Author’s Accepted Manuscript

Project Dragonfly: A Feasibility Study of

Interstellar Travel Using Laser-Powered Light Sail
Propulsion

Nikolaos Perakis, Lukas E. Schrenk, Johannes

Gutsmiedl, Artur Koop, Martin J. Losekamm
www.elsevier.com/locate/actaastro

PII: S0094-5765(16)30558-6
DOI: http://dx.doi.org/10.1016/j.actaastro.2016.09.030
Reference: AA6019
To appear in: Acta Astronautica
Received date: 9 June 2016
Accepted date: 28 September 2016
Cite this article as: Nikolaos Perakis, Lukas E. Schrenk, Johannes Gutsmiedl,
Artur Koop and Martin J. Losekamm, Project Dragonfly: A Feasibility Study of
Interstellar Travel Using Laser-Powered Light Sail Propulsion, Acta
Astronautica, http://dx.doi.org/10.1016/j.actaastro.2016.09.030
This is a PDF file of an unedited manuscript that has been accepted for
publication. As a service to our customers we are providing this early version of
the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting galley proof before it is published in its final citable form.
Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal pertain.
 Project Dragonfly: A Feasibility Study of Interstellar
Travel Using Laser-Powered Light Sail Propulsion
Nikolaos Perakisa,b,∗, Lukas E. Schrenka,b , Johannes Gutsmiedla,b , Artur
Koopa,b , Martin J. Losekamma,b
a
Technical University of Munich, Institute of Astronautics, Garching, Germany
b
Scientific Workgroup for Rocketry and Space Flight (WARR), Garching, Germany

Abstract
Light sail-based propulsion systems are a candidate technology for interplan-
etary and interstellar missions due to their flexibility and the fact that no fuel
has to be carried along. In 2014, the Initiative for Interstellar Studies (i4is)
hosted the Project Dragonfly Design Competition, which aimed at assessing
the feasibility of sending an interstellar probe propelled by a laser-powered
light sail to another star system. We analyzed and designed a mission to the
Alpha Centauri system, with the objective to carry out science operations at
the destination. Based on a comprehensive evaluation of currently available
technologies and possible locations, we selected a lunar architecture for the
laser system. It combines the advantages of surface- and space-based sys-
tems, as it requires no station keeping and suffers no atmospheric losses. We
chose a graphene-based sandwich material for the light sail because of its low
density. Deceleration of the spacecraft sufficient for science operations at the
target system is achieved using both magnetic and electric sails. Applying
these assumptions in a simulation leads to the conclusion that 250 kg of sci-
entific payload can be sent to Alpha Centauri within the Project Dragonfly
Design Competition’s constraints of 100 year travel duration and 100 GW
laser beam power. This is only sufficient to fulfill parts of the identified
scientific objectives, and therefore renders the usefulness of such a mission
questionable. A better sail material or higher laser power would improve the
acceleration behavior, an increase in the mission time would allow for larger

∗
Corresponding author
Email addresses: nikolaos.perakis@tum.de (Nikolaos Perakis ),
l.schrenk@tum.de (Lukas E. Schrenk)

Preprint submitted to Acta Astronautica September 29, 2016

spacecraft masses.
Keywords: interstellar travel, light sail, laser, mission analysis

1. Introduction
The long distances associated with interstellar travel prohibit the use of
conventional chemical or electric propulsion systems due to the extremely
large propellant masses required. More advanced propulsion methods have
been suggested to overcome this barrier. They aim to limit the propellant
mass carried on board by achieving very high specific impulses, thus enabling
shorter trip durations. Notable mission design examples use fusion-based
propulsion [1] or propellant-less methods such as large light sails [2, 3]. In
2014, the Initiative for Interstellar Studies (i4is) hosted the Dragonfly Design
Competition. Participants were asked to design an unmanned probe that
would be capable of flying to a nearby star system, and assess the technical
feasibility and scientific value of such an endeavor. The spacecraft was to be
equipped with a light sail as its primary means of propulsion. A powerful
laser system placed somewhere in the Solar System constitutes the required
light source. The competition requirements formulated by i4is were [4]:
1. To design an unmanned interstellar mission that is capable of deliver-
ing useful scientific data about the Alpha Centauri system, associated
planetary bodies, its solar environment, and the interstellar medium.
2. The spacecraft shall use current or near-future technology.
3. The Alpha Centauri system shall be reached within a century of the
probe’s launch.
4. The spacecraft propulsion for acceleration shall be mainly light sail-
based.
5. The mission shall maximize encounter time at the destination.
6. The laser beam power shall not exceed 100 GW.
7. The laser infrastructure shall be based on existing concepts for solar
power satellites.
8. The mission design should allow missions to a variety of target stars
within a 10-light-year radius.
We studied the laser system used for accelerating the probe. Its position-
ing properties are discussed in Section 2. Candidate sail materials and their
design characteristics are discussed in Section 3.

2
 The mission analysis was separated into the three distinct phases—acceleration,
cruising, and deceleration—as described in Section 4. The analysis of the de-
sign was carried out according to the simulation model presented in Section
5. The results of the mission analysis and the final spacecraft design are
outlined in Sections 6 and 7, respectively. For the design, we assessed the
feasibility of building a small interstellar probe based on current or soon-to-
be available technologies. One of the mission requirements defined by i4is
is the collection of scientifically valuable data. Maximizing the number of
instruments that can be incorporated into the spacecraft has a direct impact
on the scientific value of such a mission. We discuss the required instrumen-
tation and current trends for miniaturization briefly in Section 8. Finally,
concluding remarks on the scientific reward and the feasibility of the mission
are presented in Section 9.

2. Laser System and Optics

Due to the requirement of focusing extreme amounts of energy (up to
hundreds of GW) over large distances, the laser system used for the accel-
eration phase is one of the main challenges for the Dragonfly mission. The
first concern is the placement of the laser, with several options available. For
our study, we assumed that the laser is to remain either in a stable orbit or
on a planetary surface to allow reuse after the mission either for additional
probes or for other purposes. This should help to justify the enormous costs
for building such a system. The Alpha Centauri system has an inclination
of -60 degrees [5] and is hence visible in latitudes smaller than 30 degrees
South. While construction and energy resources are readily available on
Earth's surface—even near the South Pole where an Alpha Centauri-facing
laser would need to be built—the losses and pointing disturbances encoun-
tered while traversing the atmosphere mandate laser placement in space. A
laser system orbiting Earth or the Sun would need to constantly maintain
its orbit against the recoil caused by the emitted beam, requiring a pow-
erful propulsion system and additional energy sources. For these reasons,
we propose to position the laser on the Moon. Assuming development of
lunar infrastructure takes place until the time frame of the Dragonfly mis-
sion, the construction of large lunar structures should be possible. The re-
quired location near the Moon's south pole also enables energy production
via perpetually-lit solar arrays.

3
 While high-power scientific lasers in the desired 100 GW range are avail-
able, they are solely pulsed lasers. For the Dragonfly mission, a continuous
wave (CW) laser would be required in order to reduce loads on the sail.
It would be very desirable to find a different application for high-powered
CW lasers in order to utilize synergy effects. The most promising option
seem to be MW-class military laser systems. These lasers could be assem-
bled into very large arrays. By the time the Dragonfly mission is supposed
to be launched, CW lasers in the MW range should be mass-producible in
relatively small form factors with very rugged designs based on military re-
quirements. These properties will allow the installation of GW-class laser
arrays without extensive adaption. The free-electron lasers currently being
developed for military applications only require electric energy, without any
additional medium to be refreshed (as is the case for chemical lasers).
Depending on the mission parameters, focusing distances up to several
light years may be necessary. This requires an optical system focusing the
laser beam onto the sail of the accelerating probe. Forward [2] proposed a
Fresnel zone lens, which would have a diameter of hundreds to thousands of
kilometers at the desired light frequencies and focal distances. Unfortunately,
building this lens anywhere near a celestial body would likely cause it to be
torn apart by gravitational gradients, and building and positioning such a
large structure far away from any planets or stars was dismissed because
of the associated effort and cost. In addition, extremely accurate pointing
down to 10−12 rad, several orders of magnitude better than the capabilities of
current space telescopes, is required to keep the beam focused on the probe
over such long distances. This seems very difficult for a large and not very
stiff structure.
As an alternative to a lens-based system, a membrane reflector system as
described by Taylor [6] could be based on the Moons surface. It uses mirrors
consisting of dual reflecting aluminum sheets with a high voltage applied
between them. The resulting electrostatic forces pull the surfaces together,
creating a parabolic shape that can be controlled by changing the applied
voltage. We propose to build an array of several hundred relatively small
(100 m diameter) mirrors instead of one monolithic multi-kilometer mirror,
creating a flexible and extendable modular system (Figure 1).
A mirror array would allow repairs and upgrades during the acceleration
phase and could be re-purposed for powering multiple outposts or spacecraft
inside the solar system after the primary mission is over. By positioning
the mirrors on telescope mounts, it is easier to track the target across the

4
Figure 1: Mirror array concept

5
sky. It is not entirely clear whether the required pointing accuracy could be
achieved in this way—to our knowledge no estimation of pointing accuracy
for Moon-based telescopes exists—but due to the practically non-existent
atmosphere and low seismic activity it does not seem completely impossible.
Compared to a spacecraft-based telescope, the problems of limited accuracy
of the attitude control and jitter caused by on-board mechanical systems
can be eliminated. Since the achievable pointing accuracy is unknown, it
was used as a variable parameter in the analysis of the mission duration and
maximal achievable payload, as described in Section 6.

3. Laser Sail Material

The light sail material's properties have a great impact on the perfor-
mance of the proposed mission and influence the mass that can be sent. The
two major factors increasing a light sail’s acceleration characteristics are low
density and high reflectivity. The absorptivity and maximum operating tem-
perature influence the maximum surface power that can be applied to the sail
and thus determine the minimal sail diameter and mass. A small diameter
may be beneficial due to its low mass, however, it is not always the optimal
solution. As the laser pointing accuracy is restricted, a larger sail can be
accelerated by the full beam for a longer time and the optimum is no longer
the smallest possible sail.
A state-of-the-art solar sail material is aluminized Mylar, whereas a possi-
ble future material may be graphene. These materials have been compared by
Matloff and the results showed that a single graphene layer cannot compete
with Mylar, since its low density cannot compensate for its very low reflec-
tivity [7]. To overcome the very high transmittance τ of graphene, Matloff
proposes a sandwich structure composed of graphene superimposed between
two appropriate monolayers, which can increase the fractional absorption α
up to 40 %. Adding alkali atoms to the outer surface (the one facing the
light) increases the reflectivity ρ to 5 %. This comes at the expense of higher
density but demonstrates a much better acceleration potential.
If used in a sandwich structure, the properties of the graphene sail im-
prove, leading to a better performance than aluminized Mylar. In order to
compare the materials, we calculated the acceleration distance required to
reach 0.05 c for different spacecraft masses. Smaller distances are favorable
to avoid pointing and beam spread losses. The spread losses were assumed

6
 2.5
Aluminized mylar
Graphene monolayer
Acceleration distance to 0.05c [ly]

2 Graphene sandwich

1.5

0.5

0
1 2 3 4 5 6 7 8
3
Total system mass [10 kg]
Figure 2: Comparison of sail materials: Acceleration distance needed to reach 0.05 c as a
function of spacecraft mass.

7
to be corrected by an optical system and were therefore not considered in
the analysis, as explained in Section 4.
The results of this analysis are presented in Figure 2. The curves clearly
show the dominance of the graphene sandwich compared to Mylar and a
graphene monolayer. Therefore, a graphene sandwich light sail has been
chosen for the Dragonfly propulsion system.
In order to increase the performance of the material even further, we
examined the idea of superimposing more than one sandwich layer. This
concept reduces the transmittance of the sail even further. By placing a
second graphene sandwich sheet behind the first layer, part of the transmitted
light will be absorbed or reflected by the second sheet, thereby increasing the
total impulse transferred to the structure. Of course, one has to take into
account that reflected light is absorbed by the front sheet as well, effectively
reducing the total impulse, in order to arrive at the final values for equivalent
reflectivity and absorptivity of the system. This idea of superimposing layers
is visualized in Figure 3.
In case of two layers, the impulse component f that is transferred to the
structure is

f = (α + 2ρ) + τ (α + 2τ ρ) − τ ρ(α + 2ρ)

+ τ ρ2 (2ρ + α) − τ ρ3 (2ρ + α) + ... =
 
∞
(α + 2ρ) 1 + τ (−ρ)n = (1)
n=0
1+ρ+τ (α + 2ρ)(2 − α)
(α + 2ρ) =
1+ρ 1+ρ
For a larger number of layers, we used an approximation taking into ac-
count only first order terms, leading to Equation 2. This expression converges
asymptotically to the analytical result for high numbers of layers κ and has
a maximal error of 4 % for κ = 1 in the case of a graphene sandwich.


κ−1
f = (α + 2ρ)(1 − ρτ ) τn (2)
n=0

The force on the sail FLsail is proportional to f as shown in Equation 5 and

hence increases with κ. At the same time, every additional sheet increases
the mass of the system linearly

8
100%

ʏ
1
ʌ ʏ2
2
ʌʏ
1: 2ʌ + ɲ
ʌʏ2 2: (2ʌ + ɲ)ʏ
3 3: -(2ʌ + ɲ)ʏʌ
ʌʏ2 ʌ2ʏ2
4
4: (2ʌ + ɲ)ʏʌ2
ʌ3ʏ 5: -(2ʌ + ɲ)ʏʌ3

ʌ4ʏ
5
ʌ3ʏ2

Figure 3: The effect of two layers on the sail performance.

9
 m = mLsail + mrest = κ · mlayer + mrest (3)
where mlayer represents the mass of a single graphene sandwich layer with
alkali metals and mrest the remaining mass of the system, excluding the laser
sail.
With this consideration, the acceleration of the sail aLsail exhibits a max-
imum for a specific number of layers, as it is defined as
FLsail
aLsail = (4)
m
and both m and FLsail are proportional to κ as shown in Equations 3, and 5
respectively. The number of layers is thus an optimization parameter when
designing the mission architecture. It was included in the simulation of the
system, as described in Section 5.

4. Mission Analysis
The Dragonfly mission is designed with the purpose of reaching Alpha
Centauri and performing scientific measurements in the target star system.
This requirement adds some complexity to the mission architecture, since it
implies the necessity for deceleration of the spacecraft. The mission outline is
thus discretized into three separate phases preceding the scientific operations
within the target system: acceleration, cruising phase, and deceleration.

4.1. Acceleration
Acceleration takes place exclusively by means of laser-powered propul-
sion, using a laser module placed on the Moon and a light sail aboard the
spacecraft. As described in Section 2, the laser beam power is restricted to
100 GW. The force exerted on the light sail is described by
S
FLsail = f A (5)
c
where f = (α + 2ρ) in the case of a single layer or equal to the expression
in Equation 2 for a higher number of superimposed sheets. A is the area of
the sail, c the speed of light, and S the beam intensity at the location of the
spacecraft. It can be expressed as
P
S= 2
(6)
πrlaser

10
with rlaser being the laser beam radius at the location of the spacecraft.
Assuming perfect pointing accuracy and no divergence losses, rlaser remains
constant throughout the whole mission, thereby producing a constant force
on the sail. Realistically, this radius will increase with the distance from
the laser source due to free space losses, producing a force which decreases
quadratically with the distance. This effect was not modelled in the present
analysis for simplicity reasons.
The pointing accuracy of the laser system θ defines the maximum distance
at which the sail can still be accelerated by the laser beam according to
d
rmax = (7)
2 · tan θ
where d is the diameter of the light sail. It is clear that the concept
of laser propulsion is not effective at large distances from the source unless
pointing accuracy is sufficient. On the other hand, a larger acceleration
distance ensures a higher cruise velocity and can thus reduce the mission
duration. As described in Section 3, the reflectivity of the alkali metal doped
graphene sandwich is 5 % and its absorptivity 40 %. Larger acceleration
distances (and hence better pointing accuracies) are thus preferred to reach
high cruise speeds. It is important to mention that the highest achievable
cruise speed is not the optimal solution for the mission design. This is due to
the fact that deceleration to orbital velocities requires a deceleration system
whose mass depends directly on the cruise velocity. Therefore higher cruising
speeds increase the overall mass that needs to be brought on board. This
result will be further elaborated on in Section 5.
4.2. Cruising
The requirements of the Dragonfly competition state that the trip dura-
tion should not exceed 100 years. This means that the cruise speed should
be higher than 4.35 %c, assuming that the distance to Alpha Centauri is
4.35 light years.
After the optimal cruise speed has been achieved, the laser sail is no
longer useful for the mission and is ejected from the spacecraft. At the same
time, the laser system ceases operation. The cruise velocity is retained until
the deceleration phase begins.
4.3. Deceleration
During the deceleration phase, the speed of the spacecraft has to be grad-
ually reduced to a velocity sufficient for orbital insertion into the target sys-

11
tem. This implies that a Δv almost equal to the one gained by the laser
propulsion has to be imparted on the spacecraft.
Deceleration methods involving propellant consumption (chemical or elec-
tric) are ineffective because they require extreme amounts of propellant mass
to be stored on board and also reduce the efficiency of the laser propulsion sys-
tem, since the extra inertia upon acceleration leads to a significantly smaller
cruising speed.
The system chosen in the present analysis relies on the combination of a
magnetic sail [8] and an electric sail [9]. The operating principle of the two
sails is very similar, since they utilize magnetic or electric fields to deflect
the trajectories of incoming ions to reduce the speed of the spacecraft. In
the case of interstellar travel, the ions of the interstellar plasma (which are
travelling towards the spacecraft in its moving coordinate frame) are the ones
producing this force.
According to Freeland [10], magnetic sails produce a force equal to
 1
 23
2 2
FM sail = 0.354π mp no μ IR ν
2 (8)
where mp is the mass of the proton, no the number density of interstellar ions,
μ the free space permeability, I the current through the sail, R its radius,
and v its speed. The properties of interstellar plasma, and hence no , pose
a big source of uncertainty for the performance of the magnetic sail. In the
present analysis, a conservative value was used with no = 0.03 cm−3 [11].
Equation 8 states that magnetic sails are effective for higher speeds and
that their force drops asymptotically to zero for lower velocities, making it
difficult for the spacecraft to enter an orbit around a star system.
Electric sails on the other hand demonstrate a more complex force-velocity
dependency, according to Equation 9 [9]:

3.09 · mp no v 2 ro
FEsail = N L   2   (9)
exp meVp vo ln rrwo −1

with N being the number of tethers, L their length, Vo the voltage of the
sail, e the charge of the electron, rw the wire radius, and ro the double Debye
length λD , given by Equation 10:

o kb T e
ro = 2λD = 2 (10)
no e 2

12
 0.5
Voltage = 250 kV
0.45
Voltage = 750 kV
0.4 Voltage = 1500 kV
Voltage = 2500 kV
0.35

0.3
Force [N]

0.25

0.2

0.15

0.1

0.05

0
0 2 4 6 8 10
Velocity [x0.01 c]

Figure 4: Force on an electric sail as a function of velocity for different voltages.

13
 In the Debye length definition, o is the electric permittivity of vacuum,
kb the Boltzmann constant and Te the electron temperature of the interstellar
plasma.
A qualitative description of this profile can be seen in Figure 4. The force
at higher speeds is almost zero and increases to a maximum value as the
velocity decreases. After this peak, the force begins to decrease and reaches
zero when the sail is stationary relative to the incoming ion flux. In order
to decelerate from a large cruise velocity with an electric sail, the applied
voltage has to be high enough to ensure that the peak deceleration occurs
close to the cruise speed. This of course implies that the mass of the sail
increases, leading to a smaller acceleration magnitude.
The combination of the two components was found to be more effective
than each of the individual systems operating in the absence of the other.
The results from Perakis et al. [12] describe how the combination of the two
sails leads to a better performance and were used within the frame of the
mission design.
The combination of the two sails relies on minimizing the effects of their
respective disadvantages. The magnetic sail is used for the first stage of the
deceleration. As soon as the deceleration becomes smaller than the one that
the electric sail can produce, the magnetic sail is detached, and deceleration
using electric tethers takes over. The designs of the magnetic sail (radius and
current), of the electric sail (operating voltage and tether length), as well as
the velocity at which the latter takes over, have to be chosen in a way that
ensures the smallest deceleration duration and distance, while achieving the
lowest mass possible.
In the following, we describe a representative case demonstrating how
the combination of the two sails in series can bring additional benefits to the
mission. In this test case, the mass of the spacecraft ms/c was chosen to be
approximately equal to the launch mass of Voyager 1 (750 kg). Voyager is
a space probe which was launched to perform flybys of Jupiter, Saturn and
Titan, and continued to explore the boundaries of the outer heliosphere [13].
Voyager 1 is the man-made probe closest to entering the interstellar space
[14]. Therefore, it is relevant to consider how its deceleration would look like
in the case of a mission to another star system.
In this analysis, only the deceleration phase of the mission was exam-
ined for a given cruise speed vcruise = 0.05 c. The target speed is set to be
35 km/s, corresponding to the approximate orbital speed at a distance of one
astronomical unit around Alpha Centauri A, which has a mass of 1.1 M

14
[15]. We examined three configurations:

1. Deceleration with Msail only,

2. deceleration with Esail only, and
3. deceleration using a combination of Msail and Esail.

For an effective comparison between the separate architectures, the total

mass of the deceleration system was constrained to remain below mdecel =
7500 kg or else mdecel = 10 · ms/c . We optimized each system by minimizing
the total deceleration duration required to bring the probe from vcruise to
vtarget . The results of this analysis are presented in Table 1.
Figure 5 shows the velocity profiles for each of the deceleration architec-
tures. One observes that the combination of the two sails results in a faster
deceleration. The magnetic sail seems is very effective at higher speeds, out-
performing the other two methods for the first 27 years of operation. After
this point, it reaches its performance limit and the velocity curve flattens out.
The findings of this short test case show why the deceleration architecture
of the Dragonfly mission was based on a combination of the two sails.
It was also examined whether a solar sail deceleration in parallel to the
combination of Msail and Esail could further improve the performance of the
system. In principle, the laser-powered light sail used during the acceleration
phase could be kept on board during the cruising phase and later on utilized
for deceleration purposes using the photon pressure stemming from the Alpha
Centauri system. However, it was discovered that the efficiency of the system
is quite low. The additional force from the solar sail becomes significant only
in very short distances from the target star system and does not compensate
for the extra mass that needs to be decelerated as well (the light sail itself).
A detachment of the light sail directly after the acceleration phase was hence
found to be more effective, leading to a deceleration phase based purely on
the electric and magnetic sail.

Deceleration method Duration (years)

Pure Msail 39.7
Pure Esail 34.9
Tandem (Msail and 28.8
Esail)
Table 1: Deceleration duration for different deceleration methods.

15
 1
10
Combination Msail and Esail
Msail
Esail

0
10
Velocity [x0.01 c]

−1
10

−2
10
0 5 10 15 20 25 30 35 40
Time [years]

Figure 5: Comparison of deceleration methods: Velocity profile over time [12]

.

16
5. Simulation model
A simulation model was developed with the purpose of determining the
shortest possible mission duration for every spacecraft configuration. In this
model, the three phases described in Section 4 were simulated and optimized.
The cost function of the optimization problem was defined as the total mis-
sion duration Tmission with:

Tmission = Taccel + Tcruise + Tdecel (11)

where Taccel represents the acceleration duration, Tcruise the duration of the
cruise phase, and Tdecel the time required for deceleration. The parameters
to be optimized for the minimization of Tmission are

1. the number of light sail layers, κ,

2. the diameter of the light sail, d,
3. the distance of acceleration, raccel ,
4. the current of the magnetic sail superconductor, I,
5. the radius of the magnetic sail, R,
6. the total length of the electric sail tethers, N · L,
7. the voltage of the electric sail, Vo , and
8. the velocity at which the electric sail starts to dominate deceleration,
vswitch .

For a specific configuration of κ, d, I, R, N · L, and Vo , we define the

masses of the three sails mLsail , mM sail , and mEsail . κ and d also determine
the force on the light sail. Hence, the acceleration aLsail can be calculated
according to
FLsail
aLsail = (12)
mLsail + mM sail + mEsail + ms/c
where ms/c is the spacecraft mass being sent to the target system. It consists
of the spacecraft bus and the scientific payload. The laser sail mass mLsail is
given according to

d2
mLsail = κ · mlayer = κ · πσ (13)
4
where σ stands for the areal density of the chosen sandwich material. The
minimal diameter d is dictated by the maximal power density that the sail

17
can withstand without melting. A constraint for the acceleration distance
raccel is the pointing accuracy of the laser system θ according to
d
raccel ≤ (14)
2 · tan θ
For larger distances it cannot be ensured that the Moon-based laser sys-
tem can direct its power onto the sail and propel it further, as a better point-
ing accuracy would be required. For a specific distance of the acceleration
phase, the achievable cruise speed is expressed by
dv dv dr dv
aLsail = = = v
dt 
dr dt dr
raccel (15)
⇒ vcruise = 2 aLsail (r)dr + v(r = r0 )2
r0

where r0 is the distance of the probe from the laser source at the beginning
of the mission and v(r = r0 ) is its speed relative to the target system at
this point. The direct integration shown in Eq. 15 can be used, since a 1-D
acceleration path is assumed for the probe. The probe is accelerated along the
line of sight between the laser and the sail and travels in the absence of other
external forces. No change of direction takes place during the acceleration,
cruising and deceleration phases and hence a motion along a straight line is
modeled. Following a similar procedure, the acceleration time Taccel is given
by
raccel
dr dr dr
v= ⇒ dt = ⇒ Taccel =
dt v r0 v(r)
raccel (16)
dr
Taccel =
r
r0 2 r0
aLsail (r)dr + v(r = r0 )2
Knowing the cruise speed and the masses of the magnetic and electric
sails, we can calculate the magnitude of acceleration during the deceleration
phase. The switch from magnetic to electric sail deceleration occurs at a
speed vswitch . For a given vswitch , the acceleration becomes
⎧
⎪ FM sail
⎪
⎨m + m , for v > vswitch
s/c M sail + mEsail
adecel = (17)
⎪
⎪ FEsail
⎩ , for v < vswitch
ms/c + mEsail

18
where FM sail and FEsail are described by Equations 8 and 9, respectively. A
further constraint is that the acceleration profile should remain continuous
at the switch from magnetic to electric deceleration. Hence

aM sail (v = vswitch ) = aEsail (v = vswitch ) ⇒

FM sail (v = vswitch ) FEsail (v = vswitch ) (18)
=
mM sail + mEsail + ms/c mEsail + ms/c
With the profile of the acceleration magnitude adecel as a function of speed,
we can calculate the duration and distance required for sufficient deceleration
using
vswitch vtarget
dv dv
Tdecel = + (19)
vcruise aM sail (v) vswitch aEsail (v)
vswitch vtarget
v dv v dv
rdecel = + (20)
vcruise aM sail (v) vswitch aEsail (v)
vtarget is the predefined end velocity of the probe before it enters an orbit
in the star system. For the distances traveled in the acceleration and de-
celeration phases, a constraint has to be applied to ensure that they do not
exceed the total distance to the target system rtarget :

raccel + rdecel ≤ rtarget (21)

In the case of Alpha Centauri, rtarget is equal to 4.35 light years. Finally,
the cruising phase is modeled with the information of raccel , rdecel and vcruise :

rcruise = rtarget − raccel − rdecel (22)

rcruise
Tcruise = (23)
vcruise
With this expression, the model is complete and the cost function Tmission
is fully defined according to Equation 11. The optimization objective is
defined as:

Tmission = min! (24)

It is evident that the function for Tmission is non-linear. Since the cal-
culation of the function’s gradient would require extra computational effort,
we minimized the cost function with a pattern search method similar to the
“direct search” method proposed by Hooke [16].

19
 4
10
<100 years
100−250 years
250−500 years
3 500−1000 years
10
1000−2000 years
Spacecraft mass [kg]

>2000 years

2
10

1
10

0
10 −12 −11 −10 −9 −8
10 10 10 10 10
Pointing accuracy [rad]

Figure 6: Dependency of the mission duration on spacecraft mass and pointing accuracy.

6. Results
The design guidelines for the Dragonfly mission included a maximum mis-
sion duration of 100 years [4]. In this paper, we generalize our analysis to
include different mission classes and mission durations. We established a de-
sign map, showing the importance of the laser system’s pointing accuracy and
the spacecraft mass for the overall mission duration. The common denomi-
nator for all the combinations described here was the technical description of
the subsystems and the separation of the mission into three phases, including
the separate magnetic sail and electric sail deceleration phases. The calcu-
lations were carried out with the optimization process described in Section
5.
We already stressed the importance of the pointing accuracy for the over-
all mission design in Section 4. A better accuracy leads to a longer accelera-
tion distance and hence to a larger cruising speed. The imparted Δv is also

20
influenced by the spacecraft mass, since it defines not only the initial acceler-
ation, but also the design point of the deceleration system, which serves the
purpose of decelerating the spacecraft into orbit around the target star. In
our analysis, the term “spacecraft mass” describes the mass of the spacecraft
bus sent to Alpha Centauri with all its subsystems and scientific payload, but
excluding the laser sail, and the magnetic and electric sails. We calculated
the optimum solution (i.e. the shortest possible mission duration) for each
combination of pointing accuracy and spacecraft mass (see Figure 6).
It is evident from that missions matching the profile of the Dragonfly
design competition (i.e. duration shorter than 100 years) are possible but
limited to low spacecraft mass and high pointing accuracy of the laser system.
Given a very accurate pointing system, achieving 10−12 rad of deviation or
better, spacecraft masses up to 2000 kg can be sent to Alpha Centauri within
a century. Decreasing the pointing accuracy by an order of magnitude re-
stricts the mass of the spacecraft to a maximum of 1 kg. Lower pointing ac-
curacies fail to fulfill the 100 year requirement. If, however, this requirement
is considered to be secondary, other possible design points can be identified.
A 1000 kg spacecraft can reach the star system within 250 years after launch,
assuming that the pointing accuracy is close to 10−10 rad. Figure 7 shows
the dependency of the mission duration on the pointing accuracy.
The general trend shows that the design chosen for the Dragonfly mis-
sion promotes missions with smaller spacecraft mass. Larger mission classes,
although potentially more significant from a scientific point of view, require
much more accurate pointing, and cannot be effectively achieved with a laser-
powered propulsion system. For the propulsion to become efficient for these
mission classes, alternative sail materials will have to be utilized, combining
higher reflectivity with lower densities.

7. Spacecraft Design
In our mission analysis simulation we determined the feasible spacecraft
mass range for a 100 year long mission to Alpha Centauri to be 1 kg to 2000 kg.
Based on the current development and limitations in miniaturization of sci-
entific instrumentation and spacecraft systems explained in Section 8, we
concluded that a swarm of small, CubeSat-sized spacecraft cannot achieve
a scientific gain justifying the effort and cost required to conduct an inter-
stellar mission. Therefore, we designed a single spacecraft with a maximum

21
 4
10
Spacecraft mass = 10 kg
Spacecraft mass = 100 kg
Spacecraft mass = 1000 kg
Mission duration [years]

3
10

2
10

1
10 −12 −11 −10 −9 −8
10 10 10 10 10
Pointing accuracy [rad]

Figure 7: Mission duration as a function of pointing accuracy for three spacecraft masses.

possible mass of 2000 kg (light, magnetic, and electric sails not included).
Table 2 summarizes the mass budget for the designed spacecraft.
For the preliminary design of the spacecraft (see Figure 8) a payload

Systems Mass (kg)

Scientific payload 250
Spacecraft bus 1750
Light sail 4290
Magnetic sail 1970
Electric sail 1690
Total 9950
Table 2: Mass budget for the Dragonfly spacecraft.

22
Figure 8: Dragonfly spacecraft design. The sail diameter has been reduced for visualiza-
tion.

fraction of 12.5 % (250 kg) has been assumed. The remaining 1750 kg are
available for the spacecraft bus, containing

1. sail deployment mechanisms,

2. a thermal control system,
3. a position determination and control system,
4. a data handling system,
5. a power system, and
6. a communication system.

The sail deployment mechanism is required to deploy the light, magnetic,

and electric sails when required and ejects them after they are no longer
used. The thermal control system regulates the temperature of the Dragon-
fly systems by rejecting heat from the laser during acceleration and providing
heat for critical systems during transfer across interstellar space. The posi-
tion determination and control system uses an on-board telescope—also used
for scientific measurements—to determine its position and hall thrusters for
course corrections. The data handling system processes and stores the col-
lected data from scientific measurements and prepares them for their transfer
to Earth by the communication system.

23
 The power system is the core element of the spacecraft bus, supplying the
other systems with energy during the different operational phases, namely
acceleration, cruise, deceleration, and target star operations. For this design
we chose a combination of photovoltaic cells and electromagnetic tethers
as power supply. The photovoltaic cells are designed to provide sufficient
power for communication and scientific observations in the target star system.
During acceleration, they transform parts of the laser light into electrical
energy. During cruise the electric sail is used as an electromagnetic tether.
This concept relies on long tethers moving through the interstellar magnetic
field creating a voltage difference between the end of the tether and the
spacecraft and collecting ions from the interstellar medium [17]. With these
two systems, the scientific instruments can operate during transfer and within
the target star system.
To send the gathered data back to Earth, an interstellar communication
system is required. The ability to send data over a distance of several light
years requires either large transmitter and receiver antennas, or a large power
system. For the Dragonfly mission we chose an optical communication system
that uses the acceleration laser optics as its receiver. The diameter of the
optical transmitter on the spacecraft is limited by the achievable pointing
accuracy. With these assumptions a trade-off analysis involving laser system,
transmitter, and power system mass has been conducted. The resulting
system is able to send 100 bit/s over 4.35 light years and weighs 1000 kg,
including the power supply.

8. Science
Manned and unmanned missions to other stellar systems have—due to the
advanced technologies and long time spans required—never been undertaken
by the human civilization. A successful mission to a nearby system would
therefore be an achievement in itself, propelling humanity further ahead in
its evolutionary voyage. An interstellar mission without extensive scientific
payload would, however, be a questionable endeavor. Large-scale space pro-
grams either require the potential for commercial or scientific benefit, or
the full financial support of governments. Analyzing potential stakeholder
groups, Hein et al. come to the conclusion that political stakeholder scenar-
ios are not likely to result in continued support for major space programs
required to sustain an interstellar mission [18]. They state that “the main di-
rect output of an interstellar exploration mission will be knowledge about the

24
universe”—a conclusion that highlights the importance of identifying possi-
ble scientific areas to be investigated. For maximum return on investment,
a mission should be designed to cover as many of these areas as possible.
According to Webb [19] and Crawford [20], the main areas of interest are

1. studies of the interstellar medium,

2. astrophysical studies of the target star(s),
3. investigations of planetary system(s), and
4. biological studies of life forms.

Studies of the interstellar medium take place en route to the target star
system and provide valuable information for the understanding of physical
processes in our universe. The scientific value of astrophysical studies in-
creases with observation time. In the near future, basic observations of the
target star and its planetary system may be conducted using Earth-based
observatories that implement new technologies—such as the European Ex-
tremely Large Telescope [21]. In order to study stars and planetary systems
in detail, the spacecraft will have to decelerate or even enter a stellar or
planetary orbit in order to increase the observation time. For astrobiological
studies deceleration is inevitable as most respective measurements require
surface operations. Crawford concludes that deceleration cannot be avoided
to produce new and significant scientific results [20]. Hein et al. identified
planetary geology and astrobiology as the most relevant science stakeholder
groups, which leads to a high priority of surface missions [18].

8.1. Instrumentation
The instrumentation of the Dragonfly spacecraft is necessarily based on
the technology available today. Near-future developments will likely result in
the further miniaturization of some instruments, albeit to different degrees.
Many currently available devices are already being built close to the minimum
dimensions possible, as miniaturization is in many cases limited by physics
(e.g. optics for telescopes).
Studies of the interstellar medium require the least sophisticated instru-
ments. Simple Geiger or solid-state detectors can be used to assess the radia-
tion level during the travel phase. Much more advanced miniaturized particle
physics experiments, capable of providing more detailed information, are in
development and will be tested within the next few years [22]. Astrophysi-
cal studies of stars today are mainly focused on optical observations in the

25
ultraviolet, visible, and infrared ranges. However, miniaturization of opti-
cal instruments is severely limited by physics, as the achievable resolution
directly depends on the observable wavelength and the mirror or lens di-
ameter. At a fixed wavelength, achieving higher resolution inherently means
using either larger mirrors or lenses, or using an array of smaller instruments.
Both options result in an increased system mass, of which a large portion
are the optics themselves. Depending on the orbit that the probes may be
deployed to, decreasing resolution requirements will allow smaller systems—
but bringing the probes to lower orbits requires significantly more fuel and
sophisticated propulsion systems.
The most important instruments for the direct observation of planets are
optical telescopes. Recent developments have shown that spacecraft with a
sub-10 m resolution in low-Earth orbit can be built in small form factors.
The Dove satellites of Planet Labs are a notable example, achieving a 3 to
5 m resolution in a 3-unit CubeSat platform [23]. However, the usability of
the generated data is being questioned by the authors, both because of the
limited resolution and because the satellites do not deliver data in enough
wavelength regimes. Flying Dove-type satellites in formation to build an
array and increase their overall resolution requires a better attitude control
system and a propulsion system, which is currently completely missing. The
propellant for formation keeping is in the order of a few kilograms at best,
even if highly efficient, not-yet-available systems are assumed. Also, the Dove
satellites are designed for short lifetimes (in the order of a year or two), as
are basically all miniaturized satellites today. Even assuming that making
these systems radiation-hard and adding the required redundancies will be
possible without making them heavier, the total mass would still be in the
order of 10 kg. There is not much, if at all any, room left to build these
platforms even smaller, as the main limitations are not imposed by today’s
technology, but mostly by the laws of physics.
Staehle et al. [24] give a few examples for realistically achievable inter-
planetary science missions on 6-unit CubeSats, including imaging spectrome-
ters, magnetometers, and radio antennas. Recent advances in manufacturing
technology and high-resolution imaging sensors led to the miniaturization
of camera systems in the computer and hand-held device industry. These
cameras mostly have lenses with focal lengths smaller than 50 mm, which
render them practically useless for any scientifically meaningful observation
duty in orbit (compare to 1.14 m focal length of the Dove satellites). Mason
et al. [25] present a miniature x-ray spectrometer for solar observations in

26
3-unit CubeSat form factor, which seems to be a possible candidate tech-
nology for the Dragonfly mission. The search for biological markers and
live forms will almost certainly require to land instruments on the surface
of planetary bodies, as does the determination of surface composition. Al-
though measurements of trace gases in a planet’s atmosphere can provide
hints about the possibility of having live on the surface, many phenomena
have to be investigated in situ. This requires heat shields, rocket motors or
other means of propulsion, and parachutes, adding significantly to the mass
of the spacecraft.
From the four areas of interest identified above, studies of the interstellar
medium and astrophysical studies of the target star seem to be feasible to
some extent with a small-scale mission. We have limited ourselves to analyz-
ing these two categories of instruments, as at least some reference material
is available. A full analysis of all categories was not within the scope of this
paper. It is, however, evident that the payload mass of 250 kg is a significant
constraint that cannot be easily overcome. All of the instruments presented
here (with the exception of the x-ray spectrometer) require the spacecraft
to be in a low planetary orbit. To bring the spacecraft from the solar or-
bit assumed in Section 4 to a planetary orbit requires significant amounts of
propellant, potentially using up most of the available mass. Although propel-
lantless methods could potentially be utilized for an orbital insertion (Esail,
solar sail), they can only be applied in specific configurations of the exo-
planet’s orbital geometry. Since no information about the planets’ orbits is
known a priori, including a conventional chemical or electric propulsion sys-
tem is necessary for the flexibility of the system. We therefore conclude that
Dragonfly could, for example, be equipped with a simple particle detector,
an optical telescope, and some form of spectrometers for solar observations,
remaining in its initial solar orbit. Whether the data that can be gathered
with such instrumentation would be worth the financial investment necessary
to realize the mission is questionable from our point of view. Advances in
Earth- or Solar System-based observation technologies may provide data of
similar quality in the not-too-distant future.

9. Conclusion
For the Project Dragonfly design competition the feasibility of an inter-
stellar light sail propelled probe has been analyzed. The mass that can be
sent and decelerated into an orbit around the Alpha Centauri star system

27
depends strongly on the performance of the laser system, the sail material,
and the transfer time constraint.
The critical performance criteria identified for the laser system in this
analysis are the achievable pointing accuracy and the quality of the beam
determining the spread losses. The system does not rely on a single large
lens, because a system with a single lens that can focus the beam on the
sail over large distances would be too large and therefore require tremendous
efforts to be positioned close to the laser system.
The choice of sail material has a large impact on the acceleration behavior
of the spacecraft. A graphene sandwich structure has been selected as the
best option due to its low density. The performance may be improved if a
material will be developed that combines a higher reflectivity than graphene
with an equal or smaller density.
To ensure valuable scientific measurements, a deceleration system is nec-
essary in order to bring the probe down to orbital speeds. A combination
of magnetic and electric sails was utilized for this purpose, which was found
to outperform each of the two systems operating on their own. The use of
the light sail for deceleration purposes (using the light from Alpha Centauri)
was also examined, but was discarded due to its low efficiency.
The last parameter impacting the maximum mass of an interstellar space-
craft is the mission time. We conclude that a maximum of 250 kg of scientific
payload can be sent to Alpha Centauri within a timeframe of 100 years. In
accordance with the analysis in Section 8, an instrument capable of mea-
suring the properties of the interstellar medium should be included. As a
second instrument a telescope has been selected to observe the target star
from a stellar orbit. Sending planetary observers and landers to Alpha Cen-
tauri requires more than 250 kg of payload, assuming current and near-future
technology. Miniaturization of instruments is limited by physics or has not
progressed to a point where it can improve the scientific gain of a 250 kg
payload.
Alternative solutions to increase the payload mass are therefore a larger
laser beam power or a longer transfer duration. As the results from Section 6
show, longer mission times can significantly increase the possible spacecraft
mass. As shown in Figure 7 an increase of the mission time to 150 years
could increase the possible system mass by a factor of 100.
To summarize, we conclude that an interstellar laser-propelled light sail
probe using current and near-future technology is feasible under the assump-
tions made in this analysis. However, the scientific measurements that can

28
be conducted by a mission of this size are fairly limited and might not justify
the cost and effort that is necessary to construct and operate the propulsion
laser system.

10. Acknowledgements
The authors would like to thank the Initiative for Interstellar Studies (i4is)
for organizing the Project Dragonfly Competition, which gave the inspiration
for the present study. Specifically, the authors would like to express their
gratitude to Andreas M. Hein for the organization of the competition as well
as for his input and comments.

References
[1] K. F. Long, M. Fogg, R. Obousy, A. Tziolas, A. Mann, R. Osborne,
A. Presby, Project Icarus - Son of Daedalus - Flying Closer to Another
Star, JBIS 62 (2009) 403–414. arXiv:1005.3833.

[2] R. L. Forward, Roundtrip Interstellar Travel Using Laser-Pushed

Lightsails, Journal of Spacecraft and Rockets 21 (2) (1984) 187–195.
doi:10.2514/3.8632.

[3] R. L. Forward, Starwisp - An Ultra-Light Interstellar Probe, Journal of

Spacecraft and Rockets 22 (3) (1985) 345–350. doi:10.2514/3.25754.

[4] A. M. Hein, K. F. Long, R. Swinney, R. Osborne, A. Mann, M. Ciupa,

Project Dragonfly: Small, sail-based spacecraft for interstellar missions,
manuscript submitted for publication to the Journal of the British In-
terplanetary Society (2016).

[5] A. Boesgaard, W. Hagen, The age of alpha centauri., The Astrophysical

Journal 189 (1974) 85–87.

[6] T. Taylor, R. C. Anding, D. Halford, G. L. Matloff, Space based en-

ergy beaming requirements for interstellar laser sailing, in: BEAMED
ENERGY PROPULSION: First International Symposium on Beamed
Energy Propulsion, Vol. 664, AIP Publishing, 2003, pp. 369–381.

[7] G. L. Matloff, Graphene: The Ultimate Interstellar Solar Sail Material?,

JBIS 65 (2012) 378–381.

29
 [8] R. M. Zubrin, D. G. Andrews, Magnetic Sails and Interplanetary
Travel, Journal of Spacecraft and Rockets 28 (2) (1991) 197–203.
doi:10.2514/3.26230.

[9] P. Janhunen, A. Sandroos, Simulation study of solar wind push on a

charged wire: basis of solar wind electric sail propulsion, Annales Geo-
physicae 25 (3) (2007) 755–767. doi:10.5194/angeo-25-755-2007.

[10] R. M. Freeland, M5: Secondary Propulsion: Mathematics of Magsails,

in: Project Icarus, 2012.

[11] I. A. Crawford, Project Icarus: A review of local interstellar medium

properties of relevance for space missions to the nearest stars, Acta
Astronautica 68 (7–8) (2011) 691–699.

[12] N. Perakis, A. M. Hein, Combining Magnetic and Electric Sails

for Interstellar Deceleration, Acta Astronautica 128 (2016) 13–20.
doi:http://dx.doi.org/10.1016/j.actaastro.2016.07.005.

[13] C. E. Kohlhase, P. A. Penzo, Voyager mission description, Space Science

Reviews 21 (2) (1977) 77–101. doi:10.1007/BF00200846.
URL http://dx.doi.org/10.1007/BF00200846

[14] G. Gloeckler, L. A. Fisk, Has voyager 1 really crossed the heliopause?,

Journal of Physics: Conference Series 577 (1) (2015) 012011.

[15] P. Demarque, D. Guenther, W. F. van Altena, The case of Alpha

Centauri-Mass, age and p-mode oscillation spectrum, The Astrophys-
ical Journal 300 (1986) 773–778.

[16] R. Hooke, T. A. Jeeves, “direct search”solution of numerical and statis-

tical problems, Journal of the ACM (JACM) 8 (2) (1961) 212–229.

[17] G. L. Matloff, L. Johnson, Applications of the electrodynamic tether to

interstellar travel, Journal of the British Interplanetary Society 58.

[18] A. M. Hein, A. C. Tziolas, R. Osborne, Project Icarus: Stakeholder

Scenarios for an Interstellar Exploration Program, Journal of the British
Interplanetary Society 64 (2011) 224–233.

30
[19] G. M. Webb, Project Deadalus: Some Principles for the Design of a
Payload for a Stellar Flyby Mission, in: A. R. Martin (Ed.), Project
Daedalus - Final Report, Journal of the British Interplanetary Society
Supplement, 1978, pp. 149–161.

[20] I. A. Crawford, The Astronomical, Astrobiological and Planetary Sci-

ence Case for Interstellar Spaceflight, Journal of the British Interplane-
tary Society 62 (2010) 415–421. arXiv:1008.4893.

[21] M. Lyubenova, M. Kissler-Pating (Eds.), An Expanded View of the Uni-

verse - Science with the European Extremely Large Telescope, European
Southern Observatory, 2009.

[22] T. Pöschl, M. J. Losekamm, D. Greenwald, S. Paul, A Novel CubeSat-

Sized Antiproton Detector for Space Applications, in: 34th International
Cosmic Ray Conference, The Hague, 2015.

[23] Planet Labs, Planet Labs Specifications: Satellite Imagery Products

(2015).

[24] R. Staehle, D. Blaney, H. Hemmati, et al., Interplanetary CubeSat Ar-

chitecture and Missions, AIAA SPACE 2012.

[25] J. P. Mason, T. N. Woods, A. Caspi, et al., Miniature X-Ray Solar

Spectrometer: A Science-Oriented, University 3U CubeSat, Journal of
Spacecraft and Rockets 53 (2) (0) 1–12.

31