Microwave Synthesis of Luminescent Recycled Glass Containing Dy₂O₃ and Sm₂O₃

Abstract

This research studied the recycling of borosilicate glass wastes from damaged laboratory glassware. The luminescent glasses were prepared by doping glass waste powder with rare earth ions, namely, dysprosium ions (Dy³⁺) and samarium ions (Sm³⁺), as well as co-doping with Dy³⁺ and Sm³⁺ at a concentration of 2% by weight. The sintering process was conducted in a microwave oven for a duration of 15 min. The photoluminescence spectra of the doped glasses were obtained under excitation at 401 nm and 388 nm. The results showed that the emission characteristics depended on the doping concentrations of Dy³⁺ and Sm³⁺ and the excitation wavelengths. Upon excitation at 401 nm, the co-doped glasses exhibited the maximum emission peak of Sm³⁺ at 601 nm (yellowish and orange region in the CIE chromaticity diagram) due to the energy transition from ⁴G_5/2 to ⁶H_7/2. When excited at 388 nm, however, the emission spectra of the co-doped glasses were similar to the characteristic emission peaks of Dy³⁺ (white region in the CIE chromaticity diagram), but the peak position exhibits a red shift. This could be attributed to an increase in the amount of non-bridging oxygens (NBOs) by co-doping.

1. Introduction

For thousands of years, glass has been the most widely used packaging material. It is generally non-porous and typically inert with respect to its contents. Moreover, glass is transparent, heat-resistant, and fully recyclable. Waste glass is typically recycled by categorizing it according to its intended purpose. However, the majority of recycled waste glass has mostly been used for the production of bottles and building materials [1]. Although the melting of new glass from recycled materials requires significantly less energy than that of glass from raw materials, additional energy is still required for the recycling processes, which include sorting, cleaning, crushing/grinding, and remelting. It is, therefore, essential to continuously develop technologies to utilize waste glass. For example, it can be employed as a partial or complete replacement for certain materials such as luminescent materials [1,2,3], aggregates in concrete [4], bricks [5,6], and composite materials [7].

Color tunability of luminescent materials is very desirable in the illumination and display industry. For instance, white light emission may be produced from a combination of three color-emission centers or a combination of blue LEDs and yellow phosphors [8,9]. Previous research has demonstrated the effectiveness of co-doping with a variety of rare earth ions to achieve specific emission characteristics in luminescent materials. Dy³⁺ typically exhibits two distinct emission peaks: a blue peak at around 480 nm attributed to the ⁴F_9/2→⁶H_15/2 transition, and a yellow peak at approximately 580 nm corresponding to the ⁴F_9/2→⁶H_13/2 transition. Sm³⁺, on the other hand, exhibits an emission peak at approximately 600 nm due to the ⁴G_5/2→⁶H_7/2 transition. The concentration ratios of Dy³⁺ to Sm³⁺ can be adjusted to produce various emission colors, including the potential for white light emissions [10,11,12].

W. Yan et al. examined the synthesis and luminescent properties of NaSrPO₄ phosphors co-doped with Dy³⁺ and Sm³⁺ ions, which exhibited warm-white luminescence [11]. Upon excitation with UV light at 348 nm, the co-doped phosphors emitted white light, resulting from the combination of two primary emission peaks: one at 484 nm (blue) and another at around 575 nm (orange). The warm-white luminescence was attributed to energy transfer between Dy³⁺ and Sm³⁺ ions. Specifically, energy was transferred from the excited states of Dy³⁺ to Sm³⁺, enhancing the overall emission spectrum and enabling the production of warm-white light.

C. Chen et al. investigated the development of a specific type of glass-ceramic material for white light-emitting diode technology [12]. The study focused on borosilicate glass-ceramics co-doped with dysprosium (Dy³⁺) and samarium (Sm³⁺), which were enhanced by the incorporation of Na₅Y₉F₃₂ nanocrystals. It was found that the emission color could be tuned from yellow to white by applying the excitation at different wavelengths. With an increase in the doping concentration of Sm³⁺ and a reduction in Dy³⁺, the emission of Sm³⁺ was enhanced, resulting in the production of white light. The emission spectrum of the Sm³⁺-doped samples, excited at a wavelength of 401 nm, showed several emission bands at 563, 600, 647, and 707 nm, corresponding to the ⁴G_5/2→⁶H_J/2 (J = 5, 7, 9, 11) transitions, respectively. The emission spectra of the co-doped glass-ceramic, when excited at 350 nm, exhibited three emission bands at 484, 575, and 665 nm, corresponding to the ⁴F_9/2→⁶H_J/2 (J = 15, 13, 11) transitions of Dy³⁺, respectively. The energy transfer between Dy³⁺ and Sm³⁺ ions revealed in this study was a crucial process for producing white light in borosilicate glass.

Sintering is a process that can be used to compact or fuse glass particles by heat to form a solid mass [2]. This process is often carried out in a conventional furnace. Microwave heating is a densification sintering method that has gained a lot of interest. It is a novel method that employs microwave radiation to heat and sinter materials, predominantly metal powders and ceramics. This method differs significantly from conventional sintering processes, which depend on external heat sources. In microwave sintering, the materials absorb microwave radiation, generating heat internally rather than relying on surface heating from an external source. This results in more uniform heating throughout the material, which may enhance the properties of the final product. In addition, it also has short processing times and high energy efficiency, making it a potential method for glass manufacturing.

K. Ashis et al. investigated the application of microwave radiation for the synthesis of iron-doped alumino-phosphate glass [13]. The study showed that the microwave method reduced the melting time to less than 2 h, compared to the 6–7 h required by conventional methods. Moreover, the glass produced using microwave radiation exhibited comparable structural and optical properties to that produced by conventional heating methods, confirming the effectiveness of microwave processing. Although the microwave-assisted synthesis or the sintering of materials is faster and energy-efficient, not all materials can effectively interact with microwave radiation. Silicate glasses are typically transparent to microwaves and cannot be directly heated by them at low temperatures. At high temperatures, they may absorb microwave radiation due to structural relaxation, resulting in further heating [14]. Therefore, the microwave heating of glass frequently requires a microwave susceptor, such as silicon carbide (SiC), to initiate the heating process until the glass can absorb microwaves independently.

This research, therefore, focuses on tuning the emission color of borosilicate waste glasses by varying the doping ratios of Dy₂O₃ and Sm₂O₃ as well as the excitation wavelength. The study also addresses the use of microwave heating as an effective, rapid, and energy-efficient method for sintering glass samples, offering an alternative to conventional sintering method.

2. Materials and Methods

The glass waste was collected from broken laboratory glassware inside the department at Silpakorn University.

Table 1. Chemical composition of the glass waste.

Component	Weight%	Component	Weight%
B2O3	14.83	K2O	0.44
Na2O	3.74	CaO	0.03
Al2O3	2.53	TiO2	0.03
SiO2	78.20	Fe2O3	0.04
P2O5	0.01	PbO	0.06
MgO	0.01	Total	99.92

presents the chemical composition of the glass waste, confirming that it is borosilicate glass. After the collection of glass waste, the glass was carefully cleaned and allowed to dry in air. It was then crushed into small fragments and ground in a ball mill using alumina balls, occupying about three to four portions of the mill jar. The rotation velocity was set at 700 revolutions per minute for a duration of 4 h. Subsequent to milling, the resultant glass powder was sieved through a 325-mesh sieve (particle size of 44 µm).

Borosilicate glass powder and rare earth compounds, Dy₂O₃ (Sigma Aldrich, St. Louis, MO, USA, 99.9%) and Sm₂O₃ (Sigma Aldrich, 99.9%), were dried in an oven at 80 °C and then weighed as detailed in

Table 2. Composition of the glass samples.

Sample	Composition
	Borosilicate Powder (g)	Dy2O3 (g)	Sm2O3 (g)
MD0	3.00	0	0
MD1	2.94	0.06	0
MD2	2.94	0.04	0.02
MD3	2.94	0.03	0.03
MD4	2.94	0.0.2	0.04
MD5	2.94	0	0.06

. They were then wet-mixed together in an ethanol medium using a mortar and pestle. The mixed powder was compacted into pellets using a hydraulic press at a pressure of 1.5 tons for 90 s. The obtained pellets were about 1.5 cm in diameter. The specimens, designated MD1–MD5, were then placed in a microwave kiln (inner dimension: Ø8 × 4.5 cm) and subjected to sintering in a household microwave oven at 800 W for 15 min. The microwave kiln is internally coated with silicon carbide, allowing it to absorb microwaves and produce heat. After sintering, all samples were characterized to determine their properties. The undoped glass (MD0) was also prepared by both conventional and microwave sintering processes for comparative analysis.

The apparent and bulk densities were determined using the Archimedes principle. The following equations are used in the calculations:

Apparent density = W_D/(W_D − W_I)

(1)

Bulk density = W_D/(W_S − W_I)

(2)

where W_D is the dry weight of the glass sample, W_I is the weight of the glass sample when immersed in water, and W_S is the saturated weight after submerging in water. A Shimadzu XRD-6100 X-ray diffractometer (Kyoto, Japan) was used to analyze the crystallinity of the glass samples. The data were gathered in the 2-theta range of 10–60° with a voltage of 30 kV and a current of 20 mA at room temperature. The FTIR spectra were recorded in the range of 400–2200 cm⁻¹, using a Vertex 70v FT-IR spectrometer (Bruker, Leipzig, Germany). The surface morphology was observed with a scanning electron microscope (Tescan Mira3 FEG-SEM, Brno–Kohoutovice, Czech Republic). UV–Vis–NIR absorption spectra were measured in the wavelength range of 200–1800 nm using a UV–Vis–NIR spectrophotometer (Shimadzu UV-3600 Plus, Kyoto, Japan). Excitation and emission spectra were obtained using a Horiba Scientific FluoroMax+ spectrofluorometer (Kyoto, Japan).

3. Results and Discussion

Figure 1 illustrates the physical and structural properties of undoped glasses sintered with conventional heating at different temperatures for one hour and microwave heating at 800 W for varying durations. The results indicate that the properties of microwave- and conventionally sintered glass samples are comparable, suggesting that microwave heating can effectively be utilized for glass sintering. Based on XRD analysis, sintering via the conventional method, however, is likely to promote the crystallization of the cristobalite phase, which may affect both the mechanical and optical properties of the glasses. The SEM images also show that the samples prepared by the microwave sintering technique exhibit better quality compared to those subjected to conventional sintering [15]. The conventional heating method requires a temperature as high as 1100 °C to obtain smooth and homogeneous glass samples. Microwave heating is, therefore, a potential technique with uniform heating and low energy consumption for the sintering process. For the microwave sintering of doped glasses, a 15-min sintering time has been used.

Figure 2 compares the physical characteristics of glass samples doped with different concentrations of Dy³⁺ and Sm³⁺ and sintered in a microwave oven for 15 min under both daylight and UV irradiation. The uniformity and smooth, rounded surface of the borosilicate glass samples indicate that glass melting initiates during the sintering period. The smooth surface of the glass samples is attributed to uniform heating facilitated by microwave sintering. When examined in the dark under UV light, the samples exhibit visible luminescence with the emission color varying according to the concentrations of Dy³⁺ and Sm³⁺.

The density of the doped glasses was measured using the Archimedes method. This is an important fundamental property that can be used to determine the structure of glasses and the quantity of porosities. The apparent and bulk densities of the doped glass samples prepared in a microwave oven are shown in Figure 3. The densities vary from 1.59 to 1.92 g/cm³ and 1.58 to 1.91 g/cm³, respectively, depending on the Dy³⁺−Sm³⁺ concentrations. Although the density of glass samples increases with higher Dy³⁺ content, it is still much lower than that of undoped glass. This could be because the incorporation of Dy³⁺ and Sm³⁺ into the borosilicate glass matrix might induce the porosities and the formation of non-bridging oxygen (NBO). The high density of the glass matrix in Dy³⁺-doped glass relative to Sm³⁺-doped glass can be ascribed to the greater molecular weight of Dy₂O₃ in comparison to Sm₂O₃. Additionally, Dy³⁺ ions may be substituted inside the glass network. According to earlier studies, Dy³⁺ can replace borate ions, resulting in an increase in the density of the glass [16].

Figure 4 shows the XRD patterns of the glass samples. The powdered samples were prepared for XRD analysis by hammer crushing and grinding with a mortar and pestle. All samples exhibit a broad hump centered at about 22.5°, indicating that the glass is amorphous and lacks the long-range order. This indicates that the crystallization is not induced by the addition of doping oxides. The formation of a crystalline cristobalite phase is often observed in the current glass system prepared by the conventional sintering method, as shown in Figure 1. The results are consistent with previous studies on the microwave sintering of red-emitting glasses [2].

The FTIR spectra of the glass samples are illustrated in Figure 5. The spectra reveal peaks at 465 cm⁻¹, 670 cm⁻¹, 793 cm⁻¹, 905 cm⁻¹, 1089 cm⁻¹, and 1414 cm⁻¹, corresponding to various vibrational modes: the bending vibration of Si-O-Si, the bending vibration of bridging oxygen [BO₃], the symmetric stretching of O-Si-O, the vibration of Si-B-O bonds in [BO₄], the symmetric stretching of [BO₃] and [BO₄] groups overlapped with the symmetric stretching of Si-O-Si, and the symmetric stretching vibration of B-O bonds in [BO₃] groups, respectively [17,18]. It can be clearly seen that the spectra of the doped glasses and the undoped glasses are very similar. This suggests that the incorporation of Dy³⁺ and Sm³⁺ into the borosilicate glass matrix does not alter the glass network. The band appearing at 1089 cm⁻¹ of the undoped glasses, however, appears to be more intense, indicating a larger BO₃ content. In other words, the doping could potentially lead to the formation of non-bridging oxygens (NBOs) [19]. The appearance of NBOs in borosilicate glasses could be related to either Si and B units.

Figure 6 illustrates the absorption spectra of borosilicate glass doped with Dy³⁺ and Sm³⁺. The Dy³⁺-doped glass sample (MD1) (Figure 6a) exhibits four characteristic peaks at 894, 1081, 1256, and 1668 nm, corresponding to energy transitions from the ground state ⁶H_15/2 to states ⁶F_7/2, ⁶F_9/2, ⁶F_11/2 + ⁶H_9/2, and ⁶H_11/2 of Dy³⁺, respectively. The most prominent transition occurs at 1256 nm due to the ⁶H_15/2-6F_11/2 + ⁶H_9/2 transition. On the other hand, the light absorption spectrum of the glass sample doped with Sm³⁺ (MD5) shows seven characteristic peaks at 402, 1074, 1223, 1367, 1470, 1524, and 1597 nm. The observed peaks are ascribed to energy transitions from the ground state ⁶H_5/2 to the excited states ⁶P_3/2, ⁶F_9/2, ⁶F_7/2, ⁶F_5/2, ⁶F_3/2, ⁶H_15/2, and ⁶F_1/2 of Sm³⁺, respectively [11,12,20]. The absorption peaks at high energy (less than 400 nm) for both Dy³⁺ and Sm³⁺ are overlapped with the absorption band of glass.

When a small quantity of Sm₂O₃ is added to Dy³⁺-doped glass, the absorption peaks corresponding to Sm³⁺ are not clearly evident, as shown in Figure 6b. With an increase in Sm₂O₃ concentration in the glass, the more intense absorption peaks of Sm³⁺ are observed and attributed to the transitions from the ground state ⁶H_5/2 to various excited states. These are similar to those seen in a single Sm³⁺-doped glass samples, as depicted in Figure 6b–e.

Figure 7 presents the morphology of the fracture surfaces of the glass samples. None of the samples exhibit crystal-like structures in the glassy phase matrix. The surface is uniform and homogeneous. However, the porosity of the glass is observable at a magnification of 200X. The pores appear to expand and increase in size with the doping and the increased concentration of Sm³⁺, leading to a reduction in density. The introduction of an appropriate concentration of co-doping ions (MD4), however, may lead to a decrease in pore size. The porosity and the pore size can significantly affect light scattering and the optical properties of phosphor in glasses. Previous studies have shown that small pores (less than 2 μm) and low porosity can contribute to high luminescent efficiency [21]. The pore size and porosity can be controlled by modifying the particle size of the glass powder and the concentration of rare-earth elements in the glass matrix [21]. In addition, the mechanical properties of glass samples generally decrease with increasing porosity. As a result, it is essential to conduct a proper sintering process to ensure the full densification.

The excitation and emission spectra of the doped glasses are displayed in Figure 8 and Figure 9. In general, Dy³⁺- and Sm³⁺-doped glasses exhibit efficient excitation at wavelengths of 326–389 nm and 400–405 nm, respectively [21,22]. The excitation spectra of the MD1–MD5 samples at the emission wavelength of 601 nm, which is the optimal peak emission of Sm³⁺, are shown in Figure 8a. It is found that the excitation spectrum of the Dy³⁺-doped sample (MD1) contains a broad and weak band ranging from 300–475 nm, which corresponds to the characteristic f-f transition of Dy³⁺. The maximum intensity in the excitation spectrum appears at 347 nm for the ⁶H_5/2-to-⁶P_7/2 transition. The excitation spectra of the Sm³⁺-doped samples (MD5) shows several characteristic transitions: ⁶H_5/2 to ⁴H_9/2 + ⁴D_7/2 (344 nm), ⁶P_5/2 + ⁴D_3/2 (362 nm), ⁶P_7/2 (375 nm), ⁴F_7/2 + ⁶P_3/2 (401 nm), and ⁴I_13/2 + ⁴I_11/2 + ⁴I_9/2 (450–490 nm) [23,24,25]. The most intense excitation peak is observed at 401 nm, which is, therefore, selected as the excitation wavelength for emission spectra. The MD2–MD4 samples co-doped with Dy³⁺ and Sm³⁺ exhibit excitation peaks similar to those of Sm³⁺. This indicates that the excited states of Dy³⁺ do not significantly result in emission at 601 nm.

Under the 401 nm excitation (Figure 8b), the emission spectrum of MD5 exhibits three prominent emission peaks located at approximately 562, 601, 647, and 711 nm due to the transition from ⁴G_5/2 to ⁶H_5/2, ⁶H_7/2, ⁶H_9/2, and ⁶H_11/2 of Sm³⁺, respectively. The highest emission occurs at 601 nm, which is the characteristic and the most intense emission peak of Sm³⁺ [26]. The doublet peaks associated with each characteristic emission arises from variations in the samarium sites or different crystal fields [27]. For the co-doped glass samples, the emission spectra are similar to that of the MD5 sample. The intensity of the emission peaks increases with the increasing amount of Sm³⁺. A Dy³⁺ emission peak (575 nm) is also observed at low intensity and superimposed on the bands of the Sm³⁺ emission. For the MD1 sample, the emission spectrum in the wavelength range of 500–750 nm shows a board emission peak centered at 575 nm (yellow region) due to the ⁴F_9/2-⁶H_13/2 transition. In general, Dy³⁺ ions exhibit emission bands at 480 nm (blue) and 580 nm (yellow).

To further understand the energy transfer mechanism between Dy³⁺ and Sm³⁺, the excitation and emission spectra are investigated under the 574 nm emission and the 388 nm excitation wavelengths, respectively. Figure 9a shows the excitation spectra at an emission wavelength of 574 nm, which is the peak emission of Dy³⁺. In the MD5 sample, excitation peaks are observed at wavelengths of 344, 363, 375, 404, and 400–490 nm due to energy level transitions from ⁶H_5/2 to ⁴H_9/2 + ⁴D_7/2, ⁶P_7/2 + ⁴D_3/2, ⁶P_7/2, ⁴F_7/2 + ⁶P_3/2, and ⁴I_13/2 + ⁴I_11/2 + ⁴I_9/2 of Sm³⁺, respectively. In the MD1 sample, excitation peaks occur at wavelengths of 328, 346, 369, 388, 427, and 443 nm. These are due to energy level transitions from ⁶H_15/2 to ⁴M_17/2, ⁶P_7/2, ⁶P_5/2, ⁴I_13/2, ⁴G_11/2, and ⁴I_15/2 of Dy³⁺, respectively [23,24,25]. The most intense excitation peak at a wavelength of 388 nm was then chosen to stimulate light emission for the analysis of co-doped samples.

The emission spectra of the single doped sample under ultraviolet light emission (388 nm) exhibits characteristic peaks of doping ions, as shown in Figure 9b. For Dy³⁺, there are two emission peaks located at 483 nm (blue region) and 568 nm (yellow region), corresponding to the magnetic dipole ⁴F_9/2-⁶H_15/2 transition and electronic dipole ⁴F_9/2-⁶H_13/2 transition of Dy³⁺, respectively [28,29,30]. The single Sm³⁺-doped glass exhibits three characteristic red emission peaks at 564, 601, and 647 nm, which corresponds to the transition from ⁴G_5/2 to ⁶H_5/2, ⁶H_7/2, and ⁶H_9/2, respectively. These peaks, however, exhibit low intensity due to the unfavorable excitation at 388 nm for Sm³⁺ emission. In co-doped glasses, it is evident that the peak of Sm³⁺ emission is nearly absent. A very weak peak is observed at approximately 600 nm due to the ⁴G_5/2-⁶H_7/2 transition of Sm³⁺. The emission peak associated with the ⁴F_9/2-⁶H_13/2 transition of Dy³⁺, however, shifts to a longer wavelength (575 nm), and the emission intensity decreases with the addition of Sm³⁺. When the electrons at the ground state ⁶H_15/2 of Dy³⁺ absorb energy at 388 nm, they are excited to a higher energy level. These unstable electrons subsequently drop to the lower energy state ⁴F_9/2 through a nonradiative transition process, followed by the radiative transitions to the states ⁶H_15/2 and ⁶H_13/2, resulting in yellow and blue emissions. Lower wavelength excitation light possesses higher energy, which more effectively stimulates Dy³⁺ emission. However, the emission intensity depends on the concentration of Dy³⁺. The peak shift is likely attribute to an increase in non-bridging oxygens (NBOs) resulting from co-doping.

Based on the current findings, the excitation is most efficient when employing light of a particular wavelength. Excitation at a wavelength with high absorbance should produce multiple excited states, yielding strong fluorescence. The excitation spectra of the co-doped glass sample, however, may differ from those of single-doped glass samples. No evidence on the energy transfer from Dy³⁺ ions to Sm³⁺ ions is obtained in this study for the co-doping of Dy³⁺ and Sm³⁺. Despite the overlap between the emission peak of Dy³⁺ and the excitation peak of Sm³⁺, as shown in Figure 10, the energy transfers from Dy³⁺ to Sm³⁺ may be transferred back to the Dy³⁺, thereby inhibiting emission in Sm³⁺ [31].

Nevertheless, it can be seen that the emission color can be tuned by employing different excitation wavelengths in the co-doped glasses [20]. Figure 11 shows the CIE 1931 chromaticity diagrams corresponding to different excitation wavelengths. It can be clearly seen that the variation in excitation wavelengths significantly affects the chromaticity coordinate values. With the optimized excitation at 388 nm, the emitted light from the Dy³⁺-doped glasses is observed in white region. The chromaticity coordinates of the Sm³⁺-doped glasses, on the other hand, are located in the yellow light region of the CIE 1931 chromaticity diagram with the excitation at 401 nm. The co-doping enhances light emission predominantly in the white and yellowish-orange region for excitations at 388 nm and 401 nm, respectively. The adjustment of emitting color can be accomplished by modifying the excitation wavelength and the doping ratios.

4. Conclusions

This study investigated the application of microwave sintering for the preparation of fluorescent glasses doped with Dy³⁺ and Sm³⁺ in varying proportions. The results indicate that the surface of the prepared glass appeared smooth without the formation of crystalline phases. The study on the effects of doping Dy³⁺ and Sm³⁺ ratios suggests that the introduction of these doping ions increases the porosity and the density of the glasses. Under 401 nm excitation, emission peaks of Sm³⁺ and co-doped glasses occur at 562, 601, 647, and 711 nm, corresponding to the transitions from ⁴G_5/2 to ⁶H_5/2, ⁶H_7/2, ⁶H_9/2, and ⁶H_11/2 of Sm³⁺, respectively. However, under 388 nm excitation, no energy transfer from Dy³⁺ to Sm³⁺ is evident. The emission corresponds to the energy transitions (⁴F_9/2-⁶H_15/2, and ⁴F_9/2-⁶H_13/2) of Dy³⁺, and the intensity decreases with increasing Sm³⁺ concentrations.