You are sure it is not a bulk RI change when going back to running buffer? This would agree with the concentration dependence (if the jump is proportional). But I guess that you also referencing this, right?
You should definitely add surfactant to your running buffer. Can you tell the bulk RI change from that other RI change you mention? If so, how do you do that? Have you tried injecting just running buffer?
Problem is P20 prevents the interaction between my analyte and ligand. As it goes I am using it for Regeneration solution (it works a treat on hydrophobic interactions). The dips either side of the injection result from the delay between RI changes in FC1 and FC2 and they match up in terms of RU change. The raise in RU that I'm seeing post injection is considerably more though. But yes we also injected running buffer only as well, which correlated nicely. I wonder if this issue might be caused by adsorbed aggregates in the IFC?