Improper Background Treatment Underestimates Thermometric Performance of Rare Earth Vanadate and Phosphovanadate Nanocrystals

Abstract

Luminescence thermometry is the state-of-the-art technique for remote nanoscale temperature sensing, offering numerous promising cutting-edge applications. Advancing nanothermometry further requires rational design of phosphors and well-defined, comprehensive mathematical treatment of spectral information. However, important questions regarding improper signal processing in ratiometric luminescence thermometry are continuously overlooked in the literature. Here, we demonstrate that systematic errors arising from background/signal superposition impact the calculated thermometric quality parameters of ratiometric thermometers. We designed ultraviolet-excitable (Y,Eu)VO₄ and (Y,Eu)(P,V)O₄ nanocrystals showing overlapped VO₄^3– and Eu³⁺ emissions to discuss systematically how uncorrected background emissions cause magnified (∼10×) temperature uncertainties and undervalued (∼60%) relative thermal sensitivities. Adequate separation of spectral contributions from the VO₄^3– background and the Eu³⁺ signals via baseline correction is necessary to prevent underestimation of the thermometric performances. The described approach can be potentially extended to other luminescent thermometers to account for signal superposition, thus enabling to circumvent computation of apparent, miscalculated thermometric parameters.

Introduction

Remote temperature probes based on temperature-sensitive luminescence are highly versatile tools for nanomedicine, catalysis, and electronics.¹⁻⁴ Typical nanothermometers used in this context explore the abundant electronic level pattern of trivalent lanthanoid ions (Ln³⁺), which enable tailoring absorption and emission from the ultraviolet (UV) to the near-infrared (NIR) by the adequate choice of activators and sensitizers.^5,6 Furthermore, the narrow line width and minimal spectral overlap between 4f–4f transitions also prevents systematic errors, which provides enhanced accuracy for the thermometric determinations. While several optical parameters arising from Ln³⁺ luminescence are often available for thermal correlations,^7,8 the luminescence intensity ratio (LIR) between two electronic transitions (or Stark components of a single excited level) in thermal equilibrium is by far the preferred choice for calculating absolute temperatures in both primary and nonprimary optical thermometers.⁹⁻¹² This method is not only easily implemented but it is also considered robust against experimental or sample-related conditions, such as fluctuations in excitation intensity, sample geometry, or concentration of luminescent probes.¹³

Although ratiometric thermal correlations become progressively widespread, recent investigations have raised concerns about the precision and accuracy of temperature determination using LIR thermometry. Following the landmark work of Labrador-Páez et al.,¹⁴ several studies have demonstrated the impact of various environmental and experimental factors on the luminescence spectra and emission decays of Ln³⁺-based luminescent materials.^9,10,15 These factors include low signal-to-noise (SNR) ratios and the presence of artifacts introducing biases in the thermometric correlations. For instance, SNR quantifies the readout intensity toward the random intensity fluctuations during the measurement, which depends on excitation exposure times, brightness of the nanoprobe, and extinction of emission photons in an opaque medium.^10,13,14 In addition, reliability (i.e., difference between measured and real temperature) is affected by a distorted luminescence spectral distribution arising from wavelength-dependent transmission by the sample, wavelength-dependent self-absorption by the thermometer, or modified density of optical states around the optical probe.^9,14−17 Finally, the presence of intruding transitions within analyzed spectral ranges in LIR introduces misinterpretation, as described in cases involving Er³⁺- or Nd³⁺-based thermometry.^14,18,19 Such effects are not inherent properties of the nanothermometers, and the random nature of some of them makes it challenging to adequately process the acquired emission spectra, hindering unequivocal thermal readouts.

Here, we show that lack of background emission treatment is also an origin of inaccuracy, leading to incorrect values of relative thermal sensitivities and temperature uncertainties. Despite the high number of works dealing with potentially superposed signals for luminescence thermometry, only scarce works have investigated the artifacts arising from background emissions to date. We investigated this issue by elaborating Eu³⁺-doped yttrium vanadate (YVO₄) and yttrium phosphovanadate (Y(V,P)O₄) particles, which are UV-excited phosphors showing both narrow and broad bands arising from Eu³⁺ and VO₄^3– emissions. The intrinsic spectral overlap between ³T_1,2 → ¹A₁ (VO₄^3–) and ⁵D₀ → ⁷F_J (Eu³⁺) transitions is a fruitful example to demonstrate the general effect of background emissions in the sensor performance. The idea is not only providing insights to decide whether the temperature correlations are reliable or not but also presenting strategies to circumvent this realistic practical artifact.

Methods

Eu³⁺-doped yttrium vanadate and phosphovanadate nanoparticles were synthesized using a colloidal coprecipitation reaction under hydrothermal conditions, both with and without citrate groups as capping agents. For the preparation of vanadate or phosphovanadate nanoparticles without citrates as stabilizers, the process involved heating a mixture of aqueous rare earth chlorides (YCl₃ and EuCl₃, 99.9%Y³⁺, 0.1%Eu³⁺, mol/mol) and a combination of aqueous ammonium metavanadate (NH₄VO₃) and ammonium hydrogen phosphate ((NH₄)₂HPO₄) in the desired V/P molar ratios. The reactions were performed at 180 °C for 20 h at pH 3. In contrast, for the synthesis of citrate-stabilized vanadate/phosphovanadate nanoparticles, the processes began with mixing rare earth chlorides (99.9%Y³⁺, 0.1%Eu³⁺, mol/mol) and sodium citrate (Na₃cit·2H₂O) in water, followed by the addition of an aqueous solution containing sodium orthovanadate (Na₃VO₄) and ammonium hydrogen phosphate ((NH₄)₂HPO₄) in the desired V/P molar ratios. This synthesis was carried out at 200 °C for 24 h. A detailed description of the experimental procedures, characterization techniques, and data processing is provided in the Supporting Information.

Results and Discussion

Anhydrous YVO₄ and YPO₄ crystallize in the same tetragonal xenotime-type structure (Figure 1a) forming solid solutions at the complete range of compositions.²⁰ Powder X-ray diffraction (XRD) patterns and Raman spectra evidenced the formation of single-phase tetragonal solids (I4₁/amd space group) and the homogeneous incorporation of PO₄^3– into the YVO₄ lattice (Figure 1b,c). This was further corroborated by the linear constriction of unit cell volumes of Y(V_1–xP_x)O₄:Eu³⁺ particles upon higher PO₄^3– molar fractions (Figure 1d), which was also confirmed by Rietveld refinement of experimental XRD data (Table S1 and Figure S1). Raman and infrared (FTIR) spectra attested the occupancy of tetrahedral sites distorted to a D_2d symmetry by VO₄^3–/PO₄^3– species, in accordance with the I4₁/amd structure (Figure 1c and Figure S2). The microstructural alterations caused by PO₄^3– groups included a preferential growth in the (200) plane (Figure 1b) as well as lower crystalline coherence lengths and higher microstrains (Figure 1d). By contrast, additional inclusion of PO₄^3– ions (i.e., x > 0.3, with x = PO₄^3–/(PO₄^3– + VO₄^3–) mol/mol) led to larger crystalline domains and decreased lattice defects. These results highlight the chemical homogeneity of the particles and the effective control of structural properties through the employed colloidal synthesis.

Figure 1.

Structural properties of the (Y_0.999Eu_0.001)(V_1–xP_x)O₄ nanocrystals. (a) Representation of the unit cell and coordination polyhedra in the xenotime-type structure (I4₁/amd space group) of YVO₄ and YPO₄. (b) Powder X-ray diffraction patterns and (c) Raman spectra of (Y_0.999Eu_0.001)(V_1–xP_x)O₄ nanocrystals. (d) Evolution of unit cell volumes, coherence lengths (c. length), and microstrain with respect to PO₄^3– molar ratio. (e) Representative TEM image of (Y_0.999Eu_0.001)(V_0.8P_0.2)O₄ particles showing the internal structure of a single particle (inset).

Transmission electron microscopy (TEM) images of a representative phosphovanadate sample revealed weakly agglomerated nanocrystals with sizes of 15–40 nm (Figure 1e and Figure S3), in agreement with the dynamic light scattering (DLS) results (Figure S4). The presence of lattice fringes throughout the particle volume indicates a high crystalline quality, which is crucial for a high luminescence output.²¹ Thanks to the surface stabilization provided by citrate groups [FTIR, υ_as(COO^–) = 1560 cm^–1, υ_s(COO^–) = 1437 cm^–1, υ(CH) = 2853 + 2921 cm^–1, and υ(OH) = 3315 cm^–1, Figure S5], similar DLS-particle size distributions were achieved regardless of sample composition (Figure S4). As a matter of comparison, a hydrothermal protocol without sodium citrate resulted in Y(V_1–xP_x)O₄:Eu³⁺ nanoparticles with variable sizes (518 ± 11 nm to 63 ± 3 nm) with increasing PO₄^3– fractions (Figure S6). We therefore conclude that citrate groups efficiently provide surface stabilization and regulate growth rates during precipitation, thus yielding highly crystalline nanoparticles for UV-excited luminescent nanothermometry.

Aiming to develop a ratiometric luminescent thermometer based on VO₄^3– and Eu³⁺ emissions, a low Eu³⁺ doping ratio (x = 0.1% mol/mol with respect to Y³⁺) was selected. This enabled to achieve Eu³⁺ and VO₄^3– with similar absolute intensities for a Y_0.999Eu_0.001VO₄ sample at 77 K (Figure S7). Considering the high thermal quenching of vanadate emissions,^22,23 we also prepared samples containing varying amounts of PO₄^3– in the YVO₄ host. The partial dilution reduces the VO₄^3– → VO₄^3– energy transfer probability, thus enhancing the ³T_1,2 → ¹A₁ emissions (Figure S7). In turn, similar Eu³⁺ and VO₄^3– emission intensities were produced under 280 nm excitation even at 297 K. Using as criteria the better crystalline quality (i.e., higher coherence length and low strain, Figure 1d) combined to higher VO₄^3– emission intensities at 297 K and better signal-to-noise ratio (Figure S7), we selected a 20% mol/mol PO₄^3– concentration (i.e., Y_0.999Eu_0.001(V_0.8P_0.2)O₄) for thermometric studies in comparison to Y_0.999Eu_0.001VO₄.

Temperature-dependent luminescence spectra (Figure 2a,b) unveiled only partially altered intensities of the ⁵D₀ → ⁷F_1–4 Eu³⁺ transitions in the 77–297 K range, in contrast to the significant thermal quenching of the ³T_1,2 → ¹A₁ VO₄^3– emissions (Figure 2c,d). Such an inhomogeneous behavior arises from dissimilar thermal dependences of Eu³⁺ and VO₄^3– emitting states, which provides an attractive pathway for temperature determination, as extensively pointed out by us¹¹ and by many authors.²²⁻²⁴ However, the robustness of this approach depends critically on the absence of signal superposition and intruding emissions in the analyzed emission ranges. This is because such signals often result in systematic artifacts in ratiometric luminescence thermometry, where thermometric performances and reported readouts become erroneous.^10,14,15,18

Figure 2.

Temperature-dependent emission spectra (λ_exc.= 280 nm, 77–297 K) of the (a) Y_0.999Eu_0.001VO₄ and (b) Y_0.999Eu_0.001(V_0.8P_0.2)O₄ nanocrystals. (c, d) Integrated intensities of the VO₄^3– (³T_1,2 → ¹A₁, black diamonds) and Eu³⁺ [⁵D₀ → ⁷F_1,2,4, gray circles (J = 1), triangles (J = 2), and squares (J = 4)] transitions normalized to their corresponding values at 77 K (I₀).

The impact of background signals on the thermometry quality parameters is evidenced by comparing the luminescence of the Y_0.999Eu_0.001(V_0.8P_0.2)O₄ and Y_0.999Eu_0.001VO₄ samples (Figure 3 and Figure S8), where the phosphovanadate solid shows an even higher spectral overlap between VO₄^3– and Eu³⁺ emissions than the unmixed vanadate sample. We evaluated two approaches to assess thermometric parameters, the first one neglecting signal superposition, and the second one including a correction of the VO₄^3– emission background to compute the intensities of the Eu³⁺ signals. The thermometric parameters (Δ) were defined as the integrated intensity ratio between the VO₄^3– emission (I_V) and each of the Eu³⁺ transitions arising from the ⁵D₀ excited state, namely ⁵D₀ → ⁷F₁ (I₁), ⁵D₀ → ⁷F₂ (I₂), and ⁵D₀ → ⁷F₄ (I₄). Given the broad spectral width of the VO₄^3– emission band, spectral deconvolution of the broadband emissions was carried out considering Intensity vs wavenumbers rather than Intensity vs wavelength (Figure S9). To ensure the conservation of energy remains valid, we applied the Jacobian transformation²⁵ to all spectra. In the first analyzed approach, the vanadate signal was computed in the 380–570 nm range (17544–26316 cm^–1) while the Eu³⁺ emissions (⁵D₀ → ⁷F₁: 16722–16978 cm^–1, ⁵D₀ → ⁷F₂: 15949–16502 cm^–1, and ⁵D₀ → ⁷F₄: 14084–14451 cm^–1) were integrated without background treatment (Figure 3a). The temperature dependence of the intensity ratios displayed sigmoidal profiles and were modeled in terms of the Mott-Seitz equation,^26,27 considering two nonradiative recombination channels associated with the ³T_1,2 → ¹A₁ VO₄^3– transitions (eq 1):

1

where Δ₀ is the Δ parameter when T → 0 K, α₁ and α₂ are the ratio between the nonradiative and radiative probabilities of the deactivation channels for the ³T₁ → ¹A₁ and ³T₂ → ¹A₁ transitions, and ΔE₁ and ΔE₂ denote the activation energies for the thermal quenching of the corresponding excited states. The use of eq 1 is necessary because assuming a single thermal quenching pathway resulted in inadequate modeling of experimental points at temperatures exceeding 225 K (Figure S10). The final Δ vs temperature calibration curves showed good correlation coefficients with experimental data (r² > 0.998) regardless of the choice of the ⁵D₀ → ⁷F_J (J = 1, 2 or 4) Eu³⁺ emissions (Figure 3b and Table S2).

Figure 3.

Impact of background emission on the thermometric performance of Y_0.999Eu_0.001(V_0.8P_0.2)O₄ nanocrystals. (a, e) Emission spectra (λ_exc = 280 nm, 77 K) illustrating the two spectral processing approaches to derive the intensity ratios. I_V denotes the integrated intensity of the ³T_1,2 → ¹A₁ VO₄^3– transitions, whereas I₁, I₂, and I₄ correspond to the integrated intensities of the ⁵D₀ → ⁷F_1,2,4 Eu³⁺ transitions, respectively. Intensity ratios were calculated (a–d) neglecting signal superposition and (e–h) including a correction of the VO₄^3– emission background to compute the intensities of the Eu³⁺ signals. Temperature dependence of the (b, f) intensity ratios, (c, g) relative thermal sensitivities (S_r), and (d, h) temperature uncertainties (δT). Solid lines in panels (b, f) represent the best fits using eq 1 (r² > 0.998), while solid lines in panels (c, d, g, h) correspond to the mathematical derivation to model S_r and δT. Fitting parameters are summarized in Table S2.

The spectral superposition of the VO₄^3– band is higher for the ⁵D₀ → ⁷F₁ (594 nm) emission than for the ⁵D₀ → ⁷F₂ (619 nm) and ⁵D₀ → ⁷F₄ (698 nm) Eu³⁺ emissions. Consequently, the Δ = I_V/I₁ parameter becomes underestimated because I₁ has a large contribution arising from the VO₄^3– emission if integration is performed without correction. This effect was less pronounced for I_V/I₂ and I_V/I₄ ratios due to the high intensity of the ⁵D₀ → ⁷F₂ transition (I₂) and the lower vanadate emission intensity above 680 nm, respectively.

To quantify the thermometric performance, the relative thermal sensitivity (S_r) was calculated as a function of temperature as follows²⁸:

2

As expected, S_r presented bell-shaped profiles peaking around 192–214 K (Figure 3c). Because the different Eu³⁺ emissions used for the thermometric correlations arise from the same emitting state (⁵D₀), they should ideally yield similar thermal sensitivities after combination with the integrated VO₄^3– intensities to compute the Δ parameters. Nonetheless, maximum relative sensitivities (S_m) were not similar among analyzed ratios and decreased progressively from I_V/I₄ (S_m = 2.33 ± 0.10% K^–1) to I_V/I₂ (S_m = 2.08 ± 0.10% K^–1), and I_V/I₁ (S_m = 1.39 ± 0.52% K^–1). This trend is due to the increasing contribution of the VO₄^3– band to the integration limits of the Eu³⁺ signals, also emphasizing how background emissions may cause misleading sensitivity values in luminescence thermometry. The spectral overlap with a non-negligible broadband background introduces an additive temperature-dependent term on the Eu³⁺ integrated intensities, which results in an apparent, underestimated thermometric parameter Δ. This ultimately causes a reduction in the term of eq 2 if this superposition is not corrected, thus negatively affecting the relative thermal sensitivity values. The general effect of a spectral overlap of the background emissions is discussed mathematically in the Supporting Information (eqs S1–S11). Our conclusions also align with the discussion proposed by Brites et al.²⁸ upon band overlap on previously reported Pr³⁺/Yb³⁺/Tm³⁺-doped NaYF₄ nanocrystals.²⁹

The effects of uncorrected background emissions causing deviations in the thermometric parameters also include lower apparent signal-to-noise ratios (i.e., ratio between peak and baseline intensities), consequently causing a less precise readout. This was evaluated by determining the minimum expected statistical temperature uncertainty (δT) using eq 3(10):

3

where δΔ/Δ stands for the relative uncertainty in the determination of Δ, which has an inverse dependence on the signal-to-noise ratio (SNR) of each transition.¹⁰ For Eu³⁺ transitions (I₁, I₂ and I₄), SNR raised exponentially with temperature due to the reduced contribution of VO₄^3– emission in the monitored spectral ranges. This correlates to a lower δΔ/Δ and a decreased δT as a function of temperature (Figure 3d and Figure S11). As expected, the I_V/I₁ ratio displayed the highest δT values, ranging from 2.33 ± 2.44 K to 0.08 ± 0.02 K between 77 e 297 K. These temperature uncertainties are 1 order of magnitude higher than those observed for I_V/I₂ and I_V/I₄ ratios (Figure 3d).

To overcome these limitations, we applied a spectral separation of VO₄^3– luminescence from the Eu³⁺ emissions (Figure 3e) to calculate thermometric parameters. First, the baselines of spectral regions corresponding to the ⁵D₀ → ⁷F_J Eu³⁺ transitions were corrected to eliminate the vanadate contribution from the background. Then, the VO₄^3– emission envelope within 380–750 nm range was determined by linear interpolation after removal of Eu³⁺ signals. As a result, the intensity ratios exhibited the expected tendency for Δ₀ at 77 K: I_V/I₁ > I_V/I₄ > I_V/I₂ (Figure 3f), which is consistent with the relative intensities of VO₄^3– and Eu³⁺ transitions. The S_r values showed lower relative errors (Figure 3g), while S_m values around 194–204 K were nearly the same for the three intensity ratios (S_m = 2.18 ± 0.06% K^–1, S_m = 2.27 ± 0.01% K^–1, and S_m = 2.27 ± 0.02% K^–1 for I_V/I₁, I_V/I₂, and I_V/I₄, respectively). The background treatment resulted in an improved SNR for the ⁵D₀ → ⁷F_J Eu³⁺ transitions and consequently a reduced δΔ/Δ and δT values (Figure 3h and Figure S11). Experimental temperature uncertainties presented lower dispersions being better adjusted to the proposed model. In addition, I_V/I₂ ratio yielded δT ranging from 0.55 ± 0.04 K to 0.011 ± 0.001 K across the 77–297 K range. This represents a remarkable 10-fold reduction in temperature uncertainties compared to the previous analysis. Ultimately, the δT decreased with the increasing of the SNR involving both VO₄^3– and Eu³⁺ emissions, with the I_V/I₂ ratio producing the lowest values (Figure 3h and Figure S11). This supports the observation that ⁵D₀ → ⁷F₂ transition produces the most intense Eu³⁺ emission (Figure 2a,b). This outcome agrees with the recent analyses conducted by van Swieten et al.¹⁵ and Brites et al.,¹⁰ emphasizing that higher SNR leads to a more precise temperature assessment. Similar trends were observed for Y_0.999Eu_0.001VO₄ nanocrystals (Figure S8).

An alternative approach to deal with the background artifacts involves computing a fraction of the VO₄^3– emission band (excluding overlap with the ⁵D₀ → ⁷F₁ Eu³⁺ transition) (Figure S12). This method employs narrower VO₄^3– integration boundaries, potentially inducing higher temperature uncertainties due to decreased SNR.¹⁵ Despite this expectation, similar S_r and δT results were achieved when compared to analyzing the entire VO₄^3– band. This is also a consequence of the broad spectral width of the emission, which minimizes detrimental effects on thermometric performance resulting from different integration limits. Consequently, addressing reliability issues related to background emissions is primarily attainable through baseline corrections of the considered electronic transitions rather than selecting specific integration regions.

The above-described discussion provided solid basement for comparatively evaluating the thermal performances of the Y_0.999Eu_0.001VO₄ and Y_0.999Eu_0.001(V_0.8P_0.2)O₄ nanocrystals. This investigation focused on the I_V/I₂ ratio, as it offered superior thermometric correlations (Figure 3). The temperature calibration curves (Figure 4a) yielded the following activation energies (eq 1) for thermal quenching of the ³T_1,2 → ¹A₁ VO₄^3– transitions: ΔE₁ = 978 ± 92 cm^–1 and ΔE₂ = 328 ± 23 cm^–1 for Y_0.999Eu_0.001VO₄, and ΔE₁ = 1019 ± 88 cm^–1 and ΔE₂ = 438 ± 44 cm^–1 for Y_0.999Eu_0.001(V_0.8P_0.2)O₄ (Table S2). These nonradiative recombination channels (ΔE₁ and ΔE₂) represent the energy gap between the bottom of the potential energy curve of the ³T₂ and ³T₁ emitting states and the crossover point with the VO₄^3– ground state (¹A₁) or the Eu³⁺ excited states.¹¹ The results confirm that the barrier for thermal deactivation of the ³T_1,2 → ¹A₁ transitions is slightly higher when VO₄^3– centers are diluted in the YVO₄ lattice by the presence of PO₄^3– groups. This is because homogeneous incorporation of PO₄^3– in the crystalline lattice decreases the energy propagation rate through VO₄^3– groups and enhances radiative decay intensities.^30,31 As a result, the emission becomes less prone to nonradiative decays upon heating (Figures 2a,b and 4a).

Figure 4.

Comparative thermometric performances of the Y_0.999Eu_0.001VO₄ and Y_0.999Eu_0.001(V_0.8P_0.2)O₄ nanocrystals in terms of (a) integrated intensity ratio (I_V/I₂), (b) relative thermal sensitivity (S_r), and (c) temperature uncertainty (δT). Solid lines in panel (a) represent the best fits using eq 1 (r² > 0.998), while solid lines in panels (b, c) correspond to the mathematical derivation to model S_r and δT. Fitting parameters are summarized in Table S2.

Because relative thermal sensitivities are proportional to ΔE, a higher maximum sensitivity is expected for the Y_0.999Eu_0.001(V_0.8P_0.2)O₄ nanoparticles. Indeed, the phosphovanadate nanocrystals exhibited a S_m of 2.27 ± 0.01% K^–1 at 191 K, surpassing the 1.85 ± 0.01% K^–1 observed for vanadate nanocrystals (Figure 4b). Conversely, Y_0.999Eu_0.001VO₄ nanoparticles displayed more sensitive correlations at temperatures below 140 K due to the maximum variation of the Δ parameter occurring at lower temperatures (Figure S13). At temperatures exceeding 277 K, the low VO₄^3– emission intensities make the Δ parameter to approach to zero, and consequently, S_r to infinity. This is particularly pronounced for Y_0.999Eu_0.001VO₄ solids and obviously lacks physical significance.^7,32 Operational ranges for thermometry were therefore established based on two conditions: (i) S_r > 1% K^–1 and (ii) VO₄^3– emission intensity exceeding 3% of its integrated intensity at 77 K. The first condition is widely accepted for practical purposes^11,33 while the second one ensures reliability in the thermal performance. Accordingly, operational ranges for Y_0.999Eu_0.001VO₄ and Y_0.999Eu_0.001(V_0.8P_0.2)O₄ nanocrystals were determined as 97 to 277 K and 127 to 287 K, respectively. These ranges align perfectly with temperature requirements in superconducting magnets, aerospace, and macromolecular crystallography,³⁴⁻³⁶ showcasing potentiality of this system for sensitive and accurate temperature evaluation. Indeed, temperature uncertainties ranged between (0.35 ± 0.01) × 10^–1 K and (0.02 ± 0.01) × 10^–1 K (Figure 4c) for both compositions. The reduced δT values for Y_0.999Eu_0.001VO₄ nanocrystals below 140 K stems from higher thermal sensitivity in this range instead of lower SNR (Figure 4b and Figure S13). A comparative analysis highlights the favorably thermometric capabilities of both Y_0.999Eu_0.001VO₄ and Y_0.999Eu_0.001(V_0.8P_0.2)O₄ samples compared to other luminescent nanothermometers, whether single- or dual-center emitting (Table S3). Hence, our results suggest that prepared samples emerge as promising candidates for luminescent nanothermometry applications, specially within the cryogenic temperature range.

Conclusions

In summary, background emissions are a realistic practical problem on ratiometric optical thermometry. We hereby describe how neglecting this issue negatively affects the thermometric correlations of a dual-center thermometer based on VO₄^3– and Eu³⁺ emissions in Y_0.999Eu_0.001VO₄ and Y_0.999Eu_0.001(V_0.8P_0.2)O₄ nanocrystals. Applying appropriate baseline correction ensures reliable relative thermal sensitivities, also significantly reducing the temperature uncertainties. The Y_0.999Eu_0.001VO₄ and Y_0.999Eu_0.001(V_0.8P_0.2)O₄ solids exhibited outstanding UV-excited thermometric performance, achieving maximum relative sensitivities and temperature uncertainties around 2% K^–1 (at 191 K) and 0.03–0.002 K, respectively. The particles also showed a broad S_r > 1% K^–1 operational range (97 to 287 K), which is useful for cryogenic applications. Our work highlights that temperature determination depends not only on measurement conditions or sample characteristics but also on spectral artifacts inherent to all luminescence spectra. This insight extends beyond the specific case of vanadate or phosphovanadate particles, and similar spectral treatments is recommended to eliminate the impact of spurious signals on the thermometric performances of luminescent nanothermometers with overlapped emission bands.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.4c04835.

• Additional experimental details, X-ray diffractograms and results of data refinement, FTIR spectra, TEM images, DLS particle size distributions, emission spectra, spectral analysis of emission bands, additional temperature-dependent spectra, thermometric performance parameters (S_r, δT, and δΔ/Δ), and analysis of spectral overlap on the thermometry quality parameters (PDF)

Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

The Article Processing Charge for the publication of this research was funded by the Coordination for the Improvement of Higher Education Personnel - CAPES (ROR identifier: 00x0ma614).

The authors declare no competing financial interest.