Bifunctional Tm3+,Yb3+:GdVO4@SiO2 Core-Shell Nanoparticles in HeLa Cells: Upconversion Luminescence Nanothermometry in the First Biological Window and Biolabelling in the Visible

The bifunctional possibilities of Tm,Yb:GdVO4@SiO2 core-shell nanoparticles for temperature sensing by using the near-infrared (NIR)-excited upconversion emissions in the first biological window, and biolabeling through the visible emissions they generate, were investigated. The two emission lines located at 700 and 800 nm, that arise from the thermally coupled 3F2,3 and 3H4 energy levels of Tm3+, were used to develop a luminescent thermometer, operating through the Fluorescence Intensity Ratio (FIR) technique, with a very high thermal relative sensitivity. Moreover, since the inert shell surrounding the luminescent active core allows for dispersal of the nanoparticles in water and biological compatible fluids, we investigated the penetration depth that can be realized in biological tissues with their emissions in the NIR range, achieving a value of 0.8 mm when excited at powers of 50 mW. After their internalization in HeLa cells, a low toxicity was observed and the potentiality for biolabelling in the visible range was demonstrated, which facilitated the identification of the location of the nanoparticles inside the cells, and the temperature determination.


Introduction
Accurate temperature measurements at the nanoscale are important in many industrial processes, as well as in medicine [1,2]. Sometimes, reaching the object whose temperature has to be measured is difficult or even impossible. In this context, noncontact thermometry can be used. Among the different noncontact thermometry techniques, luminescence thermometry offers high spatial and thermal resolutions [3,4]. By using this technique, it might be possible to detect cancer cells at an early stage of development, just by monitoring the temperature rise in the body due to accelerated metabolic activities in abnormal cells [2]. Moreover, luminescence thermometry has been used to control temperature during photothermal therapy, avoiding over-heating damage [5]. All this has been that modify their physical properties. Thus, after all, it was not clear if the temperature measured inside the cells was the correct one, and the thermal resolution that could be achieved was not higher than the one obtained using luminescent nanothermometers operating in the visible range.
As pointed out above, fluorescent labelling is a widely used and indispensable tool in biology. In this context, lanthanide-doped up-converting nanoparticles, with emissions in the visible range, allow us to overcome some of the disadvantages faced by conventional bio-labelers excited in the UV or in the deep blue region of the electromagnetic spectrum. Among the advantages introduced by lanthanide-doped up-converting nanoparticles, it is important to note here very low autofluorescence, absence of photo damage to living organisms, high detection sensitivity and high depth of light penetration. All these properties make them ideal fluorescent labels for bioimaging [22][23][24][25].
Among the different host matrices for Ln 3+ ions, GdVO 4 is characterized by high optical absorption and emission cross-section values for the Ln 3+ -doping ions, high thermal stability, and moderate cut-off phonon energy (700 cm −1 ) [26].
Here we present a study of the temperature-dependent, NIR-excited upconversion emissions lying in the I-BW, grounding the development of a luminescent thermometer operating in this spectral region, consisting of Tm,Yb:GdVO 4 @SiO 2 core-shell nanoparticles prepared by a soft hydrothermal route, that could also be used as up-conversion biolabels operating in the visible range. Different concentrations of Tm 3+ as the active ion were studied in order to optimize thermal sensing. Internalization in HeLa cells, which is facilitated by the inert silica shell, was analyzed, demonstrating a low cytotoxicity. The characteristic Tm 3+ blue emission of these nanoparticles was visualized through photoluminescence microscopy. This emission can be used for biolabeling purposes since it allows easy identification of the location of the particles inside HeLa cells. This facilitated the identification of the region in which the temperature inside the cells could be measured using the luminescent thermometer, developed with these core-shell nanoparticles, operating in the I-BW. The results presented prove the potentiality of Tm,Yb:GdVO 4 @SiO 2 core-shell nanoparticles as multifunctional platforms, with applications in NIR-to-NIR upconversion temperature-sensing in the I-BW with high thermal sensitivity, and NIR-to-visible upconversion biolabeling.

Materials and Methods
2.1. Synthesis of Tm,Yb:GdVO 4 @SiO 2 Core-Shell Nanoparticles Gd 0.85−x Yb 0.15 Tm x VO 4 (x = 0.5-3 mol%) nanoparticles were prepared following a soft hydrothermal synthesis method. Starting reagents consisted of Gd 2 O 3 , Yb 2 O 3 , Tm 2 O 3 (Strem Chemicals, Newburyport, MA, USA, 99.9%) and NH 4 VO 3 (Sigma Aldrich, St. Louis, MO, USA, 99%). We prepared Ln 3+ -nitrates by dissolving the required amounts of the corresponding oxides in a nitric acid solution (50% distilled H 2 O:50% HNO 3 at 69%), and heating them until complete dryness. Then, we dissolved these Ln 3+ -nitrates in 10 mL of distilled H 2 O and added them to a solution previously prepared of NH 4 VO 3 in 20 mL of distilled water. We then adjusted the pH of the resulting solution to 7 with diluted NH 4 OH. After 15 min of magnetic stirring, the dispersion formed was heated in a Teflon-lined autoclave at 458 K for 24 h. The resulting product was collected by centrifugation, washed with distilled water several times, and dried overnight at 393 K. Finally, the samples were annealed at 873 K for 5 h to remove surface defects formed during the low-temperature synthesis process and to promote a higher degree of crystallization [27]. With the aim of achieving a further improvement of the optical emission efficiency of these nanoparticles, they were coated with a SiO 2 layer with a thickness of 5-7 nm. The coating process was performed by dispersing the Gd 0.85−x Yb 0. 15

Structural and Morphological Characterization
Powder X-ray diffraction (XRD) was performed on Tm,Yb:GdVO 4 and Tm,Yb:GdVO 4 @SiO 2 core-shell nanoparticles. For that a Bruker (Billerica, MA, USA) AXS D-8 Advance diffractometer, using Cu Kα radiation and operating at room temperature, was used.
The morphology of the nanoparticles was characterized by conventional and high-resolution transmission electron microscopy (TEM and HRTEM, respectively) with JEOL (Tokyo, Japan) JEM2100 and JEOL JEM3000F microscopes, operating at 200 kV and 300 kV accelerating voltages, respectively.
Hydrodynamic particle size distribution was measured by dynamic light scattering (DLS, Vasco 2, Cordouan Technologies, Pessac, France), and Fourier transform infrared absorption (FT-IR, Nicolet 20SXC, Nicolet Instruments Corp., Madison, WI, USA) spectra was used to investigate the formation of the Tm,Yb:GdVO 4 @SiO 2 structure and the presence of adsorbed species on the surface of the nanoparticles. These results can be consulted elsewhere [27].

Temperature-Dependent Photoluminescence Measurements
Tm,Yb:GdVO 4 @SiO 2 core-shell nanoparticles were placed in a Linkam (Tadworth, UK) THMS 600 heating stage for temperature dependent photoluminescence measurements. They were excited by a fiber-coupled laser diode with emission at 980 nm and a power of 50 mW. The laser beam was focused on the sample by a 40× microscope objective with N.A. = 0.6 providing a spot size of around 10 µm, limiting the excitation density to a value for which we did not observe local thermal effects caused by the laser, and which did not affect the temperature reading measurements by introducing artifacts. The emission signal was collected by the same microscope objective, and after passing a dichroic filter to eliminate the excitation wavelength, was sent to an AVANTES (Apeldoorn, The Netherlands) AVS-USB2000 fiberoptic spectrometer to record it.

Subtissue Spectroscopic Measurements
The subtissue penetration depth that could be achieved by using the Tm,Yb:GdVO 4 @SiO 2 core-shell nanoparticles was investigated by using a double beam fluorescence microscope. This experimental set-up allowed us to place a phantom tissue of variable thickness between the Tm,Yb:GdVO 4 @SiO 2 core-shell nanoparticles and the detection optics. The penetration depth was obtained by monitoring the infrared luminescence generated at different tissue thicknesses in Intralipid as phantom tissue. Intralipid is an absorbing and scattering medium that has been extensively used in the past to mimic the optical properties of human skin in the I-and II-BWs. In this work we used Intralipid 10%, diluted to a concentration of 2%. At this concentration, the wavelength dependence is similar to that previously reported for several human tissues [28]. The cells were incubated at 310 K in a humidified 5% CO 2 atmosphere. The medium was changed daily. For the fluorescence observation, cells were placed onto round coverslips placed into 24 wells plates.

In Vitro Experiments
For evaluation of Tm,Yb:GdVO 4 @SiO 2 core-shell nanoparticles on an in vitro thermal scanning experiment, the cells were seeded in 24 well plates containing sterile round coverslips and were kept in a sterile environment in the incubator for 24 h until the experiment started. After 2 h of incubation with Tm,Yb:GdVO 4 @SiO 2 nanoparticles, the cells were washed with phosphate-buffered saline (PBS) solution, and used in luminescence experiments.

In Vitro Cell Cytotoxicity/Viability Studies
To determine cell cytotoxicity/viability, the cells were placed in a 24 well plate at 310 K in 5% CO 2 atmosphere. After 48 h of culture, the medium in the well was replaced with a fresh medium containing the Tm,Yb:GdVO 4 @SiO 2 nanoparticles in a volume ratio ranging from 1:50 to 1:500, and cells were incubated for 2 h. After incubation, the medium was removed and a new complete medium without nanoparticles was added. After 24 h, 0.5 mL of 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium-bromide (MTT) dye solution (0.05 mg mL −1 of MTT, Sigma Aldrich, St. Louis, MO, USA) was added to each well. After 2-3 h of incubation at 310 K and 5% CO 2 , the medium was removed, and formazan crystals were solubilized with 0.5 mL of dimethylsulphoxide (DMSO). The solution was vigorously mixed to dissolve the reacted dye. The absorbance of each well was read at 540 nm on a Spectra Fluor4 (Tecan Group Ltd., Männedorf, Switzerland) microplate reader. The spectrophotometer was calibrated to zero absorbance, using culture medium without cells. The relative cell viability (%) related to control wells containing cell culture medium without nanoparticles was calculated by [A] test /[A] control × 100, where [A] test is the absorbance of the test sample and [A] control is the absorbance of the control sample. The MTT assay is a simple non-radioactive colorimetric assay to measure cell cytotoxicity, proliferation or viability. MTT is a yellow, water soluble, tetrazolium salt. Metabolically active cells are able to convert this dye into a water-insoluble, dark blue formazan, by reductive cleavage of the tetrazolium ring [29]. Formazan crystals can then be dissolved in an organic solvent such as DMSO, and quantified by measuring the absorbance of the solution at 540 nm, and the resultant value is related to the number of living cells.

Results
All pale-yellow products obtained after the hydrothermal reaction were shown to be isostructural to the tetragonal I4 1 /amd zircon-type phase of GdVO 4 (JCPDS File 86-0996), indicating a 100% synthesis yield. The crystal phase was maintained after 5 h annealing at 873 K, although with narrower Bragg peaks. This indicates an increase of the average size of individual crystalline domains that constitute vanadate nanoparticles. No additional reflections were detected for SiO 2 -coated samples, apart from a rougher background reflecting the presence of amorphous SiO 2 , as compared to the XRD of the bare Tm,Yb:GdVO 4 annealed nanoparticles (NPs), included in Figure 1a for comparison. The synthesis was repeated several times, resulting always in the same product with reproducible properties. Figure 1b shows a characteristic TEM image of an annealed sample of Tm,Yb:GdVO 4 NPs, which presents polygonal forms, mainly of square or rectangular sections, and lengths of 25-40 nm. Figure 1c,d display HRTEM images of Tm,Yb:GdVO 4 @SiO 2 NPs, with darker and lighter parts being the Tm,Yb:GdVO 4 core and the amorphous silica coating, respectively. We determined that the coating layer had an average thickness of~7 nm around the NPs. HRTEM images of discrete nanoparticles reveal a well-defined crystalline structure in the core, with a lattice fringe distance of 0.267 nm, as can be seen in Figure 1d, matching the (112) interplanar spacing of GdVO 4 , according to the JCPDS File 86-0996. Figure 2a shows a schematic representation of the Tm,Yb:GdVO 4 @SiO 2 core-shell nanoparticles to illustrate their potential use as biolabels (operating in the visible range) and luminescence thermometers (operating in the I-BW). Figure 2b shows the upconversion emission spectra of 1% Tm 3+ , 15% Yb 3+ :GdVO 4 @SiO 2 core-shell nanoparticles, under 980 nm excitation at room temperature and 333 K. In the figure, we indicate the spectral range corresponding to the I-BW in yellow. As can be seen, when the temperature increased, all emission bands drop in intensity, with the exception of the band at 700 nm that slightly increases its intensity. In order to explain why this is happening, first it is necessary to identify the energy levels assigned to each of the radiative transitions. Figure 2c shows the energy level diagram of the Tm 3+ and Yb 3+ ions, indicating the pathways for the 980 nm excited upconversion process, and the transitions involved in the generation of such spectra. In a first step, the excitation at 980 nm is absorbed by Yb 3+ promoting its electrons from the 2 F 7/2 fundamental state to the 2 F 5/2 excited state. Then, part of this energy is transferred to the 3 H 5 energy level of Tm 3+ , from which electrons relax very fast to the 3 F 4 energy level. Then, a second energy transfer from Yb 3+ promotes Tm 3+ electrons to the 3 F 2 energy level, which relaxes, populating the 3 F 3 and 3 H 4 energy levels. Finally, a third energy transfer process promotes Tm 3+ electrons in the 3 H 4 level to the 1 G 4 energy level, from which the blue and red emissions, centered at 475 and 650 nm, arise through radiative 1 G 4 → 3 H 6 and 1 G 4 → 3 F 4 transitions, respectively. The emissions at 700 nm and 800 nm arise from the 3 F 3 → 3 H 6 and 3 H 4 → 3 H 6 radiative transitions. Along with the non-radiative relaxation from the 3 F 2 level, Tm-Tm cross-relaxation processes have been previously reported which involve depopulation of the 1 G 4 level and population of the 3 F 3 and 3 H 4 levels. These mechanisms are favored as Tm 3+ concentration and temperature increase [30]. Furthermore, a non-radiative decay after the second energy transfer from the 2 F 5/2 level of Yb 3+ allows for populating the 3 H 4 level of Tm 3+ very efficiently, since its energy gap (~3400 cm −1 ) matches the phonon energy of OHgroups adsorbed on the silica surface [27]. This privileged population of the 3 H 4 level explains the strong NIR emission intensity observed at 800 nm.   Figure  1c,d display HRTEM images of Tm,Yb:GdVO4@SiO2 NPs, with darker and lighter parts being the Tm,Yb:GdVO4 core and the amorphous silica coating, respectively. We determined that the coating layer had an average thickness of ~7 nm around the NPs. HRTEM images of discrete nanoparticles reveal a well-defined crystalline structure in the core, with a lattice fringe distance of 0.267 nm, as can be seen in Figure 1d, matching the (112) interplanar spacing of GdVO4, according to the JCPDS File 86-0996. Nanomaterials 2020, 10, x; doi: www.mdpi.com/journal/nanomaterials depopulation of the 1 G4 level and population of the 3 F3 and 3 H4 levels. These mechanisms are favored as Tm 3+ concentration and temperature increase [30]. Furthermore, a non-radiative decay after the second energy transfer from the 2 F5/2 level of Yb 3+ allows for populating the 3 H4 level of Tm 3+ very efficiently, since its energy gap (3400 cm −1 ) matches the phonon energy of OHgroups adsorbed on the silica surface [27]. This privileged population of the 3 H4 level explains the strong NIR emission intensity observed at 800 nm. (b) Upconversion emission spectra of the Tm,Yb:GdVO4@SiO2 core-shell nanoparticles at room temperature and at 333 K. Note that the visible part of the spectra, from 375 to 675 nm, has been magnified 10 times to respect the NIR part. The inset shows the magnification of the peak located at 700 nm, to make more evident the variation of the intensity of this peak with temperature. (c) Energy level diagram of Tm 3+ and Yb 3+ ions in GdVO4, indicating the absorption, energy transfer and emission pathways. (b) Upconversion emission spectra of the Tm,Yb:GdVO 4 @SiO 2 core-shell nanoparticles at room temperature and at 333 K. Note that the visible part of the spectra, from 375 to 675 nm, has been magnified 10 times to respect the NIR part. The inset shows the magnification of the peak located at 700 nm, to make more evident the variation of the intensity of this peak with temperature. (c) Energy level diagram of Tm 3+ and Yb 3+ ions in GdVO 4 , indicating the absorption, energy transfer and emission pathways.
The energy gap between the 3 F 3 and 3 H 4 energy levels is 2046 cm −1 [31]. The relative low energy difference between these two electronic levels allows the existence of a thermal equilibrium between their electronic populations governed by the Boltzmann law. Thus, we can consider that the 3 F 3 and 3 H 4 energy levels are thermally coupled, and can be used for temperature determination using the fluorescence intensity ratio (FIR) technique [32]. Figure 2b shows how the intensity of the peak lying at 700 nm, and arising from the 3 F 3 → 3 H 6 transition, increases as the temperature increases, while the intensity of the peak lying at 800 nm, and arising from the 3 H 4 → 3 H 6 transition, decreases as the temperature increases. This observation would prove the thermal coupling between the 3 F 3 and 3 H 4 energy levels.
Since upconversion luminescence is a non-linear process, the intensity of the emission bands depends on the dopant concentrations, as well as on temperature. To optimize the intensity ratio between the emission lines at 700 and 800 nm, and to obtain the maximum thermal sensitivity, we analyzed the evolution of the FIR with temperature for the Tm,Yb:GdVO 4 @SiO 2 core-shell nanoparticles doped with different concentrations of Tm 3+ . In a previous study we optimized the concentration of Yb 3+ (~15 mol%) to maximize the intensity of the Tm 3+ emissions in these core-shell nanoparticles [27], thus in the present work the concentration of Yb 3+ was kept constant at 15 mol%. The normalized FIR Nanomaterials 2020, 10, 993 8 of 17 between the emission lines at 700 and 800 nm of the x mol% Tm, 15 mol% Yb:GdVO 4 @SiO 2 core-shell nanoparticles, with x = 0.5-3 mol%, is shown in Figure 3a. The experimental points were fitted to a Bolzmann distribution equation with an additional offset that takes into account the partial overlapping of the emission lines as stated in Equation (1) [32]: where ∆E is the energy gap between the two thermally coupled levels, k is the Bolzmann constant, T is the absolute temperature, and A and C are fitting constants. As can be seen in Figure 3a, the core-shell nanoparticles with a 1 mol% Tm 3+ showed the highest slope.
Since upconversion luminescence is a non-linear process, the intensity of the emission bands depends on the dopant concentrations, as well as on temperature. To optimize the intensity ratio between the emission lines at 700 and 800 nm, and to obtain the maximum thermal sensitivity, we analyzed the evolution of the FIR with temperature for the Tm,Yb:GdVO4@SiO2 core-shell nanoparticles doped with different concentrations of Tm 3+ . In a previous study we optimized the concentration of Yb 3+ (~15 mol%) to maximize the intensity of the Tm 3+ emissions in these core-shell nanoparticles [27], thus in the present work the concentration of Yb 3+ was kept constant at 15 mol%. The normalized FIR between the emission lines at 700 and 800 nm of the x mol% Tm, 15 mol% Yb:GdVO4@SiO2 core-shell nanoparticles, with x = 0.5-3 mol%, is shown in Figure 3a. The experimental points were fitted to a Bolzmann distribution equation with an additional offset that takes into account the partial overlapping of the emission lines as stated in Equation (1) [32]: where ΔE is the energy gap between the two thermally coupled levels, k is the Bolzmann constant, T is the absolute temperature, and A and C are fitting constants. As can be seen in Figure 3a, the coreshell nanoparticles with a 1 mol% Tm 3+ showed the highest slope. The absolute thermal sensitivity can be calculated from the first derivative of the FIR fittings with respect to the temperature, and might be used for comparison of the same material with different concentrations of dopants [32], as expressed in Equation (2): The thermal sensitivities calculated for the Tm,Yb:GdVO 4 @SiO 2 core-shell nanoparticles are presented in Figure 3b. As can be seen, and as expected from the slopes of FIR obtained in Figure 3a, the nanoparticles with 1 mol% Tm 3+ exhibit the highest absolute thermal sensitivity, with a maximum of 0.037 K −1 at 333 K. Table 1 shows the comparison of the results we obtained with those reported previously for other Tm 3+ -doped systems using the same intensity ratio. As can be seen, the absolute thermal sensitivity of Tm,Yb:GdVO 4 @SiO 2 core-shell nanoparticles is the highest among Tm 3+ , Yb 3+ co-doped systems, being one or two orders of magnitude higher than the S abs reported for most of the other luminescent thermometers. The only exception is Tm,Yb:YAG fiber optics [33], with a S abs of the same order of magnitude, but still 76% smaller than the one obtained in this work, being the highest ever reported for Tm 3+ , Yb 3+ co-doped systems operating in the first biological window. Table 1.
Comparison of the performance of Tm 3+ -doped systems used in luminescence nanothermometry operating totally or partially in the I-BW.

Material
Temperature The better performance of our core-shell nanoparticles in terms of thermal sensitivity might be related to the SiO 2 inert shell layer, as we showed previously in Er,Yb:GdVO 4 @SiO 2 nanoparticles [11]. The inert shell structure preserves the luminescence generated by the active core, and prevents interactions with solvent molecules that might quench the emissions of Tm 3+ . It also avoids the progressive heating of the nanoparticles when exposed to extended excitation periods with the laser emitting at 980 nm, and preserves their detrimental effects for thermometric applications [11].
Although S abs is useful for comparing luminescent thermometers operating under the same conditions, the relative thermal sensitivity (S rel ) allows for the comparison with other systems, operating in different spectral ranges and under different conditions. The relative thermal sensitivity is defined as [1] stated in Equation (3): Figure 3c shows S rel calculated for our particles, with a value of 3.3% K −1 at 298 K. Since it depends on the ∆E value, that does not change substantially when changing the doping concentration, only a single value can be given for all the nanoparticles analyzed. The S rel values reported in the literature are between 1.0 and 28% K −1 , with most of the values being in the range 2.0-3.3% K −1 [34][35][36][37][38][39][40]. The value we report here for Tm,Yb:GdVO 4 @SiO 2 nanoparticles is the highest reported up to now in the literature for Tm 3+ -doped systems, only surpassed by that of Tm:NaYbF 4 @SiO 2 nanoparticles, although in that case, the maximum thermal relative sensitivity has been found at 100 K, far away from the biological range of temperatures. In fact, S rel , when determined at RT, takes a similar value to those reported for the rest of materials listed in Table 1. A higher value was reported for Tm,Yb:Y 2 O 3 sub-micron size particles, although in this last case, the electronic levels from which the emission bands to be used as luminescent thermometers arise are not thermally coupled, and thus a modification of the FIR model has to be used [41]. This does not allow for a direct comparison with the rest of the materials included in the table, since they do not operate purely in the I-BW.
The temperature resolution can be calculated from S rel as stated in Equation (4) [20]: where δFIR/FIR is the relative uncertainty in the determination of the thermometric parameter, FIR in our case, and depends on the acquisition setup. Here, we use the typical value of 0.5% [13]. Figure 3c also shows δT for our particles. Table 1 also lists the δT values reported in the literature or calculated by us from the data reported in previous publications. As can be seen, the value of the temperature resolution we report for Tm,Yb:GdVO 4 @SiO 2 nanoparticles is among the smallest reported for Tm 3+ -doped systems, although we found several materials exhibiting a similar temperature resolution value, from 0.16 to 0.18 K, including Tm,Yb:LiNbO 3 single crystals, and Tm,Yb:YF 3 microcrystals. However, if we take into account the size of the material, the δT value we report is the smallest for Tm 3+ -doped nanoparticles operating in the I-BW.
Finally, to assess the performance of our luminescent nanothermometers, we determined their reproducibility, i.e., the change of the same measurement carried out under modified circumstances. For that, we recorded the spectra and determined FIR in 10 heating-cooling cycles between 298 and 310 K. This reproducibility is higher than 99%, as can be seen in Figure 3d, computed as stated in Equation (5) [20]: where FIR c is the FIR mean value, and FIR i is the value of FIR measured at each temperature. Motivated by the good results of the thermal sensitivity and temperature resolution obtained, we analyzed the potentiality of using Tm,Yb:GdVO 4 @SiO 2 core-shell nanoparticles in biological systems. To analyze how the fluorescence intensities of the Tm 3+ luminescence bands were attenuated by the biological tissue, we used a phantom tissue of variable thickness. A drop of a dispersion of the core-shell nanoparticles in PBS was placed in a microscope slide, and was optically excited by a continuous fiber-coupled diode laser at 980 nm. The laser beam was focused on the core-shell nanoparticles using a 20× microscope objective (N.A. = 0.4). The same objective was used to collect the Tm 3+ emission. After passing a dichroic mirror to eliminate the excitation radiation, the emission at 800 nm was analyzed using a highly sensitive Si CCD camera (Synapse, Horiba, Kyoto, Japan) attached to a high-resolution monochromator (iHR320, Horiba, Kyoto, Japan). Intralipid 10%, diluted in a concentration of 2% in water, was used as phantom tissue. Figure 4 shows the collected emission intensity for the Tm 3+ emission band as a function of the tissue thickness. To determine the penetration depth, phantom tissue slices of different thicknesses were placed between the focusing/collecting microscope objective and the Tm,Yb:GdVO 4 @SiO 2 core-shell nanoparticles dispersion. As can be seen, a monotonous decrease of the collected signal with the tissue thickness is observed, indicating that the emission band of Tm 3+ is attenuated within the phantom tissue either by absorption or scattering processes (or by a combination of them) [42]. The results show that a significant fluorescence signal could still be obtained up to a tissue thickness of 0.8 mm, defining this as the maximum penetration depth that can be achieved in this case. The results are smaller to those obtained in Tm,Yb:CaF 2 nanoparticles (maximum penetration depth of 2 mm) using emission lines in the same spectral region [18], probably because fluorides are more efficient emitters when upconversion processes are considered, due to their lower phonon energies [43,44]. However, we believe that this penetration depth might be increased by using a detection set-up with a higher sensitivity, and/or by increasing the power of the excitation radiation, which was limited to 50 mW in these experiments. The excitation power value used is still three orders of magnitude smaller than the one that would cause laser-induced local heating in water solutions [45]. Thus, this leaves room to achieve deeper penetration depths by increasing the excitation power, without causing damage to living biological tissues. However, we limited the power of the excitation radiation, in this case to 50 mW, since this is the power with which incubated HeLa cells were illuminated in the following experiments. Thus, these results demonstrate that the penetration depth achieved with this power is enough for the temperature measurements inside the cells.
Nanomaterials 2020, 10, x; doi: www.mdpi.com/journal/nanomaterials 0.8 mm, defining this as the maximum penetration depth that can be achieved in this case. The results are smaller to those obtained in Tm,Yb:CaF2 nanoparticles (maximum penetration depth of 2 mm) using emission lines in the same spectral region [18], probably because fluorides are more efficient emitters when upconversion processes are considered, due to their lower phonon energies [43,44]. However, we believe that this penetration depth might be increased by using a detection set-up with a higher sensitivity, and/or by increasing the power of the excitation radiation, which was limited to 50 mW in these experiments. The excitation power value used is still three orders of magnitude smaller than the one that would cause laser-induced local heating in water solutions [45]. Thus, this leaves room to achieve deeper penetration depths by increasing the excitation power, without causing damage to living biological tissues. However, we limited the power of the excitation radiation, in this case to 50 mW, since this is the power with which incubated HeLa cells were illuminated in the following experiments. Thus, these results demonstrate that the penetration depth achieved with this power is enough for the temperature measurements inside the cells. Normalized intensity of the upconversion emission at 800 nm, generated by Tm,Yb:GdVO 4 @SiO 2 core-shell nanoparticles in a 2% intralipid aqueous solution to respect the intensity value measured at the surface of the phantom tissue as a function of the penetration depth (continuous line was done for guiding the eyes).
For in vitro experiments, we incubated HeLa cancer cells with Tm,Yb:GdVO 4 @SiO 2 core-shell nanoparticles, as described in the Experimental Section. After incubation, the HeLa cells were excited in a confocal microscope illuminated with a continuous fiber-coupled diode laser at 980 nm. The excitation power was limited at 50 mW, and the spot size on the sample was increased to 3 mm to avoid damages to the living cells. The fluorescence was detected by the same system described before. Figure 5a shows an optical transmission microscope image of the HeLa cells after incubation. The arrow indicates the position of the nanoparticles when the excitation laser is off. Figure 5b shows an optical transmission microscope image of the same region when the excitation laser is on. The blue emission arising from the Tm,Yb:GdVO 4 @SiO 2 core-shell nanoparticles can be seen. This suggests that the nanoparticles are incorporated into vesicles. This was verified by the intensity recorded for different nanoparticles, getting a higher and clearer image for those located outside of the cells, as can be seen in Figure S1 in the Supporting Information. An enlarged image of the area in which the nanoparticle is located is shown also in Figure 5c. Here, some tiny signs of autofluorescence arising from the cell seem to be present, however, they do not prevent the observation of the emission arising from the luminescent nanoparticle. We performed an intensity scan of the blue emission along the line indicated in Figure 5c to corroborate that it was only arising from the luminescent nanoparticles. The results show an increase of the intensity in the area where the nanoparticles are located, but not in the rest of the cell. The red dashed line in the graph indicates the level of noise. These results demonstrate that the Tm,Yb:GdVO 4 @SiO 2 core-shell nanoparticles can be used as effective biolabels for HeLa cells in bioimaging experiments through an upconversion mechanism.  The visualization of the location of the luminescent nanoparticles through the emission of visible light allowed us to simplify the task of measuring the temperature inside the cells through the NIR emissions generated by the same core-shell nanoparticles. A characteristic spectrum in this region is shown in Figure S2 in the Supplementary Materials. The temperature that could be extracted from these spectra, by taking into account Equation (1) and the fitting function result in Figure 3a, was 308 ± 2 K after 5 measurements, coinciding with the incubation temperature of the HeLa cells. The spot of 3 mm used under the excitation conditions to avoid photo damages in the living cells did not allow for higher resolution temperature measurements.
At this point, we should note that we performed a number of different toxicity assays. We found that incubation of HeLa cells with medium solutions containing Tm,Yb:GdVO 4 @SiO 2 core-shell nanoparticles caused a reduced toxicity, as is shown in Figure 5d, for the different concentrations analyzed.
These results prove that Tm,Yb: GdVO 4 @SiO 2 core-shell nanoparticles can be used as thermal probes operating in the I-BW, with very good thermal sensitivity and temperature resolution, and with possible use in determining the temperature inside living HeLa cells. Table 2 shows the lanthanide-doped materials that have been used for cell thermometry up to now, to the best of our knowledge. As can be seen, a very reduced number of articles have been published with this aim. Most of them are based on metallic complexes with organic molecules [46][47][48][49], pumped in the UV range, with the subsequent autofluorescence generation in living cells. Only a couple of them used luminescent nanoparticles of NaYF 4 doped with Er 3+ and Yb 3+ ions, that can be pumped in the NIR at 980 nm, while still emitting in the visible region, especially in the green and red regions [50,51]. However, the thermal resolution that can be achieved with these materials in living cells has not been reported. Thus, there is still room to develop new luminescent materials that can be used as luminescent nanothermometers to determine temperature inside living cells, especially when exploring new spectral areas like the BWs, which are almost unexplored within this context. Thus, the use of luminescent nanoparticles emitting in the I-BW, exhibiting higher S rel and smaller δT, might improve the performance of the luminescent nanothermometers used up to now in cell thermometry. At the same time, the reduced light scattering in this spectral region would allow for a better spatial resolution, or at least better image definition, if thermal mapping in the I-BW is performed. In this context, the development of semiconducting polymers for imaging-guided photothermal therapy, using afterglow emissions pumped in the I-BW, is inspiring [52,53]. These semiconducting nanoparticles not only allowed the performance of ultrasensitive in vivo imaging by removing tissue autofluorescence, but they can also be used as afterglow luminescent thermometers and photothermal agents, being the first semiconducting polymer nanoparticles with this dual function. However, two different excitation sources had to be used, one for exciting the afterglow emissions that allowed the determination of temperature, and another one for promoting the release of heat, leaving room for future improvement of the material design.

Conclusions
In summary, we investigated the upconversion emissions generated by Tm,Yb:GdVO 4 @SiO 2 core-shell nanoparticles, in the physiological range of temperatures, doped with different concentrations of Tm 3+ after excitation at 980 nm. The two emission bands located at 700 and 800 nm, that lie in the I-BW, and arise from the thermally coupled 3 F 2,3 and 3 H 4 energy levels of Tm 3+ , respectively, allowed us to develop a luminescent thermometer operating via the FIR technique. The nanoparticles doped with 1 mol% Tm 3+ exhibited the highest thermal absolute sensitivity for this purpose, with a maximum of 0.037 K -1 and an excellent thermal relative sensitivity of 3.3% K −1 at 333 K, which are within the highest values reported for Tm 3+ , Yb 3+ co-doped systems, to the best of our knowledge. A penetration depth of 0.8 mm in biological tissues could be achieved by the 800 nm emission under excitation at 980 nm with a power of 50 mW. Furthermore, the inert silica shell surrounding the luminescent active core allowed the dispersal of the nanoparticles in biological compatible fluids like PBS, which facilitated their internalization in HeLa cells, proving their potential for biolabelling applications. Thus, Tm,Yb:GdVO 4 @SiO 2 core-shell nanoparticles can be considered as a multifunctional platform for NIR-to-visible upconversion biolabeling, and NIR-to-NIR upconversion thermal sensing. This procedure allows a high thermal resolution, of the order of 0.15 K, resulting in the high thermal sensitivity achieved when compared with the performances of other luminescent thermometers operating in the visible range, traditionally used to perform cell thermometry. An additional advantage that should be explored in detail in the future is the lower spatial resolution that can be achieved in this spectral region due to the reduced light scattering. This would allow for discriminating, with a better accuracy, the differences in temperature and the thermal gradients generated within the organelles of living cells, and the cytoplasm, the nucleus and the membrane, permitting us to better understand the processes occurring inside the cells from a thermal point of view.
Supplementary Materials: The following are available online at http://www.mdpi.com/2079-4991/10/5/993/s1, Figure S1: Microscope optical transmission image of the HeLa cells incubated with Tm,Yb:GdVO4@SiO2 core-shell nanoparticles when the excitation laser is on. Figure S2: Typical luminescence spectrum recorded for the Tm,Yb:GdVO4@SiO2 core-shell nanoparticles in the I-BW after excitation at 980 nm with a power of 50 mW, and a spot size of 3 mm in the sample.
Author Contributions: O.S. has performed the spectroscopic and thermometric characterization of the samples analyzed in this paper, prepared the figures for the paper, and designed and prepared the first drafts of the text. J.J.C.M. has coordinated the paper, working also in the discussion of the different parts, and revised it. C.C. synthesized the nanoparticles used in this paper and performed the structural characterization by XRD and the morphological characterization by TEM and HRTEM, and contributed substantially to the discussion of the mechanisms involved in the emission of the nanoparticles used. P.H.-G. designed and coordinated the spectroscopic characterization of the nanoparticles operating as luminescent thermometers in the biological window when dealing with biological specimens. F.S.-R. designed, coordinated and performed the in vitro cell toxicity/viability tests, and was also responsible for the incubation of the nanoparticles for its internalization in the cells. M.A. participated in the discussion of the structural and morphological characterization of the samples, and revised the paper. F.D. participated in the discussion of the characterization of the nanoparticles as luminescent thermometers, and revised the paper. All authors have read and agreed to the published version of the manuscript.