After Pengjia finished his BPM study, we start to use the empty target to decide the maximal raster size. We found that if we increase the slow raster size to 0.78V(1.88cm), we could see hot spots on both up and down side in 3rd arm spot++ plots. We reduced the slow raster size to 0.70V to leave some space to avoid the fluctuation on beam position. So the final setting of the slow raster size is 0.70V(1.66cm, vertical)x0.90V(2.12cm, horizontal). The new beam position is A(1.1,-6.1), B(1.2,-6.3).
We then tried to use the carbon hole target to center the beam. With a target offset 160000 (5.3mm down), we could center the carbon hole to the slow raster pattern. However, When we moved back to the C12_40 target, we saw the beam scrapping the top of the C12_40 cell. So Pengjia increased the raster size along the horizontal direction and tried to use the shape of the cell to center the beam. The final target offset is 120000 (4.0mm down). Tested on NH3 target, it works well.
Since the beam is not stable, it took us several hours to finish the procedure. We moved to production at 02:15.
Pengjia have already posted a series of halog, which contain all of the spot++ plots. You could find more information from them.