Context. Detection of electron cyclotron maser (ECM) emission from exoplanets in the 10–40 MHz radio band is likely the only way to measure an exoplanet’s magnetic field directly. However, no definitive detection of exoplanetary ECM emission has been made to date. A detection of the hot Jupiter Tau Boötis b was reported but with an observing mode that is not immune to confusion from off-axis interference, making the detection tentative. Aims. We searched for radio emissions from Tau Boötis b using the Low Frequency Array (LOFAR) in interferometric mode, which employs direction-of-arrival information to discriminate genuine signals from interference. Our aim was to confirm the previous tentative detection or establish an upper limit in the case of a non-detection. Methods. We conducted observations using LOFAR’s Low Band Antenna in interferometric mode, which totalled 64 hours spread over 8 nights. We created a custom data-processing pipeline to mitigate common challenges in decametric radio astronomy, including radio frequency interference, ionospheric distortions, and sidelobe noise from nearby bright radio sources. We used this pipeline to image the field around Tau Boötis b, searching for both quiescent and bursting emission from the direction of Tau Boötis b. Results. Despite the high sensitivity of the interferometric observations and extensive data processing, no significant emission was detected from Tau Boötis b in Stokes V. We establish an upper limit of 2 sigma at 24 mJy for any continuous emission from the exoplanet. The previous tentative detection of 400 mJy is thus not confirmed by the interferometric observations. Conclusions. The previous tentative detection is unlikely to be a bona fide astrophysical signal. Our upper limit is lower than the flux density predicted by scaling laws, meaning either the scaling laws need to be revised or the emission from this particular system is beamed away from Earth.
