Why and How Glass Breaks

I spent far too much time last week contemplating why we apply pressure on the opposite side of a score to break glass, and had to do a few more ‘GLASS IS NOT A LIQUID’ screeds. I decided this was worth a blog post. THE SCIENCE OF WHY AND HOW GLASS BREAKS.

1 - What Glass Is!

Please, repeat after me, GLASS (clap emoji) IS NOT (clap emoji) A LIQUID (clap emoji)! In grade school we learned that there are three (or four) phases of matter, right? Bad news, Mrs. Thompson lied to you. Ok, didn’t lie, just grossly oversimplified. 4th graders probably won’t understand the degradation of neutrons.

Solid – A uniform and dense configuration of tightly-bound molecules gives us rock, wood, ham sandwich, our skulls, etc.

Liquid – Less tightly bound and it’ll slosh around in your whiskey glass

Gas – Freer still, and it’ll fill up a room or otherwise expand to fill a container.

You might have also gotten to Plasma, which is the bright flash of lightning during a storm that ruins your day if it hits you.

Glass doesn’t fit neatly into any of these, so over the years various stories have developed. It must be a liquid, it must be a solid, etc. The truth is there are a lot more than 3 or 4 phases of matter, and glass is in a category of its own. It’s an ‘amorphous solid’ but that puts it in its own category along with degraded neutrons and other more unique things.

1b – Glass is NOT A LIQUID, pt 2.

The amount of time it would take for a glass object to ‘melt’ is something like ‘the age of the universe’. There is a very common misconception that old window glass is thicker on the bottom because it has flowed that way. If that were true, we wouldn’t have glass objects from Ancient Egypt and Rome. People wouldn’t have heirloom glasses older than grandma’s house with the weird windows.

The reason there is a thicker part to the window is that older methods of making glass sheets (e.g. the crown and cylinder methods) produced a sheet of irregular thickness. It’s some kind of ‘common sense’ to put the thicker part down for stability, but it’s not a rule and contrary examples exist.

1c – Wait, you didn’t really explain what glass is

Glass is solid, in the common sense that it’s not a liquid, gas, or plasma at room temperature (on earth, etc etc). But its structure is very disorganized, which is why it has similarities to a liquid. If its molecules were more orderly you would have "devitrified glass" ("de-glassed glass"), which qualifies as a ceramic.

Because of that attribute, a sheet of glass has strains within in. Various internal pulls and pushes and twists, so to speak. When glass is annealed a great deal of the stresses are able to resolve themselves, but not all. This is also why glass doesn’t have a grain like wood.

[Update: I thought it might be good to note that while current science identifies glass as a separate phase of matter, earlier academic documents would refer to it as an "amorphous solid" and, going earlier still, as a liquid. I believe that was accurate (for the time) using a very scientific definition of liquid as opposed to the common understanding. However, "Glass is NOT a liquid!" is much more concise than "Glass does NOT melt into new shapes over the course of one or two hundred years because it is not a liquid in the layperson's meaning of the term which has been superseded anyway by scientific progress and the inexorable march of time!"]

2 – How glass breaks, unsupervised

As is usually the case, force likes to take the Path of Least Resistance through glass. If a substance has a grain, like wood, you may be able to guess how it would split if hit with a general force. But in the case of glass, there is no grain and the stresses are mostly invisible[0]. As a result if you hit a piece of glass the fracture lines will follow the path of least resistance taking advantage of these faults and weaknesses until it reaches the edge of the sheet.

It's interesting to note that research (source below) shows you can apply less force than is needed to run a score if you apply it for a longer amount of time, even as much as 50% less. The authors note that if glass is under strain, it REALLY needs to be protected or minor damage may destroy it.

3 – How glass breaks, supervised

All we are doing when we break a piece of glass is coaxing the path of least resistance to be a route we want. I’d say ‘choose’ but if you’ve spent more than five minutes breaking glass you know it’s more of a coaxing and/or tearful pleading.

A typical modern glass cutter has either a steel (cheap) or tungsten carbide (much better) wheel that is shaped like a chisel. Incidentally the angle of that ‘chisel’ MAY have an influence on the cutting. I had never ever heard of this before until someone mentioned it (from a Wikipedia article). That article refers to a book on automotive glass and says a more sharply angled wheel cuts thicker glass better, from 120° - 154°. I don't work with industrial glasses so maybe that is common knowledge to them, but 154° is oddly specific to me.

Bob Beranek; Ann Schuelke (1 August 2011). The Complete Guide to Auto Glass Installation. AuthorHouse. p. 336.

Both steel and tungsten carbide are harder than glass, and so they are able to bite into the surface of a sheet. Carbide is just much harder than steel, so it is more durable. Usually steel wheels are found on cheap cutters you get at a hardware store. Carbide wheels are what you would expect from a specialty store and most anything directed at the stained glass market. Carbide cutting wheels also almost always have an oil reservoir in the handle.

Why oil? Its more to do with the tool than the glass. The stress of scoring glass builds up heat, which could eventually mess with the tempering of the metal wheel. It won’t generally be enough to cause thermal shock in the glass, but little glass particles are also caught in the deposited oil and that’s a good thing.

Fun fact, researchers (source below) discovered that the pressure applied to your cutter has an S-shaped effect on the force required to run the score (that is, break the glass). Once you reach the minimum force needed, adding pressure (pushing down your cutter harder) actually increases the torque needed to break it... until you blow past that zone in which case more pressure makes it easier again.

3b – Bringing it together

When we ‘cut’ glass we are actually introducing a known fault line into the glass in the hope and expectation that, when flexed, the path of least resistance will follow our score. And we all know that, sometimes, it doesn’t. Maybe a curve was too sharp and the PoLR was a straight line. Maybe a seed (bubble) was present that coaxed the line askew at a point, and then the PoLR is to follow the new angle instead of bending back to the score. Those invisible stresses can also be at fault, because glass is a temperamental mistress and lives off our tears. Wait, what? </Bitter Glazer>

4 – So, why DO we press from the opposite side of the cut.

It bothered me that it made sense in my mind, but I couldn’t explain it in words. I actively chewed on this for the better part of a week, thinking in terms like a race track. If you’ve seen one, you probably noticed that the starting line is staggered, and the closer to the center you are the further back you begin. This is because the center ring is smaller, and shorter, than the outer rings so that racer has to be further back to get the same distance. I was trying to leverage that rationale to explain breaking a flat sheet of glass and it didn’t quite work until I stopped thinking about a donut-shape.

 Not my best drawing, but showing that running pliers (or your thumbs) applying pressure on the cut effectively applies pressure to squeeze the score line. You might get the glass to break but it will be uncontrolled and unguided and probably isn't following your score. Note that the score is magnified A LOT in this drawing.

Another little sketch showing a greatly magnified score under no pressure, some sort of light pressure, and finally breaking pressure.

Incidentally this CMoG blog post has an even better graphic of this I found two days later, using foam.

5 – Bonus: Why your hand position is so important when cutting

I originally learned to cut glass by holding my cutter like a pencil. It's not quite that relaxed or the wheel wouldn't be able to contact the glass, but similar. I also noticed my accuracy was never where I wanted it to be, and I was always grinding pieces, finding new ways they didn't fit my pattern, grinding other areas of the piece, and eventually throwing it into scrap. After I began teaching with Molly I learned from her the "peace sign" grip that is pretty universally called for in professional-grade books.

  

My hand holding my favorite  cutter in a comically relaxed pencil grip, a much more likely pencil grip, and finally a 'peace sign' grip. The camera was not straight on, but you can still clearly see how the first two are further off-center than the last.

 I haven't found solid research into this, but I strongly believe that when you hold your cutter off-center, the score itself (and the associated "microfractures" if you believe that idea) are also at an angle. Least resistance being what it is, I suspect it would be more like a J-shaped curve than an actual diagonal, but you end up with a cut that is not perpendicular to the surface of the glass, what I know as a shelf [1]. Shelves are ridiculously sharp and seem to disappear if you even bring them near a grinder. I think that this is why my accuracy was poor for a long time;

You can see the 'shelf' on this piece, right above the little bump. It took me seven tries to hold a cutter wrong and get this piece

Holding the cutter off-center isn't going to give you a shelf every single time, or I would never have lasted this long in the hobby, but for me there was a major reduction in shelves and an increase in accuracy when I learned to use the better grip. Shelves are ridiculously sharp and will cut through your copper foil and your fingers effortlessly, not to mention (again) mess with the size of your piece. Probably a half-dozen times I have cut myself on a shelf-y piece and watched the delay as my skin realizes it's been cut, recovers from its shock for a second, and THEN starts bleeding. 

Allegedly neurosurgeons use obsidian scalpels because glass can be 100x sharper than steel. I believe it can be sharper than steel but about a year ago when I tried to investigate this, it seemed these scalpels were only an experiment.

6 - Really Cool Bonus: Glass Heals

I had heard Molly talk about this fact one summer while we were teaching. I'll be honest, I didn't believe it. How can a glass score heal and become more difficult to run if you let it sit for a few days? While doing other research, however, I found an academic paper from 1937 (Info below) that tested and proved the concept.

Their research showed that if you scored glass and let it sit, the amount of force needed to run that score increases over time, maxing out around 16 days. By the third day it's done the vast majority of its 'healing', however. Their hypothesis is that over the course of days the innate stresses in the glass will work themselves out a bit and reduce the effectiveness of your work. 

In conclusion, glass has a number of odd behaviors that bring to mind the expression "wibbly wobbly timey wimey", we work with the glass and beg it to do what we want, and windows aren't melting because GLASS IS NOT A LIQUID.

--------------------


[0] - There are polarized filters you can look through that will show the strains, but I don't THINK they are practical for non-scientists and I'm pretty sure they aren't the same thing that lets you tell which side of float glass hit the tin.

[1] - I am finally reaching that age where I can't quite remember whether I came up with this term myself or if I learned it from Norma/Danielle at GlassLink a decade+ ago.


Sources:

https://www.scientificamerican.com/article/fact-fiction-glass-liquid/

https://www.scientificamerican.com/article/is-glass-really-a-liquid/

https://en.wikipedia.org/wiki/Phase_(matter)

https://en.wikipedia.org/wiki/State_of_matter#Glass

https://io9.gizmodo.com/the-glass-is-a-liquid-myth-has-finally-been-destroyed-496190894

http://users.monash.edu.au/~ralphk/glass-cutting.html
(I'm not fully convinced on several concepts Mr. Klimek writes about, such as using spit to help run a score or the idea of microfractures, and we now know that glass is not a liquid but the anecdotal information is entertaining.)

"Concerning the Cutting of Glass" by MERTON W. JONES AND JULIAN M. BLAIR
Journal of Applied Physics 8, 627 (1937); https://doi.org/10.1063/1.1710352

https://blog.cmog.org/2015/06/03/part-2-why-does-glass-break/

A Glass Chemistry Primer

[Work in Progress! Likely revised often in the next week. My tables were broken when published, I will revise them shortly - Brynn 8/7/18]

Components of Glass
Formers - The material that makes up the bulk of the actual glass. Most commonly this is Silica Dioxide (SiO2), as in soda-lime glass. Aluminum fluoride and zirconium fluoride are also options.

Fluxes - As in other trades and crafts, fluxes lower the melting point of the former(s) involved. Feldspar, soda, natron, and potash are all fluxes.

Stabilizers - These materials impact a variety of attributes of the glass, like ability to withstand weathering. Limestone (calcium carbonate) is a stabilizer. Sodium and Magnesium are water-soluble and can leach out in water without stabilizers.

Colorants - Optional components to add (or remove) color and opacity to the glass. Mostly metal oxides. Gold, silver, copper, nickel, cobalt, iron, and many others act as colorants. Lead acts as a decolorant. Tin and antimony are opacifiers.

Fining Agents - Optional compounds to remove bubbles, e.g. arsenic trioxide.


Types of Glass
Soda-Lime - The type we encounter every day, roughly 90% of all glass produced. Silica’s very high melting point is lowered by the presence of soda (Sodium carbonate). Different formulations are used for flat (or float) glass vs other containers.
Potash-Lime/Forest glass - Potassium Carbonate (potash) is used instead of sodium cabonate (soda). The potash comes from burning inland plants, hence the term Forest glass. The ashes are soaked in water (leached) to extract the needed substances.
Borosilicate - Glass with a very low Coefficient of Expansion (around 33), which is very resistant to thermal shock. This makes it useful for cooking ware, scientific glass equipment, and artisan smoking vessels.
Fluoride glasses - Glasses which do not use silica as the predominant ingredient. Aluminum fluoride and zirconium fluoride are both options. This glass is usually used for advanced engineering and scientific applications, such as fiber optic cables.
Lead glass (Lead “crystal”) - Originally glass with a very high lead content (potentially over 50%, higher than the ~30-40% silica content). This glass has a brilliant reflective quality and is very clear, making it the best choice for cut ‘crystal’ glassware. Due to concerns over the lead content, modernly barium, zinc, and potassium oxides may be used instead of lead.
Flint glass - An early form of lead crystal. 4-60% lead, flint was used as the source of silica originally, giving it the name.
Fused quartz/fused silica - Silica fused into glass without fluxes. This requires extremely high temperatures, which were not attainable until modern technology (1650°C/3000°F). This is often used for engineering/scientific needs, though rods are also used by lampworkers.
Milk Glass, Vaseline Glass, Carnival Glass - Mostly names based off specific colors, rather than actual differences in the glass.
Porcelain, thermoplastics, glass-ceramics, etc. - There are many other things which, in strict technical terms, are glasses. They are not what any of us laypeople think of as “glass” however and are ignored for our purposes.

Specific Glass Chemistry Concepts
Striking Colors - Colorants are added to the batch of raw glass, and at the high temperatures involved they disperse well throughout the material. The colorants are able to bond with other elements in the batch and, when quickly cooled, stay in those compounds. The glass is often clear when leaving the factory.

Once the color is introduced to a torch or furnace, the temperature is hot enough to cause the colorants to break free from the compounds and form oxides or colloidal particles. Colloidal particles mean the colorant is dispersed throughout the glass without actually bonding with/truly becoming part of the glass (clouds, mayo, milk, and many other things are colloids too). This temperature is also cool enough to prevent the colorant from forming the original bonds at the factory (and/or is cooled more slowly).

Silver glass is a specific form of striking color. Particles of silver are colloidal particles in the glass. At the working temperature, larger chunks form and cause the glass to get beautiful metallic sheens.

Coefficient of Expansion (CoE) - The CoE of a glass is something that often scares people. It is a value indicating how much the glass expands when heated, and generally speaking all glass artists need to know is that you can’t join two glasses which have more than a 1-point difference in CoE. Glaskolben (a glass bulb-and-tube used to make ornaments, which borosilicate lampworkers may also call a ‘point’) that are commercially available have a CoE of 89, but are still compatible with CoE 90 glasses which are commonly available at art glass suppliers.

This value is actually an average! The CoE is tested at a number of temperatures, generating a curve. The number glass artists (and glass scientists) usually refer to is the average, roughly. While we usually refer to it by a number, in scientific terms this number is expressed in scientific notation, 10-7 per degree (in Kelvin). It’s further muddled by the fact that the change is expressed as a fraction of the length, rather than a typical unit like millimeters.

CoE 90 = 90 x 10-7 K-1 = 0.000009 change per degree Kelvin
(The space between 100 degrees Fahrenheit is roughly 55 degrees Kelvin, for comparison)

Trivia: Ceramic artists and scientists use the same values, usually expressed in terms of 10-6 x K-1, so CoE 90 in “glass” would be CoE 9.0 in “ceramic” shorthand.

But HOW is it measured? - With a “push rod dilatometer”. A sample of a material is fixed in place, a rod with a known CoE is placed against it, and a VERRRRY sensitive device is on the other end of the rod. The sensor has to be able to detect very tiny changes. As the sample is heated it will push on the rod and thus the sensor. Then “math happens” to correct for the reference material’s CoE. Fused silica/quartz rods are often the reference.

Example Coefficients of Expansion:



Satake Glass
(Used in Japanese lampworking)
CoE 120
Effetre, CiM, Vetrofond, etc
(Common in western lampworking)
CoE 104
Spectrum & Wissmach fusible glass
CoE 96
Bullseye & Uroborus fusible glass
CoE 90
Typical window/float glass
CoE 82-87
Borosilicate glass
(e.g. Pyrex pans)
CoE 30-33

List of Colorants - Below is a list of common colorants, and the basic color(s) they give to glass. I’m using the default “Crayola 12” colors rather than trying to describe the different shades, such as the green from Iron verses that from Chromium. Most of these only need to be added in tiny amounts to color glass. .001% Copper Oxide can impart an emerald green, 2% Cerium Oxide gives a light yellow.

Some of the decolorants below are also colorants. This works because the resulting colors balance/cancel each other out (e.g. Manganese brings in purple which ‘fills out’ the green from iron impurities). Sunlight could change Manganese over time, turning clear glass into purple!


Red
Copper-Tin, Cadmium-Selenium
Amber/Brown
Copper, Nickel, Sulfur
Yellow
Uranium*, Cadmium, Sulfur
Green
Iron, Chromium, Uranium*
Blue
Copper, Cobalt
Purple
Manganese, Neodymium
Black
Iron, Manganese, Cobalt, Lead
White
(Opacifiers really)
Tin, Antimony, Arsenic, Bones**
Pink
Gold, Erbium
Decolorant
Lead, Manganese, Cerium, Sodium Nitrate








* - Yes, Uranium as a colorant is as bad an idea as you think it is! Mostly discontinued
** - Bone ash, specifically. Yep, skeleton bones!


Sources:
https://www.cmog.org/article/chemistry-glass
https://en.wikipedia.org/wiki/Fluoride_glass
http://www.compoundchem.com/wp-content/uploads/2015/03/The-Chemistry-of-Coloured-Glass.pdf
https://en.wikipedia.org/wiki/Forest_glass
Introduction to Glass Science and Technology, 2nd Ed. Shelby, J.E.
http://www.tainstruments.com/wp-content/uploads/BROCH-DIL.pdf
https://www.ima-na.org/page/what_is_feldspar
http://www.chemistryexplained.com/Ge-Hy/Glass.html
http://www.compoundchem.com/2015/03/03/coloured-glass/
https://www.lehigh.edu/imi/teched/GlassProcess/Lectures/Lecture04_Shelby_ColoredGlass.pdf
https://www.mountainglass.com/tips-and-tricks
https://sha.org/bottle/pdffiles/TheColorPurpleLockhart2006.pdf
http://www.bullseyeglass.com/what-are-striking-glass-colors.html