Vacuum glazing support pillars and edge seals: the two parts that make VIG work
Almost everything remarkable about vacuum insulated glazing — the way it matches a modern double-glazed unit in the thickness of a single pane — comes down to two small, unglamorous components: a grid of microscopic support pillars scattered across the cavity, and a hermetic seal running around the edge. They are what make a hard vacuum survivable inside a window. They are also, between them, what limits how well that window can ever perform. Understand these two parts and you understand why VIG is both a genuine leap and a technology still working against its own physics.
This is the most technical of our guides, so it helps to start with the force everything else has to resist.
Ten tonnes on every square metre
Pump the air out of the gap between two panes of glass and you do not simply create a still, empty cavity. You create a pressure imbalance. Inside sits a near-perfect vacuum — the working target is around 0.1 Pa, roughly a hundred-thousandth of normal atmospheric pressure. Outside, the atmosphere presses in at about 100 kPa. That difference has to go somewhere, and it goes straight into the glass.
The numbers are startling. A pressure differential of roughly 100,000 Pa across a square metre of pane works out to about 10 tonnes per square metre pushing the two sheets of glass together — the weight of a couple of family cars spread evenly over a pane you could carry under one arm. Without something holding them apart, the panes would slam shut and the vacuum would be gone.
That “something” is the support pillar array. Every other design decision in a vacuum unit — pillar size, spacing, material, and the way the edge is sealed — is a response to this single, relentless load.
The atmosphere is not a gentle presence around a vacuum unit. It is a constant ten-tonne press, and the pillars and seal exist to hold the line against it for decades.
What a support pillar actually is
A support pillar — you will also see them called microspacers — is a tiny disc or short cylinder of hard material, set between the two panes to keep the cavity open. They are far smaller than most people expect.
A typical pillar is around 0.3 to 0.5 mm in diameter and just 0.1 to 0.2 mm high — the height being, in effect, the width of the vacuum gap itself. Pilkington’s Spacia uses pillars 0.5 mm across. At that scale a single pillar is about the thickness of a couple of sheets of paper, and it has to carry its share of that ten-tonne load without crushing, indenting the glass, or working loose.
Materials: strength first, then conductivity
The compressive stress on each pillar is enormous — well over 1 GPa at the contact point — so the material has to be genuinely hard. Two families dominate:
- High-strength metal, most commonly a stainless or high-strength steel, chosen for sheer mechanical resilience.
- Ceramic, which sacrifices a little manufacturing convenience for a useful thermal advantage (more on that below).
There is also a notable outlier. Panasonic developed a transparent pillar material, engineered to withstand around 200 MPa while being far less visible than a metal spacer — a direct attempt to design the pillar’s one cosmetic drawback out of existence.
Spacing, and the annealed-versus-tempered divide
Pillars are laid out on a regular square grid. The classic figure, used by Spacia, is a grid of about 20 mm — which means a one-square-metre unit carries roughly 2,600 pillars. That is a lot of tiny contact points, and, as we will see, each one is a small hole in the insulation.
Spacing is not a free choice, though. Push the pillars further apart and each pane has to span a wider unsupported gap under that ten-tonne load, so the bending stress in the glass rises. Annealed (ordinary) float glass can only tolerate so much, which historically forced tight grids. Fully tempered glass is far stronger, so it permits a wider array — tempered units such as LandVac use a much sparser grid, on the order of 267 pillars per square metre, roughly a quarter of the density of an annealed unit.
That difference is not cosmetic. Fewer pillars mean fewer thermal bridges, which is one reason tempered VIG reaches lower U-values than the original annealed products. But it only becomes possible if the manufacturing route keeps the glass tempered — and that, as we will come to, is a question about the edge seal, not the pillars.
The pillars are a thermal bridge — and that is the point
Here is the tension at the heart of the design. The vacuum is superb at stopping heat: it removes gas conduction and convection almost entirely, which in a normal sealed unit accounts for something like 70% of the heat lost. But the pillars are solid, and solids conduct heat readily. Each pillar is a thermal bridge — a direct little conduction path spanning the gap the vacuum was meant to interrupt.
Once the gas is gone and a good low-E coating has knocked down radiation (the topic of our how VIG is made guide), the two residual heat paths that remain are radiation and pillar conduction. The pillar array is not a manufacturing afterthought; it is one of the two things standing between current VIG and a theoretically perfect vacuum window.
This is exactly the problem the University of Sydney group set out to quantify. Their 1998 paper Heat conduction through the support pillars in vacuum glazing (Wilson, Simko and Collins) did the underlying maths on how much heat each spacer leaks, and it remains the reference point for the field — you will find it, along with the companion edge-conduction work, on our science page.
Designers have two levers to pull, and both involve compromise:
- Fewer pillars. Widening the grid cuts the number of bridges, but raises glass bending stress — which is why it depends on tempering.
- Lower-conductivity pillars. Swapping metal for a ceramic spacer can cut the conductance of each individual pillar by around 20 to 30%, shaving the U-value without changing the layout.
Neither lever is free, and neither eliminates the bridge. A vacuum unit’s centre-of-glass U-value is, in large part, a negotiation between how few pillars you dare use and how strong your glass and seal will let you be. Our guide to VIG U-values traces where the numbers land — from the classic Spacia at around 1.1 W/m²K to the best tempered units near 0.4.
Can you actually see them?
Yes — if you look for them. Against a plain bright background, the pillar array shows up as a faint, perfectly regular pattern of tiny dots, like a barely-there grid printed on the glass. At normal viewing distances, in a real window with the world behind it, they are unobtrusive and most people stop noticing them within days. Transparent-pillar designs like Panasonic’s reduce even that. It is worth being honest that the dots exist, though: a very close inspection of any VIG unit will reveal them, and they are a fair trade for the performance, not a defect.
The edge seal: holding a vacuum for decades
If the pillars fight the atmosphere across the face of the unit, the edge seal fights it around the perimeter — and it has a harder brief. It must be hermetic, holding the internal pressure near 0.1 Pa not for a warranty period but ideally for the life of the window, with no gas creeping back in through the join. A double-glazed unit that loses a little argon over the years simply gets slightly less efficient. A vacuum unit that loses its vacuum stops being vacuum glazing at all.
Two quite different sealing philosophies have emerged, and the choice between them cascades through the whole product — including whether the glass can be tempered, how the unit handles temperature swings, and how long it is likely to last.
Rigid glass-frit seals
The original approach, used in Pilkington Spacia, is a glass-frit (or solder-glass) seal. A low-melting-point glass paste is laid around the edge and the whole assembly is fired in a furnace at roughly 350 to 450 °C, fusing the two panes together with a band of solid glass. This is the seal you can often infer from the small pump-out port and its protection cap in the corner of a Spacia unit — see our materials overview for where those features sit.
A frit seal has real virtues. It is entirely inorganic, strong, and extremely long-lived — there is no organic sealant or desiccant to perish, which is one reason a well-made VIG edge can in principle outlast a conventional IGU. But it has two serious drawbacks:
- It de-tempers the glass. Firing at 350–450 °C is hot enough to anneal (relax) tempered glass, undoing its toughening. That is precisely why classic frit-sealed VIG had to be built from annealed panes — and annealed glass is both weaker and not a safety glass.
- It is rigid and brittle. A stiff band of glass cannot flex. When the two panes want to move relative to one another, the seal has almost no give, so it concentrates stress rather than absorbing it.
Flexible metal and indium seals
The newer approach bonds the edge with a metal or metal-alloy seal at much lower temperature — around 250 °C for the indium-based seals used in early research, and comparably low for the proprietary systems in current tempered products (Panasonic’s low-temperature sealing material, LandVac’s flexible lead-free alloy). Two advantages follow directly:
- The glass stays tempered. Because the seal is formed well below the temperature that would relax it, the panes can be fully toughened safety glass. This is the single enabling step behind modern tempered VIG — its wider pillar arrays, lower U-values, and safe-break behaviour all trace back to a seal that does not need a 450 °C furnace.
- It flexes. A metal seal can accommodate a degree of movement — thermal expansion, wind load, the panes bowing under pressure — where a frit seal would crack. It absorbs the differential rather than fighting it.
The two families are worth setting side by side.
| Rigid glass-frit seal | Flexible metal / indium seal | |
|---|---|---|
| Sealing temperature | ~350–450 °C (fired in furnace) | ~250 °C or lower |
| Effect on glass | De-tempers — historically forced annealed panes | Glass stays fully tempered (safety glass) |
| Mechanical behaviour | Rigid, brittle; concentrates stress | Flexes; absorbs movement and expansion |
| Thermal-stress cracking | More prone, especially with big temperature differences | Markedly reduced |
| Longevity character | Fully inorganic, very durable if uncracked | Durable; better tolerance of real-world movement |
| Example products | Pilkington Spacia | LandVac, HaanGlas |
Note the deliberate gaps in that table. Some commercial seals are proprietary and undisclosed — AGC’s Fineo, for instance, uses a seal it does not publicly characterise and has no visible pump-out port — so we do not slot every product into one column. The point is the two principles, not a claim about every unit on the market.
Why the seal decides service life
The reason the seal choice matters so much is a failure mode called thermal-stress cracking, and it is the main durability question hanging over vacuum glazing.
Picture a cold winter day. The outer pane sits near freezing; the inner pane, warmed by the room, might be 20 °C or more above it. Glass expands when warm, so the two panes want to change size by different amounts — the warm one stretches, the cold one does not. In a vacuum unit they are locked together at the edge and studded together across the face by the pillars, so they cannot move freely. The result is differential expansion: the panes try to bow and shear against one another, and that stress has to be carried by the seal and the glass.
A rigid frit seal is exactly the wrong thing to hand that stress to. Being brittle and unyielding, it concentrates the strain at the edge, and if the temperature difference is large enough — very cold climates are the classic trigger — the seal or the glass can crack. The tolerances are unforgiving: the long-term stress that annealed float glass can safely carry for a 25-year life is only around 8 to 20 MPa, so there is not much headroom. This is a well-enough understood risk that there is a dedicated international test method for it, ISO 19916-3:2021, which specifically evaluates rigid-edge-sealed VIG under temperature differences.
A flexible metal seal, paired with tempered glass, attacks the problem from both ends: the seal gives a little instead of cracking, and tempered glass tolerates far higher stress than annealed. That combination is why tempered, low-temperature-sealed VIG is considered better suited to harsh climates and large panes.
The link between sealing method and how long a unit lasts is not just intuition. The Empa durability work by Koebel and colleagues (2010) modelled how cavity pressure creeps up over time through outgassing, leakage and permeation, and tied that ageing directly to the sealing approach — a transient pressure balance that effectively sets the unit’s service life. And on the research frontier, Memon and Eames developed a low-temperature hermetic composite edge seal (a Cerasolzer-and-epoxy system curing below 200 °C) that lets delicate soft low-E coatings survive fabrication — an advance aimed squarely at triple vacuum glazing, where a prototype reached about 0.33 W/m²K centre-of-pane. Both papers are on our science page for readers who want the primary sources.
The seal is not just a way to keep air out. It is the component that most directly governs whether a vacuum unit survives twenty British winters — and the shift from rigid frit to flexible metal is the single biggest reason modern VIG can be tempered, safe, and durable.
The bottom line
Support pillars and the edge seal are the two components that make vacuum glazing physically possible, and the same two that hold it back. The pillars win the fight against ten tonnes of atmospheric pressure per square metre, but every one of them is a small thermal bridge — so, with radiation, they set the floor on how low a VIG U-value can go. Widen the array or switch to ceramic and you claw back a little performance, but only if the glass and seal will let you. The edge seal, meanwhile, decides almost everything downstream: a rigid glass-frit seal is durable and inorganic but brittle and de-tempers the glass, while a flexible low-temperature metal seal keeps the glass tempered and flexes with the thermal movement that would otherwise crack it. That is why the industry has moved decisively toward tempered, low-temperature-sealed units.
For a homeowner, none of this needs to be memorised — but it does explain the specification. When one vacuum unit quotes a lower U-value, tolerates a colder climate, or carries a safety-glass rating and another does not, the answer usually lies in these two small parts: how many pillars, made of what, and how the edge was sealed.