PROPULSION UNIT FOR CUBESATS (PUC)

Abstract

The CU Aerospace/VACCO Industries Propulsion Unit for CubeSats (PUC) was developed as a medium thrust, medium impulse thruster system to enable CubeSat orbital maneuvering, formation flying, and rendezvous. The 0.25U PUC casing is all-welded titanium, a n d comes fully integrated with all necessary propulsion subsystems, including controller, power processing unit, micro-cavity discharge thruster, propellant valves, heaters, sensors, and software. The unit is software-configurable to operate over a wide range of power, thrust, and impulse levels. System setpoints, system status, and firing telemetry are accessible and configurable through an RS422 serial interface. The baseline system fits within a compact 350 cm 3 volume (0.25U+"hockey puck"), with an 89 mm x 89 mm cross-section, leaving clearance for solar panels and magnetic torquers. PUC may be expanded from 0.25U to any desired length, providing significant potential for increased propellant capacity and ΔV capability, compared with the baseline 0.25U design. The 0.25U PUC draws 15 W when using a microcavity discharge (MCD) to heat the high-density, self-pressurizing liquid SO 2 propellant, coupled to an optimized micronozzle to provide 5 mN thrust at 70 s I sp , a 4 kg CubeSat ΔV of 48 m/s, and demonstrates negligible component wear a n d constant lifetime operations. A dedicated propellant heater provides for continuous operation below +5°C ambient temperature. Cold gas operation can be used for small impulse operations. On-orbit update of system parameters is provided, including thrust duration, plenum pressure, MCD power level temperature & fault set-points.

Full text

62nd JANNAF Propulsion Meeting (7th Spacecraft Propulsion), Nashville, TN, 1–5 June, 2015
DISTRIBUTION A: Approved for public release, distribution unlimited.
Work funded by AFRL/RQRS on contract FA9300-11-C-0007.
© Copyright by authors. Paper Tracking # 4059
PROPULSION UNIT FOR CUBESATS (PUC)
D. L. Carroll1, J. M. Cardin2, R. L. Burton1, G. F. Benavides1, N. Hejmanowski, 1 C. Woodruff1, K. Bassett1,
D. King1, J. Laystrom-Woodard1, L. Richardson1, C. Day2, K. Hageman2, and R. Bhandari2
1CU Aerospace, Champaign, IL
2VACCO Industries, South El Monte, CA
ABSTRACT
The CU Aerospace/VACCO Industries Propulsion Unit for
CubeSats (PUC) was developed
as a medium thrust, medium impulse thruster system to enable CubeSat orbital maneuvering,
formation flying, and rendezvous. The 0.25U PUC casing is all-welded titanium, and comes
fully integrated with all
necessary propulsion subsystems, including controller,
power processing
unit, micro-cavity discharge thruster,
propellant valves, heaters, sensors, and software. The unit
is software-configurable to operate over a wide range
of power, thrust, and impulse levels.
System setpoints,
system status, and firing telemetry are accessible
and configurable through
an RS422 serial interface. The baseline system fits within a compact
350 cm3 volume
(0.25U+“hockey puck”), with an 89 mm x 89 mm cross-section, leaving clearance for solar
panels and magnetic torquers. PUC may be expanded from 0.25U to any desired length,
providing
significant potential for increased propellant capacity and ΔV capability, compared with
the
baseline 0.25U design. The 0.25U PUC draws 15 W when using a microcavity discharge (MCD)
to heat the high-density, self-pressurizing liquid SO2
propellant, coupled to an optimized
micronozzle to provide 5 mN thrust at 70 s Isp, a 4 kg CubeSat ΔV of 48 m/s, and demonstrates
negligible
component wear and constant lifetime operations. A dedicated propellant heater
provides for continuous operation below +5°C ambient temperature. Cold gas operation can be used for
small impulse operations. On-orbit update of system parameters is provided, including thrust duration,
plenum pressure, MCD power level temperature & fault set-points.
1. INTRODUCTION
An emerging trend in the field of space exploration is the development and deployment of low mass
satellites, commonly referred to as micro-, nano-, or femto-satellites. These satellites are seeing
increasing use as a low-cost alternative to more traditional large-scale spacecraft, a notable example
being the CubeSat standard, and eventually as components in a durable, redundant satellite network.
Nanosats are designed for a life of 1-2 years. They often have body-mounted solar panels, which makes
them severely power limited, with usable specific power P/m of ~1 watt per kg of satellite mass (note that
the Lockheed-Martin A2100 bus used for geosynchronous satellites has 4000 W capability and a launch
mass of ~3000 kg, so the 1 W/kg rule of thumb holds over a wide range of satellite masses). CubeSats
are also volume limited (1 liter per cube), placing a severe volume constraint on the propulsion system.
Nanosats are a low-cost, easily replaced approach to satellite constellations, and as such need to be
nimble. That is, orbital maneuvers need to be accomplished relatively quickly to minimize mission control
costs and maximize the usable satellite duty cycle of 1-2 years. Although rapid orbital maneuvering can

DISTRIBUTION A: Approved for public release, distribution unlimited.
2
always be accomplished by chemical propulsion, scaling down chemical systems to nanosat size (thrust
<< 1 Newton) has proved difficult for solid and liquid propulsion systems. Thus, one impediment to wide
implementation of nanosats is the lack of a highly compact, simple, and efficient propulsion system for
primary (orbital transfer, drag makeup, and maneuvering) and secondary (attitude and trajectory control)
applications. Advances at the University of Illinois in using microcavity plasma discharges for illumination
[Eden, 2005; Park, 2005] provide a possible solution to this problem in the form of a Microcavity
Discharge (MCD) thruster [Burton, 2009; Burton, 2010; Burton, 2012]. Unlike previous attempts at
microdischarge propulsion systems that failed because they operated in an arc mode, the MCD thruster
operates in a very low electrode erosion normal or abnormal glow discharge mode at high voltage and
low current. This new technology can revolutionize low-power electric propulsion for femto-, pico-, nano-,
micro- and even larger satellites to perform various mission tasks including orbit transfer, de-orbiting,
station-keeping, position, attitude and acceleration control, and structure control.
Our motivation for adapting MCD technology to the micropropulsion field is the expected system
benefits of high specific thrust, high thrust density, and high specific power with high propellant utilization
and a simple power processor. These aspects of MCD technology are desirable for a number of reasons.
First, an MCD thruster operates at the low power levels available on nanosatellites [Ghosh, 2010].
Additionally the MCD thruster is capable of generating higher thrust than other forms of electric propulsion
which increases the operational lifetime of a typical nanosatellite. For very small scale electrothermal
thrusters gas heating to temperatures in excess of 1000 K is possible with microdischarges. The
microdischarge gas temperature can be tuned to range from ambient values to high values by changing
the power input. The overall thruster efficiency is predicted to be greater than 60% using monatomic gas
propellants, and power scalability is straightforward over a wide range by operating large numbers of
abnormal glow discharges in parallel. Additionally, the service lifetime of the thruster is expected to be
long due to operation in the glow discharge mode and the capability of operating without auxiliary
components.
Other advantages of the MCD thruster include:
1. Power is coupled via the normal or abnormal glow discharge, so electrodes remain relatively cool,
and heat loss is minimized.
2. The system incorporates capacitively-coupled and/or direct-coupled electrodes, minimizing
sheath loss and electrode ablation.
3. Electrode erosion is very low, because the ion sputtering erosion mechanism is significantly
reduced by operating in low current normal or abnormal glow discharge modes.
4. Low wall heat loss due to a wall area on the order of 0.1 – 1.0 mm2 per cavity.
5. Ionization fraction is <<1% and small resulting frozen flow losses.
6. No auxiliary systems are needed, e.g. neutralizer, heater, igniter.
7. Propellant is stored as a self-pressurizing liquid, and is evaporated to provide an operating
pressure is 0.1 – 1.0 atm, giving reasonable nozzle Reynolds numbers, low viscous losses, and
mN thrust levels [Bayt, 1999].
8. High stagnation temperatures are possible with the MCD (1500 K has been obtained with
Al/Al2O3 insulators), much higher than attainable with the resistojet, without the need for bulky,
inefficient thruster insulation.
9. The power processing unit (PPU) is comprised of a DC-AC inverter or pulsed DC power supply
with low mass (≈ 3 g/W), and with PPU efficiency exceeding 85%.
10. Throttleable by varying source pressure.
11. Very low thrust noise, making it a candidate for missions requiring extremely precise, low-noise
acceleration control.
12. Very low system mass and volume for use on low mass (<10 kg) satellites.
This paper reviews prior MCD thruster results, motivations for the technology, and presents
performance data taken with the CU Aerospace/VACCO 0.25U Propulsion Unit for CubeSats (PUC), a
robust system based upon MCD thruster technology.

DISTRIBUTION A: Approved for public release, distribution unlimited.
3
2. BACKGROUND AND DISCUSSION
EARLY MCD THRUSTER DESIGN AND EXPERIMENTS
The early version of the MCD thruster [Burton, 2009; Burton, 2010], Figs. 1 and 2, is an
electrothermal thruster composed of a gaseous propellant supply and feed system, two or more Al/Al2O3
electrodes powered by a 50-500 kHz, 400-1600 V AC supply, a 100-200 µm diameter cavity in which the
discharge plasma is created, and a moderate Reynolds number micronozzle [Bayt, 1999]. The key
technology behind the MCD Thruster is the MEMs-scale plasma discharge. Over the past decade, it has
been found that low temperature plasma confined to a microcavity has several remarkable properties,
including specific power loadings (i.e., volumetric power deposition into the plasma) of up to ~1000
W/mm3 on a steady state basis, operating pressures of less than 100 Torr to above one atmosphere, and
electron temperatures of 3-6 eV. This region of the parameter space has not previously been accessible
to plasma science but it is already clear that these parameters open entirely new applications of plasma
technology.
Fig. 1. Schematic of original design of an MCD
electrothermal microthruster.
Fig. 2. Photograph of early stage development MCD
electrothermal microthruster with foil electrodes.
Inset shows unit in operation.
In early versions, the basic structure of the MCD thruster was composed of two or more electrodes
stacked in a ‘sandwich’ configuration with a microcavity drilled or etched through the stack. The
electrodes were created by growing an oxide layer on pure aluminum foils. The microcavity can be either
drilled using a commercially available micro drill bit, or created with a wet chemical etching process, for
cavities having characteristic dimensions as small as 10 µm. The foils were then aligned and connected
together using an epoxy resin, and then connected to a chemically etched nozzle. This effort was
supported by numerical modeling of the microdischarges [Sitaraman, 2010] to better understand the
dominant physical mechanisms.
Previously published work on the MCD thruster using gaseous propellants, typically argon, neon,
nitrogen, or a mixture thereof focused on power levels in the 1 – 5 W range [Burton, 2009; Burton, 2010;
de Chadenedes, 2010]. These studies included thrust estimates and thermal efficiency measurements.
The highest efficiency reached to date is 50%, with 60% predicted [de Chadenedes, 2010]. Lifetime
issues were prevalent in the early stage development MCD thrusters with the foil structure; the thin
insulation would typically develop microcracks during operation that would ultimately result in shorting and
arcing in the flowing system. As such, CU Aerospace used Internal Research and Development funds to
explore a modification of the MCD thruster that consisted of replacing the insulator with a more robust
oxide tube. With the implementation of the dielectric tube several configurations are possible for the

DISTRIBUTION A: Approved for public release, distribution unlimited.
4
electrodes for initiating and sustaining an alternating current normal or abnormal glow discharge [Burton,
2012].
NANOSATELLITE THRUSTER CHOICES
An important question for nanosatellites is: what range of efficiency and specific impulse are
appropriate for a nanosat electric micropropulsion system? TRL 9 EP systems have flown with efficiency
η (%) and specific impulse Isp (s) including the pulsed plasma thruster (10%, 1000 s); the resistojet (50-
80%, 300 s); the Hall thruster (50%, 2000 s); and the ion thruster (70%, 3000 s). Other EP systems in
advanced development are the colloid thruster; and the FEEP thruster. Propulsion selection for nanosats
depends on the propulsion capability, expressed in terms of the maneuver time and the required orbital
maneuver expressed in terms of ΔV, and also on the mass and volume available for the propulsion
system on the nanosat.
Burton et al. [Burton, 2010] introduced an equation for a constrained maneuver time that showed ΔV
varying inversely with Ue; a priori, this is counterintuitive because high ΔV interplanetary missions typically
utilize high specific impulse systems. The conclusion is that, in order to minimize orbit transfer times,
more maneuver capability is available for propulsion systems with low exhaust velocity and specific
impulse. To insist incorrectly on a high specific impulse is to incur a long time to perform the maneuver or
to limit the ΔV capability of the nanosat.
Clearly, the maneuver time t is a fundamentally important parameter. The question then is what
maneuver time is appropriate? Because we are dealing with low-cost nanosats with limited design life (1-
2 years) in a rapid response environment, it is not useful to have maneuver times of weeks or months and
their associated delayed response, high mission control support costs and satellite downtimes. It is more
reasonable that the time to perform a maneuver should be measured in days. Figure 3 illustrates some
typical values of ΔV per day for a maneuverable nanosat, as a function of Isp. We assume that ηφ ~ 0.50
(where φ is the power fraction Pp/P,
defined in terms of the propulsive
power Pp, and the maximum nanosat
bus power P produced by the solar
panels), P/m ~ 1 W/kg, and that the
desired time for a single maneuver is
1.0 days.
As discussed by Burton et al.
[Burton, 2010], the “sweet spot” for
nanosat orbital maneuvers (shaded
region) appears to be in the 70 – 400 s
range of Isp, where ΔV is relatively large
but the fuel fraction is reasonably small,
Fig. 3. For 50 s, typical of cold gas
thrusters, ΔV is high but fuel fraction is
too large. For 2000 s, assuming a 50%
efficient Hall thruster, the ΔV per day is
only 4.3 m/s; for 3000 s, assuming a
50% efficient ion thruster, the ΔV per
day is only 2.9 m/s; and for a 6000 s
FEEP thruster, the ΔV per day is only
1.5 m/s. These latter ΔV values are too
small to be useful in time-constrained
maneuvers. The colloid thruster could
eventually be considered assuming
significant improvements in efficiency
and system volume.
We note that a nanosat propulsion system can operate from batteries. For a 5 kg, 5 W nanosat
operating for one day, the required energy is 432 kJ = 120 W-hr. Lithium-ion batteries of this size would
Fig. 3. Operating envelope for nanosat propulsion. Maneuver
time is one day, requiring high thrust and reducing specific
impulse to the electrothermal range.

DISTRIBUTION A: Approved for public release, distribution unlimited.
5
have a mass of about 1 kg, or 20% of the satellite mass, making battery operation possible, but requiring
a large fraction of the total nanosat mass. Batteries could be used in conjunction with photovoltaic cells
to increase power and decrease maneuver time, effectively providing φ > 1.
Unlike low power ion and Hall thrusters, which incur a large efficiency penalty from their neutralizers,
electrothermal thrusters in principle can operate at high efficiency at low Isp. The reason that ion and Hall
thrusters need high Isp to be efficient is that the exhaust is fully ionized, so that the kinetic energy of the
exhaust must be large compared to the energy required (ion cost) to ionize the xenon propellant. Low
power electrothermal thrusters on the other hand have no inherent requirement for highly ionized
propellant, which can be made sufficiently conductive with an ionization fraction of 10-3 – 10-6.
The conclusion from this discussion is that the best specific impulse range for nanosats is relatively
low, in a range favoring electrothermal thrusters, Fig. 3.
PROPELLANT SELECTION
Propulsion performance is critically dependent on the propellant choice. A number of propellants
have been considered for CubeSats, including isobutane (C4H10), nitrous oxide (N2O), propane, ammonia,
hydrazine, peroxide, refrigerants (R134a), etc. [London, 2010]. CU Aerospace executed a study of 350
candidate propellants for the CubeSat/nanosatellite propulsion application, and down-selected to 8
candidates. Selection is based on the following criteria, Tables 1 and 2. Another common refrigerant
R236fa is also a potential candidate; while it has several differences to R134a (most notably molecular
weight and lower vapor pressure), it is overall similar to R134a in estimated total performance numbers
and is not discussed further in this paper. (Note that SO2 has also previously been denoted as EP-13.)
Table 1: Criteria for best candidate nanosatellite propellants.
Criterion
Justification
Favorable for
Not favorable for
High liquid density ρ
max propellant mass and ΔV
Water, SO2, R134a
NH3, N2O, C4H10
High ρ x sound speed
max ΔV
H2O, N2H4, SO2, NH3, R134a
SF6, N2O, C4H10
Low heat of vaporization
low propellant heater power
SO2, R134a
H2O, N2H4, NH3
Self-pressurizing
simplifies feed system
SO2, NH3, R134a
H2O, N2H4, N2O
Critical temperature >60°C
liquid between 0°C and 60°C
H2O, SO2, NH3, R134a
N2H4, SF6, N2O, C4H10
Low freezing point
liquid between 0°C and 60°C
SO2, NH3, R134a
H2O, N2H4
Compatible with materials
& electronics
Enables location of electronics
inside storage tank
R134a, C4H10, SO2
H2O, NH3
Overall Selection
Optimizes Propulsion System
SO2, R134a
H2O, N2H4, NH3, SF6,
N2O, C4H10
Because both cold and warm gas could be used, the primary selection criterion is the product ρa of
liquid density times sound speed at 300 K, or equivalently the product of liquid density and maximum cold
Isp, Table 2. A secondary criterion is the propellant heat of vaporization.
The third criterion is self-pressurization capability, which eliminates the need for a separate
pressurization system, saves mass and volume, and therefore increases propellant mass and impulse.
Propellants are selected with sufficient vapor pressure at 0ºC and modest pressure at 60ºC to avoid
excessive tank wall thickness and mass (note that thicker tank walls can significantly reduce propellant
volume in the small tank sizes necessitated for nanosatellites). Propellants with a critical temperature
below 60ºC (SF6, N2O, C4H10) are avoided because the initial tank fill must be low to avoid over-
pressurization at 60ºC.

DISTRIBUTION A: Approved for public release, distribution unlimited.
6
Table 2: Comparison of product of liquid density and 90% of maximum Isp at 500°C for nanosatellite
propellants.
Propellant
Mol.
Weight
(g/mole)
Density
(g/cm3)
Isp at 500°C &
90% Nozzle Eff.
(s)
Density x Isp
(g-s/cm3)
Issues
H2O
18
1.002
155.1
155.4
Freezes @ 0°C, low vapor pressure
N2H4
32
1.008
116.3
117.2
Toxic, Freezes @ 2°C
SO2
64
1.381
82.2
113.6
Manageable Low toxicity
NH3
17
0.609
159.5
95.2
High P @ 60°C, thick structure
R134a
102
1.225
65.1
79.8
None
N2O
44
0.785
99.2
77.8
Critical temperature < 60°C
SF6
146
1.374
54.4
74.8
Critical temperature < 60°C
C4H10
58
0.579
86.4
50.0
Low liquid density, Tcritical < 60°C
The fourth criterion is materials compatibility with the feed system, thruster and with the control and
power electronics. While not strictly required, this capability gives the most volume-efficient way to
package electronics, inside the propellant tank, while providing waste heat to maintain propellant
pressure and temperature while evaporating. Testing studies performed by CU Aerospace have
identified materials that are compatible with R134a and SO2.
Finally, freezing is a concern for a tank temperature of 0ºC for H2O and N2H4, requiring that this risk
be mitigated by thermal management and propellant heating. These two propellants, despite high ρa, are
also contraindicated by high heat of vaporization and low self-pressurization. Because
nanosatellites/CubeSats are generally power limited, the additional heater power required during lengthy
LEO eclipse times could significantly impact these nanosatellites. Of the investigated propellants, the two
most appealing for the CubeSat operating temperature range of 0 – 60 °C are R134a and SO2. Note that
SO2 was a common refrigerant until the mid-1920’s when CFC-based refrigerants were introduced. With
the exception of its low toxicity, SO2 has excellent refrigerant characteristics and it is these characteristics
that also make it an excellent propellant that is compatible with the MCD thruster technology.
3. EXPERIMENTAL RESULTS WITH PUC PROTOTYPE
EXPERIMENTAL SETUP
Experiments were performed with a conical thruster nozzle having a 0.015” (0.38 mm) throat just
downstream of the microcavity discharge region. The basic configuration tested is illustrated in Fig. 4, for
which the plenum, MCD, PPU, and nozzle were all in a compact test housing, Fig. 5. Thrust stand
measurements were performed by placing the compact test housing on a Watt linkage pendulum thrust
platform with non-intrusive wires and gas feed line. The thrust stand has a large amount of prior historical
usage [Wilson, 1997; Laystrom, 2003]. Windage effects [Whalen, 1987] were observed from gas
recirculation in the medium sized vacuum tank when background pressures were < 10-5 Torr, so thrust
stand measurements were taken with a background pressure of approximately 160 milli-Torr. As such,
measurements of thrust include a small correction (typically < 5%) for the background pressure.
Fvacuum =Fmeasured +P
measured Aexit
A photograph of the thruster in operation is shown in Fig. 6.

DISTRIBUTION A: Approved for public release, distribution unlimited.
7
Fig. 4. Schematic of basic robust prototype PUC thruster system. The components within the dashed region
were tested inside the test housing.
Fig. 5. Photograph of prototype PUC thruster in test
housing.
Fig. 6. Photograph of prototype PUC thruster in
operation using SO2.
EXPERIMENTAL DATA
In the laboratory the MCD thruster operates at constant mass flow rate and allows the static pressure
to increase as the total temperature increases. Knowing mass flow rate and the discharge coefficient CD
allows the total temperature of the discharge to be inferred from comparison between the pressure and
mass flow measurements:
T0
*=
γ
R
γ
+1
2
"
#
$
%
&
'
p*Aeff

m
"
#
$
%
&
'
2
where R is the gas constant, γ depends upon the total temperature,
Aeff
*=CDA*
,
p0
*≅p0
, and the throat
pressure is given by:
p*=p02 /
γ
+1
( )
( )
γ
γ
−1
Experimentally, a cold gas measurement was made followed by ignition of the MCD and a warm gas
measurement taken for the same mass flow rate. Since the mass flow rate and discharge coefficient are
experimentally fixed, an increase in line pressure (assumed to be the stagnation pressure since the
incoming flow Mach number is < 0.1) corresponds to a rise in discharge stagnation temperature by the

DISTRIBUTION A: Approved for public release, distribution unlimited.
8
above equations and can therefore be determined without the need for direct measurements other than
the pressure. Thus, neglecting a small correction due to variations in γ, T0 ~ p2, or
T0,warm
*
T0,cold
*
⎛
⎝
⎜⎞
⎠
⎟=p0,warm
*
p0,cold
*
⎛
⎝
⎜⎞
⎠
⎟
2
Measurement of the cold and warm pressures along with the assumption that the cold stagnation
temperature is room temperature provides an estimated total temperature in the discharge region
downstream of the discharge.
Cold and Warm Gas Performance Data
Plenum pressures using SO2 propellant are shown in Fig. 7, as well as estimated total temperatures
in the discharge region (using the above equation) just upstream of the throat, Fig. 8. Note that Fig. 8
shows that the total temperature decreases with mass flow for approximately a constant input power of 10
W, as the same amount of power is being deposited into a progressively larger amount of gas flow.
Fig. 7. Plenum pressure of prototype PUC thruster
operating in cold gas and warm gas modes.
Fig. 8. Estimated total temperature in discharge
region of prototype PUC thruster operating in cold
gas and warm gas modes (for approximately a
constant input power of 10 W).
Thrust stand measurements including the small correction for the windage effects are shown in Figs.
9 and 10. The nominal operating flow rate of 7.6 mg/s of propellant indicates a thrust of 3.5 mN and an
Isp of 46 s in cold gas mode, and a thrust of approximately 5.4 mN and an Isp of 72 s in warm gas MCD
discharge mode. This represents a significant 54% rise in thrust and specific impulse with the highly
compact MCD configuration. Cold gas performance was also tested at a much higher flow rate of 14.6
mg/s (not shown for brevity) and produced a thrust of 7.6 mN and an Isp of 52 s. Cold flow thrust stand
measurements were also performed with R134a (molecular weight = 102), and display a similar trend to
that of SO2 (molecular weight = 64), but with the anticipated reduced specific impulse (Section 4). Note
that R134a was found to be incompatible with the MCD warm fire operation due to dissociation of the
R134a molecules and subsequent plugging of the nozzle.

DISTRIBUTION A: Approved for public release, distribution unlimited.
9
Fig. 9. Measured thrust of prototype PUC thruster
operating in cold gas and warm gas modes.
Fig. 10. Measured specific impulse of prototype PUC
thruster operating in cold gas and warm gas modes.
Dissociation of the molecular propellant almost certainly plays some role in discharge inefficiencies.
The bond energy to dissociate the first O atom from the molecule is 5.676 eV [Takacs, 1978], which
corresponds to a power loss of 8.6 W for every 1 mg/s that has the first bond broken. Experiments have
not been performed to determine quantitatively the amount of bond dissociation occurring in these MCD
tests, but we estimate that it is < 5%, i.e. greater than 95% of the flow is undissociated.
Testing of the MCD was also performed for several different input powers to the PPU. Figure 11
illustrates the general rise in total temperature of the discharge flow as a function of the ratio of power
input to the PPU to the mass flow rate. A more limited set of thrust stand measurements was made for
higher specific powers that seemed the most promising, Fig. 12. To within experimental error, the
specific impulse was approximately a constant for specific powers in the range of 1.2 – 1.8 W/mg/s.
Fig. 11. Estimated total temperature in discharge
region of prototype PUC thruster operating in warm
gas mode as a function of the specific energy in the
discharge region.
Fig. 12. Measured specific impulse of prototype PUC
thruster operating in warm gas mode as a function of
the specific energy in the discharge region.

DISTRIBUTION A: Approved for public release, distribution unlimited.
10
To test the durability of the thruster system a 19-hour continuous test was performed, Figs. 13-16,
which exceeds the available tank propellant in the PUC system described in Section 4. The flow rate for
this test was held approximately constant at 7.9 mg/s. An earlier model of our PPU with slightly lower
efficiency was utilized in this 19-hour experiment, so performance is slightly decreased from that shown in
Figs. 7-10, but the differences are < 5%. Interestingly, the thrust performance increased with operational
time, possibly due to gradual heating of the entire test housing in which the gas flows into the discharge
region. Microscope investigation of the thruster after the 19-hour test showed no plugging or erosion of
the nozzle.
Fig. 13. Plenum pressure of prototype PUC thruster
operating in cold gas and warm gas modes during a
19-hour test.
Fig. 14. Estimated total temperature in discharge
region of prototype PUC thruster operating in warm
gas mode during a 19-hour test.
Fig. 15. Measured thrust of prototype PUC thruster
operating in cold gas and warm gas modes during a
19-hour test.
Fig. 16. Measured specific impulse of prototype PUC
thruster operating in cold gas and warm gas modes
during a 19-hour test.

DISTRIBUTION A: Approved for public release, distribution unlimited.
11
Discharge coefficients typically ranged from 0.84 – 0.90 for this nozzle, depending upon flow
conditions. Nozzle efficiencies typically ranged from 0.80 – 0.90, again depending upon flow conditions.
4. PUC FLIGHT SYSTEM
DESCRIPTION OF PUC FLIGHT SYSTEM
The CU Aerospace / VACCO Propulsion Unit for
CubeSats (PUC) is a complete high-
performance and
compact small-satellite propulsion solution, Fig. 17. The all-welded titanium
PUC comes fully integrated with all
necessary propulsion subsystems, including controller,
power processing unit, micro-cavity discharge thruster,
propellant valves, heaters, sensors, and
software. PUC
is software-configurable to operate over a wide range
of power, thrust, and
impulse levels. System set-points,
system status, and firing telemetry are all accessible
and
configurable through an RS422 serial interface.
Fig. 17. Photograph of 0.25U PUC flight system showing locations of MCD thruster,
propellant fill port, firmware update header and mounting holes in a welded titanium
structure. The propellant tank is included in the structure.
The baseline 0.25U system fits within a compact
350 cm3 volume (0.25U + “hockey
puck”), providing
outstanding performance for minimal CubeSat volume
and mass fraction. The
PUC’s 89 mm x 89 mm cross-section intentionally falls well under the CubeSat 100 mm x
100 mm specification, so as to
not interfere with other CubeSat subsystems such as solar
panels and magnetic torquers. For
increased performance, or to meet customer specific mission
requirements, the tank width may be
customized. The tank may additionally be expanded from
0.25U to any desired length, providing
significant potential for increased propellant capacity, i.e.
delta-V capability, compared with the
baseline 0.25U design.
The PUC achieves its high total impulse, low-volume capability by employing CU Aerospace
Micro-Cavity Discharge (MCD) propellant heating technology, high-density and self-pressurizing
liquid
propellants, and an optimized low-mass-flow nozzle. PUC MCD thrusters demonstrate
negligible
component wear during 0.25U life cycle testing, providing constant lifetime
operations. The
robust MCD components permit extensive warm firing beyond the 0.25U
operational life, allowing
for pre-flight testing and/or increases to the PUC’s propellant tankage
without impacting MCD
performance or reliability.
PUC system performance estimates are given in
Tables 3 and 4. The system envelope and dimensions are provided in Fig. 18 along with center of
gravity estimates. Other PUC system specification (mechanical, propellant, and electrical) are provided in
the Appendix.

DISTRIBUTION A: Approved for public release, distribution unlimited.
12
Table 3. Performance specifications of 0.25U PUC MCD thruster system in cold (14.6 mg/s flow rate) and
warm (7.6 mg/s flow rate) gas modes.
!
Parameter
'
Warm'Fire'Only
'
Cold'Fire'Only
'
Unit
'
Notes
'
"#$%&'!
()*!
*)*!
+,!
,-+./01!
"-'01!.+2%1&3!
45(!
46(!
,7&!
831'079!:020;.1.'<!=(!>?!@%;3A0'B!
(5!
C6!
+D&!
831'079!:020;.1.'<!=C!>?!@%;3A0'B!
E(!
(C!
+D&!
A23:.F.:!.+2%1&3!
GH!
(G!
&3:!
,-+./01!
I0J.+%+!:-/'./%-%&!'#$%&'!'.+3!
6H!
*(!
+./!
I./.+%+!.+2%1&3!;.'!
777!
4)H!
+,7&!
!
Table 4. Performance of 0.25U PUC MCD thruster system in cold (14.6 mg/s flow rate) and warm (7.6 mg/s
flow rate) gas modes as a function of CubeSat launch mass.
CubeSat Launch Mass
(kg)
ΔV in Cold Gas Mode
(m/s)
ΔV in Warm Gas Mode
(m/s)
1.5
86
127
2.0
64
95
3.0
43
64
4.0
32
48
5.0
26
39
System Features
Operation:
K Two operational modes:
K Warm gas mode for high specific impulse, large
total impulse manuevers.
K Cold gas mode for minimum or small total impulse manuvers.
K Highly configurable controller for on-orbit update
of system parameters, including:
K Thrust duration
K Plenum pressure (thrust)
K MCD power level (specific impulse)
K Temperature set-points
K Fault set-points
K System status packets for health monitoring
K Regular telemetry packets during operation
K Dedicated propellant heater for continous operation below +5°C ambient temperature.
K Propellant temperature sensor for closed-loop propellant temperature regulation.
K Propellant vaporizer ensuring 100% vapor delivered
from liquid storage.
!

DISTRIBUTION A: Approved for public release, distribution unlimited.
13
Fig. 18. Envelope and dimensions of the 0.25U PUC flight system including center
of gravity estimates.
Sulfur dioxide (SO2) characteristics:
K High mass density
K Self-pressurizing
K Non-flammable
K Chemically stable
K High critical temperature
K Low freezing point
K Low vapor pressure
K Commonly used refrigerant prior to the development of freons
The hardware is readily adaptable to a variety of propellant tank sizes to provide a family of possible
PUC systems, Fig. 19. Table 5 provides estimates of performance with the larger tank sizes for a 4 kg
CubeSat.
!
!
8$<!L++M!
N3'!L++M
!
O!
4)6!
H)E!
P!
H)Q!
H)*!
R!
74)E!
7()5!

DISTRIBUTION A: Approved for public release, distribution unlimited.
14
Fig. 19. Family of MCD PUC thruster systems with different propellant tank sizes (0.25U, 0.5U, and 1U sizes
are illustrated).
Table 5. Estimated performance of PUC MCD thruster systems for a 4 kg CubeSat launch mass in cold (14.6
mg/s flow rate) and warm (7.6 mg/s flow rate) gas modes as a function of PUC tank size.
PUC Size
(Units)
ΔV in Cold Gas Mode
(m/s)
ΔV in Warm Gas Mode
(m/s)
0.25
32
48
0.5
51
74
0.75
74
109
1.0
97
144
PUC FLIGHT SYSTEM EXPERIMENTAL DATA
Eight PUC flight systems were fabricated,
tested and delivered to the Air Force in 2014.
Acceptance testing of each of these systems was
performed. Plenum pressures using SO2 and
R134a propellants are shown in Fig. 20. Thrust
stand measurements with both of these
propellants (including the small correction for the
windage effects) are shown in Figs. 21 and 22.
As noted earlier, R134a was found to be
incompatible with the MCD warm fire operation
due to dissociation of the R134a molecules and
subsequent plugging of the nozzle, and was
therefore only tested in cold gas mode. Operation
of all units was within desired specifications.
We note that it should be possible to push the
performance to higher thrust and Isp in the warm
gas MCD discharge mode with continued
optimization of the geometry and electronics. Our
estimates are that a total temperature of 1200 K is
achievable for which we expect an Isp > 75 s.
Fig. 20. Plenum pressure of flight PUC thruster
operating in cold gas and warm gas modes with SO2
and in cold gas mode with R134a.

DISTRIBUTION A: Approved for public release, distribution unlimited.
15
Fig. 21. Measured thrust of flight PUC thruster
operating in cold gas and warm gas modes with SO2
and in cold gas mode with R134a.
Fig. 22. Measured specific impulse of flight PUC
thruster operating in cold gas and warm gas modes
with SO2 and in cold gas mode with R134a.
5. CONCLUDING REMARKS
As nanosatellites become an increasingly utilized tool in space missions, the need for a compact and
efficient propulsion system becomes ever greater. The Propulsion Unit for CubeSats (PUC) thruster
system represents a bridge for the unique needs of the nanosatellite propulsion gap: MCD devices are
extremely compact, smaller than virtually any other propulsion system. MCD thrusters are capable of
operation in power-limited systems, and MCD thrusters are capable of providing the total impulse levels
necessary for nanosatellite missions when utilizing self-pressurizing liquid propellants.
Sulfur dioxide was chosen as an ideal propellant for CubeSats because of its excellent set of
characteristics that compensate for CubeSat limitations. Thrust stand measurements using SO2
propellant showed a thrust of approximately 4.5 mN and an Isp of 46 s in cold gas mode, and a thrust of
approximately 5.4 mN and an Isp of 70 s in warm gas MCD discharge mode. This represents a significant
>50% rise in performance with the highly compact MCD configuration as compared with cold gas
operation. The robust MCD thruster system was tested in a 19-hour experiment and demonstrated no
degradation in performance, and no erosion of the electrodes, insulator or nozzle. The PUC system
using MCD thruster technology, was fabricated, tested and delivered to the Air Force by the CU
Aerospace and VACCO Industries team and represents the first compact (< 0.75U), high total impulse
nanosatellite propulsion system. The PUC can be built as a family of thruster systems having a variety of
propellant tank sizes to meet the needs of different nanosatellite missions.
6. ACKNOWLEDGEMENTS
This work has been sponsored by the United States Air Force Research Laboratory/RQRS contract
number FA9300-11-C-0007. James Singleton and William Hargus served as Program Managers during
different phases of the project, and we would also like to acknowledge their timely and important technical
contributions.

DISTRIBUTION A: Approved for public release, distribution unlimited.
16
7. REFERENCES
Bayt, R. L., Analysis, Fabrication and Testing of a MEMS-based Micropropulsion System, PhD
thesis, Department of Aeronautics and Astronautics, MIT. (1999)
Burton, R. L., Eden, J. G., Park, S.-J., Yoon, J. K., de Chadenedes, M., Garrett, S., Raja, L. L., Sitaraman,
H., Laystrom-Woodard, J., Benavides, G., and Carroll, D., Proceedings of the 31st International
Electric Propulsion Conference, Initial Development of the Microcavity Discharge Thruster, IEPC-
2009-169. (2009)
Burton, R. L., Eden, J. G., Park, S.-J., de Chadenedes, M., Garrett, S., Raja, L. L., Sitaraman, H.,
Laystrom-Woodard, J., Benavides, G., and Carroll, D., “Development of the MCD Thruster for
Nanosat Propulsion,” JANNAF Conf., Colorado Springs, CO, Paper 1387 (2010).
Burton, R. L., Benavides, G.F., and Carroll, D.L., “Space Thruster Using Robust Microcavity
Discharge,” U.S. Patent Application No. 13/680,484 (2012).
de Chadenedes, M., Yoon, J.K., Sitaraman, H., Garrett, S., Raja, L.L., Eden, J.G., Park, S-J, Laystrom-
Woodard, J., Carroll, D.L., and Burton, R.L., “Advances in Microcavity Discharge Thruster
Technology,” AIAA Paper 2010-6616 (2010).
Eden, J. G., Park, S. –J., Ostrom, N. P., Chen, K. -F., Kim, K. S. “Large Arrays of Microcavity Plasma
Devices for Active Displays and Backlighting,” IEE/OSA Journal of Display Technology, Vol. 1,
No. 1, pp. 112-116. (2005)
Ghosh, A, Coverstone, V. “Study of Low-Thrust Trajectories for Low Orbit Multiple CubeSat
Missions,” AAS conference, paper no. 10-174 (2010).
Laystrom, J.K., Burton, R.L., and Benavides, G.F., “Geometric Optimization of a Coaxial Pulsed
Plasma Thruster,” AIAA Paper 2003-5025 (2003).
London, A. P., and Droppers, L. J., “High-Performance Liquid Propulsion For CubeSats: Requirements
and Approaches,” JANNAF Conf., Colorado Springs, CO, Paper 1428 (2010).
Park, S.-J., Kim, K. S., Eden, J. G. “Nanoporous Alumina as a Dielectric for Microcavity Plasma
Devices: Multilayer Al/Al203 Structures,” Applied Physics Letters, Vol. 86, No. 22, pp. 1-3. (2005)
Sitaraman, H. and Raja, L., Simulation Studies of Alternating-Current Microdischarges for
Microthruster Applications, AIAA Paper No. AIAA-2010-231. (2010)
Takacs, G.A., “Heats of Formation and Bond Dissociation Energies of Some Simple Sulfer- and
Halogen-Containing Molecules,” J. Chem Eng. Data, Vol. 23, No. 2, 174-175 (1978).
Whalen, M.V., “Low Reynolds Number Nozzle Flow Study,” NASA-TM-100130 (1987).
Wilson, M.J., Bushman, S.S., and Burton, R.L., “A Compact Thrust Stand for Pulsed Plasma
Thrusters,” 25th International Electric Propulsion Conf., IEPC Paper 97-122, Cleveland, OH (1997).

DISTRIBUTION A: Approved for public release, distribution unlimited.
17
APPENDIX: PUC SPECIFICATIONS
I3:#0/.:01!A23:.F.:0'.-/&!=H)6*S!A<&'3+B
!!
!
T0$0+3'3$
!
I./.+%+!
,-+./01
!
I0J.+%+!
S/.'
!
,-'3&
!
8.+3/&.-/&!
5Q!++!J!5Q!++!J!EG!++!=H)6*S!U!V#-:>3<!2%:>WB!
N3'!+0&&!
G45!
?!
/-+./01!
8$<!+0&&!
(*H!
?!
/-+./01!
"3+23$0'%$3!$0/?3&!
X23$0'./?!
74C!
U*H!
Y@!
A'-$0?3!ZD-!2$-23110/'!
U*!
U*H!
Y@!
$3:-++3/[3[!2$3F1.?#'!&'-$0?3!
A'-$0?3!ZD!2$-23110/'!
U*!
UCH!
Y@!
$3:-++3/[3[!2$3F1.?#'!&'-$0?3!
A%$\.\0;.1.'<!
7C(!
UG4!
Y@!
F1.?#'!
T$3&&%$3!$0/?3&!=2$-23110/'!'0/>B!
X23$0'./?!2$3&&%$3!
6H!
(5!
466!
2&.0!
T$--F!2$3&&%$3!
C4G!
2&.0!
]%$&'!2$3&&%$3!
*65!
2&.0!
9.;3!0::32'0/:3!13\31!
4(!
^_IA
!
`30>!$0'3&!
a&-10'.-/!\01\3!=,@B!
*!
4Q!
?D<$!
4!?D<$!b!H)H(!&::#!
T$-2-$'.-/01!\01\3!=,@B!
*!
4Q!
?D<$!
4!?D<$!b!H)H(!&::#!
I0&&!F1-Z!$3?%10'.-/!='#$-''13!:020;.1.'<B!
@-1[!F.$3!
6H!
4HH!
6HH!
c!
-F!/-+./01!
N0$+!F.$3!
5H!
4HH!
46H!
c!
-F!/-+./01!
T$-23110/'!!A23:.F.:0'.-/&
!!
!
T0$0+3'3$
!
901%3
!
S/.'
!
,-'3&
!
T$-23110/'!
AX2!
d.?#!T%$.'<!`.e%.[!A%1F%$!8.-J.[3!
T$-23110/'!+0&&!=H)6*S!A<&'3+B!
6E5!
?!
,-+./01!
@$.'.:01!'3+23$0'%$3!
4*E)Q!
Y@!
f$33g./?!2-./'!
7G*)E!
Y@!
902-$!2$3&&%$3!
(5!
2&.0!
h'!6HY@!

DISTRIBUTION A: Approved for public release, distribution unlimited.
18
i13:'$.:01!A23:.F.:0'.-/&
!!
!
T0$0+3'3$
!
I./.+%+!
,-+./01
!
I0J.+%+!
S/.'
!
,-'3&
!
X23$0'./?!\-1'0?3!
Q!
44)4!
46)E!
9!
S/$3?%10'3[!
I0J!'$0/&.3/'!\-1'0?3!
4(!
9!
I0J!./7$%&#!:%$$3/'!
4()*!
h!
j!G!0+2&!F-$!13&&!'#0/!*H!m&!
8@!.&-10'.-/!$3&.&'0/:3!
H)QQ!
4!
4)H4!
Ik!
T-Z3$!
N0$+!f.$3!
46!
4*!
45!
N!
S&3$!2$-?$0++0;13!
@-1[!f.$3!
*)Q!
5!
Q)5!
N!
a/2%'!\-1'0?3![323/[3/'!
A'0/[;<!
H)H4!
H)H*!
N!
@-++%/.:0'.-/!&'0/[0$[!
_A(66!=44*)6!>;2&B!
i13:'$.:01!N.$./?!=f1<./?!`30[&B
!
6E!hN^!T"fi!=6(W!+./)B!
]1%3!
^$33/!
X$0/?3!
P311-Z!
_3[!=J6B!
]10:>!=J6B!
T%$2-&3!
_A(66lP!
_A(66lR!
_A(66lh!
_A(66l]!
Q746)E!9!
^,8!
f.$+Z0$3!S2[0'3!d30[3$
!
T./!m!
4!
6!
C!
!
!
T%$2-&3!
C)C9!
"8X!
"8a!
T./!m!
(!
*!
E!
T%$2-&3!
"IA!
"@n!
^,8!