Glass As A Liquid

Rev. 2002-11-15, 2005-03-04 (minor), -11-29 (links), 2009-04-13, -09-09, 2011-02-05

Back to Sitemap

The argument
Why not a liquid
Flow
Supercooled
Cathedrals
Old stained
Viscosity
Slightly Old Glass
More on making Window Glass
Nice Video
Why is Glass Transparent?

Perhaps the greatest waste of ink and Internet bandwidth has to do with the arguments over whether glass is a liquid and whether it flows at room temperature. The simple answers are no, it is not, and no, it doesn't. 

Even the people who should know better will refer to glass as a supercooled liquid when they really mean it is an amorphous solid - non-crystalline. Ironically, glass can be supercooled, but it has to be molten liquid for this to occur. For a lengthy discussion of various aspects from a scientific view point see Antique Window Glass - Supercooled?  And even that title leads to confusion, since supercooled means a liquid that is below its freezing point but is still liquid.  A more casual discussion follows.

The arguments are built on a misunderstanding of two facts: glass is more like a liquid in structure than most solids and glass in Medieval cathedrals is thicker at the bottom than the top.

It is common for books to state that glass is more like, or just like, a liquid than a solid. This is moderately true. Like a liquid it just gets thinner and thinner (less viscous) as the temperature rises. Unlike most solids we encounter, there is no sharp melting point, unlike say ice which is solid at 31.9F and liquid at 32.1F and can be found with little beads of liquid on the solid.
Glass changes from a solid that will shatter to a liquid that will flow over a range of temperature that is often several hundred degrees. In between glass is soft and squishy and can be pushed around or slowly sagged.
Also, like a liquid, cold glass does not have an organized structure (crystals) like the majority of crystalline solids we deal with. It is amorphous, as are certain forms of silicon used in photo cells, many rocks, and certain ceramics. If fact, glassy is sometimes used as a state of or description of matter.

But one thing we can be absolutely certain is that glass does not flow at human temperatures. Any glass worker who has dropped a piece of glass at 700-800F and watched it shatter (while it is soft at 1200F) is fully aware that glass gets too stiff to flow long before the temperature drops past the boiling point of water. It is possible to measure the viscosity (stiffness) and strain relief of glass, and it is done to measure how to anneal the glass, and that measure gets into the millions of poises at several hundred degrees.  It is also not a super cooled liquid which is a liquid cooled below its freezing temp, but still a liquid, that instantly turns solid if a crystal is added.

The other factoid leading to the discussion is window glass found in old cathedrals, which is thicker at the bottom than the top. Some people use this to support the idea that glass flows over time. There are two obvious problems with this idea: if glass flowed, there would be an open gap at the top where glass moved down and objects much older than these windows, like the hundreds of pieces of Roman Era glass that are around, would be distorted and they aren't.

Flat glass at the time the cathedrals were built was made by the broad glass method, making a cylinder, opening it while hot on a flat surface and cutting to size or by casting the glass by pouring on a flat plate and rolling. The glass, after annealing, is cut into relatively small pieces  Further, the glass is uneven in thickness and rough because of being opened on a flat surface while hot and details of flaws show the curve of the flattening.  [Previously I had stated that crown glass was used, but crown glass was developed long after - 1696 - the cathedrals were built - Notre Dame de Paris 1250.] 2006-08-21.

When such glass is cut it is usually slightly darker near the thicker edge. Glass artists quickly learned that in windows if the thicker part is near the bottom, the resulting picture looks better to the eye. Having learned it, glass artists designed for the effect and cut the glass to match the design.

So the glass did not sag or flow over the centuries, it was thicker at the start and hasn't moved.

Schott, p.19, gives specific numbers for glass viscosity. Glass must be at 102 Poise to cook to homogenous state (the same as olive oil at 20°C/68F) The state from 104 to 108 P is the working range, 107.6 is the softening point at which it will sag of its own weight, 1013 is the annealing point and 1014.5 the strain point.  
The annealing and strain points for a common art glass are 890F and about 650F.  Thus glass at cooking temp (from batch to liquid) is 100 P, while glass at the annealing point is 10,000,000,000,000 P and it gets measurably stiffer at lower temps. Although it will probably only confuse the classroom, I will mention that the softening point (sags from its own weight when supported from the ends) and annealing point (where it is held to relieve strains developed by cooling) are only 50C (90F) apart.  At the softening point, glass moves slowly. At the annealing, virtually not at all.

Super cooledAmorphous


Here is another variation on the theme.

Glen wrote:
Yes I have heard this too. It seems the definition of a liquid states that
it can flow into any shape. Glass does flow in time. Next time you come
across a 100+ year old building, look at the windows. You will notice the
glass is distorted. If you find a broken window, pick a piece of glass from
the top of the pane and compare the thickness to a piece from the bottom of
the pane. You will find the glass from the bottom is thicker.
Gravity has caused the glass to flow down.

Old windows from 100 year old buildings have distorted glass not because the glass has flowed but because they were made that way. Between the time that glass was made by spinning a roundel, which causes a variable thickness (above) and it was made by machine by drawing it in various ways, panes of glass were made by blowing a monster cylinder - manually it might be 20 inches in diameter and 8-10 feet long - cutting off the ends, cutting the sides into curved half sections and then running the sections through a kiln where the heat allowed them to sag and then be annealed. During the sagging the inside and outside surfaces of the cylinder, being of different radii, caused ripples as they tried to be parallel. Modern cylinder blown glass - a few people still make it - has the same ripples from the start.

 

 

Contact Mike Firth