Articles Fabrication
Almost there

April 11, 2013
By Patrick Flannery


Nothing insulates like nothing. That is the simple principle behind
vacuum insulating glass, an idea that has tantalized building engineers
for a century.

Nothing insulates like nothing. That is the simple principle behind vacuum insulating glass, an idea that has tantalized building engineers for a century. But VIG that realizes its potential to outperform conventional IG solutions has proved to be a thorny problem. Fortunately, a convergence of positive factors may be bringing high-performance VIG to the point of market feasibility for the first time.

For a potentially ground-breaking innovation, it doesn’t look like much. The corrugations on this stainless-steel strip are the secret to EverSealed’s flexible mechanical seal that may make VIG practical in cold-weather climates. 


Vacuum insulating glass presents a difficult engineering challenge because solutions to one problem tend to create headaches in another area. There have been four main technical hurdles to overcome:

  • Creating and maintaining a superior hermetic seal.
  • Preventing the glass lites from bowing inward under negative pressure and touching, creating a thermal short.
  • Preventing the thermal growth and contraction of the lites from stressing the glass and seal materials to failure.
  • Obtaining the necessary hard vacuum in a timely manner.

All these obstacles have been overcome some time ago – but not in a fashion that yields insulating levels significantly better than high-quality IG units, and not in a fashion that can be warranted for cold-weather climates. Existing VIG designs are also limited in size and to using annealed glass. For many years, VIG has been stuck in this rut of having niche application for mild-climate situations where double-glazing insulating levels were required with single-pane thickness. Nippon Sheet Glass has delivered its VIG Spacia product in Japan for years, but penetration into the North American market has been low. It delivers a very respectable 1.1 U-value in a six-millimetre-thick product, but Pilkington’s product brochure describes it as a solution for heritage retrofits where standard double-glazed will not fit.  Pilkington dropped offering its Spacia products in North America about a year ago.

As with most engineering challenges, these issues might have been resolved years ago had external factors forced grant-giving bodies and industry manufacturers to invest in finding solutions. It looks like climate change may have finally delivered the political and economic conditions necessary for invention. Guardian’s David Cooper, past-president of IGMA and chair of its VIG task group, explains why we are seeing new progress on VIG technology. “Codes and Energy Star requirements,” he says. “What the industry is going to respond to are codes, and the codes are adapting to change. Energy Star is moving to higher thermal performance standards and the industry will naturally be looking for ways to incorporate new technology that does not cost an arm and a leg. You can do a lot of things – you can start making quad units – but VIG is a very compact package. Two pieces of glass that are roughly eight millimetres thick and you get the R-value performance of a superior triple-pane product. So the driver is there to get it at the same cost point as triple.”

To see how the industry is getting there, we will take a look at two groups in the U.S. – Guardian and EverSealed Windows – who are claiming to be close to bringing to market VIG that is warranted in all climate zones and delivers R-10 or better insulating values. Guardian is, of course, one of the largest glass suppliers in the world. ESW is based out of Evergreen, Colo., and is a research lab owned by a small group of co-founders led by David Stark, the president and chief technology officer. Stark is working with Pella and Cardinal with a goal to license his process to IG fabricators once it is certified. Guardian hopes to bring its VIG units to market itself. Each of these organizations have taken different approaches to solving the four problems of VIG.

The seal of approval
To create a vacuum between two lites of glass, you need a great seal. And to meet today’s expectations of IGU longevity, you need that seal to work for at least 25 years. EverSealed set a target life span for its seal of 40 years and figured it had to achieve a gas-transfer rate of no greater than 10-13 cubic centimetres per second in order to meet that target. This is a very difficult target to hit. The standard for hermeticity in the microelectrics industry is only 10-8 cc/sec..

Normal IGU edge-sealing methods are, of course, entirely useless for fabricating VIG. No butyl-based sealant is mechanically strong enough or impermeable enough to gas and water to maintain a seal against a strong vacuum. Some form of solid mechanical seal is required. NSG’s VIG uses glass solder between the lites to create a continuous glass seal. This creates an effective seal, but involves heating the solder glass to over 400 C. These high temperatures will de-temper tempered glass making the process only suitable for annealed glass. 

Another approach is to use a strip of metal as an edge seal. But the metal must be adhered to the glass much more hermetically than is possible with standard glues or epoxies. One established method for fusing glass to metal comes out of the vacuum tube industry and involves a process known as diffusion bonding, where glass and metal surfaces are forced together under temperature and pressure in a vacuum, causing the oxide layer on the metal to bond powerfully to the glass at the molecular level. Coming as he did out of the microelectronics industry (he used to build circuits for missiles, then moved to micro-windows for digital image processors), Stark was initially interested in this process as a possible way to create permanent glass/metal bonds for VIG seals. A German group called Pro-VIG initially and for years played with using ultrasonic bonding.  This uses high-frequency vibration to abrade one material’s surface into/onto the adjacent surface.  Eventually, the Pro-VIG group stated trying soldering with indium solders before it ceased operations.  But after playing around with different alloys, Stark could not find a way to make the process work with tempered glass. The temperatures required were too high. Even with annealed glass, the cycle time was up around 10 or 15 minutes, which is prohibitive for any kind of mass manufacturing process.

So EverSealed turned to ultrasonic soldering. Unable to find a lead-free solder glass that would melt at a low enough temperature in a short enough time to not affect the fully-tempered strength of tempered low-e glass, EverSealed in 2010 attempted to develop its own lead-free solder glass with appropriate reflow properties.  After almost eight months of concerted effort, EverSealed abandoned their solder-glass and reflow development efforts.

EverSealed designed, modeled and simulated several iterations of what it called its flexible asymmetrical bellows.  The metal bellows, attached to the dual-lite VIG’s surfaces 1 and 4, acted as a hermetic flexible spring, allowing lite 1 to expand and contract with changes in its temperature without imparting or transferring significant stresses to lite 2.  However, simulations performed in late 2010 showed that although the bellows provided more than sufficient compliance perpendicular to the edges of the two lites, all designs were too stiff in the direction parallel to the edges of the lites.

In a concurrent engineering brainstorming session in January 2011 in which team members from each collaboration industry as well as engineering discipline participated, twelve concepts were proposed for a replacement flexible hermetic metal seal system.  EverSealed’s mechanical engineers down-selected and to three possibilities and then designed, modeled and simulated the performance of each during extreme thermal cycling of lite 1 of the two-lite VIG.


The result is a one-piece continuous strip of very thin stainless-steel foil, roll-formed or stamped to provide the flexible form in its center portion and provided by the former on large rolls or reels.  Flat portions of the strip (which EverSealed calls the flanges) are bent 180 degrees inwards and hermetically soldered to the perimeters of lites 1 and 2 using a lead-free and flux-free solder and ultrasonic soldering equipment.  Then the remaining flange materials perpendicular to surfaces 1 and 4 are bent 90 degrees onto pre-dispensed epoxy on surfaces 1 and 4 and the epoxy is cured.  The epoxy bond of the glass to the metal seal is exposed to stresses caused by differential expansion of lites 1 and 2 before these the hermetic glass-to-metal solder bonds on the perimeter of the lites.  So the hermetic seals experience stresses that are very small compared to the shear strength of the ultrasonically-soldered bonds.

The maximum temperatures used in ultrasonically soldering the flexible metal seal to the glass are well below the temperature at which tempered glass would begin to lose any of its fully tempered strength.  EverSealed has been using Cardinal’s tempered glass with Cardinal’s Q-366 low-e coating, but Stark states that the EverSealed assembly process should work with any U.S. or foreign float glass supplier’s low-e coated tempered glass.

So EverSealed turned to ultrasonic welding, where a strip of metallic glass solder is placed between the glass and the metal edge seal foil and welded to both surfaces with powerful vibrations. In the past, lead was added to glass solder to bring its melting point down to the point where it could be used on tempered glass without deforming it. But modern building materials cannot contain any appreciable level of lead, so Stark was challenged to find a lead-free solder and a process to go with it. His friend Edward Boulos, a veteran of the automotive glass industry, came up with both, but had to go as far afield as Russia to find it. The result is a stainless-steel foil seal, wrapped around the edges with flanges ultrasonically soldered to surfaces 1 and 4 and TIG-welded at the joint. The solder’s flow point is almost too hot for the glass, but EverSealed compensates by using a robot to apply the weld, which works faster than a human and therefore heats the glass up less.

Guardian’s Andy Russo, residential segment director, is unwilling to go into the same level of detail about his process, saying “Our developments have be focused on utilizing the same proven technology that has sealed CRT tubes and televisions for decades. This utilizes a ceramic frit paste that establishes a hermetic, molecular seal.”

Can you stand it?
When VIG was first thought of, it seems doubtful that early researchers would have realized what a major obstacle the stand-offs were going to become. Negative pressure between the lites causes the glass to bend inwards, causing major structural problems and a thermal bridge if it touches. Making the cavity larger makes the unit harder to seal and somewhat defeats one of the main purposes of VIG, which is to make a thinner alternative to triple-glazed units. The only solution anyone has been able to come up with is to place a series of tiny stand-offs between the panes to prevent them from touching. But this creates new challenges. The stand-offs themselves become thermal bridges, unless the materials are carefully chosen. Plus the spacing must be carefully calculated to prevent stress to the glass as it presses against the stand-offs. Placing stand-offs could be time-consuming in the manufacturing process, but automated, highly accurate pick-and-place machines, used in electronic circuit board manufacturing for more than three decades has eliminated this labor. Finally, the stand-offs must be miniscule so that they cannot be easily seen.

The importance of the stand-offs to the VIGU thermal properties is surprising. Seth Miller, a polymer chemist contracted by EverSealed, says the stand-off system is “the single largest contributor to the VIGU’s thermal performance.  If the stand-off’s material and diameter remain constant, the thermal conductivity of the stand-off system is inversely proportional to the square of the spacing in a linear fashion – once the stand-offs get more than 50 millimetres apart, the benefits to increasing the space between the grid of stand-offs.” 

Flexible seal solutions, such as EverSealed’s, allow the lites to move practically stress-free relative to each other with thermal expansion and contraction. This creates more challenges for the stand-off system, as glass and low-e coatings can be scratched or stand-offs dislodged by this constant rubbing. EverSealed’s solution is to plate its stand-offs with a proprietary metal alloy composition that is close to the coefficient of friction of graphite on glass and about 17 per cent of the friction of glass on glass. The stand-offs are applied to the clear or low-e surface of the tempered glass and the glass is heated to melt the stand-offs coating and solder it to the glass (or low-e coating). The roller-wave distortion of horizontally-tempered glass means that all the stand-offs do not touch both inner surfaces 2 and 3 at any time, thus requiring a calculated safety factor of extra stand-offs so that no one stand-off and glass interface exceeds the compressive strength and adjacent tensile strength of the glass.

Let it slide
Because VIG insulates so well, the two lites can experience radically different temperatures, especially in extreme climates. In one of EverSealed’s tests, surface 1 inside the test chamber was nearly 125 C, while surface 4 remained at 25 C (only two degrees warmer than the room). In a traditional IGU, this would be a smaller problem because the flexible butyl seal would allow the edges of the lites to shift relative to each other without affecting the seal. However, as noted above, VIG requires a hermetic seal that is much stronger. This is why the warranty on NSG’s VIG product carries cautions for use in areas with wide temperature fluctuations.

Without going into detail, Guardian claims to have solved the issue. “The perception in the marketplace is that there may be issues with large delta-Ts across a ceramic-frit-sealed VIG unit,” Cooper says. “I am going to tell you right now that it is all perception.”

EverSealed has taken a different approach by creating a flexible mechanical seal. Stark explains it by asking you to hold a sheet of paper by its ends in both hands, then to bring your hands together so the paper is bent almost double, but not creased. Note how the paper easily opens and closes, but gives resistance to side-to-side motion of the edges past each other (the paper does not want to twist). Then Stark asks you to crumple up the paper, flatten it out and perform the same exercise. Now, the paper flexes just as easily side-to-side as it does open and closed. Stark and his team have applied this principle by roll-forming an egg-carton pattern into strips of stainless steel foil. The strips are then wrapped about the edges of the VIGU and soldered into place. Now there is a hermetic, mechanical seal, but the lites can shift toward and away from each other, side to side and up and down, or grow and contract, creating offset edges. Finding the right combination of foil and solder materials and the right process to fabricate it has taken years, but Stark says EverSealed is there and has the testing data to prove it. To demonstrate its capabilities, the company mounted a nine-by-13-inch unit in a door of a thermal-cycle chamber and cycled the heat on surface 1 of the VIG up to 130 C and down to -60 C over 2.4 hours, including a 15-minute dwell at each temperature extreme. The chamber was cycled more than 230 times without causing a detectable leak. EverSealed says this statistically-valid accelerated test certifies a 40-year minimum expected life span in any climate for the design.

Vacuum sucks
There is a sweet spot approaching 10-3 and 10-4 torr beyond which there are diminishing benefits to creating a harder vacuum between the lites. This is not a particularly hard vacuum (thermos bottles are brought down to 10-6 torr) but it is difficult to create quickly by sucking the air out through a port using a pump. When evacuating the cavity, there is a point at which there are not enough air molecules left to create a so-called “viscous flow” of air through the port.

All the pump can do at this point is create a lower pressure zone on one side of the port and wait for the remaining molecules to pass through the port by chance as part of their regular, random motion. That means it can take as long as eight hours to achieve the vacuum needed in a VIG cavity.

Until this engineering challenge can be resolved, there is little prospect of mass-producing VIGU. 

Stark thinks he has a way to perform final assembly of the seal in a vacuum chamber, but this has yet to be tested.

Here goes nothing
EverSealed Windows has recently come to the end of its government grant from the U.S. Department of Energy to research and develop a VIG that would enable a whole window to achieve R-10. EverSealed, demonstrated to its collaborators and the DOE a proof-of-concept prototype VIG with a center-of-glass U-factor of 0.07 (R-13).

This VIGU’s edge-effect thermal conductivity is minimal due to the thinness and design of the flexible seal system, producing less than a two-degree C difference from the center of surface 4’s temperature to its edge temperature for any lite 1 temperature from -60 C to 130 C. 

Pro-VIG’s German government funding was consumed by 2009 but this program continued process and equipment development until last year, when its program manager Wolfgang Friedl retired from the program’s primary equipment developer Grenzebach Maschinenbau. Guardian has been promising a commercially available product for some time now, but is apparently not quite there yet. VIG research has always proceeded in fits and starts, and right now it looks like the Great Recession has had its toll on research budgets and progress.

However, as Cooper has noted, the “new normal” in energy standards will create an inexorable pressure on fabricators to find ways to become more energy efficient without sacrificing esthetic appeal.

The Insulating Glass Manufacturers Alliance is trying to help. Under Cooper’s chairmanship, the VIG task group is close to releasing a white paper that will give the industry guidance on the various VIG technologies available and some proposed best practices for constructing and applying this revolutionary style of IG.