Just thought I’d chime in on this thread as well as I’ve been dealing with this big time. I’ve been trying to use a PLA support interface with PETG, and the PETG layers after laying down the PLA interface always break. I upped purge volume to 999 for all slots and the strength definitely does improve. Another thing that did seem to help was changing out the nozzle wiper to a silicone wiper. I tested this by using the auto calculated purge volumes with and without the silicone wiper and I had fewer failures when using the silicone wiper. You can tell that there is some definite contamination going on, because the side of the object (printed in PETG) has some slight sheen differences when changing from PLA > PETG. These sheen differences become less with a higher purge volume. I wish it was possible to increase the purge volumes even more. If you’re just using a different filament for the interface it isn’t a huge loss of material and it would ensure layer adhesion strength. For now I just run max purge volume and turn down the outer layers to 50mm so the sheen difference isn’t noticeable. This allows for strong layer adhesion and the part to look good.
Curious did you experiment with the wall order and increasing the infil percentage and line width. If for example you chose infill/inner/outer and a larger line width, the infill would help to clean any reminants from the hotend after purge and because infill is random, any remenant of the support interface material would be spread throughout the infill. Then by theory, the inner and outer walls should be clean base material with the correct strength and bond.
For example on my klipper CoreXY slicing with Superslicer, I run a 160% infill line width and 105% perimeters (using a 0.4mm nozzle) . This gives me thick strong infil and very strong parts coupled with 4-5 outer perimeters, even when printing multiple objects.