1. Electrosprayed Heavy Ion and Nanodrop Beams for Surface Engineering and Electrical Propulsion
Manuel Gamero-Castaño & Daniel Mumm
University of California, Irvine
Paulo Lozano & Markus Buehler
Massachusetts Institute of Technology
Juan Fernández de la Mora & Alessandro Gomez
Yale University

2. Project Goals
Investigate propulsive applications of electrosprayed nanodrops and ions and the interaction of these beams with surfaces:
Modeling of electrospray thruster beams
Optimum design of extracting electrodes based on beam model:
maximization of beam throughput in MEMS electrospray thrusters
elimination of beam impingement on extracting surfaces
Assessment of surface damage caused by nanodroplet and ion beams
molecular dynamics simulation of particle impact
measurement of sputtering rates, characterization of surface damage
First-principles understanding of the generation of electrosprayed nanodroplets and molecular ions in the 1 nm to 1 mm size range.
Atomization of energetic fluids for dual electric/chemical propulsion and chemically enhanced surface processing
Study of molecular fragmentation in ion beams, and its effect on thruster performance
Etching by the ionic liquid ion source (ILIS) and study the focusability of its beam

3. Modeling of Electrospray Thruster Beams (Droplet Emission Mode)
Modeling of Electrospray Thruster Beams (Droplet Emission Mode). Experimental Characterization
Experimental setup: two time of flight collectors and a retarding potential analyzers mounted on an XYZ stage are used to characterize the beams of the ionic liquid Emi-Im, electrosprayed from a single emitter/extractor source. The temperature of the emitter can be varied and controlled. This setup let us characterize beam profiles and particle composition inside the beam as a function of position, mass flow rate and working temperature. This information is in turn used to guide the model, and validate model results
Beam profiles measured at different beam currents and working temperatures. The profiles are measured m downstream from the emitter, in a region of negligible E-field. The beams are axisymmetric, and broaden up at increasing beam current. Although the beam current-mass flow rate law depends on the working temperature, the broadness of the beams remains remarkably constant at constant beam current and varying temperature
Time of flight waves of the whole beam for different beam currents at 21C. The beam is made of ions and charged droplets. Additional retarding potential and TOF measurements taken with the point detectors show that the ion population is restricted to the core of the beam, while droplets appear at all polar angles

4. Modeling of Electrospray Thruster Beams (Droplet Emission Mode)
Modeling of Electrospray Thruster Beams (Droplet Emission Mode). Beam Model and Results
The model integrates the equations of motion of droplets defining envelopes enclosing a given fraction of the beam current. Each envelope contains droplets with identical charge to mass ratio (obtained from the TOF waves). The electric field acting on the droplets is decomposed in two components, one obtained by solving Laplace equation for a given potential distribution in the electrodes, and a second associated with the distribution of charge enclosed by the envelopes (space charge). The electric field associated with the space charge is computed by the superposition of fields induced by discrete charges
Good estimates of the initial velocity and position of the trajectories are key to obtaining accurate envelopes. These parameters are estimated with information contained in the retarding potential spectra
The external electric field is computed for the emitter-extractor geometry used in the experimental characterization. This allows direct comparison between experimental and model results. The agreement is good, as shown by the similar shapes of the experimental and computed accumulated current vs polar angle curves
The beam model will be used to design emitter-extractor geometries of MEMS thrusters, with the goal of maximizing thrust density and eliminating beam impingement. It will also be extended to simulate ion beams (i.e. electrospray thrusters operating in the purely ionic regime)

5. Characterization of Thruster Surfaces Bombarded by Energetic Nanodroplets
c-Si
Pt protective
layer
bombarded surface
with amorphous layer
Silicon wafer bombarded by electrosprayed nanodroplets at varying acceleration voltages. The sputtering yields are of order one. The surfaces display craters with diameters of up to a few microns. Silicon is a key material since extracting electrodes and other components of MEMS thrusters are made of Si
Cross-sectional TEM image of single-crystal Si bombarded by nanodroplets. The damaged area is an amorphous layer with a thickness of some 20 nm
Cross-sectional TEM image of single-crystal Aluminum bombarded by nanodroplets. The damaged area is an amorphous oxide layer of approximately 40 nm. We are investigating whether the oxidation occurs during the impact, or after exposure of the sample to atmospheric conditions
Cross-sectional TEM image of a single-crystal Nickel target bombarded by nanodroplets. The damaged area remains crystalline, although different fringe patterns are indicative of a polycrystalline morphology. This may be the result of the melting and fast quenching following the nanodroplet impacts

6. Molecular Dynamics Simulations of Nanodroplet Impact
Impact of a nanodroplet (10 nm diameter, 6.4 km/s, 24 kV acceleration voltage) on single crystal Si. The density of the nanodroplet is that of EMI-Im, and is made of 1224 spheres with the EMI-Im molecule mass. The target is a nm x nm × nm slab
Temperature fields at 0 ps, 1 ps, 5 ps, 10 ps, 20 ps, and 70 ps from the time of impact. Upon impact a substantial amount of energy is dissipated in the contact region, raising its temperature above several thousand degrees (see 5 ps frame). A layer surrounding the crater remains above the normal melting point for over 10 ps, before cooling down and equilibrating with its surroundings
Pressure field at 3.5 ps. The area at the bottom of the crater is decelerating the projectile and therefore is highly compressed, while a nearly spherical elastic wave moves away from the point of impact. The jump in temperature across the shock wave averages 9 K, a marginal increase consistent with the nearly isentropic nature of shock compression.
F. Saiz and M. Gamero-Castaño, Journal of Applied Physics, in print

7. Molecular Dynamics Simulations of Nanodroplet Impact
Evolution of the total thermal, translational and potential energies of the slab, and the total kinetic energy of the projectile, ETh(t ), ECM(t ), EP(t ) and EK(t ). The projectile molecules lose 90% of their initial kinetic energy within 3 ps. By this time 12%, 18% and 60% of the projectile’s energy has been transferred to the slab in the form of translational, thermal and potential energies respectively. After peaking at 12 ps, the thermal energy in the slab is gradually lost by heat conduction with the surrounding thermal bath, while the potential energy asymptotes to a constant fraction of 24%. This potential energy excess is mostly distributed in the region near the impact, where the atomic arrangement is highly disordered by the end of the simulation.
Evolution of the temperature, pressure, coordination number and melting point in a control volume below the surface of impact. Upon impact the pressure and temperature start to increase, and rapidly plateau around 5.2 GPa and 350 K at 1.8 ps. At about 2.2 ps the temperature exceeds the melting point, and the control volume starts to undergo a solid to liquid phase transition. This is confirmed but the coordination number fluctuating around a value of 6. After 8 ps the temperature decays with cooling rates as high as 6.8x1013 K/s at the normal melting point, and 3.7x1013 K/s at the glass transition temperature of 1060 K. These cooling rates exceed by over four orders of magnitude the typical value of 109 K/s needed to prevent the regrowth of the crystalline phase, and the previously melted material solidifies as an amorphous phase.

8. Ion Emission Simulations: EMI-BF4
Ion pairs are stacked and allowed to equilibrate to form a droplet.
Largest simulation uses 117,912 atoms with a length of around 100 Å; this captures majority of Coulomb interactions.
An electric field is applied to study emission processes and emitted ion energy distributions.
Droplets
Droplet simulation are too expensive to obtain detailed fragmentation statistics.
Instead, thousands of dimer samples are generated at representative energies.
Individual Solvated Ions
Stack and equilibrate
Apply electric field
Sample individual dimers

9. Breakup Comparison for EMI-BF4 and BMI-PF6
1.5 V/nm electric field
2.3 eV excess energy above 300K
100% negative dimers broken in 200 ps
88% negative dimers broken in 200 ps
Time at breakup (ps)
pdf
EMI-BF4
BMI-PF6
10,000 negative dimer samples are generated at energies typical of those found from large droplet emission simulations; these are energies in excess of the initial droplet energy at 300K.
Typical energies are assumed similar for all ionic liquids.
Sample generation
Longer for BMI-PF6 than for EMI-BF4 at all energies in the examples shown.
Energy is distributed among more degrees of freedom for more complex molecules, giving smaller ion separations and increased Coulomb force, so greater resistance to breakup.
Consistent results seen at various energies and also experimentally for three liquids.
Time for breakup
Breakup causes efficiency losses, so complex ionic liquids appear to be the most efficient propellants.
Practical significance

10. Ion Emission Processes