ZIMPOL detects scattering polarization in He I D$_3$ during a solar flare

Abstract

Spectropolarimetric observations of solar flares in the He I D$_3$ line at 5876 Å are extremely rare, and their diagnostic potential remains largely unexplored. We report the first unambiguous detection of linear polarization in He I D$_3$ during a solar flare. Using the high-precision ZIMPOL polarimeter at the IRSOL observatory in Locarno (Switzerland), we tracked the temporal evolution of the He I D$_3$ Stokes profiles throughout the M7 GOES-class flare that occurred on 3 May 2023 at 10:45 UT. We analyzed the time evolution of the maximum in linear polarization and the absorption depth of the intensity profile. Both the fractional linear polarization, which peaks at $6\times10^{-4}$, and the absorption depth increase rapidly before gradually decaying, with their maxima occurring approximately 5 minutes after the peaks in GOES X-ray flux and SDO/AIA 304 Å emission. From the evolving He I D$_3$ core position, we also derived the temporal evolution of the plasma bulk velocities. The intensity profiles exhibiting strong absorption seems to originate from the flare ribbons. The time evolution of all Stokes parameters in the 3 May 2023 event was driven by changes in ortho-helium density prompted by the different phases of the flare. Our analysis suggests that the observed He I D$_3$ linear polarization is likely not dominated by the theorized impact polarization, as it exhibits neither spatial correspondence with electron precipitation sites nor temporal synchronization with the impulsive phase. Instead, the signals are consistent with scattering polarization produced by anisotropic radiation pumping. We conclude that scattering polarization signal on the order of $0.01\%$ can be produced in the He I D$_3$ line during solar flares. This can provide constraints for flare models.

Figures (6)

• Figure 1: GOES/XRS time profiles for the 3 May 2023 M7 event. Vertical dashed lines separate the temporal phases of the solar flare. The blue curve corresponds to the 0.5-4 Å GOES channel, the red curve to the 1-8 Å channel, and the gray curve to the derivative of the 1-8 Å curve.
• Figure 2: Location and morphology of the 3 May 2023 M7 event. Left panel: Location on the solar disk. Right panel: Context images showing (left to right, top to bottom) SDO/HMI continuum, SDO/AIA 1700 Å (pre-flare subtracted, see Sect. \ref{['aiastix']}), SDO/AIA 304 Å, SDO/HMI continuum running differences (see Eq. \ref{['rundif']}), SDO/HMI magnetogram, and SDO/AIA 131 Å. All of the images were taken at the time of the GOES X-ray flux peak, except SDO/AIA 131 Å, which was taken about 10 minutes after.
• Figure 3: Time evolution of the Stokes signals for the 3 May 2023 M7 event. From left to right, we show the $I/I_c$ (pre-flare subtracted, see Sect. \ref{['anaslit']}), $Q/I_c$, $U/I_c$, and $V/I_c$ Stokes spectra averaged along the slit at each time frame of the observation, with $\lambda_{D_3}=5875.6~\AA$. We overplot the Savitzky-Golay smoothing of data points for $Q/I_c$ and $U/I_c$ (orange curves), and for $V/I_c$ (red curves).
• Figure 4: Time series of measured $I/I_c$ spectra (black dots) and the corresponding HaZeL fit (red lines) for the 3 May 2023 M7 event. Each image was processed by subtracting the chronologically first frame (10:33 UT, not shown here). The blue and orange arrows indicate small discrepancies in the fit that might indicate the need to include in the modeling fainter components, blueshifted or redshifted, respectively, relative to the main component.
• Figure 5: Time histories of 3 May 2023 M7 event. First panel (from the top): GOES X-rays flux. Second panel: Min-max normalized counts in SDO/AIA 304 Å channel. Third panel: Maximum in wavelength of the degree of linear polarization (black dots), absorption depth (blue dots), and maximum in wavelength of circular polarization (red dots) for He i D$_3$. The value of the maximum of the degree of linear polarization was computed according to Eq. \ref{['poldeg']} and similarly the values of the maximum in circular polarization are a fraction of $I_c$. Fourth panel: LOS bulk velocities inferred by HaZel with uncertainties.