2022PASP..134k4201Z

In this study, we revisit the magnetic field (B-field) distribution of normal pulsars, motivated by the fact that the number of known pulsars has exceeded 3300. Here, we divided the normal pulsar samples into three subgroups by constant lines of characteristic age τ_ch, i.e., young, middle-aged, and old pulsars. We note that τ_ch is not used as the time indicator in this study; instead, it just served as cutting lines to divide the pulsar samples. Then, we applied several statistical tests, i.e., the Anderson-Darling, Shapiro-Wilk, Kolmogorov-Smirnov, and Mann-Whitney-Wilcoxon tests, to the selected normal pulsar samples (N = 1970) and to a data set of 32 neutron stars (NSs) in high-mass X-ray binaries (HMXBs) for comparison purposes. We obtained that (i) the conclusion on the characteristic B field (B_ch) log-normal distribution for the normal pulsars by the previous studies is no longer appropriate, while only young pulsars (N = 24, τ_ch < 15 kyr) follow a log-normal distribution, indicating that only the B_ch of young pulsars is close to real B fields. (ii) In the directly measured B-field range of NS-HMXBs (∼10¹²-10¹³ G), the B_ch of young pulsars (N = 15) and the real B fields of NS-HMXBs (N = 32) are inferred to be log-normal, and they are further verified to come from the same distribution, implying that there is no significant decay for real B fields, at least within the timescale of ∼10 Myr for normal pulsars. (iii) Statistically, young pulsars (N = 24) are inferred to be self-contained, suggesting that the initial spin period of pulsars is less than 515 ms. (iv) The B_ch distributions of three normal pulsar subsamples are different, hinting at the existence of multiple origins for NSs.