RI occurs in the fitting or while the experiment proceeding

I saw a high RI, account for over 30% in a ligand-protein binding from the "components" in a kinetics fitting. I am wondering this RI is just added into the fitting algorithm or it indeed happens in the real experimental procedure and then detected by the fitting program. If I set the RI to constant zero, why the original sensorgram was not changed? Does this mean there is no bulk effects in my samples binding? And why there are negative and positive bulk+drift?
Thanks for any comments and suggestions.

Hi,
Negative or positive bulk effects arise from differences in composition between the running buffer and the sample buffer. Positive when the sample buffer contains more salts and vice versa. Drift is almost the same but stems mainly from the running buffer. Negative or positive is the washing out or in of components out/into the matrix/sensor surface. Sometimes it is the matrix that is swelling or shrinking due to buffer differences.
You are right that bulk and drift are added to fittings to lower the overall Chi2 of the fitting. Remember, the computer has no clue what it is doing. Normally, I start fitting without bulk and drift. and only add drift when the drift is visible before the analyte injection start and only add bulk when it is apparent at the end of the analyte injection (keeping association and dissociation constant).
If the fitting is not changing I would say there is nu bulk and drift.
Kind regards
Arnoud