NIR-to-NIR and NIR-to-Vis up-conversion of SrF₂:Ho³⁺ nanoparticles under 1156 nm excitation

Holmium oxide (Ho₂O₃) (99.99%, Stanford Materials, United States) was dissolved in hydrochloric acid (HCl) (ultra-pure, POCh S.A., Poland) to obtain a 1 M solution of holmium chloride (HoCl₃), which was used as a source of Ho³⁺ ions. Strontium chloride hexahydrate (SrCl₂·6H₂O) (≥99%, Acros Organics, Poland) and ammonium fluoride (NH₄F) (≥98%, Sigma Aldrich, Poland) were used as the sources of the Sr²⁺ and F^– ions, respectively. Trisodium citrate dihydrate (NaCit·2H₂O) (≥97%, Sigma Aldrich, Poland) was added to the reaction mixture as the capping agent. All the substrates were used as received, without further purification. The reaction was carried out in deionized water. A mixture of ethanol (99.8% POCh S.A., Poland) and deionized water (1:1), was used to purify the obtained products.

To obtain SrF₂:xHo³⁺ (x = 5, 7.5, 10, 15, 20 or 25%) NPs, 5 ml SrCl₂ and HoCl₃ solutions containing in total 3.5 mmol Sr²⁺ and Ho³⁺ ions (SrCl₂: 3.325, 3.238, 3.150, 2.975, 2.800, and 2.625 mmol, and HoCl₃: 0.175, 0.262, 0.350, 0.525, 0.700, and 0.875 mmol) were prepared. Next, 20 ml of the 1 M NaCit solution was added to each mixture containing Sr²⁺ and Ho³⁺ ions and mixed for 1 h, and then, 5 ml of the 2 M NH₄F solution (1.5 excess in comparison to the stoichiometric amount) was added dropwise to each mixture and mixed for the following 15 min. The pH of the final mixtures ranged from 6.84 to 6.13 (the pH value decreased with increasing concentration of Ho³⁺ ions). The as-prepared mixtures were transferred into a 50 ml Teflon vessel and hydrothermally treated at 200°C for 12 h (Berghof DAB-2). The post-reaction mixtures were centrifuged (10 min, 9000 rpm), and the precipitates were purified three times (ethanol, mixture of ethanol and deionized water, and deionized water) by dissolution and centrifugation (5 min, 9000 rpm). The obtained products were dissolved in deionized water or dried in a desiccator for 12 h.

The compositions of the synthesized SrF₂:Ho³⁺ samples were established by energy-dispersive x-ray spectroscopy (EDXS) on the high-resolution environment scanning electron microscope (SEM) Quanta 250 FEG, FEI (resolution: 124.2 eV, amplification time: 3.84 μs). The synthesized samples were identified using Fourier transform infrared spectra (FT-IR) spectrophotometer JASCO 4200. The crystalline structures of the obtained products were analyzed with help of x-ray powder diffraction (XRD) measurements on a Bruker AXS D8 Advance Diffractometer equipped with a Johansson monochromator (λ_Cu K_α1 = 1.5406 Å) and a LynxEye strip detector (step: 0.05° 2θ, step time: 1 s, angular range: 20–100°2θ). The reference data were taken from the Inorganic Crystal Structure Database (ICSD). Hydrodynamic sizes of the prepared nanoparticles were determined by the Dynamic Light Scattering (DLS) method and zeta potential measurements using a Malvern Zetasizer Nano ZS instrument. The images of the samples, used for determination of the average sizes and size distributions, were recorded on the high-resolution transmission electron microscope Hitachi HT7700 with an accelerating voltage of 120 kV.

The spectroscopic properties of SrF₂:Ho³⁺ NPs was characterized with the use of a QuantaMaster^TM 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) (emission spectra) or Hamamatsu R928 photomultiplier (excitation spectrum) used as the detectors. The dependencies of the energy transitions intensities on the laser energy were measured using neutral filters which modified the laser energy The luminescence rise and decay lifetimes were recorded with a Mixed Domain Oscilloscope − 200 MHz - Tektronix MDO3022. The luminescence characteristics were measured at room temperature for solid SrF₂:Ho³⁺ samples (powders).

The concentrations of the Ho³⁺ ions in the prepared SrF₂:Ho³⁺ samples were similar to the intended ones (table 1). However, slight differences in the composition were observed for highly doped products (Ho³⁺ ions concentration: ≥15%), which were characterized by a lower concentration of dopant ions than assumed. These differences are probably related to the nature of the synthesis, during which the complexes of Sr²⁺ and Ho³⁺ cations and citrate C₃H₅(COO)₃ ^3- anions, present in the reaction mixture, are formed. The Sr²⁺-Cit^3– complexes (K = 6.6·10^–3 [50]) are less stable than the Ho³⁺-Cit^3– ones (K = 5.9·10^–7 [51]), therefore in the reaction medium the concentrations of free Sr²⁺ and Ho³⁺ ions is not necessary the same as the overall ones, including also chelated ions. As a result, the obtained samples are characterized by a smaller admixture of Ho³⁺ ions than intended [52].

For the subsequent SrF₂:Ho³⁺ samples, the average hydrodynamic diameters decreased (table 1), and ranged from 127.8 nm for the SrF₂:4.9%Ho³⁺ sample to 41.1 nm for the SrF₂:22.5%Ho³⁺ sample. The observed effect of decreasing the hydrodynamic sizes of the obtained particles was caused by the decrease in unit cell parameters, connected with the increasing substitution of the larger Sr²⁺ ions (${{\rm{r}}}_{{{\rm{Sr}}}^{2+}}$ = 1.18 Å) by the smaller Ho³⁺ ions (${{\rm{r}}}_{{{\rm{Ho}}}^{3+}}$ = 0.90 Å) in the particle structures of the subsequent SrF₂:Ho³⁺ samples [53]. Also another effect could be responsible for decreasing size of NPs with growing amount of Ho³⁺ ions. As mentioned before, higher stability of Ho³⁺-Cit^3– complexes than Sr²⁺-Cit^3– ones affected the precipitation reaction kinetics, as there is lower total concentration of not-complexed Sr²⁺ and Ho³⁺ ions, with growing total amount of Ho³⁺ (including free and complexed ions) in the reaction medium. The number of nucleation seeds is therefore higher, resulting in smaller NPs' sizes.

The prepared SrF₂:Ho³⁺ samples exhibited negative surface charge, which suggests the presence of residual citrate anions on their surfaces (see EDXS results in figure S1(available online at stacks.iop.org/MAF/10/024001/mmedia)). The presence of Cit^3– groups on NPs surface was confirmed by the results of the FT-IT measurements (figure S3). The determined zeta potentials took values from –22.9 mV for the sample doped with the lowest content of Ho³⁺ ions to –13.8 mV for the highest doped in the Ho³⁺ ions sample. These results indicate good stability of the low-doped SrF₂:Ho³⁺ samples in water, but a growing tendency of highly-doped SrF₂Ho³⁺ NPs to agglomerate, which is a consequence of their smaller sizes and increased specific surface area [6].

The obtained SrF₂:Ho³⁺ samples showed the crystalline single-phase structures, with the ${\rm{Fm}}\bar{3}{\rm{m}}$ space group (figure 1(a)). In the recorded XRD patterns, a shift of the diffraction peaks towards higher 2θ angles in comparison to the reference pattern (SrF₂ ICSD #40414) was observed, especially noti-ceable for the higher-doped samples. As mentioned before, the effect was due to a decrease in the unit cell parameters, resulted from substitution of the larger Sr²⁺ by the smaller Ho³⁺ ions in the crystals structure. (Table S1). The size of the obtained crystallites were determined from the Scherrer's equation [54] and ranged from 61.2 nm (SrF₂:4.9%Ho³⁺) to 18.4 nm (SrF₂:22.5%Ho³⁺) (table 1).

Zoom In Zoom Out Reset image size

The obtained SrF₂:Ho³⁺ NPs were characterized by spherical shapes and various sizes, as shown in the TEM images (figures 1(b) and (c)). In the subsequent samples with increasing content of Ho³⁺ ions, a significant decrease in the average nanoparticle sizes was observed, and their diameters varied from 82.3 nm (the sample doped with the lowest number of Ho³⁺ ions) to 16.4 nm (the sample highly doped with Ho³⁺ ions) (table 1). The observed tendency is consistent with the results of the DLS and XRD measurements.

The obtained SrF₂:Ho³⁺ NPs were subjected to spectroscopic studies, revealing interesting properties. In the excitation spectrum (figure 2(a)), the NIR band in the range from 1140 to 1210 nm (λ_max. = 1156, 1184 nm), related to the ⁵I₈ → ⁵I₆ Ho³⁺ ions energy transition, was observed [24]. On the basis of the recorded excitation spectrum, the luminescence of the obtained SrF₂:Ho³⁺ NPs was measured under 1156 nm irradiation.

Zoom In Zoom Out Reset image size

The 1156 nm excitation resulted in UC emission at 554, 653, 755, and 900 nm (figure 2(a)) and the bands were assigned to the ⁵S₂,⁵F₄ → ⁵I₈; ⁵F₅ → ⁵I₈; ⁵S₂,⁵F₄ → ⁵I₇ and ⁵I₅ → ⁵I₈ Ho³⁺ ions transitions, respectively [24]. The bands at 653 and 900 nm were characterized by intensities significantly greater than those of the 554 and 755 nm bands. The SrF₂ NPs doped with 10.0% Ho³⁺ ions were characterized by the most intense UC luminescence. Interestingly, it was observed that with increasing content of Ho³⁺ ions in the samples, the ratio of the intensities of the red emission at 653 nm and the NIR one at 900 nm decreased from 1.84 to 0.28 (figure S4). It is worth mentioning that the UC luminescence of the obtained SrF₂:Ho³⁺ NPs appearing in the range of the optical windows can be used in medical applications.

The CIE chromaticity diagram related to the emission color of the SrF₂:Ho³⁺ NPs was prepared on the basis of the corresponding UC emission spectra (figure 2(b)). The luminescence color of the obtained samples was slightly tuned due to the different concentrations of Ho³⁺ ions present in the samples and changed from red for NPs doped with low concentrations of Ho³⁺ ions to reddish-orange for NPs highly doped with Ho³⁺ ions.

The UC luminescence of Ho³⁺ ions may be a result of numerous energy transitions both from the excited states to the ground state as well as from the higher-energy excited states to the lower-energy ones, e.g. NIR emission with a wavelength of about 750 nm may result from the ⁵I₄ → ⁵I₈, ⁵S₂,⁵F₄ → ⁵I₇ and ⁵F₃ → ⁵I₆ energy transitions [33]. Therefore, to fully understand the processes occurring in the obtained SrF₂ NPs containing Ho³⁺ ions, the luminescence lifetime profiles were investigated. The results of these measurements for the SrF₂:10.0%Ho³⁺ sample, as an example, are shown in figure 3(a) (the remaining results are presented in figure S5 in ESI). Moreover, to consider the nature of the observed UC emission, the rise and decay times were estimated with the use of the fitting by the Kohlrausch function (figures 3(b) and (c), Table S3):

$$\begin{eqnarray}&&I(t)={A}_{R}exp\left[-{\left(\displaystyle \frac{t}{{\tau }_{R}}\right)}^{{\beta }_{R}}\right]+{A}_{D}exp\left[-{\left(\displaystyle \frac{t}{{\tau }_{D}}\right)}^{{\beta }_{D}}\right]\end{eqnarray} \tag{ 1 }$$

where A is the oscillator coordinate, β is the degree of non-exponentiality, and τ is the characteristic rise (R) or decay (D) time [55].

Zoom In Zoom Out Reset image size

The prepared SrF₂:Ho³⁺ NPs were characterized by similar rise time values of the UC emission recorded at 554, 755 as well as 900 nm (6.09, 5.46 and 7.50 μs, respectively, for the SrF₂:10.0%Ho³⁺ sample), which suggest, that these bands resulted from the radiative transitions from the excited energy levels similarly populated. The determined values of the rise times at 554 and 755 nm are in a very close range because they are related to the transition from the same ⁵S₂,⁵F₄ excited state. Therefore, it is concluded, that the 554 nm luminescence resulted from the radiative relaxation of the ⁵S₂,⁵F₄ excited energy levels to the ⁵I₈ ground energy level, whereas the 755 nm emission was assigned to the radiative transition from the ⁵S₂,⁵F₄ states, but to the first excited ⁵I₇ state. The above-described nature of the emission bands observed at the mentioned wavelengths is also supported by their low intensities, probably related to the low efficiency of the excitation process of the ⁵S₂,⁵F₄ energy levels. The UC emission at 900 nm is attributed to the ⁵I₅ → ⁵I₈ electronic transition. Analysis of the measured luminescence decays suggests that the excitation process, preceding the observed UC luminescence at 554, 755 and 900 nm, is associated with rapid processes, such as ground and excited state absorption (ESA and GSA).

The rise time values calculated for the ⁵F₅ → ⁵I₈ transition responsible for the emission at 653 nm are relatively long (16.01 μs for SrF₂:10.0%Ho³⁺ sample), therefore the population of the ⁵F₅ emitting level might be related to processes characterized by slower kinetics such as cross relaxations (CR) or energy transfers (ET) [31, 42]. Furthermore, for all of the SrF₂:Ho³⁺ NPs, the values of rise times decreased with growing concentration of Ho³⁺ ions in the prepared samples, while the values of decay times initially increased, achieved a maximum for the sample characterized by the most intense luminescence, then rapidly decreased. This dependence is usually found for the materials doped with Ho³⁺ ions [28] and probably results from the increased contribution of the non-radiative processes related to the concentration quenching [56].

To determine the UC mechanism, the estimation of the number of photons involved in the population of the excited states in UC process is also relevant. The values of the involved photons n can be estimated on the basis of the dependence of the UC emission intensity I_UC on the excitation power density P, or in the case of excitation with a pulsed laser, from its energy E [57]:

$$\begin{eqnarray}&&{I}_{UC}\propto {P}^{n}\,\propto {E}^{n}\end{eqnarray} \tag{ 2 }$$

further described as n coefficients or slopes (obtained from a linear fit to the plot presented on a double logarithmic scale). Therefore, the dependencies of the integral UC emission intensities on the 1156 nm laser energy were measured and analyzed. The results obtained for the SrF₂:10.0%Ho³⁺ sample are presented in figure 4(a) (the remaining results are shown in figure S6 and table S4 in ESI). The dependencies of the number of photons involved in the UC mechanism in the prepared SrF₂ NPs doped with various amounts of Ho³⁺ are compared in figure 4(b).

Zoom In Zoom Out Reset image size

The estimated n coefficients of the recorded 653 nm bands of all SrF₂:Ho³⁺ NPs were higher than 2, which implied that 3 photons were needed to excite the ⁵F₅ energy level. The actual slope values are usually lower than theoretical because of competitive to UC processes, such as cross-relaxation or multiphonon quenching. Meanwhile, the slopes of the dependences shown in figure 4 and referring to the 755 nm bands in the emission spectra (figure 2) of the samples doped with Ho³⁺ ions in 7.1, 10.0, and 14.1% are lower than 2, which suggests that 2 photons were required to excite the ⁵S₂ energy level (the intensity of this band for the remaining samples was too low to estimate the n coefficients, as were the intensities of the 554 nm bands of all obtained samples). The low doped in Ho³⁺ ions SrF₂ samples were characterized by the slopes of the analogous dependences for 900 nm bands lower than 2, while the slopes of the dependences for the bands for the highly doped SrF₂:Ho³⁺ samples were higher than 2. Thus, the number of photons, which were needed to excite the ⁵I₅ state increased from 2 to 3 with increasing content of Ho³⁺ ions, which suggests possible alternations to the mechanism of excitation of the ⁵I₅ state resulting from the processes of energy transfers (ET) induced by the Ho³⁺–Ho³⁺ interactions [43].

The spectroscopic analysis of the prepared SrF₂:Ho³⁺ NPs, allowed drawing conclusions about the mechanism responsible for the observed UC emission (see figure 5). When the samples are irradiated with the 1156 nm pulsed laser, absorption of the first NIR photon occurs due to the GSA process, which results in the excitation of Ho³⁺ ions from their ⁵I₈ ground state to the ⁵I₆ excited state. After absorption of the first photon, two competitive energy processes can take place.

Zoom In Zoom Out Reset image size

In the first one, the second NIR photon can be absorbed by the Ho³⁺ ions in their ⁵I₆ state, populating the higher-energy ⁵S₂ and ⁵F₄ states in the ESA process. Next, the absorbed energy can be released from the ⁵S₂ and ⁵F₄ energy levels to the ⁵I₈ and ⁵I₇ energy levels, generating luminescence at 554 and 755 nm, respectively.

In the second process, the Ho³⁺ ions, initially excited to the ⁵I₆ state, undergo relaxation in the non-radiative process which results in populating the ⁵I₇ excited state. Then, the absorption of the second NIR photon also by ESA process can occur leading to the excitation of the Ho³⁺ ions to their ⁵I₄ excited state. The ⁵I₄ state can undergo non-radiative relaxation process populating the ⁵I₅ state, followed with another, but this time radiative relaxation process, resulting in emission at 900 nm.

The described mechanism responsible for the UC emission at 554, 755 and 900 nm is confirmed by the measured short rise times and the calculated numbers of photons involved in the UC mechanism, close to 2 [46, 47].

Moreover, the Ho³⁺ ions in their ⁵I₇ excited state can act as an energy reservoir, from which the energy can be transferred to the other Ho³⁺ ions in the ⁵I₅ excited state. This energy exchange results in the excitation of Ho³⁺ ions to their ⁵F₅ energy state, from which depopulation to the ground state connected with red emission at 653 nm is possible. The described mechanism responsible for the red luminescence is supported by the long luminescence rise times with a rapid decrease in their values with growing Ho³⁺ concentration. Furthermore, the Ho³⁺ ions in the ⁵I₅ state can be excited by the third NIR photon in the ESA process yielding the Ho³⁺ ions in the ⁵S₂ and ⁵F₄ states, from which a relaxation to state ⁵F₅ occur, contributing to the observed at 653 nm luminescence.

The presented mechanism corresponds to the nature of the UC phenomena occurring in low-doped SrF₂:Ho³⁺ samples. However, with the increase in the concentration of Ho³⁺ admixture in subsequent samples, the n coefficient increased, which was connected with the increase in the number of photons involved in the observed UC luminescence. The similar trend of the dependence of the n coefficient on the concentration of doping Ln³⁺ ions in UC materials is explained by the fact that as the concentration of Ln³⁺ ions increases, the distance between them decreases, and this enhances the efficiency of energy transfer, which has a contribution to increasing the overall conversion efficiency [16, 58, 59]. On the other hand, the increase in the content of the Ho³⁺ ions in the prepared samples caused the decrease in the luminescence rise and decay time, which is connected with the concentration quenching effect, reducing the UC efficiency [56]. The observed increase of n coefficient may be also a consequence of higher quenching/cross-relaxation rates resulting in the need for more rapid absorption/energy transfer of another photon [60].