Multifrequency observations of integrated pulse profiles are shown in Fig. 2 (click here) for 56 pulsars. The profiles are descattered at low frequencies and aligned in pulse phase as outlined in Sect. 3 (click here). In very few cases, we added for comparison profiles, where no timing information was available; these profiles are marked with an asterix.
It is obvious from Fig. 2 (click here) that most profiles are well aligned over the 0.1 to 10 GHz frequency range considered here. This fact is remarkable in itself, since it demonstrates that the cold plasma dispersion relation is applicable over the whole frequency range and that the different phase shifts over all the measurement frequencies can be described by just one free parameter, the dispersion measure DM. The alignment was achieved either by use of the catalogued value of DM or by applying small corrections as explained in Sect. 3 (click here). Changes in the dispersion measure value influence the phase of all profiles simultaneously; low-frequency profiles are shifted quite drastically but the high-frequency phase is left nearly unchanged. Misalignments at high frequencies can therefore hardly be removed by correction of the dispersion measure.
There exists, indeed, a group of pulsars where non-removable - "non-dispersive" - phase shifts appear at high frequencies; these are the pulsars PSR 0138+59, 0355+54, 0450+55, 0540+23, 0809+74, 1541+09, 1822-09, 1839+56, 1929+10, 2021+51, and 2154+40. We will discuss the properties of these pulsars in detail below.
It should be mentioned here, that quantifying the phase shift between profiles at different frequencies is generally difficult. The reason is that a shift between unequally shaped curves has to be computed. There exist always - albeit sometimes very small - differences between the profiles at different frequencies. The apparent displacements of the measurements of some of the quoted pulsars - PSR 0355+54, 1541+09, 1839+56, and 1929+10 - can be explained by variations of the pulse shape in the sense, that some components weaken or increase drastically at higher frequencies. This interpretation becomes even more
evident, when a decomposition of the average profile into gaussian components is undertaken, as will be shown in a forthcoming paper, which is devoted to the analysis of the catalogued aligned profiles.
However, not all non-dispersive phase shifts at high frequencies can be explained easily by intensity variations of the individual components. It is much harder to explain the non-dispersive phase shift at high frequencies of PSR 0809+74 as discussed already by Bartel et al. (1981), Davies et al. (1984), and Kuzmin et al. (1986). Non-dispersive time shifts at high frequencies are indicated also for the pulsars PSR 0138+59, 1822-09, 2021+51, and 2154+40. It must be said, however, that the data for PSR 0138+59, 1822-09, and 2154+40 are based on few measurements only and need further confirmation. We have listed these pulsars with their corresponding phase shifts in Table 2 (click here), a minus sign indicating that the profile is shifted to earlier phase. A detailed interpretation and model for this non-dispersive phase shift at high frequencies may be found in Davies et al. (1984) and Kuzmin et al. (1986).
The measurements of PSR 0450+55 at 10 GHz show only a weak signal-to-noise ratio so that exact conclusions on phase shift with respect to lower frequencies are hard to make. We should also remark that three differing profiles were observed for PSR 0540+23 at 10.5 GHz on the same day at different times. We show these profiles in Fig. 2 (click here) since there exists the possibility that all three are real, the difference being produced by mode changes. Future measurements must clarify this point.
Figure 2: Aligned profiles of 56 pulsars. The error bars in the
upper left corner of each frame denote the smearing due to the sampling
interval and due to the dispersion across the bandwidth. The dispersion
measure used for the alignment is given in the upper right corner of each
frame (abscissa in degrees of pulsar rotation period)
Acknowledgements
The authors acknowledge support from their colleagues J. Gil, V.M. Malofeev, I.F. Malov, J.H. Seiradakis, and K.M. Xilouris during the "pulsar month" as well as the assistance of V.V. Ivanova and B.Ya. Losovsky from the Pushchino Radioobservatory during the observations. This work was supported in part by ESO (Grant A-04-083), INTAS (Grant 94-3097) and RFFR (Grant 95-02-0581).