next previous
Up: Pulse shapes of

4. Discussion and conclusions

In summary, we present 87 pulse shapes (see Fig. 1 (click here)), flux densities and pulse width measurements of all detected pulsars in this survey (see Table 1 (click here)). Up to now we have collected about 130 pulsar profiles at tex2html_wrap_inline1128 6 cm (see Seiradakis et al. 1995 and this paper). Following a study of pulse shapes (Fig. 1 (click here)), many pulsar profiles show a complex structure at this high frequency but 23 of them have simple Gaussian shapes. As many as 35 pulsars exhibit no changes at 4.85 GHz as compared to 1.41 GHz. Eight pulsars show slight changes in their shapes as compared to low frequencies while for eight additional sources, the simple profile observed at low frequency is resolved into several components at the high frequency.

The profile width should decrease monotonically with frequency if the emission originates from a magnetospheric region dominated by a dipole magnetic field and a radius-to-frequency mapping exists (Cordes 1978). Considering the profile-width narrowing we constructed a histogram which shows how the pulse width changes between 1.4 GHz and 4.85 GHz (see Fig. 2 (click here)). For this analysis we used pulse shapes for both frequencies (1.41 and 4.85 GHz) from Seiradakis et al. (1995) and our survey. Figure 2 (click here) shows that the pulse widths decrease or in some cases are unchanged with increasing frequency. This effect is also reported by Xilouris et al. (1996). For a few pulsars, pulse widths decreased dramatically. This behaviour is consistent with above assumptions of a hollow-cone model, a radius-to-frequency mapping and a dipolar field structure in the entire emission region.

Figure 2: A distribution of pulse width changes for 78 pulsars. For this analysis we used pulse widths at a level of 10 per cent of the pulse peak

Figure 3: Pulse shapes of PSR B0906-17 at 1.41 and 4.85 GHz. The development of conal outriders is clearly visible, causing an apparent increase in profile width

Figure 4: The distribution of flux spectral index for 141 PSRs

The increasing pulse width (see Fig. 2 (click here)) at high frequencies for some pulsars (B0402+61, B0450-18, B0626+24, B0906-17 and B1818-04) is caused by new components appearing in the profile. In Fig. 3 (click here), PSR B0906-17 is shown as a typical example. It is seen that the fairly simple profile at low frequency (1.41 GHz) is resolved into a complex one at higher frequency (4.85 GHz), which can be explained geometrically in the context of the multi cone model (Kramer et al. 1994; Sieber 1997). It is obvious that the pulse width and even more the observed pulse shape depend critically on the intensity distribution in the outer parts of the beam. Assuming a semi-Gaussian structure of nested cone beams, the conal emssion (for grazing cuts) becomes more and more dominant as the beam width shrinks at higher frequencies, so that outriders may become visible (see Fig. 3 (click here)).

Our sample also contains pulsars with very wide profiles. Two pulsars have a pulse width W10 larger than 100tex2html_wrap_inline1524 (PSR B1831-04 and PSR B1823-13) and five other pulsars have tex2html_wrap_inline1622 (see Table 1 (click here)). These pulse widths decrease by less than 10% as compared to 1.4 GHz. Moreover, two pulsars PSR B1702-19 and PSR B1855-09 have been observed with interpulses. Detection of pulsars with interpulses is not very common at high frequencies. Hankins & Fowler (1986) showed that the ratio of interpulse to main pulse intensity decreases considerably with frequency. On the other hand, Wielebinski et al. (1993) have observed an increase of this ratio at the very high frequency of 33.9 GHz for B1929+10. For the millisecond pulsars B1855+09 and J2322+2057 the same behaviour was observed (for more details see Kijak et al. 1997). Up to now we have measured only five pulsars with interpulses at tex2html_wrap_inline1128 6 cm (B0950+08, B1709-19, B1822-09, B1855+09, B1929+10; Hoensbroech & Xilouris 1997 and this paper).

We have combined the data from this survey with other published flux measurements in the frequency range between 1.41 GHz and 4.85 GHz. Figure 4 (click here) shows the distribution of spectral indices for all pulsars in the combined sample. This histogram is fairly broad, extending from tex2html_wrap_inline1626 (PSR B0834+06) to tex2html_wrap_inline1628 (PSR B1823-13). The mean spectral index of the sample is tex2html_wrap_inline1632. The overall distribution of spectral indices is symmetric. In Table 3 (click here) we compare the spectral indices for four different frequency regimes. It is cleary seen that the pulsar spectra become steeper at high frequencies.


range of tex2html_wrap_inline1634 (MHz)    tex2html_wrap_inline1598    Reference
1400 - 4850 -1.9 (this paper)
400 - 1600 -1.6 (Lorimer et al. 1995)
160 - 400 -1.2 (Slee et al. 1986)
80 - 160 -0.7 (Slee et al. 1986)
Table 3: Spectral index for several frequency regimes

The main conclusions from this analysis are as follows:

i) The profile shapes do not change much between 1.4 GHz and 4.85 GHz. About 30% of our pulsars have a simple Gaussian shape.

ii) Generally, the pulse widths decrease or are constant at high frequencies (Xilouris et al. 1996 and this paper). This is consistent with the hollow-cone model and the concept of radius-to-frequency mapping (Cordes 1978; Kijak & Gil 1997).

iii) The increasing pulse widths for a few pulsars at higher frequency is consistent with the multi-conal beam structure (Kramer et al. 1994; Sieber 1997).

iv) The flux spectral index tex2html_wrap_inline1598 has a mean value of tex2html_wrap_inline1648, with a symmetric distribution. This result is consistent with previous estimates of the mean spectral index tex2html_wrap_inline1650 (Sieber 1973; Malofeev et al. 1994). In general, pulsars show a steep emission spectrum at 4.85 GHz.

The pulse shapes from our survey and those of Seiradakis et al. (1995) are part of the European Pulsar Network Data Archive, maintained at the Max-Planck-Institut für Radioastronomie in Bonn and are available at:


We wish to thank O. Lochner and the receiver group for building the excellent 6 cm receiver. We also thank W. Sieber and D. Lorimer for helpful comments. The European Pulsar Network was funded by the European Comission under Human Capital and Mobility grant number CHRX-CT94-0622. This paper is partially supported by the Grant 2 P03D 015 12 of the Polish Committee for Scientific Research (KBN).

next previous
Up: Pulse shapes of

Copyright by the European Southern Observatory (ESO)