We have searched for continuous gravitational wave ( CGW ) signals produced by individually resolvable , circular supermassive black hole binaries ( SMBHBs ) in the latest EPTA dataset , which consists of ultra-precise timing data on 41 millisecond pulsars .
We develop frequentist and Bayesian detection algorithms to search both for monochromatic and frequency-evolving systems .
None of the adopted algorithms show evidence for the presence of such a CGW signal , indicating that the data are best described by pulsar and radiometer noise only .
Depending on the adopted detection algorithm , the 95 % upper limit on the sky-averaged strain amplitude lies in the range 6 \times 10 ^ { -15 } < A < 1.5 \times 10 ^ { -14 } at 5 { nHz } < f < 7 { nHz } .
This limit varies by a factor of five , depending on the assumed source position , and the most constraining limit is achieved towards the positions of the most sensitive pulsars in the timing array .
The most robust upper limit – obtained via a full Bayesian analysis searching simultaneously over the signal and pulsar noise on the subset of ours six best pulsars – is A \approx 10 ^ { -14 } .
These limits , the most stringent to date at f < 10 { nHz } , exclude the presence of sub-centiparsec binaries with chirp mass \mathcal { M } _ { c } > 10 ^ { 9 } M _ { \odot } out to a distance of about 25Mpc , and with \mathcal { M } _ { c } > 10 ^ { 10 } M _ { \odot } out to a distance of about 1Gpc ( z \approx 0.2 ) .
We show that state-of-the-art SMBHB population models predict < 1 \% probability of detecting a CGW with the current EPTA dataset , consistent with the reported non-detection .
We stress , however , that PTA limits on individual CGW have improved by almost an order of magnitude in the last five years .
The continuing advances in pulsar timing data acquisition and analysis techniques will allow for strong astrophysical constraints on the population of nearby SMBHBs in the coming years .