I’ve actually been looking into this quite a bit recently as part of some work on bonded titanium/carbon and Ti-Ti bonding frame construction.
My feeling is that something like a cable port is probably a very good candidate for adhesive bonding. Bottle bosses are slightly more complicated, because although they’re obviously not primary structural joints, a full bottle bouncing around on rough ground can put a surprising amount of cyclic and peel load into the boss.
The main thing I’ve taken away from looking into structural bonding is that you really need to design the part as a bonded joint rather than simply replacing the weld with glue.
Adhesives generally like large areas loaded in shear and don’t like peel loads or stress concentrations. So for a printed cable port or bottle boss I’d be looking at a decent flange/overlap, radiused transitions and ideally some geometry that gives a degree of mechanical retention as well. A sort of top-hat shaped bottle boss with a large curved flange could work nicely.
I also wouldn’t make the fit too snug. You actually want a small, controlled bond line rather than a press fit which squeezes all the adhesive out.
Surface preparation seems to be absolutely critical with titanium: degrease, mechanically abrade/grit blast the bonding surface, clean again and bond reasonably soon afterwards. There’s plenty of aerospace history of bonding titanium successfully, but the reliability seems to depend heavily on preparation and joint design rather than simply choosing the strongest epoxy.
For adhesive I’d be looking at a toughened structural 2-part epoxy rather than a general-purpose epoxy. Something along the lines of 3M DP490/DP420/DP460 or Loctite EA 9466 would be where I’d start, then speak to the adhesive manufacturer about the exact application. 3M actually publishes titanium test data for DP490, so this isn’t particularly exotic territory.
Personally I’d probably stick with titanium for the printed inserts rather than aluminium. In theory the adhesive layer can electrically isolate aluminium from titanium, but once you introduce water ingress at the edge of the joint you’ve potentially created a galvanic corrosion problem that doesn’t really need to exist. Adhesive bonding is commonly used to isolate dissimilar metals, but maintaining that isolation is part of the joint design.
The other advantage, as you say, is avoiding another weld/HAZ around a hole in an already busy area such as behind the head tube. I don’t think welding a properly designed port there is inherently problematic, but bonding gives you another way of doing it without introducing heat or distortion.
Cable ports I’d be pretty comfortable experimenting with. Bottle bosses I’d want to make some test pieces and properly abuse them first — pull-out, sideways loading and repeated cyclic loading with a weighted cage — but I don’t see anything fundamentally wrong with the idea if the boss and bond are designed together.
All this is purely theoretical from an experience perspective…but might be a good starter for others?
Adding a drop of silver to a bottleboss versus making a tiny yet tricky weld that takes quite a bit more time.