A binary model of long period radio transients and white dwarf pulsars

TL;DR

This paper addresses long-period radio transients by developing a WD binary pulsar–like emission geometry. It introduces a six-parameter geometric model linking WD magnetospheric emission to the companion wind and validates it on GPM J1839-10 using a 36-year timing baseline, revealing an orbital period $P_{\text{orb}} = 31482.4 \pm 0.2$ s and a beat period $P_2 = 1265.2197 \pm 0.0002$ s; the model reproduces the observed double-pulse structure and polarization, and extends to WD pulsar J1912-44. MCMC fits yield an almost edge-on configuration with $i = 100.1 \pm 0.6^{\circ}$ and magnetic obliquity $\alpha = 52.1 \pm 0.4^{\circ}$, supporting a emission mechanism anchored to the WD magnetic axis and modulated by the MD wind. The results unify WD binaries and LPTs under a common emission geometry and provide predictive power for identifying and interpreting future WD-LPT systems and timing analyses.

Abstract

Long-period radio transients (LPTs) represent a recently uncovered class of Galactic radio sources exhibiting minute-to-hour periodicities and highly polarised pulses of second-to-minute duration. Their phenomenology does not fit exactly in any other class, although it might resemble that of radio magnetars or white dwarf (WD) radio emitting binary systems. Notably, two LPTs with confirmed multi-wavelength counterparts have been identified as WD -- M dwarf binaries. Meanwhile, systems such as AR Scorpii and J1912-44 exhibit short-period pulsations in hrs-tight orbits, with polarised radio emission proposed to be generated by the interaction of the WD magnetosphere with the low-mass companion wind. Here, we investigate the longest-lived LPT known, GPM J1839-10, demonstrating that it has a ~8.75 hr orbital period. We show that its radio pulses can be modelled in the same geometric framework as WD binary pulsars, in which radio emission is triggered when the magnetic axis of a rotating WD intersects its companion's wind in the binary orbital plane. We use a 36-year timing baseline to infer the orbital period and binary geometry from radio data alone. The model naturally predicts its intermittent emission and double-pulse structure. Crucially, we show that the beat period between the spin and the orbit matches the observed pulse substructure and polarisation signatures, providing strong support for the model. Applying it to the WD pulsar J1912-44, it successfully reproduces the emission profile and geometry as well. Our results suggest analogous emission-site geometries in these related classes of binary system -- a possibility we extend to the broader LPT / WD pulsar population.

Figures (7)

• Figure 1: Dynamic pulse profile and polarisation position angles of GPM J1839$-$10. Profiles are folded vertically on the orbital period $P_\text{orb} = 31482.4 \pm 0.2$ s and horizontally on $P_1 = 1318.1957 \pm 0.0002$ s (left) and $P_2 = 1265.2197 \pm 0.0002$ s (right). Above, the mean flux is calculated over two boxes enclosing the pulse groups: $\lbrace-0.18 < \text{orbital phase} < 0.18\rbrace$ for period A, $(\lbrace-0.18 < \text{orbital phase} < -0.06\rbrace \cap \lbrace\text{spin phase} < 0\rbrace) \cup (\lbrace0.06 < \text{orbital phase} < 0.18\rbrace \cap \lbrace\text{spin phase} > 0\rbrace)$ for period B.
• Figure 2: Modelled dynamic pulse profile of GPM J1839$-$10. At top left, the flux density predicted by the model $I_\text{pred}$ using the best-fit parameters found using MCMC is overlaid on the real pulse profiles. At top right, the colourmap is the LOS-beam angle $\beta$ as a function of spin-orbit phase, and the contours are the beam-MD angle $\beta_\text{MD}$. The bottom panel is a full orbit recorded by ASKAP normalised to 1 GHz and the associated model prediction. The vertical lines are spaced by the spin period.
• Figure 3: Diagram of the binary system, in a moving reference frame centred on the WD. At left is a not-to-scale diagram of the geometric parameters. In green is WD with its magnetic moment vector. In red, the MD with its orbital path. The blue vector points towards Earth. At right is a to-scale projection of GPM J1839$-$10 on the y-z plane, with orbital phases marked. The cone traced by $\hat{\mu}$ crosses the y-z plane at the green full and dashed lines. The shaded red, orange, and cyan regions cover the ranges of $R_\text{MD}$, $R_{rl}$, and $R_c$ respectively, given the MD mass is $M_\text{MD} \in [0.14, 0.5] M_\odot$ and the WD mass is $M_\text{WD} \in [0.6, 1.2] M_\odot$. The green sector and solid red line are $W_\text{spin}$ and $W_\text{orb}$ respectively. The small and large black dashed circles are the Alfvén radii for the minimum and maximum WD masses respectively, assuming a stellar mass loss rate of $\dot{M}_\text{MD}$ = $10^{-14}$ M$_\odot$ yr$^{-1}$2005ApJ...628L.143W.
• Figure 4: Modelled dynamic pulse profile for J1912$-$44.
• Figure 5: Left: Envelope width for trial orbital periods. Right: Detected pulses folded on the orbital period. The y axis is the orbital period number since the most recent detection (negative means backward in time). The widths of the blue rectangles are the pulse widths. The blue lines are the uncertainties in residual given the uncertainty in period. The envelope width is measured from the leftmost pulse edge to the rightmost pulse edge.