This is not such a great article. It's true that the meme "glass is a liquid" is tiresome and incorrect, and it is true that glass doesn't flow at usual temperatures, but glass is very special and interesting because it doesn't undergo a state transition from liquid to solid state. So in a very specific technical sense, glass can be considered liquid, but there are many definitions of liquid and solids and this one doesn't align with people's naive definition.
I hoped would article would touch these special properties of glass.
I was about to say something like this in response to the parent comment. Instead of saying "glass panes are actually more like a very thick liquid than it is a solid" perhaps we should be saying "molten glass is more like a very flexible solid than it is a liquid."
This technical explanation of why glass "can not" flow, does nothing to try and explain the phenomenon where old glass panes are thicker on the bottom than the top.
A lot of things once thought impossible have been proven to be possible. Offering an explanation of why it's impossible adds nothing to the table.
[EDIT] Missed the bit in the middle. That said, I still find the use of jokes and second hand experiences a bit weak in evidentiary terms. Would have been much more convincing argument if there was a count or documentation detailing how many panels were truly installed one way or the other.
"Half of the pieces in a window are thicker at the bottom, he said, but, he added quickly, the other half are thicker at the top. My own experience has been that for earlier windows especially, there is sometimes a pronounced variation in thickness over a distance of an inch or two on individual fragments. That squares with the experience of conservators and curators who have handled hundreds of panels. Although the individual pieces of glass in a window may be uneven in thickness, and noticeably wavy, these effects result simply from the way the glasses were made."
Also, if you're a glass installer, you might feel its a best practice to put the heavier end on the bottom. It'll make the pane feel more stable because being top-heavy is like asking for gravity to take your glass and shatter it.
With today's production methods and quality control, we really don't have this issue. Glass panes are pretty consistent, but from the point of view of some 18th century carpenter, you probably had to deal with a lot of inconsistent glass panes.
The portion you quoted is a little misleading (understandably so for conciseness, but it does give the wrong impression). With a bit more context:
"Orowan had quipped that there might, indeed, be some truth to the story about glass flowing. Half of the pieces in a window are thicker at the bottom, he said, but, he added quickly, the other half are thicker at the top."
Which reads to me as a joke. I note this simply because it seems sub-optimal to introduce myths about the thickness of glass in old cathedrals when correcting a myth about glass in old cathedrals :)
The explanation is rather simple: glassmaking techniques were too primitive to achieve a uniform thickness. It wasn't just "a bit in the middle;" this was addressed pretty extensively beyond the quote someone already gave you.
BTW, a little bit of "gedankenexperiment" would show that, if glass really did flow, the panes would not end up so uniformly thin at the top, and so uniformly thick at the bottom, with a smooth transition in between.
Rather, they would sag and warp, eventually becoming non-flat if there was any small deviation from verticality. They would remain thicker for longer around the edges (assuming the pane is embedded solidly in a frame) and any support elements that hold it firmly. The top edge would sag more quickly in the middle. Heck, prevailing winds, if any, would slowly push the pane inwards, like a soap bubble.
The neat shallow wedge shape simply cannot be obtained via regular viscous flow in a near-vertical pane.
---
Also, if mere gravity could change the shape that much, any load-bearing glass element supporting a heavy item would sag visibly in a much shorter time. I'm thinking of the glass table supporting my very large, very heavy plasma TV.
So if time to form a drop is linear in viscosity (I have no idea if it is or not, but it seems like a plausible guess) then it would take longer than the present age of the universe for glass to form a single drop. In those circumstances, saying "glass doesn't flow" seems like a pretty darn good approximation to the truth.
Well, we could test it. Put lead into a funnel, suspend the funnel above a beaker, and wait for a drop. Assuming the scale is linear, you should only have to wait around 80 to 130 years based on pitch experiments:
More seriously, there's possibly a viscosity at which gravity is no longer enough acceleration to overcome whatever mechanisms are holding the atoms together. If we assume that metallic lead does not exhibit any flow on Earth, then we can estimate this as being between 10^10 and 10^11P at g.
It seems like from the definition of viscosity and the estimate of glass viscosity provided we should easily be able to observe some glass flow over the course of 10^12 years or so.
It's kind of funny that even in glassblowing circles this myth gets propagated...where the old methods of making panes are generally well known and understood (see the picture of the crown glass in the article).
If glass did flow at room temperature, telescopes would be a mess.
All the error you're allowed when making the mirror of a reflecting telescope for visible light is 0.1 microns, or 100 nm. That's the top limit; a good telescope needs to push that down at least 2x or 3x, if not more. The mirror must remain within that limit for the entire lifetime of the instrument.
Now picture this big hunk of glass, resting on 3, 9, or 27 support points (or much more than that for the large professional scopes). It's not sitting on a flat table; it's sitting only on a few narrow points on the back, with weight distributed equally among all points. It is so thick so as to prevent it from sagging under its own weight and distorting the active surface via elastic deformation.
If glass was a "liquid" to the extent that cathedral windows would thicken noticeably at the bottom over centuries, how long would a telescope mirror remain within 100nm? A few minutes before it got irreversibly warped like icecream in the summer? Yet there are many telescopes still in use after more than 100 years - and the optics are just as good now as they were in the beginning.
---
Before anyone tries to respond with a counter-argument: first try and realize what 100 nm really is. If you take one strand of human hair and cut it in two, and look with a microscope at the fresh section - if the surface of that fresh cut was the size of the rug in your living room, 100 nm would be the size of a large coin on that rug.
Think about that.
---
(I make telescope mirrors and I've lost track how many times I've told this argument whenever there's a discussion about "glass is a liquid".)
My girlfriend and I met another couple to go canoeing one day on town lake. The guy in the other boat was a chemist. He trotted out the old "glass is a liquid" thing, and I explained no, that's not true, and I went into great detail. But he stuck to his guns and insisted, and then played the "professional chemist" card, saying "I think I know a little bit more about chemistry than a guy who makes webpages."
I was furious, but my dogged insistence that I was right was starting to draw glares from both of the women, so reluctantly, in the name of peace, I dropped it. Then the ass had the gall to say "Haha, don't worry, I'm sure there's a lot of things about webpages that you could correct me on!"
The worst part is, if I sent him this article, and any number of others proving the same point, he'd just say "what? why are we talking about glass?" That conversation was nothing to him, and I'm sure he's already forgotten it.
> But he stuck to his guns and insisted, and then played the "professional chemist" card, saying "I think I know a little bit more about chemistry than a guy who makes webpages."
This is precisely why authority has no role in science, why evidence outranks eminence.
"Science is the organized skepticism in the reliability of expert opinion." — Richard Feynman
I think it is better to say that glass DOES flow, but it is a solid! For example, it would take several universe lifetimes to observe any appreciable flow (or more quantifiably, the viscosity is circa 10^19 Pa-s). Similarly, lead flows and to a much lesser degree, steel too.
28 comments
[ 3.1 ms ] story [ 75.6 ms ] threadI hoped would article would touch these special properties of glass.
http://en.wikipedia.org/wiki/Glass#Formation_from_a_supercoo...
http://en.wikipedia.org/wiki/Glass_transition
http://en.wikipedia.org/wiki/Amorphous_solid
A lot of things once thought impossible have been proven to be possible. Offering an explanation of why it's impossible adds nothing to the table.
[EDIT] Missed the bit in the middle. That said, I still find the use of jokes and second hand experiences a bit weak in evidentiary terms. Would have been much more convincing argument if there was a count or documentation detailing how many panels were truly installed one way or the other.
"Half of the pieces in a window are thicker at the bottom, he said, but, he added quickly, the other half are thicker at the top. My own experience has been that for earlier windows especially, there is sometimes a pronounced variation in thickness over a distance of an inch or two on individual fragments. That squares with the experience of conservators and curators who have handled hundreds of panels. Although the individual pieces of glass in a window may be uneven in thickness, and noticeably wavy, these effects result simply from the way the glasses were made."
With today's production methods and quality control, we really don't have this issue. Glass panes are pretty consistent, but from the point of view of some 18th century carpenter, you probably had to deal with a lot of inconsistent glass panes.
"Orowan had quipped that there might, indeed, be some truth to the story about glass flowing. Half of the pieces in a window are thicker at the bottom, he said, but, he added quickly, the other half are thicker at the top."
Which reads to me as a joke. I note this simply because it seems sub-optimal to introduce myths about the thickness of glass in old cathedrals when correcting a myth about glass in old cathedrals :)
Rather, they would sag and warp, eventually becoming non-flat if there was any small deviation from verticality. They would remain thicker for longer around the edges (assuming the pane is embedded solidly in a frame) and any support elements that hold it firmly. The top edge would sag more quickly in the middle. Heck, prevailing winds, if any, would slowly push the pane inwards, like a soap bubble.
The neat shallow wedge shape simply cannot be obtained via regular viscous flow in a near-vertical pane.
---
Also, if mere gravity could change the shape that much, any load-bearing glass element supporting a heavy item would sag visibly in a much shorter time. I'm thinking of the glass table supporting my very large, very heavy plasma TV.
http://en.wikipedia.org/wiki/Pitch_drop_experiment
Does lead flow?
The article dismisses it but only in a casual way.
http://en.wikipedia.org/wiki/Pitch_drop_experiment
More seriously, there's possibly a viscosity at which gravity is no longer enough acceleration to overcome whatever mechanisms are holding the atoms together. If we assume that metallic lead does not exhibit any flow on Earth, then we can estimate this as being between 10^10 and 10^11P at g.
All the error you're allowed when making the mirror of a reflecting telescope for visible light is 0.1 microns, or 100 nm. That's the top limit; a good telescope needs to push that down at least 2x or 3x, if not more. The mirror must remain within that limit for the entire lifetime of the instrument.
Now picture this big hunk of glass, resting on 3, 9, or 27 support points (or much more than that for the large professional scopes). It's not sitting on a flat table; it's sitting only on a few narrow points on the back, with weight distributed equally among all points. It is so thick so as to prevent it from sagging under its own weight and distorting the active surface via elastic deformation.
If glass was a "liquid" to the extent that cathedral windows would thicken noticeably at the bottom over centuries, how long would a telescope mirror remain within 100nm? A few minutes before it got irreversibly warped like icecream in the summer? Yet there are many telescopes still in use after more than 100 years - and the optics are just as good now as they were in the beginning.
---
Before anyone tries to respond with a counter-argument: first try and realize what 100 nm really is. If you take one strand of human hair and cut it in two, and look with a microscope at the fresh section - if the surface of that fresh cut was the size of the rug in your living room, 100 nm would be the size of a large coin on that rug.
Think about that.
---
(I make telescope mirrors and I've lost track how many times I've told this argument whenever there's a discussion about "glass is a liquid".)
a few minutes: I'd guess more like a few days: 1 mm/100 years = 10 micron/year = .1 micron / 3.65 days.
I was furious, but my dogged insistence that I was right was starting to draw glares from both of the women, so reluctantly, in the name of peace, I dropped it. Then the ass had the gall to say "Haha, don't worry, I'm sure there's a lot of things about webpages that you could correct me on!"
The worst part is, if I sent him this article, and any number of others proving the same point, he'd just say "what? why are we talking about glass?" That conversation was nothing to him, and I'm sure he's already forgotten it.
This is precisely why authority has no role in science, why evidence outranks eminence.
"Science is the organized skepticism in the reliability of expert opinion." — Richard Feynman