My experience with carbon fiber is mostly in cycling.

One of the drawbacks with carbon fiber is its tendency to fail catastrophically.

By that I don't mean that it tends to fail per se, but when it does it often doesn't do so gradually with warning like steel or aluminum (which warps, dents, and/or cracks visibly). So there's a certain fear about hidden cracks and failure points.

This gets exacerbated when talking about used items.

When talking about recycling or repairing carbon fiber, then, I think you'd have to convince people that it's just as stable as the original product, and that there aren't any hidden failure points. This obviously wouldn't apply to everything carbon fiber, but for things where it's often used, like in transportation, I think you'd have to have an inexpensive, foolproof, airtight way to verify integrity.

Another fellow cyclist here. Aluminium in fact has a very similar failure mode. It also fails catastrophically without warning. This was also brought against aluminium frames when they first started coming up to replace steel.

The bigger issue with carbon (in the context of cycling) is that there can be significant fabrication defects (voids) that are impossible to detect without special equipment. On aluminium frames you really only have to look and the welds to see if there are defects. Together with the fact that most (even premium brands like trek, canyon...) manufacture ultra cheap in China with very little QC (due to the race to the bottom regarding costs) , means you could get a significantly weakened frame straight from the dealer.

Trek had an 'affordable' hybrid composite bike, the Trek 2100, that merely twice as expensive as any reasonable hobbyist would pay for a bike instead of 5x as expensive. It was carbon tubes (they had patents on making them) and aluminum lugs, but the bond wasn't always good and so they would crack, chip, or delaminate. Big recall. The one guy we knew who had one rode it until it broke and then sent it back. That was a pretty long day for him (I can't recall if he rode it back gingerly or had someone pick him up).

I've done carbon fiber boat repair, and removal is pretty easy - lots of cutting and sanding. Applying a patch works pretty well - mechanical bonding to the roughed-up surface is pretty good, and even if the patch uses 2x the material of the original, it doesn't affect overall weight too much.

There are two major complication to repair.

First, carbon fiber items are made in a mold. Repair doesn't have access to that mold, so you've got to build your own. You can get a good approximation from casting an adjacent part, or the opposite side - otherwise you have to improvise and sand it to match. Catastrophic failure, like when the bow is broken more than a few degrees, requires an alignment jig to keep the boat straight.

Second, carbon fiber works best when it's a laminate - e.g. a carbon/foam/carbon sandwich. So most designs are tubular or involve a stackup. This is tricky if the inside layer isn't accessible. Or a high-performance inner layer is used, like delicate honeycomb aluminum. Rebuilding something that was custom-made to tight tolerances is difficult.

I'd think that a degradable epoxy would be best for recycling applications; repair has a lot of other challenges.

You would have to design the article to be repaired. The core material would need to be dissolvable so it can be refreshed along with the epoxy.

I imagine the repair process would be something like:

  1. Place bike/boat in mold/alignment frame. 
  2. Dissolve core then epoxy.
  3. Refill and cure core.
  4. Reapply epoxy to carbon fiber.
  5. Vacuum bag the bike/boat then cure epoxy.

A lot of composite artifacts are fairly easy to repair already, assuming you accept some degradation of performance since the newly bonded media and resin won't be chemically attached, but mechanically instead.

How often would you need to repair CF vs recycling it though? Most cars don't get involved in wrecks over time enough to compare with the number of cars that end up getting recycles as in probably 99% of them. Seems like that would be the real win. I'm just wondering how many solvents are created to complete this process and how they are disposed of.