Sm3+/Gd3+ Co-doped Lead/Bismuth Borogermanate Glasses – New Materials for Luminescent Thermometry

1. Introduction

Temperature (T) is one of the most fundamental parameter in all technological processes. Thermosensors are widely used in metrology, environment, high-temperature engineering, chemistry, medicine, biology, physiology, air conditioning, in production plants and the storage of goods etc. T-sensors are estimated to fill as much as 75–80% of the world’s sensor market [1]. The first wide spread types of thermometers were liquid-filled glass thermometers [2]. In the 19th century, the real breakthrough in thermometry was made using the Seebeck effect, which gave rise to the era of thermocouples. The next step in thermometry dealt with optical sensors of various types [3], which cross the threshold of 3700 K and made it possible to apply the thermometry in plasma technologies.

In the last 20 years, there has been growing interest in luminescent thermometry (LT), due to its higher efficiency in accurately measuring the temperature of objects with sizes less than 10 μm [4], fast response, high thermal and spatial resolution, as well as a wide range of operating temperatures from 10 K up to 1300 K [5]. Luminescence thermometry can operate in extremely high electric fields [6], unlike thermocouples. This opens up opportunities for its application in industry [7], scientific research [8] and biomedicine [9]. However, in recent decades, the most popular LT temperature range has been the measurement of physiological temperature over a narrow temperature range of 300–330 K with high spatial resolution and accuracy better than 0.1 °C.

There are many scientific applications that require fast sensing with high spatial resolution, unaffected by strong electromagnetic fields, in the temperature range from RT to 673 K and above. For example, for studying heat and mass transfer in liquids in various technological and chemical processes, measuring local temperature is very important. Correct 3D mapping of temperature inside chemical reactors in which processes occur during ultrasonic or microwave activation is a very urgent task for substantiating a numerical model of the technological process. This problem could only be solved using fluorescent thermometry. Sometimes, in order to develop a correct numerical model for the growth of crystals from a melt at high temperatures (1000-3000 K) [10], information on 3D temperature mapping in the volume of the melt is very important to correct the growth process. To solve the problem, scientists use numerical modeling, which requires confirmation by physical experiments at relatively low temperatures [11]. The application of LT for 3D temperature mapping of melt volume is a promising tool for improving numerical models.

Over the decades, two classes of methods: time-resolved methods and spectral-resolved methods, using both emission lifetimes and spectral responses, have found application in remote thermometry.

Temperature determination through LT can be based on monitoring various photoluminescence (PL) temperature-sensitive parameters [12,13,14], including absolute PL intensity (Figure 1a), spectral width of the PL band (Figure 1b), fluorescence intensity ratio of two bands (FIR) (Figure 1c), spectral shift of the PL band maximum (Figure 1d), and excited-state lifetime (Figure 1e).

LT does not provide the value of the temperature directly; instead, it provides an indication (Q). Depending on Q the luminescent thermometry is classified as fluorescent intensity thermometry, imaging luminescent thermometry, fluorescent intensity ratio thermometry (FIR), luminescent decay time thermometry.

The most important parameters to quantifying measurement performance are the measurement temperature range, absolute (S_a) and relative (S_r ) sensitivities, temperature resolution (δT), temporal resolution (δx_min), spatial resolution (δt_min), the repeatability and reproducibility (R).

$$S_a=\left|\frac{dQ}{dT}\right|,$$

(1)

$$S_r=\left|\frac{1}{Q}\times \frac{dQ}{dT}\right|\times 100\%,$$

(2)

$$\delta {\rm T}=\frac{\sigma }{S_a}=\frac{{\sigma }_r}{S_r},$$

(3)

$${\delta x}_{min}=\frac{\delta T}{\left|dT/dx\right|},$$

(4)

$${\delta t}_{min}=\frac{\delta T}{\left|dT/dt\right|},$$

(5)

$$R=1-\frac{max\left|\overline{Q}-Q_1\right|}{Q}$$

(6)

The advantages of LT are high sensitivity, independence of the measured quantity from absolute value of PL intensities and excitation energy [15]. From the technical point of view, FIR could be easy implemented by comparing the intensities of only selected bands using narrowband filters and sensitive semiconducting detectors.

A variety of materials both well-ordered such as inorganic single crystals and strongly disordered such as organic dyes as well as nanostructured materials such as nanogels and quantum dots are used for sensitive elements of luminescent thermometers.

Among the luminescent materials investigated as potential LT sensors, special attention is given to materials doped with lanthanides due to their non-toxicity, thermal stability, high quantum yield, and emission independence from matrix [16]. In LT materials co-doped with different Ln³⁺ ions, e.g., Eu³⁺ and Tb³⁺, the ions are binding between each other and energy-transfer from excited levels may followed by a strong temperature dependence [5,17,18]. This mechanism does not lead to thermal equilibrium of populations in the excited state, which makes possible higher relative sensitivity if the luminescence intensity of various ions is sufficient enough [19,20]. The operating mechanism of energy transfer luminescent thermometers is complex because it depends not only on relaxation processes within the ion, but also involves interactions between different ions as well as the influence of the crystal field of the matrix.

Metalorganic Ln-ion based on β-diketonate chelates demonstrated S_r=4,9 %×K^–1 in the 10-350 K temperature range [5]. High temperature LT sensors based on YAG:Dy and YAG:Dy, Er single crystals [21] performed an excellent S_r=0,9 %×K^–1 in the temperature range up to 1973 K.

Optical quality glasses (contains commonly network former, intermediates, and modifiers) are considered a cheap alternative for creation a uniform doped compact material with adjustable PL, mechanical and chemical properties. For the last decade there were published a number of articles on the problem of functional glasses in the systems of B₂O₃-PbO-Bi₂O₃-GeO₂ [22], Bi₂O₃-GeO₂ [23], and B₂O₃-GeO₂-Gd₂O₃ [24], doped with Sm³⁺ and/or Gd³⁺. The above investigations have proved the perspective of such glasses application as orange lighting sources [25] and key-feature components of optoelectronic devices [26]. Nevertheless, for LT application B₂O₃-PbO-Bi₂O₃-GeO₂ glasses doped by RE has not been reported yet.

In this study, we have investigated for the first time luminescent properties of glasses in the systems of B₂O₃–GeO₂–PbO and B₂O₃–GeO₂–Bi₂O₃ co-doped by Sm³⁺/Gd³⁺ vs temperature. Particular attention was paid to the search for a new function to increase the sensitivity of LT materials.

2. Materials and Methods

2.1. Materials and synthesis

We used Bi₂O₃ 99.999 wt%, GeO₂ 99.995 wt% (LANHIT LTD, Moscow, Russia), PbO 99.5 wt%, B₂O₃ 98.5 wt% (Khimkraft, Kaliningrad, Russia), Sm₂O₃ 99,9 wt%, Gd₂O₃ 99.9 wt% (POLARIT LTD, Moscow, Russia). We synthesized glasses with the general formulas 40B₂O₃–40GeO₂–xPbO(Bi₂O₃)–ySm₂O₃–zGd₂O₃, where x = 15, 16, 17; y = 3, 2.5, 2; z = 2, 1.5, 1 mol% and 42.5B₂O₃–42.5GeO₂–xPbO(Bi₂O₃)–ySm₂O₃–zGd₂O₃, where x = 11.25, 12, 12.75; y = 2.25, 1.875, 1.5; z = 1.5, 1.125, 0.075 mol%.

For the easy presentation of the results, we made the following notation of the glass samples (see Table 1).

Glasses were synthesized in corundum crucibles at 1373 K for 30 min by the standard melt-quenching technique with casting onto a glassy carbon substrate at room temperature. Thermal stresses were removed by annealing at 573 K for 3 h, followed by cooling at ~100 K/h rate. Polished plane-parallel plates with 1.8 mm thickness were fabricated from glasses for further studies.

2.2. Characterization techniques

Structure analysis

The glass samples were grinded to 20 μm powder and their structure was examined by X-ray diffraction using an Equinox-2000 diffractometer (Inel SAS, Artnay, France) with CuKα radiation (λ=1.54056 Å). The element analysis of the synthesized glasses was carried out using X-ray spectral energy-dispersive microanalyzer (EDS Oxford Instruments X-MAX-50, Abingdon, Oxfordshire, UK) on the base of Tescan VEGA3-LMU scanning electron microscope (TESCAN ORSAY HOLDING, Brno, Czech Republic).

Raman spectra were recorded on a QE65000 spectrophotometer (Ocean Optics, Largo, FL, USA) using a 785 nm excitation laser in the frequency shift range of 200–2000 cm^{̶ 1} in backscattering geometry. FT-IR absorption spectra were recorded on a Tensor 27 IR-Fourier spectrometer (Bruker, Ettlingen, Germany) in 400–8000 cm ^{̶ 1} range.

Glass characterization

The glass transition temperatures of the samples were determined by differential scanning calorimetry methods using a NETZSCH STA 449 F3 Jupiter instrument (Erich Netzsch GmbH & Co. Holding KG, Selb, Germany) with 1 K accuracy. The measurements were carried out for 15 mg bulk samples placed in platinum crucibles at 10 K/min heating rate. The density of the samples was determined by the hydrostatic method using a MER123 ACF JR-150.005 TFT balance (Mercury WP Tech Group Co., Ltd., Incheon, The Republic of Korea) with the accuracy of 0.005 g/cm³.

Optical characterization techniques

The refractive index was determined using a MEGEON 72022 gemological refractometer (MEGEON, Moscow, Russia) with 1.78 refractive index liquid (sample n_D < 1.78) and a MIN-8 optical microscope (LOMO JSC, Saint Peterburg, Russia) by measuring the shift of the refracted beam at different preset tilt angles of sample plates located on a special stage (sample n_D > 1.78). The measuring accuracy was 0.01.

The absorption spectra were recorded on a JASCO V-770 spectrophotometer (JASCO Corporation, Tokyo, Japan) in the 190–2700 nm wavelength range with 1 nm step.

The photoluminescence and excitation spectra at room temperature were recorded on an Fluorolog FL3-22 (Horiba Jobin Yvon, Longjumeau, France) with double monohromatization of both excitation and emission radiation in 400–720 nm wavelength range with 1 nm step; the PL decay kinetics were measured on a Fluorolog FL3-22 (Horiba Jobin Yvon, Longjumeau, France) with 370 nm a pulsed diode excitation and Δτ = 1.5 ns. The decay kinetics was carried out using the OriginPro 8 SR4 program (OriginLab Corp., Northampton, MA, USA) using the Fit Exponential procedure.

The thermoluminescence spectra were recorded on an OpticInsight OCEAN-HDX-UV-VIS fiber optic spectrometer (Ocean Insight, Orlando, FL, USA) under excitation with an XML T6 10W 365 nm laser LED (TiaoChongYi, Changchun, China). Heating and cooling of glass samples were carried out using Optical cryo and high temperature stage (SIMTRUM Pte. Ltd., Singapore, Singapore) with temperature control range from room temperature to 673 K.

In all PL measurements the exciting beam spot was 0.5 mm.

3. Experimental results

3.1. Glass samples

Pb-based glasses have a pale-yellow color, which is attributed to the presence of lead oxide (Figure 2). Pb-based glasses with a matrix composition of 40–40 have a more intense yellow color, due to an increase in the content of lead oxide shifts the short-wavelength absorption edge to the long-wavelength region of the spectrum [27].

Bi-based glasses had a ruby-red color due to the absorption shoulder in the region of 500 nm attributed with bismuth active centers [28], which becomes more saturated at bismuth oxide content increase.

Figure 2. The photographs of B₂O₃–GeO₂–PbO and B₂O₃–GeO₂–Bi₂O₃ glasses co-doped by Sm³⁺/Gd³⁺.

Figure 2. The photographs of B₂O₃–GeO₂–PbO and B₂O₃–GeO₂–Bi₂O₃ glasses co-doped by Sm³⁺/Gd³⁺.

The results of the elemental analysis (XFS-EDS) of the glasses showed that all samples were characterized by homogeneous composition and all of them contained an aluminum impurity due to the synthesis in corundum crucibles, which was explained by the chemical action of fluid PbO or Bi₂O₃ with the crucible material (for details see Figure S31-S42 and Table S1).

3.2. Glass structure characterisation

X-ray diffraction patterns of all synthesized glasses demonstrated a wide halo in 21-36 degree 2θ range (Figure S1) which proved the amorphous nature of all glass samples.

Samples of the two systems under consideration have common vibron modes in all frequency regions of Raman spectra (Figure S2-S3) and the corresponding interpretations are presented in Table 2, which correlates with various literature data.

In the low-frequency range (300–700 cm ^{̶ 1}) we observed a band in the 460-465 cm ^{̶ 1} range which was referred to the symmetrical stretching of Ge–O–Ge bonds in 4- or 6-membered Ge⁴⁺ rings in a tetrahedral arrangement [29,30,31]. The 800 cm ^{̶ 1} band referred to vibrations of oxygen atoms in the boroxol ring with the participation of a very small movement of boron atoms [32,33]. The presence of broad bands in the high-frequency region (1250–1500 cm ^{̶ 1}) were associated with stretching of the B–O^– terminal bonds in borate structures containing metaborate triangles [33,34,35,36].

We observed four additional bands: at 390, 650, 715, and 890 cm ^{̶ 1} for Bi-based glasses. The 390 cm^–1 band was associated with the bending modes of the Ge–O–Ge [37]. The 650 cm ^{̶ 1} band was referred as vibrations of the 6-coordinated germanium atom in the glass structure [30,38]. The 890 cm ^{̶ 1} band is generally attributed with a double degenerate vibration [39]. According to the Ref. [31] the 715 cm ^{̶ 1} band characterized tensile vibrations of the B-sublattice relative to the O-sublattice and out-of-plane ring B–O–B vibration.

Table 2. Interpretation of Raman bands in Pb-based and Bi-based glasses.

Raman shift, cm–1		Interpretation	Reference
Pb-based	Bi-based
	390	Bending modes of Ge–O–Ge bonds	[37]
465	460	Symmetrical stretching of Ge−O−Ge bonds in 4- or 6-membered rings of Ge4+ in a tetrahedral arrangement	[29,30,31]
	650	Vibrations of 6-coordinated atom Ge (VI)	[30,38]
	715	Stretching vibration of the B-sublattice against the O-sublattice, a B–O–B out-of-plane ring vibration	[31]
800	800	Vibration of the oxygen atoms in the boroxol ring	[32,33]
	890	Ge–O–Ge double degenerate vibration	[39]
1315	1240	B−O− stretching mode in metaborate structures	[33,34,35]
1500		Overlap of following modes: B–O– in BØ2O-, BØ2O- triangle linked to BØ4 unit, and stretching in BØ3 triangles	[36]

Table 2. Interpretation of Raman bands in Pb-based and Bi-based glasses.

Raman shift, cm–1		Interpretation	Reference
Pb-based	Bi-based
	390	Bending modes of Ge–O–Ge bonds	[37]
465	460	Symmetrical stretching of Ge−O−Ge bonds in 4- or 6-membered rings of Ge4+ in a tetrahedral arrangement	[29,30,31]
	650	Vibrations of 6-coordinated atom Ge (VI)	[30,38]
	715	Stretching vibration of the B-sublattice against the O-sublattice, a B–O–B out-of-plane ring vibration	[31]
800	800	Vibration of the oxygen atoms in the boroxol ring	[32,33]
	890	Ge–O–Ge double degenerate vibration	[39]
1315	1240	B−O− stretching mode in metaborate structures	[33,34,35]
1500		Overlap of following modes: B–O– in BØ2O-, BØ2O- triangle linked to BØ4 unit, and stretching in BØ3 triangles	[36]

The Fourier transform infrared spectra were interpreted based on the assumption that the B–O and O–Ge–O bonds play a key role in the glass structure. The FT-IR spectra of bismuth/lead borogermanate glasses doped with Sm³⁺ and Gd³⁺ ions were investigated in the 500–2000 cm ^{̶ 1} range (Figure S4-S5). All analyzed glass samples demonstrated active vibrational modes, indicating the presence of typical characteristics for a borogermanate glass network.

The 540 and 544 cm ^{̶ 1} bands referred to the deforming vibrations of the Ge–O bond [40]. The 773 cm ^{̶ 1} band was related to the Ge–O stretching vibrations in the structural units [GeO₄] [40,41]. Bands in the mid-frequency region (>800 cm ^{̶ 1}) at 805, 1170, 1185, 1450, and 1547 cm ^{̶ 1} described the stretching of the B–O bond. The 805 cm ^{̶ 1} band is attributed with [BO₄] unit stretching [34,41], while the bands at 1170, 1185, 1450, 1547 cm ^{̶ 1} are responsible for [BO₃] unit stretching [41,42].

An analysis of IR spectra obtained by FT-IR and Raman spectroscopies did not reveal any contradictions. Hypotheses regarding the structure of glass presented earlier [29,30,31,32,33,34,35,36,37,38,39,40,41,42] were confirmed.

3.3. DSC characterisation and physical properties

To determine the maximum operating temperature for luminescent thermometry for the obtained glasses we measured the glass transition temperatures by the DSC technique (Table 3, Figures S6-S17). To the date as an upper limit for the maximum operating temperature for luminescent thermometry we chose the 673 K temperature which was ~100 K lower the glass transition temperatures for all glasses.

The density of the samples naturally increases with an increase in the content of heavy metal oxide and a decrease in the content of B₂O₃ and GeO₂ (Table 4). The densities of Bi-based samples were higher compared to Pb-based ones, which could be associated with a higher Bi molecular weight [12]. A similar trend we observed for the refractive index (Table 4). It should be noted that the change in the content of lead oxide by ±1 mol. % did not significantly affect the refractive index values.

3.4. Spectral-luminescent properties

We observed that the absorption spectrum of Pb-based glasses (Figure 3) included 11 main bands associated with f–f transitions from the ground state (⁶H_5/2) to various excited states of Sm³⁺. The transition to the ⁶P_3/2 state at 402 nm is the most intense, since the 4f electrons are screened by the 5s and 5p shells [43]. The remaining absorption bands are observed in the IR region. The most intense band corresponds to the ⁶H_5/2 → ⁶F_7/2 transition. The intensity of all absorption bands increased with increasing content of Sm³⁺ in the obtained Pb-based glasses.

The absorption spectra of Bi-based samples similarly included the above 11 bands specific for Sm³⁺ (Figure 4). The bands attributed to the ⁴I_9/2, ⁴I_11/2 and ⁴I_13/2 transitions (~475 nm) are invisible, since the absorption in the region of 500 nm of bismuth active centers (BACs) overlaps these bands [28,45]. An increase in the Bi₂O₃ concentration resulted in the intensity growth of 475 nm band, and in a shift of the short-wavelength absorption edge to the long-wavelength region, which led to a narrowing of the transparency window of the Bi-based glasses.

The intensive excitation band for all glasses was observed around 400 nm (Figure 5, S18). The intensity of the ⁶H_5/2 → ⁴K_11/2 transition at 404 nm is the highest regardless of the composition.

The photoluminescence (PL) spectra of glasses (Figure 6) revealed the Sm³⁺ bands at 562 nm (⁴G_5/2 → ⁶H_5/2), 597 nm (⁴G_5/2 → ⁶H_7/2), 645 nm (⁴G_5/2 → ⁶H_9/2) and 704 nm (⁴G_5/2 → ⁶H_11/2) resulted from the transitions between 4f–4f levels. The most intense transition is ⁴G_5/2 → ⁶H_7/2, which corresponds to the orange luminescence. An increase in the Sm³⁺ ion concentration led to concentration quenching. In the obtained glasses Gd³⁺ played a role of a sensitizer with respect to Sm³⁺ [46]. The presence of Gd³⁺ increased the PL intensity in times. The most intense luminescence was obtained for the (12.75)Pb(1.5)Sm(0.75)Gd-42.5 and (12.75)Bi(1.5)Sm(0.75)Gd-42.5 glass samples.

An analysis of PL kinetics (Figure 7, for details see Figure S19-S30) showed that the decay curves could be described by bi-exponential decay kinetics. The fast PL centers had the lifetime of 90-100 ns while the slow PL centers had the lifetime of 590-620 ns (Table 5).

The (12.75)Pb(1.5)Sm(0.75)Gd-42.5 and (12.75)Bi(1.5)Sm(0.75)Gd-42.5 glass samples demonstrated the most intensive PL (Figure 7, 8) and they were chosen for the testing of LT properties in the 298–673 K temperature range.

At temperature rise, the (12.75)Pb(1.5)Sm(0.75)Gd-42.5 sample demonstrated an appearance of 525–542 nm band starting with 423 K and its further growth up to 673 K, which could be associated with an increase in the population of the overlying ⁴F_3/2 state due to transitions from the ⁴G_5/2 level in samarium ions under the influence of temperature [47]. Simultaneously, the most intensive 597 and 645 nm bands gradually reduced at temperature rise from 423 to 673 K in 1.4 times maximum (Figure 8).

It is of common use that FIR is calculated using the ratio of normalized intensities of the band (Figure 9a). So, calculate FIR in the 423 to 673 K temperature range we used PL intensity of the ⁴F_3/2 → ⁶H_5/2 transition (${\lambda }_{PL}^{max}=$530 nm) as a constant value of numerator while for a denominator the PL intensities of the ⁴G_5/2 → ⁶H_5/2 (${\lambda }_{PL}^{max}=$ 562 nm); ⁴G_5/2 → ⁶H_7/2 (${\lambda }_{PL}^{max}=$597 nm); ⁴G_5/2 → ⁶H_9/2 (${\lambda }_{PL}^{max}=$645 nm) transitions were used (Figure 9b). At temperature increase starting from 423 K the PL peak at 535 nm attributed with ⁴F_3/2 → ⁶H_5/2 transition has been growing up. Using this thermosensitive transition for the numerator and thermostable transition ⁴G_5/2 → ⁶H_7/2 as the denominator we calculated FIR according to the equation (7).

$$FIR=\frac{I_{{}^{4}F_{3/2}}\to {}_{{}^{6}H_{5/2}}}{I_{{}^{4}G_{5/2}}\to {}_{{}^{6}H_{7/2}}},$$

(7)

The temperature dependence of FIR was approximated by a linear function (FIR = $a+b\times T$) (Figure 10a) and an exponential Boltzmann function (FIR= $a\times {exp}^{-∆E/kT}+b$) (Figure 10b). Regardless of the function the best descriptions of the obtained experimental data of FIR was calculated for the transitions I(⁴F_3/2 → ⁶H_5/2)/I(⁴G_5/2 → ⁶H_7/2) FIR=I₅₃₅/I₆₄₅ . We obtained Adj. R²=0.90644 for linear function and Adj. R²=0.97815 for the Boltzmann function.

The relative thermal sensitivity (S_r) for the parameter under consideration was obtained S_r ~0.20 %×K ^{̶ 1} for the temperature range 423–673 K.

We tried to increase FIR sensitivity by manipulation of different peak intensities. Indeed, when calculating FIR as a ratio

$$FIR=Q=\frac{I_1}{I_2},$$

(8)

to increase the sensitivity S_r we need to reduce the denominator at temperature growth.

In the case of (12.75)Pb(1.5)Sm(0.75)Gd-42.5 sample the generalized picture of relative PL intensities redistribution with temperature are presented in Figure 11b. The PL intensity of ⁴F_3/2 → ⁶H_5/2 transition is growing up in 2 times. And this resulted to increase of the nominator in Eq. 7. But the denominator associated with PL intensity of ⁴G_5/2 → ⁶H_7/2 transition stayed unchangeable. We need to reduce the denominator. To do this we can use the difference of the constant PL intensity of ⁴G_5/2 → ⁶H_7/2 transition and growing up PL intensity of ⁴G_5/2 → ⁶H_9/2 or ⁴G_5/2 → ⁶H_5/2. Both of the latters are temperature sensitive transitions. In this case new efficient FIR_eff could be calculated from the expression

$${FIR}_{eff}(1)=\frac{I_{{}^{4}F_{3/2}}\to {}_{{}^{6}H_{5/2}}}{I_{{}^{4}G_{5/2}}\to {}_{{}^{6}H_{7/2}}-I_{{}^{4}G_{5/2}}\to {}_{{}^{6}H_{5/2}}},$$

(9)

$${FIR}_{eff}(2)=\frac{I_{{}^{4}F_{3/2}}\to {}_{{}^{6}H_{5/2}}}{I_{{}^{4}G_{5/2}}\to {}_{{}^{6}H_{7/2}}-I_{{}^{4}G_{5/2}}\to {}_{{}^{6}H_{9/2}}},$$

(10)

Figure 11 demonstrated the privilege of such approach when using the same raw data one can increase S_r in several times by manipulations of the PL intensities from different T-sensitive and non-sensitive transitions.

Figure 11. The temperature dependence of FIR for (12.75)Pb(1.5)Sm(0.75)Gd-42.5 sample calculated using the Eq.(7) red dots (1); Eq.(9) green dots (2), Eq.(10) blue dots (3).

Figure 11. The temperature dependence of FIR for (12.75)Pb(1.5)Sm(0.75)Gd-42.5 sample calculated using the Eq.(7) red dots (1); Eq.(9) green dots (2), Eq.(10) blue dots (3).

For the (12.75)Pb(1.5)Sm(0.75)Gd-42.5 sample the corresponding sensitivities for different FIRs occurred to be S_r(1) = 0.20 %×K^–1; S_r(2) = 0.68 %×K^–1; S_r(3) = 0.61 %×K^–1. The better achieved sensitivity was more than in 3 times higher than that determined by the classical expression (7).

In the series of Bi-based samples, the most interesting was the (17)Bi(2)Sm(1)Gd-40 glass. We detected a redistribution of the intensities of the peaks corresponding to the ⁴G_5/2 → ⁶H_7/2 and ⁴G_5/2 → ⁶H_9/2 transitions (Figure 12, 13a). This redistribution is a temperature-sensitive parameter, which could be used for temperature determination. Such redistribution is usually observed for dyes [48], but we found out it for the glass sample for the first time.

So, calculating FIR in 298–673 K temperature range we tried to choose the most sensitive parameter based on PL intensities of the bands which had been redistributed: ⁴G_5/2 → ⁶H_9/2 (${\lambda }_{PL}^{max}=$ 645 nm) or ⁴G_5/2 → ⁶H_7/2 (${\lambda }_{PL}^{max}=$ 597 nm) (Figure 13b).

We varied the PL intensities ratio by changing nominator and denominator vice versa. The better approximation was derived for linear function when the I(⁴G_5/2 → ⁶H_7/2)/I(⁴G_5/2 → ⁶H_9/2) ratio was used (Figure 14a). Contrary, for the Boltzmann function the better result was achieved for the I(⁴G_5/2 → ⁶H_9/2)/I(⁴G_5/2 → ⁶H_7/2) ratio (Figure 14b). But taking into consideration the accuracy of the experimental data both these functions could be used in the case of (17)Bi(2)Sm(1)Gd-40 glass for luminescent thermometry. The relative thermal sensitivity for the considered parameter was ~0.105 %×K ^{̶ 1} for 298-673 K temperature range.

4. Discussion

Decay kinetics studies let us better understand the nature of PL mechanism in the analyzed samples. The obtained values of life-time in our research are in order smaller than those described in Ref. [46,49,50,51,52]. And in the majority of publications the authors presented one-exponential decay kinetics when measured the millisecond range [48,49,50]. The transition to microsecond range resulted to detecting two-exponential kinetics [46,52]. In our experiments we investigated the nanosecond decay range, in which we found out two-exponential kinetics with the contribution of the fast component was ~40%.

Decay kinetics studies demonstrate that the addition of Gd³⁺ ions provides an additional energy transfer mechanism, resulting in a decay time increase. Initially, the absorbed energy is transferred from the matrix to Gd³⁺ ions, and then further to the Sm³⁺ ions, which results to two different luminescent centers and affects the total decay time. This mechanism was also discussed in Ref. [46] and the authors came to the same conclusions.

In Pb-based glasses there is no specific dependence on the decay time, which correlates with the absence of absorption bands in 500-600 nm region (Figure 3). While in Bi-based glasses, an increase in the matrix content reduces the short life-time value, since the BACs additionally interacts with Sm³⁺ ions and in 500-600 nm region, the Bi-contained matrix has a characteristic absorption band (Figure 4).

In previous studies [43] devoted to the use of similar glasses as materials for non-contact thermometers, a relatively narrow operating temperature range (288-333 K) was considered with S_r = 1.68%×K ^{̶ 1} (288 K) and S_r = 1.25 %×K ^{̶ 1} (333 K). We have expanded the working range to 298-673 K and found out that Pb-based glasses could give S_r = 0.68%×K ^{̶ 1} (423-673 K) while Bi-based glasses had rather low S_r = 0.105%×K ^{̶ 1}. So, we could not replace the toxic Pb-based glasses on non-toxic Bi-based glasses in LT applications for the large temperature range.

In the case of Bi-based glass one of the problem for LT is the interference of BACs with PL band if RE³⁺ ions. To solve the problem one can suppress BACs by introducing of alkali ions in a large amount [37]. But the LT properties of such glasses need the further investigation.

The achieved values of S_r in a wide temperature range open up real prospects for the use of the Pb-based glasses for thermal detection in medicine, energy and industry. In addition, in contrast to the previous study [43], we recorded not only changes in the PL peak intensity, but also in the redistribution of intensities between peaks, as well as the appearance of new peaks for the temperature measurement. Using the multiply function for FIR, which taking into account both thermal sensitive and non-sensitive transitions we have succeeded to increase the sensitivity of LT materials in 3.4 times.

It is worth emphasizing that the borogermanate matrix, thanks to GeO₂, is characterized by low phonon energy, while the presence of B₂O₃ enhances the glass’s resistance to thermal influences, which is critical for thermoluminescence detector operating in a wide temperature range.

5. Conclusions

Sm³⁺/Gd³⁺ co-doped glasses in the B₂O₃–GeO₂–PbO and B₂O₃–GeO₂–Bi₂O₃ systems were fabricated by conventional glass technology. The structural, optical and luminescence properties of the glasses have been studied. It was demonstrated that the obtained glasses revealed the photoluminescence in the wide temperature range of 298-673 K. The best temperature-sensitive luminescence was found out for (12.75)Pb(1.5)Sm(0.75)Gd-42.5 and (17)Bi(2)Sm(1)Gd-40 glass samples. An increase in temperature led to an increase in the population of the ⁴F_3/2 Sm³⁺ state due to transitions from the lower ⁴S_3/2 level for glasses under study.

The thermal sensitivity for the (12.75)Pb(1.5)Sm(0.75)Gd-42.5 glass sample was estimated to be about 0.250 %×K ^{̶ 1} and for the (17)Bi(2)Sm(1)Gd-40 glass sample ~0.105 %×K ^{̶ 1} based on the ratio of the luminescence intensities of transitions from different Stark sublevels.

For the first time we found out the effect of Bi-based samples, the most interesting was. We detected a redistribution of the intensities of the peaks corresponding to the ⁴G_5/2 → ⁶H_7/2 and ⁴G_5/2 → ⁶H_9/2 transitions in the inorganic phosphor (17)Bi(2)Sm(1)Gd-40 glass which is usually observed for dyes.

We used two temperature sensitive and one non-sensitive transitions to calculate FIR. This made it possible to increase the sensitivity of LT in several times. Such method could be very useful for the LT materials based on Ln ions with multiplied transitions.

The results showed that the produced glasses are perspective for photonics applications, in particular in luminescence thermometry.