"Let's put carbon fibre in there," says the marketing director. "That stuff's stronger. And I can sell it as a high-end feature."
"Yup, yup," replies the shop foreman. "We can do that. It's a bit pricey though, maybe we could use just a bit of it mixed with the fibreglass."
Fast forward three years, and both men are scratching their heads over why the component- which was, according to the designer, more than strong enough in fibreglass alone- has failed catastrophically even though they added a "better" material.
TL;DR: Mixing different fibres in the same load path can lead to a component being weaker than it would be if only one type of fibre had been used.
To understand why this is, we must consider the raw mechanical properties of the individual materials. For this example, we'll use E-glass (the standard grade of fibreglass used for boat construction) and Hexcel UHM 3K carbon fibre (a grade whose properties are typical of the carbons used in boats, but that is of considerably lower quality than aerospace grades). The properties of interest are:
- Modulus of elasticity (Young's modulus), which describes how stiff a material is. Doubling the Young's modulus means that it'll stretch half as much under a given load.
- Ultimate tensile strength, which is the stress (force per unit cross-sectional area) that will cause the fibre to fail in tension.
- Elongation at break, which describes how much (as a percentage of original length) the fibre can stretch when pulled before it breaks. (Exercise for any sudents reading this: You should be able to figure out how to predict elongation at break if you know the Young's modulus and the ultimate tensile strength.)
|Carbon fibre (Hexcel UHM 3K)
|Modulus of elasticity
|Ultimate tensile strength
|Elongation at break
(I should insert a certain disclaimer here: These tables are not for design purposes, and you should be seeking project-specific engineering advice if you're actually building a high-tech composite assembly. This blog is not an engineering manual.)
Consider two identical components, both made from unidirectional fibres and both loaded primarily in tension. One is pure fibreglass; the other is 50% glass and 50% carbon. (We can, for the moment, neglect the epoxy matrix that forms roughly 60% of the volume of the part, as it is the fibres themselves that will bear almost all of the load.)
In the fibreglass part, the tension force is spread over the entire cross-sectional area of the part, leading to a stress (= force / area) on those fibres. If the part was properly designed and built, this stress will be much less than the ultimate tensile strength of the fibreglass, and the load can be safely carried.
In the mixed carbon/fibreglass part, the tension force is not spread uniformly over the entire cross-section. The fibreglass, which has to 'give' a bit to accept a load, wants to stretch, but it's held back by the six-times-stiffer carbon fibre. The carbon, then, takes almost all of the load, leaving the fibreglass nearly unloaded. In essence, the cross-sectional area of the part is reduced to the cross-sectional area of the carbon fibre; the fibreglass can't come into play unless it is allowed to stretch slightly, and the stiff carbon has already taken most of the load before that can happen.
We are applying the same force in both cases, but the stresses are not the same. If the stress in the pure-fibreglass part is a nice uniform $X$, the stress in the mixed-fibre part alternates between nearly $2X$ in the carbon regions and minimal in the glass regions.
As the force increases, the part will elongate. The strain (the elongation relative to no-load condition) reaches 0.8%, and while the fibreglass half is working at about one-sixth of its maximum stress, the stress on the carbon half edges past 3700 MPa. BANG! The carbon components of the mixed-fibre part fail, suddenly transferring all of their load to the fibreglass. In a tiny fraction of a second, the (now grossly overloaded) fibreglass also fails.
Meanwhile, the same force on the pure-fibreglass part is safely distributed, uniformly, over all the fibreglass strands.
The moral of the story? Don't mix materials with radically different mechanical properties in parallel in the same load path. The load won't be distributed evenly; rather, the stiffest material will take almost all of the load until it fails, transferring the load to the next-stiffest material.
Carbon is still pretty awesome
This should not be taken as a slight against carbon fibre. Carbon is an amazing material and, when used properly, produces some of the lightest and stiffest parts available. (Indeed, its incredible stiffness is carbon's single biggest advantage; it is this property, in combination with its light weight, that allows carbon parts to be so much lighter than conventional materials in stiffness-limited applications like masts and wing spars.)
I'm also not intending to knock the idea of using multiple composites in one boat. There are many cases where carbon and fibreglass components can get along very well together- a rudder with a carbon stock and a fibreglass foil, for example, can be a very sturdy and reliable combination. (The key here is that the two materials are handling different types of loads along different load paths; the forces are transferred from one to the other rather than being shared in parallel.)
If you're going to start mixing materials, though, you have to understand the load paths involved. Each material has to be used where its properties will do the most good, and the way loads will be shared between materials must be carefully analyzed; mixing them willy-nilly is likely to lead to unexpected failures like the one described above.