Carbon Steel surface prep for Carbon Fiber Composite

[phoenix221]

Hi all,

I am facing a problem with ensuring that a multi layer carbon fiber composite properly bonds to a carbon steel surface (one side only) and it stays bonded while going through termal cycles going as low as -25F to 150F. I would also like to avoid any cathodic corrosion between the steel and the carbon fiber while rataining a high heat transfer rate through the bond.

What mechanical/chemical surface prep should be done, and do I have to use a special adhesive as the bonding layer or can I just use the epoxy used as the matrix?

Thanks,
M

 

[Compositepro]

The main difference between a matrix resin and an adhesive is the resin modulus or stiffness. Matrix resins fill microscopic gaps (on the order of one micron) between filaments. There is relatively little strain on the resin since it cannot strain more than the fiber. An adhesive is designed to carry load across large area interfaces and stresses in bondlines are often non-uniform. Epoxy bondlines are usually 5 to 10 mils thick (125 to 250 microns). This is to allow for some deformation in the adhesive while under stress so that stress can be spread over a larger area. It is also ideal for an adhesive to be capable of large plastic deformation after the yield stress is exceeded. A one mil bondline is much weaker than 5 mils. Very thick bondlines of high modulus adhesives will also be weak. Adhesive also generally have higher viscosity than matrix resin to help control bondline thickness, but this is not always true.

You can cure directly to steel but in most cases the stresses are such that an adhesive layer should be used. The streses caused by thermal expansion difference can be very great and may be beyond the ability of any adhesive. If the parts are large and there will be large strain differences between the parts it might require using a low modulus adhesive like silicone or urethane and very thick glue line, like 1/8".

As for electrical isolation of the parts, spacers can be included in the glue line. For thin glue lines spunbonded polyester like Remay is used. For thicker glue lines Open weave fabrics or glass beads in the glue are used.

 

[phoenix221]

Compositepro, thanks for the answer...

> If the parts are large and there will be large strain
> differences between the parts

The part is a steel rod of .56" diameter surrounded by a carbon fiber sleeve which cycles through a range of -20F to +200F. While it heats up the carbon fiber should contract with a thermal expansion near 0 or just slighly negative so I don't foresee the problem there. What concerns me is the -20F end of the operational range where steel should contract faster/more than the carbon fiber sleeve.

> An adhesive is designed to carry load across large area
> interfaces and stresses in bondlines are often non-uniform

I am still unsure whether the adhesive may be a way to go. Will the adhesives mentioned adversely effect thermal transfer? It is a thin layer of adhesive for sure but one never knows...

> As for electrical isolation of the parts, spacers can be
> included in the glue line

This all sounds very complicated and probably beyond my manufacturing capacity. Is there no way to coat the steel first with something, e.g. blue it. I've also been suggested to use stainless steel which should not have the same problem, is this true? I could use a layer of fiberglass between the carbon and the steel but I am worried about that approach compromizing the good heat transfer that I am looking for :-(

 

[Compositepro]

It sounds like a gun barrel. In any case your big problem will be the length dimension which is probably much longer than your diamater. Fiber orientation and ply thickness will be major factors in residual stresses. Making thick-walled carbon fiber cylinders is difficult. I've seen many parts at customer companies that delamintated on cool-down from cure because of internal stresses. There is a big difference in CTE between fiber direction (hoop or axial) and through thickness (radial) CTE. Resin shrinkage during cure is also mainly radial.

You can put a rough surface on the steel interface to get a strong mechanical lock with the rod. You can eliminate any corrosion issues by coating the outside surfaces of the steel and composite and keeping thing dry. Steel will rust without contact with carbon if it isn't coated or kept dry. Carbon contact will only accelerate the problem, and only if there is a complete electrical path through electrolyte from carbon to steel.

In another post I believe it was you who asked about improving through thickness thermal conductivity. You might consider twisting the carbon fiber tows. Axial conductivity of carbon fiber is much greater than radial conuctivity. Twisting the fiber will mean that axial heat flow will cause heat to flow though the ply thickness as well. Twist also makes the fiber easier to wrap around a rod. Twisted fibers "rope-up" under tension, however.

 

[phoenix221]

Thanks Compositepro.

I was trying to keep the issues in separate posts for future searchers of the forum who may not be interested in the compound problem...

> You can put a rough surface on the steel interface to get a
> strong mechanical lock with the rod.
Will sand/bead blasting do it? ... then, I am thinking of skipping the whole adhesive idea and just going epoxy all the way :-0

 

[Compositepro]

I would suggest knurling as the easiest for a rod and also strongest bond. Coarse sand or bead-blasting will also work, or dimpling with a punch. Even putting a fine thread groove woul work.