Soft Muon Tagging

$b$-jets can be identified by taking advantage of the presence of a muon coming from the semi-leptonic decay of the $b$ hadrons. This note describes a soft muon tagging algorithm and its performance when applied to different Monte Carlo physics samples in ATLAS. A $b$-jet tagging efficiency of about 10% (including the inclusive b to mu nu X branching ratio of ~20$%) is achieved for a light jet rejection of better than 300. The effect of pile-up events and cavern background on the performance is also considered and is found to be significant but manageable.