materials existing in the Earth and in space were formed under extreme conditions of high pressure and temperature.⁷ That is why researchers have tried to reproduce such extreme conditions occurring during the formation of stars and planets on a laboratory scale, to understand the mechanisms governing these processes, as well as to obtain new materials with the desired properties and functionalities.^3−6,8−14 However, to imitate and simulate such conditions on a laboratory scale, namely, to achieve very high pressure values, from several tens to even hundreds of GPa, it is necessary to generate a very high force acting per small area.¹⁻³

Nowadays, the most commonly used device to generate high pressures, which allows at the same time for structural and spectroscopic measurements, is a diamond anvil cell (DAC), which was mainly developed by Piermarini et al.,¹⁵ Bassett et al.,¹⁶ and Barnett,¹⁷ allowing the generation of pressure up to several 100 GPa.¹ The principle of operation of the DAC is based on the use of two diamond anvils, to squeeze and compress the investigated material or new material precursors. Such diamond anvils are transparent, are ultrahard, and have a very small culet, contact surface with the sample, and their use allows the generation of very high force and simultaneous observation of the effect of compression (pressure) on the properties of the material.¹⁻³

A very important issue during the high-pressure measurements is the determination of the pressure value in the DAC’s hydrostatic chamber. For technical reasons, i.e., the small size of the pressure chamber, typically with a diameter of about 100 μm and a height of 50 μm, the values of pressure are monitored remotely using the optical detection of the luminescence at about 700 nm of a small piece of ruby (Al₂O₃:Cr³⁺) or Sm²⁺-doped SrB₄O₇.¹⁷⁻²⁴ Such materials can be easily excited by UV or visible light, generating a strong luminescence signal. Pressure is determined by comparing the spectral shift of the ruby R₁ fluorescence line or the Sm²⁺⁵D₀ → ⁷F₀ narrow emission band with available calibration curves,^19,21 including temperature correction factors.^22,25 Hence, for an experienced researcher, the determination of pressure in a DAC is usually relatively easy. The recent report, which shows how to enhance the signal intensity of the SrB₄O₇:Sm²⁺ pressure sensor by about 60 times, incorporating Eu²⁺ ions in its structure, is worth noting.²⁶

Similarly, optical determination of temperature is nowadays commonly performed using luminescent thermometers.^2,27−30 Their operation principle is based on the use of some spectroscopic parameters, such as band intensity ratio, line shift, bandwidth, luminescence lifetimes, etc., which monotonously change with temperature.^2,31−33 Inorganic materials based on lanthanides (Ln³⁺) are usually used for these purposes, e.g., fluorides, oxides, phosphates, vanadate, or borates doped with Pr³⁺, Nd³⁺, Ho³⁺, Er³⁺, Tm³⁺, and Yb³⁺.^2,27−34 This is mainly due to the fact that Ln³⁺ ions have a ladder-like electronic structure and exhibit narrow emission lines, large Stokes shifts, thermally coupled levels (TCLs), photostability, and resistance to high temperature and pressure treatments.^29,35−38 Moreover, materials codoped with Yb³⁺ and Ho³⁺, Er³⁺, or Tm³⁺ ions may exhibit the conversion of infrared-to-visible photon energy that gives rise to what is called upconverted (UC) luminescence, i.e., anti-Stokes emission under the influence of NIR irradiation.³⁹⁻⁴² UC emission is beneficial for various applications requiring the use of a low-energy NIR light source instead of UV light, as well as for use in nontransparent systems requiring highly penetrable NIR light.^6,43−47

The advantage of luminescence, contactless thermometers/manometers over their conventional analogues is the rapid, noninvasive, precise, and accurate remote detection of temperature/pressure, allowing sensing in the nano-sized and micron-sized areas.^2,27−30

However, when experiments are performed simultaneously as a function of pressure and temperature, the determination of both quantities becomes more problematic due to the two-sided pressure–temperature interdependences, originated from the fundamental properties of materials, namely, the decreasing volume of material under pressure and its expansion at elevated temperatures.¹⁻⁷ Such changes in the unit cell volume are associated with the shortening/elongation of interatomic distances, i.e., variations in bond lengths and angles, which have a direct impact on their chemical and physical properties. Therefore, both compression and heating of the material affect the phonon energies of the crystal lattice, the energy level structure of optically active ions, radiative and nonradiative transition rates, and may also lead to the formation of defects and strains in crystals.^1,2,6,43,48

Because of these reasons, the determination of the temperature/pressure value (calibrated only vs temperature or pressure) may become challenging, because both factors affect the spectroscopic properties of the sensors (luminescent ions). Despite the fact that for pressure sensors temperature corrections are available, allowing pressure sensing at elevated temperature,^22,24,25 such sensors are unable to simultaneously detect the local temperature value that the compressed material senses. In fact, the temperature inside the DAC pressure chamber and outside, on the DAC surface, may be different due to local heat fluctuation and nonuniform cooling of the system. Moreover, in the case of experiments with laser heating under high-pressure conditions, heating of the compressed material is very localized, so it is not possible to use a thermocouple or any other contact thermometer to accurately monitor the temperature of the sample.⁴⁹

To the best of our knowledge, there are no reports concerning accurate, simultaneous temperature–pressure sensing under extreme conditions of high pressure and temperature via UC luminescence thermometry–manometry, except for our recent work showing the possibility of high-temperature determination under mild pressure (up to ≈5 GPa), thanks to the use of Tm³⁺-doped LaPO₄ nanoparticles.⁶ However, in this work, we could only estimate one high-temperature value (≈468 K), being unable to detect lower temperature values (due to the low signal intensity, as Tm³⁺ ions are thermalized at a higher temperature). Moreover, using the same luminescence sensor, we could not simultaneously detect the pressure in the system subjected to high-temperature treatments. Other currently available reports on luminescence temperature sensing under pressure and vice versa either use the same parameter for sensing of both state functions or do not consider or clearly indicate the pressure–temperature interdependences, leading to significant inaccuracy/uncertainty, hampering their simultaneous determination under extreme conditions of both factors.^{19,25,50−53}

In this work, we report about the possibility of simultaneous optical temperature and pressure sensing in the micron-sized range (≈100 μm) of the DAC’s chamber, via luminescence thermometry and manometry. The developed upconverting YF₃:Yb³⁺-Er³⁺ microsensor can effectively work being simultaneously compressed and heated, allowing accurate temperature and pressure sensing even under extreme conditions and in the small-sized regions.

Experimental Section

Materials

Y(NO₃)₃, Yb(NO₃)₃, and Er(NO₃)₃ were prepared by dissolving Y₂O₃, Yb₂O₃, and Er₂O₃ (99.99%, Stanford Materials) in concentrated nitric acid (65%, p.a., Avantor Performance Materials Poland).

Synthesis of YF₃:20% Yb³⁺,2% Er³⁺

The synthesis of 0.3 g of YF₃:20% Yb³⁺, 2% Er³⁺ was performed by mixing 2.1807 mmol Y(NO₃)₃, 0.5592 mmol Yb(NO₃)₃, and 0.0559 mmol Er(NO₃)₃ aqueous solutions together and filling the glass up to 40 mL with water. Next, 310.6 mg of NH₄F (98%, Sigma-Aldrich) was dissolved in 40 mL of water and slowly added dropwise to the lanthanide solution. After completion of the dropwise addition, the prepared mixture was transferred to a 100 mL Teflon-lined stainless-steel autoclave and processed at 513 K for 36 h. The white precipitate was centrifuged at 6000 rpm and rinsed three times with water. The prepared microcrystals were dried overnight in an oven at 343 K. Deionized water (Milli-Q quality) was used for all experiments.

Characterization

XRD powder patterns were recorded using a Bruker AXS D8 Advance diffractometer, operating with Cu Kα radiation (λ = 1.5406 Å). Scanning electron microscopy (SEM) and energy-dispersive X-ray (EDX) analyses were performed using a scanning electron microscope FEI Quanta 250 FEG, equipped with an EDAX detector. The UV–vis–NIR absorption spectrum of the sample was measured using a JASCO V-770 spectrophotometer, equipped with a spherical integrator ILN-925. UC emission spectra at high-pressure conditions were collected using a Hamamatsu R928P photomultiplier coupled with a 750 mm focal length monochromator, Jobin-Yvon Spex 750 M, and a tunable CW Ti:sapphire laser system, Spectra Physics 3900-S pumped with a 15 W 532 nm Spectra Physics Millennia, fixed at 975 nm as an excitation source (0.5 W; spot size ∼0.2 mm). The spectra at high-temperature conditions were measured using an Andor Newton CCD camera (silicon), coupled to an Andor Shamrock 500 spectrometer. All spectra were corrected for the apparatus response. Ruby emission was monitored using a Renishaw InVia confocal micro-Raman system with a power-controlled 100 mW 532 nm laser diode.

Loading DAC Procedure and Pressure Calibration

For measurements of high pressure at ambient temperature, a membrane DAC made at Sorbonne Université (Paris, France) was used. Tungsten carbide sheets with a thickness of 200 μm were used as gaskets. The gaskets were preindented to ∼50 μm thick and then drilled with a spark erosion machine to make a hole with a diameter of ∼100 μm, which will be the hydrostatic chamber. After mounting the metal gasket on a diamond, a small piece of ruby and the sample were placed in the gasket hole, which was finally filled with a methanol/ethanol/water (16:3:1) solution, working as a hydrostatic transmitting medium up to ∼10 GPa. The membrane DAC was quickly closed to prevent the evaporation of the solution. For the measurements under extreme conditions of pressure and temperature, experiments were carried out using a miniature body-piston type DAC (ϕ = 26 mm, h = 18 mm) made at Universität Paderborn (Germany) (see Figure 2 in ref (1)), in which the pressure values were adjusted using four metal screws. In both cases, pressure values were determined using ruby R₁ and R₂ fluorescence line shifts, excited by a 532 nm laser and a temperature-corrected ruby calibration curve.²¹

Figure 2.

(a) Normalized upconverted emission spectra of the YF₃:Yb³⁺, Er³⁺ material, measured at different pressure values and (b) spectral position of the ⁴F_9/2 → ⁴I_15/2 transition (Stark sublevel) as a function of pressure; λ_ex = 975 nm.

Luminescence Measurements

High-pressure UC luminescence measurements of the material placed in the DAC were performed in a back illuminated configuration with 180° detection geometry. The laser beam was focused on the sample placed in a gasket hole, and the emission signal was collected from the opposite site of the DAC. The spectroscopic measurements at high-temperature conditions were performed by placing a thin piece of the sample between two microscope glasses, at the center of a tube furnace (±1 K). Optical geometry was the same as in the case of high-pressure measurements Appropriate power of the laser was adjusted before measurements began to avoid the laser-induced heating of the sample. The same procedure was applied for the UC emission measurements in DAC, performed simultaneously at elevated temperature and under high-pressure conditions using a home-made furnace and placing the tip of the thermocouple deeply inside the DAC housing.

Results and Discussion

YF₃ irregular, rhombus-like microcrystalline particles (∼1–2 μm size), as observed in Figure 1a and Figure S1, doped with the 20% Yb³⁺ and 2% Er³⁺ sample, further used for sensing purposes, were prepared via a facile hydrothermal synthesis (see the Methods section). Our intention was to use a simple, inorganic micron-sized material that is easy to synthesize and reproduce and exhibits intense and bright UC emission, which is crucial for sensing applications. Hence, we used YF₃ of the low phonon energy of the crystal lattice (to avoid multiphonon relaxation) and the commonly used doping concentration of 20% Yb³⁺, 2% Er³⁺, which is usually the optimal amount of these ions, providing a strong UC luminescence signal.^{32,33,43,48,54} The relatively large particle size is the result of hydrothermal synthesis conditions, which induces the growth of the crystals, resulting in a product of higher crystallinity and improved luminescence properties, compared with the materials prepared via other methods.⁵⁵ Moreover, the obtained micron-sized particles were also chosen for further sensing studies because of their lower surface-to-volume ratio compared to nanoparticles, making them less susceptible to the negative effects of the external environment, e.g., potential damage/decomposition and photobleaching under high-pressure and temperature conditions, which might affect and deteriorate the luminescence-based sensing properties of the material. Energy-dispersive X-ray spectroscopy (EDX) mapping (Figure S2) confirms the uniform distribution of Y³⁺, Yb³⁺, and Er³⁺ ions in all particles of the prepared material. The powder XRD pattern (Figure S3) of the prepared YF₃:Yb³⁺-Er³⁺ sample matches with the reference pattern of orthorhombic YF₃, with the Pnma space group (ICDD #01-074-0911). ICP-MS analysis revealed that the exact composition of the obtained product was YF₃:18.92% Yb³⁺, 2.92% Er³⁺. The investigated micron-sized luminophore is resistant to high-temperature treatment at least up to above ∼700 K,⁵⁶ and it is stable under high-pressure conditions, preserving its orthorhombic structure up to above ∼9 GPa, with a bulk modulus of ∼146(6) GPa.⁵⁷

Figure 1.

(a) Scheme of the experimental setup for high-pressure measurements. (b) Er³⁺ upconverted green and red luminescence obtained exciting the Er³⁺ and Yb³⁺ ions with a laser at 975 nm in the YF₃ micromaterial; the inset presents the luminescence of the sample under 975 nm laser irradiation. (c) Energy-level diagram of the studied material.

The studied material shows green and red UC luminescence occurring from Er³⁺ activator ions sensitized by NIR excitation at 975 nm with Yb³⁺ ions. In this system, Yb³⁺ species absorb NIR irradiation and subsequently transfers it via the energy transfer upconversion mechanism (ETU) to Er³⁺ emitters.^39,43 The diffuse-reflectance absorption spectrum presented in Figure S4 shows absorption bands corresponding to Er³⁺ (240–850 nm) and Yb³⁺ (850–1100 nm) ions, confirming the possibility of effective absorption of the incident photons by the sample. The UC emission spectrum of the obtained YF₃:Yb³⁺, Er³⁺ material, shown in Figure 1b, presents three bands resulting from the 4f–4f intraconfigurational electronic transitions in Er³⁺ ions: ²H_11/2 → ⁴I_15/2 at ∼525 nm, ⁴S_3/2 → ⁴I_15/2 at ∼545 nm, and ⁴F_9/2 → ⁴I_15/2 at ∼660 nm. All bands are split into several Stark sublevels, due to the crystal-field effects. The energy level diagram presented in Figure 1c schematically shows the main energy transfer (ETU), nonradiative (dotted arrows), and radiative processes occurring in this system. The UC emission quantum yield was measured in a spherical integrator, resulting in the value of ≈0.71 ± 0.09%, determined at the same pump power as used during the luminescence measurements (0.5 W; spot size ∼0.2 mm).

The luminescence properties of the prepared material were investigated under high-pressure conditions using a diamond anvil cell (DAC) up to around 8.4 GPa. To determine the pressure value, a small ruby sphere of about a few microns was placed together with the sample in the hydrostatic chamber of the DAC. The detailed procedures of the DAC loading and pressure calibration are described in the “Methods” section, while the experimental setup used for the measurements is schematically presented in Figure 1a.

Figure 2a shows the normalized UC emission spectra of the YF₃:Yb³⁺, Er³⁺ sample, measured under high-pressure conditions. As the pressure increases, the overall luminescence intensity decreases, as observed in the non-normalized UC emission spectra shown in Figure S5. This effect is due to the decrease in interionic distances in the compressed material, resulting in the enhanced cross-relaxation and multiphonon relaxation, stronger electron–phonon coupling, changes in radiative probabilities, and formation of crystal defects and strains in the material structure.^2,6,43,58,59

Based on a more detailed analysis, it can be observed that the spectral positions of the observed bands shift with pressure, as observed in Figure S6, showing spectral positions of the emission bands of Er³⁺ as a function of pressure. This is due to the fact that shortening the interionic distances between the Er³⁺ and F^– ligand ions leads to a decrease in the energy separation between the ground and excited multiplets of Er³⁺, which typically translates into the spectral red-shift of the emission lines.^2,43 In more detail, the decreasing distance between the luminescent ion and its surrounding ligands causes stronger covalent interactions, as well as a decrease in the Coulomb and spin–orbit interactions, known as the nephelauxetic effect.⁶ Since the compression of the material also leads to the enhanced crystal-field strength, manifested as the increasing splitting of the bands (see the magnified spectra in Figure S7), which leads to the blue- and red-shifts of different Stark sublevels.⁶⁰ Due to the superposition of these effects, the resulting, average spectral shifts of the whole bands are different, i.e., they usually have lower rates and are less monotonic (Figure S6) compared to the spectral shifts of the individual lines (Figure 2b), as each Stark sublevel can move in a different direction with a different rate.

Because of this, for the purpose of pressure sensing, we selected the emission line located at ∼664.6 nm, which is sharp and well separated from adjacent sublevels, its shift is monotonic and reversible, and it shows the highest shift rate among the others, as well as almost temperature-independent behavior, which will be discussed in the next paragraphs. This peak red-shifts with a rate of ∼0.1855 nm·GPa^–1 (−4.191 cm^–1·GPa^–1) and shows a linear pressure dependence (line shift [cm^–1] = −4.191P + 15046.819, where P is the pressure in GPa; R² ≈0.997), shown in Figure 2b, enabling the accurate pressure sensing of ±0.1–0.2 GPa. This value is relatively high compared to other shifts of various Ln³⁺ narrow emission lines reported in the literature as potential pressure sensors, such as 0.1 nm·GPa^–1 for Tm³⁺ in LaPO₄⁶ and −0.13 nm·GPa^–1 for Nd³⁺ in YAlO₃,¹⁷ and is also comparable or lower than other commonly used pressure sensors, such as 0.197 nm·GPa^–1 for Eu³⁺ in Y₃Al₅O₁₂,⁶¹ 0.35 nm·GPa^–1 of Cr³⁺ in ruby,^18,19 and 0.25 nm·GPa^–1 of Sm²⁺ in strontium borates.^20,23Table 1 shows a comparison of various luminescent pressure sensors, based on the spectral shift of Ln³⁺ emission lines.

Table 1. Comparison of Pressure-Sensing Properties Determined Based on Peak Shift Observations between Selected Lanthanide-Based Luminophores^a.

dopant ions	host	transition	ν0 (cm–1)	dν/d0 (cm–1/GPa)	pressure range (GPa)	ref.
Sm2+	SrFCl	5D0 → 7F0	14,493	–22.8	0–8	(62)
Sm2+	SrB4O7	5D0 → 7F0	14,599	–5.41	0–130	(20)
Sm2+	SrB2O4	5D0 → 7F0	14,599	–5.11	0–25	(24)
Nd3+	YAlO3	4F7/2, 4S3/2 → 4I9/2	13,650	3.15	0–10	(60)
		4F9/2 → 4I9/2	14,450	–7.21
Tb3+	Y3Al5O12	5D4 → 7F6	15,823	–3.5	0–8	(63)
Ce3+	Y6Ba4(SiO4)6F2	5d → 4f	21,459	–28.62	0–30	(64)
Ce3+Ce3+	(NH4)4Ce(SO4)4·2H2O	5d → 4f	29,411	–9.9	0–60	(65)
	Ce(NO3)3·6H2O	5d → 4f	30,534	–24.8	0–63
Eu2+	BaLi2Al2Si2N6	5d → 4f	18,797	–58.05	0–20	(66)
Pr3+	LaCl3	3P0 → 3H4	20,473	–22.9	0–8	(67)
Pr3+	LaOCl	3P0 → 3H4	20,268	–20.5	0–21	(68)
Eu2+	KMgF3	6P7/2 → 8S7/2	27,846	–8.15	0–30	(69)
Eu3+	α-GdBO3	5D0 → 7F2	15,974	–6.4	0–8	(70)
Eu3+	β-GdBO3	5D0 → 7F1	16,920	–10.0	0–8	(70)
Eu3+	EuPO4	5D0 → 7F0	17,274	–2.1	0–6	(71)
Eu3+	Y3Al5O12	5D0 → 7F1	16,932	–5.7	0–7	(61)
Yb3+-Er3+	NaBiF4	4I13/2 → 4I15/2	6,653	3.54	0–16	(53)
Yb3+-Er3+	YF3	4F9/2 → 4I15/2	15,038	–4.18	0–8	this work

^aν₀ is the center of the observed transition at ambient pressure and d_ν/d₀ is the shift of the observed transition under high-pressure conditions.

Figure 2a clearly shows that as the pressure increases, the relative intensities of the upconverted emission bands change, and the band at around 660 nm, associated with the ⁴F_9/2 → ⁴I_15/2 transition, changes most significantly. The determined luminescence intensity ratios, taken as the area of the bands, increase with pressure for 545/660 nm (⁴S_3/2 → ⁴I_15/2/⁴F_9/2 → ⁴I_15/2) and 525/660 nm (²H_11/2 → ⁴I_15/2/⁴F_9/2 → ⁴I_15/2) from about 0.55 to 1.4 and from 0.055 to 0.125, respectively (see Figure S8). Fortunately, the changes are negligibly small for the thermalized bands at 525/545 nm (²H_11/2 → ⁴I_15/2/⁴S_3/2 → ⁴I_15/2), enabling accurate optical temperature sensing under elevated pressure, as discussed further in the text. The observed changes in the relative intensities between the bands along with pressure lead to luminescence color tuning from yellow-green to green (Figure S9).

Normalized UC emission spectra of the material studied, recorded in the temperature range from 295 to 478 K, are presented in Figure 3a. It can be clearly seen that the relative intensity of the thermalized ²H_11/2 → ⁴I_15/2 transition at ∼525 nm increases significantly with temperature compared to other bands of Er³⁺. On the other hand, the band intensities at ∼545 and ∼660 nm decrease, as observed in Figure S10, due to thermal quenching processes, such as enhanced multiphonon relaxation and, in the case of the ⁴S_3/2 → ⁴I_15/2 transition, the mentioned thermalization processes.^2,31,33 Thanks to the thermally coupled nature of the ²H_11/2 and ⁴S_3/2 energy levels in Er³⁺ ions, the corresponding band intensity ratio ²H_11/2 → ⁴I_15/2/⁴S_3/2 → ⁴I_15/2 (525/545 nm) is commonly used for optical temperature sensing purposes.^{2,27−33,54,72,73} In other words, elevation of temperature increases the population of the ²H_11/2 multiplet and at the same time decreases the population of the ⁴S_3/2 multiplet. These processes are consistent with the Boltzmann type distribution:²⁹

1

where LIR is the luminescence intensity ratio or band ratio; ΔE is the energy difference between the centroids of the ²H_11/2 → ⁴I_15/2 and ⁴S_3/2 → ⁴I_15/2 emission bands; I₅₂₅ and I₅₄₅ are their integrated intensities, respectively; k_B is the Boltzmann constant; T is the absolute temperature; and B is a constant, associated with the state degeneracies, branching ratio of the transitions in relation to the ground state, energy of the transitions, and rates of total spontaneous emission.²⁹

Figure 3.

(a) Normalized upconverted emission spectra for the YF₃:Yb³⁺, Er³⁺ material, measured at different temperature values, using λ_ex = 975 nm; (b, d) determined bands intensity ratios for the Er³⁺ transitions ²H_11/2 → ⁴I_15/2/⁴S_3/2 → ⁴I_15/2 (525/545 nm) and ²H_11/2 → ⁴I_15/2/⁴F_9/2 → ⁴I_15/2 (525/660 nm); and (c, e) corresponding relative sensitivities as a function of temperature.

Based on the integrated areas under the curves of the bands, we determined the I₅₂₅/I₅₄₅ band ratio, and then fitted to eq 1, resulting in ΔE ≈706.6 cm^–1, with R² ≈0.999 (Figure 3b), which is very close to the value determined on the basis of UC emission spectra, using the spectral positions (peak centroids) of these bands, i.e., ∼702 cm^–1. Based on the determined band ratios, we calculated the relative thermal sensitivity S_r as a function of temperature, according to the formula:

2

which shows how the LIR parameter changes per 1 K of absolute temperature. With the increase of temperature, the S_r values decrease from ∼1.2% K^–1 at 295 K down to ∼0.4% K^–1 at 478 K, as observed in Figure 3c.

Additionally, we determined the band intensity ratio of the nonthermally coupled levels (non-TCLs) ²H_11/2 → ⁴I_15/2/⁴F_9/2 → ⁴I_15/2 (525/660 nm), and fitted it to the second-order polynomial function: LIR = −4.14 × 10^–7T² + 8.94 × 10^–4T −0.18, with R² ≈0.994 (Figure 3d), as they do not conform to the Boltzmann distribution. Please note that despite the fact that there is no thermalization between these two bands, they apparently exhibit different temperature-dependent quenching and energy transfer rates; hence they can be correlated with temperature and used for sensing purposes.^2,31,33 The calculated S_r values for the 525/660 nm band ratios decrease from ≈1.4% K^–1 (at 295 K) down to ≈0.3% K^–1 (at 478 K), as observed in Figure 3e.

Finally, using the determined band ratios and S_r values, we have calculated the temperature resolution, i.e., the uncertainty of temperature determination (δT), defined as

3

where δLIR is the uncertainty of determination of the LIR parameter.^{2,29,33,54,74} Under ambient conditions, the δT values are around 0.4 K for both band ratios, whereas with temperature elevation they monotonously increase up to around ≈1.1 and 1.5 K (at 478 K) for 525/545 and 525/660 nm band ratios, respectively (see Figure S11), ensuring high resolution of temperature sensing in the whole measured range. Table 2 shows the comparison of the performance of various upconverting thermometers, using the TCLs of Ln³⁺ for temperature sensing. To check the thermal stability of the sample and repeatability of temperature sensing, we cycled the sample between low and high temperatures (Figure S12), confirming that both the determined LIR parameters reversibly change within the measured temperature range.

Table 2. Comparison of Different Upconverting Luminophores Used for Thermometry Based on the Ratio of TCLs Bands^a.

dopant ions	host	Sr max (% K–1)	T (K)	T-range (K)	transitions	λ (nm)	ref.
Yb3+-Ho3+	KLu(WO4)2	0.54	300	297–673	5F5 → 5I8/2S2,5F4 → 5I8	545/650	(75)
Yb3+-Ho3+	β-NaYF4	0.90	300	300–500	5F5 → 5I8/2S2,5F4 → 5I8	545/650	(76)
Yb3+-Tm3+	LaPO4	3.00	293	293–773	3F2,3 → 3H6/3H4 → 3H6	700/800	(6)
Yb3+-Tm3+	YPO4	2.34	293	293–773	3F2,3 → 3H6/3H4 → 3H6	700/800	(6)
Yb3+-Er3+-Tm3+	YF3	1.01	293	293–563	2H11/2 → 4I15/2/4S3/2 → 4I15/2	520/540	(77)
Yb3+-Nd3+-Er3+	GdOF/SiO2	1.60	260	260–490	2H11/2 → 4I15/2/4S3/2 → 4I15/2	534/543	(78)
Yb3+-Er3+	SrF2	1.20	298	298–383	2H11/2 → 4I15/2/4S3/2 → 4I15/2	525/545	(54)
Yb3+-Er3+	Gd2O3	0.85	300	300–900	2H11/2 → 4I15/2/4S3/2 → 4I15/2	523/548	(79)
Yb3+-Er3+	Tellurite glass	0.53	298	298–473	2H11/2 → 4I15/2/4S3/2 → 4I15/2	525/548	(80)
Yb3+-Er3+	KBaY(MoO4)3	1.80	250	250–460	2H11/2 → 4I15/2/4S3/2 → 4I15/2	519/550	(81)
Yb3+-Er3+	SrWO4	0.96	300	300–518	2H11/2 → 4I15/2/4S3/2 → 4I15/2	525/547	(82)
Yb3+-Er3+	NaYbF4	3.46	175	175–475	2H11/2 → 4I15/2/4S3/2 → 4I15/2	525/545	(83)
Yb3+-Er3+	Gd2(MoO4)3	1.34	295	295–660	2H11/2 → 4I15/2/4S3/2 → 4I15/2	530/550	(84)
Yb3+-Er3+	Bi4Ti3O12	1.19	293	293–573	2H11/2 → 4I15/2/4S3/2 → 4I15/2	530/554	(85)
Yb3+-Er3+	YF3	1.20	295	295–478	2H11/2 → 4I15/2/4S3/2 → 4I15/2	525/545	this work

^aS_r, max is the maximum observed value of relative sensitivity.

It is worth noting that along with an increase in temperature, a slight blue shift of the peak centroids was observed for all of the transitions, i.e., about −5 × 10^–3 nm K^–1 (see Figure S13). This effect is exactly opposite to the previously described behavior of the Er³⁺ radiative transitions in pressure (spectral shift of the bands), as in the pressure, the material is compressed (decreasing volume), whereas in temperature its unit cell volume increases, due to the well-known temperature expansion of the materials.⁶ It should be emphasized that, favorably, the temperature shift of the selected (for pressure sensing) Stark sublevel is 1 order of magnitude lower (−3 × 10^–4 nm K^–1), compared to the thermal shift of the whole band (⁴F_9/2 → ⁴I_15/2). Moreover, due to variations in the bands ratio, a slight change in the luminescence color was noticed, which can be observed in the CIE diagram in Figure S14.

The investigated sample was tested for the possibility of using it simultaneously as a temperature and pressure sensor. To do this, we performed two series of measurements, loading the DAC with the developed, bifunctional optical sensor of pressure and temperature (YF₃:Yb³⁺, Er³⁺) and adjusting the initial pressure values to ≈3.5 and 8 GPa, respectively. Afterward, we measured its UC emission spectra every ≈25 K, up to 473 K. Due to the different thermal expansion of the DAC material (parts of the chamber, pressure-transmitting medium, and the sample itself), the initially set pressure values decreased with increasing temperature. As described previously, we used the ratio of the ²H_11/2 → ⁴I_15/2/⁴S_3/2 → ⁴I_15/2 band for temperature determination and the shift of the selected Stark sublevel (≈664.6 nm) for pressure determination. The reference pressure values were calculated on the basis of the positions of R₁ and R₂ ruby fluorescence lines and the Mao scale,¹⁹ taking into account the temperature correction made by Rekhi et al.²⁵ A thermocouple mounted in the DAC metal housing was used as a temperature reference. The recorded UC emission spectra for the YF₃:Yb³⁺, Er³⁺ material, measured simultaneously at different high temperature and pressure values, are presented in Figures S15 and S16 in the Supporting Information.

Based on the determined shift rate (∼0.1855 nm GPa^–1) of the mentioned Stark sublevel of Er³⁺, we successfully estimated the pressure values in the DAC chamber at elevated temperature, both at lower and higher pressures (Figure 4a,b). The determined pressure values with the use of ruby and the developed upconverting sensor are in good agreement because the accuracy of our method was about 0.2–0.3 GPa, and decreases with increasing temperature. Please note that the method used includes the determined temperature correction, i.e., a small spectral blue-shift with the temperature of the Stark line discussed, i.e., −1.62 × 10^–3 GPa K^–1.

Figure 4.

(a, b) Pressure measurements conducted as a function of temperature for various initial pressure values, using ruby as a reference and spectral shift of the Er³⁺ Stark line for experimental pressure determination; (c, d) temperature measurements conducted as a function of pressure, using thermocouple mounted in the DAC as a reference and the LIR technique for optical temperature determination; the developed upconverting YF₃:Yb³⁺, Er³⁺ optical sensor was used as an experimental probe of pressure and temperature.

Using the determined band ratio 525/545 nm of Er³⁺ TCLs, we can optically estimate the local temperature value in the pressure chamber of the DAC, i.e., the temperature of the sample placed in a tiny ≈100 μm hole of a metal gasket, situated between two diamond anvils. For lower temperature values under both the lower and the higher pressure ranges (Figure 4c,d), the temperature readouts from the sample (optical) and the reference (thermocouple) are almost identical, with small variations ≤5 K. At higher temperature and pressure values (Figure 4d), larger differences can be noticed between the readouts (≈10 K).

The differences in pressure and temperature readouts between the sample and the ruby/thermocouple are mainly caused by the following factors: (1) pressure–temperature interdependences; (2) precision of the band spectral position and intensity ratio readouts, associated with the resolution and sensitivity of the setup used for the UC emission spectra measurements as well as the intensity of the luminescence signal from the sample; (3) the accuracy of the method used; and (4) variations in the temperature outside and inside the DAC pressure chamber, affecting the direct readouts from the thermocouple and the sample, respectively.

Conclusions

We have developed a bifunctional optical sensor of temperature and pressure, working under extreme conditions of both state functions. The optical probe is based on the upconverting YF₃:Yb³⁺, Er³⁺ micron-sized particles, exhibiting a bright yellow-green UC emission under the influence of NIR (975 nm) laser irradiation. By monitoring the spectral shift of the selected narrow emission line corresponding to the Stark sublevel around 665 nm of Er³⁺, it is possible to accurately determine high pressure values in the system (±0.1–0.2 GPa). On the other hand, applying the LIR technique, i.e., by monitoring the change of the 525/545 nm band intensity ratio, corresponding to the Er³⁺ TCLs ²H_11/2 → ⁴I_15/2/⁴S_3/2 → ⁴I_15/2, it is possible to determine the temperature in the system (±1 K). Favorably, the pressure negligibly influences the mentioned TCL band ratio; hence we were able to determine the local temperature values that the compressed material achieved under high-pressure (up to ≈8 GPa) in the DAC. Moreover, thanks to the very low temperature dependence of the selected Stark line, we also successfully determined the pressure values in the system subjected to the high-temperature treatment (up to ≈473 K). Hence, using a single material, it is possible to simultaneously monitor the high temperature and pressure of the system, under extreme conditions of both factors, with a relatively small error. The presented results play an invaluable role in the research performed simultaneously under high pressure and temperature conditions as well as in field of the synthesis of novel materials under extreme conditions, requiring precise and accurate monitoring of temperature and pressure in the system.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsami.0c09882.

• SEM and TEM images; EDX mapping; XRD pattern; diffuse-reflectance absorption spectrum; normalized and non-normalized emission spectra; calculated band ratios and temperature resolutions; temperature cycling; determined peak centroids; and CIE 1931 diagrams (PDF)

The authors declare no competing financial interest.