Evan Goetz, Alan Knee, Ansel Neunzert (on behalf of line list contributors, see LIGO-T2400204)

Summary:
Version 1 of the lines and combs list of non-astrophysical artifacts yields insights into the O4a H1 data about long duration narrow spectral artifacts. In summary, H1 data has nearly 7x more of these artifacts than L1, mostly due to the much higher proliferation of comb artifacts. H1 data also suffers from a large number of narrow artifacts in the ~100 Hz bands near test mass violin resonance harmonics, usually during times of rung up violin resonances. Without careful handling of the data in frequency bands near the violin modes, nearly 18% of the 10 Hz - 2 kHz band is contaminated, though more careful handling results in about 6% of the band being contaminated (roughly equivalent to L1). The L1 detector does not suffer from these same artifacts.

Details:
After compiling a list of vetted, non-astrophysical artifacts into the version 1 lines and combs list for O4a, we extracted some statistics for each detector in order to compare and contrast the data quality from this perspective, and to determine which of the artifacts are most problematic for long duration narrowband searches (CW or narrowband directed stochastic searches).

New for this observing run, we have two new features for the lines list: 1) artifacts with known time variation (turn on/turn off) where we can establish reasonable metrics for their presence in the data we also provide a segment list for times those artifacts are present; 2) a new type 3 artifact that is not simply a narrow frequency band for a single artifact, but rather a band (typically 1 Hz or larger) with many narrow artifacts degrading the data quality within the band.

Type                         H1        L1
-----------------------------------------
Total number of artifacts   1721      251
Number of vetted combs        26       16
Comb lines                  1510      168
Type 3 artifacts              29        0
Vetoed band percentage
(excl. type 3 artifacts)    5.8%      6.3%
Vetoed band percentage
(incl. type 3 artifacts)   17.9%      6.3%

Most problematic H1 artifacts:
1) Type 3 artifacts: non-stationary line artifacts impacting ~100 Hz bands around TM violin resonances - partial understanding of some of these artifacts due to calibration line mixing with violin resonances (future aLOG), but not all artifacts are understood
2) 9.5 Hz comb triplet: comb of line triplets present from 10 Hz to 2000 Hz (seen weakly at L1) - potential cause identified in PSL flow sensors (LHO aLOG 79533)
3) ~11.11 Hz comb seen in both H1 and L1 - not understood. It was also seen at H1 in O3.
4) ~11.9 Hz comb seen in both H1 and L1 - not understood. See also checks on different changes. Seen also in O3 H1 data.

H1 artifacts that have been mitigated:
1) Hartmann wavefront sensor (HWS) comb mitigated near 1 Hz, near 5 Hz, and near 7 Hz combs
2) OM2 heater driver comb mitigated ~1.66 Hz comb (also 1.1086 Hz briefly appearing)

Ongoing detailed analysis (future aLOG) of the violin mode region contamination indicates that at least some of the lines are calibration line + violin mode mixing. We are not yet sure whether this explains all the contaminated time periods in O4a, nor to all the contaminated lines, and we have not done a systematic comparison with L1. However, at the overview level, the mixing of calibration lines and violin resonances seems much worse in H1 than in L1. Perhaps this is somehow related to the new signal readout chain installed ahead of O4 that is different from the readout at L1. The number of lines also should instill some caution into further efforts to put additional calibration monitoring lines into the DARM control loop as the added lines would greatly increase the number of artifacts in the spectrum. This needs to be solved before further lines are added.

We anticipate further understanding of the O4a data with continued investigation, so the number of artifacts known to be non-astrophysical may go up.