How Is Flat Window Glass Made?

Look through a modern window and the view is almost perfectly undistorted. Old windows were different: the glass often had ripples and waves, and large clear panes were expensive because each sheet had to be ground and polished flat.

Almost all flat glass today, from windows to mirrors and car windscreens, starts life the same way. Molten glass is poured onto a long bath of molten tin and allowed to float. The surface of a liquid is naturally flat, and the float process takes advantage of that to make glass sheets by the kilometre. Here is how it works and why tin is the key ingredient.

The basic principle: let gravity do the polishing

A liquid at rest forms a level surface. If you pour one liquid onto another, denser liquid that it does not mix with, the lighter one floats and spreads into a layer with two smooth faces: the bottom face takes the shape of the liquid underneath, and the top face is flattened by gravity and surface tension.

The float process applies this to glass. The molten tin underneath is flat, so the underside of the glass is flat. The top surface is fire-polished by the heat above it. Once the glass cools enough to hold its shape, both faces are smooth and parallel without any grinding or polishing.

The process was developed at Pilkington in the UK between 1953 and 1957 by Alastair Pilkington and Kenneth Bickerstaff. It became commercially profitable around 1960 and was licensed worldwide during the 1960s, replacing the older approach of grinding and polishing plate glass.

How it works, step by step

1. Batching. Fine sand, soda ash, limestone, dolomite and smaller additions are weighed and mixed. Crushed waste glass, called cullet, is usually added as well.
2. Melting and refining. The batch is fed into a furnace where the glass is at temperatures of around 1,500 to 1,700 °C. Melting is continuous and can take up to about 50 hours, during which bubbles and unmelted particles are removed.
3. Conditioning. The melt is cooled to roughly 1,000 to 1,200 °C so it has the right thickness of flow (viscosity) for the next stage.
4. Floating. A gate called a tweel controls how much molten glass pours over a refractory lip onto the tin bath. The glass spreads out and moves along the bath as a continuous ribbon.
5. Forming. Rollers at the edges of the ribbon and the speed at which it is drawn off control its width and thickness. The ribbon cools from about 1,100 °C to about 600 °C along the bath.
6. Coating (optional). Some lines apply very thin coatings to the hot ribbon by chemical vapour deposition, for example to reflect heat.
7. Annealing. The ribbon passes through a lehr, a long oven that cools it slowly so internal stresses do not build up.
8. Inspection and cutting. Automated systems scan for flaws, then diamond wheels trim the stressed edges and cut the ribbon to size.

The science behind it

Why tin?

The liquid underneath has to meet several demanding conditions at once. It must be denser than glass so the glass floats. It must stay liquid at the temperature at which the glass becomes solid, around 600 °C, so the glass can be lifted off a still-liquid surface. And it must not boil or give off large amounts of vapour at the hottest end of the bath, around 1,100 °C.

Tin melts at about 232 °C and boils at about 2,600 °C, so it is liquid across the whole working range of the bath with a wide margin. Its weakness is that it reacts with oxygen to form tin oxide, which sticks to the glass. To prevent this, the bath is kept under a slight positive pressure of a protective nitrogen and hydrogen atmosphere.

How deep does the glass sit in the tin?

A floating object sinks until it displaces its own weight of liquid. For a flat sheet, the fraction of its thickness below the liquid surface roughly equals the ratio of the two densities:

fraction submerged ≈ ρ_glass / ρ_tin

Approximate worked example, using the room-temperature density of soda-lime glass (molten glass is slightly less dense, so the real figure is a little lower):

2.5 / 6.99 ≈ 0.36

So a 6 mm ribbon would sit with roughly 2 mm below the level of the surrounding tin and about 4 mm above it.

Controlling thickness

Left alone, molten glass on tin spreads until gravity and surface tension balance, which gives a natural thickness of roughly 6 to 7 mm. The original float lines could only make this one thickness. Modern lines use edge rollers and draw-off speed to stretch or hold back the ribbon. Pulling faster makes thinner glass; restraining the spread makes thicker glass. Pilkington reports that the process now produces glass from about 0.4 mm to 25 mm thick.

Why slow cooling matters

As the ribbon cools, the outside cools before the inside. If this happened quickly, the glass would lock in uneven stresses and could crack during cutting or later in use. The lehr brings the temperature down gradually so those stresses relax. This step is the same physics that, used in reverse, makes toughened glass: tempered glass is deliberately cooled fast to lock in useful compressive stress at the surface.

Real-world applications

• Windows and façades. Float glass is the base material for most architectural glazing, including double-glazed units.
• Safety glass. Toughened and laminated glass both start as float glass that is then processed further.
• Coated glass. Low-emissivity and solar-control coatings, which reflect infrared heat, can be applied on the float line or afterwards.
• Mirrors. The flatness of float glass is what makes a modern mirror reflect without distortion.

Common misconceptions

“Window glass is polished smooth by machines”

That was true of older plate glass. Float glass needs no grinding or polishing: the bottom face takes the flatness of the tin, and the top face is smoothed by heat.

“The glass is cooled in water or air to set it”

Fast cooling would build in stress. The ribbon is cooled gradually on the tin and then annealed slowly in the lehr. Rapid cooling is used only on purpose, later, to make tempered glass.

“Any heavy liquid metal would do”

Density alone is not enough. The metal must stay liquid below the temperature at which glass sets, must not boil at the hot end of the bath, and must not stain or react with the glass. Lead was used in the past, and earlier attempts to float glass on tin were not workable until Pilkington solved the practical problems of feeding and controlling the ribbon.

Limitations and exceptions

• Not all glass is float glass. Container glass for bottles and jars is shaped in moulds, and its composition differs slightly from flat glass.
• Defects still happen. Bubbles, unmelted grains or ripples from disturbances in the tin can occasionally slip through, which is why continuous automated inspection is used.
• Energy use. Furnaces run continuously at very high temperatures, so flat-glass making is energy-intensive.
• Safety. Cutting, handling and installing large sheets of glass carries a real risk of injury. Glazing work, especially overhead or in doors and wet areas where safety glass is required, should be done by a qualified glazier.

Header photo: Christian Ladewig on Unsplash.