In 2017, my collaborators and I examined all Rossi X-ray Timing Explorer observations of fifteen accreting black holes in our Galaxy. We discovered that the lag of the QPO - whether the signal arrives first at high or low X-ray energies - depends on our viewing angle of the binary. This simple but striking effect presents strong evidence that the QPO is caused by a geometric process, such as General Relativistic Lense-Thirring precession close to the black hole.

The famous accreting black hole GRS 1915+105 shows odd behaviour in many ways; one is that its QPO frequency displays a dependence on energy. As this does not fit with canonical QPO models, we set out to test whether this may be a selection effect, due to a correlation between hardness of the emission and QPO frequency, or could be an intrinsic QPO property. Surprisingly, we find that the latter is the case. This lead us to argue that differential precession, which was not seen before, may explain this variability pattern in GRS 1915+105.

Many advanced studies of QPO try to find changes in the X-ray spectrum, as a function of phase in the QPO signal. However, as QPOs are not perfectly periodic, subsequent cycles of the signal do not necessarily have the same waveform. In this work, lead by Iris de Ruiter (now PhD student at the University of Amsterdam), we explicitly tested whether accreting black holes show consistent waveforms. We found that this is indeed the case, which supports the idea that we can systematically search for changes within a QPO cycle by averaging cycles together.