1. Introduction
With the development of the economy and the rapid consumption of fossil fuels, increases in carbon emissions have become a concern in terms of climate threats, i.e., the greenhouse effect and global warming [
1,
2]. Many efforts have been made to reduce carbon emissions. For example, President Xi Jinping of China has declared that the carbon emissions in China will reach their peak in 2030 and decrease to net-zero in 2060 [
3,
4]. As for the carbon emissions in buildings, it is known that windows are a key factor in carbon emission reduction and energy savings due to excessive thermal energy transfer through their transparent surfaces.
A variety of glazing technologies have been reported for carbon emission reductions and energy savings. For example, tinted glass [
5,
6,
7], vacuum glazing [
8], low-emissivity (low-E) coated glass [
9,
10], and various transparent thermal shielding coatings are used for solar glazing [
11]. Among these glazing technologies, tinted glass usually contains a higher Fe[II] content in the glass to achieve infrared adsorption properties [
6,
7]. Common tinted glass, especially tinted silicate glass, usually suffers from low visible transmittance, heavy color, and poor heat-shielding performance [
12]. A. K. Mandal and coworkers reported a higher Fe[II] content and colorless phosphate glass fabricated via microwave heating [
6]. However, the hydrolysis of phosphate and the higher cost compared to common silicate glass also limited its further application. Vacuum glazing can reduce heat transfer via low air conduction and convection in the vacuum gap but is limited by the potential risk of self-explosion and significantly higher costs than standard triple-glazed windows [
13,
14]. Low-E coated glass is the most widely used solar glazing technology in energy-saving buildings, which can efficiently reflect far infrared radiation and allow visible light through the window. However, low-E coated glass is also criticized for its poor stability due to oxidation of the Ag layer, light pollution caused by reflection, low near-infrared and solar shielding performance, and need for expense vacuum sputtering equipment [
15,
16,
17,
18].
In contrast, transparent thermal shielding coatings can selectively absorb near-infrared rays in sunlight via the local surface plasmon resonance (LSPR) effect of nanomaterials, such as Au/Ag nanoparticles [
19,
20], ATO/ITO [
21,
22,
23], Cs-doped WO
3 [
24,
25], rare-earth borides [
26,
27,
28], etc. These nanoparticles possess excellent solar-shielding properties and are easy to prepare but also have poor chemical stability, which results in the protective resins usually being indispensable [
29]. Obviously, common polymer resins are unable to bear extreme heat-treated conditions, such as 640 °C for curved tempered glass and 730–750 °C for tempered glass. Poor aging resistance is another issue of thermal shielding coatings due to the degradation of the resins under solar radiation after long-term application. A coated glass that integrates thermal stability and good solar shielding performance, as well as easy production, is thus desired.
It is commonly known that nanomaterials, such as Au/Ag nanoparticles, ATO/ITO, Cs-doped WO
3, and rare-earth borides, etc., are difficult to directly calcine into glass glaze at high temperatures and in air during the glass tempering process. In general, the temperature for glass tempering is ca. 650–750 °C, depending on the thickness of the glass. This is mainly because these materials are easily oxidized at such a temperature and in air [
29,
30]. For example, K. Adachi and coworkers reported an oxidation reaction of Cs-doped WO
3 into WO
3 under heat with or without water, which resulted in the loss of its solar-shielding property [
29]. Obviously, the high temperature for glass tempering will accelerate this oxidation reaction. Fabricating Cs-doped WO
3 nanoparticles with a ZnO or SiO
2 shell may partly improve the thermal stability but also result in an increase in costs and preparation difficulties in the industry [
31,
32]. Rare-earth boride can be oxidized into borate at high temperatures and in air [
30], and the silver layer can also be easily oxidized into black oxidates [
17]. In contrast, ATO/ITO nanoparticles, as an oxidate themselves, have better oxidation resistance. However, ATO and ITO nanoparticles have poor UV adsorption properties, which are desired for solar glazing.
The inhibition of Ostwald ripening is another puzzle to be solved when these nanoparticles, such as Au nanoparticles, are directly calcined in air. This is mainly due to the increase in surface energy and the decrease in melting point with the decrease in particle size [
33,
34]. Doping with lanthanide ions is a good way to inhibit ripening of the nanoparticles. H. Parangusan and coworkers reported that Ce ions could efficiently slow down the motion of the ZnO grain boundary [
35]. P. Diko and coworkers reported that the addition of CeO
2 could also inhibit the ripening of Y
2BaCuO
5 particles [
36].
Herein, we present a novel, simple, and low-cost method to fabricate thermally stable heat-shielding coated glass for solar glazing by directly calcining Ce and Sb co-doped SnO2 nanoparticles on a glass surface in air. The Ce and Sb co-doped SnO2 is synthesized via a coprecipitation method and sintered at 1000 °C for 4 h, and it is further fabricated into nanoparticles via direct grinding with a bead mill, which is the most suitable fabrication method for industrial production. Then, the obtained nano-slurry is mixed with a certain amount of polysilazanes (PSZs) and sprayed on glass, followed by directly calcining in air. PSZs are polymers with backbones consisting of alternating silicon and nitrogen atoms and pendent carbon-containing groups, which are widely used as precursors. When calcined at a high temperature and in air, PSZs can form a transparent “Si-O-Si” layer which inhibits the aggregation of Ce and Sb co-doped SnO2 nanoparticles and also supports the adhesion of nanoparticles to a glass surface. In this method, no complex vacuum coating equipment is required. Meanwhile, the utilization of Ce ions can inhibit the ripening of nanoparticles, refine the grain, and produce coated glass with excellent UV adsorption properties. Moreover, we also fabricated coated glass with a high surface pencil hardness of 9H by further spraying a PSZ coating as the cover layer and calcinating in air, which is the hardest coated solar glazing to our knowledge. As a result, the Ce and Sb co-doped SnO2-coated glass shows excellent heat-shielding performance, with transmittances of 4.7% at 250–380 nm, 59.3% at 380–780 nm, and 9.7% at 780–2500 nm. In addition, the coated glass shows excellent aging resistance and thermal resistance. The coated glass passed a 350 h accelerated aging experiment without any performance degradation and could also endure the extreme conditions of tempered glass and curved tempered glass, thus exhibiting its potential for application in energy-saving buildings and cars.
2. Materials and Methods
2.1. Materials
Hydrated tin chloride (SnCl4·5H2O, AR), antimony chloride (SbCl3, AR), aqueous ammonia (25%–28% in water), cerium nitrate hexahydrate (Ce(NO3)3·6H2O, AR), and hydrochloric acid (HCl, AR) were purchased from Macklin Biochemical Technology Co., Ltd., Shanghai, China. Ethanol (95%), 2-acetoxy-1-methoxypropane (PMA, 99%), and n-butyl acetate (99%) were purchased from Zhonghe Shengtai Chemical Co., Ltd., Tianjin, China. WA-8108 surfactant and polysilazanes (60 wt% in n-butyl acetate, PSZs) were purchased from EASYTO Technology Development Co., Ltd., Tianjin, China. All the chemicals were used as received, without further purification.
2.2. Synthesis of Ce and Sb Co-Doped SnO2 Powder
Ce and Sb co-doped SnO
2 powder was synthesized using a dual-titration co-precipitation method [
37]. In this work, hydrated tin chloride (2.1 g, 6 mmol), antimony chloride (0.163 g, 0.71 mmol), cerium nitrate hexahydrate (0.22 g, 0.51 mmol), ethanol (25 mL), and hydrochloric acid (0.5 mL) were mixed and stirred in a dry beaker for 30 min to form a clear Sn
4 + /Sb
3 + /Ce
3+ solution. Then, the mixed solution and aqueous ammonia (25%–28% in water, 25 mL) were loaded into two separatory funnels. The two solutions were simultaneously added dropwise into a three-necked flask to form precipitates. After the titration process, the obtained mixture was heated at 60 °C for another 2 h, filtered, and washed with ethanol and then water. After being fully dried in a vacuum oven at 120 °C, the obtained powder was placed in crucibles and sintered at 1000 °C for 4 h to form a blue Ce and Sb co-doped SnO
2 powder.
2.3. Fabrication of Ce and Sb Co-Doped SnO2 Nano-Slurry
The obtained Ce and Sb co-doped SnO2 powder (100 g) and PMA (500 g) were ground in a planetary ball mill machine at a speed of 300 rpm for 4 h, followed by mixing with 30 g of WA-8108 surfactant and further grinding into a nano-slurry with a sand mill machine at 1500 rpm for 33 h. The diameter of the zirconium bead was about 0.3 mm.
2.4. Fabrication of Ce and Sb Co-Doped SnO2 Heat-Shielding Coated Glass (CeCG)
The resulting Ce and Sb co-doped SnO2 nano-slurry (10 g) was transferred into a one-necked flask and concentrated into a viscous nano-slurry weighing 2.5 g. Then, the obtained nano-slurry was mixed well with n-butyl acetate (2 g) and polysilazanes (0.5 g). The mixture was sprayed onto a clean, low-iron glass measuring 5 × 5 cm. Then, the coated glass was directly calcinated at 736 °C for 100 s, followed by fast and smooth cooling in air in a glass tempering furnace.
2.5. Fabrication of Two-Layered Coated Glass with High Surface Hardness (CeCG-HH)
CeCG-HH coated glass is a double-coated glass consisting of a nano-slurry mixed with PSZs as the bottom layer and another PSZ coating as the cover layer. For the bottom layer, 2.5 g of the concentrated Ce and Sb co-doped SnO2 mentioned above was mixed with n-butyl acetate (2 g) and PSZs (0.5 g), followed by spraying onto glass. Then, the glass was calcinated at 450 °C for 10 min and cooled down to room temperature slowly in a furnace.
As for the cover layer, PSZs (5 g) were mixed with n-butyl acetate (5 g) and sprayed on the resulting glass, followed by calcination at 736 °C for 100 s and fast cooling in air in a glass tempering furnace.
2.6. Fabrication of CE-1 Coated Glass
CE-1 is the control coated glass. The CE-1 coated glass was fabricated in the absence of PSZs. In a typical experiment, the Ce and Sb co-doped SnO2 nano-slurry (10 g) was concentrated into a viscous nano-slurry weighting 2.5 g. Then, the concentrated nano-slurry was directly mixed with n-butyl acetate (2.5 g), followed by spraying onto the glass surface, calcination at 736 °C for 100 s, and fast cooling in air in a glass tempering furnace.
2.7. Characterization
The crystalline structure of the powders was investigated using X-ray diffraction (XRD) (40 kV, 40 mA) with a graphite-monochromatized CuKα source (λ = 0.154056 nm) and a 2Ɵ scanning angle range of 20° to 80°. An X-ray photoelectron spectroscopy (XPS) analysis was performed using a Thermo ESCALAB 250XI system (Waltham, MA, USA) with an Al Kα source and a 500 μm X-ray spot size. Scanning electron microscopy (SEM) images were obtained and energy-dispersive X-ray spectroscopy (EDX) mapping of the glass coating was performed using a Thermos Scientific Apreo 2S field-emission scanning electron microscope (Waltham, MA, USA). Atomic force microscope (AFM) measurements were performed using a Bruker Dimension Icon microscope (Billerica, MA, USA) in tapping mode. Thermogravimetric analysis (TGA) (Waltham, MA, USA) was performed with a TA Q600 thermogravimetric analyzer under an air atmosphere at a heating rate of 20 °Ϲ/min. The ultraviolet–visible–near-infrared (UV-Vis-NIR) absorption spectra were obtained using a PerkinElmer Lambda 1050 spectrophotometer (Waltham, Massachusetts, USA). The particle size of the Ce and Sb co-doped SnO2 nano-slurry was determined using a Malvin 3000 particle size analyzer (Egham, Surrey, UK) with ethyl alcohol as the solvent.
According to the ISO 9050-2003 international standard [
38], the transmittance for two different wavelengths of incident light is calculated as follows:
where λ is the wavelength of incident light, with subscripts 1 and 2 referring to the two different wavelengths, and
T(λ) refers to the transmittance at the wavelength of λ.
The heat-shielding performance was tested with a Linshang LS301 solar film meter purchased from Linshang Technology Co., Ltd. (Shenzhen, China). The test box was divided into two columns for measuring the coated glass and the control sample. Each column had a 150 W Philips infrared lamp located at the top to emit infrared light with a wavelength ranging from 780 to 5000 nm, a bracket to place the samples, a thermocouple at the bottom to measure the temperature of the box, and two LEDs to display the temperature and heat-shielding rate. The heat-shielding rate was measured by the cut-off rate of the 940 nm infrared light. When the two glass samples were exposed to the infrared lamps at the same time, the infrared light passing through the glass sample could heat the thermocouples at the bottom of the box. Therefore, the difference in the box temperature could be used to characterize the heat-shielding performance of the glass.
An accelerated aging test was conducted in an aging test chamber with a Xenon lamp as the source, purchased from Jingyu Environment Test Equipment Co., Ltd. (Dongguan, Guangdong, China). The aging experiment was performed at an irradiation intensity of 1000 W/m2 and at 50% humidity. The transmittance properties of the coated glass at 250–380 nm, 380–780 nm, and 780–2500 nm were measured. The normalized transmittances at these bands were calculated using the formula Xt/X0 ×100%, where X is the transmittance of the coated glass, with the subscripts 0 and t referring to the initial time and the aging time.
3. Results and Discussion
Ce and Sb co-doped SnO
2 was synthesized via a dual-titration co-precipitation method. We also synthesized Sb-doped SnO
2 via common precipitation as a control sample. The X-ray diffraction (XRD) patterns of the Ce and Sb co-doped SnO
2 and 10% Sb-doped SnO
2 powders were characterized (
Figure 1). The results show that both the Ce and Sb co-doped SnO
2 powder and the Sb-doped SnO
2 powder had the standard tetragonal crystalline structure of SnO
2 (JCPDS card no. 99-0024). It can also be observed that the peaks slightly shift at the (110), (101), and (211) planes (shown in
Figure 1a′). For example, the (110) plane shifts from 26.66° in the Sb-doped SnO
2 to 26.54° in the Ce and Sb co-doped SnO
2, while the (101) plane shifts from 33.93° in the Sb-doped SnO
2 to 33.79° in the Ce and Sb co-doped SnO
2. The results indicate the successful synthesis of Ce and Sb co-doped SnO
2, and also the lattice distortion induced by the doped Ce ions. We can also see additional peaks at 2theta = 27° and 47°, which can be attributed to the structure of CeO
2 (JCPDS card no. 43-1022), indicating the existence of doped Ce ions in the form of oxides. This can be attributed to excess doped Ce ions and their low solubility limit in SnO
2 grains [
39]. The results also imply that the segregation and enrichment of Ce ions at the SnO
2 grain boundary and surface may be beneficial for thermal stability at high temperatures by inhibiting the coalescence and growth of the SnO
2 grains.
We also calculated the average crystal size of the two samples from the (110) plane via Scherrer’s formula [
40]. We see that the Ce and Sb co-doped SnO
2 had a smaller average crystal size of ca. 42.7 nm compared to that of the Sb-doped SnO
2, which was 84.9 nm. It can be attributed to the increase in lattice distortion and inner stress induced by the co-doped Ce ions inhibiting the grain growth of the SnO
2 [
41].
The Ce and Sb co-doped SnO
2 nano-slurry was directly fabricated using a mechanical grinding method. Compared with other methods for fabricating nanomaterials, such as hydrothermal or solvothermal synthesis methods, the mechanical grinding method is more suitable for industrial production.
Figure 2 shows the grinding curve of the particle size vs. grinding time. It is shown that the grinding process becomes increasingly time-consuming with the decrease in particle size, and the particle size has a linear relationship with the logarithm of grinding time. It should be noted that the change in slope of the grinding curve is an important parameter for real-time evaluation of the grinding efficiency in practical production processes. For instance, the abrasion of the zirconium bead, the consumption of surfactants, or potential equipment failure, etc., can lead to a decrease in the grinding efficiency and a change in the slope. Therefore, no obvious change in the slope of the grinding curve means a smooth and stable grinding process.
After mixing with PSZs and other solvents, the mixture was sprayed onto glass and calcinated in air to obtain Ce and Sb co-doped SnO
2 heat-shielding (CeCG) coated glass. The chemical structure was examined using X-ray photoelectron spectroscopy (XPS), with the spectrum shown in
Figure 3. In the Ce3d XPS spectrum of the CeCG coated glass (
Figure 3b), Ce ions with two valences were observed. The peaks of Ce
3 + 3d
5/2 and Ce
3 + 3d
3/2 are shown at binding energies of 885.1 eV and 904.4 eV, respectively. The peaks of Ce
4 + 3d
5/2 and Ce
4 + 3d
3/2 are indicated at 900.6 eV and 916 eV, respectively. We also calculated the ratio of Ce
3+ ions according to the method reported in the literature [
42]. The ratio of Ce
3+ ions was calculated as 73.4%, while the ratio of Ce
4+ was ca. 26.6%. The absorption band centered at ca. 303 nm for Ce
+3 ions is due to the
2F
5/2 -> 5d transition of the Ce ions, while the wavelength ranges from 200 nm to 270 nm for Ce
4+ ions [
43]. It is known that no UV irradiation at a wavelength lower than 280 nm exists in the AM 1.5 standard solar spectrum due to the absorption of this UV irradiation by the ozonosphere [
44]. Therefore, the high content of Ce
3+ ions will benefit the UV and heat shielding performance of the CeCG coated glass.
The bonding configurations of O, Sb, Sn, and Si atoms were also characterized. The peaks of Sb-O bonds are indicated at 530.6 and 540.1 eV, while the peaks of metallic Sb are indicated at 528.4 and 537.5 eV [
45]. As shown in
Figure 3c, the peaks of Sb3d
5/2 and Sb3d
3/2 are shown at binding energies of 530.6 eV and 540.5 eV, respectively. The results indicate that the Sb element existed in the form of Sb-O bonds. It is also noted that no Sb
2O
3-related impurity peaks are observed in
Figure 1a, which suggests that the Sb element was fully doped into the lattices of SnO
2 instead of forming antimony oxides or metallic Sb, which also implies the successful synthesis of Ce and Sb co-doped SnO
2. The peaks at 531.4 eV and 532.4 eV are assigned to SnO
2 and SiO
x formed by the PSZs, respectively. In the Sn3d XPS spectrum (
Figure 3d), the peaks of Sn3d
5/2 and Sn3d
3/2 are indicated at 487.1 eV and 495.5 eV. In the Si2p XPS spectrum (
Figure 3e), the peak of Si2p at 102.6 eV also clearly demonstrates the formation of SiO
x.
The thermal stability of the CeCG coated glass was examined using thermogravimetric analysis under an air atmosphere (TGA) (
Figure S1). No obvious decomposition was observed below 800 °C, indicating excellent thermal stability and confirming the suitability of the CeCG coated glass for tempered glass processing at 730 °C for 100 s, as well as for curved tempered glass that is typically processed at 640 °C for 10 min.
In order to explore the formation mechanism of the CeCG coated glass, we fabricated a control experiment sample named CE-1, which had a monolayer Ce and Sb co-doped SnO
2 coating in the absence of PSZs. We also fabricated a double-glazed coated glass with the CeCG coated glass as the bottom layer, followed by spraying PSZs on the surface and calcination in air, named CeCG-HH. The morphologies of the CeCG, CE-1, and CeCG-HH samples were revealed using scanning electron microscopy (SEM) and an atomic force microscope (AFM) (
Figure 4). In
Figure 4a–c, we see that all these samples are fully coated with massive irregular spherical particles, with sizes of ca. 37.2 ± 9.9 nm for CeCG, 34.1 ± 4.3 nm for CE-1, and 38.6 ± 8.8 nm for CeCG-HH, which are close to the D
50 particle size of the nano-slurry of 51.8 nm. This indicates that the nanoparticles of the Ce and Sb co-doped SnO
2 were successfully calcinated on the glass surface rather than ripening or agglomerating. As for the CE-1 coated glass, it is indicated that the Ce and Sb co-doped SnO
2 nanoparticles also had excellent ripening resistance, even in the absence of the PSZs, which may be attributed to the inhibition of Ce ions [
39]. We also observed some cracks on the surface of the CE-1 sample, which may have resulted from the pyrolysis of organic components, such as surfactants (marked with red arrow in
Figure 4b). In comparison, these cracks were hardly observed on the CeCG (
Figure 4a) and CeCG-HH samples (
Figure 4c). A possible explanation could be that the amorphous silicon oxide formed by the decomposition of PSZs could fill the cracks and gaps between the nanoparticles and the glass surface, as described in the schemes of
Figure 4g,h, which may support the adhesion of nanoparticles on the glass surface and the wear-resistance of the coated glass.
We further calculated the surface roughness via AFM images (
Figure 4d–f). The CeCG-HH sample had the smoothest surface, with an average roughness (Ra) of 1.33 nm, followed by CeCG with an Ra of 4.81 nm and the CE-1 sample with an Ra of 5.38 nm. A possible explanation could be that the formed amorphous “Si-O-Si” structure filled the cracks and defects between the nanoparticles and even covered the nanoparticles themselves in the CeCG-HH sample, as described in the scheme of
Figure 4i. Interestingly, we compared the topography image of CeCG-HH (
Figure 4f) with its phase image (
Figure S2) at the same location. In the area marked with a blue dashed square, we observed no obvious surface morphology in the topography image, but a phase difference in the phase image, indicating that the nanoparticles were shallowly embedded under the amorphous silicon oxide.
It is known that silver particles are deposited, nucleated, and grown on the surface of the glass via a bottom-up magnetron sputtering method for common low-E coated glass [
9,
10]. As a result, the surface chemical environment of the glass has great effects on the final performance of low-E coated glass. For example, in some factories of Saint-Gobain, it is necessary to distinguish the tin surface of the glass before magnetron sputtering, which is the undersurface of the glass suspended in the tin bath during the float glass production process. In contrast to common low-E glass, fabricating CeCG coated glass is a top-down approach. The nanoparticles are fabricated in advance, and their particle sizes are largely independent of other industrial processes. Therefore, the surface chemical environment of the glass has less impact on the final performance, which simplifies and improves the repeatability of the process.
The element content and distribution of the CeCG, CE-1, and CeCG-HH samples were examined via EDX mapping (
Figure S3 and
Table S1). The results show a uniform distribution of Ce, Sb, and Sn. No obvious Ce and Sb aggregation was observed, indicating the homogenous distribution of the Ce and Sb co-doped SnO
2 nanoparticles on the glass. The element molar ratio of Ce:Sb:Sn was calculated as ca. 0.7:1:8, which coincides with the molar ratio of the reactants. It was also found that the Si molar content largely increased, ranging from 0% in CE-1 to 8.6% in CeCG and 21.1% in CeCG-HH, indicating that the Ce and Sb co-doped SnO
2 nanoparticles were coated and covered by the “Si-O-Si” structures formed by the PSZs.
The CeCG coated glass can be used for energy-saving buildings and cars. The transmittance of the CeCG glass between 250 nm and 2500 nm was examined (
Figure 5a). We found that the CeCG coated glass had transmittances of 4.7% at 250–380 nm, 59.3% at 380–780 nm, and 9.7% at 780–2500 nm. When compared with the uncoated glass, the heat-shielding performance of the CeCG coated glass was measured by exposing the two samples to the same infrared radiation for the same duration (
Figure 5b,c). The results showed that the temperature on the back of the CeCG coated glass increased from 30.5 °C to 33 °C, which is lower than that of the uncoated glass of 43.7 °C (
Figure 5b). Therefore, the results demonstrate that the CeCG coated glass has an excellent thermal shielding performance.
We also compared the heat shielding performance of the CeCG coated glass with a commercial double-glazed low-E coated glass by calculating the light transmittance (LT) and the transmission of total solar energy (TTS) according to the ISO 9050-2003 international standard [
38]. The commercial double-glazed low-E coated glass was obtained from Saint-Gobain with the product number SGG COOL- LITE
® SKN 154II [
46]. The CeCG coated glass in this work had an LT of 73.1 and a TTS of 50.8, which are close to the values of the commercial double-glazed low-E glass with an LT of 73.6 and a TTS of 50.5. It indicates that the CeCG coated glass offers a similar thermal shielding performance to commercial coated glass, despite its simpler production process.
The surface hardness of the CeCG and CeCG-HH coated glass samples was examined according to the ISO 15184-1998 international standard [
47]. We found that the CeCG sample had a surface pencil hardness of less than 6B, which is similar to that of double-glazed low-E glass. However, as for the CeCG-HH sample, the results showed a high pencil hardness of 9H (
Figure 6), which may be attributed to its smooth surface with fewer defects (
Figure 4f) and the stiff glass-like composite containing nanoparticles and an amorphous “Si-O-Si” molecular structure. As we know, the CeCG-HH sample also has the hardest surface compared to other coated glass for solar glazing. The CeCG-HH coated glass can thus be installed on the outside surface of the window glazing, which is beneficial for heat dissipation through thermal convection on the outside surface of the glass.
The stability of CeCG and CeCG-HH glass in the long term is also of importance in practical applications. Hence, an accelerated aging test was conducted at an irradiation intensity of 1000 W/m
2 with a Xenon lamp source and at 50% humidity to examine the stability of these coated glass samples (
Figure 7). The results show that both the CeCG and CeCG-HH coated glass samples had no obvious degradation of the heat-shielding performance over 350 h, indicating the excellent long-term environmental stability of the CeCG and CeCG-HH coated glass samples.
The economic analysis of the CeCG and CeCG-HH coated glass samples considered the fixed costs of the equipment and the unit costs of the coated glass per m
2, and the results were used to determine their industrial application potential (
Table 1). We learned from Saint-Gobain that the production line for double-glazed curved tempered low-E glass costs ca. USD 9.5–11 million, due to the massive and expensive vacuum deposition equipment. In comparison, the production line for CeCG would require USD 0.04–0.07 million for the spray equipment, 0.1–0.12 million for the tempering furnace, and 0.1–0.15 million for transmission and other equipment, including the loading and unloading machines, the edging machine, the cleaning machine, etc. The fixed cost for the CeCG production line is between 2% and 3% of that of low-E glass production lines. In addition to the fixed costs, we also calculated the unit cost per m
2 for CeCG and CeCG-HH coated glass based on the raw material costs and water and electricity expenses. We learned that the unit cost of the double-glazed low-E glass per m
2 is equal to the cost of blank glass per m
2 plus USD 2.74, which is higher than the price for CeCG of ca. USD 1.68 and that for CeCG-HH of ca. USD 2.15. Therefore, the CeCG and CeCG-HH coated glasses offer excellent heat-shielding performance as well as a very low cost.
In addition, we established a production line for the heat-shielding coated glass in Foshan, Guangdong Province (
Figure 8a). It is designed to produce ca. 600,000 m
2 of coated glass per year, with a maximum panel size of 2440 × 3660 mm (
Figure 8b). In this production line, uncoated glass is loaded on the production line (
Figure 8c), and the edges are ground to reduce defects on the glass (
Figure 8d). After edging, the glass is cleaned using water, an air knife, and air plasma treatment. Then, the mixed nano-slurry is sprayed on the cleaned glass (
Figure 8e), followed by calcination in a tempering furnace (
Figure 8f) and rapid cooling by air (
Figure 8g). After unloading the glass from the production line (
Figure 8h), coated glass can finally be obtained (
Figure 8i). The produced heat-shielding coated glass will be utilized in energy-saving buildings and cars.