Announcing Float Glass: Forming a Flat Sheet on Liquid Tin

Pilkington publicly announced the float-glass process on 20 January 1959 after years of experiments by Alastair Pilkington and research staff in St Helens. Hot glass passed onto liquid tin, forming and cooling gradually in a controlled environment before further transport and annealing. The objective was continuous, smooth, uniformly thick flat glass without expensive grinding and polishing to correct forming defects.

The company history dates experiments to December 1952. Pilkington’s 1969 account distinguishes a full-scale installation starting in May 1957 from saleable glass achieved in July 1958. Announcement and patent grant were later events, with repeated failures and modifications between the stages.

References: [2] [4]

Earlier rolled-sheet processes pressed molten glass into a ribbon. Contact with cool rollers lowered the surface temperature, hardened it and left marks needing removal. Pilkington and Kenneth Bickerstaff’s U.S. patent addresses that problem by passing a preformed ribbon into a metal bath to improve its surfaces through reheating and gradual cooling.

The patent claims priority from a British application of 10 December 1953. It was filed in the United States on 6 December 1954 and granted on 10 November 1959. Its roller-preforming arrangement preceded the mature direct-pour route that eliminated that step. The grant date and final process thus describe different parts of development.

References: [1] [2]

Tin is denser than glass, remains liquid in its working temperature range, reacts relatively little with it and causes limited disturbance through evaporation. Floating glass tends toward flat surfaces through gravity and interfacial tension. It does not spread indefinitely into a thinner layer: ordinary window glass has an equilibrium thickness of roughly seven millimeters in this system.

Charnock’s 1970 paper explains the demanding controls. Glass needs sufficiently low viscosity and enough time to remove irregularities, then must cool enough before leaving the bath to withstand roller support without surface damage. Faster pulling can narrow the ribbon without producing the desired thinner sheet. Later thin and thick products used stretching or restriction of lateral spreading.

References: [3]

Continuous production exposed problems less visible in small trials. Even with a protective atmosphere, oxygen and other impurities had to be excluded to prevent defects. The bath enclosure, heating and cooling zones and exit required coordinated adjustment. Gas passages and temperature-control arrangements in the patent show a controlled environment for reactions and heat transfer within the line.

Direct pouring encountered refractory contamination. Slight dissolution of the delivery spout changed local composition and produced fine optical streaks. Staff adjusted the spout and flow to direct glass touching refractories toward the ribbon edges. Pilkington recounts a repair that rebuilt a spout to match its apparent form yet again produced unsuitable glass. Models and production-line tests identified critical flow conditions.

References: [1] [2] [3]

The economic attraction was forming control that removed much later finishing, also reducing surface damage from grinding and polishing. Good samples nevertheless preceded dependable commercial supply. In his 1969 retrospective, Pilkington separately discusses saleable glass and fulfillment of the aim of replacing polished plate glass, with cumulative expenditure of approximately £4 million and £7 million. These were retrospective totals, not interchangeable annual investments.

After announcement, licensing spread the process to other manufacturers. By the 1969 report, fifteen overseas manufacturers held licenses and more than twenty installations were operating or being built. The 1959 node marks movement from internal development to public demonstration and technical dissemination, with capacity, thickness ranges and additional processes continuing to develop.

References: [2] [4]