New Observations of the Strongly Magnetic O-star NGC 1624-2 Reveal Its Magnetic South Pole

Abstract

NCG 1624-2 has the strongest detected magnetic field of all known main-sequence O-type stars. It was originally found that its magnetospheric emission lines followed a $\sim$5 month periodicity, and the existing line-of-sight magnetic measurements were predominantly of strong positive (north) polarity. As such, the field's geometric interpretation has been a mildly tilted (with respect to the rotational axis) dipole, such that only the magnetic north pole is visible during a rotation cycle. However, S. Seadrow et al. (2026) recently reported that new magnetospheric observations no longer phased with the established ephemeris and that the period had to be decreased by a few days. S. Seadrow et al. (2026) also found that existing magnetic measurements did not rule out a period twice as long (306.56 d). This period suggests a different magnetic configuration with a larger dipolar tilt, making both magnetic poles visible over a single rotation. Because previous spectropolarimetric observations did not have sufficient phase coverage to distinguish between the geometries, both were equally viable. In this paper, we present new spectropolarimetric observations obtained specifically to resolve this ambiguity. Our new magnetic measurements have a strong negative (south) polarity, confirming that the rotational period of NGC 1624-2 is indeed nearly twice as long as previously thought. Our measurements show that both poles come within a similar angle to our line of sight and likely have roughly the same local magnetic field strength (with a dipolar strength of 15-20 kG or more depending on the inclination angle).

Figures (4)

• Figure 1: Intensity and Stokes $V/I_C$ profiles of Hei$\lambda$5015 (left) and Civ$\lambda$5812 (right). The top panels compare Stokes $V/I_C$ (scaled by 100) from two observations at phases 0.98 (blue) and 0.41 (purple; Jan 2025), according to the double-wave ephemeris of 2026arXiv260215321S. The bottom panels show intensity and Stokes $V/I_C$ profiles for each line, which are offset vertically according to their phase with the same ephemeris. The new observations are shown in purple (Jan 2025) and red (Dec 2025). The grey curves in the Stokes $V$ panels show the diagnostic Null profile. The profiles have been normalized to their local intensity continuum, as described in §\ref{['sec:Bz']}. The profiles have been shifted to the stellar rest frame by subtracting the radial velocity ($-30$ km/s). For display purposes only, we apply a small boxcar averaging to smooth profiles.
• Figure 2: Left: Longitudinal magnetic field measurements phased with the single-wave (top) and double-wave (bottom) periods of 2026arXiv260215321S. In the top panel, the two new observations are indicated. In the bottom panel, the labeled observations correspond to those shown in Figure \ref{['fig:CIV_split']} The colored curves show a cosine model fit (that only includes the previous observations for the top panel, but all observations for the bottom panel). Right: The dipolar geometry constraints for the obliquity angle (top) and the dipolar field strength (bottom) that stem from the cosine fit as a function of the rotational inclination angle, as described in \ref{['sec:Bz']}. The colors match the fit curves in the left panels.
• Figure 3: The intensity profiles for Civ$\lambda$5812. The top panel and bottom panels compare two pole-on observations (labeled N and S in Figure \ref{['fig:Bz']}) and two equator-on observations (labeled E1 and E2 in Figure \ref{['fig:Bz']}), respectively. The middle panel shows the region between the two pole-on profiles (blue) and the region between the equator-on profiles (red). These profiles have been shifted to the stellar rest frame (by subtracting the radial velocity $-32$ km/s).
• Figure 4: The $B_\ell$ (units are kG) measurements from individual spectral lines and the weighted mean measurements, $\langle B_\ell \rangle$ measurements. We list the the observations' HJD's (minus 2450000.0 days) and phases that were computed using the revised spectroscopic ephemeris, $\mathrm{HJD} = (2455967.0 \pm 10.0) + (306.56 \mathrm{d} \pm 1.19 \mathrm{d}) \times E$