Near-infrared optical nanothermometry via upconversion of Ho3+-sensitized nanoparticles

Recently, materials revealing the upconversion (UC) phenomenon, which is a conversion of low-energy photons to higher-energy ones, have attracted considerable attention in luminescence thermometry due to the possibility of precise and remote optical thermal sensing. The most widely studied type of luminescent thermometry uses a ratiometric approach based on changes in the UC luminescence intensity, mainly of lanthanide ions’ thermally coupled energy levels. In this work, NaYF4:Ho3+@NaYF4, and NaYF4:Ho3+, Er3+@NaYF4 nanoparticles (NPs) were synthesized by the controlled reaction in oleic acid and octadecene at 573 K. The obtained nanoparticles had hexagonal structures, oval shapes, and average sizes of 22.5 ± 2.2 nm and 22.2 ± 2.0 nm, respectively. The spectroscopic properties of the products were investigated by measurements of the UC emission under 1151 nm laser excitation in the temperature range between 295 to 378 K. The sample doped with Ho3+ and Er3+ ions showed unique behavior of enhancing emission intensity with the temperature. The relative sensitivity determined for the NPs containing Ho3+ and Er3+ ions, reached the maximum value of 1.80%/K at 378 K. Here, we prove that the NaYF4:Ho3+, Er3+@NaYF4 system presents unique and excellent optical temperature sensing properties based on the luminescence intensity ratios of the near-infrared bands of both Ho3+ and Er3+ ions.

Sigma Aldrich), and ammonium fluoride (≥ 98%, Fluka) were used as a source of Y 3+ , Ho 3+ , Er 3+ , Na + and F -ions, respectively.The chlorides were placed in the dryer at 348 K for a week to remove the water (the residual water content was determined by TGA analysis).The reaction was carried out in n-octadecene (90% Alfa Aesar) and oleic acid (70% Fisher Chemicals).Ethanol (99.8% POCh S.A.) and n-hexane (≥ 99% POCh S.A.) were used to purify the post-reaction products.
NaYF 4 :7.5%Ho3+ ,7.5%Er 3+ @NaYF 4 preparation: • β-core To obtain 5.5 mmol of β-NaYF 4 :7.5%Ho3+ ,7.5%Er NPs, 110 mL of n-octadecene and oleic acid mixture (1:1), 4.6750 mmol of yttrium chloride, 0.4125 mmol of both holmium and erbium chlorides were purified at 373 K under vacuum for 2.5 h.11.0 mmol of sodium oleate (2× excess) and 33.0 mmol of ammonium fluoride (1.5× excess) were separately added to the heated mixture under nitrogen flow and purified at 373 K under vacuum for 30 and 5 min, respectively.The mixture was heated at 573 K with vigorous stirring, under nitrogen flow for 1 h, and cooled down.The post-reaction product was purified five times by sequential dispersing in n-hexane and precipitating by ethanol (5 min, 8000 rpm).The obtained NPs were dispersed in n-hexane and air-dried for 24 h.

• α-shell
To obtain 15 mmol of α-NaYF 4 NPs, 300 mL of n-octadecene and oleic acid mixture (1:1) and 15 mmol of yttrium chloride were purified at 373 K under vacuum for 3 h.22.5 mmol of sodium oleate (1.5× excess) and 60.0 mmol of ammonium fluoride were separately added to the heated mixture under nitrogen flow and purified at 373 K under vacuum for 45 and 10 min, respectively.The mixture was heated at 473 K with vigorous stirring, under nitrogen flow for 1 h, and cooled down.The post-reaction mixture was centrifuged (10 min, 9000 rpm), the product was precipitated by adding ethanol and purified three times by sequential disperse in n-hexane and precipitated by ethanol (5 min, 8000 rpm).The obtained NPs were air-dried for 36 h.

Characterization. The purity of the products obtained at individual synthesis stages was determined by
Thermogravimetric Analysis (TGA) on Netzsch TG 209 Libra in the temperature range from 298 to 878 K under nitrogen flow (see Fig. S1).The crystalline structures and phase purity of the prepared samples were specified by X-ray Powder Diffraction (XRD) measurements on a Bruker AXS D8 Advance Diffractometer equipped with a Johansson monochromator (λ Cu K α1 = 1.5406Å) and a LynxEye strip detector (step: 0.05° 2θ, step time: 1 s, angular range: 20-100° 2θ).The reference data was taken from JCPDS (00-016-0334).The images of synthesized NPs, based on which the average sizes and size distributions were determined, were recorded on the high-resolution transmission electron microscope Hitachi HT7700 with an accelerating voltage of 120 kV.
Measurements from 295 to 378 K were carried out in a tubular electric furnace (Gero RES-E 230/3), where the sample temperature was controlled via a type K thermocouple in contact with it.The temperature-dependent UC emission spectra of the NaYF 4 :Ho 3+ @NaYF 4, and NaYF 4 :Ho 3+ , Er 3+ @NaYF 4 NPs were obtained using a 10 ns pulsed optical parametric oscillator OPO (EKSPLA/NT342/3/UVE) as the laser source with energy 0.5 mJ.Emissions from the oven were focused on the entrance slit of a spectrograph (Andor SR-303i-A) equipped with a cooled CCD camera (Andor Newton).All spectra were corrected from the spectral response of the equipment.The QuantaMasterTM 40 spectrophotometer equipped with an Opolette 355LD UVDM tunable laser (with a repetition rate of 20 Hz) and a PIXIS:256E digital CCD camera with an SP-2156 imaging spectrograph (Princeton Instruments) was used to measure the dependencies of the energy transitions intensities on the laser energy.The luminescence rise and decay lifetimes were recorded with a Mixed Domain Oscilloscope-200 MHz-Tektronix MDO3022.These measurements were carried out for solid samples at room temperature.

Results and discussion
Structural and morphological properties.The prepared NaYF 4 :Ho 3+ @NaYF 4, and NaYF 4 :Ho 3+ , Er 3+ @ NaYF 4 NPs crystallized as a single hexagonal phase, with the P6 space group (Fig. 1a).The diffraction peaks of the obtained structures align well with the reference patterns (JCPDS 00-016-0334).No significant shifts in the registered diffractograms were observed because the substitution of Y 3+ ions with Ho 3+ or Er 3+ ions did not affect the unit cell parameters, as all ions are of similar size ( r Y 3+ = 1.040Å, r Ho 3+ = 1.041Å, r Er 3+ = 1.030Å, see Table S1) 41 .The samples were characterized by small sizes, around 21 nm (by Scherrer equation), as evidenced by broad peaks in the measured diffractograms.The obtained results agree with the TEM images in Fig. 1b,c.
The obtained NPs had oval shapes and average sizes of 22.2/22.5 nm (calculated based on the TEM results) with narrow size distributions (Fig. 1b,c).In the TEM images, a slightly darker region in the centers of the NPs is visible due to about two times higher densities of the Ho 3+ and Er 3+ ions added to the core compared to Y 3+ ions in the shell.Ultimately, the presence of the core@shell structure of the prepared samples was confirmed by the observed increase in their sizes compared with the core-only NPs (see also Fig. S2).
The UC luminescence spectra of the prepared NPs were measured under 1151 nm excitation in the temperature range from 295 to 378 K (Fig. 3).When the temperature increased, the emission intensity of the NaYF 4 :Ho 3+ @NaYF 4 sample decreased.Such behavior is consistent with the general tendency of thermal quenching of luminescence due to the intensified non-radiative relaxation processes 11,17 .In contrast, the emission intensity of the NaYF 4 :Ho 3+ , Er 3+ @NaYF 4 NPs increased with increasing temperature.The luminescence of this sample resulted from certain energy transfers from Ho 3+ ions to Er 3+ ions, so the increase of their efficiency with increasing temperature could cause the enhancement of their emission.It is worth noting that the thermal increase of Er 3+ ions' luminescence intensity, particularly visible at 982 nm, was more significant than the increase of the emission intensity from only Ho 3+ ions (752, 898 nm).
The parameters characterizing prepared optical temperature sensors were determined using the ratiometric approach.To describe the properties of the NaYF 4 :Ho 3+ @NaYF 4 sample, the luminescence intensity ratio (LIR), which is a ratio of the luminescence intensity from the upper (I U ) and lower (I L ) states, was used 4,42,43 : (1) The luminescence spectra of the NaYF 4 :Ho 3+ , Er 3+ @NaYF 4 sample, resulted from the overlapped emission from the energy levels of both Ho 3+ and Er 3+ ions.In that case, the LIRs were calculated as a ratio of the shorter wavelength peaks' intensities (I s ) to the intensities of the peaks with the longer wavelength (I l ): The most important parameter related to temperature-dependent luminescence, especially for the application in optical temperature sensors, is the relative sensitivity (S R ) of the material to the temperature changes, determined as the rate of LIR changes with the temperature 4,43,44 : The S R curves were plotted based on the LIRs' temperature changes of the selected luminescence bands of both Ho 3+ and Er 3+ ions (NaYF 4 :Ho 3+ , Er 3+ @NaYF 4 NPs) or only Ho 3+ ions (NaYF 4 :Ho 3+ @NaYF 4 NPs).The LIR and S R dependencies were estimated for the NPs' luminescence peaks with the different temperature behaviors and are shown in Fig. 4. In the case of the sample doped solely with Ho 3+ ions, we selected the emission peaks at 489 and 544 nm, 648 and 898 nm, as well as 898 and 970 nm (Fig. 4a,c).When the Ho 3+ and Er 3+ ions were dopants, we also took into account the additional peaks that occurred at similar wavelengths: 489 and 523 + 544 nm, 648 + 672 and 898 nm, as well as 898 and 970 + 982 nm (Fig. 4b,d).
The NaYF 4 :Ho 3+ , Er 3+ @NaYF 4 sample presented relatively high S R values with a maximum equal to 1.80 (378 K)%/K for the NIR peaks at 898 and 970 + 982 nm.In contrast, all determined S R values of the NaYF 4 :Ho 3+ @ NaYF 4 sample were below 1.00%/K.The results indicate that the co-doping with Er 3+ ions to a system based on the Ho 3+ ions significantly improves the temperature sensing properties, particularly in the NIR range.The prepared NPs exhibit minimum temperature uncertainty around 1.08 K (Fig. S4).
Determining the number of photons involved in populating the excited states of the emitting ions is crucial for explaining the mechanism behind the observed spectroscopic properties of the NPs.The photon's numbers (further described as n coefficients) are determined from the dependencies of the luminescence intensities I UC on the excitation power densities P, or in the case of the pulsed laser excitation, from its energies E 45: Luminescence peaks recorded at similar wavelengths showed significant differences in the n coefficient values between both samples (Figs. 5 and S4).The NaYF 4 :Ho 3+ @NaYF 4 NPs had n values mostly between 2.0 and 3.0, which implied that mainly 3 photons are needed to obtain UC emission of Ho 3+ ions.Only the emission at 544 nm was related to the absorption of 4 photons (n = 3.29).The n coefficients for the NaYF 4 :Ho 3+ , Er 3+ @ NaYF 4 NPs were mainly below 2.0, suggesting that the observed UC was primarily influenced by processes that necessitated the absorption of only 2 photons.The exception was green emission, which resulted from 3 photons process (n = 2.14).( 4) To fully understand the nature of the observed UC phenomenon, we measured the luminescence lifetimes of the selected Ho 3+ and Er 3+ transitions (Fig. 6) and used the following equations to calculate average (effective) rise and decay times 46 : where t eff is the effective rise (R) or decay (D) time, t p is the time when the lifetime trend changes from rise to decay and I is the intensity at time t (see insets in Fig. 6).
The UC luminescence of the NaYF 4 :Ho 3+ @NaYF 4 NPs results from energy transfer (ET) processes between Ho 3+ ions, as evidenced by the visible rise times of the registered transitions (Fig. 6a).Moreover, the similar rise times of the transitions responsible for the 1250 752 nm emission (t D values around 20 μs) indicate that the 5 I 6 and 5 I 4 Ho 3+ levels are likewise populated.The 5 F 5 , 5 S 2, and 5 F 3 states related to the emission at 648, 544, and 489 nm are also populated similarly, slower than the previously mentioned, as evidenced by the congruous rise times close to 40 μs.In the case of the NaYF 4 :Ho 3+ , Er 3+ @NaYF 4 sample, the UC luminescence at 1250 nm results from quick processes since the rise time of the 5 I 6 → 5 I 8 transition is not visible.The other emissions, which have rise times below 20 μs, are connected with the energy transfer processes between Ho 3+ and Er 3+ ions.In the recorded lifetime profiles of the Ho 3+ ion transitions, the Er 3+ ion influence is visible, especially for the red emission (see Fig. 6b).
The results of the spectroscopic measurements of the prepared NPs became the basis for the proposed mechanism responsible for the observed UC emission under 1151 nm excitation (Fig. 7).
The irradiation of the NaYF 4 :Ho 3+ @NaYF 4 NPs with the 1151 nm pulsed laser produces Ho 3+ ions in their 5 I 6 excited state via ground state absorption (GSA) process (Fig. 7a).The initially excited Ho 3+ ions exchange energy with each other by the ET processes.The absorption of the subsequent photon leads to the population of the Ho 3+ ions into the 5 I 4 levels and the weak emission at 752 nm.From this state, there is also relaxation to the 5 I 5 state, from which the emission at 898 nm occurs.Simultaneously, the absorption of another photon by the ET process produces the Ho 3+ ions in their 5 F 5 state.The emission at 648 nm ( 5 F 5 → 5 I 8 ) and 970 nm ( 5 F 5 → 5 I 7 ) (5) t eff R =   is from this energy level.The Ho 3+ ions, previously populated to the 5 I 4 level, absorb another photon by the quick ESA process, whereby the Ho 3+ ions are excited to their 5 F 3 energy levels.Hence, the sample presented emissions at 489 nm and after relaxation at 544 nm.
In the case of the NaYF 4 :Ho 3+ , Er 3+ @NaYF 4 NPs, most of the energy absorbed by Ho 3+ ions is transferred to Er 3+ ions (see Fig. 7b).The Er 3+ ions are excited to their 4 F 9/2 levels from where an additional emission is possible giving a peak at 672 nm.Further, intense emission at 982 nm occurs after the relaxation to the 4 I 11/2 state.Another photon absorbed within the Er 3+ ion leads to a population of the 2 H 11/2 , 4 S 3/2 Er 3+ ions levels and an additional emission band at 523 nm.The emission of Ho 3+ ions at 890 and 743 nm results from energy back transfer from the excited Er 3+ ions and is weaker than in the case of a system containing only Ho 3+ ions 47,48 .The emission of Ho 3+ ions in NaYF 4 :Ho 3+ , Er 3+ @NaYF 4 sample probably also consists of the processes occurring in sample NaYF 4 :Ho 3+ @NaYF 4 sample, however, they are less intense because the pre-excited Ho 3+ ions transferred most of the energy to Er 3+ ions.

Conclusions
Using the precipitation reaction in the oleic acid/octadecene solution, we successfully obtained core@shell UCNPs based on sodium yttrium fluorides doped with either Ho 3+ ions or both Ho 3+ and Er 3+ ions.The prepared UCNPs exhibited an oval shape and average sizes of approximately 22.5 nm.
The NaYF 4 :Ho 3+ @NaYF 4 and NaYF 4 :Ho 3+ , Er 3+ @NaYF 4 samples showed UC emission under 1151 nm pulsed laser excitation.We registered the emission of the products in the 295 to 378 K temperature range to determine their temperature-sensing properties.
The NPs containing Ho 3+ and Er 3+ ions revealed unusual behavior manifested by increased luminescence intensity with the temperature increase.This observation can be attributed to the specificity of the UC mechanism based on energy transfers from Ho 3+ to Er 3+ ions.Upon the research, we discovered that the NaYF 4 :Ho 3+ , Er 3+ @NaYF 4 NPs have great potential as a temperature sensor based on the excitation and emission in the range of biological windows.This sample shows intense NIR luminescence from Ho 3+ ions at 899 and 970 nm and Er 3+ ions at 982 nm.The relative sensitivity determined for these peaks reached the maximum value of 1.80%/K at 378 K.This optical temperature sensor based on the NIR UC emission of the system containing Ho 3+ and Er 3+ ions has been reported for the first time.The possibility of excitation within the second biological window and detecting temperature changes in emission intensity around the first biological window make our UCNPs promising candidates for biomedical applications.However, the obtained UCNPs presented high sensitivities not only in the NIR range, which generally makes them excellent candidates for temperature sensing applications not only limited to biological ones.