Joshua Wiggs
The Jovian magnetosphere is loaded internally with material from the volcanic moon of Io, this is ionised and brought into co-rotation to form the Io plasma torus. Plasma is removed from the torus mainly via ejection as energetic neutrals and by bulk transport into sink regions in the outer magnetosphere. The physical process generally considered to be responsible for bulk transport is the centrifugal-interchange instability, analogous to the Rayleigh-Taylor instability but with centrifugal force taking the place of gravity. This allows hot, tenuous plasma to exchange with cool, dense plasma, moving material from the inner to outer magnetosphere whilst returning magnetic flux to the inner magnetosphere. Although centrifugal interchange is the favoured process, modelling and interpretation of data often assume a diffusive-type process. In this work we have developed a full hybrid kinetic-ion, fluid-electron model to examine bulk plasma transport. The technique of hybrid modelling allows for probing of plasma motions down to individual ions, allowing for insights into particle motions on spatial scales below the size of magnetic flux tubes. It also provides a computational framework capable of capturing large-scale flow dynamics, on the order of a planetary radii in some cases. Results from this model will allow for the examination of bulk transport on spatial scales not currently accessible with state-of-the-art models, improving understanding of mechanisms responsible for moving particles between flux tubes and from the inner to the outer magnetosphere. In this poster a series of benchmarks will be summarised validating the performance and accuracy of the model.
The model considers ions using a Particle-In-Cell (PIC) description and the electrons are a neutralising magnetohydrodynamic (MHD) fluid. A Cartesian grid is overlaid across the simulation region on the vertices of which the electromagnetic (EM) fields are calculated. The model is advanced through time numerically.
Examining the computational performance of the model using a 10x10m surface with a 51x51 grid overlaid, it is determined that as the number of particles in the domain is increased, the time taken to perform each temporal step increases linearly. However, when the time taken per step is examined per particle it is found that as the number of particles is increased the time taken per particle decreases, this trend continues until a critical value of 47 μs is reached. When compared to a highly optimised PIC model this critical value is found to be 2 orders of magnitude greater, highlighting the need for further work.
Results obtained from the model require validation to ensure they are accurately reproducing plasma motions. Therefore, a range of test simulations are performed. From these comparisons a set of benchmarks are constructed.
First, a 240s ray-trace of an ion’s path is shown. The region through which the particle travels contains a uniform magnetic field of 1nT. Close agreement between values calculated and those observed in the model is found. There are 4 separate ray traces, each is for a separate simulation with the size of the temporal step equal to a proportion of the particles' gyro-frequency. This shows that the temporal resolution of the model must be at least an order of magnitude below the gyro-frequency to obtain accurate results.
Then particle diffusion is examined to ensure the ion’s are correctly interacting via the PIC grid. The ions diffuse from an initially compressed distribution to occupy all space available. 400 particles were initialised in a 1x1m area at the centre of the model. The particle positions on each second are plotted over a diffusive fluid model of the same region. It is seen that the particle distribution matches well with the contours of the fluid.
Rotational forces are tested by turning off the EM fields. Examining the path of a single ion over 3 hours reveals it moving radially outwards with a small deflection in the azimuthal direction, as expected. It is initialised with a position that would be expected to be within Io’s plasma tours.
Finally, by perturbing the velocity of the ions in a weakly magnetised domain an ion-acoustic wave is launched. The accuracy of complex plasma dynamics is ensured by comparing the observed wave speed in the model to that obtained analytically. Using multiple simulation runs it is possible to obtain the wave speed as a function of plasma pressure, which is used to calculate the adiabatic index of the modelled electron fluid.