Pulsars identified in the LOFAR Two-metre Sky Survey at 144 MHz

TL;DR

LoTSS DR2 enables astrometric identification of 80 known radio pulsars at 144 MHz through cross-matching with the ATNF catalog, achieving high recovery within the DR2 footprint. Flux densities are largely consistent with literature, with a couple of outliers explained by intrinsic pulsar behavior or propagation effects; polarization greatly boosts detectability, with 35 pulsars identified via polarization. The study demonstrates that polarization-based candidate selection, together with cross-matching to LoLSS at 54 MHz, is a powerful approach for finding new pulsars and constraining spectra, and it outlines a path for expanding the census with DR3/LoLSS data. The results underscore the value of wide-area imaging surveys for pulsar science at low frequencies and pave the way for spectral studies of a large pulsar sample from 54 to 168 MHz.

Abstract

We present the astrometric identification of 80 known radio pulsars as unresolved continuum sources detected at 144 MHz in the second data release of the LOFAR Two-metre Sky Survey (LoTSS DR2), which covers 27% of the Northern hemisphere. These identifications represent the majority ($\geq$ 86%) of radio pulsars present in the LoTSS DR2 footprint and provide independent celestial positions and flux densities at 144 MHz. We compare LoTSS flux densities with literature values from various image and time-domain observations and find good agreement for all but two pulsars. We attribute these flux density deviations to intrinsic pulsar properties (nulling and off-pulse emission). We investigate criteria to select promising pulsar candidates using data from the upcoming LoTSS release of the entire Northern sky ($δ>0^\circ$), as well as the LOFAR LBA Sky Survey (LoLSS) at 54 MHz (covering $δ>24^\circ$). Of the 80 detections, 35 (44%) were blindly redetected based on their linear or circular polarization. Therefore we conclude that candidate selection based on polarization properties is a promising approach. Candidate selection can be supplemented with spectral indices via cross-matching to LoLSS sources at 54 MHz, as the high sensitivity of LoTSS is not matched by image-domain surveys at higher frequencies.

Figures (4)

• Figure 1: LoTSS DR2 images in Stokes I for nine example pulsars. All images have a field-of-view of $90\arcsec\,\times\,90\arcsec$, except for panel a) which has $8\arcmin\,\times\,8\arcmin$ and panel b) which has $2\farcm5 \times 2\farcm5$. Positions of LoTSS sources are indicated with a black square, while positions from the pulsar catalog are denoted with the red cross. The extent of LoTSS sources is shown with the black dashed ellipses, depicting the major- and minor axes resulting from fitting to the LoTSS sources Shimwell+22. The FWHM of the LoTSS beam is shown in the bottom-right corners. Flux density contours in mJy beam$^{-1}$ are plotted for panel i) to highlight the structure within bright sources. Panel d) and g) show example pulsars with accurate position measurements, leading to a unique match to a LoTSS source, while panel a) shows an example where the positional uncertainty of the pulsar is large and several LoTSS sources coincide with the 99% confidence error region (denoted by the red dotted ellipse). In panel b) a field of LoTSS sources are shown, where the Northern-most source is matched to a pulsar. All visible components show source extension to the East, indicating an imaging problem. Pulsar J1336+3414 is dimly visible as a point source in panel e), and due to its low brightness and/or the proximity of the Eastern bright source it is a $\sim4\sigma$ detection that is not catalogued in LoTSS DR2. We estimate its flux density directly from the image. Panel c) shows a pulsar with a separation of $8\farcs4$ to a LoTSS source, but does not contain it within its 99% uncertainty ellipse. In panel f) a pulsar is matched to a LoTSS source that was classified as a single extended object, while it in fact consists of two unrelated objects, one of which is the pulsar counterpart. Finally, panel i) shows a compact radio source that is present at the position of PSR B1112+50, but due to the proximity of a brighter radio source, it is catalogued in LoTSS as an extended source with an offset position.
• Figure 2: Comparison of LoTSS flux densities from Table \ref{['tab:detections']} at 144 MHz (horizontal) to 150 MHz flux density measurements from image-domain observations by Frail+2016 and time-domain observations by Bilous+2016Kondratiev+20162020MNRAS.492.5878TMichilli+20202023AA...669A.160V and LOTAAS at 135 MHz. The grey region represent fluxes ratios between factors 0.5 and 2.
• Figure 3: LoTSS DR2 images in Stokes I for 7 globular clusters. The image in panel a) has a field-of-view of 2$\times$ 2. Panel b) has a field-of-view of 35 $\times$ 35, and panels c) and d) have a field-of-view of 2.5$\times$ 2.5 and 4$\times$ 4 respectively. Positions of LoTSS sources are indicated with a black box, pulsar positions are denoted with the red cross and positions of VLA sources 2020ApJ...903...73S are shown with magenta triangles. The core- and half-light radii of globular clusters are plotted in salmon with solid and dash-dotted circles respectively. The FWHM of the synthesized LoTSS beam of 6$\times$ 6 is shown in the bottom-right corners.
• Figure 4: A comparison of survey sensitivity and frequency coverage of various time domain pulsar surveys (green) and wide area continuum image domain surveys (red) against distributions of pulsar flux density measurements from the pulsar catalog (open circles). An example power law spectrum $S_\nu\propto\nu^\alpha$ with $\alpha=-1.6\pm1.0$ for $S_{144}=0.9$ mJy is shown for reference. (References: NPBS; 2025AA...693A..96B, LOTAAS; LOTAAS, SMART; 2023PASA...40...20B, AO327; 2013ApJ...775...51D, GBNCC; 2014ApJ...791...67S, FAST GPPS; 2021RAA....21..107H, PALFA; 2006ApJ...637..446C, HTRU; 2010MNRAS.409..619K, LoLLS; 2021AA...648A.104D, VLSSr; 2014MNRAS.440..327L, LoTSS; Shimwell+17, TGSS; 2017AA...598A..78I, WENSS; 1997AAS..124..259R, RACS; 2020PASA...37...48M, NVSS; 1998AJ....115.1693C, VLASS; Lacy+2020).