I'm having some issues with COPs where as soon as I try to perform a channel copy, with both images set to 16 bit integer depth, the resulting ROP file output changes the image to 64bit. Does anyone know a fix for this as I would like to keep it at 16.

In my case it always outputs 16 bit floats regardless of settings. Can you post a file? It's entirely possible there's a bug. I've found a few in COPs and if you submit them they usually fix it quickly.

Turn off "Linearize Images" and you'll get the native 8b depth. This option promotes the image to 16b FP when enabled, otherwise you'd get banding. Note that it will appear differently due to the lack of linearization. The other approach is to force the bit depth back to 8b in the Composite ROP.