Measuring SOI device layer thicknesses

Hi all,

     I'm trying to measure the Si device layer thickness on an SOI wafer with a known 1-micron BOX layer.  Using the litho nanospec program #4 (default RI=3.973), I'm getting values that I suspect are 10-15% too high compared to step-height test measurements on the same wafer.  Does anyone have a different methodology or "special" RI that they regularly use for 1-micron BOX layers?

     Due to the densely patterned nature of my wafers, I can't really use the ellipsometer or the Woollam.

Thanks,

Ben