ASTM C1048, Heat-treated flat glass
Heat-strengthened and fully tempered glass come off the same furnace, split apart by one thing: how hard you quench them. ASTM C1048 is the standard that draws the line, spelling out the surface compression, break pattern, and tests each has to hit before you are allowed to call it by either name.
Two kinds of heat-treated glass
Same furnace, same heating cycle, same glass going in. The one thing that separates the two products is how hard you blast the glass with air on the way out, and how much surface compression that locks in permanently:
- Surface compression: 3,500–7,500 psi (24–52 MPa)
- Approximately 2× the load resistance of annealed glass
- Fractures in large irregular pieces similar to annealed (not safety glazing)
- Lower NiS risk than fully tempered
- Suitable for spandrel, overhead structural applications where large-fragment breakage is preferred
- Surface compression: ≥10,000 psi (≥69 MPa)
- Approximately 4× the load resistance of annealed glass
- Fractures into small, relatively blunt fragments, meets safety glazing requirements
- Required by building codes in hazardous locations (per CPSC 16 CFR 1201 and ANSI Z97.1)
- Higher NiS risk, heat soak testing recommended for overhead or high-consequence use
Surface compression testing methods
Compression is the whole ballgame, so C1048 sets minimum values and accepts two ways of proving you hit them:
| Method | Principle | Notes |
|---|---|---|
| Surface stress measurement (GASP) | Scattered light polariscope measures stress optically at the glass surface | Non-destructive; can be performed on production lites |
| Fragmentation test | A sample lite is fractured and the fragment count in a defined area (50mm × 50mm) is counted | Destructive sampling; FT glass must have ≥40 fragments in 50mm × 50mm area; HS glass has no minimum fragment count requirement |
Bow, warp, and surface flatness
Heat treating bends glass slightly, and there is no way around it. The lite goes soft in the furnace and sags a fraction under its own weight before it sets again. So C1048 does not pretend the glass comes out dead flat. It allows bow, scaled to the length of the lite, and allows noticeably more of it than C1036 does for annealed glass. That gap is not the standard going easy on anyone; it is an honest acknowledgement of what the process physically does to glass.
Roller wave is a different beast, handled separately. It is the gentle repeating ripple the furnace rollers press into the glass, and head-on it is close to invisible. Stand at an angle on a bright day, though, and the reflections go wavy right across the elevation. That is why it gets judged optically rather than with a gauge on the bench. Good heat treaters chase it hard through furnace control and roller setup, because on a large facade it is the defect everyone spots from the car park.
Coated and processed glass
Coatings complicate the order of operations but not the requirements. Hard-coat pyrolytic coatings go through tempering fine; soft-coat sputtered ones generally do not, which is why they get applied afterward or deleted at the edges. Either way, C1048 still governs the heat treatment result, and the coating does not soften the compression numbers you have to hit. Spandrel glass falls under C1048 too when the substrate needs heat treating, since the ceramic frit is fired on and the glass is heat treated in the same pass.
Edge compression
Edges get their own requirements, because the stress profile at an edge is not the same as the one across the face. This matters most on fully tempered structural glass with exposed edges, whether polished, beveled or ground. The rule that catches people out is one of sequence: the edge has to be finished before the glass is tempered. Do it in the wrong order and every chip and grind mark you leave behind becomes a stress concentration sitting exactly where the glass is most likely to start a fracture. See the edge finishing article for fabrication details.