We report on the analisys and projections of B4-GHOST flange scattering. The TF and projection is computed for three different target channels: MICH, DARM and Hrec. Also, 50Hz and multiples are notched in the accelerometer

• Figure 1 shows that coupling between DARM and MICH and hrec and MICH was low both right before the injection started and at the time chosen for the projection.
• Figure 2 shows the measured TF between the accelerometer on the flange and MICH (the TF is only computed where the excess noise in the target channel is at least a factor two)
• Figure 3 shows the measured TF between the accelerometer on the flange and DARM (the TF is only computed where the excess noise in the target channel is at least a factor two)
• Figure 4 shows the projected noise onto MICH  (in the red regions the projection is done using interpolated values of TF)
• Figure 5 shows the projected noise onto DARM (in the red regions the projection is done using interpolated values of TF)
• Figure 6 shows the projected noise onto Hrec (in the red regions the projection is done using interpolated values of TF)

The computation has been repeated with two independent codes giving similar results.

At some frequencies projections clearly exceed quiet spectra. We do not have a clear explanation for this. It can be that the TF is not stationary or there is something to be better understood.

With this caveat, the indications from these projections are:

• scattered light from the flange compromises MICH at most frequencies
• DARM and Hrec are compromised at 90Hz (significantly) and below 25Hz, and close to be limited around 130-140Hz.