Technology
Glass gets strong
by being squeezed.
Toughened glass isn't a thicker sheet — it's a sheet whose surface is permanently under compression. Here's how we put it there, and how we prove it worked.
The principle
Compression has to be beaten before a crack can start.
Glass almost never fails from the middle. It fails from a microscopic flaw at the surface that opens under tension and runs. Put the surface under permanent compression and that flaw has to be pulled open before it can propagate at all.
Ion exchange does this chemically rather than thermally. Soaking the sheet in molten potassium salt swaps small sodium ions in the glass for larger potassium ions. The larger ions don't fit, so they crowd the outer 40 microns into a permanently compressed skin.
The result is a sheet a third of a millimetre thick that takes a 128 gram steel ball from 1.2 metres without letting anything through to your display.
Cross-section
Six layers, 0.33 mm.
Hover any layer to isolate it. Total stack height is about a third of a millimetre — roughly three sheets of paper.
Total stack
0.33
millimetres, ±0.02 across the full sheet
Oleophobic nano-coating
0.02 mm
AF anti-fingerprint film
0.01 mm
Chemically toughened glass
0.21 mm
PET shatter-containment
0.02 mm
Optical silicone adhesive
0.04 mm
Removal liner
—
Process
Six stages, start to sealed box.
Roughly 19 hours from raw sheet to packed unit, most of it spent in the salt bath.
Float and cut
Aluminosilicate sheet is floated to an optical flatness of λ/4, then CNC-cut to the device's own CAD outline — cut-outs, radii and all.
Ion exchange
14 hours in a molten potassium salt bath at 420 °C. Larger potassium ions displace sodium at the surface and wedge themselves in, leaving the outer 40 microns under permanent compression.
Edge arc and polish
The perimeter is ground to a 2.5D arc and diamond-polished. A polished edge is not cosmetic — a rough edge is where 90% of glass fractures begin.
Lamination
PET containment film and index-matched silicone adhesive are laminated under vacuum, so there is no trapped air to scatter light or seed a bubble.
Nano-coating
A fluoropolymer oleophobic layer is vapour-deposited at 20 nm. Thin enough to be optically invisible, durable enough to survive years of thumbs.
Destructive QC
Samples pulled at random from every run go through drop, hardness, abrasion and bend testing. One failure scraps the run.
Test protocol
Four tests. One failure scraps the run.
Samples are drawn at random from each production batch. There is no downgrade path — a batch either ships as Krystio or it doesn't ship.
Steel ball drop
128 g from 1.2 mA 128 gram steel ball is dropped onto the centre of a mounted sheet from 1.2 metres. Pass requires no penetration to the substrate beneath. We run five drops per sample.
Pencil hardness
9H, 750 g loadA calibrated 9H pencil is drawn across the surface under a 750 gram load at 45°. Pass requires no visible mark under raking light after cleaning.
Sand abrasion
5,000 cycles, ISO 9211-4Standardised silica is oscillated across the surface for 5,000 cycles. Pass requires haze increase below 1% and no measurable loss of oleophobic performance.
Bend and torsion
220 mm radiusThe sheet is flexed to a 220 mm radius and held. Pass requires no fracture and no delamination at the PET interface after release.
Bill of materials
Nothing proprietary. Nothing hidden.
Every material in the stack is named, along with what it's there to do. If a competitor wants to copy it, the hard part was never the recipe.
- Glass
- Aluminosilicate, ion-exchanged
- Compression depth
- 40 microns
- Containment
- PET, 0.02 mm
- Adhesive
- Index-matched optical silicone
- Coating
- Fluoropolymer, 20 nm
- Edge
- 2.5D arc, diamond polished