The very first phase (typ. < 100 ms) of the start-up of rocket engines in microgravity implies the injection of liquid propellant in partial vacuum. The knowledge of the ensuing flow structure is required to master the ignition transient. In that scope, experiments have been undertaken to investigate the flow resulting from the sudden opening of a ball valve and the injection of liquid in a low pressure environment. In the present study, we used an almost pure (less than 5ppm of dissolved oxygen) fluorocarbon, namely tetradecafluorohexane C6F14. The experimental conditions were controlled by two main parameters: the initial superheat governed by the initial fluid temperature and the initial pressure in the tank (accounted for by a Jakob number) and the incoming liquid flux governed by the liquid pressure in the circuitry and the valve discharge coefficient (accounted for by a Weber number). A large range of superheat has been considered by varying the initial tank pressure from 0.3 to 2 mbar. In most conditions, explosive vaporization has been observed (high speed imaging) that leads to the formation of dense sprays. From pressure and temperature signatures, the influence of the two control parameters on the vaporized fraction of the incoming liquid flux has been investigated and will be discussed. We will also report on preliminary measurements of the dispersed phase characteristics such as the size and velocity of drops (measured by Phase Doppler Anemometry) and their volume concentration (measured by optical probes). In particular, for a constant Weber number, the droplet velocity was found proportional to the initial superheat while their mean size remained almost unaffected. The later result markedly differs from most previous experiments, and in particular from those by Lecourt et al. , who found a strong correlation between the droplet diameter and the superheat. A possible explanation of that difference may be due to the injector geometry as small orifices or opening were used in most previous experiments. Instead, our experimental conditions seem to be similar to the boiling fronts investigated notably by Reinke and Yadigaroglu and by Hahne and Barthau . For such situations, the drop formation mechanism is not well understood, and possible scenarii will be discussed based on the measured dispersed phase characteristics.
