Microarcsecond VLBI pulsar astrometry with PSR$\pi$ II. parallax distances for 57 pulsars

TL;DR

This paper presents precise VLBI parallax measurements for 57 pulsars, significantly expanding the sample of pulsars with reliable, model-independent distances, and highlights discrepancies in Galactic electron density models and pulsar proper motions.

Contribution

It provides the largest set of VLBI-based pulsar distances to date, improving distance constraints and testing Galactic electron density models and pulsar timing accuracy.

Findings

Doubled the number of pulsars with reliable distance measurements.

Identified shortcomings in Galactic electron density models at high latitudes.

Detected proper motion discrepancies in millisecond pulsars compared to timing estimates.

Abstract

We present the results of PSR$\pi$, a large astrometric project targeting radio pulsars using the Very Long Baseline Array (VLBA). From our astrometric database of 60 pulsars, we have obtained parallax-based distance measurements for all but 3, with a parallax precision of typically 40 $\mu$as and approaching 10 $\mu$as in the best cases. Our full sample doubles the number of radio pulsars with a reliable ($\gtrsim$5$\sigma$) model-independent distance constraint. Importantly, many of the newly measured pulsars are well outside the solar neighbourhood, and so PSR$\pi$ brings a near-tenfold increase in the number of pulsars with a reliable model-independent distance at $d>2$ kpc. Using our sample along with previously published results, we show that even the most recent models of the Galactic electron density distribution model contain significant shortcomings, particularly at high Galactic latitudes. When comparing our results to pulsar timing, two of the four millisecond pulsars in our sample exhibit significant discrepancies in the estimates of proper motion obtained by at least one pulsar timing array. With additional VLBI observations to improve the absolute positional accuracy of our reference sources and an expansion of the number of millisecond pulsars, we will be able to extend the comparison of proper motion discrepancies to a larger sample of pulsar reference positions, which will provide a much more sensitive test of the applicability of the solar system ephemerides used for pulsar timing. Finally, we use our large sample to estimate the typical accuracy attainable for differential astrometry with the VLBA when observing pulsars, showing that for sufficiently bright targets observed 8 times over 18 months, a parallax uncertainty of 4 $\mu$as per arcminute of separation between the pulsar and calibrator can be expected.