- Email: [email protected]
Effect of sealing temperature on the sealing edge performance of vacuum glazing
Vacuum 116 (2015) 7e12
Contents lists available at ScienceDirect
Vacuum journal homepage: www.elsevier.com/locate/vacuum
Effect of sealing temperature on the sealing edge performance of vacuum glazing Hong Miao*, Xiang Shan, Jianfeng Zhang, Juan Sun, Hongjun Wang College of Mechanical Engineering, Yangzhou University, No. 196 Huayang West Road, Yangzhou 225127, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 15 December 2014 Received in revised form 30 January 2015 Accepted 7 February 2015 Available online 20 February 2015
Because inadequate control of the temperature of vacuum glazing may cause changes in form, stress, microstructure, or performance and thereby affect its lifetime, vacuum welding has been adopted to seal the vacuum glazing from the side, and experiments using different sealing temperatures have been executed. The impact of different sealing temperatures on the microstructure of the sealing layer was analyzed to evaluate the combination of material science features on the interface and to reveal the influence of the sealing temperature on the hardness and residual stress of the sealing layer. The results show that a sufficiently high sealing temperature will drive the sealing layer to transform from a hybrid structure to the liquid phase, accelerate the element migration, eliminate pores, stabilize and compact the structure and improve the sealing performance of the vacuum glazing. As the sealing temperature increases, the residual stress and hardness substantially increase. However, when the sealing temperature reaches 460 C, the residual stress and hardness begin to plateau, and when the sealing temperature reaches 470 C, no further change can be detected. Therefore, a sufficiently high sealing temperature is beneficial to the bonding of glass and sealing solder and can promote the sealing performance of the vacuum glazing. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Vacuum glazing Sealing edge Microstructure Residual stress Hardness
1. Introduction Vacuum glazing is a new energy-saving transparent construction material with excellent thermal insulation, ageing resistance, visible light transmission and infrared light reflection properties. These properties result in good development potential, and have made vacuum glazing a research focus of glass processing technology [1e3]. In 1989, the first successful vacuum glazing sample was fabricated by Collins and Robinson at Sydney University [4]. Vacuum glazing is produced by coating all around two glass panes with sealing solder, and then put them into the vacuum heating furnace for sealing. To avoid deformation from the high temperature, the sealing should be performed below the softening temperature of the glass, which is generally low (for example, the softening temperature of common soda-lime glass is lower than 600 C) [5]. The sealing of vacuum glazing is a low-temperature process, and the vacuum glazing sealed in a heating furnace
* Corresponding author. Tel.: þ86 13852706878. E-mail addresses: [email protected] (H. Miao), [email protected] (X. Shan), [email protected] (J. Zhang), [email protected] (J. Sun), [email protected] (H. Wang). http://dx.doi.org/10.1016/j.vacuum.2015.02.009 0042-207X/© 2015 Elsevier Ltd. All rights reserved.
should maintain its shrinkage during the cooling process [6]. This shrinkage could generate stress on the sealing edge of the vacuum glazing, and if the stress exceeds the allowable stress of the glass or sealing layer, the glass, the sealing layer or the sealing interface could break. Therefore, the sealing solder must have a thermal expansion coefficient that matches that of the glass. Different sealing temperatures may change the growth form and the grain size of the solder, and the element diffusion caused by the element affinity on the bonding interface between the solder and the glass substrate may further change the sealing form, stress, microstructure and performance of the glass and thereby lifetime [7,8]. Thus vacuum glazing has a very great need for high-performance sealing solder and sealing technology. Due to the difficulty of the manufacturing vacuum glazing and the few people who have mastered this technology, research has mainly focuses on the bearing stress, the residual stress of the sealing edge and heat transfer mechanisms of the vacuum glazing [9e12]. Manufacturing theory and design theory remain to be established, as the theoretical research in this field is still at an exploratory stage and achievements are few. Based on the vacuum welding method of side sealing vacuum glazing proposed at Yangzhou University, this paper takes full advantage of the excellent performance of PbOeTiO2eSiO2eRxOy
8
H. Miao et al. / Vacuum 116 (2015) 7e12
Table 1 Composition and various performance indexes of the sealing solder. Composition/wt%
PbO
TiO2
SiO2
CuO
75
15
10
<2.5
Sintering temperature/ C 420e520
Properties
Coefficient of thermal expansion/ C 90e94 107
system welding solder. It discusses the manufacture of the solder and the sealing process [13,14]. Then it describes a series of trails on vacuum glazing at different sealing temperatures, analyses the impact of the sealing temperature on the microstructure of the sealing layer, studies the diffusion of the affinity element on the sealing interface, evaluates the effectiveness of the combination of material science features on the interface and reveals the influence of the sealing temperature on the hardness and residual stress of the sealing layer. This research not only provides significant data related to the application and verification of vacuum glazing, the prediction of its lifetime and the design and improvement of its structure, but also lays the foundation for the manufacturing and commercialization of the vacuum glazing.
2. Testing material and method 2.1. Testing material The material used in the experiment is two 800 600 mm sodalime glasses, each 4 mm thick with an elasticity of 72.45 MPa, a Poisson's ratio 0.22, a tensile strength of 40 MPa and a compressive strength of 880 MPa (both the tensile strength and the compressive strength were selected at their lower limits). The spacing distance between each pillar is 40 mm. The diameter of each pillar is 0.6 mm, with a height of 0.2 mm, an elasticity modulus of 55 GPa, Poisson's ratio of 0.25 and a density of 2500 kg/m3. Mix 99% pure and 30-mm-diameter PbO and TiO2 powders in a proportion of 70:15, and then add 10%, <2.5% and <2.5% SiO2, CuO and Fe2O3 respectively (purity 99.8% and average size 50 nm). Place the powder into a stainless steel ball grinder, add some dispersing agent, and mix the powder by the wet milling method, with a powder quality ratio of 10:1, a powder milling time of 0.5 h and a revolving speed of 250 r/min. Then heat the powder in a vacuum drying oven to obtain the system welding solder PbOeTiO2eSiO2eRxOy. The ingredients and properties of the final mix powder are as shown in Table 1. The grain size of the welding
Fig. 1. SEM photo of the sealing solder.
1
Fe2O3 <2.5
Melting temperature/ C min1 460e480
solder for the experiment is within 1e25 mm, with an average size of 12.15 mm, as shown in Fig. 1. 2.2. Testing The vacuum glazing is performed by placing one clean glass pane over another clean glass pane with pillar array, and then coating around the two glass panes with welding solder. The panes are than heated them in a 420e520 C vacuum environment for 15e60 min. The two glass panes are sealed with welding solder as the water and organic mater are removed from the glass to form an integral part of the structure [6e8]. The structure of the Low-E vacuum glazing is as shown in Fig. 2. Using a diamond wire saw cutting machine at the Nanjing University of Aeronautics and Astronautics, we cut a 20 20 8.2 mm sample from the sealing edge of the sealed 800 600 8.2 mm vacuum glazing, removed impurities such as oil contamination from the sample, and conducted ultrasonic cleaning in acetone for 15 min. The samples was then repeatedly washed with alcohol and dried and the microstructure of the section and the phase structure of the sealing layer were analyzed using a Hitachi S-4800 thermal field emission scanning electron microscope (SEM) and a D8-Advance poly-crystal X-ray diffraction instrument, Bruker AXS, German [14]. The X-ray testing conditions are a Cu target, a working voltage of 60 kV, a working current of 450 mA and a scanned area of 10 e80 . A hardness test was conducted by using an MHV-1000 digital Vickers micro-hardness test to determine the hardness of the sealing layer of the vacuum glazing following different sealing temperatures. The hardness measurement conditions were a load of 200 g load and a duration of 15 s. The average value was taken after measuring several points, and the hardness testing position of the sealing layer is as shown in Fig. 3. The residual stress of the sealing layer was tested using an electrical measuring method, adopting the KHCR-5-120-G16 weldable high-temperature strain gage produced by the Japan KYOWA Company [3]. The high-temperature strain gages were Pre-buried in the welding solder between the two glasses at a distance of 100 mm. Then the two wiring ends of the strain gage were set aside, and they were placed in the vacuum sealing furnace for sealing. The cooled and educed strain gage was connected to the deformation instrument to test the residual stress of each point, and use the average value was used.
Fig. 2. Schematic diagram of vacuum glazing with a fused edge seal.
H. Miao et al. / Vacuum 116 (2015) 7e12
9
Fig. 3. Hardness testing position.
3. Results and discussion 3.1. Sealing structure Fig. 4 shows the SEM diagram of the sealing section of vacuum glazing under different sealing temperatures. When the sealing temperature is 440 C, the connection between the sealing layer and the glass is contact and clear, with an obvious boundary, as a bulk hybrid. There are some pores and cracks between the bulks and some solder in the sealing layer that has not yet melted because the connection temperature had just reached the softening temperature of the solder. The viscosity of the fusion solder was still relatively large, which is unfavourable to the overflow of the residual gas between the particles that had not yet been melted. Therefore, a hybrid structure with independent pores and cracks is generated. During the melting, some elements will migrate to the interface and form a reactive wetting layer on the interface, but the element migration will be slow due to the low sealing temperature and the impedance of the solid grain. At this time, the reactive wetting layer is not that obvious but has a compact structure, as shown in Fig. 4(a). From Fig. 4 we can see that the liquid phase group mainly contains Pb, Ti and O, while the grain group mainly contains Cu, Si, Pb, Ti and O, so the addition of CuO in the welding material makes the sealing temperature of the solder higher. At a sealing temperature of 450 C, the bulks in the sealing layer is reduced and the pores and cracks in the hybrid structure are also reduced. As the solder viscosity is decreased, many pores will overflow to the outside from the melting layer due to the internalexternal pressure difference, while some cannot overflow. The increase in the sealing temperature will accelerate the element migration, enabling an obvious reactive wetting layer to continuously distribute on the interface as a lamellar with a certain thickness, as shown in Fig. 4(b). When the sealing temperature reaches 460 C, the bonding between the sealing layer and the glass is even tighter, and the structure stably compacts (see Fig. 5). The sealing performance is good, there is no obvious bulk, and it mainly exists in the form of a glass liquid. Through XRD analysis, the main crystal phase generated is PbTiO3 (as shown in Fig. 6). Most of the pores disappear due to substantial micro-pore. convergence as the sealing temperature increases and most bubbles converge under the internal-external pressure differential overflow from the sealing layer, as shown in Fig. 4(c). The relatively high sealing temperature may cause some chemical and physical changes in the glass and sealing solder, and the element dispersion and migration on the sealing interface may affect the sealing performance of the vacuum glazing to some degree. During the element migration, the affinity element undergo chemical and physical reactions on the sealing interface and generates new matter, and the new chemical bond of the new matter can strengthen the sealing between glasses and solder and improve the bonding performance between the glass and the sealing layer.
Fig. 4. SEM diagram of sealing section under different sealing temperatures. (a) Sealing temperature as 440 C. (b) Sealing temperature as 450 C. (c) Sealing temperature as 460 C.
However, excessive element migration may generate too much new matter, which may affect the performance of the solder and glass themselves. Fig. 6 shows the energy spectrum on the sealing interface. The sealing solder mainly contains Pb, Ti, Al, Si and O. The glass used is common soda-lime glass, which mainly contains Si, Na, Ca and O. Among these elements, Na and Ca are the main elements unique to the glass material, and Pb, Ti and Al are unique to the sealing solder. As Au is left in the sample due to the metal
10
H. Miao et al. / Vacuum 116 (2015) 7e12
sealing solder at a certain depth of the glass basis with a content of 15.3%. Therefore, Pb migration on the interface will cause some reaction on the interface between the glass and the sealing solder, which is favourable for the bonding between the glass and the sealing solder, because that reactive wetting layer can help improve the bonding performance. As the sealing temperature increases, the reactive wetting layer becomes more obvious and the bonding performance between the reactive wetting layer and the glass improves. Hence the use of a higher use of sealing temperature can improve the sealing performance of vacuum glazing. 3.2. Hardness of the sealing layer
Fig. 5. Energy spectrum of hybrid structure in sealing layer. (a) Bulks. (b) Liquid phase.
spraying before observation, its impact is omitted. From the figure we can see that the elements between the reactive wetting layer and intermediate melting layer mainly include Pb, Si and Al, and the main ingredients between the two are PbO, SiO2 and Al2O3. Compared to the intermediate melting layer, the content of Pb and Si in the reactive wetting layer is substantially improved, reaching 36.61% and 11.9%,respectively, while the change in Al is subtle, which means that the migration of Pb is the strongest, and that of Si is next. Al has high thermal stability and the contents of PbO and SiO2 are much closer to the content of the basis glass. All of these features make the interface more compact. The absence of Na and Ca which are unique to the glass material in the reactive wetting layer indicates that the glass basis elements undergo no obvious dispersal or migration during sealing, while Pb appears in the
Fig. 7 shows the hardness of the sealing layer at different sealing temperatures. As show in the figure, when the sealing temperature is 440 C, the hardness of the sealing layer is 279HV, and when the temperature is 460 C, the hardness also significantly increases to 298HV. This is because the sealing layer mainly exists in the form of a hybrid when the temperature is lower, and the higher sealing temperature induces the transformation from the hybrid to the liquid phase, resulting in a much higher hardness than that of the bulk solid. If the sealing temperature increases to 470 C, the hardness of the sealing layer is 300HV, which is a slight increase. This is because the sealing layer at this time exists in the form of a liquid phase with a stable structure, so a further increase in the temperature would not substantially change the performance or hardness of the sealing layer (see Fig. 8). 3.3. Residual stress of the sealing edge The residual stress of the sealing layer of the vacuum glazing is mainly the residual tension caused by the thermal residual stress from the following three causes: (1) As the high temperature at the sealing edge cools, the heat dissipation speed of the surface between the sealing layers is different from that of the glass basis, leading to a temperature gradient inside the sealing layer. Moreover, the heat conductivity coefficient of both the sealing solder and the glass basis is small and the material is a poor conductor of heat, so the temperature gradient is even larger. During the sealing, the thermal expansion of the solder and the basis glass are different
Fig. 6. XRD spectrum of sealing layer at 460 C.
H. Miao et al. / Vacuum 116 (2015) 7e12
11
Fig. 9. Residual stress of sealing layer under different sealing temperatures.
Fig. 7. Energy spectrum of sealing layer of vacuum glazing. (a) Intermediate Melting Layer. (b) Reactive Wetting Layer (c) Glass Substrate.
and must have different yield behaviours, causing the parts in the sealing layer to have different plastic deformation behaviours and further causing thermal gradient residual stress in the sealing layer. (2) Low-melting-point sealing solder is a compound material, with different thermal expansion coefficients, and the thermal expansion coefficient between the basis glass and the sealing solder is also different. These coefficient differences make the deformation of materials different when cooling thus leading to thermal residual
Fig. 8. Hardness of the sealing layer at different sealing temperatures.
stress mismatch in the sealing layer. The research shows that even if the thermal expansion coefficient difference is smaller, it may still cause a large thermal residual stress mismatch [4e10]. (3) When the sealing layer of the vacuum glazing cools down, the transformation of the sealing layer from a hybrid structure to the liquid phase will cause a volume change. Whether the phase change part is bigger or smaller, stress will be produced between the phase change part and the unchanged part, leading to a phase change residual stress in the sealing layer. To study the impacts of different sealing temperatures on the residual stress of the vacuum glazing sealing layer, the samples sealed under different temperatures are cooled to room temperature 20 C in 24 h. The samples at four different sealing temperatures in the experiment were compared, as shown in Fig. 9. As the figure shows, when the sealing temperature is 440 C, the residual stress is 40 MPa, and when the temperature increases to 460 C, the residual stress also increases to 55 MPa. The residual stress was significantly higher at 460 C. When the temperature increases to 470 C, the residual stress increases to 56 Mpa, which is a negligible change. The different cooling rates for the vacuum glazing sealing layers of different residual stresses, according to a sealing temperature of 460 C forcibly cooled to a room temperature of 20 C, extracted six different cooling speeds that are shown in Fig. 10. The figure shows that with the acceleration of the cooling rate, the residual stress of the sealing layer of shows a rising trend. At a cooling rate of 4 h, the residual stress is 69 MPa, but the sealing layer has many
Fig. 10. Residual stress of vacuum plate glass sealing layer under different cooling speed.
12
H. Miao et al. / Vacuum 116 (2015) 7e12
Acknowledgement The authors acknowledge the support from the National Youth Science Fund Project (No. 51302242) of China, the National Natural Science Foundation(No. 51172199, No. 51372216) of China, the Natural Science Foundation (No. BK2012530) of Jiangsu Province, and the Construction and Science and Technology Plan Project (No. 2012BAD32B11) of Yangzhou. References
Fig. 11. Partial cracks in vacuum glazing sealing layer.
microcracks (Fig. 11), so the cooling rate is not too fast. To prevent the sealing layer from being too large due to the residual stress microcrack formation, resulting in the failure of the vacuum chamber, the cooling rate is generally maintained at more than 12 h. 4. Conclusion Solder sealing performance is critical to the performance of the vacuum glazing. The variations of the microstructure, hardness and residual stress at different temperatures were analyzed. The sealing solder is effective when it is dense and well-connected with the glass substrate, and then it has excellent anti-ageing properties (the first generation of vacuum glazing samples we produced in 2005 is valid proof of this). For PbOeTiO2eSiO2eRxOy sealing solder, we suggest maximizing the sealing temperature and maintaining an adequate cooling time so as not to affect the performance of the vacuum glazing. At the same time, this approach will aid in identifying the solder sealing performance and improving the manufacturing processes of vacuum glazing.
[1] Synowicki RA, Johs BD, Martin AC. Optical properties of soda-lime float glass from spectroscopic ellipsometry. Thin Solid Films 2011;519(9):2907e13. [2] Luo LH, Huang ZZ, Wu YF, Cheng D, Sun LL, Cheng L, et al. Sealing properties of BaOeCaOeAl2O3eB2O3eSiO2 glass ceramics to large size planar IT-SOFC. J Chin Ceram Soc 2014;42(1):101e7. ar FO, Revel B. Synergic effect of V2O5 [3] Sharma K, Kothiyal GP, Montagne L, Me and P2O5 on the sealing properties of barium-strontium-alumino-silicate glass/glass-ceramics. Int J Hydrogen Energy 2013;38:15542e52. [4] Robinson SJ, Collins RE. Evacuated windows-theory and practice. In: Proceeding of ISES solar world congress; 1989. Kobe, Japan: Oxford: Pergamon; 1989. [5] Griffiths PW, di Leo M, Cartwright P, Eames PC, Yianoulis P, Leftheriotis G, et al. Fabrication of evacuated glazing at low temperature. Sol Energy 1998;63:243e9. [6] Miao H, Xi XB, Zhang RH, Zuo DW, Wang HL. Numerical calculation and experimental research on wind load of vacuum plate glass. Acta Energiae Solaris Sin 2013;34(11):1888e94. [7] Fang Y, Hyde TJ, Ayra F, Hewitt N, Eames PC, Norton B, et al. Indium alloysealed vacuum glazing development and context. Renew Sustain Energy Rev 2014;37:480e501. [8] Ng N, Collins RE, So L. Thermal conductance measurement on vacuum glazing. Int J Heat Mass Transf 2006;49(25e26):4877e85. [9] Yu Li A, Du JP, Han MF, Peng SP. Structure and sealing properties of Bi2O3eBaOeSiO2eRxOy glass. J Univ Sci Technol Beijing 2011;33(12):1529e33. [10] Zhang LY, Qin ML, Qu XH, Lu YJ, Zhang XY. Study on active brazing of AlN ceramics and MoeNieCu alloy. Rare Metal Mater Eng 2009;38(12):2159e62. [11] Zhang RH, Ma CW, Miao H, Zhang Y, Kong DJ. Theoretical analysis and experimental study on heat transfer in vacuum plate glass. Trans Chin Soc Agric Mach 2006;12:134e8. [12] Miao H, Zhang RH, Zuo DW, Wang HJ, Wang HY. Finite-element analysis of static and modal response of He vacuum plate glass. J Yangzhou Univ 2012;15(1):69e72. [13] Luo XL, Yu XL, Wang YL, Gu L, Li F, Li QL, et al. Sealing technology of the SiO2eBaO system glass and metal. J Chin Ceram Soc 2013;41(6):858e62. [14] Cheng Y, Xiao HN, Guo WM. Effects of GeO2 on structure and properties of PbOeB2O3eZnO. J Inorg Mater 2006;21(3):533e8.
Contents lists available at ScienceDirect
Vacuum journal homepage: www.elsevier.com/locate/vacuum
Effect of sealing temperature on the sealing edge performance of vacuum glazing Hong Miao*, Xiang Shan, Jianfeng Zhang, Juan Sun, Hongjun Wang College of Mechanical Engineering, Yangzhou University, No. 196 Huayang West Road, Yangzhou 225127, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 15 December 2014 Received in revised form 30 January 2015 Accepted 7 February 2015 Available online 20 February 2015
Because inadequate control of the temperature of vacuum glazing may cause changes in form, stress, microstructure, or performance and thereby affect its lifetime, vacuum welding has been adopted to seal the vacuum glazing from the side, and experiments using different sealing temperatures have been executed. The impact of different sealing temperatures on the microstructure of the sealing layer was analyzed to evaluate the combination of material science features on the interface and to reveal the influence of the sealing temperature on the hardness and residual stress of the sealing layer. The results show that a sufficiently high sealing temperature will drive the sealing layer to transform from a hybrid structure to the liquid phase, accelerate the element migration, eliminate pores, stabilize and compact the structure and improve the sealing performance of the vacuum glazing. As the sealing temperature increases, the residual stress and hardness substantially increase. However, when the sealing temperature reaches 460 C, the residual stress and hardness begin to plateau, and when the sealing temperature reaches 470 C, no further change can be detected. Therefore, a sufficiently high sealing temperature is beneficial to the bonding of glass and sealing solder and can promote the sealing performance of the vacuum glazing. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Vacuum glazing Sealing edge Microstructure Residual stress Hardness
1. Introduction Vacuum glazing is a new energy-saving transparent construction material with excellent thermal insulation, ageing resistance, visible light transmission and infrared light reflection properties. These properties result in good development potential, and have made vacuum glazing a research focus of glass processing technology [1e3]. In 1989, the first successful vacuum glazing sample was fabricated by Collins and Robinson at Sydney University [4]. Vacuum glazing is produced by coating all around two glass panes with sealing solder, and then put them into the vacuum heating furnace for sealing. To avoid deformation from the high temperature, the sealing should be performed below the softening temperature of the glass, which is generally low (for example, the softening temperature of common soda-lime glass is lower than 600 C) [5]. The sealing of vacuum glazing is a low-temperature process, and the vacuum glazing sealed in a heating furnace
* Corresponding author. Tel.: þ86 13852706878. E-mail addresses: [email protected] (H. Miao), [email protected] (X. Shan), [email protected] (J. Zhang), [email protected] (J. Sun), [email protected] (H. Wang). http://dx.doi.org/10.1016/j.vacuum.2015.02.009 0042-207X/© 2015 Elsevier Ltd. All rights reserved.
should maintain its shrinkage during the cooling process [6]. This shrinkage could generate stress on the sealing edge of the vacuum glazing, and if the stress exceeds the allowable stress of the glass or sealing layer, the glass, the sealing layer or the sealing interface could break. Therefore, the sealing solder must have a thermal expansion coefficient that matches that of the glass. Different sealing temperatures may change the growth form and the grain size of the solder, and the element diffusion caused by the element affinity on the bonding interface between the solder and the glass substrate may further change the sealing form, stress, microstructure and performance of the glass and thereby lifetime [7,8]. Thus vacuum glazing has a very great need for high-performance sealing solder and sealing technology. Due to the difficulty of the manufacturing vacuum glazing and the few people who have mastered this technology, research has mainly focuses on the bearing stress, the residual stress of the sealing edge and heat transfer mechanisms of the vacuum glazing [9e12]. Manufacturing theory and design theory remain to be established, as the theoretical research in this field is still at an exploratory stage and achievements are few. Based on the vacuum welding method of side sealing vacuum glazing proposed at Yangzhou University, this paper takes full advantage of the excellent performance of PbOeTiO2eSiO2eRxOy
8
H. Miao et al. / Vacuum 116 (2015) 7e12
Table 1 Composition and various performance indexes of the sealing solder. Composition/wt%
PbO
TiO2
SiO2
CuO
75
15
10
<2.5
Sintering temperature/ C 420e520
Properties
Coefficient of thermal expansion/ C 90e94 107
system welding solder. It discusses the manufacture of the solder and the sealing process [13,14]. Then it describes a series of trails on vacuum glazing at different sealing temperatures, analyses the impact of the sealing temperature on the microstructure of the sealing layer, studies the diffusion of the affinity element on the sealing interface, evaluates the effectiveness of the combination of material science features on the interface and reveals the influence of the sealing temperature on the hardness and residual stress of the sealing layer. This research not only provides significant data related to the application and verification of vacuum glazing, the prediction of its lifetime and the design and improvement of its structure, but also lays the foundation for the manufacturing and commercialization of the vacuum glazing.
2. Testing material and method 2.1. Testing material The material used in the experiment is two 800 600 mm sodalime glasses, each 4 mm thick with an elasticity of 72.45 MPa, a Poisson's ratio 0.22, a tensile strength of 40 MPa and a compressive strength of 880 MPa (both the tensile strength and the compressive strength were selected at their lower limits). The spacing distance between each pillar is 40 mm. The diameter of each pillar is 0.6 mm, with a height of 0.2 mm, an elasticity modulus of 55 GPa, Poisson's ratio of 0.25 and a density of 2500 kg/m3. Mix 99% pure and 30-mm-diameter PbO and TiO2 powders in a proportion of 70:15, and then add 10%, <2.5% and <2.5% SiO2, CuO and Fe2O3 respectively (purity 99.8% and average size 50 nm). Place the powder into a stainless steel ball grinder, add some dispersing agent, and mix the powder by the wet milling method, with a powder quality ratio of 10:1, a powder milling time of 0.5 h and a revolving speed of 250 r/min. Then heat the powder in a vacuum drying oven to obtain the system welding solder PbOeTiO2eSiO2eRxOy. The ingredients and properties of the final mix powder are as shown in Table 1. The grain size of the welding
Fig. 1. SEM photo of the sealing solder.
1
Fe2O3 <2.5
Melting temperature/ C min1 460e480
solder for the experiment is within 1e25 mm, with an average size of 12.15 mm, as shown in Fig. 1. 2.2. Testing The vacuum glazing is performed by placing one clean glass pane over another clean glass pane with pillar array, and then coating around the two glass panes with welding solder. The panes are than heated them in a 420e520 C vacuum environment for 15e60 min. The two glass panes are sealed with welding solder as the water and organic mater are removed from the glass to form an integral part of the structure [6e8]. The structure of the Low-E vacuum glazing is as shown in Fig. 2. Using a diamond wire saw cutting machine at the Nanjing University of Aeronautics and Astronautics, we cut a 20 20 8.2 mm sample from the sealing edge of the sealed 800 600 8.2 mm vacuum glazing, removed impurities such as oil contamination from the sample, and conducted ultrasonic cleaning in acetone for 15 min. The samples was then repeatedly washed with alcohol and dried and the microstructure of the section and the phase structure of the sealing layer were analyzed using a Hitachi S-4800 thermal field emission scanning electron microscope (SEM) and a D8-Advance poly-crystal X-ray diffraction instrument, Bruker AXS, German [14]. The X-ray testing conditions are a Cu target, a working voltage of 60 kV, a working current of 450 mA and a scanned area of 10 e80 . A hardness test was conducted by using an MHV-1000 digital Vickers micro-hardness test to determine the hardness of the sealing layer of the vacuum glazing following different sealing temperatures. The hardness measurement conditions were a load of 200 g load and a duration of 15 s. The average value was taken after measuring several points, and the hardness testing position of the sealing layer is as shown in Fig. 3. The residual stress of the sealing layer was tested using an electrical measuring method, adopting the KHCR-5-120-G16 weldable high-temperature strain gage produced by the Japan KYOWA Company [3]. The high-temperature strain gages were Pre-buried in the welding solder between the two glasses at a distance of 100 mm. Then the two wiring ends of the strain gage were set aside, and they were placed in the vacuum sealing furnace for sealing. The cooled and educed strain gage was connected to the deformation instrument to test the residual stress of each point, and use the average value was used.
Fig. 2. Schematic diagram of vacuum glazing with a fused edge seal.
H. Miao et al. / Vacuum 116 (2015) 7e12
9
Fig. 3. Hardness testing position.
3. Results and discussion 3.1. Sealing structure Fig. 4 shows the SEM diagram of the sealing section of vacuum glazing under different sealing temperatures. When the sealing temperature is 440 C, the connection between the sealing layer and the glass is contact and clear, with an obvious boundary, as a bulk hybrid. There are some pores and cracks between the bulks and some solder in the sealing layer that has not yet melted because the connection temperature had just reached the softening temperature of the solder. The viscosity of the fusion solder was still relatively large, which is unfavourable to the overflow of the residual gas between the particles that had not yet been melted. Therefore, a hybrid structure with independent pores and cracks is generated. During the melting, some elements will migrate to the interface and form a reactive wetting layer on the interface, but the element migration will be slow due to the low sealing temperature and the impedance of the solid grain. At this time, the reactive wetting layer is not that obvious but has a compact structure, as shown in Fig. 4(a). From Fig. 4 we can see that the liquid phase group mainly contains Pb, Ti and O, while the grain group mainly contains Cu, Si, Pb, Ti and O, so the addition of CuO in the welding material makes the sealing temperature of the solder higher. At a sealing temperature of 450 C, the bulks in the sealing layer is reduced and the pores and cracks in the hybrid structure are also reduced. As the solder viscosity is decreased, many pores will overflow to the outside from the melting layer due to the internalexternal pressure difference, while some cannot overflow. The increase in the sealing temperature will accelerate the element migration, enabling an obvious reactive wetting layer to continuously distribute on the interface as a lamellar with a certain thickness, as shown in Fig. 4(b). When the sealing temperature reaches 460 C, the bonding between the sealing layer and the glass is even tighter, and the structure stably compacts (see Fig. 5). The sealing performance is good, there is no obvious bulk, and it mainly exists in the form of a glass liquid. Through XRD analysis, the main crystal phase generated is PbTiO3 (as shown in Fig. 6). Most of the pores disappear due to substantial micro-pore. convergence as the sealing temperature increases and most bubbles converge under the internal-external pressure differential overflow from the sealing layer, as shown in Fig. 4(c). The relatively high sealing temperature may cause some chemical and physical changes in the glass and sealing solder, and the element dispersion and migration on the sealing interface may affect the sealing performance of the vacuum glazing to some degree. During the element migration, the affinity element undergo chemical and physical reactions on the sealing interface and generates new matter, and the new chemical bond of the new matter can strengthen the sealing between glasses and solder and improve the bonding performance between the glass and the sealing layer.
Fig. 4. SEM diagram of sealing section under different sealing temperatures. (a) Sealing temperature as 440 C. (b) Sealing temperature as 450 C. (c) Sealing temperature as 460 C.
However, excessive element migration may generate too much new matter, which may affect the performance of the solder and glass themselves. Fig. 6 shows the energy spectrum on the sealing interface. The sealing solder mainly contains Pb, Ti, Al, Si and O. The glass used is common soda-lime glass, which mainly contains Si, Na, Ca and O. Among these elements, Na and Ca are the main elements unique to the glass material, and Pb, Ti and Al are unique to the sealing solder. As Au is left in the sample due to the metal
10
H. Miao et al. / Vacuum 116 (2015) 7e12
sealing solder at a certain depth of the glass basis with a content of 15.3%. Therefore, Pb migration on the interface will cause some reaction on the interface between the glass and the sealing solder, which is favourable for the bonding between the glass and the sealing solder, because that reactive wetting layer can help improve the bonding performance. As the sealing temperature increases, the reactive wetting layer becomes more obvious and the bonding performance between the reactive wetting layer and the glass improves. Hence the use of a higher use of sealing temperature can improve the sealing performance of vacuum glazing. 3.2. Hardness of the sealing layer
Fig. 5. Energy spectrum of hybrid structure in sealing layer. (a) Bulks. (b) Liquid phase.
spraying before observation, its impact is omitted. From the figure we can see that the elements between the reactive wetting layer and intermediate melting layer mainly include Pb, Si and Al, and the main ingredients between the two are PbO, SiO2 and Al2O3. Compared to the intermediate melting layer, the content of Pb and Si in the reactive wetting layer is substantially improved, reaching 36.61% and 11.9%,respectively, while the change in Al is subtle, which means that the migration of Pb is the strongest, and that of Si is next. Al has high thermal stability and the contents of PbO and SiO2 are much closer to the content of the basis glass. All of these features make the interface more compact. The absence of Na and Ca which are unique to the glass material in the reactive wetting layer indicates that the glass basis elements undergo no obvious dispersal or migration during sealing, while Pb appears in the
Fig. 7 shows the hardness of the sealing layer at different sealing temperatures. As show in the figure, when the sealing temperature is 440 C, the hardness of the sealing layer is 279HV, and when the temperature is 460 C, the hardness also significantly increases to 298HV. This is because the sealing layer mainly exists in the form of a hybrid when the temperature is lower, and the higher sealing temperature induces the transformation from the hybrid to the liquid phase, resulting in a much higher hardness than that of the bulk solid. If the sealing temperature increases to 470 C, the hardness of the sealing layer is 300HV, which is a slight increase. This is because the sealing layer at this time exists in the form of a liquid phase with a stable structure, so a further increase in the temperature would not substantially change the performance or hardness of the sealing layer (see Fig. 8). 3.3. Residual stress of the sealing edge The residual stress of the sealing layer of the vacuum glazing is mainly the residual tension caused by the thermal residual stress from the following three causes: (1) As the high temperature at the sealing edge cools, the heat dissipation speed of the surface between the sealing layers is different from that of the glass basis, leading to a temperature gradient inside the sealing layer. Moreover, the heat conductivity coefficient of both the sealing solder and the glass basis is small and the material is a poor conductor of heat, so the temperature gradient is even larger. During the sealing, the thermal expansion of the solder and the basis glass are different
Fig. 6. XRD spectrum of sealing layer at 460 C.
H. Miao et al. / Vacuum 116 (2015) 7e12
11
Fig. 9. Residual stress of sealing layer under different sealing temperatures.
Fig. 7. Energy spectrum of sealing layer of vacuum glazing. (a) Intermediate Melting Layer. (b) Reactive Wetting Layer (c) Glass Substrate.
and must have different yield behaviours, causing the parts in the sealing layer to have different plastic deformation behaviours and further causing thermal gradient residual stress in the sealing layer. (2) Low-melting-point sealing solder is a compound material, with different thermal expansion coefficients, and the thermal expansion coefficient between the basis glass and the sealing solder is also different. These coefficient differences make the deformation of materials different when cooling thus leading to thermal residual
Fig. 8. Hardness of the sealing layer at different sealing temperatures.
stress mismatch in the sealing layer. The research shows that even if the thermal expansion coefficient difference is smaller, it may still cause a large thermal residual stress mismatch [4e10]. (3) When the sealing layer of the vacuum glazing cools down, the transformation of the sealing layer from a hybrid structure to the liquid phase will cause a volume change. Whether the phase change part is bigger or smaller, stress will be produced between the phase change part and the unchanged part, leading to a phase change residual stress in the sealing layer. To study the impacts of different sealing temperatures on the residual stress of the vacuum glazing sealing layer, the samples sealed under different temperatures are cooled to room temperature 20 C in 24 h. The samples at four different sealing temperatures in the experiment were compared, as shown in Fig. 9. As the figure shows, when the sealing temperature is 440 C, the residual stress is 40 MPa, and when the temperature increases to 460 C, the residual stress also increases to 55 MPa. The residual stress was significantly higher at 460 C. When the temperature increases to 470 C, the residual stress increases to 56 Mpa, which is a negligible change. The different cooling rates for the vacuum glazing sealing layers of different residual stresses, according to a sealing temperature of 460 C forcibly cooled to a room temperature of 20 C, extracted six different cooling speeds that are shown in Fig. 10. The figure shows that with the acceleration of the cooling rate, the residual stress of the sealing layer of shows a rising trend. At a cooling rate of 4 h, the residual stress is 69 MPa, but the sealing layer has many
Fig. 10. Residual stress of vacuum plate glass sealing layer under different cooling speed.
12
H. Miao et al. / Vacuum 116 (2015) 7e12
Acknowledgement The authors acknowledge the support from the National Youth Science Fund Project (No. 51302242) of China, the National Natural Science Foundation(No. 51172199, No. 51372216) of China, the Natural Science Foundation (No. BK2012530) of Jiangsu Province, and the Construction and Science and Technology Plan Project (No. 2012BAD32B11) of Yangzhou. References
Fig. 11. Partial cracks in vacuum glazing sealing layer.
microcracks (Fig. 11), so the cooling rate is not too fast. To prevent the sealing layer from being too large due to the residual stress microcrack formation, resulting in the failure of the vacuum chamber, the cooling rate is generally maintained at more than 12 h. 4. Conclusion Solder sealing performance is critical to the performance of the vacuum glazing. The variations of the microstructure, hardness and residual stress at different temperatures were analyzed. The sealing solder is effective when it is dense and well-connected with the glass substrate, and then it has excellent anti-ageing properties (the first generation of vacuum glazing samples we produced in 2005 is valid proof of this). For PbOeTiO2eSiO2eRxOy sealing solder, we suggest maximizing the sealing temperature and maintaining an adequate cooling time so as not to affect the performance of the vacuum glazing. At the same time, this approach will aid in identifying the solder sealing performance and improving the manufacturing processes of vacuum glazing.
[1] Synowicki RA, Johs BD, Martin AC. Optical properties of soda-lime float glass from spectroscopic ellipsometry. Thin Solid Films 2011;519(9):2907e13. [2] Luo LH, Huang ZZ, Wu YF, Cheng D, Sun LL, Cheng L, et al. Sealing properties of BaOeCaOeAl2O3eB2O3eSiO2 glass ceramics to large size planar IT-SOFC. J Chin Ceram Soc 2014;42(1):101e7. ar FO, Revel B. Synergic effect of V2O5 [3] Sharma K, Kothiyal GP, Montagne L, Me and P2O5 on the sealing properties of barium-strontium-alumino-silicate glass/glass-ceramics. Int J Hydrogen Energy 2013;38:15542e52. [4] Robinson SJ, Collins RE. Evacuated windows-theory and practice. In: Proceeding of ISES solar world congress; 1989. Kobe, Japan: Oxford: Pergamon; 1989. [5] Griffiths PW, di Leo M, Cartwright P, Eames PC, Yianoulis P, Leftheriotis G, et al. Fabrication of evacuated glazing at low temperature. Sol Energy 1998;63:243e9. [6] Miao H, Xi XB, Zhang RH, Zuo DW, Wang HL. Numerical calculation and experimental research on wind load of vacuum plate glass. Acta Energiae Solaris Sin 2013;34(11):1888e94. [7] Fang Y, Hyde TJ, Ayra F, Hewitt N, Eames PC, Norton B, et al. Indium alloysealed vacuum glazing development and context. Renew Sustain Energy Rev 2014;37:480e501. [8] Ng N, Collins RE, So L. Thermal conductance measurement on vacuum glazing. Int J Heat Mass Transf 2006;49(25e26):4877e85. [9] Yu Li A, Du JP, Han MF, Peng SP. Structure and sealing properties of Bi2O3eBaOeSiO2eRxOy glass. J Univ Sci Technol Beijing 2011;33(12):1529e33. [10] Zhang LY, Qin ML, Qu XH, Lu YJ, Zhang XY. Study on active brazing of AlN ceramics and MoeNieCu alloy. Rare Metal Mater Eng 2009;38(12):2159e62. [11] Zhang RH, Ma CW, Miao H, Zhang Y, Kong DJ. Theoretical analysis and experimental study on heat transfer in vacuum plate glass. Trans Chin Soc Agric Mach 2006;12:134e8. [12] Miao H, Zhang RH, Zuo DW, Wang HJ, Wang HY. Finite-element analysis of static and modal response of He vacuum plate glass. J Yangzhou Univ 2012;15(1):69e72. [13] Luo XL, Yu XL, Wang YL, Gu L, Li F, Li QL, et al. Sealing technology of the SiO2eBaO system glass and metal. J Chin Ceram Soc 2013;41(6):858e62. [14] Cheng Y, Xiao HN, Guo WM. Effects of GeO2 on structure and properties of PbOeB2O3eZnO. J Inorg Mater 2006;21(3):533e8.