The LISA Pathfinder (LPF) mission has demonstrated the ability to limit and measure the fluctuations in acceleration between two free falling test masses down to sub-femto-g levels. One of the key elements to achieve such a level of residual acceleration is the drag free control. In this scheme the spacecraft is used as a shield against any external disturbances by adjusting its relative position to a reference test mass. The actuators used to move the spacecraft are cold gas micropropulsion thrusters. In this paper, we report in-flight characterization of these thrusters in term of noise and artefacts during science operations using all the metrology capabilities of LISA Pathfinder. Using the LISA Pathfinder test masses as an inertial reference frame, an average thruster noise of $\sim 0.17\mu \mathrm{N}/\mathrm{Hz}$ is observed and decomposed into a common (coherent) and an uncorrelated component. The very low noise and stability of the onboard metrology system associated with the quietness of the space environment allowed the measurement of the thruster noise down to $\sim 20\mu \mathrm{Hz}$, more than an order of magnitude below any ground measurement. Spectral lines were observed around $\sim 1.5\mathrm{mHz}$ and its harmonics and around 55 and 70 mHz. They are associated with the cold gas system itself and possibly to a clock synchronization issue. The thruster noise-floor exhibits an excess of $\sim 70\%$ compared to characterization that have been made on ground on a single unit and without the feeding system. However this small excess has no impact on the LPF mission performance and is compatible with the noise budget for the upcoming LISA gravitational wave observatory. Over the whole mission, nominal, and extension, the thrusters showed remarkable stability for both the science operations and the different maneuvers necessary to maintain LPF on its orbit around L1. It is therefore concluded that a similar cold gas system would be a viable propulsion system for the future LISA mission.

• Received 29 April 2019

DOI:https://doi.org/10.1103/PhysRevD.99.122003