One of the main coupling sites for >1 Hz vibration (e.g. anthropogenic noise) at both LHO and LLO is at EY. While coupling was reduced by work on the p-cal periscopes, there appears to be a secondary source that dominated after the improvements.

I made a movie of the inside of the LHO EY test mass chamber while I was making end-of-run PEM injections (at full sensitivity), and noticed light from the cryo-baffle barrel flickering as I made impulse injections (see figure below, and movie clip: https://youtu.be/ZSNVuvWRpl0 ). 

Most of the barrel, including the brightly lit region, is visible from the beam spot on the ETM (see beam spot view photos on first page of figure). In these beam spot view photos, there was a bright reflection on the barrel of the closed gate valve, so I did not notice that the barrel was also directly reflecting light from the flash, and thus that it might directly reflect light scattered from the beam spot on the test mass. The light is at grazing incidence in the barrel cylinder, so it is mainly the imperfections that would retro-reflect. Unfortunately, the imperfect fusion of the two pieces making the cylinder is right where the main beam is closest to the barrel, increasing the reflected light. 

The second page of the figure shows that movie pixel value fluctuations increase during impulses and persist for a long time, like the noise in DARM. As a control for things like impulse-induced camera movement, I compared pixel variation in the barrel region of the movie to the ACB region of the movie, and found that the correlation with impulses was much better for the barrel region, suggesting that impulses cause the barrel to flicker more than other regions.

The evidence that the barrel is moving on wavelength-scales (producing visible flickering), when other surfaces aren’t, and that it has a high Q, like the noise in DARM, and that it has a bright spot on it that is visible from the test mass, all point to the barrel as being a strong candidate for our residual scattering noise at EY. The barrel may also be the source of EY scattering noise at LLO.

Extending the P-cal baffle piece to cover the barrel on the side where the beam is closest, might work to block the scattered light, but we should be careful that this doesn’t cause clipping noise. The high-Q of the baffle at relevant frequencies also means that it is a good candidate for damping, and that damping alone may make a significant improvement, as it did for the Swiss cheese baffle.