4Map Engineering Run 5262013

I performed the standard donut analysis, and compared the astigmatisms against a reference image to derive an overall change in the value ofthe two astigmatism Zernikes. These are combined to calculate a magnitude
of Astigmatism, changing units to [nm] to match CT2TWK. I do also adjust the zero point of the astigmatism using the first 2 measurements taken at CT2TWK=0.

The donut measured astigmatism vs. exposure id clearly shows the run up and then down in the astigmatism tweak. The first two 1d56 points (at ct2twk of 1500) were marked as bad - these show up clearly in the plot. There are also two other problem points - those are the 1d64 points (at ct2twk of 1000) --
I wonder if these also had the same 4MAP problem?

Now plotting the Donut values vs. the Ct2Twk shows for the run up and
run down separately:

So there is nice linear dependence between the donut measurements and the Ct2Twk!
I would conclude from this that there is certainly sufficient resolution in the
4MAP system to make fine corrections to the astigmatism.

Then looking at the trip down:

The behavior looks linear for these values as well (removing the pair of questionable points). But the measured astigmatism at CT2TWK=0 is noticeably different than the starting value, with a value around 500nm -- indicating that there is either some hysteresis or drift in the system. Most of the pairs of points agree to better than this level, so it is unlikely to be due to donut measurement error alone.