Just before the pandemic shutdown, I noticed that scattered light from the EY manifold/cryopump baffle (cryobaffle for short) modulated with impulse injections, and this modulation lasted as long as the noise in DARM (https://alog.ligo-wa.caltech.edu/aLOG/index.php?callRep=55927 ). This is pretty strong evidence that the cryobaffle is the source of the scattering noise at LHO EY because not many objects have Qs high enough to account for the persistence and slow decay of the scattering noise in DARM after an impulse, but the movies showed that the cryobaffle motion did persist long enough.

Since the recovery, Anamaria, Valera and others have been studying the cryobaffles at LLO as potentially the dominant source of anthropogenic noise at LLO. They have found that scattering sources at several cryobaffle locations have dominant resonances at about 4 Hz and high Qs: EY (4.6, Q~400 https://alog.ligo-la.caltech.edu/aLOG/index.php?callRep=53069 ), EX (4.1, Q~2000 https://alog.ligo-la.caltech.edu/aLOG/index.php?callRep=53185  and ITMY (4.19 Hz, Q~2000 https://alog.ligo-la.caltech.edu/aLOG/index.php?callRep=53364 ). The detection of scattering and ~4 Hz resonances at each of these cryobaffle locations, especially for the very different environment at ITMY, increases the confidence that the cryobaffles are the scattering source at ~4 Hz. The scattered light may come from the edges of the barrel (https://alog.ligo-la.caltech.edu/aLOG/index.php?callRep=53361) or from the inner surface since the barrel would only have to be cocked by about 1 degree in order to put a specular reflection of the ITM onto the ETM.

There was little time before the shutdown when I first investigated the cryobaffle at LHO EY, and so, in the past week, I used slow sweeps to identify the resonances of the scatter source at LHO EY (Figure 1), and have more precisely measured the Q for comparison to LLO.  The dominant noise-producing resonance was at 3.85 Hz, with a Q of 969, but there are other lower-Q resonances at 5.3, 6.4, 7.1, and 9.7 Hz (Figure 1).

In addition, I made movies while injecting at the dominant resonance, 3.85 Hz. Figure 2 shows a spectrum of the standard deviation of movie pixel values that has a strong peak at 3.85 Hz. This demonstrates that the movie analysis can detect cryobaffle motion and strengthens the previous observation that variation in light from the baffle lasted as long as the noise in DARM, pointing strongly to the cryobaffle barrel.  The light modulation at 3.85 Hz is quite visible in the movie: https://youtu.be/743GMCZXY7Y

As with the similarly high-Q swiss cheese baffle (https://alog.ligo-wa.caltech.edu/aLOG/index.php?callRep=35166), good damping may be enough to reduce the cryobaffle scattering shelves to an acceptable peak frequency. Along with the swiss cheese baffle, this is another reminder that we should be sure to damp all new baffles so that their motion doesn’t get amplified by a high Q.