'double sandwich' is mostly or 'sandwich' is experimental at best, supporting and enhancing rigidity of such parts and do distribute the overall load stress a part can take before distorting or disintegrating structurally on exceeded load
The problem is carbon fibre, it is strong and ridged but does not flex well and cannot deal with shock forces as much as glass fibre or Kevlar or Aramid

That means if it flexes, twists, bulked out on load, it will break without warning

As for fatigue, I have seen load of carbon fibre bike frames separating the re-enforcement (breaking the epoxy bond) on high stress areas such as fork heads which have turned into balsa wood like substance

Have a look at that, I found it interesting
https://oscarliang.com/carbon-vs-fibreglass/

Hi Mike
That really depends on what axial, vertical and horizontal load, distributed over the length of centre point that beam in gone have to take
You are creating a box profile

Your bean has total area of 1800 square inches, 1.2sqare meter
Looking at your requirement of mx 0.8kg carbon fibre that makes no more than 900gsm material for the outside wall without internal bulkheads and pvc core installed.

Even that it is carbon fibre and assuming that both sides are anchored solid to prevent twisting I would estimate 150kg to 200 kg horizontal load on the centre before distortion and twisting will occur, gradual deforming at max load support can be ruled out, when carbon goes, it goes at a bang 

Overall hard to tell without knowing the use of the final beam

Hi fellow members,

I am building a structural rectangular carbon box beam 2.5” thick x 6.0” wide and 60” long. The structural objective is to equally maximize the specific bending (in both the vertical and horizontal axis) and torsional stiffness of the beam using no more than 800 grams of cured material. In the application for the beam there are no significant point loads and load distribution can be considered to be uniform across the entire surface of the object . Other factors such as the ultimate strength, fatigue strength, surface finish, layup, sourcing material, etc can be ignored for the purposes of this inquiry. At this point in the process I am only concerned with identifying the form that will produce the stiffest and lightest beam. As my background is more in the area of materials engineering, I am seeking the advice of forum members with structural experience. Having said that, my sense is a closed edge sandwich panel with uni-carbon faceplates and a carbon honeycomb core may be the best approach to achieve this objective. Would anyone disagree or suggest another approach?

Also, if a sandwich panel construction is the best approach, owing to the 2.5” thickness of the panel, I am having doubts about whether the standard “thin face sheet” sandwich panel construction will produce the lifghtest and stiffest result as opposed to a “double sandwich” design in which the face sheets of the main panel are also sandwich (perhaps with nomex core) panels albeit very thin - say 0.25” thick. My theory is that since “face panels” will be 10x stiffer thin face skins and only marginally heavier, this approach will allow the cell size of the main 2.5” thick panel core to be much larger and therefore much lighter. For the 2.5” core I am considering using “end-grain” pre-cured laminate sheet so my second question is whether the cell shape should be a hexagon. My sense is that a set of circular cells (short tubes) bonded to adjacent cells along their tangents would be lighter, have better bucking strength and provide superior shear transfer?

Thanks in advance for your consideration and assistance.