How many times do you reuse your bottles?

I'll take a stab at trying to explain the difference between glass and metals, and also try to keep it as brief as possible (although brief is a relative term.) In order to keep it brief, it cannot be comprehensive. So please, no berry punches for leaving out special cases or second/third order effects. For those of you who get bored reading this, feel free to move on to the next post. There will not be a test at the end

To understand what fatigue is, we must understand the difference between ductile and brittle behavior. In general, metals are ductile, and glass and ceramics are brittle (I'll leave polymers and composites out of the discussion.)

All materials deform when stressed. As long as the deformation is limited to the stretching/compressing of the microstructure, the deformations are reversible, and will disappear when the stress is removed. Fully reversible deformation is known as "elastic" deformation. Elastic deformation does not damage the material in any way.

When stress levels get high enough, the limits of microstructure stretching/compressing are exceeded. For brittle materials, once this stretching/compressing limit is exceeded, the atomic bonds rupture, and the material fails catastrophically (ie it shatters.) For ductile materials on the other hand, the microstructure can undergo additional deformation by altering the microstructure. These deformations are usually irreversible, and are known as "plastic" deformation. Plastic deformation represents damage to the microstructure of the material, and generally results in a weakening of the material. Over time and multiple stress cycles, the plastic damage will build up in the material, continuing to weaken the material. This build up of plastic damage and weakening is what we call "fatigue." Fatigued materials will fail at lower stresses than undamaged materials, and these lower stress failures are called "fatigue fails." If enough plastic deformation stress cycles are accumulated, the material can fail without additional stress.

In some cases the altered microstructure leads to a stronger/harder material. This is known as work hardening. This work hardened material is usually less ductile than the original material and more prone to cracking when the (higher) elastic stress limit is reached.

Now to the punch line: Since brittle materials fail before undergoing any plastic deformation, they don't build up plastic damage in the microstructure. Therefore they cannot "fatigue" in the classical sense.

Now some more on brittle failure:

Brittle materials normally fail at macroscopic stresses much lower than what is required to rupture the atomic bonds. IE their strength is much lower than would be predicted based on the strength of the bonds between the constituents. Why is this? The answer is cracks, most often microscopic cracks.

Stress is load divided by the cross sectional area supporting the load. In the presence of a crack, the cross sectional area is reduced, but the load remains constant, thus the stress is increased wherever there are cracks. And, this stress increase is not uniform. The stress is concentrated at the tip of the crack. The amount of stress concentration is determined by the length of the crack and the radius of the crack tip. Longer cracks and smaller radii lead to higher stress concentrations. The atomic bonds start to rupture when the stress at a crack tip exceeds the bond strength. Once the material starts to rupture, the crack gets longer, which causes the stress at the crack tip to increase even more. The result is sudden and catastrophic fail.

In practice, all brittle materials contain microscopic cracks which determine the engineering (usable) strength of the material. These are taken into account when designing objects made from brittle materials. The objects will continue to function as designed unless a much larger crack develops, and the stress concentration at the tip of that crack exceeds the working stress.

Now, there is a phenomenon known as "stress corrosion" that can affect glass and other brittle materials. Stress corrosion is often referred to as "static fatigue," but that is just lazy and confusing terminology. (The term "static fatigue" was coined before the mechanism of failure under static load was understood.) Stress corrosion is a different mechanism than cyclic stress fatigue or creep fatigue (which hasn't been discussed.)

Stress corrosion is the result of corrosion that is accelerated by the stress concentration at the tip of microcracks in brittle materials. The corrosion at the crack tip allows the crack to grow in length over time. If the crack ever grows to a length where the stress concentration at the tip exceeds the atomic bond strength, then rapid catastrophic failure can occur. Rates of stress corrosion are affected by availability of moisture at the crack tip and pH of the crack tip environment, as well as the particular material.

In theory, beverage bottles can be subject to stress corrosion. However, if the mechanism were operative in practice, we should see older bottles of carbonated beverages exploding on a regular basis, especially bottles with higher internal pressures. Since exploding bottles are rarely observed in the absence of massive overpressure. It is safe to assume that stress corrosion is not a significant factor in the life expectancy of beverage bottles. I did do a limited amount of searching for reports of stress corrosion bottle failures, but was unable to locate any.

As to observable stretching of beer bottles, I really doubt that you have ever seen this. Soda lime glass has a typical Young's modulus of 72 GPa or 10.4 million psi. This means that for every 1000 psi tensile stress in the glass, the glass would stretch ~1/10,000 th of an inch along each 1 inch of length (or 0.01%.) I know I could never detect this without special instruments.

Single use bottles are designed to be cheaper and lighter than multiuse bottles, so they are manufactured with significantly thinner glass. Because of this they are less able to withstand the rigors of being handled thru recycling and refilling. They are by design, crappier bottles.

Hope the above helps, and sorry it's so long. Feel free to ask questions.

Brew on

 

You very likely know more about structures than I do. Not sure off the top of my head about "elastic fatigue." My specialization was ceramics, not metals. It's possible that the bridge members could be undergoing a (stress) corrosion process, or maybe there is some creep going on at less than the measured yield stress. I'll see if I can find anything that looks like a good explanation.

EDIT: Ok I found something on Wikipedia that sheds some light on this:

Because of the type of atomic bonding in metals vs. glass and ceramics, these molecular level faults are more likely to occur in metals.

They had to have some kind of crack in order to fail. The instigating crack could have been too small for the naked eye to see. Without having good pics it's hard to say where the original crack was. With conchoidal fractures, the ridges often radiate outward from the original crack.

Brew on

 

Remember, that in the days of commercially refilling returned bottles, it would have been logistically impossible for a refiller to know how many cycles were on a bottle, or how badly they had been handled. If the bottles had molded in manufacturing date stamps, they would have known how old they were. But, it would have been very expensive to inspect and cull bottles by date codes (and automated vision systems didn't exist back then.)

Brew on

 

While I've added bottles, I'm pretty sure I still have the first case of used bottles from 36 batches ago. I have no intention of replacing them. I'm still using my grandmother's pie plate, and some lanterns that are decades old. While capping is stressful, so is baking and containing flames. Glass windows last centuries.

 

Yes, when they break

Is there a reliable way to determine when they will break? No.

Brew on

 

Definitely briefer, but I'm not sure it's as satisfying.

Brew on

 

Thanks, I'm glad at least a few people got something worthwhile out of it.

Brew on

 

Be careful of absolutes. Gases and liquids don't fatigue. Gold is so ductile that it doesn't fatigue when being deformed. That's why you can make gold leaf by pounding it out to ridiculously low thicknesses.

I think the highlighted statement conflicts with what most people who have learned about fatigue were taught. Do you have any support for that statement, and can you provide references? Would be nice if the references explained on a molecular level what leads to fatigue at very low stress levels. Not saying you're wrong, it's just not the way I learned it.

Brew on

 

Can you explain how you used gravity measurements to determine carb levels? It's not immediately clear to me how this would be done, and it sounds like it might be a useful method to know.

Thanks, and brew on

 

Ok, but how do you get from ABV to carbonation level?

Brew on

 

Yield stress is an engineering definition, conventionally set at some fixed amount of deviation from the linear portion of a stress/strain curve resulting from a standardized test. Plastic deformation starts as soon as stress is sufficient to cause any change in the microstructure of a solid material, even if the deformation is not detectable at a macroscopic level. These microstructural changes can occur at much lower stresses than the yield stress, as you have noted. It's the microstructural changes that accumulate over time, and multiple cycles, which eventually lead to fatigue and failure. I'm no expert on aluminum, so have no idea at what stress level microstructural changes can occur.

Ok, I get the general idea. Thanks.

Brew on

 

I hope not, since I graduated ~40 yrs ago

Brew on

 

Sorry about the self quote, but on further reflection, there is a way to tell if a bottle is strong enough to use safely. The method is a hydrostatic test, similar to what we have to have done to our CO2 cylinders periodically.

A hydro test is a type of proof test, which simply tests if the bottle is strong enough to withstand expected maximum use pressure + a safety margin. It the bottle doesn't break during the test, then all of the preexisting defects are small enough that no stress concentration due to them, at working pressure, will be enough to cause failure. The test is short enough that no significant stress corrosion (crack growth) can occur during the test, and the test pressure is low enough that it will not cause microstructure deformation (fatigue) on it's own. So, if the bottle survives the hydro test it is safe to use, and if not, then you can't use it anyway.

To select the pressure for a hydro test: pick a maximum possible carb level (let's say 10 volumes to keep Calder out of trouble ), and a maximum possible storage temperature (let's say 110°F). So, we have to figure out the pressure that corresponds to 10 volumes of CO2 @ 110°F, and that is the proof pressure. Of course, if all or most of bottles break at this pressure, then we would have to settle on a lower pressure with a smaller safety factor (no more priming to 9 volumes.)

Pressure should be applied using a liquid (water is probably best) rather than air, as the driving force behind shrapnel acceleration drops off much faster with an incompressible fluid than with a gas. Still, a shrapnel shielding enclosure for the apparatus is a must. You'd want to suspend the bottle from the neck, but with jaws specifically designed to prevent possible damage to the neck (unlike wing capper jaws. ) Then we just need a stopper with plumbing connecting it to a high pressure, flow restricted, water source.

Any demand for such a tester? Any DIY'er want to take it on?

Brew on

 

10 volumes was a little tongue in cheek, but the example works anyway for illustration. 5 volumes might be more reasonable for proof testing with some safety margin for normal cases. I've seen the 3.5 volumes number mentioned seveal times, but never seen an authoritative reference for the number. And a volume number by itself is not sufficient. We also need to know the max assumed temperature, since the pressure increases significantly with temperature for a given volume.

Brew on