First Year Performance Paper - Section 4.3
4.3 Systematic timing offset study with downgoing muons
To measure systematic timing offsets in the IceCube string we repeated 60 times the single-string reconstruction for one day of data. Each of the 60 DOMs was successively removed during the reconstruction, and the time delays of the hits in each DOM with respect to the expected direct (unscattered) hit times from the reconstructed tracks (called time residuals) were evaluated. The residual time distributions are consistent with our expectation, with a narrow peak of nearly unscattered photons and a long tail of photons delayed by scattering, as illustrated for a particular DOM in Fig. 20. The maxima of such distributions indicate the time residuals of the most probable hits.
Systematic displacement of time residuals from zero could indicate errors in the time calibration procedure and possible variations in internal delays in different modules. They could also be influenced by uncertainties in the DOM positions and by photon scattering in the ice. The peaks in residual distributions of most DOMs are within 3 ns of each other, i.e., the DOM clock times for the whole In-Ice array (currently 60 DOMs) are calibrated to within 3 ns of each other (with a mean of 1.1 ns and rms of 2.4 ns). DOMs 35-43 are 23 located in dustier ice (the region between 2000 m and 2140 m in Fig. 15) and have the highest residuals; DOMs 53-60 are located in the clearest ice (the region between 2300 m and 2430 m) and have the lowest residuals (Fig. 21). (Negative residuals are an artifact of the fitting procedure.)