Radio emission from beyond the light cylinder in millisecond pulsars

TL;DR

This study tackles where radio emission in millisecond pulsars (MSPs) originates relative to gamma-ray emission by applying a robust, data-driven method to classify profiles into disjoint versus contiguous emission and by fitting magnetospheric geometry with the Rotating Vector Model (RVM). It finds that about $35\%$ of MSPs exhibit disjoint radio components, and many gamma-ray MSPs show radio components co-located with gamma-ray emission in the current sheet beyond the light cylinder (LC), supporting a two-region emission picture combining polar-cap (PC) and LC contributions. LC components are often highly polarized with flat position-angle traverses, and LC emission tends to have lower luminosity than PC emission, affecting detectability and beaming. The results imply a larger beaming fraction for MSPs than previously thought, with important consequences for the Galactic MSP population, radio-loud versus radio-quiet fractions, and long-term timing stability in pulsar timing arrays.

Abstract

A striking aspect of the radio profiles of many millisecond pulsars (MSPs) is that they consist of components separated from each other by regions lacking in emission. We devise a technique for determining "disjoint" from "contiguous" components and show that 35% of MSPs have disjoint components as opposed to only 3% of the slow pulsar population. We surmise that the pulsars with these disjoint components show evidence for both emission above the polar cap and from the current sheet beyond the light cylinder (LC), co-located with gamma-ray emission. For some of the radio MSPs only the LC emission is being observed. It is our contention that almost all of the current population of gamma-ray MSPs show evidence for co-located radio emission. A simple geometric explanation allows the presence (or not) of LC emission and the relationship (or not) between the gamma-ray and radio profiles to be determined. The LC components have frequently very high polarization and typically flat position-angle traverses thus helping to explain the difficulties in determining the geometry of MSPs. In cases where the geometry can be determined the values broadly align with expectations. In this picture, the number of potentially detectable radio MSPs is higher than previously thought, although the actual detectability of LC components depends on their luminosity function. A mechanism is required to produce coherent radio emission far from the stellar surface. These ideas have implications for our understanding of the populations of radio-loud and radio-quiet rotation-powered millisecond pulsars, and may have implications for the long-term timing stability of some of these sources.

Figures (7)

• Figure 1: Example outputs from the finding technique for six millisecond pulsars. The left hand column shows pulsars classified as contiguous (class C) and the right hand column shows pulsars with disjoint (class D) profiles. The three panels show the profile, the cumulative distribution of the amplitudes and the polar plot. See text for details.
• Figure 2: Example outputs from the finding technique for three slow pulsars with disjoint profiles, PSRs J0401--7608 (top), J1851+0418 (middle) and J1852--0118 (bottom). See Figure \ref{['fig:msp']} and text for details.
• Figure 3: PSR J1125$-$5825. The top panel shows the total intensity (black), linear polarization (red) and circular polarization (blue) for the radio emission. The light-blue line indicates the gamma-ray profile. In the middle panel the PAs are shown along with the best fit RVM. PA values are shown as measured in black and offset by 90 in blue. The dashed lines correspond to a RVM solution separated in PA by 90. PA values in gray have been not been modeled. The bottom panel shows the residuals between model and data.
• Figure 4: PSR J1939+2134 is shown in a similar way to Figure \ref{['fig:1125']}. The additional inset in the top panel top panel highlights the low-level emission.
• Figure 5: Regions of the $\alpha$-$\zeta$ plane showing detectability of emission regions. In the green regions only the LC components are visible, in the blue region both LC and PC components are visible. Only the PC emission is visible from the red region and in the yellow areas neither the LC nor PC emission is visible. A spin period of 3.5 ms and an emission height for the PC emission of 30 km has been assumed. The source geometries derived and given in the tables in the Appendix are shown: Sources from Table \ref{['tab:LC']} as blue, from \ref{['tab:NOLC']} as green, from \ref{['tab:DNO']} as orange and from \ref{['tab:CLUM']} as red symbols. Values of $\alpha$ that are larger than 90 have been folded into the plotted range.