The Molten Tin Ribbon
How Alastair Pilkington Floated Perfect Plate Glass on Liquid Metal
Until the late 1950s, architects and automakers had to choose between two flawed ways of making flat glass. Cheap window glass was drawn vertically from a furnace tank through asbestos rollers. Gravity and small viscosity ripples left waves across the cooled pane, distorting distant objects like a funhouse mirror. Whenever distortion-free optical flatness was required for shopfronts, mirrors, or windshields, factories had to cast thick glass between water-cooled steel rollers. Huge twin grinding machines then ground away twenty percent of the glass on both faces using wet sand and jeweler's rouge.
In October 1952, British engineer Alastair Pilkington stood at his kitchen sink washing greasy dinner plates. Watching soap and oil float into a smooth film across the rinse water, he conceived a radical alternative. What if a ribbon of molten glass were poured onto a bath of denser liquid metal? Gravity and surface tension would flatten both surfaces simultaneously without touching a solid roller. Pilkington brought the sketch to the board of Pilkington Brothers in St. Helens, Lancashire.
Selecting the liquid metal bath imposed strict physical and chemical limits. The metal had to be denser than molten soda-lime glass ($2.5\text{ grams per cubic centimeter}$) so the glass would float effortlessly. It had to melt well below $600^\circ\text{C}$, the temperature where glass solidifies enough to lift onto steel exit rollers without denting. At the same time, the metal could not boil or bubble at $1,050^\circ\text{C}$, where liquid glass flows out of the furnace. Across the periodic table, only metallic tin satisfied every boundary condition. Tin melts at $232^\circ\text{C}$, boils at $2,602^\circ\text{C}$, has a density of $6.5\text{ grams per cubic centimeter}$, and does not react with liquid silica.
Turning that thermodynamic match into a working factory line nearly bankrupted the company. When Pilkington and Kenneth Bickerstaff poured $1,050^\circ\text{C}$ glass over a refractory lip onto molten tin, trace oxygen in the air immediately attacked the metal bath. A crust of tin dioxide ($\text{SnO}_2$) formed on the liquid tin and stuck to the underside of the moving glass ribbon. Furthermore, stannous ions ($\text{Sn}^{2+}$) diffused into the bottom micrometers of the glass, causing a milky bloom whenever the finished sheet was reheated for curved windshields.
To stop oxidation, Pilkington encased the tin bath inside a sealed steel roof pressurized with a reducing atmosphere of ninety percent nitrogen and ten percent hydrogen. The hydrogen scavenged stray oxygen atoms and reduced any tin oxide film back into bright liquid metal. Inside that oxygen-free chamber, molten glass spreads across the liquid tin until downward hydrostatic gravity balances inward capillary surface tension. Left untouched by top rollers, a floating soda-lime glass ribbon on molten tin naturally settles to an equilibrium thickness of $6.8\text{ millimeters}$.
To make thinner windowpanes or thicker slabs, Pilkington added toothed top edge-rollers that grip the margins of the hot ribbon. Speeding up the exit rollers stretches the viscous ribbon down to $0.4\text{ millimeters}$, while pushing inward thickens it up to $25\text{ millimeters}$. As the ribbon travels down the fifty-meter tin bath, its temperature drops smoothly from $1,050^\circ\text{C}$ to $600^\circ\text{C}$. The hardened glass emerges fire-polished on both faces without ever touching a grinding disk. During a fourteen-month trial on a full-scale production tank in St. Helens, the crew ran the furnace twenty-four hours a day while producing nothing but broken scrap glass before the flow geometry stabilized. After producing more than one hundred thousand tons of cullet scrap during fourteen months of trial runs in St. Helens, Pilkington unveiled commercial float glass in January 1959 and cut plate-glass manufacturing costs by eighty percent.