Optimal Frequency Ranges for Sub-Microsecond Precision Pulsar Timing

Precision pulsar timing requires optimization against measurement errors and astrophysical variance from the neutron stars themselves and the interstellar medium. We investigate optimization of arrival time precision as a function of radio frequency and bandwidth. We find that increases in bandwidth that reduce the contribution from receiver noise are countered by the strong chromatic dependence of interstellar effects and intrinsic pulse-profile evolution. The resulting optimal frequency range is therefore telescope and pulsar dependent. We demonstrate the results for five pulsars included in current pulsar timing arrays and determine that they are not optimally observed at current center frequencies. For those objects, we find that better choices of total bandwidth as well as center frequency can improve the arrival-time precision. Wideband receivers centered at somewhat higher frequencies with respect to the currently adopted receivers can reduce required overall integration times and provide significant improvements in arrival time uncertainty by a factor of ~sqrt(2) in most cases, assuming a fixed integration time. We also discuss how timing programs can be extended to pulsars with larger dispersion measures through the use of higher-frequency observations.