First Mid-infrared Detection and Modeling of a Flare from Sgr A*. II. Mid-IR Spectral Energy Distribution and Millimeter Polarimetry

Abstract

S. D. von Fellenberg et al. (2025a, Paper I) reported the first mid-infrared detection of a flare from Sgr A*. The JWST/MIRI/MRS observations were consistent with an orbiting hotspot undergoing electron injection with a spectrum that subsequently breaks from synchrotron cooling. However, mid-infrared extinction measurements appropriate for these data were not yet determined, and therefore the temporal evolution of the absolute spectral index remained unknown. This work applies new Galactic Center extinction measurements to the flare observations. The evolution of the spectral index after the peak is fully consistent with that reported in Paper I with a maximum absolute mid-infrared spectral index $α_{\rm{MIR}}=0.45\pm0.01_{\rm{stat}}\pm0.08_{\rm{sys}}$ during the second mid-infrared flare peak, matching the known near-infrared spectral index during bright states ($α_{\rm{NIR}}\approx0.5$). There was a near-instantaneous change in the mid-infrared spectral index of $Δα_{\rm{MIR}}=0.33\pm0.06_{\rm{stat}}\pm0.11_{\rm{sys}}$ at the flare onset. We propose this as a quantitative definition for this infrared flare's beginning, physically interpreted as the underlying electron distribution's transition into a hard power-law distribution. This paper also reports the SMA millimeter polarization during the flare, which shows a small, distorted, but overall clockwise-oriented Stokes Q--U loop during the third mid-infrared peak. Extrapolating the mid-infrared flux power law to the millimeter yields a variable flux consistent with the observed 220 GHz emission. These results, together with the Paper I modeling, plausibly suggest a single hotspot produced both the mid-infrared and millimeter variability during this event. However, additional flares are required to make a general statement about the millimeter and mid-infrared connection.

Figures (8)

• Figure 1: Radio to X-ray SED of Sgr A*, including mid-IR JWST/MIRI lower limits (red and blue points; corresponding times are denoted in Figure \ref{['fig:timed_sed']}), observed absolute maximum (light purple point) and variable (purple point) mm fluxes measured with the SMA, and X-ray flare upper limit from Chandra (blue-violet line) during this event. The first mid-IR peak's best-fit power law is extrapolated, and the 68% (dark gold) and 90% (light yellow) credible intervals are plotted. The variable mm data are fully within the 90% credible range of the single power law stretching between mm to mid-IR. The binned, historic SED data are shown in the radio and mm Bower2015Bower2019liu_linearly_2016Brinkerink2016, far-IR Stone2016VonFellenberg2018, and X-ray Baganoff2003_spectrum. Spitzer Witzel2018, GRAVITY GravityCollaboration2020flux, and post-2019 Keck Weldon2023 points show the median flux and 5% to 95% percentile ranges. We have recalculated the ordinate using a standard distance of $d = 8.178$ kpc; for the Spitzer data, we have dereddened the data with the extrapolated extinction from the [][$\rm{A}_{4.5~\mu\rm{m}} = 1.14$ mag]vonFellenberg2025_extinction.
• Figure 2: Temporal evolution of the mid-IR SED during the flaring event. Top Left: Spectral index variations. The dark gray shaded region is the total statistical uncertainty, while lighter gray denotes the total statistical + systematic range. Bottom Left: Intrinsic mid-IR light curves of Sgr A* in all four channels; the gray shaded region shows the statistical photometric noise in the measurement. The red and blue vertical bars denote times at which the time-resolved SEDs are plotted in Figure \ref{['fig:mwl_sed']}. Right panels: Residual-extinction-corrected SEDs (see Appendix \ref{['appx:residcorr']}) at different times during the observation. Each MIRI/MRS channel has been broken up into three equally-wide frequency bins (connected points), and a best-fit power-law (solid lines) has been plotted. Shaded regions correspond to the statistical uncertainty in the measurement. The colors of data points in all panels correspond to the measured spectral indices.
• Figure 3: Top left: Multiwavelength plot of SMA and dereddened JWST/MIRI light curves. Colored points in the SMA light curve correspond to the barycentric time of each 60-second point. Top right: Time-resolved mm Stokes $Q$--$U$ plot for Sgr A* obtained with the SMA at 192-second binning. Point colors correspond to the barycentric UT time in the top left panel. The green box corresponds to the zoom-in of the small $Q$--$U$ loop during the third mid-IR peak, and arrows show the overall orientation of the two observed loops. Bottom left: Mean-centered zoom-in of the clockwise-oriented $Q$--$U$ loop before 13:00 UT at 192-second binning during the JWST/MIRI flare. Bottom right: EVPA of the mean-centered Stokes $Q$--$U$ loop as a function of time, colored with the same scheme as the lower right panel.
• Figure 4: Left: Stokes $I$, spectral index, debiased linear polarization percent, and EVPA light curves for Sgr A* (red) and the main gain calibrator J1733 (blue) on 2024 April 6. The flux density of J1733 is shifted up by 2.75 Jy. The gray shaded region denotes the time range where JWST/MIRI was observing Sgr A*. Right: RM variations as a function of time. Native resolution (1-minute) values have been binned to 15 minutes using uncertainty-weighted averages. The SMA light curve is also plotted in the bottom panel.
• Figure 5: Comparisons of Stokes $I$, linear polarization, and EVPA (top, center, and bottom rows, respectively) for the four calibrators in this track using AMAPOLA (ALMA, black dots) and SMA (calibrated through COMPASS and CASA, red crosses) during 2024, showing that the calibrator properties are consistent with each other. The RMS-based error bars for SMA are plotted but smaller than the marker size.