It is possible in the case of glasses because they can change from a brittle to a plastic state: at the softening temperature Tg, the state of the glass changes, and only for this reason it can be "hardened". During the "quenching" of glass (rapid cooling of temperatures higher than Tg to temperatures lower than Tg), residual compressive stresses are "frozen" in the surfaces.

Crystalline ceramics do not have Tg. If you cool them quickly, the surface of the ceramic will also get stress, but these are only temporary. No residual stresses remain in the material.

It is possible in the case of glasses because they can change from a brittle to a plastic state: at the softening temperature Tg, the state of the glass changes, and only for this reason it can be "hardened". During the "quenching" of glass (rapid cooling of temperatures higher than Tg to temperatures lower than Tg), residual compressive stresses are "frozen" in the surfaces.

Crystalline ceramics do not have Tg. If you cool them quickly, the surface of the ceramic will also get stress, but these are only temporary. No residual stresses remain in the material.

I don't have experience in quchening ceramics, like Alumina, but seem that if done right, it can be Quench-Strengthening.

http://link.springer.com/chapter/10.1007%2F978-94-015-8200-1_38

Structure of the glass is very specific to its thermal history. As Vadim pointed out, that is why it could be tempered/hardened. Ceramic can't be processed this way. That said, there is a method similar to thermal quenching that is used in glass that can, at least in theory, be applied to ceramics. In Gorilla glass -- the Corning product that is used in cell phone screens -- the compression on the outside and tension on the inside is imparted via ion exchange. If I remember correctly, the sodium ions are swapped out for potassium by submerging the glass into a molten potassium salt. So in theory, something similar can be done with a ceramic.

Although that brings in another consideration: the reason why tempered glass is resilient to failure is that the part of it under tension is "inside" and doesn't have anything to nucleate a crack, it's quite homogeneous. In a ceramic, you have grain boundaries, and these inhomogeneities could serve the purpose of a crack nucleus. This would defeat the purpose of outside compression-inside tension for strengthening.

Hi Erik

In general in quenching processes of materials, it is know that "what cools first is in a residual tension". So, in the case of quenching a glass, the surface of the glass is in residual tension and the bulk of the glass is in residual compression.   Thereafter, tempering processes of the glass would balance the mechanical properties across the structure of the glass in a way that a less stress gradient will develop between the surface & the bulk of the glass.   

Same things can be conducted on other ceramics materials in order to strengthening the ceramic with the following cautions :

1-The temperature of the quenching medium should not lead to cracking of the surface of  ceramic materials.

2-The  temperature range of  tempering processes of the ceramic should not lead to cracking of the interface between the surface & the bulk of the  ceramic materials.    

@ Khaled: By tempering or quenching/hardening of glass the (first cooled) surface is in compressive state and the bulk is in tension, not opposite. If the surface would be in tension cracks would easier/earlier start to grow and the glass would be weakened.

@ Pavel: Yes, the bulk is in tension. But if the cross-section of the bulk is large compared to the compressive stressed surface layer, the intensity of the tensile stress would be low and stay below the materials strength.

Good point.

I repeat again;

In general in quenching processes of materials, it is well known that "what cools first is in a residual tension".   Indeed, the surface of  a quenched glass,  first goes through a shrinkage stage as the surface touches the quenching medium.   Then, because of the presence of a temperature gradient  between the surface & the bulk of the glass, i.e., the bulk is still hot in a comparison  to the surface, the surface would gradually attain a residual tension state as the bulk slowly cools down as a function of time.

In addition, I have already mentioned the above cautions in my 1st answer to avoid cracks of the quenched surface.   

@ Khaled: I'm afraid that in sequential quenching of plates (the typical example would be steel, but the same reasoning goes for glass) the firstly quenched side will typically be in compression in its final state. This is due to the fact that the quenched side will cool down and shrink while the uncooled side will simply follow without developing large internal stresses due to the easy plastic deformation at high temperature for the unquenched side. For the following quenching of the opposite side, the initially quenched material can no longer accommodate the shrinkage plastically due to the fact that it is in its quenched hence hardened state. The only way to accommodate will be by developing large elastic compressive stresses.

The phenomenon is sometimes observed in quenching of continuously casted steel slabs. If the upper quenching installation is not working properly (and hence not cooling as rapid as the other side), the steel slab will bent towards the upper side, and not towards the lower side as many people think intuitively. I've put a document in attachment illustrating the effect.

Ceramics are crystalline structures with very low elasticity, and as a result are quite prone to thermal shock. In general, by trying to quench a ceramic, you are more than likely to cause cracks and failure of the material.

What are you trying to achieve (in theory) by quenching? Ceramics are hard to begin with.

As an example, I work a lot with alumina and it will loose a significant part of its strength as a result of a thermal shock caused by quenching from 250°C to RT.

Dear Erik:  You have not heard of Rupert's tears.  If you melt the end of a glass rod so that the tear drop falls into water, you form a tear-shaped particle.  This is so hard and residually stressed that it can resist being struck with a hammer.   Unfortunately it has has a tail.  If you cut off the tail with scissors, the tear drop spontaneously explodes.

@ Yevgeni:

Yes, quenching Alumina from T=250°C to room temperature (RT) will lead to damage due to thermal shock. You need to be above a minimum temperature (much higher than 250°C) in order to get some plasticity in the bulk of this material.

The required (minimum) temperature could be well above 1200°C for such materials, in fact you may have to be above some fraction of the melting point (in K) of the material.

I was hoping that other researchers could help me on these conditions

@ Campbell: Yes I know the Rupert's tears, there are nice high speed video's on the internet about the failure wave running through them when failure is initiated (which is not too easy considering the huge compressive pre-stress in them. 

@ Eric:

In the case of ceramics, is there distinction between pure crystalline materials and silicatic ceramic materials, which have been sintered with the aid of a glass phase. Ceramics with the glass phase can be deform plastically at high temperatures. This happens far below the melting point of main components of the material. Such traditional ceramics as, for example, porcelain behave almost like glasses and can, at least purely theoretically, be quenched. Although I know no "hardened" porcelain, but I can me imagine that a rapid cooling from 1300°C to about 800°C so an effect could be evoke (porcelain contains about 70% of glass phase).

On the other hand, the purely crystalline ceramics without a glass phase soften only shortly before the melting point.

Hi Erick

The example of Professor Laird on "Ruperts's tears" is a good way to illustrate a compressive stress , by the scissors, at the tail of the quenched tear drop of the glass would induce an explosive tensile stress at the tear drop, indicating that the surface of the quenched tear drop was under a residual tensile stress. 

I disagree with the view of Khaled Habib. 'Ruperts's tears' (thermally stressed glass) is very difficult to break due to compressive stress on the outer (first cooled) surface and has a tensile stress in the bulk in order to get force/stress equivalence. In fact the glass has been made such strong that even a strike with a steel hammer is not able to initiate fracture (you can find nice video's of such experiments on You Tube). That is because the compressive stress in the surface below the impacted area is so high that it overcompensates the tensile stress that occurs next to the impact area (see Herzian stress theory). All materials stay in their elastic regime hence no plastic deformations and no cracking occurs (although all glass surfaces have surface flaws from which cracks normally easily grow).

The large amount of elastic (tensile and compressive) energy in the hardened glass will be released once a crack starts growing. This is easier happening from the tail of the drop, as here the cross-section of the glass is much reduced this reduces the stiffness of the glass (stiffness relates to thickness to the third order). Possibly also the hardening effect is less at this tail. When overloaded a crack starts to grow from the tail and this releases the stored elastic energy. This energy release is high enough to have (bifurcating) cracks self-propagate through the whole of the material.

Nothing new, as all cars have side windows made of a single hardened glass layer that is quite strong, but the window shatters completely into small glass fragments. This helps to provide an escape route in emergency situations (like a car under water). All it takes is to initiate a small crack anywhere on the (compressively pre-stressed) glass surface.

So, this has been done on glass (tears and safety windows) for decades, my question remains: can this pre-stress also be applied to ceramic tiles like Alumina and SiC?

Dear Erik: I agree with your interpretation of the behavior of Rupert's tears.  Whether or not this behavior can be obtained for alumina and SiC will depend on the structure of the material.  A molten drop of glass is reasonably expected to be homogeneous and defect free internally, so as to allow the surface compressive stresses to keep any surface microcrack closed (barring the tail, of course, per your interpration).  Inside your ceramics, that likelihood of the material being free of defects in the interior where there are tensile stresses produced by quenching will depend on the history of its formation, and is unlikely.  I do not know if your ceramics can be prepared so defect free.  The Oxford experiments to produce "defect free" concrete did achieve remarkable tensile properties in that material.

Hi folks

As I said before "what cools first is in a residual tensile state".  

Also, I would say

"what cracks first and crack propagates  is only in a tensile state".     

Dear Khaled:  What cools first is in a tensile state, and that of course is the outside, but the inside cools second, goes into tension and puts the surface in compression.  Only then can Rupert's tear resist hammer blows.

I found an old (1975) patent that shows quenching can be applied to alumina (Al2O3):

Method of strengthening alumina ceramic by quenching with liquid medium

US 3887524 A

I wonder if this process has been researched and applied commercially

Well found, Eric.  Keep looking!

Defect free ceramics do (and will?) not exist. But this does not means that they can not support tensile stresses, especially in their bulk. When the compressive stress is only applied in a small part of the cross-section (near the free surface) the tensile stress can be kept sufficiently low. 

There is also this book by Henry Paul Kirchner - 1979:

Strengthening of Ceramics: Treatments: Tests, and Design Applications

https://books.google.nl/books?isbn=0824768515 -

- ‎Technology & Engineering

Chapter 2.2 STRENGTHENING HOT PRESSED ALUMINA BY QUENCHING Strengthening of pure hot pressed (H.P.) alumina by quenching differs in some respects ...

Dears all

I am glad that Erik has found a documented answer to his question.

Professor Laird,

Sir/ it is an immense honor to have a chance of a technical communication with a mentor , we all look up to, like yourself.   I have really enjoyed your works on the micro-plasticity ,i.e., dislocations and so on, of fatigue phenomenon of aluminum alloys while I was working on my MS thesis in early eighties with professors Raynold  Daniels , William Upthegrove , and Robert Block in U. of Oklahoma, Norman, OK.     

According to Prof. seen babiniec...

However, you should also consider how quenching may influence the crystallographic nature of the material. Have you run XRD before and after quenching to see if there is any change in lattice parameter, possibly due to some phase transition that didn't have time to relax upon cooling? Things of that sort can change the dipole moment, and thus the piezoelectric effect. Also, if you rely on point defects to bolster the effect, they can be largely influenced by annealing processes.

What is the influence of quenching on the ceramics? - ResearchGate. Available from: https://www.researchgate.net/post/What_is_the_influence_of_quenching_on_the_ceramics [accessed Feb 9, 2017].

Hi Erik! Just an idea I had: wouldn't it be easier to develop a bilayer solution where you put the front ceramic layer under compression using an adhesive ductile back layer (steel?)) that would be put under tensile stress (similar to the bended steel slabs I showed in the document attached to my previous post)? You would of course have to watch out for the parasitic bending stresses not to lead to fracture of your ceramic, but it might be feasible... It would of course in a way be similar to the steel-encased ceramic armour tiles already used and probably you are looking for an easier ceramic-only manufacturing route... Or maybe you can develop something similar to a prestressed reinforced concrete slab?

Extrinsic solutions (using other materials/systems around the ceramic itself) come with additional mass. The nice potential of pre-stressed (tempered) ceramics would be that the stress does not require added mass.