Upconversion luminescence and favorable temperature sensing performance of eulytite-type Sr3Y(PO4)3:Yb3+/Ln3+ phosphors (Ln=Ho, Er, Tm).

Phase-pure eulytite-type Sr3Y0.88(PO4)3:0.10Yb3+,0.02Ln3+ upconversion (UC) phosphors (Ln = Ho, Er, Tm) were synthesized via gel-combustion and subsequent calcination at 1250°C. Their UC luminescence, temperature-dependent fluorescence intensity ratio of thermally and/or non-thermally coupled energy levels, and performance of optical temperature sensing were systematically investigated. The phosphors typically exhibit green, orange-red and blue luminescence under 978 nm NIR laser excitation for Ln = Er, Ho and Tm, respectively, which were discussed from two- and three-photon processes. The 524 nm green (Er3+), 657 nm red (Ho3+) and 476 nm blue (Tm3+) main emissions were analyzed to have average decay times of ~52 ± 2, 260.6 ± 0.7 and 117 ± 1 μs, respectively. It was shown that (1) the Er3+ doped phosphor has a better overall performance of temperature sensing with thermally coupled 2H11/2 and 4S3/2 energy levels, whose maximum absolute (S A) and relative (SR ) sensitivities are ~5.07 × 10-3 K-1 at 523 K and ~1.16% at 298 K, respectively; (2) the Ho3+ doped phosphor shows maximum S A and SR values of ~0.019 K-1 (298-573 K) and 0.42% at 573 K for the non-thermally coupled energy pairs of 5F5/(5F4,5S2) and 5I4/5F5, respectively; (3) the Tm3+ doped phosphor has a maximum S A of ~12.74 × 10-3 K-1 at 573 K for the non-thermally coupled 3F2,3/1G4 energy levels and a maximum SR of ~1.74% K-1 at 298 K for the thermally coupled 3F2,3/3H4 levels. Advantages of the current phosphors in optical temperature sensing were also revealed by comparing with other typical UC phosphors.

Introduction

Upconversion (UC) luminescence is a process of converting low-energy light, usually near-infrared (NIR) or infrared, into high-energy light (ultraviolet or visible) through multiple absorption and/or energy transfer [1-3]. UC materials are drawing extensive attention due to their wide applications in the fields of solid-state lasers, multi-color displays, optical communication, wavelength converters for solar cells, bio-imaging, optical temperature sensors, and so forth [1-4]. A UC phosphor is usually formed by doping a host lattice with a sensitizer/activator pair, and the Yb3+/Ln3+ combination (Ln = Ho, Er, Tm) is the most popular since the 2F7/2-2F5/2 transition of Yb3+ possesses a large absorption cross-section for ~980 nm NIR excitation light and can well resonate with the ladder-like energy levels of Ho3+, Er3+ and Tm3+ activators [1-3]. The host lattice for UC should assure not only a satisfactory luminescence efficiency, but also excellent physicochemical stability, safety, and low toxicity. A handful of inorganic compounds have been developed as UC host so far, typically including fluorides [5,6], oxides [7,8], oxysulfides [9], phosphates [10,11] and other oxygenates [12-15], and new hosts are also under active exploration and/or perfection.

Optical temperature sensing with UC phosphor, most frequently investigated in the range of ~293–573 K, gained increasing research interest during recent years, which utilizes the fluorescence intensity ratio (FIR) of two emission bands that involve either thermally coupled or non-thermally coupled energy levels of the luminescent ion [4,16]. The FIR technique is usually independent of spectrum loss and excitation-power fluctuation, and may thus provide a high detection resolution and excellent sensitivity [17-19]. The FIR of thermally coupled emission levels is a function of temperature and obeys the Boltzmann distribution of electrons [11]. Since the energy separation ΔE of such levels is usually restricted to 200–2000 cm−1 to avoid strong overlapping of two emission bands [16-19], the sensitivity of temperature sensing, which is proportional to ΔE, can hardly be further improved to a higher level according to the Boltzmann distribution. For this reason, the use of non-thermally coupled energy levels is being considered as an effective complement to a better sensitivity of FIR. Wang et al. [4] have recently reviewed the rare-earth ions, host lattices, electronic transitions and emission wavelengths that have been used for the purpose of optical temperature sensing, together with the temperature ranges of sensing and the absolute/relative sensitivities of FIR.

Eulytite-type orthophosphate M3A(PO4)3 (M = Ca, Sr or Ba; A = La, Gd, Y or Lu) possesses high physical, chemical and structure stabilities [20,21], and may thus serve as an important family of phosphor hosts. It should be noted that many other types of inorganic compounds, such as GdPO4 orthophosphate [22] and NaLn(WO4)2 tungstate (Ln = La-Lu, and Y) [23], also draw great interest for phosphor applications. For downconversion (DC) luminescence, You et al. [24] prepared Eu2+/Mn2+ co-doped Sr3Lu(PO4)3 by solid-state reaction at 1300°C for 3 h in a CO atmosphere and investigated its luminescence and Eu2+-Mn2+ energy transfer; Liang et al. [25] produced Ba3La(PO4)3:Ln3+ (Ln = Tb, Eu) phosphors via solid reaction at 1200–1250°C for 5–8 h in a thermal carbon atmosphere for color-tunable luminescence; Xia et al. [26] synthesized a series of (Ba,Sr)3Lu(PO4)3:Eu2+ phosphors by solid reaction at 1300°C for 4 h under a 10%H2-90%N2 gas mixture, and the blue shift of Eu2+ emission with increasing Sr/Ba ratio was discussed in detail; Xia et al. [27] also synthesized eulytite-type Ba3Eu(PO4)3 and Sr3Eu(PO4)3 compounds via solid reaction at 1250°C for 10 h in air, and systematically compared their crystal structures and photoluminescence; Guo et al. [28] prepared Ba3Y(PO4)3:Eu2+/Mn2+ phosphors by solid reaction at 1300°C for 4 h under a 10%H2-90%N2 gas mixture, and thoroughly investigated their phase formation, luminescence properties and Eu2+→Mn2+ energy transfer. We have recently synthesized by gel-combustion a series of Ba3La(PO4)3:Ce3+/Mn2+ and Sr3Y(PO4)3:Eu phosphors [20,21] and evaluated their luminescence. Our thorough literature survey, however, found that the UC luminescence of Yb3+/Ln3+ pair in eulytite-type M3A(PO4)3 has rarely been investigated up to date, though examples exist for other phosphate hosts such as K3Y(PO4)2 [11] and BiPO4 [29]. It is noteworthy that Zhang et al. [14] synthesized in 2018 eulytite-type Ba3La(PO4)3:Yb3+/Ln3+ phosphors (Ln = Er, Tm) via solid reaction at 1360°C for 5 h in air and studied their properties of temperature sensing with the thermally coupled energy levels of 2H11/2/4S3/2 (Er3+) and 3F2,3/3H4 (Tm3+). It was demonstrated that the absolute and relative sensitivities of FIR successively increase (maximum ~4.38 × 10−3 K−1 at 498 K for Er3+ and ~1.31 × 10−4 K−1 at 503 K for Tm3+) and decrease with increasing temperature for both Er3+ and Tm3+, respectively [14]. While not mentioned for Tm3+, a two-photon process was discussed for the UC luminescence of Er3+ [14]. The performance of temperature sensing with non-thermally coupled energy levels, however, was not investigated therein for the Ba3La(PO4)3:Yb3+/Ln3+ phosphors. Compared with Ba3La(PO4)3, the isomorphic Sr3Y(PO4)3 compound can be a better host for UC luminescence, since Y3+ is closer to Yb3+ and Ln3+ (Ln = Ho, Er, Tm) than La3+ in ionic radius, which may minimize the lattice distortion upon Yb3+/Ln3+ doping, and Y is among the most abundant rare-earth elements.

We thus originally synthesized in this work a series of eulytite-type Sr3Y(PO4)3:Yb3+/Ln3+ UC phosphors (Ln = Ho, Er, Tm) via gel-combustion, followed by a thorough investigation of their UC luminescence and performance of optical temperature sensing with thermally coupled and/or non-thermally coupled energy levels. The high cation homogeneity of sol-gel processing allowed phase-pure products to form by calcination at the lower temperature of 1250°C for 4 h. The current phosphors were also compared with other typical UC systems to show their advantages in optical temperature sensing. In the following sections, we report the synthesis, characterization and optical properties of Sr3Y(PO4)3:Yb3+/Ln3+ UC phosphors.

Experimental details

Materials and synthesis

The starting reagents are 99.99% pure RE2O3 (RE = Y, Ho, Er or Tm; Huizhou Ruier Rare-Chem. Hi-Tech. Co. Ltd., Huizhou, China) and analytical grade ethylenediamine tetraacetic acid (C10H16N2O8, EDTA), ammonium hydroxide solution (25 wt%), nitric acid (65 wt%), NH4H2PO4 and Sr(NO3)2 (Shenyang Chemical Reagent Factory, Shenyang, China). The nitrate solution of RE3+ was prepared by dissolving the corresponding oxide with a proper amount of nitric acid, followed by evaporation at 95°C to remove the superfluous HNO3 and dilution with distilled water.

For gel-combustion synthesis of Sr3Y0.88(PO4)3:0.10Yb3+,0.02Ln3+, stoichiometric amounts of RE(NO3)3 and Sr(NO3)2 were dissolved in an aqueous solution (20 ml) of EDTA-NH4OH to chelate the RE3+ and Sr2+ cations (total cation to EDTA molar ratio = 1:1), followed by the addition of a stoichiometric amount of NH4H2PO4. The mixture was evaporated by heating at 85°C under continuous magnetic stirring to form a sol and then a viscous white gel. Auto-ignition of the gel took place upon increasing the temperature to ~300°C on a resistance oven, which produced a loosely packed black precursor powder. The targeted phosphor was then produced by calcining the precursor in flowing oxygen (200 ml/min) at 1250°C for 4 h [20,21], using a heating rate of 8°C/min at the ramp stage.

Characterization

Phase identification was performed via X-ray diffractometry (XRD, SmartLab, Rigaku, Tokyo, Japan) under 40 kV/200 mA, using nickel-filtered Cu-Kα radiation (λ = 0.15406 nm) and a scanning speed of 4.0º 2θ per minute. Crystal structure refinement of the product was carried out using the TOPAS 3.0 program [13], and the XRD data for this purpose were acquired in the step scan mode using a step size of 0.02° and an accumulation time of 1.8 s per step. Powder morphology was analyzed via field-emission scanning electron microscopy (FE-SEM, Model S-5000, Hitachi, Tokyo) under an acceleration voltage of 10 kV. UV-Vis spectroscopy was performed at room temperature on a UV-VIS-NIR spectrometer (Model UV-3600 Plus, Shimadzu Co., Kyoto, Japan) equipped with a 150-mm diameter integrating sphere (Model ISR-1503, Shimadzu Co.). UC luminescence of the phosphor was analyzed under 978 nm CW-laser excitation (Model KS3-12322-105, BWT Beijing Ltd., Beijing, China) on an FP-8600 fluorospectrophotometer (JASCO, Tokyo) installed with a heating controller (Model HPC-836, JASCO). Fluorescence decay kinetics was analyzed under 980 nm pulsed laser excitation on a steady-state and transient photoluminescence spectrometer (Model FLS1000, Edinburgh Instruments Ltd., Livingston, UK).

Results and discussion

Phase analysis and morphology

Figure 1(a) presents the powder XRD profiles for the three Sr3Y0.88(PO4)3:0.10Yb3+,0.02Ln3+ phosphors (Ln = Ho, Er or Tm), where it is evident that the patterns are similar to each other and match well with that of cubic structured eulytite-type Sr3Y(PO4)3 in the standard diffraction file (JCPDS No. 00-44-0320; space group: I-43d). The sharp reflections also indicate that the products were well crystallized. The synthesis temperature is about100°C lower than that (1360°C) needed for the synthesis of Ba3La(PO4)3:Yb3+/Ln3+ (Ln = Er, Tm) via solid reaction [14], which may originate from the better cation homogeneity of sol-gel processing. The crystal structure of Sr3Y(PO4)3 can be viewed as a three-dimensional connection of [PO4]3- tetrahedrons and [(Sr/Y)-O] polyhedrons via corner sharing, where all the [PO4]3- are totally independent while the Sr/Y polyhedrons share edges with each other to form a three-dimensional network [26]. In such a structure, the Sr2+/Y3+ cations are randomly disordered over a single 16c crystallographic site (C3 point symmetry) [30] but the [PO4]3- tetrahedrons show three different orientations in response to three sets of partially occupied oxygen positions O1, O2 and O3 [31]. It is noteworthy that Sr2+ and Y3+ have different coordination environments although they occupy the same lattice site. Specifically, the Y3+ ion resides in YO6 octahedron distorted by three equally short and three equally long Y-O bonds [32] while the Sr2+ ions have the two coordination environments of CN = 6 and CN = 9 (CN: coordination number) [20]. Accordingly, Sr3Y(PO4)3 presents not only cation disorder but also disorder in the oxygen sublattice [20,32]. In this work, the Yb3+ and Ln3+ dopants were expected to replace Y3+ by valence and size preference (ionic radius r= 90.0, 90.1, 89.0, 88.0 and 86.8 pm for Y3+, Ho3+, Er3+, Tm3+ and Yb3+ under CN = 6; r= 118 and 131 pm for Sr2+ under CN = 6 and 9, respectively) [33]. Based on this information, Rietveld refinement of the XRD pattern was performed using the standard crystallographic data of isostructural Sr3La(PO4)3 (ICSD No. 69432) as initial structure model. Figure 1(b) shows the experimental and calculated XRD profiles for the Sr3Y0.88(PO4)3:0.10Yb3+,0.02Er3+ representative, while the derived coordinates and site occupancy factors (SOF) of atoms are summarized in Table S1. The refinement was ended up with the well-acceptable reliability factors of R= 8.28%, R= 6.12%, R= 3.83% and χ= 2.16, and yielded a lattice parameter (a= b= c) of ~10.1043 ± 0.0002 Å and cell volume V of ~1031.62 ± 0.05 Å3. Similar analysis found the a and V values of ~10.1045 ± 0.0004 Å and 1031.68 ± 0.12Å3 for the Yb3+/Ho3+ doped and ~10.1024 ± 0.0006 Å and 1031.04 ± 0.18 Å3 for the Yb3+/Tm3+ doped phosphor powders. The cell constants are all smaller than that (10.1091 Å) of Sr3Y(PO4)3 in the standard diffraction file, owing to the smaller average ionic radius of Yb3+/Ln3+ pair, and tend to decrease toward a smaller Ln3+. The above results thus provided persuasive evidence of solid-solution formation.

Figure 1.

(a) XRD patterns of the as-synthesized Sr3Y0.88(PO4)3:0.10Yb3+,0.02Ln3+ powders, with the standard diffractions of Sr3Y(PO4)3 included as bars for comparison, and (b) Rietveld refinement of the XRD pattern for the Yb3+/Er3+ doped sample, where the observed, calculated and difference XRD patterns and the positions of Bragg reflections are shown by the red solid line, black crosses, green solid line and blue vertical bars, respectively.

FE-SEM observations (Figure 2(a–c)) reveal that the Sr3Y0.88(PO4)3:0.10Yb3+,0.02Ln3+ products contain aggregated primary particles/crystallites of ~2.0–6.0 μm, which is typical of a gel-combustion product [20,21], and the type of Ln has no appreciable influence on overall morphology of the powder. Elemental mapping via energy dispersive X-ray spectroscopy (EDS), with the Sr3Y0.88(PO4)3:0.10Yb3+,0.02Er3+ sample as a representative, found that all the elements of concern are quite evenly distributed among the particles (Figure 2(d–i)). EDS analysis (Figure S1) also indicated that the sample contains ~15.73 at% of Sr, 4.60 at% of Y, 0.54 at% of Yb, 0.11 at% of Er, 15.81 at% of P and 63.21 at% of O, and the derived Sr:Y:Yb:Er:P:O atomic ratio of 3:0.877:0.103:0.021:3.015:12.055 is very close to the theoretical value of 3:0.88:0.10:0.02:3:12. The above EDS and XRD analyses confirmed that a solid-solution product with the intended chemical composition has been formed.

Figure 2.

FE-SEM morphologies (upper row) of the Sr3Y0.88(PO4)3:0.10Yb3+,0.02Ln3+ powders, where Ln = Ho (a), Er (b) and Tm (c). The bottom row presents elemental mapping of the red-framed area in (b), with parts (d), (e), (f), (g), (h) and (i) corresponding to elements Sr, Y, P, O, Yb and Er, respectively.

Figure 3 shows the UV-Vis diffuse reflectance spectra of the Sr3Y(PO4)3:Yb3+/Ln3+ powders, where the broad band in the spectral range of ~200-360 nm and that centered at ~978 nm, which are common to the three samples, can be assigned to absorption by the Sr3Y(PO4)3 host and 2F7/2→2F5/2 transition of Yb3+, respectively. In addition, the Ho3 + activator clearly shows transitions its 5F→5I8 (i= 2, 3, 4 and 5) transitions at ~453, 486, 541 and 644/657 nm (Figure 3(a)), Er3+ exhibits transitions from the 4F7/2, 2H11/2, 4S3/2 and 4F9/2 energy states to 4I15/2 ground state at ~489, 523, 546, 654 nm (Figure 3(b)), and Tm3+ shows transitions from the 1G4, 3F2,3 and 3H4 levels to 3H6 ground state at ~474, 690 and 796 nm (Figure 3(c)), respectively. The results thus imply that the Sr3Y(PO4)3:Yb3+/Ln3+ powders can effectively absorb 978 nm laser excitation for UC luminescence. The energy bandgap of Sr3Y(PO4)3:Yb3+/Ln3+ can be estimated from the reflectance spectra according to Equation (1) [34,35]

Figure 3.

UV-Vis diffuse reflectance spectra of the Sr3Y0.88(PO4)3:0.10Yb3+,0.02Ln3+ powders.

where hv is the incident photon energy, A is a proportional constant, Eg is the value of bandgap, n= 2 for a direct transition or 1/2 for an indirect transition, and F(R∞) is the Kubelka-Munk function which is defined as [36,37]

where R, K and S are the reflection, absorption and scattering coefficients, respectively. The [F(R∞)hv]1/2 vs hv plots are shown in Figure S2, where extrapolating the linear portions to [F(R∞)hv]1/2 = 0 yielded the similar Eg values of ~3.37 eV. The Eg value is also close to those reported for the isostructural Ba3La(PO4)3 (3.46 eV) [38], Ba3Y(PO4)3 (3.15 eV) [39] and Sr3Gd(PO4)3 (3.49 eV) [40] compounds synthesized by solid reaction.

Upconversion luminescence of the Sr3Y(PO4)3:Yb3+/Ln3+ phosphors

Yb3+/Er3+ is the most widely investigated sensitizer/activator pair for UC luminescence in various types of host lattices [9-19], since the 2F5/2→2F7/2 emission of Yb3+ and the 4I15/2→4I11/2 excitation transition of Er3+ have well-matching energies. Figure 4(a) shows the UC luminescence spectra of Sr3Y0.88(PO4)3:0.10Yb3+,0.02Er3+ under varying pumping power of 978 nm laser. It is seen that, in each case, the spectrum includes a blue (~486 nm, negligibly weak), green (~524/547 nm, dominantly strong) and red (~655 nm, strong) band in the visible-light region, which are assignable to transitions from the 4F7/2, 2H11/2/4S3/2 and 4F9/2 excited states to the4I15/2 ground state of Er3+ [41,42], respectively. Increasing power of excitation did not bring about any change to peak position but monotonically raised the emission intensity of each band. The Commission International de L’Eclairage (CIE) chromaticity coordinates of UC luminescence are summarized in Figure 4(d) and Table S2, where it is clear that the emission color steadily drifted from yellowish green [color coordinates: (0.3288, 0.5102)] to green [color coordinates: (0.2671, 0.6276)] with increasing excitation power from 1.00 to 3.00 W. The color change agrees with the almost linearly increasing intensity ratio of green to red emission (I524/I655 and I547/I655, Figure S3(a)). Under 2.00 W laser pumping, vivid and strong green emission was observed for Sr3Y0.88(PO4)3:0.10Yb3+,0.02Er3+ with naked eyes, as shown by the inset photograph taken for the appearance of luminescence in Figure 4(d).

Figure 4.

UC luminescence spectra under different excitation power levels (a), the relationship between log(Iem) and logP (in Watt, b), and a scheme showing the energy levels and UC process (c) for the Sr3Y0.88(PO4)3:0.10Yb3+,0.02Er3+ phosphor. Part (d) shows the CIE chromaticity coordinates of Sr3Y0.88(PO4)3:0.10Yb3+,0.02Ln3+ under varying excitation power, where the inset photographs are for the appearances of UC luminescence under 2.00 W of 978 nm laser excitation.

The number of excitation photons required to populate the upper emitting state under unsaturated condition can be obtained from the relation Iem∝P [9,13], where Iem is the emission intensity, P is the pumping power, and n is the number of low-energy photons required to convert to one high-energy photon in the UC process. Figure 4(b) shows the log(Iem)-log(P) plot of the above relation, from which the n values were determined from the slope of the linear fitting to be ~2.58, 2.78, 2.50 and 1.61 for the UC peaks at ~486, 524, 547 and 655 nm, respectively. The results thus suggest that a three-photon process is primarily responsible for the observed 4F7/2→4I15/2 (486 nm), 2H11/2→4I15/2 (524 nm) and 4S3/2→4I15/2 (547 nm) emissions and a two-photon process for the 4F9/2→4I15/2 (655 nm) one. The process of UC is known to be virtually affected a number of factors, such as host composition, lattice defects, crystallinity, the content, distribution uniformity and actual lattice site of the sensitizer/activator pair, and so forth [1-4,12]. The results of this work are consistent with those obtained from YbPO4:Er [10] and NaLu(WO4)2:Yb/Er [12] UC phosphors, though a three-photon process was reported for all the 2H11/2/4S3/2→4I15/2 and 4F9/2→4I15/2 emissions of La2O2SO4:Yb/Er [13] and a two-photon mechanism for each emission of La2O2S:Yb/Er [9], KMgF3:Yb/Er [43], Ba5Gd8Zn4O21:Yb/Er [44], α-NaYF4:Yb/Er [45] and Na2Y2B2O7:Yb/Er [46]. The three basic mechanisms of excited state absorption (ESA), energy transfer (ET) and photon avalanche (PA) have been proposed for UC luminescence [1-3]. Since no power threshold was observed in the range of this study, the avalanche mechanism can be neglected. With the energy diagram constructed in Figure 4(c) by referring to previous studies, the UC luminescence of Sr3Y0.88(PO4)3:0.10Yb3+,0.02Er3+ was proposed to occur via the following processes:

(1) The Yb3+ electrons are excited from the 2F7/2 ground state to the 2F5/2 level by absorbing the energy of one laser photon [ESA, 2F7/2 (Yb3+) + hν (978 nm) → 2F5/2 (Yb3+)], which transfer energy to Er3+ while returning to the 2F7/2 ground state and thus promotes Er3+ electrons from the 4I15/2 ground state to the 4I11/2 level [ET1, 2F5/2 (Yb3+) + 4I15/2 (Er3+) → 2F7/2 (Yb3+) + 4I11/2 (Er3+)]; (2) After nonradiative relaxation to 4I13/2 [NR, 4I11/2 (Er3+) ~ 4I13/2 (Er3+)], Er3+ electrons are raised to the 4F9/2 level via energy transfer of a second laser photon [ET2, 2F5/2 (Yb3+) + 4I13/2 (Er3+) → 2F7/2 (Yb3+) + 4F9/2 (Er3+)]; (3) A part of the 4F9/2 electrons radiatively relax to the 4I15/2 ground state, which produces the ~655 nm red emission (4F9/2→4I15/2), and the other part relax to the 4I11/2 state via NR [4I9/2 (Er3+) ~ 4I11/2 (Er3+)], followed by further excitation to the 4F7/2 level with a third laser photon [ET3, 2F5/2 (Yb3+) + 4I11/2 (Er3+) → 2F7/2 (Yb3+) + 4F7/2 (Er3+)]. The 4F7/2 electrons may directly jump back to the 4I15/2 ground state to produce the ~486 nm blue emission (4F7/2→4I15/2) and may relax to the 2H11/2/4S3/2 levels via NR, from which the ~524/547 nm green emissions can be resulted upon back-jumping of the electrons to the 4I15/2 ground level (2H11/2/4S3/2→4I15/2). The very weak 4F7/2→4I15/2 blue emission suggests that most of the excitation energy accumulated by ET3 relaxes to the 2H11/2/4S3/2 states to result in the strong green emission (Figure 4(a)). The gradually larger green to red intensity ratio (Figure S3(a)) may imply that the 2H11/2/4S3/2 levels gain population faster than 4F9/2 under a higher excitation power.

Figure 5(a) shows the UC luminescence spectra of Yb3+/Ho3+ codoped Sr3Y(PO4)3 under varying excitation power, where the four groups of emission bands centered at ~485 nm (blue, negligible), 545 nm (green, weak), 657 nm (red, overwhelmingly strong) and 767 nm (red in NIR, weak) can be assigned to 5F3→5I8, 5F4/5S2→5I8, 5F5→5I8 and 5I4→5I8 transitions of Ho3+ [9-12], respectively. Raising excitation power from 1.00 to 3.00 W did not produce any new emission but successively improved the intensity of the existing luminescence. The CIE chromaticity coordinates of UC luminescence gradually drifted from orange [(0.5503,0.4318)] to orange-red [(0.6235,0.3676)] with increasing excitation power (Figure 4(d) and Table S2), which is due to the gradually larger red to green intensity ratio (I657/I545, Figure S3(b)). Under 2.00 W laser pumping, the phosphor exhibits a vivid and strong orange-red emission visible to naked eyes, as shown by the inset in Figure 4(d).

Figure 5.

UC luminescence spectra under different excitation power levels (a), the relationship between log(Iem) and logP (in Watt, b), and a scheme showing the energy levels and UC process (c) for the Sr3Y0.88(PO4)3:0.10Yb3+,0.02Ho3+ phosphor.

Analysis of the log(Iem)-log(P) plots found n values of ~2.55, 2.88, 1.93 and 1.51 for the ~485, 545, 657 and 767 nm UC bands (Figure 5(b)), respectively, which indicate that a three-photon process is largely responsible for the 5F3→5I8 (485 nm) and 5F4/5S2→5I8 (545 nm) transitions while a two-photon process for the 5F5→5I8 (657 nm) and 5I4→5I8 (767 nm) transitions of Ho3+. The UC luminescence of Y2O3:Yb3+/Ho3+/Zn2+(YOZ) [8] and BaZrO3:Yb3+/Ho3+(BZ) [47] were also reported to involve three- (5F3→5I8 of YOZ; 5F4/5S2→5I8 of BZ) and two- (5F4/5S2→5I8 and 5F5→5I8 of YOZ; 5F5→5I8 of BZ) photon processes, although a three-photon process was suggested for all the emissions of La2O2S:Yb/Ho [9], YbPO4:Ho [10], NaLu(WO4)2:Yb/Ho [12] and La2O2SO4:Yb/Ho [13] and a two-photon process for those of Sr5(PO4)3Cl:Yb/Ho [48]. With the energy level diagram constructed in Figure 5(c), the UC process of Sr3Y0.88(PO4)3:0.10Yb3+,0.02Ho3+ can be described with the following photon reactions:

(1) ESA: 2F7/2 (Yb3+) + hν (978 nm) → 2F5/2 (Yb3+);

(2) ET1: 2F5/2 (Yb3+) + 5I8 (Ho3+) → 2F7/2 (Yb3+) + 5I6 (Ho3+);

(3) NR: 5I6 (Ho3+) ~ 5I7 (Ho3+);

(4) ET2: 2F5/2 (Yb3+) + 5I7 (Ho3+) → 2F7/2 (Yb3+) + 5F5 (Ho3+);

(5) Emission: 5F5 (Ho3+) → 5I8 (Ho3+) + hν (657 nm);

(6) NR: 5F5 (Ho3+) ~ 5I4 (Ho3+);

(7) Emission: 5I4 (Ho3+) → 5I8 (Ho3+) + hν (767 nm);

(8) NR: 5F5 (Ho3+) ~ 5I5 (Ho3+);

(9) ET3: 2F5/2 (Yb3+) + 5I5 (Ho3+) → 2F7/2 (Yb3+) + 5F3 (Ho3+);

(10) Emission: 5F3 (Ho3+) → 5I8 (Ho3+) + hν (485 nm);

(11) NR: 5F3 (Ho3+) ~ 5F4/5S2 (Ho3+);

(12) Emission: 5F4/5S2 (Ho3+) → 5I8 (Ho3+) + hν (545 nm).

The overwhelmingly strong red emission (~657 nm) may have two origins: (1) most of the 5F5 electrons excited by ET2 directly transit back to the 5I8 ground state (the n= 2 channels), and (2) the 5F3 electrons excited by ET3 decay to the 5F5 level, followed by radiative transition to the 5I8 state (the n= 3 channel). The linearly increasing I657/I485 and I657/I545 intensity ratios (Figure S3(b)) may imply that a higher excitation power leads to a faster population of the 5F5 energy state.

Under 978 nm laser excitation, Sr3Y0.88(PO4)3:0.10Yb3+, 0.02Tm3+ phosphor exhibits four groups of emissions at ~476 nm (blue), ~650 nm (red), 695 nm (red, negligible) and ~795 nm (NIR) as shown in Figure 6(a), which correspond to the 1G4→3H6, 1G4→3F4, 3F2,3 →3H6 and 3H4→3H6 transitions of Tm3+, respectively. Increasing excitation power led to faster enhancement of blue emission, which became dominant when P reached ~1.50 W. The strong blue emission is evident from the appearance of UC luminescence under 2.00 W of laser pumping (Figure 4(d), the inset). It is also seen from Figure 4(d) and Table S2 that the emission color gradually drifted from light blue [(0.1764, 0.1781)] to deep blue [(0.1323, 0.1194)] with increasing excitation power from 1.00 to 3.00 W, which conforms to the gradually larger blue to red intensity ratio (I476/I650, Figure S3(c)). Analyzing the emission intensity against excitation power yielded slope (n) values of ~2.99 and 2.71 (around 3) for the ~476 nm (1G4→3H6) and 650 nm (1G4→3F4) UC bands and ~2.25 and 1.89 (around 2) for the ~695 nm (3F2,3 →3H6) and 795 nm (3H4→3H6) ones (Figure 6(b)), which correspond to three- and two-photon UC mechanisms, respectively. Similar results were reported in the literature for β-NaLuF4:Yb/Tm [10], Ba5Gd8Zn4O21:Yb/Tm [49] and LiLa(MoO4)2:Yb/Tm [50], though a three-photon process was proposed for the UC emissions of NaLu(WO4)2:Yb/Tm [12], La2O2SO4:Yb/Tm [13] and La2O2S:Yb/Tm [51]. By referring to the energy level diagram shown in Figure 6(c), the processes that led to the observed UC luminescence may be presented as follows:

Figure 6.

UC luminescence spectra under different excitation power levels (a), the relationship between log(Iem) and logP (in Watt, b), and a scheme showing the energy levels and UC process (c) for the Sr3Y0.88(PO4)3:0.10Yb3+,0.02Tm3+ phosphor.

(1) ESA: [2F7/2 (Yb3+) + hν (978 nm) → 2F5/2 (Yb3+)];

(2) ET1: [2F5/2 (Yb3+) + 3H6 (Tm3+) → 2F7/2 (Yb3+) + 3H5 (Tm3+)];

(3) NR: [3H5 (Tm3+) ~ 3F4 (Tm3+)];

(4) ET2: [2F5/2 (Yb3+) + 3F4 (Tm3+) → 2F7/2 (Yb3+) + 3F2,3 (Tm3+)];

(5) Emission: 3F2,3 (Tm3+) → 3H6 (Tm3+) + hν (695 nm);

(6) NR: 3F2,3 (Tm3+) ~ 3H4 (Tm3+)];

(7) Emission: 3H4 (Tm3+) → 3H6 (Tm3+) + hν (795 nm);

(7) ET3: 2F5/2 (Yb3+) +3H4 (Tm3+) → 2F7/2 (Yb3+) + 1G4 (Tm3+);

(8) Emission: 1G4 (Tm3+) → 3H6 (Tm3+) +hν (476 nm) and 1G4 (Tm3+) → 3F4 (Tm3+) + hν (650nm).

The negligibly weak 695 nm red emission indicates that only a very limited number of the electrons excited by ET2 to the 3F2,3 level directly decay to the 3H6 ground state. On the other hand, the much faster intensity increase of 476 nm blue emission suggests a preferential population of the 1G4 energy level under increasing excitation power.

Figure 7 exhibits fluorescence decay curves for the green emission of Er3+ (524 nm, 2H11/2→4I15/2 transition), red emission of Ho3+ (657 nm, 5F5→4I8 transition) and blue emission of Tm3+ (476 nm, 1G4→3H6). It was found that the decay curve can be well fitted with the second-order exponential equation I(t) = A1exp(-t/τ1)+A2exp(-t/τ2)+B in each case, where I(t) is the fluorescence intensity at time t, A1 and A2 are pre-exponential constants, τ1 and τ2 stand for the decay time of exponential components, and B is a constant. The average lifetime (τ*) can be calculated with the following formula [20,41]:

Figure 7.

Temporal evolution of the 524 nm green, 657 nm red and 476 nm blue emissions of Er3+, Ho3+ and Tm3+, respectively.

The derived τ1 and τ2 values and their weights are tabulated in Table S3, from which τ* values of ~52 ± 2, 260.6 ± 0.7 and 117 ± 1 μs were obtained for the aforesaid emissions of Er3+, Ho3+ and Tm3+, respectively. The short fluorescence lifetime would be beneficial to temporal and spatial resolution of temperature measurement.

Temperature sensing performance of Sr3Y(PO4)3:Yb3+/Ln3+ UC phosphors

Figure 8(a,b) present the temperature-dependent UC spectra and relative emission intensity of Sr3Y0.88(PO4)3:0.10Yb3+,0.02Er3+ phosphor under 1.00 W of 978 nm laser excitation, respectively. It is clear that the 524 nm green emission (2H11/2→4I15/2) gains intensity while the 547 nm green (4S3/2→4I15/2) and 655 nm red (4F9/2→4I15/2) emissions lose intensity with increasing temperature. The opposite trends observed for the two green bands can be ascribed to thermal coupling of the 2H11/2 and 4S3/2 levels [16-18], while intensity loss of the red emission is mostly due to enhanced non-radiative relaxation from the 4F9/2 level by intensified lattice vibration at a higher temperature.16 The fluorescence intensity ratio (FIR) of thermally coupled 2H11/2/4S3/2 levels follows Boltzmann distribution, and can be described as [11,16-19,44,52]

Figure 8.

Temperature-dependent emission spectra under 1.00 W of 978 nm laser excitation (a), relative intensities of the 524, 547 and 655 nm emissions as a function of measurement temperature (b), the dependences of I524/I547 (c) and I524/I655 (d) FIRs on the absolute temperature, the absolute sensitivity of I524/I547 FIR (e) and the relative sensitivities of I524/I547 and I524/I655 FIRs (f) of the Sr3Y0.88(PO4)3:0.10Yb3+,0.02Er3+ phosphor.

where ΔE is the energy gap between the 2H11/2 and 4S3/2 levels, k is the Boltzmann constant (0.695 K−1cm−1), T is the absolute temperature, and N is a proportionality constant. Figure 8(c) shows the temperature dependence of I524/I547 FIR, where it was found that the experimental data can be well fitted with the single-exponential equation of FIR(I524/I547) = 9.7exp(−1026.5/T). The derived ΔE of ~713 ± 4 cm−1 agrees with the energy gap (700–800 cm−1) between 2H11/2 and 4S3/2 [11]. Figure 8(d) shows I524/I655 FIR for the non-thermally coupled levels of 2H11/2 and 4F9/2, where it was found that the experimental data can be satisfactorily fitted with the linear equation of FIR(I524/I655) = 0.0036*T-0.716.

Absolute (SA) and relative (S) sensitivities are two indispensable parameters for temperature sensing, which can be calculated using the following formulas [14,16-18,53]:

For the thermally coupled 2H11/2/4S3/2 levels, it was observed that the SA of I524/I547 FIR first increases to reach its maximum of ~5.07 × 10−3 K−1 at 523 K and then slightly decreases (Figure 8(e)). The non-thermally coupled 2H11/2 and 4F9/2 levels have an SA of ~3.6 × 10−3 K−1 for I524/I655 FIR, according to the linear fitting in Figure 8(d), which is generally smaller than the SA of I524/I547 FIR (Figure 8(e)). As compared in Table 1, our phosphor has a significantly better absolute sensitivity (SA) than Yb3+/Er3+ codoped Y2O3 [7], K3Y(PO4)2 [11], Ca3La6Si6O24 [18], NaYF4 [52], and Ba3La(PO4)3 [14].

Table 1.

A summary of SA and S values, electronic transitions and temperature sensing ranges for some typical temperature sensing UC phosphors doped with Yb3+/Ln3+ pair.

Ln3+	Host	Transition/wavelength (nm)	Range (K)	SA (K−1) (maximum)	SR (K−1)	Ref.
Er3+	Y2O3	2H11/2/4S3/2→4I15/2(525/550)	93–613	4.4 × 10−3(427K)	886.08/T2	[7]
Er3+	K3Y(PO4)2	2H11/2/4S3/2→4I15/2(520/555)	293–553	3.04 × 10−3(553K)	1127.5/T2	[11]
Er3+	Ba3La(PO4)3	2H11/2/4S3/2→4I15/2(522/545)	298–498	4.38 × 10−3(498K)	1002/T2	[14]
Er3+	Ca3La6Si6O24	2H11/2/4S3/2→4I15/2(522/548)	293–573	3.91 × 10−3(500K)	1008/T2	[18]
Er3+	α-NaYF4	2H11/2/4S3/2→4I15/2(524/545)	303–573	4.54 × 10−3(541K)	1085.3/T2	[52]
Er3+	β-NaYF4	2H11/2/4S3/2→4I15/2(524/545)	303–573	4.84 × 10−3(515K)	1025.8/T2	[52]
Ho3+	Y2O3	5F3/3K8 →5I8(465/491)	299–673	3.02 × 10−3 (673 K)	1067.76/T2	[8]
Ho3+	K3Y(PO4)2	5F5/(5F4,5S2)→5I8(659/545)	303–523	0.078(303-523K)	0.20%(303K)	[11]
Ho3+	BaY2Si3O10	5F5/(5F4,5S2)→5I8(662/548)	303–523	0.023(298-448K)	0.49%(298K)	[16]
Ho3+	Ca3La6Si6O24	5F5/(5F4,5S2)→5I8(658/546)	293–533	0.03(293-533K)	0.15%(293K)	[18]
Ho3+	CaMoO4	5F3/3K8 →5I8(460/489)	303–543	6.6 × 10−3(353K)	648.8/T2	[54]
Tm3+	K3Y(PO4)2	3F2,3/3H4→3H6(688/790)	293–553	0.304 × 10−3(553K)	1910.1/T2	[11]
Tm3+	Ba3La(PO4)3	3F2,3/3H4→3H6(690/792)	303–503	0.131 × 10−3(503K)	2.11%(303K)	[14]
Tm3+	KLuF4	3F2,3/3H4→3H6(690/795)	303–503	0.145 × 10−3(503K)	1249.85/T2	[56]
Er3+	Sr3Y(PO4)3	2H11/2/4S3/2→4I15/2(524/547)	298–573	5.07 × 10−3(523K)	1026.5/T2	This work
Er3+		2H11/2/4F9/2→4I15/2(524/655)		3.6 × 10−3 (298-573K)	1.11%(298K)
Ho3+		5F5/(5F4,5S2)→5I8(656/543)		0.019(298-573K)	0.16%(298K)
Ho3+		5I4/5F5→5I8(801/656)		0.46 × 10−3(573K)	0.35%(573K)
Ho3		5I4/(5F4,5S2)→5I8(801/543)		9.39 × 10−3(573K)	0.42%(573K)
Tm3+		3F2,3/3H4→3H6(695/795)		0.82 × 10−3(573K)	1547.7/T2
Tm3+		3F2,3/1G4→3H6(695/476)		1.53 × 10−3(573K)	0.92%(298K)
Tm3+		3F2,3→3H6/1G4→3F4(695/650)		12.74 × 10−3(573K)	1.52%(298K)

The relative sensitivity (S) determined with Equation (6) presents a continuous decrease with increasing temperature for both I524/I547 and I524/I655 FIRs (Figure 8(f)), but the use of I524/I547 produced a larger S than I524/I655 on the whole. As presented in Table 1, the Sr3Y0.88(PO4)3:0.10Yb3+,0.02Er3+ phosphor has maximum S values of ~1.16% (298 K) and 1.11% (298 K) for the I524/I547 and I524/I655 FIRs, respectively, which are slightly smaller than those of K3Y(PO4)2:Yb/Er [11] and NaYF4:Yb/Er [52] but are higher than those of Y2O3:Yb/Er [7], Ca3La6Si6O24:Yb/Er [18] and Ba3La(PO4)3:Yb/Er [14]. Judged from SA and S values, it can be concluded that Sr3Y0.88(PO4)3:0.10Yb3+,0.02Er3+ has a better performance of temperature sensing with thermally coupled 2H11/2/4S3/2 instead of non-thermally coupled 2H11/2/4F9/2 levels.

Figure 9(a) presents the temperature-dependent UC spectra of Sr3Y0.88(PO4)3:0.10Yb3+,0.02Ho3+ under 1.00 W of 978 nm laser pumping, where it is seen that the green (543 nm; 5F4/5S2→5I8), red (656 nm; 5F5→5I8) and NIR (801 nm; 5I4 →5I8) bands continuously lose intensity towards a higher temperature but at different rates (Figure 9(a), the inset). As analyzed in Figure 9(b,c), I656/I543 FIR follows the linear equation of FIR(I656/I543) = 0.019*T+ 6.7, while I801/I656 and I801/I543 FIRs can be fitted with the single-exponential equations of FIR(I801/I656) = 0.0007exp(T/130.04)+0.073 and FIR(I801/I543) = 0.058exp(T/173.01)+0.655, respectively.

Figure 9.

Temperature-dependent emission spectra under 1.00 W of 978 nm laser excitation (a), the dependences of I656/I543 (b) and I801/I656 and I801/I543 (c) FIRs on the absolute temperature, the absolute sensitivity (SA) of I801/I656 and I801/I543 FIRs (d), and the relative sensitivities (SR) of I656/I543, I801/I656 and I801/I543 FIRs (e) of the Sr3Y0.88(PO4)3:0.10Yb3+,0.02Ho3+ phosphor. The inset on part (a) shows relative intensities of the 543, 656 and 801 nm emissions as a function of the absolute temperature. Note the different scales of the vertical axes in parts (d) and (e).

Though the SA of I801/I656 and I801/I543 FIRs steadily increased from ~0.056 × 10−3 to 0.46 × 10−3 K−1 and from ~1.68 × 10−3 to 9.39 × 10−3 K−1 with the temperature increasing from 298 to 573 K (Figure 9(d)), respectively, the maximum values are yet lower than that of I656/I543 FIR (~0.019 K−1, Figure 9(b)). Accordingly, the maximum SA is ~0.019 K−1 for the Sr3Y0.88(PO4)3:0.10Yb3+,0.02Ho3+ phosphor in the tested temperature range, which is lower than the values of K3Y(PO4)2:Yb/Ho [11], BaY2Si3O10:Yb/Ho [16] and Ca3La6Si6O24:Yb/Ho [18] but is higher than those of Y2O3:Yb/Ho [8] and CaMoO4:Yb/Ho [54] (Table 1). Figure 9(e) demonstrates the relative sensitivities derived with Equation (6). While the S of I656/I543 FIR gradually decreased from ~0.16% to 0.12% K−1, those of I801/I656 and I801/I543 FIRs monotonically increased from ~0.059% to 0.35% K−1 and from ~0.19% to 0.42% K−1, respectively. The maximum S (0.42% K−1) of this phosphor is higher than those of Yb3+/Ho3+ co-doped K3Y(PO4)2 [11] and Ca3La6Si6O24 [18], as compared in Table 1.

Figure 10(a) and (b), respectively, show temperature-dependent UC spectra and relative intensities of the emission bands for the Sr3Y0.88(PO4)3:0.10Yb3+,0.02Tm3+ phosphor under 1.00 W of 978 nm laser excitation. It was noticed that the intensities of 1G4→3H6 (476 nm), 1G4→3F4 (650 nm) and 3H4→3H6 (795 nm) emissions continuously decrease while that of 3F2,3→3H6 (695 nm) increases towards a higher temperature. Specifically, the 795 and 695 nm emissions were lowered by ~60% and enhanced by ~4.5 times at 573 K, respectively (Figure 10(b)), which is owing to the fact that 3F2,3 and 3H4 energy levels are thermally coupled and a higher temperature enhances population of the upper-lying 3F2,3 state [11]. Figure 10(c) plots the I695/I795 FIR as a function of the absolute temperature, where it was found that the data can be satisfactorily fitted with the single-exponential equation of FIR(I695/I795) = 2.6exp(−1547.7/T). The derived ΔE value (1075.7 cm−1) is significantly larger than that (713.42 cm−1) between the thermally coupled 2H11/2/4S3/2 levels of Er3+, which suggests that Yb3+/Tm3+ pair can be much better than Yb3+/Er3+ in Sr3Y(PO4)3 for temperature sensing. For the thermally coupled 3F2,3/3H4 energy levels, the error (δ) between estimated energy-gap (ΔEe~1075.7 cm−1) and experimentally measured energy-gap (ΔEm~1709.9 cm−1 from Figure 10(c)) can be calculated with the following expression [4,46,55] to be ~37.1%.

Figure 10.

Temperature-dependent emission spectra under 1.00 W of 978 nm laser excitation (a), relative intensities of the 476, 650, 695 and 795 nm emissions as a function of the measurement temperature (b), the dependences of I695/I795 (c) and I695/I476 and I695/I650 (d) FIRs on the absolute temperature, and the absolute (e) and relative (f) sensitivities for the I695/I476, I695/I650 and I695/I795 FIRs of the Sr3Y0.88(PO4)3:0.10Yb3+,0.02Tm3+ phosphor. Note the different scales of the vertical axes in parts (b), (d), (e) and (f).

The δ of this work is larger than the value (below 20%) [4] reported for Na(Lu,Gd)F4:Tm3+/Yb3+ but is smaller than those (above 40%) [4] for Tm3+/Yb3+ codoped NaNbO3 and Y2O3.

Non-thermally coupled energy levels were also analyzed to see their performance of temperature sensing, and Figure 10(d) shows the dependences of I695/I476 and I695/I650 FIRs on absolute temperature. It is encouraging to see that both the FIRs continuously increase with increasing temperature and the experimental data can be well fitted with the single-exponential equations of FIR(I695/I476) = 0.001exp(T/111.7)+0.00014 and FIR(I695/I650) = 0.053exp(T/158.7)-0.224. As analyzed in Figure 10(e,f), the SA and SR values gradually increase and decrease with increasing temperature, respectively. It is also seen that I695/I650 FIR has the largest SA and I695/I476 FIR has the smallest SR on the whole. It is also seen from the Figures that the 1G4 →3F4 (650 nm)/3F2,3→3H6 (695 nm) non-thermally coupled emissions have the largest SA of ~12.74 × 10−3 K−1 at 573 K while the thermally coupled 3F2,3→3H6 (695 nm)/3H4→3H6 (795 nm) emissions have the largest S value of ~1.74% K−1 at 298 K. As compared in Table 1, the Sr3Y0.88(PO4)3:0.10Yb3+,0.02Tm3+ phosphor has lower S but significant higher SA than K3Y(PO4)2:Yb/Tm [11], Ba3La(PO4)3:Yb/Tm [14] and KLuF4:Yb/Tm [56].

Conclusions

Eulytite-type Sr3Y0.88(PO4)3:0.10Yb3+,0.02Ln3+ phosphors (Ln = Ho, Er, Tm) were synthesized via gel-combustion, and their properties and mechanisms of UC luminescence as well as performances of optical temperature sensing with both thermally coupled and non-thermally coupled energy levels were systematically investigated. The main conclusions are summarized as follows:

(1) The Sr3Y(PO4)3:Yb3+/Ln3+ phosphors exhibit green (Ln = Er), orange-red (Ln = Ho) and blue (Ln = Tm) UC luminescence via two- and three-photon processes under 978 nm NIR laser excitation. The phosphors were analyzed to have the average decay times of ~52 ± 2, 260.6 ± 0.7 and 117 ± 1 μs for the 524 nm green, 657 nm red and 476 nm blue emissions of Er3+, Ho3+ and Tm3+, respectively.

(2) Sr3Y0.88(PO4)3:Yb3+/Er3+ exhibits a better performance of temperature sensing with the thermally coupled 2H11/2 and 4S3/2 energy levels, whose maximum absolute (SA) and relative (S) sensitivities are ~5.07 × 10−3 K−1 at 523 K and ~1.16% at 298 K, respectively.

(3) Sr3Y0.88(PO4)3:Yb3+/Ho3+ shows maximum SA and S values of ~0.019 K−1 (298–573 K) and ~0.42% at 573 K for the non-thermally coupled energy pairs of 5F5/(5F4,5S2) and 5I4/5F5, respectively.

(4) Sr3Y0.88(PO4)3:Yb3+/Tm3+ has a maximum SA of ~12.74 × 10−3 K−1 at 573 K for the non-thermally coupled 3F2,3→3H6/1G4→3F4 emissions and a maximum S of ~1.74% K−1 at 298 K for the thermally coupled 3F2,3→3H6/3H4→3H6 emissions.