University of Southern California Dissertations and Theses

An investigation into magnetically induced extractor-less electrospray propulsion devices

(USC Thesis Other)

An investigation into magnetically induced extractor-less electrospray propulsion devices

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

An Investigation into Magnetically Induced Extractor-less Electrospray Propulsion
Devices

by

Robert Antypas

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(ASTRONAUTICAL ENGINEERING)

August 2021

Copyright 2021 Robert Antypas

ii
Dedication

To my loving wife Mary, son Blake and newborn daughter Astrid, thank you for always
supporting me and being there when I needed a break, especially when I could not realize It was
time to take a step back. Thank you to my mom and dad; your commitment to education and
expanding one’s horizon is contagious and commendable. To my dad, our talks and discussions
on the theory of everything from electromagnetism to gravitational warping led me to realize this
innovative discovery.

iii
Acknowledgments

To my Advisor: Thank you, Dr. Wang, for being the sounding board I needed to find
direction and conviction on this long path. Jeff, Kevin, Nick, thank you for keeping a good
atmosphere in the lab and for the fun times we had opining the new space and getting the LAPD
up and running as an electrospray laboratory. A special thanks to; Mike Holmes, Mike Natisin,
Dan Eckhardt, and Justin Koo at the Air Force Research Laboratory.
This research was made possible by the DOD Science Mathematics and Research for
Transformation (SMART) Scholarship program.

iv
Table of Contents
Dedication ...................................................................................................................... ii
Acknowledgments ........................................................................................................ iii
List of Figures ................................................................................................................. vi
List of Tables ................................................................................................................... ix
List of Nomenclature ......................................................................................................... x
Abstract ......................................................................................................................... xi
CHAPTER 1: INTRODUCTION ...................................................................................... 1
1.1 Background ....................................................................................................................... 1
1.3 Outline and approach ....................................................................................................... 7
CHAPTER 2: LITERATURE REVIEW .......................................................................... 11
2.1 Electrostatic Electrospray .............................................................................................. 11
2.1.1 Colloid Thrusters ..................................................................................................................... 11
2.1.2 FEEP or LMIS Thrusters ......................................................................................................... 12
2.1.3 PIR or ILIS Thrusters ............................................................................................................... 13
2.2 Magnetically Induced Ion Mobility ................................................................................. 19
CHAPTER 3: METHODOLOGY ................................................................................... 22
3.1 Laboratory and Facility Equipment ............................................................................... 23
3.1.1 Three-Axis Probe System and LabVIEW Control ................................................................. 24
3.1.2 Faraday Probe/Micro Faraday Probe ..................................................................................... 28
3.1.3 Langmuir Probe ....................................................................................................................... 30
3.1.4 Retarding Potential Analyzer ................................................................................................. 31
3.1.5 Faraday Plate ........................................................................................................................... 34
3.1.6 High-speed TOF Diagnostics ................................................................................................. 34
3.2 Data Acquisition and Reduction .................................................................................... 39
3.2.1 Noise ......................................................................................................................................... 40
3.2.2 Grounding ................................................................................................................................ 43
Chapter 4: The USC Testbed Thruster ...................................................................... 47
4.1 Thruster Design and fabrication .................................................................................... 47
4.1.1 Emitter and Extractor .............................................................................................................. 49
4.1.2 Propellant Loading .................................................................................................................. 52
4.2 Test Results ..................................................................................................................... 53
4.2.1 Thruster One Testing .............................................................................................................. 53
4.2.2 Thruster 2 Testing ................................................................................................................... 55
a. Plume Characterization ........................................................................................................... 57
b. Thruster 2’s Disassembly ....................................................................................................... 60
4.2.3 Thruster 3/4 Testing ................................................................................................................ 63
4.2.4 TOF Results ............................................................................................................................. 69
Chapter 5: The Magnetically induced Test Thruster ................................................ 77
5.1 Thruster Development .................................................................................................... 77
5.2 Thruster Test Results ..................................................................................................... 81

v
Chapter 6: Conclusions and Proposed Future Work ............................................... 87
Appendix A: Electrospray Thruster Mathematics Principles. ................................. 93
A.1 Fundamental Physics of Electrospray Operation ....................................................... 93
A.2 Performance characteristic equations ......................................................................... 93
Appendix B: BNG Development ................................................................................. 95
Bibliography .............................................................................................................. 100

vi
List of Figures

Figure 1: (Left) Example of 4 emitter PIR Thruster Block Diagram. (Right) Example of Single
capillary fed Colloid Thruster block diagram ................................................................................ 3
Figure 2: Geometric relationship between the electrospray tip and the extraction electrode ....... 4
Figure 3: Top left: perfect symmetrical ion emission. Top right: clipped/broken emitter tip ........ 5
Figure 4: Top left: perfect symmetrical ion emission. Top center: .001” grid miss alignment ..... 5
Figure 5: (Left) Dark spots are areas of the USC emitter array where arc events shorted the
emitter to the extractor and scorched an emitter tip. (Right) build up on MIT TILE Extractor
after 172 hrs. [7] ............................................................................................................................. 6
Figure 6: USC Testbed Thruster CAD drawing ............................................................................. 8
Figure 7: As-built assembled USC Testbed Thruster (UTT) .......................................................... 9
Figure 8: MTT CAD rendering and wire wrap of the first coil pair .............................................. 9
Figure 9: An SEM image of the UTT emitter substrate ................................................................ 15
Figure 10: Ion wear pattern and grid deterioration due to ion sputtering on the acceleration grid
[70] ............................................................................................................................................... 18
Figure 11: Helmholtz coil pair with driver schematic .................................................................. 19
Figure 12: LAPD primary vacuum chamber ................................................................................ 22
Figure 13: Probe setup inside LAPD main chamber .................................................................... 23
Figure 14: Tri-axis data traversing system .................................................................................. 24
Figure 15: Arduino MEGA with RAMPS 1.4 stepper driver for three axis control and three I2C
0-5V control chips. ........................................................................................................................ 25
Figure 16: LAPD LabVIEW user interface. ................................................................................. 25
Figure 17: Voltage inverter for command signal of RPA ............................................................. 26
Figure 18: LabVIEW block diagram for thruster operation and probe positioning. ................... 27
Figure 19: Conditional loops for user interface ........................................................................... 27
Figure 20: Limit switch (left) and stepper motor (right) initialization for Arduino LabVIEW
interface ........................................................................................................................................ 28
Figure 21: Faraday probe/micro-Faraday probe and electrical schematic ................................ 28
Figure 22: Heat map of MFP with .75” Dis. to thruster face. ..................................................... 30
Figure 23: Langmuir probe and electrical schematic .................................................................. 30
Figure 24: An example of an LP measurement and analysis [81] ............................................... 31
Figure 25: Retarding potential analyzer and electrical schematic .............................................. 32
Figure 26: RPA intercepted current vs Voltage applied to sweep grid ........................................ 33
Figure 27: 6” large collection plate and electrical schematic ..................................................... 34
Figure 28: BNG Test setup, chamber geometry, where LBC = 30 in., HT = 0.106 in., LTB = 0.5
in., and HFP = 5.9 in. Simplified ion beams are also depicted. ................................................... 36
Figure 29: BNG projected deflection with VBNG =1020V and a 12° thruster half-angle at
Tthruster=1400V ............................................................................................................................... 37
Figure 30: reflection gate test setup, chamber geometry, where LBC = 28 in., HT = 0.106 in.,
LTB = 2 in., and HFP = 5.9 in. Simplified ion beams are also depicted. .................................... 38
Figure 31: High frequency analysis of lab components contribution to background noise ......... 42
Figure 32: Example time of flight run with ground induced noise ............................................... 43
Figure 33: Faulty chamber ground .............................................................................................. 44
Figure 34: New dedicated facilities ground in LAPD .................................................................. 45

vii
Figure 35: High frequency analysis of lab components contribution to ground floor noise with
the newly installed dedicated data ground. .................................................................................. 45
Figure 36: USC Thruster V1 Expanded Rendering. ..................................................................... 48
Figure 37: Extractors ................................................................................................................... 50
Figure 38: Additively manufactured grids with insufficient support ............................................ 51
Figure 39: CAD renderings of novel extractor profiles, .............................................................. 51
Figure 40: COMSOL modeling of a novel extractor design ......................................................... 52
Figure 41: Propellant loading of first (left) and second thrusters(right). .................................... 53
Figure 42: Thruster 1 Emitter profile. Irregular emitter tips and a row of clipped emitters ....... 54
Figure 43: Source Voltage and Current of Thruster 1 Initial Test. .............................................. 55
Figure 44: Thruster 2 machined emitters ..................................................................................... 56
Figure 45: Startup Source Voltage v Current of Thruster 2. ........................................................ 57
Figure 46: Heat maps of MFP with 4” dis. (left) .75” dis. (right) to thruster face. .................... 58
Figure 47: MFP heat map at .25” from thruster face .................................................................. 58
Figure 48: FP heat maps of the horizontal plane. ........................................................................ 59
Figure 49: FP heat map of the vertical plane. .............................................................................. 60
Figure 50: Thruster 2’s disassembly pictures .............................................................................. 61
Figure 51: SEM Image of carbon structure on the top of the emitter. ......................................... 61
Figure 52: New emitters for thruster 3 and 4 ............................................................................... 63
Figure 53: Plume scan of thruster 3 operation showing 14.47°plume half angle ........................ 64
Figure 54: RPA Plume analysis of thruster 3 negative mode operation ...................................... 64
Figure 55: Large structure on the tip of the second emitter from the top left .............................. 65
Figure 56: Carbon Build up on fired emitter (left) and melted borosilicate pores from arc events
(right) ............................................................................................................................................ 65
Figure 57: Reloading of thruster 3 ............................................................................................... 66
Figure 58: Thruster 4 loading ...................................................................................................... 66
Figure 59: simultaneous quasi-neutral bipolar thruster operation of thrusters 3 and 4. ............ 67
Figure 60: Thruster 4 start up and shut down .............................................................................. 68
Figure 61: RPA sweep of thruster 4 plume potential ................................................................... 68
Figure 62: Collected TOF raw data (left) and Time Synchronized Averaged data (right) .......... 70
Figure 63: Top-down schematic of test setup in the chamber ...................................................... 70
Figure 64: Axial plume scans of BNG off and BNG on ................................................................ 71
Figure 65: Plume analysis for BNG thruster tuning .................................................................... 72
Figure 66: Idealized and Expected Time of Flight data ............................................................... 73
Figure 67: Reflection gate normalized collected signal: gate voltage 1650V Thruster voltage
1450V. Negative BF4 collection (left) Positive EMI collection (right) ......................................... 74
Figure 68: BNG normalized collected signal: gate voltage 720V Thruster voltage 1450V.
Negative BF4 collection (left) Positive EMI collection (right) ...................................................... 75
Figure 69: CAD blow out of MTT-3 with propellant reservoir. ................................................... 77
Figure 70: MTT versions as built, (left) MTT-1, (middle) MTT-2, (right) MTT-3 ....................... 78
Figure 71: Static depiction of the magnetic field generated in the emitter region of the MTT .... 79
Figure 72: Line diagram of Helmholtz driver [79] ...................................................................... 80
Figure 73: Electric Displacement field generated by 156kHz operation of the MTT. ................. 80
Figure 74: MTT Surface induced current density at emitter tips ................................................. 81
Figure 75: Bench top test of MTT ................................................................................................. 82

viii
Figure 76: a) A simplified representation of a firing thruster’s expected results. b) Results from
3 separate test firings overlayed to show consistent operation with respect to theoretical. ........ 83
Figure 77: Histogram of RPA RMS with MTT-3 off (left), Histogram of RPA RMS with MTT-2
Firing (right) ................................................................................................................................. 84
Figure 78: RPA sweep of MTT-2 plume spectrum ........................................................................ 84
Figure 79: New Robu P5 Borosilicate glass Left and USC Emitter ROBU Borosilicate glass
Right. ............................................................................................................................................. 85
Figure 80: AMPS Thruster depiction with integrated extractor .................................................. 91
Figure 81: Thruster performance with extrapolated growth for 200𝜇𝜇 m pitch and 35𝜇𝜇 m pitch tip
density ........................................................................................................................................... 91

ix
List of Tables

Table 1: Arduino RAMPS 1.4 stepper control pins ...................................................................... 28
Table 2: Expected Constituents from EMI
+
BF4
-
........................................................................... 52
Table 3: Thruster 2 firing times and mass change ....................................................................... 62
Table 4: Expected Constituents and flight times from EMI
+
BF4
-
................................................. 69
Table 5: Comparison of drive specifications for MTT-2 and MTT-3 ........................................... 82
Table 6: Comparison of UTT-2,3,4 and MTT-2 ISP and thrust ................................................... 89

x
List of Nomenclature

Vth/start = estimated starting potential for emission (V)
h = emitter height (μm)
Rc = emitter apex radius of curvature (m)
ds = emitter apex to extractor grid separation (μm)
θ = beam angle measured
θeff = effective beam angle
θ0 = beam angle offset
Vem = emitter voltage (V)
qi = species i charge (C)
ci = species i particle velocity (m/s)
mi = mass of species i (kg)
𝑚𝑚 ̇ = mass flow rate (kg/s)
𝑁𝑁 ̇ = particle flow rate (particles/s)
Va = average beam potential (V)
Ib = beam current (A)
Iem = emitted current (A)
Iex = extractor grid current (A)
T = calculated thrust (N)
Isp = specific impulse (s)
ηprop. = propulsive efficiency (%)
j(θ) = collected current density (A/m
2
)
Lp = distance from the source to the probe (m)
ΔV = energy deficit (V)
En = characteristic activation energy
θT = Taylor angle
ϵ
0
= permittivity of free space
Φ = potential (V)
∆P = hydrostatic pressure
𝛾𝛾 = surface energy of fluid (J/m
2
)
kb = Boltzmann’s constant
μ
0
= Permeability of free space 4 π x 10
− 7
(𝑇𝑇 𝑚𝑚 / 𝐴𝐴 )
B = Magnetic field strength (T)
L = Inductance (H)
ω = Angular frequency = 2 πf

xi
Abstract

Electrospray research has made immense progress since its infancy in the 1960’s
and its rebirth in the early 2000’s. Successfully flight tested, this technology shows great
promises for high efficiency and low thrust applications, with the ultimate high thrust
utility of these devices relying on the elusive promise of emitter tip densification. While the
physics that enable conventional electrospray devices is simple, imposing engineering
challenges still persist in the relationship between the extraction plane and the emitter tips.
These engineering challenges practically limit the size and densification of conventional
electrospray devices. The USC Magnetic Testbed Thruster (MTT) was developed as a
passively fed, porous media, pure ionic, extractor-less electrospray thruster. The MTT
eliminates the need for an extractor entirely by utilizing dynamic magnetic induction to
create a virtual electric field that accelerates bipolar species of an ionic liquid. The virtual
electric field that is induced in the conductive propellant takes the place of the fixed field
between a conventional extractor and the emitter array thus eliminating the largest barrier
to scalability and densification.
To baseline the novel MTT performance characteristics, a state-of-the-art
electrospray testing facility was constructed to evaluate conventional and novel
electrospray devices. The Laboratory of Astronautical Plasma Dynamics was retrofitted
with a multitude of Faraday and retarding potential analyzer probes as well as a full 3 axis
automated probe sweep system with both high and low speed data acquisition. ISP and
thrust were calculated with Time-of-Flight Spectroscopy utilizing a Bradbury Neilson
refraction ion diversion gate instead of a conventional reflection suppression grid. This
alternate approach to ion diversion illustrated drawbacks associated with the industry

xii
standard reflection approach. The differences in spectrographic results between the two
methods show a clear difference in emission characteristics of thrusters under reflected
investigation and not just differences in results due to the two methods, the differences can
be attributed to the deleterious effects of back streaming large quantities of high energy
ions on the emitter surface for extended periods.
Four conventional configuration 25 emitter USC Testbed Thrusters (UTT) were
designed/constructed and tested, utilizing EMI-BF
4
electrospray emission from 1cm
borosilicate disks. The UTT was used to baseline the diagnostic tools in the laboratory with
respect to literary peers and to determine the densification threshold of both the current
extractor design as well as the theoretical density of the porous structure of the borosilicate
glass substrate materiel. The UTT tested extractor configurations with pitch densities of
533𝝁𝝁 m to 432𝝁𝝁 m. The various prototyped of the UTT displayed performance ranging from
4.2x10
-8
N to 4.5x10
-7
N of thrust and ISP from 2300s to 6300s. Three comparable MTT
prototypes were designed/constructed and tested utilizing the same 1cm 25 emitter disks
used in the UTTs. Utilizing orthogonal matched Helmholtz coil pairs operating at a
dynamic frequency of ~150kHz, the MTT voltage was tuned to +/- 1500V to match the
optimum performance of the UTT for comparison. At the comparable extraction voltages,
the MTT produces 5.5-8.6x10
-8
N of thrust with an ISP range of 3600-5700s. Showing that
removal of the extractor can eliminate the constrains of conventional electrospray thrusters
such as service life and overall component size without sacrificing peak thrust and specific
impulse. Currently, the sizeable unobstructed extraction window in the USC MTT is 1
cm
2
. This window is scalable as a function of the radius of the matched Helmholtz coils.
Modeling of the prototype thruster configuration shows the promise for both non-planar

xiii
(curved) and ultra-dense emitter arrays. Projections on the same 1cm extraction discs at
the substrate porosity limited pitch of 200 𝝁𝝁 m shows a two-order magnitude increase in
thrust with constant specific impulse. New substrate materials can be investigated to
increase this densification further which would vastly increase the potential scalability and
utility of electrospray devices.

1
CHAPTER 1: INTRODUCTION

1.1 Background
Colloid Spray, Field Emission Electric Propulsion (FEEP), Liquid Metal Ion Sources
(LMIS), Pure Ion Regime (PIR), and Ionic Liquid Ion Sources (ILIS) all share the broad
umbrella term of electrospray devices. Each technological subset operates by accelerating
different types of ions in a static electric field. Free ions are emitted when a significant electric
field is applied to a conductive fluid. The force acting on the ions creates a Taylor cone when the
potential in the fluid reaches the Rayleigh limit. At this point, ions are ejected at high velocity,
and this ensuing beam can be used to generate propulsion. This method for the acceleration of
ions was theorized by Sir Rayleigh in 1882 [1], experimentally verified by Zeleny in 1914 [2]
and mathematically derived by G. I. Taylor in 1964 [3], [4]. Many devices have been constructed
to utilize this phenomenon. These electrospray devices have been used in a myriad of fields such
as spacecraft propulsion, medical diagnostic devices, and industrial spray applicators. It is the
former application, propulsion, that this thesis intends to investigate. A Taylor cone can be
formed on the molecular scale, and because of this, these ES thrusters can be extremely small.
These miniature propulsion devices show great promise in the small and microsatellite field.
They are mechanically simple and can be scaled simply by adding more emission sites.
Electrospray thrusters show promise for high ISP (>3000) high trust outputs, when densely
gridded with thousands to millions of individual emitters. Until now, all of the propulsion
devices utilizing the electrospray phenomenon have relied on a static electric field generated
between two charged surfaces with an ionic, sprayable, fluid held in liquid form. The practical
engineering challenges that arise from the implementation of a static electric field create
limitations in these devices' practical density and configuration.

2
The propulsion physics of all classes of electrospray thrusters work on the same
principles. The simplest goal of electrospray is to accelerate some form of ion to high velocities.
Whether liquid metal or ionic liquid high velocities can be obtained by differential charge
acceleration in a static electric field. Equation 1 illustrates the fundamental simplicity of the
electrospray operation. At equilibrium and prior to emission, the electrical pressure and
hydrostatic pressure equal the surface energy across the area normal to the surface. Once the
electrical pressure breaks the surface tension, ion evaporation occurs to restore equilibrium.
Equation 2 is the simplest form of the Taylor principle and guides the Taylor cone formation
with its defining 49.29-degree half-angle [4]. This half-angle is achieved when the static electric
field normal to the cone is at equipotential and only has one solution at the Taylor cone angle of
49.29-degrees. Appendix A goes through the mathematical relationships and derivations of these
principles.
1
2
ϵ
0
( ∇ Φ)
2
+ ∆ 𝑃𝑃 = 𝛾𝛾 ∇ ∙ n � ⃗ Eq. 1
𝐸𝐸 𝑛𝑛 = �
2 𝛾𝛾 cot 𝜃𝜃 𝑇𝑇 𝜖𝜖 0
𝑟𝑟 Eq. 2
The simplicity of these physical properties led to a boom in Electrospray development in
the 1960s-1980s. Liquid metal ions were the primary constituents used in this period, due to their
high conductivity. The high power required to accelerate the high molecular density of ionic
fluids ultimately proved too challenging to miniaturize and too heavy to insulate. These factors
caused the technology to remain shelved until the utility of the ionic liquid EMI-BF4 in 2003 for
use in electrosprays was discovered [5]. This discovery led to the revitalization of interest in the
field and its rapid expansion to today.

3
Since a static electric field cannot be created on such a small scale without a physical
structure to hold a charge (electrons) in place, the nuance of thruster operation lies in each
thruster's physical construction. Figure 1 illustrates very simple PIR and capillary fed
electrospray thruster block diagrams. Each thruster consists of three main parts; the emitter tips
or capillaries, the extractor electrode, and a biased ionic liquid or liquid metal in a propellant
reservoir.

Figure 1: (Left) Example of 4 emitter PIR Thruster Block Diagram. (Right) Example of Single capillary fed Colloid
Thruster block diagram

It is in the structure of the extraction electrode that problems with the simplicity of
operation begin to arise. Equation 3 shows the relationship between the Taylor cone tip and the
structure that creates the static electric field. Where Rc is the emitter radius of curvature, ds is the
1D separation distance between the emitter and the extractor, and Vth is the starting or threshold
voltage needed to create emission.
𝑉𝑉 𝑡𝑡 ℎ
≅ �
𝛾𝛾 𝑅𝑅 𝑐𝑐 𝜖𝜖 0
� 𝑙𝑙𝑙𝑙 �
4 𝑑𝑑 𝑠𝑠 𝑅𝑅 𝑐𝑐 � Eq. 3
Figure 2 is an illustration of a PIR emitter tip and the physical relationship to the
extraction plane. To reduce the starting or threshold voltage, one needs to reduce the radius of
curvature (Rc), reduce the separation distance ds, or reduce 𝛾𝛾 by changing the ion source.

4

Figure 2: Geometric relationship between the electrospray tip and the extraction electrode

Using a static electric field means that the escaping ions must pass through the extractor
to escape the thruster and provide thrust. As Seen in Figure 3, the ions must pass from the
“Emitter Tips” through the “Extractor Grid” to generate thrust. Perfect emission occurs when the
extractor is perfectly centered above the emitter tips, and each individual emitter tip is the same
height and shape (Figure 3a). These thrusters are expected to fire for thousands of hours and
extract hundreds of trillions (10^14) of ions every second. If even a tiny percentage of the ions
hit the extractor gird, thrust is lost, and over time ionic liquid buildup shorts out the electrodes.
Mis-matched tip height, tip shape, and overall tip to emitter alignment can lead to extractor
impingement. Miniscule offsets in grid alignment can lead to wide degrees of off-axis emission.
These problems can be visualized in Figure 3 and 4 below. Particle interception is also dependent
on the uniformity of the emitters' array and the consistency of the emitter radius of curvature.

5

Figure 3: Top left: perfect symmetrical ion emission. Top right: clipped/broken emitter tip

Figure 4: Top left: perfect symmetrical ion emission. Top center: .001” grid miss alignment
Top right: .004” grid miss alignment. Bottom left: .006” grid miss alignment.
Bottom center: .008” grid miss alignment. Bottom right: .009” grid miss alignment.

1.2 Motivation and Objectives
All electrospray propulsion derivatives have the same limiting characteristic; they must
all spray ions through a grid of windows. These windows must be held at a potential high enough
to create an electric field that will generate a Taylor cone in the ionic fluid. The emission
characteristics and the emitters' densification are directly impacted by the size and shape of the
ionic windows. As such, the ultimate densification of an electrospray thruster becomes finite.

6
The grids also have so far proved to be the life-limiting component of these thrusters due to
buildup from overspray or off-axis spray.
The current state of the art thrusters suffers anywhere from .1%-10% loss due to
interception [6]. This loss limits current thrusters that are supposed to run for tens of thousands
of hours to between 1-2000 hours. Figure 5 is an example of damage to an extractor due to
prolonged electrical bridging between an extractor and the emitter tips.

Figure 5: (Left) Dark spots are areas of the USC emitter array where arc events shorted the emitter to the extractor
and scorched an emitter tip. (Right) build up on MIT TILE Extractor after 172 hrs. [7]

Some mission profiles require short periods of increased thrust; electrospray can
accommodate this by increasing the extractor's potential and thus increasing the radius of
curvature on the emitter tip. This increase opens more emission sites and increases thrust at a
cost. Running the thrusters at a high current output, by raising the extractor potential voltage,
increases the overall thrust generated by increasing the number of Taylor cones per thruster tip
and, as a result, can increase ion impingent on the extractor the same way a clipped or short tip
increases impingement.
In addition to the above problem, the extractor is also the limiting factor in scalability;
micromachined holes need to be precisely drilled in exotic materials to create these ionic
windows. The flatness of the extractor is crucial to the operation of the thruster. Currently, the

7
limit to the size of a single thruster face is about 2 square centimeters due to extractor deflection.
If the extractor deflects by even a fraction of a millimeter over its face, the thruster loses
performance. This loss of performance is due to the above depiction of non-uniform emitter
heights extrapolated to the entire thruster face (Figure 3). To reduce the interception problem, the
extractors must be extremely thin (approximately .003” or the thickness of a sheet of standard
paper). Any changes in temperature or vibration can warp the extractor at this thickness, causing
flatness issues and hurting electrospray performance. To mitigate these problems, the current
state of the art thrusters combine many smaller thruster chips into one big thruster. While this
reduces the warping and extractor deformation, it increases complexity, weight, and inert mass.
This thesis will seek to solve the problems that arise as a consequence of practical
engineering and materiel property constraints, impacting the extractor by eliminating it. Suppose
the ions are free to escape the reservoir and emitter tips without minimal conical angle
restrictions from a physical extraction plane. In this case the thruster’s lifetime and scalability
issues can be solved, and electrospray thrusters can move from small satellite propulsion into
larger and more practical applications like their larger cousin, the conventional ion engine.
1.3 Outline and approach
This thesis will consist of five main parts. Chapter 2 will discuss the current state of
industry and academia with an in-depth literature review. Chapter 3 will lay out the diagnostic
requirements and data reduction necessary to parameterize and quantify the performance of a
PIR electrospray thruster. Chapter 4 will go into the detailed construction and testing of a state of
the industry Pure Ion Regime USC Testbed Thruster (UTT). Chapter 5 will outline the initial
construction and testing of a novel electrospray device that induces electrospray emission

8
through an extractor less process driven by dynamic magnetic induction. Chapter 6 will present
results and future considerations for the field of electrospray technology.
The USC Testbed Thruster (UTT) depicted in Figure 6 is a passively fed electrospray
thruster utilizing a porous borosilicate reservoir and emitter layer. It is designed to utilize 1-
ethyl-3-methylimidazolium-tetrafluoroborate (EMI-BF4) propellant and operate in the Pure Ionic
Regime (PIR). The USC electrospray thruster program started in 2018 with the design of the US
Air Force Research Laboratory’s (AFRL) Air Force Electrospray Thruster II (AFET-II)
developed by Mike Natisin

[8].

The AFET-II thruster was based on Dan Courtney’s novel
borosilicate

thruster from 2013 [9]–[11]. The UTT is a simplified 50% scale AFRL AFET-II,
modified to have only 25 emitters on a 1 cm diameter, interchangeable borosilicate emitter disk.

Figure 6: USC Testbed Thruster CAD drawing

The downsizing of the thruster was done in an attempt to fit the thruster in the USC
plasma laboratory’s 940-liter test chamber, without saturating the chamber’s diagnostic and
pumping capabilities. At the current pitch, the extrapolated thruster tip density is 345
emitters/cm^2. The emitter array and extractor are the most critical components in the UTT.
They have the tightest tolerance requirements of 25 μm. Any misalignment or discrepancies in

9
manufacturing lead to increased particle impingement and a loss of efficiency. Figure 7 shows
the UTT and its relative scale.

Figure 7: As-built assembled USC Testbed Thruster (UTT)

Four UTT baseline thrusters have been built, tested, and evaluated. This data allows for
the comparison of performance data with a new gridless thruster. This new Magnetically Induced
Electrospray Thruster (MIET) will be built, tested, and its performance relative to the state of the
industry will be quantified and presented for review. The USC Magnetic Test Thruster (MTT),
Figure 8, is the first iteration of a proposed technological advancement in Electrospray Thrusters.

Figure 8: MTT CAD rendering and wire wrap of the first coil pair

This proposed technology eliminates the need for the extractor allowing the emitted ions
to escape directly and produce thrust without any interception and loss. With the removal of the
extractor, this technology can be scaled up as a function of how many emitters can be placed on

10
a surface, not by how many holes can be consistently drilled in a flat extractor. Similar to the
conventional electrostatically propelled thrusters, the physics does not limit the size of the
thruster that can be created using this technology. Large thrusters with billions of emitters would
be able to operate for many thousands of hours. A side benefit of this technology is its inherent
redundancy. Conventional thrusters die when one emitter tip (out of hundreds/thousands) is
shorted to the extractor. With this technology, any emitter that fails is just lost and no longer
contributes to thrust, but it does not short the whole thruster out. Off-axis emission may still hit
the thruster’s body or Helmholtz coils. The thruster body could be designed with this limitation
in mind, because, unlike the extractor, it does not have to be unpractically thin.

11
CHAPTER 2: LITERATURE REVIEW

2.1 Electrostatic Electrospray
While each technological subset of electrospray devices shares a standard electrostatic
acceleration, each device took a distinct evolutionary path. Colloid Spray, Field Emission
Electric Propulsion (FEEP), Liquid Metal Ion Sources (LMIS), Pure Ion Regime (PIR), and Ionic
Liquid Ion Sources (ILIS) all have their evolutionary process that must be investigated to
understand their distinct differences.
2.1.1 Colloid Thrusters

Colloid Thrusters are the oldest and most studied branch of electrospray
thrusters. The utility in the field of thrust began even before Taylor quantified the
angle at which emission occurs. Taylor wrote his paper on “Electrically Driven
Jets” in 1969 [4]. Before this, the US Air Force was investigating arrays of 10’s
of capillary tube emission. Cohen et al. proved the functionality of ‘charged
particle electrostatic thrusters’ in 1966 [12][13].
Fundamentally a colloid thruster uses a small capillary and an electrically
conductive fluid to create a spray of charged droplets. Unlike pure ion emission,
colloid thrusters emit collections of high mass molecules that are singly or doubly
charged [14]. These droplets have very high masses when compared to their
purely ionic counterparts [15] and while this leads to increased thrust potential, it
drives up the operating voltage. The fluid used in colloid thrusters can be any
conducting fluid from water to sulfuric acid to exotic dissolved ionic liquids [16].
In the early days of investigation, the use of high molecular mass solvents led to

12
the eventual shelving of the technology due to the requirement for 10s-100s of kV
to extract a useful amount of propulsion [17] . The power supplies and insulation
proved too heavy for the small thrust generated by these thrusters to find utility in
the 1960’s technology. Additionally, colloid capillary tube emitters require
minuscule pumps that finely regulate the flow required to emit ions stably. If this
flow rate is not tuned to extractor voltage and emission characteristics, then the
thruster will not operate or leak [18].
The requirement for constant stabilization at low thrust levels led to a
resurgence of funding for colloid thrusters in the early 2000s, and the NASA
Laser Interferometer Space Antenna (LISA) Pathfinder flew a set of Busek Space
Systems’ Colloid thrusters to prove the technology in 2015-2017 [19]–[22]. While
there have been many grounded studied of colloid thrusters, this is the only flight
heritage that they have. This funding and the subsequent follow up LISA mission
have led to a wide array of papers and investigations into the charging and
emission characteristics of colloid thrusters [14], [19], [23]–[25].
2.1.2 FEEP or LMIS Thrusters

Field Emission Electric Propulsion and its synonymous Liquid Metal Ion
Source propulsion have seen a large boom since 2002. This branch of electrospray
thrusters has the most flight heritage to date, with over 17 thrusters flown as of
2019. Due to the use of Liquid metal as the ionic fluid, FEEP/LMIS thrusters
must use an external neutralizer to overcome significant spacecraft charging [26].
Often molten Indium or Cesium are used as a propellant, and as such, the thrusters
only release positive ions [27]. These thrusters can be actively fed or passively fed

13
with most of the flight heritage coming from passively fed porous emitters
operating in the Pure Ion Regime [28].
In support of small satellite propulsion and the ESA LISA mission, Tajmar
et al. [28]–[31] and the Austrian Institute of Technology, developed an Indium
based FEEP thruster that has since found a commercial home at Enpulsion [32]
under the name IFM Nano. Biagioni et al. in Italy are continuing development on
a Cesium-based electrospray thruster for use in small satellite propulsion [33].
Enpulsion has flown 17 IFM Nano thrusters that are pure ion emission
FEEP thrusters [34] [35]. They use Indium as a propellant and only release
positive ions. ENPULSION uses a novel approach to electrospray with their IFM
Nano FEEP thruster. Based on the FOTEC design for ESA, the “crown emitter
does not have a classical extractor [30]. It uses a toroid of discrete emitters
arranged in a crown with an extractor on the outside. This is similar to the linear
emitters, but each discrete emitter anchors emission, thus reducing loss.
ENPULSION has demonstrated 6000+ hours of firing time on this design using
melted Indium propellant. One drawback to this flight-proven design is that the
toroid has no inner extractor plate and thus suffers from upwards of a 60-degree
plume half angle. While this thruster does not suffer lifetime issues in the form of
failure, it is wasting a lot of propellant to off-axis thrust [32].
2.1.3 PIR or ILIS Thrusters

While it was important to discuss Colloid Thrusters and FEEP/LMIS
because of their heritage, thrusters operating in the Pure Ionic Regime make up
the branch of electrospray thrusters with the most potential, and this thesis will

14
look in-depth into this field of electrospray. Chapter 4 discusses the creation and
testing of 4 thrusters’ which operate in the PIR with passively fed porous media.
The Explosion into the research of PIR thrusters can be traced back to the
Ramero-Sanz et al. discovery of the utility for 1-ethyl-3-methyl imidazolium
tetrafluoroborate (EMI
+
BF4
-
) [5] to be used as a propellant in what started as a
colloid thruster investigation by Velisquez-Garcia in 2001 and quickly evolved
into investigations into externally wetted and passively fed PIR thrusters [36].
a. Passively fed externally wetted
The utility of Ionic liquids for electrospray thrusters moved from capillary
fed to externally wetted passive feed pure ionic thrusters because of the
characteristics of ionic liquids [16],[37]. These fluids are stable in vacuum
due to low vapor pressures and are viscus enough to wick up a surface or
through pores due to Laplace pressure [38]. In 2004-2006 Velisquez-
Garcia et al. developed a planar array of externally fed emitters on silicon
chips [39],[40]. In 2007 Gassend et al. integrated extractor grids with 500
emitters [41],[42]. This design led to the investigation into porous media
PIR thrusters because of inconsistent operation due to inconsistent wetting
of the emitters. With no way to actively control this process and the
impracticality of using this on a spacecraft, this technological subset has
fallen out of favor.
b. Passive fed porous media
Figure 9 is a Scanning Electron Microscope (SEM) image of a passively
fed emitter tip microfabricated in borosilicate glass with 50% void space

15
available for propellant. Multiple openings are visible at the tip of the
emitter.

Figure 9: An SEM image of the UTT emitter substrate

The radius of curvature of the tips can be artificially influenced by
mechanically forming emitter tips out of this substrate or another porous
substrate with a porous interface layer [7]. Emission, just like for the other
types of electrospray, occurs when the electric fields between the voids
that are holding propellant in the emitter tips and the extractor grids reach
a critical point and a Taylor Cone forms. This action releases a stream of
ions, and in the case of passively fed porous media emitters, it has been
noted that multiple emission sites can open up on a single tip causing off-
axis emission [43]. Pure Ionic Regime thrusters accelerate single monomer
or dimer molecules.

16
As opposed to externally wetted emitters the propellant, in most
cases an Ionic Liquid, is held in solution in the pores of the substrate
instead of on the surface. This allows for the wicking of propellant
through a material as opposed to around or up the length of it. Dan
Cortney and Chase Coffman investigated the use of porous substrates in
2009-2011[9],[44]. Metal substrates such as porous tungsten were
investigated, initially, due to their heritage with FEEP thrusters [44],[45].
Dan Courtney found that electrochemical etching over the life of the
thruster was found to dull any conductive emitter tip [44]. Porous
borosilicate has been the main focus of research in passively fed porous
media emitters. Busek [46] [47], Accion [9] [48], AFRL [49][50], and the
university of South Hampton [51] all use Porous borosilicate glass
emitters [22], [8]. For this reason and its availability, USC has chosen to
utilize porous borosilicate disks in the first iteration of the Testbed
Thruster.
Recent research into the field has stagnated on the tried-and-true
approach of centering fabricated emitters in an external extractor. Accion
has moved to using liquid reservoirs and an array of laser etched porous
media emitters [48]. The pivot to a liquid reservoir has increased
propellant volume by up to 50% but has introduced problems in the
storage of the emitters in a flight configuration. Propellant has wicked out
of the emitter when in a horizontal launch configuration or in the vibration
environment of launch [6], [43], [52]. These thrusters show great promise

17
in the micro propulsion realm but are still limited in size by the extractor
to emitter deformation and centering problems [52][53].
The discovery of new ionic liquids is an ongoing endeavor. The
use of EMI-BF4 for electrosprays seems to be a lucky happenstance.
There are other propellants that are currently being investigated like the
Green monopropellant [54][55][56] that the US AF is developing with
partners in industry. New studies on mixing ionic liquids [57]–[59] has
shown promises for creating custom properties that may be more suitable
to electrospray. The physical properties of ionic liquids [37], [60] are
being investigated as well as their evaporation [61] characteristics. Natbity
et al. is investigating the Effect of IL composition [14] on electrospray
performance. While all this work is ongoing there is great potential for
further research and a detailed study of electrospray performance on
various IL compositions.
c. PIR Extractor Failure Modes
PIR electrospray extraction shares common characteristics with,
the close electric propulsion counterpart, conventional ion thrusters. The
extraction of an individual ES emitter tip through an ion window looks
like a micro scale of an ion thruster performance through an acceleration
grid. While the ion generation may be different between noble gas and
ionic liquids the ultimate extraction and acceleration of the ions is similar.
As a result, they share some of the same failure modes. As extensive
research into the life of Ion thrusters has been performed [62]–[68] plenty

18
of knowledge can be gleaned on the performance of ES extractors. High
energy ion impingement or sputtering on the acceleration grid of ion
thrusters has been shown to wear on the grids and cause ion back
streaming. Figure 10 shows the wear and the potential for structural grid
failure. To combat these issues ion thruster grids have been constructed
with wear in mind and are thick enough for the thrusters life [65], [69],
[70].

Figure 10: Ion wear pattern and grid deterioration due to ion sputtering on the
acceleration grid [70]
While most PIR IL ES thrusters firing times have been in the 100’s
of hours the IFM nano firing times are approaching that of flight proven
ion thrusters [35], [48]. Similarities can be found in the pitting and
deterioration of ion thruster accelerator grids and the electrospray
extractor failure modes [24], [71], [72]. The properties and fluidic nature
of ionic liquids pose additional concerns that ion thrusters do not have to
contend with. The addition of liquid bridging to grid deterioration adds an
additional failure mode to ES thrusters that currently drastically shortens
their possible life. Extensive work has been done to reduce ES extractor

19
impingement, this work is ongoing and this thesis will show an alternative
ion extraction mode that eliminates the need for an ES extractor.
2.2 Magnetically Induced Ion Mobility
The magnetically induced extractor-less electrospray does not have a documented
evolution, as it is a novel idea, but the underlying magnetic field generation technology, The
Helmholtz coil, has been used in the medical field and the physical investigation of magnetic
field theory. This coil configuration was chosen because multiple pairs can be used to induce
magnetic field variations in different planes. The ultimate MTT thruster uses two orthogonal coil
pairs for thrust vector control. The Helmholtz coils that drive the high-frequency oscillation
necessary to induce an electric field in the propellant disc are a well understood and studied
technology [73]–[76]. Theorized by Herman von Helmholtz in the late 1800s the coils are a
simple N turn symmetric winding of two coils spaced one radius apart, that generate a uniform
magnetic field in-between the coils (Figure 11). These coils are commonly used with constant
currant to generate static magnetic fields or at high frequency operation to induce biomedical
response patterns in neurons and the circulatory system [77][78].

Figure 11: Helmholtz coil pair with driver schematic

20
If properly spaced the magnetic field strength in the center of the coils will be uniform
and can be found can be found using Eq. 4.
𝐵𝐵 = �
4
5
�
3
2
�
𝜇𝜇 0
𝑁𝑁 𝑁𝑁 𝑟𝑟 � Eq. 4
Conversely the current needed to drive a magnetic field of particular magnitude, in the center of
the coil, can be found using Eq. 5.
𝐼𝐼 = �
5
4
�
3
2
�
𝐵𝐵 ∗ 𝑟𝑟 𝜇𝜇 0
𝑁𝑁 � Eq. 5
Where I is the peak to peak coil current (A). r is coil radius and spacing between the two
coils. N is the number of turns in each coil. 𝜇𝜇 0
= the permeability constant. ω is the angular
frequency or ω = 2πf. L1 and L2 are the inductances of the two Helmholtz coils, and R1 and R2 are
the resistances of the two coils. Drive voltage can be found using the expanded form of Ohms
law Eq. 6 and the inductance in each coil, which is a function of the number of turns (N), the coil
diameter (D), wire diameter (d) and permeability of the wire ( 𝜇𝜇 𝑟𝑟 ) Eq. 7.
𝑉𝑉 = 𝐼𝐼 �[ 𝜛𝜛 ( 𝐿𝐿 1
+ 𝐿𝐿 2
)]
2
+ ( 𝑅𝑅 1
+ 𝑅𝑅 2
) Eq. 6
𝐿𝐿 𝑖𝑖 = 𝑁𝑁 2
𝜇𝜇 0
𝜇𝜇 𝑟𝑟 𝐷𝐷 2
� 𝑙𝑙𝑙𝑙 �
8 ∗ 𝐷𝐷 𝑑𝑑 � � − 2 Eq. 7
When generating a dynamic magnetic field, the inductance of the coil must be taken into
account. The dynamic magnetic field induces current in the adjacent wires and drives voltage
proportional to the frequency. Inductance drives the required driver voltage as a function of
frequency and can quickly overcome the voltage required due to wire resistance alone. This
problem can be overcome with a series resonant circuit. A capacitor of capacitance found in Eq.
8 needs to be added into the coil driver circuit in series [79]. The frequency of oscillation drives

21
𝐶𝐶 =
1
( 2 𝜋𝜋𝜋𝜋 )
2
( 𝐿𝐿 1
+ 𝐿𝐿 2
)
, 𝑓𝑓 0
=
1
2 𝜋𝜋 �( 𝐿𝐿 1
+ 𝐿𝐿 2
) 𝐶𝐶 Eq. 8
the capacitance required and as the frequency changes so does the resonant capacitance required
to drive the excitation voltage down to the static resistance of the wires. This drive voltage can
be reduced by matching the inductance to a capacitance value and using a series capacitor as
illustrated in Figure 11.

22
CHAPTER 3: METHODOLOGY

The USC Laboratory of Astronautical Plasma Dynamics (LAPD) is a plasma research
laboratory built around a 940-liter vacuum chamber Figure 12. Historically the laboratory has
utilized this chamber alongside a mesothermal plasma source to investigate solar wind dynamics,
spacecraft charging/interactions, and lunar dust charging/transport [80]–[82]. The chamber
operates primarily on a CVI TM-500 helium cryopump with a floor vacuum pressure of 1x10
-7

Torr. In late 2018 the focus of LAPD expanded to include electrospray thruster development and
testing, the pumping capacity of the chamber was doubled, and the diagnostics were overhauled
to enable the collection of data relevant to electrospray performance specifically.

Figure 12: LAPD primary vacuum chamber
This overhaul maintained the ability to perform the traditional operations with the plasma
source while enabling the expansion of research into the ever-growing field of ILIS electrospray
technology. An Alcatel roughing pump seen in the bottom right of Figure 12 will take the
chamber to 20 mTorr. A newly added T-1000 turbopump can maintain a chamber pressure of
less than 1x10
-6
Torr for extended periods and during thruster operations. If necessary, a gate
valve, blocking the cryopump, can be opened to pull the chamber down to 2x10
-7
for

23
approximately 4 hours. This pumping capacity has proven adequate and stable during all
electrospray testing.
3.1 Laboratory and Facility Equipment
In addition to the doubling of pumping capability, a new probe suit was explicitly
developed for the electrospray program. A Retarding Potential Analyzer (RPA), Faraday Probe
(FP), and Micro Faraday Probe (MFP) were all explicitly developed for spatial analysis of a non-
neutral electrospray plume. Figure 13 shows the probe suite in its final configuration. A
Langmuir probe is included in the suite as a legacy diagnostic device for use with the quasi-
neutral plasma source. This probe suit is mounted on a 3-axis traversing system that can scan the
full volume of the inside of the test chamber.

Figure 13: Probe setup inside LAPD main chamber
left to right on three axes spatial mount: phosphor screen, Retarding Potential Analyzer (RPA), Micro
Faraday Probe (MFP), Langmuir Probe (LP), Faraday Probe (FP). Back of chamber 6” Large Collection Plate
(CP)
All probes can be utilized simultaneously; the MFP is situated at the bottom of the suite
in front of the other probes. This was done in an effort to be able to get the MFP as close to the

24
USC Testbed Thruster, during operation, as possible. The MFP was developed with an inner cup
diameter of 1.28 mm. This small-diameter inner cup, coupled with close standoff distances,
allows the MFP to resolve the current measurements of individual emission sites. The collection
surfaces of the entire instrument suite can be biased to reduce secondary electron emission within
individual collection devices.
3.1.1 Three-Axis Probe System and LabVIEW Control

The three-axis probe system seen in Figure 14 was an upgrade of John
Polanski’s original two-axis system [80] developed for his 2013 thesis.

Figure 14: Tri-axis data traversing system

The new three-axis traversing system was rebuilt from the ground up with
new electronics and control interfaces. The system is controlled with a LabVIEW
interface that runs an Arduino Mega and a Ramps 1.4 stepper board, Figure 15.

25

Figure 15: Arduino MEGA with RAMPS 1.4 stepper driver for three axis control and three I2C 0-
5V control chips.
This system can be programmed to run sweep patterns and routes independent of
the user interface. This system leverages modern additive manufacturing control
and flexibility while allowing for sub-millimeter movement and control of the
system. All three axis have independent control and can be rehomed without
breaking the vacuum to increase accuracy and measurement flexibility between
runs. The same LabVIEW traversing control runs all probe voltage sweeps, as
well as the thruster voltage controller. Figure 16 shows the LabVIEW control
interface.

Figure 16: LAPD LabVIEW user interface.

The above interface provides control and recording of
thruster/RPA voltage settings as well as probe positions. The probe traversing
system positions are recorded continually to give an accurate representation of

26
where each measurement was taken and by what probe. Limit switches were
installed on each axis of the traversing system so that the probes can be homed to
re-zero the probe suite without breaking the vacuum in the event of a stepping
error. Preset scan profiles can be run for a 3-dimensional plume scan, a BNG 2
axis plume scan, or an MFP close thruster face two-axis scan. Thruster voltage
and RPA voltage can be manually controlled or stepped up and down via a preset
profile. This code has automated almost all of the daily thruster operations in the
LAPD. The LabVIEW interface for Arduino had been tested on this scale
multiple times without success [83][84]. Custom scripts had to be written for the
limit switch functionality and LabVIEW base code had to be changed to add
Arduino MEGA functionality over the inherent Uno interface. Pin outs had to be
mapped between LabVIEW, the Arduino Mega, the Ramps 1.4 board and the
probe traversing system. I2C capability had to be added to the LabVIEW Arduino
interface to control the thruster and RPA power supplies. The RPA power supply
requires a 0to-5V command signal and so the I2C 0-5V signal had to be inverted
using a MCIGICM dip lm358 operational amplifier and a 10-volt bipolar power
supply seen in Figure 17.

Figure 17: Voltage inverter for command signal of RPA

27
Lab View block diagrams for the thruster control, probe positioning, and
data acquisition can be seen in Figure 18 below. Once running, the stepper motor
and I2C communication to the Arduino are set up and the while loop for user

Figure 18: LabVIEW block diagram for thruster operation and probe positioning.
inputs waits for command input in the front panel (Figure 16). This iteration of
the code has 15 conditional control loops to control everything, a 16
th
control loop
was developed to run automated BNG plume scans (Figure 19).

Figure 19: Conditional loops for user interface
LabVIEW block diagrams for the limit switches and stepper motor sub VI
initialization can be seen in Figure 20. Each of these sub-VIs was replicated for
the Y and Z axes according to the parameters in Table 1. These initializations

28
were not done properly in other attempts to interface LabVIEW with Arduino.
This caused a reoccurring problem that when the VI was run these values would

Figure 20: Limit switch (left) and stepper motor (right) initialization for Arduino LabVIEW interface

Table 1: Arduino RAMPS 1.4 stepper control pins
revert to initial values that would not communicate. Additionally, LabVIEW code
had to be changed in the LabVIEW Interface For Arduino (LIFA) toolkit to allow
LabVIEW to talk to pins higher than 60 on the RAMPS board.

3.1.2 Faraday Probe/Micro Faraday Probe

Figure 21: Faraday probe/micro-Faraday probe and electrical schematic

29
Two Faraday Cup (FC) type probes were created for the express purpose
of diagnostic testing of the electrospray thrusters in LAPD (Figure 21). The
Faraday cup was chosen over a Faraday Plate (FP) with a guard ring because the
FC responds better under high energy ion bombardment than a 2D Faraday plate
[85]. A higher fraction of secondary electron losses is recaptured by the
conducting cup. Off perpendicular ions that make it through the plane of the cup
opening also get collected [86]. The larger FC is a stainless-steel cup with a
graphite collection plate at the end. It has a diameter of 9.57mm and a length of
35mm. The probe is surrounded by a stainless-steel guard tube that can be
grounded or biased to -40 V to suppress electron collection and charging in the
plasma environment. The interface between the two layers is dielectrically
shielded with a Teflon sheet. The Micro Faraday Probe (MFP) follows the same
construction with a cup diameter of 1.59mm and a shield diameter of 3.02mm.
The Faraday probes are swept horizontally and vertically across the electrospray
plume to capture current and visualize plume angle divergence. The intercepted
current is then divided by the probe’s circular facial area to calculate the current
density of the plume. The size of the MFP was chosen so that at a distance of .5-
.75 inches, the emission of single emitter tips on the UTT could be resolved from
the probe area. Figure 22 shows a thruster face scan of the MFP at .75 inches
away from the thruster. Individual emitter tip emission can be resolved, and hot
spots can be seen where more efficient emission is visualized.
𝑣𝑣 𝑖𝑖 = �
2 𝑒𝑒 (𝜙𝜙 0
− 𝜙𝜙 𝑝𝑝 )
𝑚𝑚 𝑖𝑖 Eq. 9

𝑙𝑙 𝑖𝑖 =
𝐽𝐽 𝑖𝑖 𝑒𝑒𝑒𝑒 𝑖𝑖 Eq. 10

30

Figure 22: Heat map of MFP with .75” Dis. to thruster face.
Thruster extractor position in the background is relative to the scanning probe location and scaled for
distance

3.1.3 Langmuir Probe

Figure 23: Langmuir probe and electrical schematic

The Langmuir probe is a relic of the lab’s heritage as a quasi-neutral
plasma physics laboratory. Since the 1960’s the Langmuir probe has been one of
the most extensively used diagnostics tools in plasma physics[87]. A basic
Langmuir probe is a bare wire immersed in a plasma. Through probe theory, one
can obtain the current-voltage (I-V) characteristics as an applied bias voltage for
the probe is swept from a low to a high potential. The measured I-V curve can
then be used to determine the electron temperature, Te, and number density, ne of
the immersed plasma. The characteristic length of the heritage cylindrical
Langmuir probe used in this study was 6.35 mm, with a probe radius of 0.5 mm.

31
Primarily the Langmuir probe was used as a baseline to make sure the other
biased probes were working. Since the electrospray thruster does not produce a
neutral plasma, the Langmuir probe was not used conventionally. In the event,
two thrusters are used to produce a quasi-neutral ion field, the probe will be
biased from -400 V to +600 V, and the voltage drop across a resistor will be
recorded over the voltage sweep, as shown in Figure 23. By taking the natural log
of the recorded curve, Te in eV is calculated from the slope of the “Te line” by Eq.
(11), as shown in Figure 24. Then, ne is calculated from Eq. (12), where Ie;sat is the
electron saturation current, and Aprobe is the probe collecting surface area. The
plasma potential, Φ p, can also be taken from the Langmuir probe; In a flowing
plasma, such as the ion beam from an electrospray, the potential measurement can
be skewed, requiring the use of an emissive probe for better accuracy.
𝑇𝑇 𝑒𝑒 =
1
𝑠𝑠 𝑠𝑠𝑠𝑠 𝑠𝑠 𝑒𝑒 ( 𝑇𝑇 𝑒𝑒 𝑠𝑠𝑖𝑖 𝑙𝑙 𝑒𝑒 )
Eq. 11
𝑙𝑙 𝑒𝑒 =
𝑁𝑁 𝑒𝑒 ,𝑠𝑠 𝑠𝑠𝑠𝑠 𝑒𝑒 𝐴𝐴 𝑝𝑝𝑝𝑝𝑝𝑝 𝑝𝑝 𝑒𝑒 �
2 𝜋𝜋 𝑚𝑚 𝑒𝑒 𝑘𝑘 𝑇𝑇 𝑒𝑒 Eq. 12

Figure 24: An example of an LP measurement and analysis [81]

3.1.4 Retarding Potential Analyzer

32

Figure 25: Retarding potential analyzer and electrical schematic

The Retarding Potential Analyzer (RPA) was created to perform where the
LP could not (Figure 25). The collection component of the RPA is a Faraday cup
collector with a stainless-steel cup and a graphite plate at the end. The probe has
the same collection diameter of 9.57mm and a length of 35mm as the Faraday
probe. The probe is surrounded by a stainless-steel guard tube that can be
grounded or biased to -40 V to suppress electron collection and charging in the
plasma environment. The interface between the two layers is dielectrically
shielded with a Teflon sheet. 3 Grids are aligned and placed in front of the
collector cup to manipulate the flow of ions into the collector. Each mesh grid has
a 78% transparency. This was chosen because even if all three stacked grids were
entirely miss-aligned, they would still have 47% transparency. The three grids
consist of a floating grid, a sweep voltage grid, and an electron suppression grid.
The sweep voltage screen is the most important part of the RPA. The sweep
screen allows for high voltage (-5kV-5kV) bias. This bias is significant enough to
retard the flow of ions emitted from the electrospray thruster. The floating grid is
placed in the front of the RPA (in the direction of flow) so that the incoming ions
are not diverted or accelerated by the high potential of the sweeping screen. This
floating screen allows for stable measurements of ions that do pass through all

33
three screens to the collection plate tube and plate. The last screen is the electron
suppression screen; this screen is after the high voltage sweep screen and returns
secondary electron emission that can be emitted when high energy ions impact the
collector plate [88].
The RPA allows for the beam potential to be resolved and graphically
illustrates the beam potential being emitted from the electrospray thruster. IV
curves can be established similar to the LP. Figure 26 is a depiction of an IV
curve from a UTT test run with the thruster and RPA running in negative mode.
As RPA voltage is increased intercepted current is decreased until no current can
enter the RPA.

Figure 26: RPA intercepted current vs Voltage applied to sweep grid

34
3.1.5 Faraday Plate

Figure 27: 6” large collection plate and electrical schematic

Figure 27 is a depiction and electrical schematic for the large Faraday
plate at the rear of the chamber. This plate is used to collect larger portions of the
electrospray plume than the other diagnostic devices in the chamber. The plate is
a 6” x6” receiving plate with a 5” x5” mesh window. This mesh can be left
floating, grounded, or biased to -40V to suppress secondary electron emission.
The size of the plate and its distance from the thruster face allow for a large
enough signal collection at a significant enough distance (30”) to resolve Time of
Flight (TOF) spectroscopy. The plate has two mounting positions to enable
readings when the cryopump door is in the open (up) position or the closed
(down) position.
3.1.6 High-speed TOF Diagnostics

Time of Flight (TOF) spectroscopy is the leading tool in empirically
determining electrospray thruster ISP [8], [10], [89], [90]. Four main components,
a gate, a collector plate, an operational current amplifier, and a high-frequency
digital oscilloscope are needed to get a reliable TOF signal. This TOF signal can
then be combined with emitted current to derive a mass spectrum of an emitted

35
electrospray plume. The variability in emitted ion mass spectra by polarity and
excitation voltage leads to the need for a reliable, repeatable, non-destructive TOF
measurement. This measurement must be taken often throughout a testing
campaign. It is imperative that any diagnostic tool used to determine the ejected
mass spectra does not create undue noise or wear on the diagnostic tools or
thruster itself. TOF diagnostics require a gate to either reflect or deflect the ions
so that the intercepted current at the collector plate can be controlled reliably.
Two main paths are to use high voltage across a mesh to reflect the ions or use
lower voltage across parallel wires to deflect the ions [91].
Standard reflection gates used for TOF generate high voltage noise that is
caused by the high-frequency cycling required to generate the averaged TOF
signal. Often upwards of 50 HV pulse cycles are needed to time average the TOF
signal and filter induced variation and noise. This cycling creates induction in all
other probes and raises the ground voltage temporarily, creating a high floor of
high-frequency noise that must be post-processed. In addition to the unwanted
noise, the reflection of high energy ions (1-3keV) back on the firing thruster
creates undue wear and accelerated aging of the thruster. This accelerated aging
process is visible when using reflection diagnostics in high energy arcing and
discoloration of the thruster face over a very short period of reflection. To combat
these two detrimental problems, a Bradbury Nielsen Gate (BNG) deflection gate
will be used in combination with the other components required for TOF
spectroscopy (Figure 28).

36

Figure 28: BNG Test setup, chamber geometry, where LBC = 30 in., HT = 0.106 in., LTB = 0.5 in., and HFP
= 5.9 in. Simplified ion beams are also depicted.

The BNG does not reflect on a firing thruster but instead temporarily
deflects them away from the downstream collector plate. This deflection uses
significantly less voltage, and therefore less noise is induced when the gate is
cycled multiple times at high frequency. Figure 29 illustrates the BNG deflection
of a thruster operating with a 1400V Extraction and a plume half angle of 12°.
Given these constraints, the BNG must operate at a 15° deflection to clear the 4”
downstream collection plate at a distance of 26” from the BNG. A script was
written to calculate the required deflection voltage for the BNG based on the test
set up and individual thruster performance parameters to include operation
voltage and plume half angle. Appendix B goes into the detail of the construction
and testing of the BNG.

37

Figure 29: BNG projected deflection with V
BNG
=1020V and a 12° thruster half-angle at T
thruster
=1400V

Conventional TOF gates use high voltage pulses across a central grid
placed in front of a thruster Figure 30 to reflect the electrospray plume and
temporarily halt downstream collection of the signal. To reduce unintended
influence on the ions by the high energy plane a ground plane is placed in front of
and behind the reflection gate. The stacking of these 3 grids can greatly reduce the
transparency of the gate to the thruster plume. For our case the stacking reduced
transparency to 12% when two 52% grounding grids and a 45% transparent
reflection grid were stacked. The transparency was increased to 30% when two
74%grounding grids sandwiched a 52% reflection grid. For a conventional gate
to “close,” the high voltage grid needs to be held above the thruster excitation
voltage to retard and reflect the incoming ions. This High Voltage reflection leads

38

Figure 30: reflection gate test setup, chamber geometry, where LBC = 28 in., HT = 0.106 in., LTB = 2 in.,
and HFP = 5.9 in. Simplified ion beams are also depicted.

to unintended noise and wears on the thrusters under diagnostic test [11], [21],
[50], [92] . High voltage noise is induced in data collection devices by the high-
frequency cycling required to generate the averaged TOF signal. Often upwards
of 50 HV pulse cycles in the range of 15-20 kHz are needed to time average the
TOF signal and filter induced variation and noise. This cycling creates induced
noise in all other probes and raises the ground voltage temporarily, creating a high
floor of high-frequency noise that must be post-processed. In addition to the
unwanted noise, the reflection of high energy ions (1-3keV) back on the firing
thruster creates undue wear and accelerated aging. This accelerated aging process
is visible when using reflection gate testing, in the form of high energy arcing
visible on the thruster extractor during reflection as well as discoloration of the
thruster face observed after short TOF testing durations [26], [44], [92], [93].

39
Additionally, thruster extractor current readings rise as a result of returned current
impacting the thruster face.
A large format BNG tailored for electrospray operation is a less
destructive, more passive time of flight spectroscopy ion diverting device than the
current academic standard of using a conventional ion repulsion gate. Thruster
plume free stream ion transparency through the BNG gate was increased from 12-
30% in 3 grid repulsion gates to 75% with the BNG. While total BNG ion
diversion efficiency at the TOF collection plane is 66-93% of the total beam as
compared to 99% with a conventional gate, the increased current collected due to
the improved transparency of the BNG overcomes this limitation. If during initial
construction, wire tension and expected thermal cycling are properly matched, the
BNG shows long term utility and parallelization stability, after being vacuum
cycled hundreds of times in a standard lab environment. Returning current to the
thrusters emitting surface was eliminated by the use of the BNG. As return current
can damage the thruster emitters and extractor by causing shorts or high energy
arc discharges, this advantage will increase long term measurement accuracy as
well as thruster lifetime during investigation. The virtual elimination of return
current, reflected back on the operational thrusters, by the BNG should make this
device the industry standard for TOF-S of electrospray thrusters.
3.2 Data Acquisition and Reduction
Voltage and current data from the thruster’s positive and negative power supplies are fed
along with shunted current data from the RPA, LP, CP, MFP, and FP to an Agilent 34970A 20
channel DAQ. The data, along with -40V bias voltage and RPA high voltage bias states, are all

40
measured at a 1 Hz data sample rate. This data is combined with the LabVIEW position data
collected from the control code for post-processing. High-speed data is collected from an NI
virtual bench with a maximum of 1 GHz sample rate. This high-speed data is primarily utilized
for the TOF system but can be used with any single probe to investigate thruster transient effects.
The UTT extractor is grounded through a shunt resistor that allows for the measurement of
extracted current. This, in conjunction with the current data from the power supplies, gives the
thruster percent fraction of beam interception. This %int is the leading diagnostic tool for
understanding thruster health and instantaneous performance.
Utilizing performance equations in Appendix A, thruster key performance parameters are
calculated based on TOF derived ion mass distribution, power source acceleration voltage, and
probe confirmed plume divergence. All of the probes on the special traversing system measure
plume current density with respect to spatial location. This data is valuable for diagnostic
information and tuning of each thruster. Since large plume half angles decrease axial thrust and
contribute to life-limiting thruster extractor degradation, a lot of effort is devoted to each
individualized thruster’s plume characteristics over their entire life. Each thruster’s overall
performance is measured and compared to baseline results from previous thrusters and literature
in chapter 4.
3.2.1 Noise

Noise has been a significant problem in the development and testing of
both the UTT and the MTT. Chamber wall effects due to the relatively high
power of the thruster and small nature of the test cell led to secondary anodes as
well as plume characterization challenges. A significant amount of laboratory
time was spent chasing down sources of induced noise. When the measured probe

41
current for the expected signals is in the range of 10s-100s of nanoamps small
noise sources can overwhelm the desired signal.
Possible noise constituents that were investigated were: the traverse
system stepper motors, the turbo pump, the diffusion pump, ion gages, thruster
power supplies, chamber wall effects as well as high energy ion bombardment
[94]. The Operational amplifiers that were used for TOF signal were particularly
susceptible to induced noise. An in-depth noise analysis was performed on each
piece of equipment in the lab to reduce the noise floor. Data was taken with a
National Instruments Virtual bench with 1GS/s and either an SRS-SR570 or a
FEMPTO-DHPCA-100 operational amplifier. This data was then analyzed using
a Fast Fourier Transform (FFT) script. A small sample of 3 test runs shown in
Figure 31 below illustrate the varying frequency of noise for the diffusion pump,
turbo pump/thruster power supply and a test with the thruster running.

42

Figure 31: High frequency analysis of lab components contribution to background noise
The various components contributed to low band and high band frequency
noise. Thruster operation was shown to increase mid band noise over the test
duration. Analysis showed that the largest contributor of noise was the stepper
motors for the traversing system. When they were active in the locked position,
they maintain control by rapidly stepping between 2 positions thus inducing
significant current into the ground plane of the chamber. These motors
contributed 30-60 microamps of noise, 3 orders of magnitude above the expected
thruster signal. The TOF high frequency gate initially induced 10-50 nanoamps of
noise because of the long 20-foot wire that carried the high voltage pulses. The
shielded ground on this cable over the long run was inducing noise in the ground
plane. This was reduced to 1-5 nanoamps when the equipment was moved and

43
further reduced when the ground was moved to a separate ground along with the
power supplies and other lab components.
3.2.2 Grounding

Noise on the data channels seemed to increase over time and
diagnostically looked like it built up during long duration thruster runs (Figure
32). This indicated a grounding problem with the lab facilities ground. An

Figure 32: Example time of flight run with ground induced noise
investigation showed that the ground wire from the chamber and equipment ran to
an old water drain line that was cut and no longer properly conducting. The
resistance between actual ground and the chamber was 120k ohms (Figure 33).
This was enough resistance at the thruster operational potential to add a
cumulative 11.6-21 mA/s of charge capacity on the chamber.

44

Figure 33: Faulty chamber ground
Facilities was very aminiable to dedicating a ground line to the lab (Figure
34), thank you FMS! The new ground was installed with a 4 AWG conducting
line to the main, earth burried, ground line of the building at the base of the
electrical pole. This new ground showed less than 5 ohm resistance to to other
Earth grounded objects. One note is that this is still an electrical line ground and
not a dedicated instramentation line to Earth, as such 60 hz noise is still prevelent.
With the large gauge wire that was used the current is minisule and this noise can
be filtered out in post proseccing data reduction.

45

Figure 34: New dedicated facilities ground in LAPD
In addition to the addition of a dedicated ground the large facility
components were isolated from what became the chamber and data ground and
grounded to another ground line at an adjacent electrical panel. These two main
effects enabled a significant reduction in the noise floor for the LAPD lab. Figure
35 below shows this drastic reduction in noise compared to Figure 31c. The
thruster and

Figure 35: High frequency analysis of lab components contribution to ground floor noise with the newly
installed dedicated data ground.

46
pumping devices were on and this test was taken after 6 min of what would have
previously raised the ground floor. The 0 to .08V ultra high frequency noise is bit
error noise cause by the 8-bit nature of the NI Virtual bench. This reading was
taken with the traversing system stepper motors off. This configuration could only
be maintained during negative mode operation as the Arduino was the only way
to control the Glassman positive emission HV power supply. For Positive mode
TOF the stepper noise was removed via post processing.

47
Chapter 4: The USC Testbed Thruster

The Laboratory of Astronautical Plasma Dynamics at the University of Southern
California has designed, developed, and tested an Ion liquid Ion Source (ILIS) operating in the
Pure Ionic Regime (PIR) utilizing 1-ethyl-3-methylimidazolium-tetrafluoroborate. The USC
Testbed Thruster exploits the repeatability and stability of replaceable 1-cm machined
borosilicate emitter and reservoir layers. The workhorse thruster employs a convenient modular
design which permits the introduction and testing of rapidly developed prototyped parts.
Increasing emitter tip density is a critical development goal that will lead to expansions in the
capability and utility of electrospray thrusters. Extractor design and manufacturing is a major
limiting factor in the current densification of emitters. At higher density, there is simply not
enough material between ion grid windows to ensure the structural integrity of the extractor at
the operating potentials required to emit in the PIR. To provide the groundwork for the
investigation of novel extractor designs, the USC Testbed Thruster performance was baselined
and is presented below. Overall, the USC Testbed Thruster performs well with minimal losses
due to grid interception.
4.1 Thruster Design and fabrication
The USC Testbed Thruster was designed as a ½ scale AFET-II thruster to take advantage
of readily available materials, namely 1 cm P4 and P5 borosilicate disks. The UTT is 2.54 cm L
x 2.54 cm W x 1.27 cm tall. An expanded view of the thruster can be seen in Figure 36. The
UTT consists of 5 main parts; the aluminum body, a Polyether Ether Ketone (PEEK) propellant
housing, a borosilicate reservoir, a borosilicate emitter layer, a distal electrode, and an extractor
grid. There is also a tension spring in the propellant reservoir to ensure contact between the
emitter layer and the distal electrode is maintained at a constant pressure. The body of the

48
thrusters’ made of aluminum and is connected electrically to the extractor grid. The extractor
grid is designed to sit at the top plane of the emitter tips. This reduces ds to zero in an ideal
configuration. Variation in emitter height due to manufacturing tolerances can cause ds to vary
by approximately 24 µm.

Figure 36: USC Thruster V1 Expanded Rendering.

The propellant module sits inside the body and is made of Polyether Ether Ketone
(PEEK). The high dielectric strength of the PEEK reduces the bulk and weight of the thruster
and allows for tighter tolerances in manufacturing. The modular design of the propellant
reservoir allows for efficient propellant loading operations and quick replacement of faulty or
broken components. The propellant and borosilicate disks are encased in the propellant module

49
with the distal electrode. This electrode is a 254 μm thin sheet of stainless steel milled to
surround the emitter plane fully.
Borosilicate glass was chosen as the reservoir and emitter substrate due to its
nonconductive nature and its inherent resistance to electrochemical degradation [9]. The
selection of borosilicate glass leads to the need for a reliable distal electrode to raise the
conductive EMI-BF4 propellant to high potentials evenly.
4.1.1 Emitter and Extractor

The emitter array and extractor are the most critical components in the UTT. They have
the tightest tolerance requirements of 25 μm. Any misalignment or discrepancies in
manufacturing lead to increased particle impingement and a loss of efficiency. The design of the
thruster allows for this inefficiency to be directly measured. During normal operations, the body
is grounded through the data acquisition system. This allows for a direct measurement of current
intercepted by the extractor grid. This mode of operation allows for the intercepted current to be
turned into a weighting factor in the design and evaluation of extractor grids. The intercepted
current is an important metric in the repeatability of individual thruster operation. It is also the
defining metric in the comparison of the efficiency of individual thruster designs.
Four extractors have been conventionally machined with grids of 25 ion windows,
measuring 381μm and 508μm. Figure 37 displays the manufactured extractor grids. These grids
suffer from pitfalls and structural stability of micromachining at such tight tolerances. Figure
37(a and b) show grids of 508μm holes with 37.5 μm of material separating the voids. Grids with
such tightly packed holes are on the very edge of the practical tolerance provided by
conventionally machining techniques. Figure 37(d-e) show more conservatively machined grids
of twenty-five 381μm holes. While these grids are easier to manufacture, they suffer from higher

50
particle interception due to the reduction in allowable plume angular expansion. Figure 37(g) is
the distal electrode for comparison.

Figure 37: Extractors
(a) and (b) grids of 25 508μm holes, (c) Filament irregularities from the manufacturing process, (d) and (e) grids of
25 381μm holes, (f) 381μm holes post deburring process

A byproduct of the conventional machining process can be minor irregularities in the
finished product. These irregularities can be seen in Figure 37(c) as metal filaments that bridge
across individual ion windows. As displayed, these filaments would lead to a higher current
interception and reduced thruster efficiency. More critically, these filaments can lead to a
shorting of the thruster if the ends become detached. This shorting will most likely end the life of
the thruster if it cannot be removed. To reduce this risk, a process of deburring was utilized to
eliminate the filaments from the extractor; Figure 37(f) shows the 381μm holes after the
deburring was performed.
Multiple attempts were made to manufacture new grids using Direct Metal Laser
Sintering on an EOS M290 machine Figure 38. The machine had the resolution to make the
e
f
a b
g

51
.015” window grids but the process required support below the extractor and the support could
not be printed at a resolution that enabled grid stability. A second attempt was made to

Figure 38: Additively manufactured grids with insufficient support
manufacture the parts at an angle that would allow the grids to be self-supporting but this process
resulted in a part that needed significant post processing and did not have the flatness of the
conventionally manufactured grids.
Various grids were designed and modeled in COMSOL. These grids ranged from slits
and square extractors with multiple emitters tips beneath each extraction opening to novel
designs that had tapered extractor openings to allow for expanded plume angles with thicker,
more ridged extractors. These novel designs can be seen in Figure 39. An attempt was made to
manufacture the tapered emitter with additive manufacturing, but due to the tight tolerances and
the lack of inherent grid support, the proses proved out of reach of current sintering technology.

Figure 39: CAD renderings of novel extractor profiles,
items from left to right: tapered extractor, quad extraction configuration for UTT, 5 381 μm slits, 5 508 μm slits.

52

When modeled, the multiple emitter designs, whether slots or square, showed large
plume half angles toward the edges of the extraction frame (Figure 40). This finding is
corroborated in the development of the IFM nano FEEP thruster, which uses a toroid of emitters
and a circular extractor [32]. Benefits from reduced extraction impingement at this scale would
be negated by the loss of thrust over the lifetime of the thruster.

Figure 40: COMSOL modeling of a novel extractor design
4.1.2 Propellant Loading

Propellant loading is currently done in the atmosphere with a microsyringe dropper and
an Ohaus AP250D scale. The propellant chosen for all thruster firings is 1-ethyl-3-methyl
imidazolium tetrafluoroborate (EMI
+
BF4
-
) properties for the emitted species can be seen in Table
2.

Table 2: Expected Constituents from EMI
+
BF
4
-

53
The propellant module and the propellant are outgassed for at least 24 hours before the
loading operations begin. During the course of this effort, issues with the propellant loading of
the first thruster led to an optimization of the loading process. Initially, the propellant was loaded
after the propellant module was fully assembled. The propellant was applied to the emitter
surface and subsequently absorbed into the reservoir through capillary forces. This process led to
pooling and uneven absorption of propellant into the reservoir layer. The results of this process
are displayed in Figure 41-left. When loading the propellant through the emitter layer, the
saturation of the emitters occurred at 24% of the calculated porous maximum allowed. The
yellowing of the propellant was due to the age of the initial batch that was utilized for thruster
testing. All subsequent thrusters were loaded with a modified procedure that proved to show
better saturation results. Under this new procedure, the reservoir is filled and weighed before the
emitter layer is filled. The distal electrode is installed after all loading operations are completed.
This new process enabled an 85% fill rate for Thruster 2 and a 92% fill rate for Thruster 3.

Figure 41: Propellant loading of first (left) and second thrusters(right).
Stills from the absorption of the propellant. Saturation and pooling can be seen in the last frame of
Thruster 1. Even spreading and no pooling can be observed for the loading of Thruster 2

4.2 Test Results
4.2.1 Thruster One Testing

Thruster 1 was assembled using the 508 μm 25 emitter grid, with the emitter tips centered
in the middle of the extractor windows and flush with the extractor’ bottom. It was fired in

54
positive ion mode, emitting EMI
+
, for a duration of only 8 minutes. Startup voltage was
approximately 2000 V, 250% of the design Vstart of 800 V. The discrepancies in performance can
be attributed to damage to the emitter tips (Figure 42) that happened during the installation of the
distal electrode. This damage, alongside the propellant loading problems identified earlier, led to
significant shorting and higher operating voltages. Extractor intercepted current was 7% of
emitted current. This number is high and would have led to premature failure of the thruster if
the testing continued. Considering the severely damaged condition of the emitter tips, the
measured intercepted current was lower than what one would expect. An analysis of the startup
voltage using Eq. 13 shows that the separation distance for the emitters (ds) was approximately
20𝜇𝜇 m. This is significantly more than the design of 5 𝜇𝜇 m and is evidence of the extent of damage
to the emitter tips.

Figure 42: Thruster 1 Emitter profile. Irregular emitter tips and a row of clipped emitters

𝑉𝑉 𝑡𝑡 ℎ
≅ �
𝛾𝛾 𝑅𝑅 𝑐𝑐 𝜖𝜖 0
� 𝑙𝑙𝑙𝑙 �
4 𝑑𝑑 𝑠𝑠 𝑅𝑅 𝑐𝑐 � Eq. 13
Figure 43 shows the voltage and current profile of the full duration of the Thruster 1 test
set. The test for Thruster 1 consisted of three gradual rises in applied voltage resulting in three
thruster startups. The first startup was terminated by an unexpected short at approximately t =
500s. This short is believed to be from liquid bridging between the emitter layer and the extractor

55
grid due to the pooling of propellant after loading operations. The short burnt itself out, and the
thruster began normal operations for a short period during startup two. A few major shorts in the
middle of operations led to the shutdown of the voltage source. The thruster was started for the
third time and was subsequently shut down due to the inconsistent operation.

Figure 43: Source Voltage and Current of Thruster 1 Initial Test.
Emitter source voltage (V
em
), emitted current (I
em)
), and extractor intercepted current (I
ex
) from the full test
campaign of Thruster 1

Although the test of Thruster 1 was inconsistent and short, it showed that the data
acquisition system could sufficiently measure the vital parameters needed to investigate future
thruster performance and that the overall design and assembly produced a viable thruster.
Measurements taken during Thruster 1 operations were sourced voltage, source current,
intercepted current on the extractor, and three spatial probe current measurements.
4.2.2 Thruster 2 Testing

Thruster 2 was assembled using the 508 μm 25 emitter grid, with an expected distance to
the emitter tips of 0mm (Figure 44). Initially it was fired in negative ion mode, emitting BF4
-
.
The thruster was run intermittently over three days for over 2.5 hours of emission time. Thruster
2 showed a 420% improvement in performance and intercepted current over Thruster 1. The
thruster operated with an average extractor interception of 1.66% and an initial high of 4.5%.

56
The third operation performed the most remarkable with a 0.34% average interception over 30
min of 1700 volt emission. Utilizing equations 2-7 in appendix A, thruster performance was
established over 20 tests and 20.74 combined hours of operational time, in both positive (6.87
hrs.) and negative (13.87 hrs.) emission modes.

Figure 44: Thruster 2 machined emitters

The paper on “Comparing Direct and Indirect Thrust Measurements from Passively Fed
Ionic Electrospray Thrusters” [10] used a Time of Flight mass spectrometer to determine the
percentage of monomers and dimers in the plume of a passively fed borosilicate glass
electrospray. Utilizing EMI-BF4, Dan Courtney found a 44% monomer and dimer population
with 4% of trimers and 4% of larger mass constituents. This information was used to make
performance estimates of the UTT across its operational voltage range of 1300V-1750V Figure
45. Thrust was found to be between 42 and 76 nN and ISP was estimated to be between 4355 and
6340s. This Isp measurement overestimates performance as it does not account for neutral mass
loss from the electrochemical breakdown of EMI-BF4. Mike Natisin speculates that
electrochemical breakdown in the propellant leads to large neutral losses during thruster
operation [50]. Using information on the true mass loss of the thruster form Table 3 With 44%
monomer and dimer constituents, the thrust was found to be between 74nN and 4.6µN, and the
ISP was estimated to be between 1650s and 2374s. If the thruster mass loss is due to neutral loss
and not propellant exhaust, then the thrust would still be in the 42-76nN range.

57

Figure 45: Startup Source Voltage v Current of Thruster 2.
Initial startup at 1300 V with the stable startup of 1400 V and maximum operation of 1750 V

a. Plume Characterization

Multiple Faraday Probe and Micro Faraday Probe measurements were
made of Thruster 2’s plume during operation. Figure 46a/b show Micro Faraday
Probe scans of the emission surface of the thruster. Scans were performed at
standoff distances of 4” and 0.75”. Limitations in the three-axis traversing system
and the probe configuration initially led to a minimum scan distance of 4 inches
from the truster’s face. As seen in Figure 46a, the scan distance was too far away
from the truster to resolve individual emitter sites, but a clear plume can be
resolved. For clarification, an image of the thruster has been overlayed on top of
the sensor heatmaps to display the thruster’s relative position to the scan area.
Figure 46b illustrates better emitter site emission resolution after the traversing
system was moved closer to the thruster face. The plume appears to be saturated
even at this 0.75” distance. After further refinement to the traversing system, the
MFP was moved to 0.25” away from the thruster, and a closer map with smaller
steps sizes was performed. The emitter tip resolution can clearly be seen in the 3D
thruster two extractor map seen in Figure 47.

58

Figure 46: Heat maps of MFP with 4” dis. (left) .75” dis. (right) to thruster face.
Thruster position is relative to the scanning probe location and scaled for distance

Figure 47: MFP heat map at .25” from thruster face
Thruster 2 Test 5 @ 1731V (left) and Thruster 2 Test 7 @ 1767V (right) Extractor Potential. Current
Measurement at 6.35mm from Extractor Face

Faraday probe data was collected across two of Thruster 2’s operation
cycles. Heatmaps of the horizontal plane with respect to the thruster are shown in
Figure 48a/b. These heatmaps are renderings of the plume charge in the Y
direction, the horizontal chamber axis WRT the thruster, as Z is moved away
from the thruster. The thruster position relative to the scan area would be centered
on the x-axis starting at 4 inches away from the thruster.

59

Figure 48: FP heat maps of the horizontal plane.
Z distance is the distance to the thruster, Y distance is from the center of the thruster face (2.5”)

Analysis of the first rendering was initially thought to be erroneous. A
second map was taken along with a scan of the vertical plane with respect to the
distance from the thruster Figure 48b. Although the magnitude is slightly different
due to different operating parameters of the thruster, a secondary anode is
consistently found at 6 inches from the thruster plane. The XZ plume scan in
Figure 49. shows the same secondary anode at a Z distance of 6 inches from the
thruster. It is believed that the leading cause of error in these probe scans is the
chamber environment. The chamber is a non-coated stainless-steel chamber that is
susceptible to very high secondary electron/ion emission under high energy
negative ion bombardment [94]–[96]. This diagnostic anomaly seems to be a
consequence of the experimental setup and maybe unavoidable with the current
chamber.

60

Figure 49: FP heat map of the vertical plane.
Z distance is the distance to the thruster, X distance is from the center of the thruster face (4.55”)

b. Thruster 2’s Disassembly

After Thruster 2’s performance was severely degraded and startup
voltages were approaching 2500V, the thruster was decommissioned and
disassembled. Severe arching can be seen in the photographs from this
disassembly. Figure 50a shows the emitter under the extractor with charring
toward the tip of the emitters. This charring was investigated under SEM,
significant concentrations of Carbon can be seen on the emitter tip in Figure 51.
Additionally, stainless tendrils look like they have broken free of the extractor.
After disassembly, the arching events across the thruster face were evident, Figure
50b. The distal electrode was seen to have arching damage from extractor to
distal electrode breakdown Figure 50c.

61

Figure 50: Thruster 2’s disassembly pictures
(Left) emitter charring under extractor, (center) damage to emitters from arcing, (right) arching damage
on the bottom of the extractor

Figure 51: SEM Image of carbon structure on the top of the emitter.

Thruster 2 still had propellant remaining in the emitter layer, but the
reservoir was almost entirely void of propellant (21.9% by volume remaining).
Table 3 shows the propellant usage across all testing of Thruster 2. Weights were
taken before and after each test, to determine a propellant usage for each test run.

62

Table 3: Thruster 2 firing times and mass change
It was observed that the thruster would gain mass up to 30 minutes after the
chamber was pumped down. Great care was taken to keep the thruster in vacuum
when it was not in use and to weigh it promptly upon removal from the chamber.
A couple instances where this could not happen are seen in firings 4, 12 and 15.
These data points are preserved for posterity but not counted in the average mass
flow rate. Upon disassembly, it was noted that the emitter layer had more
propellant than was initially loaded, but the revisor was completely drained.
Because these thrusters operate due to capillary wicking of propellant from the
larger reservoir pores to the smaller emitter pores; it is believed that with an
empty reservoir, the emitter has virtually no backpressure and the electrostatic
force required to extract propellant increases. This seems to happen because,
without the minuscule back pressure of the capillary force from larger pores to
smaller ones, the electrostatic force must overcome the same capillary action
holding propellant in place [38]. Reduction in propellant viscosity and a thinning
of the emitter layer may overcome this engineering problem.

63
4.2.3 Thruster 3/4 Testing

Three new sets of emitters were manufactured for the baseline testing of Thruster 3 and 4.
The thrusters were assembled and primarily used for TOF gate development testing and
simultaneous quasi-neutral bipolar thruster operation. Figure 52 shows the dimensions and
profiles of these emitters.

Figure 52: New emitters for thruster 3 and 4

Thruster 3 was constructed with emitter 3 in Figure 52 and the .015” window extractor
grid e from Figure 37. Emitter 3 has 1 clipped emitter tip that was damaged during transport.
Thruster 3 has been fired for over 15 hours, across 16 test runs in both positive and negative
modes of operation. Thruster 3 was filled with more propellant in the reservoir due to the lessons
learned from Thruster 2’s premature failure. The reservoir was filled to 97% capacity and the
emitter was filled to 98.9% capacity for a thruster total of 98.3%.
The thruster was fired 12 times and then was disassembled, inspected and rebuilt. Initial
startup voltage for the thruster was low at 1000V but % interception was high at ~20%. Thruster
plume scans showed a 14.5-degree plume half angle (Figure 53). This half angle is favorable

64

Figure 53: Plume scan of thruster 3 operation showing 14.47°plume half angle
considering on orbit IFM Nano FEEP thrusters have 60-degree plume half angle. RPA plume
potential scans shows a linear distribution of potential from 100V to 94% of the thruster
operating voltage (Figure 54). This can be attributed to the fragmentation of ions as well as
variable heights of emission sites relative to the extraction plane.

Figure 54: RPA Plume analysis of thruster 3 negative mode operation

65
Performance data for thruster 3 looked slightly worse than thruster 2, this is believed to
be because of the higher % int which reduces the number of ions escaping and therefore reduces
thrust. Interception due to the smaller window size of the grid was between .8% and 5%, after 1
hour of consistent operation. This interception spiked to 8%-11% during the 12
th
firing ~11 hours
into operation. Investigation of the thruster showed a large build up on one of the emitters
(Figure 55). The ISP during the operational potential of +/- 1340-1910V ranged from 3960-
4660s. The thrust range was 38-68 nN.

Figure 55: Large structure on the tip of the second emitter from the top left
Build up on emitter tip at the top right of Figure 55 caused concern and thruster 3 was
disassembled. An investigation into the phenomenon showed a large carbon structure had built
up at the tip of the emitter Figure 56. This carbon structure was removed but it shows that

Figure 56: Carbon Build up on fired emitter (left) and melted borosilicate pores from arc events (right)

66

impurities in the emitters and propellant can cause serious problems. While this structure did not
short out the thruster it led to the increased extractor impingement that warranted investigation.
If this structure were to fall off in a 0g environment it could have floated into the grid and
shorted.
A small amount of propellant was added to the rebuilt thruster (Figure 57), bringing it to
within 98.45% of being full. This thruster has been fired in 3 simultaneous quasi-neutral bipolar
thruster firings and is still in developmental testing at the time of writing.

Figure 57: Reloading of thruster 3

Thruster 4 was constructed with emitter 2 in Figure 52 and the .02” window extractor
grid a from Figure 37. Thruster 4 has been fired in bi-polar operation mode for over 12 hours,
across 11 test runs. Initial testing of thruster 4 was unsuccessful due to an overfilling of
propellant and a liquid bridge that formed between the emitters and the extractors (Figure 58).

Figure 58: Thruster 4 loading

67
The reservoir was filled to 100.8% capacity and the emitter was filled to 100% capacity. This
was remedied after a grid disassembly and a wicking of a small amount of propellant from the
emitter face. The thruster is the best performing version of the UTT so far with thrust in the 267-
885nN range and ISP values in the 3600-4050s range. The thruster has been fired along with
thruster 3 in quasi-neutral bipolar operation and under single operation for further TOF

Figure 59: simultaneous quasi-neutral bipolar thruster operation of thrusters 3 and 4.
Diagnostics (Figure 59). Figure 60 illustrates thruster start up and shut down. It can be seen that
start up occurs at 1100 Volts and shut down occurs at 1180 Volts. While this is slightly higher
than thruster 3 it is significantly lower than the 1300 V startup of thruster 2.

68

Figure 60: Thruster 4 start up and shut down
Probe Extractor, RPA and FP current is multiplied by 100 for ease of readability
Thruster 4 has continually operated with extremely low % interception in the range of
.08-.5% through the full range of positive and negative operation. RPA Plume analysis of
Thruster 4 shows the same relatively linear regression in the plume constituents as the RPA
voltage is raised to the thruster operating potential (Figure 61).

Figure 61: RPA sweep of thruster 4 plume potential

Thruster 4 should serve the LAPD and important research well into the future.

69
4.2.4 TOF Results

Time of Flight spectroscopy data was taken for both the conventional gate and the BNG
gate (Figure 30, Figure 28), the gates were positioned at 2 inches (conventional) and .5 inches
(BNG) from the thruster face. This was done to minimize the back scattering of ions on the
thruster face when the conventional gate was closed and to maximize deflection angle distance
for the BNG when it was closed. Expected time of flights, with varying thruster extraction
voltages, for the propellant EMI
+
BF
-
4 with a collector plate distance of 30 inches can be seen in
Table 4. The monomer and dimer weights can also be seen, it was assumed that all ions were
singly charged, this is consistent with the literature [6].

Table 4: Expected Constituents and flight times from EMI
+
BF
4
-

To get a reliable time averaged TOF signal 10-50 samples had to be taken at a time for
each given configuration. These signals were then time synchronized and averaged to get 1
baseline signal. The signal was then post-processed and normalized for TOF reduction. This
process can be seen in Figure 62. Data was taken with a National Instruments Virtual bench with

70

Figure 62: Collected TOF raw data (left) and Time Synchronized Averaged data (right)
1GS/s and either an SRS-SR570 or a FEMPTO-DHPCA-100 operational amplifier. Both the
BNG and conventional gate were powered with a Hewlett Packard 6522A +/- 2000V power
supply and cycled with a DEI PVX 4140 high voltage pulse generator. Large current spikes can
be seen in Figure 58(right) due to the PVX-4140 dropping high potential current from the gate to
ground. This noise subsides before the monomer flight time and thus can be ignored. The
thruster and test apparatus were placed in a 1 cubic meter vacuum chamber operating at 5E-6
torr. A probe sweep system with 2 Faraday probes and an RPA was utilized to take background
plume measurements when the BNG was in operation. Figure 63 illustrates the chamber setup
and component placement.

Figure 63: Top-down schematic of test setup in the chamber

71
As stated before, plume scans were taken to examine gate operational effectiveness.
Figure 64 below shows the reduction in current in the downstream plume region due to gate

Figure 64: Axial plume scans of BNG off and BNG on
operation. This 500V BNG operation was initially tailored to the 14.5-degree plume half angle of
thruster 3. The expected deflection of the BNG was 7.8-degrees but was observed to be 9.6-
degrees in practice. This increase is due to a slightly lower than expected thruster plume half
angle. Throughout the testing of the BNG it was observed that one drawback of the BNG is that
its operational voltage must be tuned to each individual thruster and updated throughout the life
of the thruster. Figure 65 illustrates the plume to BNG voltage analysis for each individual
thruster. Because the BNG alternates polarity of deflection between each successive high low
wire pair the plume angle of the thruster and the distance to the collector plate must be take into

72

Figure 65: Plume analysis for BNG thruster tuning
account in addition to the ion potential of the thruster’s plume. The top 2 graphs show an
idealized deflection with no plume spreading. A 6-degree BNG addition is necessary for the
plume to clear the collector plate. The middle two graphs show the scan results of thruster 2 and
3’s plume expansion in the chamber. This additional beam spreading (bottom 2 graphs) requires
an additional BNG plume diversion to clear the collector plate.
Weather using the BNG or the conventional reflection gate, the expected TOF results
were the same. Figure 66 illustrates an idealized and notionally expected TOF signal with 54%
monomers 44% dimers and 2% trimers. The difference in the two expected results comes from
the addition of variable velocity (cj) distribution due to a spectrum of potentials of each

73

Figure 66: Idealized and Expected Time of Flight data
individual constituent. The RPA data corroborates the expectation that the monomers, dimers
and trimers, will not have a homogeneous velocity and will instead have a spectrum due to
varying potential because of the locations of extraction relative to the extractor as well as
fragmentation in the extraction region.
TOF results with the conventional gate were not conclusive and suffered from a severe
degradation in signal. The initial gate described in chapter 3 was only 12% transparent because
of the need to stack multiple grids of low transparency. The conventional gate was increased to
30% transparency but with the initial grounding noise and signal degradation TOF results were
non conclusive. Figure 67 shows the time synchronized average results of a conventional gate
TOF spectroscopy for a thruster operating with +/- 1450 V extraction and a gate repulsion
voltage of +/- 1650V. Noise is driving the initial slope before monomer flight time and chamber

74

Figure 67: Reflection gate normalized collected signal: gate voltage 1650V Thruster voltage 1450V. Negative BF
4
collection (left) Positive EMI collection (right)
wall effects (and reflection) are driving the rise in intercepted current over time towards the
trimer flight time. A longer chamber with a more transparent gate would allow for better signal
to noise ratio as the flight times would be after the initial gate noise has fully subsided. The
higher the gate voltage the longer duration the PVX-4140 rings when it cycles the gate. The gate
closed signal from the right half of Figure 67(right) shows the best promise for a reliable TOF
reading, monomers level off at .7 dimers fall off as expected at .9 and trimers rise peculiarly up
to the full duration of the test. If these results are true the thruster has ~78% monomers and
~22% dimers.
The BNG operation, when properly tuned was significantly more reliable. The lower gate
voltage allowed for the PVX 4140 noise to subside well before the monomer flight time. While
the time synchronized averaging removes most of this noise, it is still important for it to subside
before the expected flight time to ensure a reliable reading. Figure 68 displays TOF results for
Negative mode (left) and positive mode (right) thruster operation with the BNG deflection gate
deflection. The thruster operated at +/- 1450V and the BNG was optimized at 720 V wire pair

75

Figure 68: BNG normalized collected signal: gate voltage 720V Thruster voltage 1450V. Negative BF
4
collection
(left) Positive EMI collection (right)
difference. 10 samples were averaged for each signal. Negative mode (BF4) operation showed
~20% monomers ~80% dimers. Positive mode (EMI) operation showed 33% monomers 77%
dimers. Tail off after monomers and dimers may be due to reflected ions from the small
chamber. The difference between the monomer and dimer constituencies from the two gates
indicate that the methods of plume diversion are altering the signals. A look at the extractor
intercepted current shows a possible detrimental effect of conventional gate reflection back on
the thruster. Standard operation of the UTT has been baselined with .08-.4% impingement as
compared to total emitted current. When the conventional gate was placed in front of the thruster
in the “Open” configuration extractor current increased to 4% of the total beam current and when
the gate was operating in the “Closed” position the extractor current increased to 45% of the total
beam current. While the conventional gate reduces downstream collected current on the collector
plate by 99% in the closed position, the gate increases thruster impingement by a factor of 562.
This impingement caused visible thruster arcing on the surface of the extractor and when high
energy return ions impact the emitters degradation and adverse performance effects can be seen
in thruster operation. The BNG showed no such rise in intercepted current during operation. The
return current from the conventional thruster could be fragmenting or altering the plume

76
constituents during thruster operation. For this reason, we have included BNG results in the
expected thrust of the UTT.

77
Chapter 5: The Magnetically induced Test Thruster

5.1 Thruster Development
The main difference in how the MTT thruster operates with respect to the UTT is in its
acceleration of ions, instead of using a static electric field, this thruster uses a dynamic magnetic
field to generate a quasi-static electric field. Currently, the dynamic magnetic field is generated
by one Helmholtz coil that is oscillated at high-frequency (Figure 69). This potentially could also
be done with a spinning magnet if the structure of the magnet could withstand the force of

Figure 69: CAD blow out of MTT-3 with propellant reservoir.
spinning at hundreds of kHz. Three versions of the MTT have been built (Figure 70). MTT-1
was a demonstration coil pair to experiment with driver voltage and see if a high current would
heat and melt the additively manufactured PLA body of the MTT. MTT-2 was a tech
demonstration thruster with one coil using the previously fired emitter from UTT-2. This emitter

78

Figure 70: MTT versions as built, (left) MTT-1, (middle) MTT-2, (right) MTT-3
was refilled and utilized in MTT-2 without a propellant reservoir. MTT-3 was constructed with a
new propellant reservoir and emitter one from Figure 48.
While conventional electrospray thrusters utilize a static electric field, the MTT takes
advantage of Faraday’s law ( Eq. 14, Eq. 15) by dynamically changing a magnetic field that runs
through the emitters (Figure 71) to generate a virtual electric field to accelerate ions without the
need for an extractor. In order to generate a high enough gradient in electric field the change in
magnetic field has to be large a 1500V electric potential requires 153kHz magnetic oscillation.

𝜀𝜀 = −
𝑑𝑑 𝜙𝜙 𝐵𝐵 𝑑𝑑𝑑𝑑
Eq. 14

𝛻𝛻 𝛻𝛻𝛻𝛻 = −
𝜕𝜕𝜕𝜕
𝜕𝜕 𝑑𝑑 Eq. 15

79

Figure 71: Static depiction of the magnetic field generated in the emitter region of the MTT
The current design for the MTT Figure 69 shows a structure of two orthogonal
Helmholtz coil pairs. This design allows for inherent beam stabilization or future thrust
vectoring. This design of the thruster utilizes a matched Helmholtz pair with a coil driver that
utilizes a series capacitor. This series capacitor is required because at the excitation frequency of
the MTT the inductance of the wire, or more accurately the resistance induced in the coil, can
overshadow the natural wire resistance by two orders of magnitude. The series capacitor
performs the function of capacitive matching the inductance produced by high-frequency
amperage oscillation. This capacitive matching reduces the voltage required to drive the coil pair
and brings it in line with the voltage required with no inductance. For the initial MTT this
reduction in driver voltage at 2.9A was from 323V to 2.29V. A wiring diagram of the thruster
electronics can be seen in Figure 72.

80

Figure 72: Line diagram of Helmholtz driver [79]
Figure 73 is a COMSOL model of thruster operation at 156kHz. The electric
displacement field from this operation produces a force that propels the ions away from the
thruster orthogonal to both coil pairs. Figure 74 is a representation of the surface induced current

Figure 73: Electric Displacement field generated by 156kHz operation of the MTT.
at the emitter tips. Displacement is orthogonal to both matched coil pairs and in the desired
direction of thrust. This displacement field oscillates in polarity along the z axis. The magnitude
of the displacement field is a function of the sinewave that is fed into the driver coils.

81

Figure 74: MTT Surface induced current density at emitter tips
Surface current is the highest at the emitter tips where we expect it to be.

5.2 Thruster Test Results
Before placing the MTT into the chamber some initial benchtop inductance tests were
performed. A small conducting plate was placed in the place of the emitters and current was
measured as the frequency and capacitance of the driver were changed. The initial testing
allowed for the capacitance to be fine-tuned. MTT-1 was a 30 winding coil pair with a radius of
2 cm. Utilizing equations 4-8 the required magnetic field to generate a 1500V potential
difference at the extraction tips was calculated to be 9.77 mT. This field required a 153.5 kHz
oscillation at a current of 3.62A. With no capacitive matching the driver voltage would have
been 255V with matching the voltage was reduced to 2.9 V. Optimal capacitance for MTT-1 at
30 windings was found to be 14700 pF. In practice this needed to be almost doubled to initiate
oscillation. Table 5 shows a comparison of MTT-2 and MTT-3 coil specifications. The goal of
MTT-3 was not only to add a reservoir but also lower the required driver amperage to what was

82
observed to be the maximum of testing for MTT-2. This is believed to be a driver limitation but
may be a capacitance charge/discharge frequency limitation.

Table 5: Comparison of drive specifications for MTT-2 and MTT-3
Bench top test results (Figure 75) from MTT-1 show some interesting correlations
between the function generator signal, waveform amplifier output and the induced current in the
MTT propellant cavity. An SRS DG535 function generator was used to send a square wave
(yellow Ch1) with varying frequency to the an ACCEL TS250 waveform amplifier (blue ch2).

Figure 75: Bench top test of MTT
The ACCEL TS250 has a high frequency current output at 100mV per Amp. Figure 75 shows a
correlation in function frequency and induced current at the propellant reservoir (pink ch3). The
driver is fed a square wave that is turned into a SIN wave by the capacitance charge and
discharge time, better capacitors would reduce this response.
The MTT has been fired in the USC vacuum chamber, showing the emission of both
positive and negative ions from the same thruster emitter chip that is used in the UTT. Figure 76
shows a simplified version of a diagnostic test used to confirm the emission of ions at high-

83
frequency oscillation. The thruster operates as the current oscillates, in a sinewave, between -3
and 3 Amps, at high frequency (~150kHz). This high-frequency oscillation makes diagnostics

Figure 76: a) A simplified representation of a firing thruster’s expected results. b) Results from 3 separate test
firings overlayed to show consistent operation with respect to theoretical.
challenging. Figure 76a shows the thruster oscillation (red line) and current collected (blue line)
by probes in the chamber. The probe that is used to collect this measurement is a Retarding
Potential Analyzer (RPA) that can bias a grid in front of the collection device to high potentials.
The action of biasing the retarding grid on the RPA limits the collection of a certain species of
emitted Ions. Positive bias precludes positive ions from reaching the plate, and this is the same
for negative bias. The green area in Figure 76a/b shows a negative bias of 1000 volts, while the
blue area shows a positive 1000-volt bias. We expect to receive less current, only in the polarity
of retardation, on the collector of the RPA when the bias is on; this is the difference between the
red and blue lines. Figure 76b shows actual data collected during a test run of the thruster. We
can see from the RMS line that the current intercepted by the RPA increases appropriately when
RPA retarding voltage is applied. This indicated a consistent thruster operation.
The RPA operates in a static configuration with the high voltage retarding grids set to
various voltages while data is taken. Because the thruster is operating in bipolar operation the
RPA still receives current of the opposite polarity from retarding potential. A frequency analysis

84
of the RMS of each signal was performed on MTT-2 and MTT-3 with and without emission.
These histograms show a clear pattern of emission when the thruster is operating (Figure 77).
The base coil still induces current flow in the RPA collector plate but when the thruster is firing
current above the baseline can clearly be seen up to the retarding voltage that precludes current
collection.

Figure 77: Histogram of RPA RMS with MTT-3 off (left), Histogram of RPA RMS with MTT-2 Firing (right)

Figure 78: RPA sweep of MTT-2 plume spectrum
MTT-2 actual sustained drive current was 2.47Amp peak to peak with a 4 Volt function
at 155kHz sine wave. The induced ion acceleration potential was -1560V for Negative BF4
emission and 1625 V for positive EMI emission. This induced acceleration potential was derived
from the current collected by the RPA probe sweep of the MTT (Figure 78). This is higher than
the 1500 expected voltage but can be attributed to the larger drive voltage than needed. ISP was

85
calculated to be between 4270s and 5996s for Negative emission and between 4940s and 5418s
for positive emission. Monomer Thrust was calculated to be between 52nN-59nN with
Monomer/Dimer Thrust in the range of 76nN-81nN [97].

Figure 79: New Robu P5 Borosilicate glass Left and USC Emitter ROBU Borosilicate glass Right.

A problem with the vendor matirial quality for the emitters has been identified. Large
boulders can bee seen in the left image of Figure 79. These bolders in the substrate are larger
than the peaks of the UTT and will perclude the manufaturing of emitters at the current pitch and
density. This problem in large paticulate contamination of the base material seems to be a
manufacturing problem in quality control. The average pore size and consistancy will be a
limiting factor in the densification of the MTT. Peaks can not be smaller than the pores at the tip
and boulders or large agregate matirial will limit the ability to make consistant small emitter tips.
The consistant matiriel in the right image of Figure 79 was used in all of the UTT and MTT
emitters, this material has a pore size that limits tip density to aproximatly 200𝜇𝜇 m. A
densification of emitters to the aformentioned pitch would be a 7.8 fold increase over the current
demonstrated thrust.

86
The MTT has the potential to vastly alter the trade space of where electrospray thruster
utility falls. Currently, electrospray thrusters are more efficient than Hall effect and conventional
Ion thrusters, but they do not produce nearly as much thrust because they are limited by their
scale. The MTT is an enabling technology that would push the Electrospray utility up to higher
power and thrust/size ratios while maintaining the high ISP electrospray thrusters are known for.

87
Chapter 6: Conclusions and Proposed Future Work

Densification remains the primary goal of a myriad of electrospray thruster development
paths. With densification comes a broadening of the technological utility and with utility will
come mass adoption. Currently these thrusters are only suited for small thrust, small satellite
missions but that could change with large arrays of dense tips. The UTT pushes the bounds of
conventional machining techniques. The structural stability and material stiffness of the extractor
currently limits densification and scalability of a conventionally manufactured electrospray. The
UTT emitter pitch could be scaled from 533 𝜇𝜇 m to 432𝜇𝜇 m by utilizing different end mills and
closing the gap between the 381𝜇𝜇 m extraction windows. Doing this would result in high
impingement and premature thruster failure. The current practical limit of the density of the UTT
is ~500𝜇𝜇 m this is based on the relationship between the extraction voltage at the current tip
geometry and the required extractor window needed to keep the impingement as low as possible
to prolong thruster life. With current manufacturing techniques the practical limit on thruster size
and extrapolated thrust per unit would be ~4cm
2
, 56mN respectively. The practical size
limitation of the units is driven by the requirement to maintain a uniform separation between the
extractor plane and emitter tip plane. A miniscule deviation in extractor flatness will drive
asymmetric thruster operation and non-uniform tip emission due to the deviation in the static
electric fields.
With the current manufacturing practices and emitter structure the pitch could be shrunk
further to 267 𝜇𝜇 m if a 254 𝜇𝜇 m grid was used. Unfortunately, while the emitters could be
machined the extractor would have negative margin and cannot be manufactured. With the
increased intercepted current that the 381 𝜇𝜇 m window extractor suffered, due to plume
impingement at higher operational potential, the thruster would have to be operated at

88
prohibitively low extraction potentials to mitigate detrimental effects. This reduction in the
operational potential voltage would mitigate the added thrust from increasing pitch of the
thruster. The MTT can be the answer, with the elimination of the extractor 267𝜇𝜇 m or smaller
pitch can be achieved. The current pitch limit of the MTT, due to substrate pore dimensions,
would be ~200𝜇𝜇 m. At this pitch the currently used 1cm emitter disk would have area for 1400
tips and the MTT extrapolated thrust would be 25-80 𝜇𝜇 N. This is a significant improvement over
a larger UTT that could only support 180 tips on the same disk, at the current 533 𝜇𝜇 m pitch, with
an extrapolated thrust of 2-3.2 𝜇𝜇 N. While the UTT is limited by the ability to maintain
parallelism between the extractor and the emitters the MTT is currently limited by the
necessitated Helmholtz coil size to emitter area ratio. The elimination of the extractor allows for
the scaling of the emitter surface area to previously unreachable scales. It also eliminates a single
point of failure that affects conventional thrusters. In a conventional configuration if a single tip
out of the entire array shorts via conductive liquid build up it can raise the extraction plane to the
same high potential and destroy the thruster. In rare events this conductive liquid bridging can be
burned off by raising the voltage and ablating the material, but this is not a repeatable and
reliable on orbit mitigation practice. Liquid bridging is not a concern in the magnetic induction
configuration, liquid may still build up on the emitter tips but there is nothing for them to short
too, as the body of the MTT must be non-conductive for the magnetic induction to be focused in
the propellant. Plume impingement on the Helmholtz coil structure may be a long-term issue but
the structure of these coil supports could be built with mitigation in mind.
Performance and specific impulse were comparable between the MTT and the UTT,
Table 6 shows a summary of the performance metrics for the various thrusters. These

89

Table 6: Comparison of UTT-2,3,4 and MTT-2 ISP and thrust
performance metrics show the future utility of the MTT technology. The MTT-2 used the same
emitters that were previously utilized on the UTT Thruster 2. The difference in the ISP between
the two thrusters can be attributed to the lower extraction potential that was used to operate the
two thrusters. Thruster 2 was operated at ~1800V until it shorted due to liquid bridging. MTT-2
was operated at ~1550V extraction.
Monomer thrust of MTT with a previously baselined UTT emitter chip showed a
substantial increase of 30% while monomer/dimer thrust (at the expected 30% monomers and
70% dimers) was increased by greater than 13%. This increase in thrust can be partially
attributed to lower loss due to interception but indicates a broader overall increase in the number
of emitted ions. In comparison to conventional thrusters the MTT shows growth potential in
scalability as a function of the radius of the Helmholtz coil instead of the flatness and stability of
the extractor. Multiplexing will not be necessary with larger emitter arrays, reducing overall
complexity. The MTT eliminates the primary driver to conventional thruster service life
degradation and failure, single emitter problems (such as arching or liquid bridging) will no
longer render an entire emitter array defunct. The MTT shows a robustness for further
development and growth potential to warrant future investigation.

90
Future Work

A study on the MTT with a non-planar emitter should be performed. This would illustrate
further utility of the technology. The flatness of the emitter should not detrimentally affect the
extraction of ions from its surface if the tip sharpness relative to Lenz force induction is
maintained. Off axis thrust and expanded exhaust plume could be limited by only varying the tip
height relative to the extraction plane and not the tip orientation or the tip parallelism. A concave
emitter surface theoretically would focus thrust in the axial direction and eliminate losses due to
beam spreading. COMSOL Modeling of this proposal shows no discernable difference in
extraction potential so long as the tips are distributed in the central region between the Helmholtz
coils [97].
Scalability of the MTT should be shown: A large 4cm MTT has been additively
manufactured, this MTT should be wound and tested to demonstrate scalability beyond that of
standard electrospray. This thruster would require additional ACCEL drivers as the drive current
will need to be upped to 6 Amps to induce significant enough current in the larger conducting
surface area. This thruster will exceed the test capabilities of the LAPD facility and will saturate
the small chamber. Another option would be A micromachining of emitters using MEMS with a
pitch of ~30𝜇𝜇 m. Nick Christiansen’s Advanced Microfabricated Electrospray Propulsion System
(AMPS) proposal [98] devised an electrospray with 32 𝜇𝜇 m pitch with an integrated extractor
(Figure 80). The demonstrated emitter tip fabrication could be performed without the integrated

91

Figure 80: AMPS Thruster depiction with integrated extractor

extractor that was described in the proposal and allow integration into the existing MTT-3
thruster. The fabrication process for the emitters would enable a full order of magnitude
improvement in pitch and a greater than 300-fold projected increase in thrust of the 3.3x3.3mm
emitter area. Figure 81 shows the increase in thrust on a logarithmic scale. For comparative
results the area of the emission surface was not changed, the emission area for all thrusters and
proposed technological improvements is 3.3x3.3mm on a 1 cm disk. Further improvements could
be realized if the entire surface area of the 1 cm emitter chip was utilized but this would skew the
comparison and was not addressed.

Figure 81: Thruster performance with extrapolated growth for 200 𝜇𝜇 m pitch and 35 𝜇𝜇 m pitch tip density

92
Exponential growth to the space charge limit of each individual tip can be achieved with
the removal of the extractor. Provided the substrate porosity can support the tip geometry
multiplexed arrays can be created with exceedingly small tips located closely together. A 4 cm
MTT with 35 𝜇𝜇 m pitch with comparable 1500V extraction and ~3000-5000s of ISP would
produce 10-16mn of thrust. This thrust density on a ~1U size thruster would take the pure ionic
electrospray from micro stabilizing thrust levels to orbit maintenance and avoidance utility
range.

93
Appendix A: Electrospray Thruster Mathematics Principles.

A.1 Fundamental Physics of Electrospray Operation
The voltage required to accelerate ions is given by 𝑉𝑉 =
𝑚𝑚 𝑐𝑐 𝑖𝑖 2
2 𝑞𝑞 where ci is the velocity of the ions,
m is their mass and q the charge. The ion velocity is given by

Emitted current (I) is a function of the number density N of the emitted ions and their respective
charge states (q). For PIR Electrosprays the charge states are expected to be 1/-1.

The Starting voltage for an electrospray relies on the fluid properties and the shape of the
electrode. The radius of curvature (Rc) acts as a lens based on the distance (ds) to between the anode
and cathode of the emitter.

𝑉𝑉 𝑡𝑡 ℎ
≅ �
𝛾𝛾 𝑅𝑅 𝑐𝑐 𝜖𝜖 0
� 𝑙𝑙𝑙𝑙 �
4 𝑑𝑑 𝑠𝑠 𝑅𝑅 𝑐𝑐 � A3

The maximum radius of curvature for emission can be estimated to be:
𝑅𝑅 𝑐𝑐 ≅
4 𝛾𝛾 𝜀𝜀 0
𝐸𝐸 𝑐𝑐 2
�
𝜀𝜀 𝜀𝜀 − 1
� A4

Field-evaporated current per unit area can be found by:

𝑗𝑗 = 𝜎𝜎 𝑘𝑘𝑘𝑘
ℎ
𝑒𝑒 −
∆ 𝐺𝐺 − 𝐺𝐺 ( 𝐸𝐸 )
𝑘𝑘 𝑘𝑘 A5

Where k is the Boltzmann’s constant, h is the Planck’s constant, 𝜎𝜎 is the surface charge density,
∆ 𝐺𝐺 is the free energy of solvation for the ion specific extraction, and G(E) is the normal E field
reduction of the free energy.

A.2 Performance characteristic equations
Thrust of an electrospray can be estimated as 𝐹𝐹 𝑘𝑘 = 𝐴𝐴 ∗
𝜀𝜀 0
2
�
4 𝑉𝑉 3 𝑑𝑑 �
2
This is specific to gridded
electrosprays where d (the pitch) is a subset of the larger Emitter area (A). This thrust is a
summation of the number of ions emitted per emitter times their mass and velocity. Thrust
increases as a direct function of velocity while velocity increases as a square root of acceleration
Voltage.
𝑐𝑐 𝑖𝑖 =
�
2 𝑞𝑞 𝑖𝑖 𝑉𝑉 𝑚𝑚 𝑗𝑗 A1

𝐼𝐼 𝑒𝑒 𝑚𝑚 = ∑ 𝑁𝑁 ̇ 𝑖𝑖 𝑞𝑞 𝑖𝑖 i
A2

94

ISP can be estimated using the mass spectrum derived from TOF measurements. Current
utilized from the power supply and intercepted current at the extractor are used to create a reliable
estimation for emitted current. Acceleration voltage from the distal electrode is utilized as the
measure of potential. The extractor is held close to ground through the data system.

The plume efficiency is a function of the mass and charge difference between monomers and
dimers in each polarity of operation. The thruster 𝜂𝜂 𝑝𝑝 efficiency is a function thruster drive and
plume current.

𝜖𝜖 ≡
( 𝑞𝑞 𝑚𝑚 ⁄ )
2
( 𝑞𝑞 𝑚𝑚 ⁄ )
1
𝐴𝐴 8

𝜷𝜷 𝟏𝟏 ≡
𝐼𝐼 0
𝐼𝐼 0
− 𝐼𝐼 1
=
𝐼𝐼 1
𝐼𝐼 0
. 𝜷𝜷 𝟐𝟐 ≡
𝐼𝐼 2
𝐼𝐼 0
A9

𝜂𝜂 𝑝𝑝 =
� 1 − � 1 − √ 𝜖𝜖 � 𝛽𝛽 2
�
2
1 −( 1 − 𝜖𝜖 ) 𝛽𝛽 2
=
𝐹𝐹 2
2 𝑚𝑚 ̇ 𝐼𝐼 𝑉𝑉 A10

An estimation of the thruster plume interception can be made with the expected beam angle
and the thruster geometry. This intercepted current is measured through the extractor ground line
and can be used to make assessments of the true plume angle.

%𝐼𝐼 =
� (t an(𝜃𝜃 𝑏𝑏 )∗ 𝑑𝑑 𝑠𝑠 )+
𝑡𝑡 𝑒𝑒𝑒𝑒 𝑡𝑡 2
∗ t a n 𝜃𝜃 𝑏𝑏 2
� −
𝐷𝐷 𝑒𝑒𝑒𝑒 𝑡𝑡 2
(𝐷𝐷 𝑒𝑒𝑒𝑒 𝑡𝑡 )/ 2
A11

𝐹𝐹 𝑘𝑘 = ∑ 𝑁𝑁 ̇ 𝑖𝑖 m
𝑖𝑖 c
i

i
A6

𝐼𝐼 𝑆𝑆 𝑆𝑆 =
1
𝑔𝑔
�
2 𝑉𝑉 𝑞𝑞 1
𝑚𝑚 1
� 1 − � 1 − √ 𝜖𝜖 � 𝛽𝛽 2
�
2
1 − (1 − 𝜖𝜖 ) 𝛽𝛽 2
=
1
𝑔𝑔
�
2 𝑉𝑉 𝑞𝑞 1
𝑚𝑚 1
𝜂𝜂 𝑝𝑝 𝐴𝐴 7

95
Appendix B: BNG Development

This work was performed by an undergraduate research team for AME 441 under
guidance from myself and was continued after graduation by Kevin Sampson and I in the Fall
semester of 2020.
A BNG utilized for TOF spectroscopy consists of N number of parallel wire sets with a
varying binary potential difference of 0V “open” and VBNG “closed.” To minimize induced noise
and design a functional Bradbury Nielsen Gate, the relationship between operating voltage, wire
diameter, and wire spacing was optimized using Eq B1 and B2. Both dwire and dspace could be
varied to satisfy requirements for an operating voltage, VBNG, fixed under 600 V with 75%
transparency. These parameters are set to produce the deflection angle, 𝛂𝛂 . The relative spacing
between the BNG and thruster fed to the need for an 𝛂𝛂 > 7.6°. Figure B1 shows the electric field
generated by VBNG with theoretical ion beam deflection behavior.

tan(𝛼𝛼 ) =
2 𝐸𝐸 𝑘𝑘 𝑖𝑖 𝑘𝑘 𝑉𝑉 𝐵𝐵 𝐵𝐵 𝐺𝐺 ln(cot(
𝜋𝜋 𝑑𝑑 𝑠𝑠𝑠𝑠 𝑠𝑠𝑠𝑠 𝑒𝑒 4 𝑑𝑑 𝑤𝑤 𝑖𝑖 𝑤𝑤 𝑒𝑒 ))
𝜋𝜋 𝑞𝑞 B1

𝐸𝐸 𝑘𝑘 𝑖𝑖 𝑘𝑘 =
𝑚𝑚 𝑣𝑣 2
2
B2

Figure B1: Cross-sectional view of charged BNG wires (after reaching steady-state), illustrating deflected
monomers of EMI or BF4 ion trajectories at αth = 20.0°.

The optimum BNG wire spacing and voltage were computed to be 1 ± .02 mm, 600 ± 2.8 V
respectively by using a readily available 30 AWG gauge wire and a thruster performance
window with an initially assumed 50% monomer and dimer plume.

1. The frame’s thermal expansion is directly correlated to further wire deformation.
Therefore, it was necessary to consider the effect of ambient laboratory temperature

96
fluctuations on the frame’s thermal expansion. For a 5.08 x 5.08 cm inner window, this
expansion created enough slack for adjacent wire contact when the device returned to its
initial temperature. (Although Delrin’s thermal expansion proved unsuitable for these
inner window dimensions, this factor may be negligible for scaled-down versions.)

These findings motivated the use of PolyEther Ether Ketone (PEEK) for the frame material due
to its lower CTE. A comparison of both materials is shown below in Table B1.

Table B1 Frame behavior considering temperature fluctuations of considered BNG
materials.
Material CTE
[cm/cm/C]
Initial Material
Length, [cm]
DeltaT
[ºC]
Thermal Frame
Deformation
[cm]
Delrin 1.23E-4 5.08 5.6 3.45E-3
PEEK 4.68E-5 1.32E-3

Adjacent wire contact may occur with wire deformations of at least 1.23E-3 cm. It is important
to note that this value represents a worst-case condition, where adjacent wires arc directly
towards each other.

To evaluate their limits of acceptability during operation, both designs were thermally cycled
within a temperature range of 20 ± 6°C. After a temperature history of two ± 6°C cycles, results
for BNGs made from both materials are shown in Figure B2.

Figure B2 Visual comparison of Delrin (a) and PEEK (b) framed BNGs with enhanced sections to illustrate
wire contact.

The PEEK-framed BNG underwent slight wire deformations without compromising the BNG
design or wire spacing. One wire was slightly deformed, without adjacent wire contact, resulting

97
in 98% wire parallelism. In comparison, Delrin underwent significant wire deformation, resulting
in adjacent wire contact, deeming it unsuitable. The lifetime of this PEEK design currently
surpasses two months at the time of writing.

Lessons Learned

The PEEK-based frame proved much more robust than the Delrin-based frame. This can
be attributed mainly to the lower CTE of PEEK and uniformly tensioned wires. For the PEEK-
based BNG, constant wire loads ensured that wire behavior was consistently held under ultimate
stress. In contrast, previous Delrin-based BNGs relied on manual tensioning, which caused an
unknown, non-uniform tension distribution among the wires that resulted in wire deformation.
While the wire tension (for the PEEK-based frame) was held at 109.7 MPa (which surpassed the
yield stress of 70 MPa), the wire stress was maintained below ultimate stress (220 MPa). In that
range, some chains of the annealed copper molecule chains remained unbroken, leaving a portion
of elastic forces at play in the wires. In an ideal scenario, wires should be tensioned under the
yield stress to avoid deformation. This would allow for absolute wire elasticity when the frame
thermally contracts, making it more practical for daily use and storage. Throughout daily
ambient temperature fluctuations of ± 10° F, as well as countless exposures to both vacuum and
atmospheric conditions, the gate’s wire parallelism has persisted for over 11 months at the time
of writing.

Beam Spread Angle

A Retarding Potential Analyzer (RPA), on a three-dimensional traverse system within the
vacuum chamber, scanned across the span of the thruster’s ion plume. Starting from the
thruster’s centerline (where x-distance is zero), the RPA scans occurred at a distance from the
thruster of 3.0 in and 4.8 in, respectively. Although only one side of the plume is swept, it is
assumed that the plume behavior is axisymmetric. During these scans, 500V was applied to the
BNG, deflecting the ion beam away from the centerline, as shown in Figure B3 below.

98

Figure B3: Transverse probe sweep results with a zero-current baseline for comparison. From these results,
a deflection angle of =9.6°was calculated.

The RPA’s intercepted current (dscan) was expected to be zero near the centerline, increasing
when entering the region of the deflected plume at some x-distance beyond this centerline. This
plot demonstrates the device's efficacy, displays a general behavior of the system, and confirms
that a necessary deflection angle for clearing the Faraday plate was obtained. While some signal
remained at the centerline, the signal reduction is significant enough to resolve at the collector
plate.

Gate Efficiency at Different Voltages

To observe BNG performance at different setpoints, baseline thruster readings were taken
for approximately 20 seconds. The thruster operated at a relatively steady baseline current for the
duration of each test. When the BNG was activated, deflecting the ion beam, the collected
current was reduced at the Faraday Plate. By setting the gate voltage to 420V, 530V, 605V, the
signal readings dropped by 40%, 70%, and 95%, respectively. To illustrate the effect of VBNG on
the signal response, current readings were taken, normalized, and are shown below in Figure
B4. Theoretical reduction lines for the various collected current were calculated based on the
Thrusters 36-degree plume half angle caused by a relatively high extraction voltage of 2242V.
The theoretical reduction for BNG voltages of 420V, 530V, and 605V were 37%, 66%, and 93%,
respectively. Due to the inconsistent nature of the electrospray thruster used in this experiment,
the collected current emitted from the thruster varies slightly with time.

99

Figure B4: CP current at various BNG voltages

A large format BNG tailored for electrospray operation is a less destructive, more passive
time of flight spectroscopy ion diverting device than the current academic standard of using a
conventional ion repulsion gate. While total BNG ion diversion efficiency at the TOF collection
plane is 66-93% of the total beam as compared to 99% with a conventional gate, the increased
current collected due to the improved transparency of the BNG overcomes this limitation. If
during initial construction, wire tension and expected thermal cycling are properly matched, the
BNG shows long term utility and parallelization stability, after being vacuum cycled hundreds of
times in a standard lab environment. Returning current to the thrusters emitting surface was
eliminated by the use of the BNG. As return current can damage the thruster emitters and
extractor by causing shorts or high energy arc discharges, this advantage will increase long term
measurement accuracy as well as thruster lifetime during investigation. The virtual elimination
of return current, reflected back on the operational thrusters, by the BNG should make this
device the industry standard for TOF-S of electrospray thrusters.

100
Bibliography

[1] Lord Rayleigh, “On the equilibrium of liquid conducting masses charged with electricity,”
London, Edinburgh, Dublin Philos. Mag. J. Sci., vol. 14, no. 87, pp. 184–186, 1882, doi:
10.1080/14786448208628425.
[2] J. Zeleny, “The electrical discharge from liquid points, and a hydrostatic method of
measuring the electric intensity at their surfaces,” Phys. Rev., vol. 3, no. 2, pp. 69–91,
1914, doi: 10.1103/PhysRev.3.69.
[3] G. Taylor, “Disintegration of Water Drops in an Electric,” Proc. R. Soc. London. A. Math.
Phys. Sci., vol. 280, no. 1382, pp. 383–397, 1964.
[4] G. Taylor, “Electrically driven jets,” Proc. R. Soc. London. A. Math. Phys. Sci., vol. 313,
no. 1515, pp. 453–475, 1969, doi: 10.1098/rspa.1969.0205.
[5] I. Romero-Sanz, R. Bocanegra, J. Fernandez de la Mora, and M. Gamero-Castaño,
“Source of heavy molecular ions based on Taylor cones of ionic liquids operating in the
pure ion evaporation regime,” J. Appl. Phys., vol. 94, no. 5, pp. 3599–3605, 2003, doi:
10.1063/1.1598281.
[6] D. Krejci and P. Lozano, “Space Propulsion Technology for Small Spacecraft,” Proc.
IEEE, vol. 106, no. 3, pp. 362–378, 2018, doi: 10.1109/JPROC.2017.2778747.
[7] D. Krejci and P. Lozano, “Micro-machined ionic liquid electrospray thrusters for Cubesat
applications,” Proc. 35th Int. Electr. Propuls. Conf., no. October, 2017.
[8] M. R. Natisin and H. L. Zamora, “Performance of a Fully Conventionally Machined
Liquid-Ion Electrospray Thruster Operated in PIR,” Int. Electr. Propuls. Conf. 2019, no.
19443, pp. 1–16, 2019.
[9] C. Coffman, L. Perna, H. Li, and P. C. Lozano, “On the manufacturing and emission
characteristics of a novel borosilicate electrospray source,” 49th AIAA/ASME/SAE/ASEE
Jt. Propuls. Conf., vol. 1 PartF, pp. 1–12, 2013, doi: 10.2514/6.2013-4035.
[10] D. G. Courtney, S. Dandavino, and H. Shea, “Comparing direct and indirect thrust
measurements from passively fed ionic electrospray thrusters,” J. Propuls. Power, vol. 32,
no. 2, pp. 392–407, 2016, doi: 10.2514/1.B35836.
[11] Daniel George Courtney, “Ionic Liquid Ion Source Emitter Arrays Fabricated on Bulk
Porous Substrates for Spacecraft Propulsion,” J. Chem. Inf. Model., vol. 53, no. 9, pp.
1689–1699, 2013, doi: 10.1017/CBO9781107415324.004.
[12] M. N. Huberman, J. C. Beynon, E. Cohen, D. S. Goldin, P. W. Kidd, and S. Zafran,
“Present status of colloid microthruster technology,” J. Spacecr. Rockets, vol. 5, no. 11,

101
pp. 1319–1324, 1968, doi: 10.2514/3.29476.
[13] E. Cohen and M. H. Huberman, “Research on charged particle electrostatic thrustors,”
Technical Report, Air Force Aero Propulsion Lab, AFAPL-TR-66-94, 1966.
https://www.google.com/search?q=Cohen%2C+E.+and+Huberman%2C+M.+H.%2C+%2
2Research+on+charged+particle+electrostatic+thrustors%2C%22+Technical+Report%2C
+Air+Force+Aero+Propulsion+Lab.%2C+1966%2C+AFAPL-TR-66-
94&rlz=1C5CHFA_enDE867DE867&oq=Cohen%2C+E.+and+Hub (accessed Sep. 23,
2020).
[14] J. Nabity and J. Daily, “Effect of ionic liquid composition on colloid thruster emission and
thrust performance,” J. Propuls. Power, vol. 34, no. 1, 2018, doi: 10.2514/1.B36376.
[15] P. Wang, A. Borner, B. Korkut, Z. Li, and D. A. Levin, “Simulations of electrospray in a
colloid thruster with high resolution particle-in-cell method,” 44th AIAA Plasmadynamics
Lasers Conf., 2013, doi: 10.2514/6.2013-2629.
[16] A. B. McEwen, “Electrochemical Properties of Imidazolium Salt Electrolytes for
Electrochemical Capacitor Applications,” J. Electrochem. Soc., vol. 146, no. 5, p. 1687,
1999, doi: 10.1149/1.1391827.
[17] D. G. COURTNEY and P. LOZANO, “Characterization of Conical Ionic Liquid Ion
Sources for 2-D Electrospray Thruster Arrays on Porous Substrates,” Trans. Japan Soc.
Aeronaut. Sp. Sci. Aerosp. Technol. Japan, vol. 8, no. ists27, p. Pb_73-Pb_78, 2010, doi:
10.2322/tastj.8.pb_73.
[18] M. S. Alexander, K. L. Smith, M. D. Paine, and J. P. W. Stark, “Voltage-modulated flow
rate for precise thrust control in colloid electrospray propulsion,” J. Propuls. Power, vol.
23, no. 5, pp. 1042–1048, May 2007, doi: 10.2514/1.25309.
[19] P. Lezono and P. C. Lozano, “Studies on the ion-droplet mixed regime in colloid
thrusters,” Physics (College. Park. Md)., p. 222, 2003, [Online]. Available:
http://ssl.mit.edu/publications/theses/PhD-2003-Lozano-TovarPaulo.pdf.
[20] J. Ziemer et al., “Colloid Microthruster Flight Performance Results from Space
Technology 7 Disturbance Reduction System.”
[21] G. Anderson et al., “Experimental results from the ST7 mission on LISA Pathfinder,”
Phys. Rev. D, vol. 98, no. 10, p. 102005, Nov. 2018, doi: 10.1103/PhysRevD.98.102005.
[22] D. Courtney and others, “Busek SmallSat Technologies,” Planet. CubeSats Symp., 2018.
[23] M. S. Alexander, J. Stark, K. L. Smith, B. Stevens, and B. Kent, “Electrospray
performance of microfabricated colloid thruster arrays,” J. Propuls. Power, vol. 22, no. 3,
pp. 620–627, 2006, doi: 10.2514/1.15190.

102
[24] P. L. Wright, H. Huh, and M. J. Davis, “Assessment of Grid Impingement for
Electrospray Thruster Lifetime,” 36th Int. Electr. Propuls. Conf., no. September, pp. 1–18,
2019.
[25] N. M. Uchizono et al., “Emission Modes in Electrospray Thrusters Operating with High
Conductivity Ionic Liquids,” Aerospace, vol. 7, no. 10, p. 141, Sep. 2020, doi:
10.3390/aerospace7100141.
[26] F. Mier-Hicks and P. C. Lozano, “Spacecraft-Charging characteristics induced by the
operation of electrospray thrusters,” J. Propuls. Power, vol. 33, no. 2, pp. 456–467, 2017,
doi: 10.2514/1.B36292.
[27] A. Kramer, P. Bangert, and K. Schilling, “UWE-4: First electric propulsion on a 1U
cubesat-in-orbit experiments and characterization,” Aerospace, vol. 7, no. 7, pp. 1–21,
2020, doi: 10.3390/AEROSPACE7070098.
[28] I. Vasiljevich et al., “Development of an Indium mN-FEEP thruster,” 2008, doi:
10.2514/6.2008-4534.
[29] M. Tajmar, A. Genovese, and W. Steiger, “Indium Field Emission Electric Propulsion
Microthruster Experimental Characterization,” J. Propuls. Power, vol. 20, no. 2, pp. 211–
218, May 2004, doi: 10.2514/1.9247.
[30] M. Tajmar, I. Vasiljevich, and W. Grienauer, “High current liquid metal ion source using
porous tungsten multiemitters,” Ultramicroscopy, vol. 111, no. 1, pp. 1–4, Dec. 2010, doi:
10.1016/j.ultramic.2010.09.005.
[31] M. Tajmar, A. Genovese, N. Buldrini, and W. Steiger, “Miniaturized indium-FEEP
multiemitter design and performance,” 2002, doi: 10.2514/6.2002-5718.
[32] D. Jelem, A. Reissner, B. Seifert, N. Buldrini, L. Wilding, and D. Krejci, “Direct thrust
and plume divergence measurements of the IFM Nano Thruster,” Adv. Sp. Res., vol. 62,
no. 12, pp. 3398–3404, 2018, doi: 10.1016/j.asr.2018.06.028.
[33] L. Biagioni, F. Ceccanti, M. Saverdi, M. Saviozzi, and M. Andrenucci, “Qualification
Status of the FEEP-150 Electric Micropropulsion Subsystem,” Jul. 2005, doi:
10.2514/6.2005-4261.
[34] B. Seifert, A. Reissner, D. Jelem, and T. Hörbe, “Development of A Low Cost PPU for
Feep Electric Propulsion Using Cots Components,” E3S Web Conf., vol. 16, pp. 1–7,
2017, doi: 10.1051/e3sconf/20171615003.
[35] D. Krejci, A. Reissner, T. Schoenherr, B. Seifert, Z. Saleem, and R. Alejos, “Recent flight
data from IFM Nano Thrusters in a low earth orbit,” Int. Electr. Propuls. Conf., no.
October, p. IEPC-2019-A724, 2019.

103
[36] P. Lozano and M. Martinez-Sanchez, “Efficiency estimation of EMI-BF 4 ionic liquid
electrospray thrusters,” 41st AIAA/ASME/SAE/ASEE Jt. Propuls. Conf. Exhib., no. July,
pp. 1–7, 2005, doi: 10.2514/6.2005-4388.
[37] S. Zhang, N. Sun, X. He, X. Lu, and X. Zhang, “Physical properties of ionic liquids:
Database and evaluation,” J. Phys. Chem. Ref. Data, vol. 35, no. 4, pp. 1475–1517, 2006,
doi: 10.1063/1.2204959.
[38] D. G. Courtney and H. Shea, “Influences of porous reservoir Laplace pressure on
emissions from passively fed ionic liquid electrospray sources,” Appl. Phys. Lett., vol.
107, no. 10, pp. 2–7, 2015, doi: 10.1063/1.4930231.
[39] L. F. Velásquez-García et al., “A planar array of micro-fabricated electrospray emitters for
thruster applications,” J. Microelectromechanical Syst., vol. 15, no. 5, pp. 1272–1280,
2006, doi: 10.1109/JMEMS.2006.879710.
[40] C. Guerra-Garcia, D. Krejci, and P. Lozano, “Spatial uniformity of the current emitted by
an array of passively fed electrospray porous emitters,” J. Phys. D. Appl. Phys., vol. 49,
no. 11, 2016, doi: 10.1088/0022-3727/49/11/115503.
[41] P. Lozano and M. Martínez-Sánchez, “Ionic liquid ion sources: Characterization of
externally wetted emitters,” J. Colloid Interface Sci., vol. 282, no. 2, pp. 415–421, 2005,
doi: 10.1016/j.jcis.2004.08.132.
[42] B. Gassend, L. F. Veĺasquez-Garćia, A. I. Akinwande, and M. Mart́inez-Śanchez, “A
microfabricated planar electrospray array ionic liquid ion source with integrated
extractor,” J. Microelectromechanical Syst., vol. 18, no. 3, pp. 679–694, 2009, doi:
10.1109/JMEMS.2009.2015475.
[43] D. Krejci, F. Mier-Hicks, R. Thomas, T. Haag, and P. Lozano, “Emission characteristics
of passively fed electrospray microthrusters with propellant reservoirs,” J. Spacecr.
Rockets, vol. 54, no. 2, pp. 447–458, Jan. 2017, doi: 10.2514/1.A33531.
[44] P. Lozano and D. G. Courtney, “Ionic Liquid Ion Source Emitter Arrays Fabricated on
Bulk Porous Substrates for Spacecraft Propulsion,” 2011.
[45] R. S. Legge, “Fabrication and Characterization of Porous Metal Emitters for Electrospray
Applications C>’,” Massachusetts Institute of Technology, 2006. Accessed: Sep. 23,
2020. [Online]. Available: https://dspace.mit.edu/handle/1721.1/45301.
[46] D. G. C. and Z. W. and S. G. and J. Model§, “High-Speed Transient Characterization of
the Busek BET-300-P Electrospray Thruster,” pp. 1–18, 2019.
[47] T. F. Daniel G. Courtney, Zackary Wood, “Reconstructing Electrospray Plume Current
Spatial Distributions using Computed Tomography,” 36th Int. Electr. Propuls. Conf., p.

104
16, 2019.
[48] E. M. Petro, A. R. Bruno, P. C. Lozano, L. E. Perna, and D. S. Freeman, “Characterization
of the tile electrospray emitters,” AIAA Propuls. Energy 2020 Forum, pp. 1–13, 2020, doi:
10.2514/6.2020-3612.
[49] M. R. Gilpin, M. R. Natisin, C. Marin, M. D. Eschen, and H. L. Zamora, “Dual-Axis
Torsional Thrust Stand for Simultaneous Direct Measurement of Thrust and Mass Loss,”
pp. 1–13, 2019.
[50] M. R. Natisin et al., “Fabrication and characterization of a fully conventionally machined,
high-performance porous-media electrospray thruster,” J. Micromechanics
Microengineering, vol. 30, no. 11, Sep. 2020, doi: 10.1088/1361-6439/abb8c3.
[51] C. Ma and C. N. Ryan, “The Design and Characterization of a Porous-emitter
Electrospray Thruster (PET-100) for Interplanetary CubeSats,” Proceeding 7th
Interplanet. CubeSat Work., no. May, pp. 1–9, 2018.
[52] A. Reddy Tummala and A. Dutta, “An Overview of Cube-Satellite Propulsion
Technologies and Trends,” doi: 10.3390/aerospace4040058.
[53] M. Gomez Jenkins, D. Krejci, and P. Lozano, “CubeSat constellation management using
Ionic Liquid Electrospray Propulsion,” Acta Astronaut., vol. 151, no. June, pp. 243–252,
2018, doi: 10.1016/j.actaastro.2018.06.007.
[54] N. Rasmont, E. J. Broemmelsiek, and J. L. Rovey, “Linear burn rate of green ionic liquid
multimode monopropellant,” Combust. Flame, vol. 219, pp. 212–224, 2020, doi:
10.1016/j.combustflame.2020.04.014.
[55] Ahmed E. S. Nosseir, Angelo Cervone, and Angelo Pasini, “Review of State-of-the-Art
Green Monopropellants: For Propulsion Systems Analysts and Designers,” Aerospace,
vol. 8, no. 1, p. 20, 2021, doi: 10.3390/aerospace8010020.
[56] M. J. Wainwright, J. L. Rovey, S. W. Miller, B. D. Prince, and S. P. Berg,
“Hydroxylammonium Nitrate Species in a Monopropellant Electrospray Plume,” J.
Propuls. power, vol. 35, no. 5, pp. 922–929, 2019, doi: 10.2514/1.B37471.
[57] H. Niedermeyer, J. P. Hallett, I. J. Villar-Garcia, P. A. Hunt, and T. Welton, “Mixtures of
ionic liquids,” Chem. Soc. Rev., vol. 41, no. 23, pp. 7780–7802, Nov. 2012, doi:
10.1039/c2cs35177c.
[58] M. J. Wainwright and J. L. Rovey, “Effect of nonlinear mixing on electrospray propulsion
predictions,” Journal of Propulsion and Power, vol. 37, no. 1. AIAA International, pp.
167–170, 2021, doi: 10.2514/1.B38045.
[59] G. Chatel, J. F. B. Pereira, V. Debbeti, H. Wang, and R. D. Rogers, “Mixing ionic liquids-

105
"simple mixtures" or ‘double salts’?,” Green Chemistry, vol. 16, no. 4. Royal Society of
Chemistry, pp. 2051–2083, 2014, doi: 10.1039/c3gc41389f.
[60] V. H. Paschoal, L. F. O. Faria, and M. C. C. Ribeiro, “Vibrational Spectroscopy of Ionic
Liquids,” Chem. Rev., vol. 117, no. 10, pp. 7053–7112, 2017, doi:
10.1021/acs.chemrev.6b00461.
[61] C. Coffman, “Electrically-Assisted Evaporation of Charged Fluids: Fundamental
Modeling and Studies on Ionic Liquids,” 2016.
[62] I. Funaki, H. Kuninaka, K. Toki, Y. Shimizu, K. Nishiyama, and Y. Horiuchi,
“Verification tests of carbon-carbon composite grids for microwave discharge ion
thruster,” J. Propuls. Power, vol. 18, no. 1, pp. 169–175, May 2002, doi: 10.2514/2.5913.
[63] J. Yuan, P. Dai, D. Liang, S. Zhou, S. Xiao, and X. Liang, “Grid deformation real-time
measurement system of Ion thruster based on videometrics,” Appl. Sci., vol. 9, no. 9,
2019, doi: 10.3390/app9091759.
[64] J. E. Polk et al., “An overview of the results from an 8200 hour wear test of the nstar ion
thruster,” 1999, doi: 10.2514/6.1999-2446.
[65] A. Sengupta et al., “Deep space 1 flight spare ion thruster 30,000-hour life test,” J.
Propuls. Power, vol. 25, no. 1, pp. 105–117, May 2009, doi: 10.2514/1.36549.
[66] J. R. Anderson, K. D. Goodfellow, J. E. Polk, V. K. Rawlin, and J. S. Sovey,
“Performance characteristics of the NSTAR ion thruster during an on-going long duration
ground test,” IEEE Aerosp. Conf. Proc., vol. 4, pp. 99–121, 2000, doi:
10.1109/AERO.2000.878371.
[67] R. I. Samanta Roy, D. E. Hastings, and N. A. Gatsonis, “Ion-thruster plume modeling for
backflow contamination,” J. Spacecr. Rockets, vol. 33, no. 4, pp. 525–534, May 1996, doi:
10.2514/3.26795.
[68] J. L. Van Noord and D. A. Herman, “Application of the NEXT ion thruster lifetime
assessment to thruster throttling,” 2008, doi: 10.2514/6.2008-4526.
[69] G. R. Schmidt, nasagov J. Michael Patterson, and S. W. Benson, “IAC-08-C4.4.2 THE
NASA EVOLUTIONARY XENON THRUSTER (NEXT): THE NEXT STEP FOR U.S.
DEEP SPACE PROPULSION.”
[70] M. SANGREGORIO, K. XIE, N. WANG, N. GUO, and Z. ZHANG, “Ion engine grids:
Function, main parameters, issues, configurations, geometries, materials and fabrication
methods,” Chinese Journal of Aeronautics, vol. 31, no. 8. Chinese Journal of Aeronautics,
pp. 1635–1649, Aug. 01, 2018, doi: 10.1016/j.cja.2018.06.005.
[71] R. J. Antypas and J. J. Wang, “Pure Ionic Electrospray Extractor Design Optimization,”

106
Int. Electr. Propuls. Conf., pp. 1–9, 2019.
[72] A. Thuppul, P. L. Wright, and R. E. Wirz, “Lifetime considerations and estimation for
electrospray thrusters,” 2018 Jt. Propuls. Conf., 2018, doi: 10.2514/6.2018-4652.
[73] S. Ciurlo, A. Mariscotti, and A. Viacava, “A Helmholtz coil for high frequency high field
intensity applications,” 16th IMEKO TC4 Int. Symp. Explor. New Front. Instrum. Methods
Electr. Electron. Meas. 13th TC21 Int. Work. ADC Model. Test. - Jt. Sess. Proc., pp. 842–
846, 2008.
[74] J. J. Abbott, “Parametric design of tri-axial nested Helmholtz coils,” Rev. Sci. Instrum.,
vol. 86, no. 5, pp. 1–10, 2015, doi: 10.1063/1.4919400.
[75] “High frequency Helmholtz coils generate magnetic fields - EDN.”
https://www.edn.com/high-frequency-helmholtz-coils-generate-magnetic-fields/ (accessed
Mar. 05, 2021).
[76] A. Haghnegahdar et al., “Design and fabrication of helmholtz coils to study the effects of
pulsed electromagnetic fields on the healing process in periodontitis: preliminary animal
results.,” J. Biomed. Phys. Eng., vol. 4, no. 3, pp. 83–90, 2014, [Online]. Available:
http://www.ncbi.nlm.nih.gov/pubmed/25505775%0Ahttp://www.pubmedcentral.nih.gov/a
rticlerender.fcgi?artid=PMC4258865.
[77] T. S. Tenforde, “Magnetically induced electric fields and currents in the circulatory
system,” in Progress in Biophysics and Molecular Biology, Feb. 2005, vol. 87, no. 2-3
SPEC. ISS., pp. 279–288, doi: 10.1016/j.pbiomolbio.2004.08.003.
[78] B. Wang, W. M. Grill, and A. V. Peterchev, “Coupling Magnetically Induced Electric
Fields to Neurons: Longitudinal and Transverse Activation,” Biophys. J., vol. 115, no. 1,
pp. 95–107, Jul. 2018, doi: 10.1016/j.bpj.2018.06.004.
[79] “High-Frequency Electromagnetic Coil Resonant.”
https://www.accelinstruments.com/Applications/WaveformAmp/Electromagnetic-Coil-
Resonant.html (accessed Mar. 05, 2021).
[80] John L. Polansky, “LABORATORY INVESTIGATIONS OF THE NEAR SURFACE
PLASMA FIELD AND CHARGING AT THE LUNAR TERMINATOR,” 2013.
[81] W. Yu, “NUMERICAL AND EXPERIMENTAL INVESTIGATIONS OF DUST-
PLASMA-ASTEROID,” USC Diss., no. August, 2018.
[82] K. Chou, “EXPERIMENTAL INVESTIGATION ON DUSTY SURFACE CHARGING
IN PLASMA,” no. December, 2017.
[83] G. Members, J. Dipaola, W. Fackelman, and J. Young, “Integration of the National
Instruments MyRIO FPGA & LabVIEW software with a 3D Printer,” p. 17, 2012.

107
[84] Vipul Kumar Sharma, “System Integration & Programming of EHD 3D printer,” Drexel
Univ., no. June, 2017.
[85] D. L. Brown and A. D. Gallimore, “Evaluation of ion collection area in Faraday probes,”
Rev. Sci. Instrum., vol. 81, no. 6, 2010, doi: 10.1063/1.3449541.
[86] D. L. Brown, M. L. R. Walker, J. Szabo, W. Huang, and J. E. Foster, “Recommended
practice for use of faraday probes in electric propulsion testing,” J. Propuls. Power, vol.
33, no. 3, pp. 582–613, 2017, doi: 10.2514/1.B35696.
[87] D. A. Lister, “Investigation of Langmuir Probes in Non-Maxwellian Plasma Using
Particle-in-Cell (Pic) Modeling,” 2016.
[88] Z. Zhang, H. Tang, Z. Zhang, J. Wang, and S. Cao, “A retarding potential analyzer design
for keV-level ion thruster beams,” Rev. Sci. Instrum., vol. 87, no. 12, p. 123510, Dec.
2016, doi: 10.1063/1.4972345.
[89] D. G. Courtney and H. Shea, “Fragmentation in time-of-flight spectrometry-based
calculations of ionic electrospray thruster performance,” J. Guid. Control. Dyn., vol. 38,
no. 6, pp. 1500–1504, 2015, doi: 10.2514/1.B35837.
[90] E. Grustan-Gutierrez and M. Gamero-Castaño, “Time-of-Flight Characterization of a
Microfabricated Electrospray Thruster Emitter Array,” Proc. 35th IEPC, October, 2017,
Georg. Inst. Technol. USA, no. December, pp. 1–6, 2017.
[91] R. Antypas, K. Sampson, D. Torre, and J. Wang, “Large Bradbury Nielsen Gate
Optimized to Reduce Noise for Electrospray Time of Flight Spectroscopy,” submitted to
Journal of Propulsion and Power, 2021.
[92] L. F. Velásquez-García, M. Martinez-Sanchez, L. Fernando Velásquez-García, A. I.
Akinwande, and M. Martínez-Sánchez, “A Micro-Fabricated Linear Array of Electrospray
Emitters for Thruster Applications NASA NNC08CA58C View project MIT mini-
Helicion Thruster Experiment View project A Micro-Fabricated Linear Array of
Electrospray Emitters for Thruster Applications,” Artic. J. Microelectromechanical Syst.,
vol. 15, no. 5, 2006, doi: 10.1109/JMEMS.2006.879707.
[93] S. Dandavino et al., “Microfabricated electrospray emitter arrays with integrated extractor
and accelerator electrodes for the propulsion of small spacecraft,” J. Micromechanics
Microengineering, vol. 24, no. 7, 2014, doi: 10.1088/0960-1317/24/7/075011.
[94] M. Kaminsky, Atomic and Ionic Impact Phenomena on Metal Surfaces. Springer Berlin
Heidelberg, 1965.
[95] M. Kaminsky and M. Kaminsky, “The Nature of the Metal Surface,” in Atomic and Ionic
Impact Phenomena on Metal Surfaces, Springer Berlin Heidelberg, 1965, pp. 1–16.

108
[96] M. Kaminsky and M. Kaminsky, “Binding Forces Effective in the Collision of Atoms and
Molecules with Metal Surfaces,” in Atomic and Ionic Impact Phenomena on Metal
Surfaces, Springer Berlin Heidelberg, 1965, pp. 33–42.
[97] R. J. Antypas and J. Wang, “Magnetically Induced Extractor-Less Electrospray
Emission,” submitted to Journal of Propulsion and Power, 2021.
[98] N. Christensen, “The Advanced Microfabricated Electrospray Propulsion System
(AMPS),” private communication, 2020.

Abstract (if available)

Abstract Electrospray research has made immense progress since its infancy in the 1960抯 and its rebirth in the early 2000抯. Successfully flight tested, this technology shows great promises for high efficiency and low thrust applications, with the ultimate high thrust utility of these devices relying on the elusive promise of emitter tip densification. While the physics that enable conventional electrospray devices is simple, imposing engineering challenges still persist in the relationship between the extraction plane and the emitter tips. These engineering challenges practically limit the size and densification of conventional electrospray devices. The USC Magnetic Testbed Thruster (MTT) was developed as a passively fed, porous media, pure ionic, extractor-less electrospray thruster. The MTT eliminates the need for an extractor entirely by utilizing dynamic magnetic induction to create a virtual electric field that accelerates bipolar species of an ionic liquid. The virtual electric field that is induced in the conductive propellant takes the place of the fixed field between a conventional extractor and the emitter array thus eliminating the largest barrier to scalability and densification. ? To baseline the novel MTT performance characteristics, a state-of-the-art electrospray testing facility was constructed to evaluate conventional and novel electrospray devices. The Laboratory of Astronautical Plasma Dynamics was retrofitted with a multitude of Faraday and retarding potential analyzer probes as well as a full 3 axis automated probe sweep system with both high and low speed data acquisition. ISP and thrust were calculated with Time-of-Flight Spectroscopy utilizing a Bradbury Neilson refraction ion diversion gate instead of a conventional reflection suppression grid. This alternate approach to ion diversion illustrated drawbacks associated with the industry standard reflection approach. The differences in spectrographic results between the two methods show a clear difference in emission characteristics of thrusters under reflected investigation and not just differences in results due to the two methods, the differences can be attributed to the deleterious effects of back streaming large quantities of high energy ions on the emitter surface for extended periods. ? Four conventional configuration 25 emitter USC Testbed Thrusters (UTT) were designed/constructed and tested, utilizing EMI-BF? electrospray emission from 1cm borosilicate disks. The UTT was used to baseline the diagnostic tools in the laboratory with respect to literary peers and to determine the densification threshold of both the current extractor design as well as the theoretical density of the porous structure of the borosilicate glass substrate materiel. The UTT tested extractor configurations with pitch densities of 533?m to 432?m. The various prototyped of the UTT displayed performance ranging from 4.2?10??N to 4.5?10??N of thrust and ISP from 2300s to 6300s. Three comparable MTT prototypes were designed/constructed and tested utilizing the same 1cm 25 emitter disks used in the UTTs. Utilizing orthogonal matched Helmholtz coil pairs operating at a dynamic frequency of ~150kHz, the MTT voltage was tuned to +/? 1500V to match the optimum performance of the UTT for comparison. At the comparable extraction voltages, the MTT produces 5.5?8.6?10??N of thrust with an ISP range of 3600?5700s. Showing that removal of the extractor can eliminate the constraints of conventional electrospray thrusters such as service life and overall component size without sacrificing peak thrust and specific impulse. Currently, the sizeable unobstructed extraction window in the USC MTT is 1 cm?. This window is scalable as a function of the radius of the matched Helmholtz coils. Modeling of the prototype thruster configuration shows the promise for both non-planar (curved) and ultra-dense emitter arrays. Projections on the same 1cm extraction discs at the substrate porosity limited pitch of 200 ?m shows a two-order magnitude increase in thrust with constant specific impulse. New substrate materials can be investigated to increase this densification further which would vastly increase the potential scalability and utility of electrospray devices.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Numerical and experimental investigations of ionic electrospray thruster plume

PDF

An experimental investigation of cathode erosion in high current magnetoplasmadynamic arc discharges

PDF

Grid-based Vlasov method for kinetic plasma simulations

PDF

Laboratory investigations of the near surface plasma field and charging at the lunar terminator

PDF

Experimental investigation on dusty surface charging in plasma

Asset Metadata

Creator Antypas, Robert (author)

Core Title An investigation into magnetically induced extractor-less electrospray propulsion devices

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Astronautical Engineering

Degree Conferral Date 2021-08

Publication Date 07/24/2021

Defense Date 05/18/2021

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag densification,electric propulsion,electrospray,extractor,extractorless emission,Laboratory of Astronautical Plasma Dynamics,magnetic induction,magnetically induced Testbed Thruster,MTT,OAI-PMH Harvest,pure ion emission,satellite propulsion,USC Testbed Thruster,UTT

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Wang, Joseph (committee chair), Barnhart, David (committee member), Erwin, Dan (committee member), Gupta, Satyandra (committee member)

Creator Email antypas@usc.edu,Robert.antypas@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC15619740

Unique identifier UC15619740

Legacy Identifier etd-AntypasRob-9852

Document Type Dissertation

Format application/pdf (imt)

Rights Antypas, Robert

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu