Yesterday we found the small mirror surface of the beam expander was damaged, we figured out that it was caused by the laser beam filament or self-focusing. We improved the spatial filter vacuum, covered the optical tables, cleaned the windows, and filled the new gases into TWINAMP. All these behaviors might increase the laser output energy, which finally provoked the strong nonlinear effect such as filament or self-focusing.
We used the photo papers to imprint the laser beam patterns, one was put before the beam expander, the other one was put in front of the convex mirror in the beam expander box. Each print recorded about 300 laser shots at ~24 mJ. It's obvious that the beam print before the telescope was smoother than the one in front of the convex mirror. Some dark spots were recorded at the edge or in the bottom of the laser beam, which almost matched the damaged spots on the convex mirror.