I have first of all to make a note. Brittleness is not related to the value of strain-to-failure. Brittleness is related to the difference between strain-to-yielding and strain-to-failure. A stress-strain graph of a uniaxial tensile test for a metal (I refer to metals as historically these concepts of mechanics of materials were developed with respect to them) presents for increasing strain:

1. an initial linear elastic range;

2. a (usually short) non-linear elastic range;

3. a plastic range (with or without hardening);

4. a more or less developed softening region;

5. total failure of the specimen.

The point of change from 2 to 3, i.e. from elasticity to plasticity, is the yield point. Point 5 is total failure. The more point 5 is far on the strain axis from the yield point, the more the material is ductile. Viceversa, the more the total failure takes place close to the end of the elastic region, the more the material is brittle. It doesn't depend on the absolute value of the strain-to-failure, you do not compare different materials in terms of ductility/brittleness based on this single value.

Carbon fibers and carbon fiber composites present a strong brittle behaviour. What does that mean? That often, the end of the linear elastic region coincides with the point of complete failure. Why is this important? Structures (buildings, bridges, ships, cars, airplanes) are designed to remain in the linear elastic region while sustaining the loads due to normal operative conditions, i.e. the loads predicted to be applied to the system during service life. However, these estimations are of statistical nature. It means that there is always a chance that the structure is subjected to a load exceeding the maximum designed value. Such load will cause the structure to exit the linear elastic range. And now brittleness/ductility enters the stage. If the material is brittle, total failure of the most loaded part will soon follow after the linear elastic region, which might lead to a complete and sudden structural collapse. If the material is ductile, plasticity-related mechanisms will absorb the load in excess, avoiding collapse and allowing the survival of the structure.

To obtain a better understanding of your results, you should compare more data, not just the strain-to-failure. I expect the stress-to-failure (or load-to-failure) also to be greater for the single fiber than the woven laminate. The same for Young's modulus, i.e. the angular coefficient of the stress-displacement graph for a uni-axial test (uniaxiality is an important feature, as we are dealing with non-isotropic materials). Such a behaviour would be common: carbon fiber properties are quite exceptional due to its microscopic structure (crystalline structure stretched in the axial direction) and are in some way "lost" due to the mixing with the resin. To verify your results for the woven composites, you should make a theoretical evaluation of its properties. To do this, just apply the Rule of Mixtures for the plies and Classical Laminate Theory (CLT) for the whole laminate. They are widely used and provide good results on average (you can find a lot on them on Google or I can give you some material, if you need). In this way you can do a more meaningful analysis of your experiments and spot problems, if there are.

I hope this long answer might be of some help for you.

Best,

Luca

Dear Will,

The single fiber is subjected to only unidirectional strain, whereas the woven fiber is subjected to  multi directional strain.The woven fiber will try to straighten at the crossover points in the weaving. Also due to frequent  changes in the fiber direction  shear forces are introduced at the crossover points. ( By Cross over points I mean the location at which the fibers cross each other ). Hence single fiber will always give higher strain to failure than the woven fabric. For tensile loading this effect will remain the same irrespective of the matrix  as I think that the failure is governed by the weaving pattern.

A great detail about the subject is given in the above two answers. I just want to add the effect of rubbing of fibers against each other in the weave that weakens them.

Very good contributions above.  It is also important to understand your failure criteria. 

As mentioned above, due to the statistical nature of strength of fibers and the weave, the full strain achieved is expected to be lower.  However, failure of individual fibers within a ply (or even failure of plies within a laminate) is not necessarily considered failure of the laminate. 

When a fiber breaks, its load is redistributed around the location of the fracture by the matrix.  Thus the two fibers created from the single original fiber continue to carry load, increasing from zero at the fiber ends to approximately the fiber stresses in the rest of the laminate (subject to the statistical variance mentioned above). 

The first fiber breaks quite early in the loading, but because it continues to carry load, the only effect at the laminate level is an infinitesimal reduction in elastic modulus.  Under normal operation, the first fiber failure is not noticeable, and typical loading of a laminate may be above the first fiber failure. 

That said, one would expect the rupture strain of a fiber/epoxy composite to be greater than that of an individual carbon fiber, subject to the elongation available in the epoxy and the fiber/epoxy interface strength.

I should add that failure analysis is important here.  In particular, you should ensure that

1. Failure is not occurring near the machine grips.

2. There is some retained epoxy on fiber ends (cohesive failure rather than adhesive failure).

It may also be worth noting the length of fibers in the failed composite.  These ideally would be near the theoretical "critical length" where the load transferred by the epoxy in shear is approximately equal to the load-carrying capacity of the fiber (or in this case, weave).  This would indicate that the bond between the fibers and the epoxy is a strong one.