Design and development of lead-free glass-metallic vacuum materials for the construction and thermal performance of smart fusion edge-sealed vacuum glazing

Highlights

• Novel concept of forming glass-metallic textured layer for hermetic fusion seal.

• Fusion edge-seal invented when Sn62-B₂O₃38 wt% fused with Sn90-In10 wt% alloy.

• Fusion edge-sealed vacuum glazing (FSVG) achieved vacuum pressure of 8.2 · 810⁻⁴ Pa.

• Validated FEM predicted central thermal performance of FSVG to be 1.039 Wm⁻²K⁻¹.

• FSVG is cost-effective, ultrasonic-soldering free and Pd-free solution.

Abstract

Advancement in hermetic (vacuum-tight) edge-sealing materials has been one of the challenges since decades because of the existing cost, use of hazardous substance and complexity-to-construct issues in vacuum glazing. This paper presents novel experimental findings with designs and methods developed to construct and analyse thermal performance of the fusion edge-sealed vacuum glazing. The novel concept of fusion edge-seal consists of forming a thin glass-metallic rigid textured layer, in which the formation processes and experimental glass-metallic textured surface bonding property tests of 15 samples are microstructurally analysed using FIB-SEM and optical microscopy and succeeded the correct mixture of B₂O₃38-Sn62 wt%. Experimental analyses of at least 60 samples conducted using different techniques and Pb-free materials, among which five vacuum glazing samples of various designs and techniques discussed in this paper. The fusion edge-sealed vacuum glazing, constructed with bonded Sn62-B₂O₃38 wt% surface textured fused with Sn90-In10 wt% alloy at 450 °C, achieved at the hot-plate surface heat induction of 50 ± 5 °C and the cavity vacuum pressure of 8.2 · 10⁻⁴ Pa. Validated 3D FEM employed and the centre-of-sheet and total thermal transmittance values of fusion edge-sealed vacuum glazing (sample ‘A5’), area of 300 · 8300 mm with 10 mm wide fusion edge-seal, predicted to be 1.039 and 1.4038 Wm⁻²K⁻¹, respectively.

Introduction

The perception at which, nowadays, buildings cannot be imagined without glass is because glazed windows play an imperious role of allowing natural daylighting [1], [2], but with repercussion of space-heating energy loss in cold-arid and space-cooling energy loss in hot-arid areas [3]. As such, evitable energy losses through glazed windows of buildings indirectly contributing to carbon emissions and, thus, impelling climate change [4]. This is, predominantly, due to an inadequate thermal transmittance value (U-value) of a glazing. In spite of the retrofits of a number of glazing technologies to building sector that usually do not compromise natural daylighting [5], such as cavity/cavities of two glass sheets (glazing) filled with air and/or heavy gases with low-e coatings, and could achieve the U-value of up to 1.4 Wm⁻²K⁻¹ [6]; However, it does not account the increase of cavity thickness that affects window frame sizes and limits retrofit options on existing buildings. To achieve U-value of less than 1.1 Wm⁻²K⁻¹ [7] without compromising the natural daylighting and cavity thickness, a vacuum glazing is a feasible solution [8]. Vacuum glazing has a 0.15 mm thick cavity between two sheets of glass, supported with an array of tiny stainless-steel pillars i.e. 0.15 mm high and 0.3 wide [9]. This cavity entails reduced atmospheric-air pressure down to high-vacuum pressure (between 0.13 Pa and 1.33 · 10⁻⁴ Pa). Subsequently, it suppresses gaseous heat conduction and convection to infinitesimal level because of an increase of mean-free path between two air molecules beyond 3000 m. The high-vacuum pressure must be maintained with a hermetic (vacuum-tight) edge-seal around the periphery of two sheets of glass, whilst avoiding the problems of future gas leaks, outgassing and absorption of moisture in order to provide long-term durability. Thus, the U value is dependent on the vacuum-tight edge seal of the vacuum glazing.

Advancement in hermetic (vacuum-tight) edge-sealing materials is one of the challenges since decades because of the existing cost [10], use of hazardous substance (Pb) [11] and complexity-to-construct [12] issues in vacuum glazing that this paper addresses with empirical evidences. However, it is important to appreciate and briefly discuss the prior research art on the past inventions of vacuum glazing and discuss, here, the realism of the challenges exist the need of a new hermetic (vacuum-tight) edge-seal. The first successful construction of vacuum glazing [13] utilises the method of using Schott solder glass type 8467 (lead borate glass) [14] to hermetically seal the edges of two glass sheets at the high-temperature of around 450 °C, this achieved the U-value of 0.8 Wm⁻²K⁻¹ [15]. Subsequently, this method attained an attention at mass manufacturing level under the trade name ‘SPACIA’ in Japan by Nippon Sheet Glass (NSG) group [16]. It is often desirable to have lead (Pb)-free materials in building components, because the use of Pb is restricted in buildings and the allowable concentration quantity of not more than 0.1 wt% in homogeneous materials [17]. Though, the solder glass powder (Schott 8467) has more percentage of hazardous-substance (Pb) than regulated. This high-temperature solder glass edge-sealed vacuum glazing also impose restriction on the use of tempered glass and soft (low-emittance) coatings such as silver (Ag) [16]. The second successful construction of vacuum glazing utilises the method of ultrasonically soldering of pure indium 99.99% (In) around the periphery of two glass sheets at the low-temperature of around 156 °C achieved U value of 0.86 Wm⁻²K⁻¹ [18]. This method overcomes the restrictions of tempered glass and soft (low-emittance) coatings but inefficacious at the production level, because of the cost and scarcity of semi-precious metal (In). The third successful construction of triple vacuum glazing and vacuum glazing, invented by Dr Memon [19], was based on ultrasonically soldering the primary seal, at low-temperature around 200 °C, made of composite CS-186 or Sn-Pb-Zn-Sb-AlTiSiCu in the proportion ratio of 56:39:3:1:1 by wt% and the secondary seal made of reinforced steel epoxy [20]. This composite hermetically sealed the edges of glass sheets and predicted the U-value of 0.33 Wm⁻²K⁻¹ and 0.91 Wm⁻²K⁻¹ for triple vacuum glazing and vacuum glazing, respectively. This method overcomes the cost issue, without compromising the scope of using tempered glass and soft coatings, but its composite edge seal has higher percentage of lead (Pb) and it is complex-to-construct, mainly due to the need of precision in ultrasonically soldering the edges of glass sheets. The prior research methods have, however, quite successfully solved the problem of effective hermetic seal, design and construction improvements, fabrication processes and the significant predictions in the reduction of U value of less than 0.5 Wm⁻²K⁻¹ for vacuum glazing units. But the current research challenges are the cost-effectiveness, not use of hazardous restricted metals such Pb, reduction in complexity-to-construct issue and the long-term stability of vacuum pressure in the cavity of the vacuum glazing.

It is perceptible from the above that there exists a need of the invention in the art for the development of an alternative cost-effective Pb-free edge sealing material for vacuum glazing, and corresponding novel method of making the same, including an improved evacuated cavity pressure with their stability over long term between opposing glass sheets.

To achieve hermeticity or vacuum-tight edge seal, a high vacuum pressure (0.13 Pa to 1.33 · 10⁻⁴ Pa) is required in the cavity between two glass sheets supported by tiny pillars and must be maintained by the edge seal providing a mechanically strong bond, be free from gas molecules, match the thermal expansion characteristics between the glass and metals/alloys. It is, therefore, important to discuss, as follows, the (i) thermal expansion of sealing materials and glass; (ii) incidence of the interaction between glass and metal seal; (iii) importance of vapour pressure in choosing metals to form glass-to-metal seal and; and (iv) viscous behaviour of the glass and sealing materials.

Glass-to-metal seals often have mismatch in their thermal expansion coefficients. This causes stresses and strains with uneven temperature distributions or contact with other metallic materials and composites [21]. To better match the thermal coefficients to the glass and the metal seal, an extra layer of different metal or layer of oxide could possibly produce the desired sealing composition [22]. The whole expansion characteristics of the sealing material (metal or solder glass) and the glass should also be considered [23]. For example, a certain amount of mismatch from 1 to 2 · 10⁻⁷/°C or even up to 6 · 10⁻⁷/°C (set-point temperature range of 300 °C to 600 °C) [24] could be tolerated in a solder glass to glass seal. Solder glasses are classified, according to their thermal expansion coefficients, into three categories: hard solder glasses with thermal expansion of 25 · 10⁻⁷/°C; semi-hard solder glasses with thermal expansion of 25 to 50 · 10⁻⁷/°C; and soft solder glasses with thermal expansion of more than 50 · 10⁻⁷/°C [24], [25]. The thermal expansion of pure metals (such as Sn, Zn, Al, Cu, Bi, Pb or In) is linear with temperature, and that of its alloys the thermal expansion is linear up to the point of inflection. After which, their expansion increases more rapidly [26]. The solder glass seal becomes rigid, fragile, brittle and inelastic up to the inflection point and on further cooling to the ambient temperature. Substantial cracking in the solder glass bond will appear if the contraction rate of the glass and solder glass differs extensively [27]. A glass-to-metal seal could be under compressive stresses when the thermal expansion coefficient of metal is greater than the glass. The glass-to-metal seal is under tensile stress when the thermal expansion coefficient is less than the glass [24]. To substantiate stresses normal to the glass-to-metal interface, an approach suggested in [27] that when the metals of larger thermal expansion coefficient than the glass, the metal part needed to be soft and thin to deform elastically in order to allow changes in the dimensions of the glass when it cools. The total stress scenario in the glass-to-metal seal is far more complicated than in the case of fusion of different composition of materials to the glass surface. For the strong bond between solder glass and the glass, the thermal expansion coefficient of the solder glass must either be similar to or less than the glass.

In general, a good glass-to-metal seal can be achieved by either creating a layer of materials or interactions between the contact surfaces. The nature of the interaction and its mechanism vary dependent on the type of metal and/or its alloys. It was suggested in Ref. [28] that a good seal can be achieved when the glass is saturated with the appropriate metal oxide layer as a transitional character. At elevated temperature when the atoms of metal-wire and the metal ions of the soda-lime glass are moveable, because the chemical composition of soda-lime glass [29] has different weight % of SiO₂, Al₂O₃, TiO₃ and Fe₂O₃ and others. So, there will be a continuous exchange at the glass-to-metal interface. In an ideal case scenario for the formation of hermetic seal, metal ions from the soda-lime glass surface will diffuse into the metal-wire where they will gain electrons and become zero valence metal atoms. Whilst, the metal-wire atoms will diffuse into the soda-lime glass surface and become ionized. The reliability of the seal would then be assessed determined by thickness, uniformity and types of oxide layer [26]. The thickness of the layer is critical to avoid the oxide layer being porous or mechanically weak or brittle. In high vacuum applications, the presence of the oxide layer in the glass-to-metal seal should be avoided. Since, the seal enters into the outgassing processes such as surface desorption, diffusion and permeation of gases [30]. The surface appearance, such as colour change to a darker shade in the glass-metal after sealing, generally indicates the level of oxidation and the reliability of the seal. In some cases, however, e.g. tungsten, this is not a definite indication as to whether the seal contains or does not contain the required oxides. The level of oxidation is influenced by the type of metal used and the processing temperatures. Any gas molecules that remain trapped in the glass-to-metal seal usually form micro air gaps or pinholes in the seal. The processing temperature at which the oxide-layer is formed plays an important part in reducing oxygen molecules in the seal to a negligible level. For metals, if the surface is contaminated by a thin layer of oxide, the initial outgassing rate at ambient temperature is of the order of 10⁻⁴ Pa m³ s⁻¹m⁻² [26]. Nevertheless, the outgassing rate decreases with time and subject to temperature regimes applied. The rate of outgassing increases due to the presence of water vapour between the seal and the glass. This can be decreased exponentially to a much lower value by continuing evacuation for more than 10 h. The gases left within the seal are commonly CO, CO₂, O₂, N₂ and H₂ which can be removed during the molten stage of the seal [21] by making it under high-vacuum environment. Some of these gases react with metals to form compounds. The type and composition of the materials used for sealing and the cleaning methods used before fabricating the sample will determine the gases that are left.

The vapour pressure of metals indicates the evaporation rate in the liquid phase and is a tendency of molecules in a liquid phase, at a given temperature, to escape within a closed system. For example, the vapour pressure of In increases from 10⁻² Pa to around 1 Pa with a temperature increase from around 650 °C to 850 °C. Comparing this with Sn for a similar temperature increase, the vapour pressure increases from 10⁻⁵Pa to 8 · 10⁻³Pa. At ambient temperatures metals, such as Pb, In, Sn, Bi, Sb, Zn, or Cd, have a vapour pressure above 10⁻⁹ Pa [21]. Among these metals, Cd and Zn have a higher vapour pressure and should be avoided, and not as elements for the formation of glass-metal seals. Pb, Sn and In can be used in high vacuum applications due to their low vapour pressures [26], [31]. Bi is brittle in nature and individually this element cannot be used in high vacuum applications for the glass-metal seals [32]. The basis of the method and techniques explored in this research is the use of B₂O₃ to form a layer of oxides on the glass surface, which become glassy and rigid in shape after heating to its melting temperature. It is worth noting that the formation of B₂O₃ was initially reported in the scientific paper no. 857 from research laboratories of Westinghouse Electric & IC manufacturing company in 1937 and then patented in Ref. [33]. B₂O₃ has increasing applications as a fluxing agent for glass and enamels [34]. The melting point of crystalline B₂O₃ is between 450 °C and 470 °C.

One of the essential characteristics, when forming a glass-to-metal seal, is the viscous dynamic behaviour at their straining and setting point. In addition to the thermal expansion of the materials, the viscosity at melting temperature of the material also influences the mechanical rigidity of the seal. According to the ASTM definition [35], glass is an inorganic liquid product which has cooled off and solidified, with very high viscosity, without crystallizing. When the temperature increases, the viscosity decreases and the glass gradually assume the character of a liquid. Glasses are traditionally classified into two categories. The first category is the ‘hard’ or borosilicate, in which the main additive to the silica is B₂O₃. The second category is the ‘soft’ in which the main additive is either Na₂O to give soda glass or PbO to give lead glass [21]. At the temperature of 470 °C [27] when all movement of the glass molecules at which a certain glass viscosity is reached and no more strain can be introduced into the hot soda glass (strain point) the viscosity is ≈ 10^13.5 Pa.s [21]. The viscosity of 10¹⁴ Pa.s [36] is believed to be the limit for the solid (fragile) body of the glass. For temperatures above 300 °C, a tempered glass begins to lose its temper and because of this it could not be used for the high temperature construction process. Glasses with low softening temperatures (<600 °C) are often referred to as solder glasses. In general, the set point temperature of the glass to solder glass seal is ≈ 510 °C. In which the viscosity of the glass being ≈ 10¹² Pa.s and the viscosity of solder glass is in between 10⁴ and 10⁵ Pa.s [26]. The pure metals or alloys of In, Sn, Bi or Pb, the viscosity is above 10¹⁴ Pa.s can be achieved at the temperature less than the strain point of the glass, this could be the fragile or solid phase point i.e. below 350 °C [27].

The concept of producing alloys for the glass seals has, however, offer a number of solutions. An alloy of Bi (53–76 wt%), Sn (22–35 wt%) and In (2–12 wt%) was claimed to have an excellent oxidation resistance and sealing property suitable for rotating plug of a nuclear reactor [37]. Pb-B glasses or Pb-Zn-B glasses in which the Zn-Pb (mol ratio below 1:2) was claimed to be used in sealing semiconductor packages at the temperature of 450 °C [38]. A glass solder consists of B₂O₃ (20–30 mol %), PbO (60–69 mol %), ZnO (0–10 mol %), CuO (0–6 mol %) and Bi₂O₃ (0.5–2.0 mol %). In which the mixture of these composites with SiO₂ and Al₂O₃ could be a suitable solution to seal with soda-lime-silica glass [39]. A lead-free composite material or a glass sealing paste consists of Bi₂O₃ (70–90 wt%), ZnO (1–20 wt%), B₂O₃ (2–12 wt%), Al₂O₃ (0.1–5 wt%), CeO₂ (0.1–5 wt%), CuO (0–5 wt%), Fe₂O₃ (0–0.2 wt%) and CuO + Fe₂O₃ (0.05–5 wt%) was claimed to be a reasonable sealing composite for the glass seal [40].

To date, there has been no reported study which has identified a cost-effective, Pb-free and ultrasonic-soldering free high temperature sealing method for vacuum glazing construction and to triumph these challenges, this paper presents the novel experimental findings that addresses these issues. In this paper the design, methods and techniques to form a hermetic edge seal, named as a fusion edge-seal, are presented for the successful construction of fusion edge-sealed vacuum glazing. Fusion edge-seal is constructed with a process invented after a series of experiments that involves the formation of the a thin glass-metallic surface texture, made of the particles of B₂O₃ and/or Sn, and successfully fused with the hard Sn or Sn-In alloy to form an hermetic (vacuum-tight) edge-seal. This requires the formation of a textured layer on the glass surface using an appropriate mix of B₂O₃ and/or Sn. This method, as shown in Fig. 1, was achieved after detailed study and experimentations in understanding the glass components and metals behaviour at different temperature regimes. This paper also contributes to the thermal performance prediction of the fusion edge-sealed vacuum glazing with a validated finite element model of the fabricated (sample ‘A5’). One of the main drifts of this invention is its cost-effectiveness, if compared to composite or indium sealed vacuum glazing, discussed in this paper, and it doesn’t contain any hazardous Pb metal in it when compared to solder glass sealed vacuum glazing. Other features and advantages of the fusion edge-seal will become apparent upon examining the following detailed experimental analyses of the fusion edge-sealed vacuum glazing.