Research Summary

Group Member: Ben Winjum

My doctoral research focused on the computer simulation of stimulated Raman scattering (SRS), a laser-plasma instability in which a light wave decays into a scattered light wave and an electron plasma wave. SRS is a potential threat to inertial fusion energy (IFE) devices such as the National Ignition Facility (NIF). If the laser is scattered backwards due to SRS, energy will be directly lost from the fusion drive. Furthermore, the plasma waves generated by SRS can produce hot electrons that threaten to preheat the fuel before it is sufficiently compressed for fusion to occur.

SRS is a complex instability in this regime due to the nonlinear interaction of the scattered plasma wave with resonant electrons in the plasma. Using the electromagnetic Particle-In-Cell (PIC) code OSIRIS, I simulated SRS in 1 and 2D under conditions relevant to Inertial Fusion Energy (IFE) and the National Ignition Facility (NIF). In this regime, the plasma wave has a kλD ~ 0.3 (k is the plasma wave wavenumber and λD is the plasma Debye length) and kinetic effects are important even for small amounts of growth. OSIRIS simulations showed that inflation, frequency shifts, sideband instabilities, beam modes, pump depletion, plasma wave convection, plasma length, and ion motion all played a role in SRS dynamics. While each nonlinearity is a subject in its own right, an array of simulations were used to study the dynamics of SRS behavior as a whole.

In my dissertation, I presented a comprehensive picture of the onset, saturation, and recurrence of SRS. The onset was shown to depend on the convective gain length in the strongly damped regime. Even though SRS was below the absolute threshold, it was shown to grow at the undamped absolute growth rate due to the effect of trapped particles. Saturation was shown to depend on both the plasma wave’s nonlinear frequency shift and pump depletion, with sidebands and beam modes growing significantly only after saturation. Following saturation, plasma waves were shown to convect as a packet which Raman scatters at a shifted frequency, with recurrence depending on the nonlinear packet speed, shifted frequencies, and pump depletion. (See Figure for packet evolution during SRS).

 

alt
The evolution of plasma wave packets generated
from Stimulated Raman Scattering. The packets
erode as they propagate to the right. Results from
an OSIRIS simulation.

The 1D simulations indicated that total time-averaged reflectivities less than 1% required the ratio of the laser speckle length to the convective gain length to be <~ O(1). The total time-averaged reflectivity was shown to be lower when the ratio of the growth rate to the detuning rate due to the nonlinear frequency shift was lower (<~ O (1)) and lower for shorter simulated laser speckle lengths. Instantaneous reflectivity levels were shown to increase in time when multiple plasma packets existed simultaneously within the simulated space. 2D reflectivity was shown to be lower in comparison with 1D due to transverse localization of the plasma wave, but the same dependence of the instantaneous reflectivity on the nonlinear frequency shift and plasma packet effects was shown. The results indicated that mesoscale models that incorporate kinetic effects must include the effects of plasma packets and a nonlinear frequency shift, albeit the frequency shift was shown to be larger than theoretically expected.

My postdoctoral research continues to explore SRS, but with a focus on the hot electrons that are generated.  In addition to the primary SRS processes (which can include either a forward scattered or backward scattered light wave), the scattered light waves from primary SRS can also decay via SRS.  These rescattering processes produce plasma waves that have intermediate phase velocities between those of the primary forward and backward SRS plasma waves.  Forward scattering by itself has a phase velocity which is normally too high to interact with the background electron distribution, but when all of these scattering processes occur simultaneously, electrons from the backscatter get bootstrapped into the rescattered plasma wave and then into the forward scattered plasma wave.  The result is a cascading acceleration of electrons to very high energies – this cascade is evident in the second Figure.

alt

Evolution of the distribution function due to stimulated
Raman scattering. Early in time only backscatter has
occurred which flattens the distribution function. Late in
time forward and rescatter of forward scatter occurs
generating super thermal electrons to 1 MeV. Results from
an OSIRIS simulation.