With many thanks I refer to:
Kondratiev et al. 2015, "A LOFAR Census of Millisecond Pulsars"
Abstract: "We report the detection of 48 millisecond pulsars (MSPs) out of 75 observed thus far using the LOFAR in the frequency range 110-188 MHz. We have also detected three MSPs out of nine observed in the frequency range 38-77 MHz. This is the largest sample of MSPs ever observed at these low frequencies, and half of the detected MSPs were observed for the first time at frequencies below 200 MHz. We present the average pulse profiles of the detected MSPs, their effective pulse widths and flux densities, and compare these with higher observing frequencies. The LOFAR pulse profiles will be publicly available via the EPN Database of Pulsar Profiles. We also present average values of dispersion measures (DM) and discuss DM and profile variations. About 35% of the MSPs show strong narrow profiles, another 25% exhibit scattered profiles, and the rest are only weakly detected. A qualitative comparison of the LOFAR MSP profiles with those at higher radio frequencies shows constant separation between profile components. Similarly, the profile widths are consistent with those observed at higher frequencies, unless scattering dominates at the lowest frequencies. This is very different from what is observed for normal pulsars and suggests a compact emission region in the MSP magnetosphere. The amplitude ratio of the profile components, on the other hand, can dramatically change towards low frequencies, often with the trailing component becoming dominant. As demonstrated by Dyks et al. (2010) this can be caused by aberration and retardation. This data set enables high-precision studies of pulse profile evolution with frequency, dispersion, Faraday rotation, and scattering in the interstellar medium. Characterizing and correcting these systematic effects may improve pulsar-timing precision at higher observing frequencies, where pulsar timing array projects aim to directly detect gravitational waves."