Glass structural design & thickness
Glass is not a material you pick from a catalogue - it is a structural plate you size, like a beam, to carry the wind without breaking or bowing too far.
The same wind that needs 6 mm of glass at the centre of a wall can need 12 mm at the corner - and choosing wrong is a panel that explodes in a storm.
Most people think glass thickness is about looks or insulation. For a facade engineer it is a structural calculation as real as sizing a beam. A pane of glass spanning between its four edges is a **plate in bending**: the wind pushes it into a shallow dome, tension builds on the back face, and if that tension passes what the glass can bear, it breaks - suddenly, completely, and from the surface flaws every piece of glass carries. So we size glass to the wind pressure from Lesson 4.1, against the deflection limit from Lesson 4.2, choosing the make-up - thickness, heat treatment, lamination, single pane or IGU - that resists the load with margin. This is where glass stops being a window and becomes a structural element you are responsible for.
Glass as a structural plate - strength, stiffness and the load-resistance method
A plate in bending, limited by surface tension - and by how far it bows
A glass pane fixed on four edges and pushed by wind behaves as a thin plate in bending. Two things can fail it. Strength: the wind creates tensile stress on the back (leeward) face, and glass - strong in compression, weak and brittle in tension - breaks when that stress exceeds its allowable. Because glass strength comes from microscopic surface condition, it is probabilistic: design is to an accepted probability of breakage, conventionally 8 lites per 1000 (0.008) at the design load. Stiffness: even unbroken, the pane must not bow too far - a deflection limit of roughly span/60 to span/90 of the short edge, or about 19-25 mm, protects the edge seals and the IGU.
Thickness drives both. A plate's bending stiffness goes with thickness cubed, so going from 6 mm to 8 mm raises stiffness by (8/6)-cubed ~ 2.4x. Its stress capacity rises roughly with thickness squared. This is why a small increase in glass thickness buys a large increase in wind resistance - and why corner panels, which see the worst suction, jump a thickness or two above field panels.
Glass is strong in compression, weak in tension, and breaks from its surface flaws. Size it for the back-face tension and the bow - and remember every pane is a little weaker than the last.
Annealed, heat-strengthened, toughened, laminated - and load resistance by glass type factor
How glass is processed changes its strength. Annealed float glass is the baseline (and breaks into dangerous shards). Heat-strengthened glass is ~2x stronger than annealed. Fully toughened (tempered) glass is ~4x stronger and breaks safely into small dice. Laminated glass bonds two plies with an interlayer (PVB/SGP) - it shares load between plies and, crucially, holds together when broken (it is the safety and overhead choice).
The global method for sizing is ASTM E1300, the load-resistance approach. You compute the glass's Load Resistance (LR) - the uniform pressure it can safely carry at the 0.008 breakage probability - from a Non-Factored Load (NFL) for the pane's size and thickness, multiplied by a Glass Type Factor (GTF): GTF ~ 1.0 for annealed, ~2.0 for heat-strengthened, ~4.0 for fully toughened. The design check is simply LR >= design wind pressure. Run both the strength check (LR vs load) and a deflection check, and pick the make-up that passes both with margin.
Two panes, one sealed air gap - and the load splits between them
Most modern facade glass is an insulated glass unit (IGU): two (or three) panes separated by a sealed, gas-filled cavity. Structurally an IGU is clever - because the cavity is sealed, a wind pressure on the outer pane compresses the trapped gas, which then pushes on the inner pane, so the two panes share the load. The split depends on relative stiffness (thickness cubed): for two equal panes the load splits roughly 50/50; for unequal panes the stiffer pane takes the larger share.
This load sharing is real benefit - two 6 mm panes in an IGU resist more than a single 6 mm pane - but it comes with traps. The sealed cavity also reacts to temperature and altitude changes (the gas expands and contracts, bowing the panes - this is why large IGUs can look slightly concave or convex). And the IGU's structural performance depends on the edge seal staying intact for decades. So an IGU is sized by sharing the design wind pressure between its panes per their stiffness, checking each pane against its own LR and deflection limit, while remembering the cavity's thermal breathing and the seal's finite life. Glass selection is never just a thickness - it is a make-up: type, treatment, lamination and IGU build, sized to the load.
Glass thickness is a cost and a look you partly control. Bigger panes and more exposed corners both push thickness up (and price with it). If you want very large, very thin, very flush glass, expect heat-strengthened or toughened lites, possibly laminated, and a structural-silicone detail - all of which the engineer sizes to your wind pressure. Specify the glass _performance_ (and safety - laminated where there is a fall risk) and let the engineer set the structural make-up; a thickness picked for looks alone is how panels end up under-designed at the corners.
Size glass by the load-resistance method (ASTM E1300, or the IS/EN equivalent): compute LR from the pane's NFL and glass-type factor and check LR >= design wind pressure at the 0.008 breakage probability, then check deflection (short-span/60 to /90, ~19-25 mm cap). For IGUs, split the wind between panes by relative stiffness (thickness cubed) and verify each pane and its seal, accounting for cavity thermal/altitude effects. Use heat-strengthened over annealed to beat thermal-stress breakage, and laminated where safety and post-breakage retention matter. Always design the worst corner/edge suction case, not the field panel.
Every lite on a facade has a specified make-up - thickness, heat treatment (annealed/HS/toughened), laminated or not, and the IGU build - and it is on the glass schedule for a structural reason. The corner and top-floor lites are often thicker or toughened; that is deliberate. Never swap a marked lite for a thinner or cheaper one, never store toughened glass so its edges chip (a chipped edge is a weak point), and check the IGU edge seal is intact - a broken seal fogs the unit and, over time, weakens the structural sharing the engineer counted on.
ASTM E1300
Glass thickness / load resistance
The dominant load-resistance method - gives Non-Factored Load and Glass Type Factors to find the load resistance at a 0.008 breakage probability. It is for rectangular vertical glass under uniform load; complex shapes or point fixings need FE analysis.
IS 2553 / IS 16231 / IS 875-3
Safety glass & loads (India)
IS 2553 covers safety glass (toughened/laminated); IS 16231 covers fenestration performance; the wind pressure to design against comes from IS 875-3. India has no single E1300-equivalent design chart, so E1300 or EN 16612 is widely used on premium projects.
EN 16612 / EN 16613
Glass design (Europe)
The European method for lateral load resistance of glass and the interlayer behaviour of laminated glass. Differs from E1300 in its partial factors and load-duration treatment - don't mix the two methods in one calculation.
ASTM E2188 / EN 1279
IGU durability
Govern insulated-glass-unit testing and the edge-seal life on which an IGU's long-term load sharing depends. They prove durability, not structural capacity - the structural check is still E1300/EN 16612.
“Thicker glass is stronger, so for safety just specify the thickest glass you can afford everywhere.”
Thickness helps, but it is the wrong lever to reach for first - and uniform 'thickest everywhere' is wasteful and still unsafe in the wrong way. Glass strength is mainly governed by **heat treatment** (toughened is ~4x annealed) and by **breakage behaviour** (laminated holds together; annealed rains shards). A correctly chosen 8 mm heat-strengthened laminated lite can be both stronger and far safer than a thick annealed monolithic pane. Size by the load-resistance method to the actual local wind pressure, choose the type and lamination for strength and safety, and vary thickness by zone - don't blanket the building in thick annealed glass.
Worked example - select glass thickness for a curtain-wall lite
We have a wind pressure and a deflection limit. Now choose a glass make-up that carries it - the load-resistance check at the heart of glass design.
The design wind pressure (Lesson 4.1), ASTM E1300 NFL charts and glass type factors (or EN 16612), the pane geometry from the facade grid, and the glass supplier's make-up options.
GIVEN (a vision lite in the curtain wall, four edges supported):
Pane size a x b = 1.5 m x 2.0 m (3.0 m2)
Design wind (ULS) p = 3.0 kPa (factored corner case, Lesson 4.1)
Breakage probability 8 lites / 1000 (0.008) per ASTM E1300
Deflection limit short-span / 60 (short edge = 1.5 m)
Method (ASTM E1300, load resistance):
LR = NFL x GTF must satisfy LR >= p
Glass Type Factors: annealed 1.0 . heat-strengthened ~2.0 . toughened ~4.0
Trial NFL (6 mm pane, 1.5 x 2.0 m, from E1300 charts): NFL ~ 1.6 kPa- 1Try 6 mm annealed. LR = NFL x GTF = 1.6 kPa x 1.0 = 1.6 kPa. Required = 3.0 kPa. 1.6 < 3.0 -> FAILS. Annealed at 6 mm cannot carry this corner pressure.
- 2Try 6 mm heat-strengthened. Same pane, GTF ~ 2.0: LR = 1.6 x 2.0 = 3.2 kPa. Required = 3.0 kPa. 3.2 > 3.0 -> PASSES strength with a slim ~7% margin.
- 3Or try 8 mm. A thicker pane raises NFL (stiffness scales with thickness cubed). For 8 mm, NFL ~ 2.6 kPa; annealed LR = 2.6 kPa (still < 3.0, fails), but heat-strengthened LR = 2.6 x 2.0 = 5.2 kPa - comfortable margin.
- 4Check deflection. Limit = short span / 60 = 1500 / 60 = 25 mm. The 8 mm heat-strengthened pane under 3.0 kPa deflects on the order of ~15-20 mm (from the E1300 deflection chart) - within 25 mm. The 6 mm pane, near its strength limit, would bow more and sit closer to the cap.
- 5Account for the IGU. This lite is actually an IGU (6 mm outer + 12 mm cavity + 6 mm inner, both heat-strengthened). The wind splits ~50/50 between equal panes, so each pane carries ~1.5 kPa - well inside a single 6 mm HS pane's 3.2 kPa LR. The IGU's load sharing is what lets two modest panes do the job.
- 6Select and record. Choose a 6 mm HS / 12 mm cavity / 6 mm HS laminated-inner IGU for the corner zone: strength margin secured by heat strengthening + load sharing, safety by lamination of the inner pane. Field (mid-wall) lites, seeing perhaps 1.5 kPa, can drop to 6 mm annealed or thinner - vary the make-up by zone.
- 7Sanity-check thermal stress. A dark or partially shaded corner lite risks thermal-stress cracking; the heat-strengthening chosen for wind also covers that - one decision solving two problems.
You’ll walk away with
A defensible glass selection - a 6/12/6 heat-strengthened, laminated-inner IGU for the 3.0 kPa corner zone passing both the ASTM E1300 strength check (LR 3.2 kPa > 3.0 kPa) and the deflection limit (~25 mm) - plus the principle that you vary glass make-up by wind zone rather than blanket the building in one thickness.
Two quick glass checks.
- 01Take the field-panel pressure (~1.5 kPa) and confirm a single 6 mm heat-strengthened lite (LR ~3.2 kPa) passes with a 2x margin - showing why the same building uses thicker glass only at the corners and top floors.
- 02For an IGU of unequal panes (6 mm outer, 10 mm inner), estimate the load split by thickness cubed: outer share ~ 6^3 / (6^3 + 10^3) = 216 / 1216 ~ 18%; the stiffer 10 mm inner takes ~82%. Stiffness, not position, decides the share.
Glass is a structural plate sized to the wind: check that its load resistance (ASTM E1300 - non-factored load times the glass type factor for annealed/HS/toughened) exceeds the design pressure at a 0.008 breakage probability, and that its bow stays inside ~span/60-90. In an IGU the sealed cavity shares the load between panes by their relative stiffness. Vary the make-up by wind zone, and choose treatment and lamination for strength and safety, not thickness alone.
Glass is a plate in bending: fails by back-face tension (probabilistic, design to 0.008 breakage) or by excess bow (~span/60-90, ~19-25 mm). Stiffness scales with thickness cubed. ASTM E1300 load-resistance: LR = NFL x GTF (annealed 1.0, HS ~2.0, toughened ~4.0), need LR >= design pressure. IGU panes share load by relative stiffness; the sealed cavity breathes thermally and the edge seal has finite life. Vary make-up by zone.
How do you choose glass thickness for wind load?
Treat the glass as a structural plate and use the load-resistance method (ASTM E1300, or EN 16612 in Europe). For the pane's size and a trial thickness, read the Non-Factored Load and multiply by the Glass Type Factor (annealed ~1.0, heat-strengthened ~2.0, toughened ~4.0) to get the Load Resistance at a 0.008 breakage probability, then require Load Resistance >= the design wind pressure from IS 875-3. Also check the deflection stays within roughly the short span over 60 to 90 (about 19-25 mm). Increase thickness or upgrade the heat treatment until both pass with margin.
How do the two panes of an insulated glass unit share wind load?
Because the cavity is sealed, wind pushing on the outer pane compresses the trapped gas, which pushes on the inner pane - so the two panes share the load rather than one carrying it all. The split follows relative stiffness, which scales with thickness cubed: two equal panes split roughly 50/50, while an unequal pair has the thicker, stiffer pane take the larger share. This sharing lets two modest panes resist more wind than a single pane of the same thickness, but the benefit depends on the edge seal staying intact for the unit's life.
What is the difference between annealed, heat-strengthened and toughened glass structurally?
They differ mainly in strength and breakage. Annealed is the baseline strength and breaks into large dangerous shards. Heat-strengthened is about twice as strong as annealed and breaks into larger fragments (good against thermal-stress cracking). Fully toughened (tempered) is about four times annealed and breaks safely into small blunt dice. In the ASTM E1300 method these map to Glass Type Factors of roughly 1.0, 2.0 and 4.0. Laminated glass, separately, bonds plies so the unit holds together when broken - the safety choice for fall risk and overhead glazing.
Peer-reviewed journals & authoritative standards
- 01Bedon, C. et al. Performance of structural glass facades under extreme loads - design methods, existing research, current issues and trends. Construction and Building Materials, 163. — Construction and Building Materials (Elsevier), 2018.
- 02Review on Glass Curtain Walls under Different Dynamic Mechanical Loads: Regulations, Experimental Methods and Numerical Tools. IntechOpen. — IntechOpen (peer-reviewed chapter), 2023.
- 03Li, X. & Wu, Y. A review of complex window-glazing systems for building energy saving and daylight comfort. — Journal of Building Physics (SAGE), 2025.
_The glass is sized, the frame is stiff enough, and the panel knows how to move. One question remains: how does all this load - dead weight down, wind suction out - actually reach the building? Through brackets, anchors and embeds. The load path to the slab, next._