Haven't you pretty much answered your own question with the first post? I.e, that the reason is that compression itself does not cause failure, but that the tension perpendicular to the applied load due to Poisson's ratio (nu) causes failure, which means the fracture toughness should be 1/nu times higher in compression than tension? Or are you asking for a different explanation because you observe an even bigger difference in fracture toughness experimentally?

By "stronger", we shall assume here that brittle materials fail, by the propagation of cracks, at a stress level higher in compression than in tension.

      In tension, the propagating crack tends to orientate perpendicularly to the applied tension direction; in equilibrium, denote by SigmaT the stress at the beginning of crack motion. In compression, the identical crack tends to align parallel to the applied compression direction. The stress to move the crack is SigmaC=sigmaT/nu, where nu is Poisson's ratio (see P.N.B. Anongba, J. Bonneville and A. Joulain, Brittle cracks under compression: introducing Poisson effect, Revue Ivoirienne des Sciences et Technologie 17 (2011) 37-53) . In isotropic materials, nu the Poisson's ratio is equal to 1/3, so that SigmaC=sigmaT/nu=3*sigmaT. Consequently, in isotropic materials, three times larger stress is required to break a fracture specimen in compression as compared to tension.

     By passing, let us say that Poisson's effect, common to most materials, whose importance is demonstrated in two recent papers on the propagation of the interface crack loaded in tension and in mixed mode I+II (P.N.B. Anongba, ResearchGate) would be considered seriously in the fracture of composites under load because it induces internal shear stresses in the surrounding media about the interface.

This is something to do with the microstructure, atomic distace etc.

Just as Poisson's effect whose physical origin is associated to the existence of "transverse" microscopic bonds (transverse with respect to the applied loading direction).

Embrittlement of materials has to do with impurities in the materials that

migrate to grain boundaries, and the shape of those impurities.  The 

molecules are either oblate , ( like the shape of the spinning earth ), or 

they are prolate ( the shape of a medical capsule ). Oblate molecules

embrittle the grain boundaries, while prolate molecules strengthen the

grain bondaries.   This has to do with the change in the electrical forces

at the boundaries as the molecules roll over during shearing. oblate

molecules wedge the grain boundary apart when they roll over, while

prolate molecules pull the grain boundaries together when they roll over.

The Book " WHY THINGS BREAK " gives a history of one man's study

of this during his lifetime.

Your proposal deals with the origin of fracture initiation but the failure of the material is concerned with the propagation over large distance of the crack under an externally applied stress. The origin of "stronger" is in the latter.

We should add to your proposal a mechanism that is associated with the propagating crack over time both in compression and tension.

Cracks have very high stress concentrations at the tip of the crack where

only a few molecules are carrying the majority of the load ( applied force ).

The uncracked areas distribute the load over millions of molecules such

that each one carries a small portion of the load.

I recommend the book:

WHY THINGS BREAK

UNDERSTANDING THE WORLD BY THE WAY IT COMES APART

Mark E. Eberhart

Professor of Chemistry and Geochemistry at the Colorado School of Mines

ISBN 1-4000-4883-4

bar code: 9 781400 048830        51295    ( $ 12.95 )    ( 17.95 Canada )

Three Rivers Press - New York

The alternation of compression and tension is termed cyclical

loading, which leads to fatigue failures especially in certain

materials like the aluminum skin of aircraft,  when it is repeatedly

pressurized and depressurized , or when you bend a paper

clip back and forth until it breaks.

I would not recommend this book if it would not answer all of your 

questions much better than I ever could answer them.

We are not dealing with alternation of the load between tension and compression, don't forget the Q&A question: Why are brittle materials stronger in compression than in tension? Conditions for crack propagation in compression must be asserted and compared with those in tension. In compression, cracks are aligned parallel to the direction of the applied loading after fracture propagation over large distance;  in tension in the same situation, the crack is perpendicular to the applied loading. We have to compare the stress at fracture in both cases.

What are the possible physical origins of the crack extension force in compression, when the crack is aligned parallel to the applied compression direction, here are the keys for our question.

Haven't you pretty much answered your own question with the first post? I.e, that the reason is that compression itself does not cause failure, but that the tension perpendicular to the applied load due to Poisson's ratio (nu) causes failure, which means the fracture toughness should be 1/nu times higher in compression than tension? Or are you asking for a different explanation because you observe an even bigger difference in fracture toughness experimentally?

Now that I understand your question.

Lets look at it as a short fat  rod in compression.  Take a unit volume

at any location along the rod.  The unit volume remains a unit volume.

The length gets shorter, and the cross-sectional area get larger.

When the area increases, the radius is partially in compression, while the 

circumference is in tension,  Cracks form at the perimeter parallel to

the axis of the rod since the maximum force in tension is at the surface

of the rod.    Looking across the diameter, the  internal forces actually

go from tension to compression, and then back to tension in any and every

diameter direction.  The  perimeter tension forces are quite high on the

outer surface of the rod.

Rolled up sheets of two colors of modeling clay will help in understanding this

when an axial load is applied.  actually a laminate of modeling clay,

and thin aluminum foil will work better. The aluminum foil should rip

in the outer layers.

This give an idea as to why you previously saw tools with wooden

handles that has wire wound around the handle near the head of the tool

to help prevent the cracking of the wood .

To Manuel Schnabel: What happens with zero or negative Poisson's ratio? Poisson's effect contributes certainly but I wonder whether there are other possibilities.

This is related to the chemical bond potential well shape and its asymmetry: under compression, the atom behaviour is similar to hard balls and hence a high level of compression is possible; in tension the lengthening of the iono-covalent bond is very limited; see e.g.

Ph. COLOMBAN, Analysis of Strain and Stress in Ceramic, Polymer and Metal Matrix Composites by Raman Spectroscopy, Advanced Engineering Materials 4 [8] (2002) 535-42

The presence of a crack in a material modifies completely the applied stress field. What is the stress state in the cracked material? What is the relation between the stresses at failure in both cases (i.e. in tension and compression) once an internal crack has been created? Again, it seems to me that the possible physical origins of the crack extension force in compression, when the crack is aligned parallel to the applied compression direction, are the keys of our Q&A question. It also seems to me that our Q&A question involves the propagation of cracks, i.e. the failure of the material.

I would also like to say that Poisson's effect has a link with microscopic bonds (atoms, molecules ...)

To Philippe Colomban: can I have a copy of your paper : Ph. COLOMBAN, Analysis of Strain and Stress in Ceramic, Polymer and Metal Matrix Composites by Raman Spectroscopy, Advanced Engineering Materials 4 [8] (2002) 535-42

Let us follow our Q&A: How can one break brittle materials (with zero or negative Poisson's ratio) when they are loaded in compression? What are the associated expression of the crack extension force and its physical origin?

Tractions in the direction perpendicular to the compression direction are mostly due to material heterogeneities, not Poisson ratio (think of the case of a material with zero Poisson ratio and a whole in it).

Assume that we start with an homogeneous state of the material; what is the origin of the crack extension force? These heterogeneities, you suggest, would be distributed throughout the broken solid (at least along the path of crack propagation) with again an associated Poisson's tension. Actually Poisson's effect can have different possible origins.

Patrick,

I would argue that the answer is much more fundamental than everything that is being discussed here. At least in bcc materials, which are known to be brittle below room temperature, the tension/compression discrepancy is due to the twinning/antitwinning asymmetry. This asymmetry makes the sign of the resolved shear stress matter, as dislocation glide planes offer different resistance to positive applied stresses than to negative ones. Having more difficulty moving, screw dislocations create back stresses on crack tips that make them go easier.

I find no much more thing fundamental than Poisson's effect in materials physics. Our proposal is based on the way cracks behave under stress, after fracture propagation over macroscopic distances, in brittle materials in both tension and compression. I have nothing more to argue in addition, personally. May be there are other reasons in some particular cases.

To Jaime: Have you an answer to: What is the physical origin of the crack extension force when the crack is aligned parallel to the applied compression direction, in compression testing of brittle materials?

Dear

As per my research experience i observed that at micro level crack is initiate and further propagated through weak region, in tension porosity at micro or even nano level crack start and in brittle material it further propagates, whereas in compression test presence of porosity effect less when compared with tensile testing.

Closure (tentative?)

We are looking at cracks that have propagated over large (macroscopic) distance in large brittle isotropic materials under an externally applied loading. The natural orientation of the cracks depends on the applied loading direction.

a) Applied tension

The crack is perpendicular to the direction of the applied tension.

b) Applied compression

The crack is aligned parallel to the compression direction.

What is common to both situations?

The crack is perpendicular to the internal tension field. The internal tension field in compression originates from the Poisson's effect.

When crack propagation is the controlling mechanism of failure, three times larger stress is required in order to bring to failure isotropic material in compression (when comparison is made with tension test).

The discrepancy between tensile and compressive strengths is in part due to the brittle nature of ceramics. When subjected to a tensile load, ceramics, unlike metals, are unable to yield and relieve the stress. Another important factor is the presence of internal flaws from which cracks can propagate in tension, but not in compression

Right,equally well, cracks can propagate from flaws in ceramics under compression. The crack opening displacement is due to internal tension stresses. This includes necessarily internal Poisson tension when the crack is aligned parallel to the applied compression direction, as observed after fracture propagation over large distance in brittle materials including ceramics.

Patrick N.B. Anongba     can you give me your email or facebook 

Please,take it, from recent papers.

A quick question about compression and crack propagation if you can.

I am currently doing my thesis and looking into compressive cracking in aluminium alloy. Compressive buckling at cell level keeps coming up as a reason for cracking and ultimate failure.

Any comments for additional avenues to explore?

Thanks for any help in advance.

Look at concrete, it is ten times as strong in compression than tension

Please, mention the experiments! Poisson's extension may be inhibited in appreciable parts of the fracture specimen adjacent to the grips. Ashby and Hallam (1986) noted with insistance that the behaviour of the growing crack is extremely sensitive to the condition of the ends of the sample. End effects on samples (about the grips for instance) may prevent Poisson's extension and increase the applied stress level at failure (comments have been provided in our paper).