Millisecond Pulsar Scintillation Studies with LOFAR: Initial Results

High-precision timing of millisecond pulsars (MSPs) over years to decades is a promising technique for direct detection of gravitational waves at nanohertz frequencies. Time-variable, multi-path scattering in the interstellar medium is a significant source of noise for this detector, particularly as timing precision approaches 10 ns or better for MSPs in the pulsar timing array. For many MSPs the scattering delay above 1 GHz is at the limit of detectability; therefore, we study it at lower frequencies. Using the LOFAR (LOw-Frequency ARray) radio telescope we have analyzed short (5-20 min) observations of three MSPs in order to estimate the scattering delay at 110-190 MHz, where the number of scintles is large and, hence, the statistical uncertainty in the scattering delay is small. We used cyclic spectroscopy, still relatively novel in radio astronomy, on baseband-sampled data to achieve unprecedented frequency resolution while retaining adequate pulse phase resolution. We detected scintillation structure in the spectra of the MSPs PSR B1257+12, PSR J1810+1744, and PSR J2317+1439 with diffractive bandwidths of $6pm 3$, $2.0pm 0.3$, and $sim 7$ kHz, respectively, where the estimate for PSR J2317+1439 is reliable to about a factor of 2. For the brightest of the three pulsars, PSR J1810+1744, we found that the diffractive bandwidth has a power-law behavior $Delta u_d propto u^{alpha}$, where $ u$ is the observing frequency and $alpha = 4.5pm 0.5$, consistent with a Kolmogorov inhomogeneity spectrum. We conclude that this technique holds promise for monitoring the scattering delay of MSPs with LOFAR and other high-sensitivity, low-frequency arrays like SKA-Low.

Long-Term Radio Timing Observations of the Transition Millisecond Pulsar PSR~J1023+0038

The radio millisecond pulsar PSR J1023+0038 exhibits complex timing and eclipse behavior. Here we analyze four years worth of radio monitoring observations of this object. We obtain a long-term timing solution, albeit with large residual timing errors as a result of apparent orbital period variations. We also observe variable eclipses when the companion passes near our line of sight, excess dispersion measure near the eclipses and at random orbital phases, and short-term disappearances of signal at random orbital phases. We interpret the eclipses as possibly due to material in the companions magnetosphere supported by magnetic pressure, and the orbital period variations as possibly due to a gravitational quadrupole coupling mechanism. Both of these mechanisms would be the result of magnetic activity in the companion, in conflict with evolutionary models that predict it should be fully convective and hence non-magnetic. We also use our timing data to test for orbital and rotational modulation of the systems $gamma$-ray emission, finding no evidence for orbital modulation and $3.7sigma$ evidence for modulation at the pulsar period. The energetics of the system make it plausible that the $gamma$-ray emission we observe is entirely from the millisecond pulsar itself, but it seems unlikely for these $gamma$-rays to provide the irradiation of the companion, which we attribute instead to X-ray heating from a shock powered by a particle wind.