At What Temperature Does Glass Crack From Heat?

Ordinary window glass can crack from a temperature difference of roughly 80 °C (about 150 °F) between one area of the pane and another, not an absolute temperature on a thermometer. That distinction matters enormously: glass does not have a single cracking temperature the way water has a freezing point. The number shifts depending on the type of glass, the condition of its surface, how quickly the heat arrives, and even the humidity in the air. Borosilicate glass, the kind used in lab equipment and some cookware, withstands a temperature swing roughly twice as large before it fails. Tempered glass pushes the threshold even higher. Understanding what actually triggers the crack, rather than memorizing a single number, is what keeps a baking dish or a car windshield from surprising you.

Why Temperature Difference Matters More Than Absolute Temperature

Glass does not crack because it reaches some magic number on a thermometer. It cracks because different parts of the same piece are at different temperatures at the same moment. When one region is hot and an adjacent region is cold, the hot zone tries to expand while the cold zone resists that expansion. The cold zone effectively pins the hot zone in place, generating tensile stress at the boundary. Once that stress exceeds the strength of the glass at its weakest point, a crack forms.

This is thermal shock, and the physics have been studied for the better part of a century. Classic analyses of the stresses caused by thermal gradients in plates and cylinders show that the maximum tension often appears immediately at the surface exposed to the temperature change, from pure stretch, and can be larger than the bending stresses that develop later.¹ In plain terms, the very first instant of a sudden temperature change is often the most dangerous, because the surface gets all the stress before the heat has had time to spread inward.

Because thermal shock depends on a temperature difference rather than an absolute value, the same glass pane that survives sitting in a 200 °C oven for an hour can shatter if you splash cold water on it. The oven heated it slowly and uniformly, so no significant gradient ever built up. The cold water created a steep gradient in a fraction of a second. The material itself was fine at the high temperature; the sudden, lopsided cooling was the killer.

Threshold Numbers for Common Glass Types

For ordinary soda-lime float glass, the kind found in the vast majority of house windows, a theoretical value of about 80 °C of temperature difference is widely cited as the threshold for initial fracture.² That means if one region of the pane is at room temperature (around 20 °C) and another region reaches 100 °C, you are at the danger line. In practice, partially shaded window panes on a sunny day can build up gradients approaching this range, which is why thermal cracking in architectural glazing is a real design concern, not just a laboratory curiosity.

Borosilicate glass performs substantially better. Testing of BK-7 optical-grade borosilicate found that the critical temperature drop for catastrophic failure ranged from about 124 to 140 °C, depending on the surface finish.³ That is roughly 50–75 percent more headroom than ordinary window glass, which explains why Pyrex-style borosilicate bakeware can handle oven-to-countertop transitions that would shatter a regular glass dish.

Tempered (toughened) glass, the type used in car side windows and shower doors, is stronger still. The tempering process locks compressive stress into the surfaces, so a thermal gradient has to overcome that built-in compression before it can put the surface into tension. Tempered glass can generally tolerate temperature differences of around 150–200 °C before cracking, though the exact number depends on the tempering level and pane geometry. When tempered glass does finally crack, it shatters into small, relatively harmless cubes rather than long shards.

At the far end of the spectrum, quartz glass (fused silica) is in a different league. Its thermal expansion is extremely low, so even large temperature differences produce comparatively little stress. Research on quartz glass exposed to extreme thermal radiation found that at a surface temperature equivalent to roughly 3,500 K sustained for over fifteen minutes, a 10 mm thick pane developed a softening zone of about 5.3 mm, while at that temperature for just three seconds the softening zone was only 0.14 mm.⁴ At those extreme exposures, quartz softens rather than cracking, because the material’s low expansion means thermal shock stress never becomes the dominant failure mode. This is why fused-silica crucibles survive furnace temperatures that would pulverize other glasses.

Surface Condition and Edge Quality

The threshold numbers above assume laboratory conditions: clean samples with well-characterized surfaces. In real life, surface flaws are the wild card. Every glass surface carries microscopic scratches, chips, and pits from handling, cutting, and everyday use. These flaws act as stress concentrators. A crack does not need to overcome the theoretical strength of flawless glass; it only has to exceed the much lower effective strength around the tip of an existing flaw.

The borosilicate study mentioned above illustrated this directly. Samples lapped with coarser abrasives (larger grit, rougher surface) had lower thermal shock resistance, while finer-finished surfaces survived larger temperature swings before failing.³ The critical temperature drop shifted by more than 15 °C just from changes in surface roughness, all on the same glass composition.

Edge condition tells a similar story for window glass in buildings. Research on float glass in fire conditions found that ordinary as-cut edges actually performed better than ground or polished edges under thermal loading.⁵ That sounds counterintuitive, since ground or polished edges seem like they should be smoother and therefore stronger. The likely explanation is that grinding and polishing can introduce a population of subsurface micro-cracks oriented in ways that happen to align badly with the thermal stress field at the pane’s edge. The takeaway for anyone working with glass: a pristine-looking edge is not always a strong edge.

For your kitchen, this means an old baking dish with years of tiny scratches will be less tolerant of thermal shock than the same dish was when it was new. A chip on the rim of a glass casserole dish is not just cosmetic; it is a stress riser waiting for the right temperature gradient.

How Quickly Heat Arrives

Speed of heating changes the equation dramatically. When heat arrives slowly, it has time to conduct through the glass, keeping the temperature fairly uniform across the thickness. The gradient stays small and the stress stays low. When heat arrives rapidly, the exposed surface jumps in temperature while the interior and far side remain cool, generating a steep gradient through a very thin layer.

Work on thermal shock and crack propagation in glass panes found that under a rapid rate of temperature rise, a smaller overall temperature increase could produce a bigger dynamic stress and still initiate cracking.⁶ The maximum stress concentration appeared at the boundary between the hot and cool layers within the glass, which is where the crack started. So it is not just the total temperature difference that matters but how quickly that difference develops. A 60 °C difference applied in two seconds can be more dangerous than a 100 °C difference applied over five minutes.

This is the core reason why pouring boiling water into a cold glass is so risky while heating the same glass gradually in an oven is usually safe. In the boiling-water scenario, the inner surface of the glass might reach 80–90 °C almost instantly while the outer surface is still near room temperature. The gradient through the wall of the glass spikes to a level that exceeds the failure threshold before the heat has a chance to even out.

Thickness, Geometry, and Where the Gradient Lives

You might expect thicker glass to be tougher, and in some ways it is: thicker glass can absorb more total energy before it heats through. But thickness also changes the stress picture. A study examining temperature gradients in glass panes under thermal radiation found that the gradient through the glass thickness was much larger than the gradients in the planar directions (left-to-right or top-to-bottom). However, the increase in stress and the cracking behavior were mainly driven by the planar temperature gradients, not the through-thickness gradient.⁷

In practical terms, this means that for a window pane exposed to fire or intense sunlight on one side, what drives cracking is not so much the difference between the hot face and the cool face (through-thickness), but the difference between the heated center of the pane and its cooler edges hidden behind the frame (planar). The frame shades the edges, keeping them cool while the exposed center heats up. That planar mismatch is what generates the tensile stress at the edge that eventually starts a crack. It is also why cracks in thermally stressed windows almost always start at the edge and run inward, rather than starting in the center of the pane.

Geometry matters in another way: curved glass, like that found in some architectural facades and automobile windshields, distributes stress differently than flat panes. A curve can either help or hurt depending on whether the thermal gradient works with or against the existing shape stresses. Laminated windshields, which sandwich a polymer interlayer between two sheets of glass, have their own behavior: the inner sheet may crack first because it faces the cabin-side temperature while the outer sheet faces the weather.

Humidity and Stress Corrosion

This one catches people off guard. Moisture in the air can make glass more vulnerable to cracking, even at temperatures that would normally be safe. Water molecules attack the silicon-oxygen bonds at the tip of a flaw under stress, a process known in materials science as stress corrosion cracking. It means that a flaw under moderate thermal stress might slowly grow into a full crack if the air is humid, even though the same stress in dry air would not have been enough.

Research on crack propagation during scribing of glass sheets found that the median crack growth rate increased with relative humidity, consistent with moisture-assisted stress corrosion.⁸ While that study focused on mechanical scribing rather than thermal shock, the underlying mechanism is the same: water at the crack tip accelerates bond breaking under stress. In a thermal shock scenario, this means that a glass pane on a muggy summer day might fail at a slightly lower temperature difference than the same pane on a dry winter day, all else being equal. The effect is not dramatic enough to halve the threshold, but it can shave a meaningful margin off an already borderline situation.

Manufacturing Processes That Raise the Bar

Much of modern glass engineering is about pushing the thermal shock threshold higher. Two main strategies dominate: reducing the residual stress left over from manufacturing, and deliberately building beneficial compressive stress into the surface.

Annealing addresses the first strategy. When glass cools from its molten state, the outside solidifies before the inside, locking in internal stresses. A well-controlled annealing process reheats the glass and then cools it very slowly, allowing those stresses to relax. Research on large silica glass pieces showed that both the annealing temperature and the time spent at temperature significantly affect the residual stress, with higher temperatures and longer holds reducing it further. A two-step annealing approach, using a higher-temperature stage followed by a lower one, reduced both the overall residual stress and the stress peaks more efficiently than a single-stage anneal.⁹ Lower residual stress means more of the glass’s theoretical strength is available to resist thermal shock rather than being eaten up by internal tension left over from cooling.

Tempering takes the opposite approach. Instead of eliminating internal stress, it deliberately creates a stress profile where the surfaces are in strong compression and the interior is in tension. Because thermal shock generates surface tension, and the surface is already pre-compressed, the thermal stress has to cancel out all the built-in compression before it can start pulling the surface apart. This is why tempered glass tolerates temperature differences roughly two to three times larger than annealed glass of the same composition.

Chemical strengthening works on a similar principle but through a different route: the glass is soaked in a molten salt bath, and larger ions from the salt replace smaller ions near the glass surface, squeezing the surface layer into compression. Chemically strengthened glass is commonly used in smartphone screens and aircraft windows. Its thermal shock resistance falls somewhere between ordinary annealed and fully tempered glass, depending on the depth and magnitude of the compressive layer.

The Role of Material Properties Beyond Expansion

When engineers assess how well a glass will resist thermal shock, they do not look at a single number. Analysis of thermal shock resistance in brittle solids identifies a cluster of properties that all contribute: the thermal expansion coefficient, the thermal conductivity, the elastic modulus, and the fracture toughness, among others, along with external factors like the heat transfer rate and the size of the piece.¹⁰ A glass with low expansion generates less stress per degree of temperature difference. A glass with high thermal conductivity spreads heat faster, reducing the gradient. A glass with high fracture toughness can tolerate more stress before a flaw starts to grow. The interplay among all of these is why two glasses with the same expansion coefficient can still have very different cracking thresholds.

This also explains why some glass-ceramics, like those used in cooktop surfaces, can go from a red-hot burner to a wet sponge without cracking. These materials are engineered to have near-zero thermal expansion, which means almost no stress develops regardless of the gradient. The tradeoff is that engineering such low expansion into a transparent or translucent material is difficult and expensive, which is why your kitchen windows are not made of the same stuff as your stovetop.

Common Real-World Scenarios

Knowing the principles helps, but most people encounter glass cracking in a handful of specific situations. Here is how the physics play out in each:

• Baking dishes: Most modern glass bakeware is made of tempered soda-lime glass, not borosilicate. It can handle the oven-to-countertop transition if you set it on a dry towel or trivet, but placing it on a cold, wet surface creates a sudden cooling gradient on the bottom face that can exceed the threshold. Chipped or scratched dishes are at higher risk.
• Window panes in fires: The edge-to-center gradient drives failure. In room fire conditions, ordinary float glass panes typically crack within the first few minutes as the exposed center heats rapidly while the frame-shaded edges stay cool. The initial cracks run from the edge inward.
• Pouring boiling water into a glass: A thin-walled drinking glass can crack from a temperature difference well under 80 °C if it has surface flaws. Pre-warming the glass with lukewarm water reduces the gradient and lowers the risk significantly.
• Car windshields in winter: Laminated automotive glass is more resistant than single-pane window glass, but an existing chip or crack from a stone impact acts as a stress riser. Blasting the defroster at full heat on a sub-zero morning creates a localized gradient right where the heated air hits the glass, which can turn a small chip into a spreading crack.
• Glass cooktops: These are made from glass-ceramics with near-zero expansion, so thermal shock from cooking is essentially a non-issue. The danger is mechanical: dropping a heavy pan can chip the surface, and that chip can become a crack origin if the cooktop later experiences an unusual thermal event.

When Softening Replaces Cracking

At high enough temperatures, glass stops behaving as a brittle solid and starts to soften. Soda-lime glass begins to soften around 700 °C, while borosilicate softens around 820 °C and fused silica around 1,665 °C. Once the glass reaches its softening point, it deforms rather than cracking, because it can flow to relieve the stress. This is why glassblowers can heat glass to orange-hot temperatures without it shattering: the material is soft enough that thermal gradients simply cause it to warp rather than fracture.

The extreme-temperature research on quartz glass showed this transition clearly. Under sustained temperatures equivalent to roughly 5,000 K, a 10 mm quartz pane softened completely through its thickness in under two minutes.⁴ At those levels, the failure mode is not cracking but deformation and eventual melting. For everyday purposes, though, you will hit the cracking threshold long before you approach the softening point. The danger zone is the range where glass is still rigid enough to be brittle but hot enough (or, more precisely, unevenly hot enough) to be stressed past its breaking point.