BLAZAR 3C 454.3 IN OUTBURST AND QUIESCENCE DURING 2005-2007: TWO VARIABLE SYNCHROTRON EMISSION PEAKS

We monitored the flaring blazar 3C 454.3 during 2005 June-July with the Spitzer Infrared Spectrograph (IRS: 15 epochs), Infrared Array Camera (IRAC: 12 epochs), and Multiband Imaging Photometer (MIPS: 2 epochs). We also made Spitzer IRS, IRAC, and MIPS observations from 2006 December to 2007 January when the source was in a low state, the latter simultaneous with a single Chandra X-ray observation. In addition, we present optical and submillimeter (sub-mm) monitoring data. The 2005-2007 period saw three major outbursts. We present evidence that the radio-optical spectral energy distribution (SED) actually consists of two variable synchrotron peaks, the primary at IR and the secondary at sub-mm wavelengths. The lag between the optical and sub-mm outbursts may indicate that these two peaks arise from two distinct regions along the jet separated by a distance of 0.9-3 pc. The flux at 5-35 mu m varied by a factor of 40 and the IR peak varied in frequency from 4 x 10(12) Hz to 4 x 10(13) Hz between the highest and lowest states in 2005 and 2006, respectively. Variability was well correlated across the mid-IR band, with no measurable lag. Flares that doubled in flux occurred on a timescale of similar to 5 days, yielding a variability size of <0.05 pc. The IR SED peak moved to higher frequency as a flare brightened, then returned to lower frequency as it decayed. The fractional variability amplitude increased with frequency, which we attribute to decreasing synchrotron self-absorption optical depth. Mid-IR flares may signal the re-energization of a shock that runs into inhomogeneities along the pre-existing jet or in the external medium. The synchrotron peak frequencies during each major outburst may depend upon both the distance from the jet apex and the physical conditions in the shocks. Variation of the Doppler parameter along a curved or helical jet is another possibility. Frequency variability of the IR synchrotron peak may have important consequences for the interpretation of the blazar sequence, and the presence of a secondary peak may give insight into jet structure.