Nanoprobe Network

While acquiring f-d curves in tapping mode, recently, I have been seeing a continuous reduction in the oscillation amplitude while approaching to and retracting from the sample over a range of 5 micron distance. This reduction in amplitude happens before the tip and sample begin interaction with each other; where we see a sharp drop in amplitude. These probes that I'm using have kind of low aspect ration tips than the regular silicon probes. It seems that there is an unknown source of damping for these probes that doesn't exist for the regular silicon probes. I'm sure the AFM itself has no problem as the regular silicon probes f-d curves show a constant amplitude before the initiation of the tip-sample interaction. I can't figure out what could be the source of this energy dissipation. I have tuned these probes far from and at distances close to the sample and realized that the closer the probe gets to the sample, the lower is the quality factor, and the higher is the drive amplitude required to achieve a certain oscillation amplitude. I am wondering if anybody has seen this before or there is a publication that deals with this problem.

Thanks
Vahid