Content uploaded by Marshall Eubanks
Author content
All content in this area was uploaded by Marshall Eubanks on Sep 11, 2023
Content may be subject to copyright.
Highlights
Swarming Proxima Centauri: Optical Communications Over Interstellar Distances
T. Marshall Eubanks,W. Paul Blase,Andreas M. Hein,Adam Hibberd,Robert G. Kennedy III
•We are developing a means of interstellar travel for fast flyby missions combined with optical communications over
interstellar range, based on swarms. of picospacecraft accelerated with laser power beaming.
•Our plan brings a sequentially-launched string of probes together into a swarm using time- and velocity-on-target
dynamic techniques.
•Coherent swarms of order 1000 picospacecraft can greatly increase the data return rate over interstellar distances.
•We develop means to communicate within swarms extending over 100,000s of km.
•This laser intraswarm communication is used to synchronize clocks across the swarm.
•We discovered a simple new cost-effective type of betavoltaic isotopic power to provide onboard electricity sufficient
to drive optical coms at interstellar range for decades.
Swarming Proxima Centauri: Optical Communications Over
Interstellar Distances
T. Marshall Eubanksa,∗,W. Paul Blasea,Andreas M. Heinb,Adam Hibberdcand Robert G.
Kennedy IIId
aSpace Initiatives Inc, 106 Mahood Avenue, Princeton, 24740, West Virginia, USA
bLuxembourg University, Luxembourg, Luxembourg
cInitiative for Interstellar Studies (i4is), 27/29 South Lambeth Road, London, SW8 1SZ, United Kingdom
dInstitute for Interstellar Studies (i4is US), 112 Mason Lane, Oak Ridge, Tennessee, 37830, USA
ARTICLE INFO
Keywords:
interstellar travel
Proxima Centauri b
swarm spacecraft
laser communications
betavoltaic power
ABSTRACT
Interstellar communications are achievable with gram-scale spacecraft using swarm techniques
introduced herein if an adequate energy source, clocks and a suitable communications protocol
exist. The essence of our approach to the Breakthrough Starshot challenge is to launch a long string
of 100s of gram-scale interstellar probes at 0.2c in a firing campaign up to a year long, maintain
continuous contact with them (directly amongst each other and via Earth utilizing the launch laser),
and gradually, during the 20-year cruise, dynamically coalesce the long string into a lens-shaped
mesh network ∼100,000 km across centered on the target planet Proxima b at the time of fly-by.
In-flight formation would be accomplished using the “time on target” technique of grossly
modulating the initial launch velocity between the head and the tail of the string, and combined with
continual fine control or “velocity on target” by adjusting the attitude of selected probes, exploiting
the drag imparted by the ISM.
Such a swarm could tolerate significant attrition, e.g., by collisions enroute with interstellar dust
grains, thus mitigating the risk that comes with “putting all your eggs in one basket”. It would
also enable the observation of Proxima b at close range from a multiplicity of viewpoints. Swarm
synchronization with state-of-the-art space-rated clocks would enable operational coherence if not
actual phase coherence in the swarm optical communications. Betavoltaic technology, which should
be commercialized and space-rated in the next decade, can provide an adequate primary energy storage
for these swarms. The combination would thus enable data return rates orders of magnitude greater
than possible from a single probe.
1. Introduction
Per the direction of Breakthrough Starshot (BTS), we
conducted a systems engineering study of the challenge of
obtaining information from swarms of gram-scale probes
across 4 light years from an interstellar target, hereinafter
the planet Proxima b. We find that this task is not impossible,
and could be accomplished with six broad innovations:
∗Corresponding author
tme@space-initiatives.com (T.M. Eubanks)
ORCID (s):
T.M. Eubanks et al. Page 1 of 24
Figure 1: Artist’s impression of swarm passing by Proxima Centauri and Proxima b. The swarm’s extent is ∼10 larger than the
planet’s, yet the ∼5000-km spacing is such that one or more probes will come close to or even impact the planet (flare on
limb). It should be possible to do transmission spectroscopy with such swarms. Green 432/539-nm beams are coms to Earth; red
12,000-nm laser beacons are for intra-swarm probe-to-probe coms. Conceptual artwork courtesy of Michel Lamontagne, P.Eng.
1. Rather than individual 1-g spacecraft launched in
isolation, instead send an “operationally coherent”
(synchronized if not phase-coherent), swarm of 100s–
1000s of spacecraft in order to simultaneously and fea-
sibly transmit a reasonable number of signal photons
to Earth.
2. Forming effective swarms of order 100s–1000s of
probes at planetary encounter with a gross “time-on-
target” (ToT) technique, consisting of modulating the
initial velocity of each probe by the launch laser such
that the tail catches up with the head, in terms of
their position relative to each other, not with Earth or
Proxima b.
3. A finer “velocity-on-target” (VoT) technique based on
controlled drag imparted by the interstellar medium
(ISM) by altering the attitude of individual probes
with respect to the ISM, thus keeping swarm together
in relative and absolute position once formed. At-
titude adjustment is also necessary to minimize the
extremely high radiation dose induced by traveling
through the ISM at 0.2c. See Figure 2c. To minimize
frontal area, hence dose and erosion, we anticipate
flying edge-on most of the way, without rotation about
the roll axis, hence leading and trailing edges are
distinct, and configured differently, even though the
probe is mostly symmetric.
4. Intraswarm communication via optical means, to form
up and then use the swarm as a signaling array.
5. Applying state-of-the-art (with frequency stability
goal of 10−19 at Earth & 10−13 at Proxima) optical
clock metrology combined with a clever scheme of
time- and frequency-bandpass filtering to improve
data collection and signal-to-noise ratio (SNR) of data
return to Earth.
6. A method onboard carrying stored energy sufficient
for the entire mission, well-matched to its 2-decade
duration, and in a compact form, i.e. at nuclear energy
densities orders of magnitude greater than achievable
with any known chemical method or in-flight electro-
magnetic or photonic method, to power computation
and communication within the rigorous constraint of
total payload not to exceed (NTE) one gram. This
particular betavoltaic source material, 90Sr, is many
orders of magnitude cheaper than any other candidate,
has a moderately high technology readiness level and
could attain commercial-off-the-shelf (COTS) status
within a decade with the right incentives.
Finally, we have identified a fundamental operational
unknown that must be solved well in advance: accurately
determining the orbital position of Proxima b at least
8.5 Earth years (∼300 revolutions) ahead of launch. We
find that this can be overcome at reasonable cost through the
targeted use of gravitational microlensing observed by small
telescopes in low Earth orbit (LEO).
We also have a few lesser innovations, including:
Swarming Proxima Centauri
(a) A flotilla (sub-fleet) of probes (far left), individually fired at the maximum tempo of once per 9 minutes, departs Earth (blue) daily.
The planets pass in rapid succession. Launched with the primary ToT technique, the individual probes draw closer to one another inside
the flotilla, while the flotilla itself catches up with previously-launched flotillas exiting the outer Solar system (middle) ∼100 AU. For the
animation go to https://www.youtube.com/watch?v=jMgfVMNxNQs Hibberd (2022).
(b) Time sped up by a scale factor of 30. The last flotilla launched draws closer to the earlier flotillas; the full fleet begins to coalesce
(middle), now under both the primary ToT and secondary VoT techniques, beyond the Kuiper-Edgeworth Belt and entry into the Oort
Cloud ∼1000–10,000 AU.
(c) Time sped up by a scale factor of 3000. The fully-formed fleet, having reconfigured itself into a mesh network by autonomously applying
the secondary VoT technique over decades of cruising, has exited Sol’s Oort Cloud and flies by Proxima b at 0.2c. ∼300,000 AU.
Figure 2: Screenshots from CG animation at https://www.youtube.com/watch?v=jMgfVMNxNQs depicting launch and forming-
up of the swarm.
1. A novel repurposing of the 100-gigawatt (GW) drive
laser as an “interstellar flashlight,” illuminating ob-
jects in the path of the swarm (which we assess to
be quite feasible). This external light would aid in
spotting small bodies at encounter, serve as a adjunct
monochromatic light source for spectroscopy by the
fleet during flyby, as well as provide an additional
marginal but possibly useful controlled source of illu-
mination for photography of dark objects in Proxima
system.
2. A novel structural concept, wherein most of the
heart of the device (batteries, ultracapacitors, small-
aperture inter-probe laser communication, computa-
tion) is concentrated in a 2-cm high thickened rim,
while the central disk consists of a thin but large-
aperture phase-coherent meta-material disk of flat
optics similar to a fresnel lens for both imaging the
target and communicating with Earth. (Since valuable
laser energy has been invested in accelerating both the
sail and payload to 0.2c, we shall retain the dielectric
layer on the aft face of the probe to serve as armor
when the probe has to turn full-face on to the ISM
in order to send and receive to Earth.) The layout is
reminiscent of a red corpuscle rather than the simple
featureless flat disk which seems to be the default in
the community. Refer to Figure 3a.
2. Approach
The broad goal is to send information from one or more
gram-scale spacecraft across 4+ light years. The default
reference mission for the current BTS work is contained in
Parkin (2018a). Basic assigned constraints are: 0.2c cruise
T.M. Eubanks et al. Page 3 of 24
Swarming Proxima Centauri
(a) Oblique view of the top/forward of a probe (side facing away from the launch laser) depicting array of phase-coherent apertures for
sending data back to Earth, and optical transceivers in the rim for communication with each other.
(b) Cross-sectional close-up view of one annular aperture in array of phase-coherent elements depicting ray trace from annular opening to
sensor / emitter.
(c) Oblique view of bottom / aft of a probe (side facing the launch laser) depicting dielectric boost layer and and optical transceivers in the
rim.
(d) Cross-sections of protective leading (to left) edge with sacrificial barrier, and instrumented trailing edge (to right) that contains the
betavoltaic battery sandwiches and the probe-to-probe optical transceivers. Note the electronic layer and ultracapacitor layer span the entire
perimeter to assure connectivity.
Figure 3: Various views of Probe concept.
T.M. Eubanks et al. Page 4 of 24
Swarming Proxima Centauri
velocity, 200-nm sail thickness, 4.1-m sail diameter, 3.6-
gram mass.
2.1. Our Mission Vision
Our swarming approach relies on launching a pipeline of
hundreds of probes serially, maintaining continuous contact
with each other and with Earth, and gradually, over the
course of the cruise phase, reforming the long string into a
flat lens-shaped mesh network ∼100,000 km across by the
time of encounter. Continual pinging of swarm by Earth with
the boost laser will keep mission director continually ap-
prised of the attrition rate due to collisions with microscopic
interstellar objects or other hazards. These spacecraft would
be lower-mass and faster than the femtospacecraft of Project
Andromeda Hein, Long, Fries, Perakis, Genovese, Zeidler,
Langer, Osborne, Swinney, Davies, Cress, Casson, Mann
and Armstrong (2017). Novel dynamic techniques described
below would be utilized to cohere the swarm.
We note for the record that although all probes are
assumed to be identical, implicitly in the community and
explicitly in the baseline study, there is in fact no necessity
for them to be “cookie cutter” copies, since the launch
laser must be exquisitely tunable in the first place, capable
of providing a boost tailored to every individual probe.
At minimum, probes can be configured and assigned for
different operations while remaining dynamically identical,
or they can be made truly heterogeneous wherein each probe
could be rather different in form and function, if not overall
mass and size.
2.1.1. Time on Target
“Time on target” (ToT) means that either velocities, or
paths, or both, are adjusted to bring multiple projectiles on a
target at the same time. For relativistic spacecraft the paths
are largely fixed, so ToT relies on variation of velocity; in
the remainder of this section velocities are assumed to be
scalar quantities. Suppose two probes are fired at different
times with v1being the velocity of the first probe, launched
at time T. The target is distance 𝐿away and it thus takes,
in the Newtonian approximation, the first probe a duration,
D, = L∕𝐯1to arrive, which thus occurs at a time T + D.
Suppose that the second probe is launched a time Δtafter
the first, at a velocity v2=v1+Δv. The second probe’s
velocity is adjusted by Δv so they both reach the target at
the same time, albeit with different velocities. That can be
solved, yielding Δ𝑣=𝑣(Δt / D).
If D = 20 years and Δt = 1 hour, then Δ𝑣=0.3 km/s. Δv
scales linearly with Δt, so that 10 probes launched at longer
1-hour intervals would need, for example, a 3 km/s velocity
difference between the head and tail.
2.1.2. Time and Velocity on Target
Time and Velocity on Target (VoT), adjusts both acceler-
ation and velocity, and is the fundamental means of forming
a coherent swarm in flight. We achieve this by launching the
last probes with higher speeds and using variable geometry
to give them a higher drag (deceleration), so that they slow
down to the speed of slower leading probes as they reach
their location.
A string of probes relying on the ToT technique only could
indeed form a swarm coincident with the Proxima Cen-
tauri system, or any other arbitrary point, albeit briefly.
But then absent any other forces it would quickly disperse
afterwards. Post-encounter dispersion of the swarm is highly
undesirable, but can be eliminated with the VoT technique by
changing the attitude of the spacecraft such that the leading
edge points at an angle to the flight direction, increasing
the drag induced by the ISM, and slowing the faster swarm
members as they approach the slower ones. Furthermore,
this approach does not require substantial additional changes
to the baseline BTS architecture. The analytic approximation
for exploiting the drag from the incoming neutral hydrogen
ISM is Hoang, Lazarian, Burkhart and Loeb (2017):
Δv
v≈2.5 × 10−6
𝑀 NH
1018 cm−2
0.2𝑐
v2.6Asail
1 m2 l
1𝜇𝑚 (1)
where ΔV/v is the ratio between the velocity change
due to the ISM and the initial launch velocity, 𝑀the
spacecraft mass, N𝐻the column density of interstellar
hydrogen between Earth and the target star, 𝑣the initial
spacecraft velocity, A∕𝑚𝑎𝑡ℎ𝑟𝑚𝑠𝑎𝑖𝑙 the sail surface area and
l the sail thickness. Using the analytic approximation and
baseline dimensions (v= 0.2c; M= 3.6 g; d= 4.1 m; l=
100 nm) from the Starshot RFP, we get ratios on the order
of 10−6. It can be seen in the table that the launch tempo
and thickness of aerographene are inverse linearly related
quantities.
As before, suppose two probes are fired at different times
with v1being the velocity of the first probe, launched at time
T, and the second probe be launched a time Δt after the first,
with a velocity difference of Δv. In this case, the target is a
nominal point on the trajectory where the swarm can cohere;
this formation is then retained for the rest of the mission.
Suppose that this target point is mid-way to the target. A
point 2.12 light years from Earth would be reached at a time
𝜏= 10.6 years into the mission for the first probe. Assume
that the imposed drag differential acceleration is a, and is
constant over the first half of the mission. This can be solved,
yielding
a=−2 v1Δt
(𝜏− Δt)2(2)
and
𝛿v = 2 v1Δt
(𝜏− Δt) .(3)
The operational objective is to dissipate a portion of
the velocity of leading probes by continually adjusting their
attitude hence aspect ratio and sectional density with respect
to the oncoming ISM (edge on, fully face on, or something
T.M. Eubanks et al. Page 5 of 24
Swarming Proxima Centauri
Figure 4: Commercially available magnetorquers for Cubesats
(left) and implemented as chips (right)
in between), in pitch and yaw axes, such that the hindmost
members catch up with but do not overtake the leading
members.
If the collisions with the ISM are elastic, then the reaction
would also generate some useful cross-range velocity of
order ∼1 km/sec, transverse to the direction of travel. The
VoT-Attitude Adjustment method could be initially be under
Earth’s control, but soon (due to communication latency well
before hitting the Oort Cloud), it would have to become
fully autonomous, i.e., under the control of individual probes
and eventually that of the fleet, in effect a “hive mind”.
With virtually no mass allowance for shielding, attitude
adjustment is the only practical means to minimize the
extreme radiation damage induced by traveling through
the ISM at 0.2c. Moreover, lacking the mass budget for
mechanical gimbals or other means to point instruments,
then controlling attitude and rate changes of the entire craft
in pitch, yaw, roll, is the only practical way aim onboard
sensors for intra-swarm communications, interstellar coms
with Earth and imagery acquisition / distributed processing
at encounter.
Magnetorquers. In addition to interacting with the ISM,
adjusting attitude with onboard “magnetorquers” (𝜏∼I×B)
might be feasible, as is done for 1-kg Cubesats in low Earth
orbit (LEO) right now with commercially-available space-
rated products CubesatMarket (2022). However, applying
this attitude adjustment technique to a 1-gram probe for
BTS would require reducing the mass and size of presently
available magnetorquers by many orders of magnitude.
Assessing the feasibility of this method within the mass and
time constraints of BTS will be left to the next phase of work.
Interaction with B-field in the ISM. If a probe accumulates
an electric charge, q, then Lorentz force is expected to be a
factor (F=q(v×B). Although the magnetic fields in the
ISM are thought to be very weak, the velocity vwould be
very high, so the net effect may provide a basis for attitude
control. Assessing the feasibility of this method within the
mass constraint will be left to the next phase of work.
Other Means of Interaction with ISM Assessed. Given that
the luminosity of Proxima Centauri is but ∼0.1% of Sol’s, in
other words ∼1 W m−2 at 1 AU, then photonic deceleration
by that star alone would be ineffective. For thoroughness,
we also examined and modeled the stellar radiation of 𝛼
Centauri A and B as a means of photonic deceleration, as
has been proposed by Heller, Anglada-Escudé, Hippke and
Kervella (2020). However, the bright pair A/B are a fifth of
a light-year away from Proxima, the deceleration would be
infinitesimal. We assess the photonic deceleration technique
as insufficient to the BTS task.
MEMS trim tabs. Although we envision our conceptual
probe as being entirely solid state, with no moving parts,
a simple mechanism may become necessary in order to pro-
vide a rapid way to adjust attitude. Consider the common air-
craft “trim tab” ∼but 1 cm square, actuated by microscopic
electromechanical lever or other simple MEMS machine.
This is conceptually similar to the trim tabs proposed by
R.Angel in 2006 for the school of trillions of 1-gram flyers
near the Sun-Earth Lagrange 1 (SEL1) point. Angel (2006)
A few such trim tabs could be spaced symmetrically in
pairs around the Probe’s rim to provide a means of dynamic
control. When it is raised, it interacts with the ISM to
generate an asymmetric torque on the probe; to halt the
torque, the tab is lowered; to stop the rotation caused by the
torque, the tab opposite the first tab is raised then lowered. Of
course, the “drag-brakes” will be withdrawn after the probes
reach the target zone, so that the differential acceleration a
goes to zero and the velocities thereafter remain the same.
2.1.3. Optical Communications by a Swarm
Transverse laser links between probes, which we have
modeled, would inform an even finer degree of autonomous
station-keeping. But a swarm of probes has a coordination
problem after it is launched—at first, its members will not
know where the other probes are. We have developed prelim-
inary protocols to get a coherent swarm, or a well behaved
synchronized one at any rate, and in the process, discover
how large a swarm might reasonably be. Result: A swarm
of order ∼100,000 km diameter can be established in deep
space! This will take time, but time is something we have a
lot of. “Self-knowledge” will occur in several distinct phases:
1. Discovery Like fireflies, members have to find and
establish communications with each other. We as-
sume that the swarm have been placed into a clus-
ter by the drive process, which is now over. As-
sume each probe is slowly rotating (order 0.5 rpm)
about its line of flight, defined as X-axis, per left-
hand side of the figure. For the animation go to
https://www.youtube.com/watch?v=iD3S5Qj5sR4 Blase
(2022)). The probe has multiple 1-mW pulsed optical
transceivers around its rim with 20-mm flat lenses and
quantum dot lasers, for sending and receiving to other
members, and a mutual view period (due to rotation)
of order 15 msec, then each probe should should be
detectable by its closest neighbor out to ∼5,700 km.
2. Probes as Beacons During the initial link-up phase,
the probes will have to find each other to form a
swarm. Their approximate locations can be uplinked
from Earth, but the probes will have to locate each
other for communications using internal resources.
For this, we assume that the intra-swarm commu-
nications receive optics can be adjusted for a much
wider field of view of up to 45 degrees and that
the transmit beam is formed into a thin vertical fan,
T.M. Eubanks et al. Page 6 of 24
Swarming Proxima Centauri
Figure 5: 1-gram, 70-cm glass diffraction gratings with photonic pressure trim tabs. Angel (2006)
Figure 6: Rotations for Intra-Swarm Probe-to-Probe Com, and for Swarm-to-Earth Com. Note the rotation alternately presents
the forward coms then the gold dielectric boost layer.
several arc-seconds wide by 45 degrees high. Thus,
as the probe rotates during cruise (again, the optics
are along the bottom half of the rim, over a span of
160 degrees) each probe transmits a pulsed beacon
while simultaneously scanning for its neighbors. The
output power, again, is calculated by assuming that
the total generated power of 4 mW is stored in a
super-capacitor and available for laser pulsing. Any
one probe should be able to detect another probe at
a range of ∼1 million km.
3. Convergence. Setting up a mesh network using a mod-
ification of Mobile Ad Hoc Networking is based on
Manet. Low-bit identity tokens can be passed between
member pairs. Eventually, a distance vector “map” of
visibility of other nodes can be iteratively established
by each node. This map might include angles as well
as actual distance and SNR. Nodes can then efficiently
transmit distance vector information to other nodes
such that the swarm becomes aware of itself. The goal
of convergence is to establish mesh communications
between all nodes in the swarm. The goal of alignment
is to optimize that mesh communication.
4. Alignment is still a work in progress.
5. Intra-Swarm Probe-to-Probe Link Budget
The communications link budget for intra-swarm
communications is calculated using the link-budget
method. The equations are the same as for the swarm-
Earth budget below, but only one transmitter and one
receiver are used. The transmit/receive apertures are
2-cm diameter flat-optic devices spread around the
back, or trailing, half of the outer rim of the probe,
which flies edge-on most of the time to minimize
radiation dose and erosion by particle flux induced
by the ISM at 0.2c. We assume that the optics are
electronically steerable, so that high gain may be
achieved. Again, the beam is collimated to the degree
T.M. Eubanks et al. Page 7 of 24
Swarming Proxima Centauri
Figure 7: Worksheet for Probe-to-Probe Communication.
allowed by the diffraction limit of the aperture.
Instead of being visible, the intra-fleet transmit laser
wavelength is ∼12,000 nm, or long-wave infrared
(IR). This greatly reduces the path loss, drawing less
power.
We don’t know what direction the side-looking rim
ports on the probe will be pointing at any given time,
but we can be sure the Centauri system stars will
not be directly in the field of view. Therefore, the
background is assumed to be the general sky noise
used by astronomy. The receivers are equivalent to
avalanche photo diodes (APD). Given a swarm of 300
probes with a diameter of 100,000 km, and assuming
that the probes are evenly spaced, the average distance
between probes will be about 5000 km. If we allow the
maximum path to be triple this, so that each probe can
talk to as many neighbors as possible, then with a bit
rate of 10 kHz we get 2.5 photons per time slot and a
SNR of 27 dB.
6. Swarm-to-Earth Communication using Photon Count-
ing and Link Budget Methods
2-symbol Pulse Position Modulation (2-PPM), as
shown in figure 8is widely used in optical commu-
nications links. 2-PPM uses synchronous time slots,
with two adjacent time slots for each bit. A “0” value
is sent with a pulse in slot 1 and no pulse in slot 2;
and vice versa to send a “1” value. During processing
the first slot is subtracted from the second slot; the
background will cancel out. A positive result indicates
a “1” value, a negative indicates a “0”. Pulses are 1 ns
in duration and the integration slots are 10 ns in length.
Figure 8: Pulse Position Modulation
Symbols are transmitted at 10 Hz. This technique
also has the desirable effect of greatly lowering the
unwanted background noise, since the integration time
is very short.
(Higher order PPM symbols may be used. For instance
4-PPM encodes two bits, with four discrete values,
into each symbol using a laser pulse in one of four time
slots. However, for simplicity, we did not consider
these in this paper.)
Given a final velocity of 0.2c, a transmitted wave-
length of 432 nm (blue), is Doppler-shifted to 539 nm
(green) when received at Earth.
The first table below depicts a method based on esti-
mating direct photon transmission from the swarm to
Earth. The photon flux for the laser pulse is calculated
and summed for the swarm. We assume that the optics
collimate the output beam to the degree allowed by
the Airy disk of the optics; for simplicity we assume
T.M. Eubanks et al. Page 8 of 24
Swarming Proxima Centauri
Figure 9: Worksheet for Swarm-to-Earth Communication, Photon Counting Method.
that the beam flux is uniform across the beam’s
width. Once the beam reaches Earth, we factor in the
atmospheric transmission, based on the average value
for the laser wavelength, and calculate the number of
photons captured by the entire array, within the 2-PPM
slot time. We also sum the background flux from both
the Milky Way galaxy, which is directly behind the
probe swarm in the field of view, taking into account
the Airy disk angle of the receiving telescope, and the
three stars in the Centauri system, which will also be in
the field of view. By calculating the number of photons
arriving during the 2-PPM time slot, we can estimate
the noise levels on top of the signal. Applying this
method, we end up with 724 signal photons captured
by the array per pulse, from an original count per pulse
of 2.6x1016 with an average background level of less
than one noise photon per time slot.
The second method is the traditional link-budget algo-
rithm used in both optical and RF communications.
𝑃𝑟=𝑃𝑡𝐺𝑡𝐺𝑟𝐿𝑠𝐿𝑎𝜂𝑝𝑡𝜂𝑡𝜂𝑟,(4)
where P𝑡is the transmitted power;
G𝑡is the transmitter gain;
G𝑟is the receiver gain;
L𝑠is the path loss;
L𝑎is the atmospheric loss;
𝜂𝑝𝑡 is the pointing efficiency;
𝜂𝑡is the transmitter efficiency and 𝜂𝑟is the receiver
efficiency.
The transmitter and receiver gain are calculated from
the diffraction limits of the telescope apertures for
the laser bean wavelength and are expressed relative
to an omnidirectional “antenna”. Marshall and Burk
(1986); Youinou and Lin (2022); Wang, Guo, Zhang
and Liu (2014) . For simplicity and because we do
not have reliable estimates, in this method, we as-
sume that the transceiver and the receivers both have
perfect pointing accuracy. Using this method we get
386 received photons per PPM time slot and again,
an average background level of less than one noise
photon per slot.
For each probe, the electrical power is provided by
a betavoltaic battery generating a total of ∼10 mW
at launch, decaying to 6 mW at the time of flyby, of
which 4 mW is partitioned to the main laser com sys-
tem. This is converted into 0.4 mW of optical power,
assuming today’s mere 10 percent electric-to-photonic
conversion efficiency. (We expect this efficiency to im-
prove dramatically over the next few decades.) Elec-
tricity can be stored in a rapid-discharge ultracapaci-
tor. By concentrating power into 1-ns pulses, the aver-
age power of each laser pulse is 4x104W, containing
40 micro-joules per pulse.
2.1.4. Onboard Clock Metrology and Science Return,
or The Role of Really Good Clocks in Swarming
Proxima
Better timekeeping has always been a disruptive tech-
nology throughout human history Kennedy III (2021). How-
ever, the rate of technological progress since World War II
has been exceptional, literally a decade of frequency stability
per decade on the calendar. Furthermore, the differential
cost trends are almost as attractive as Moore’s Law—while
∼$2000 buys an instrument that can measure voltage or
current to five places, the same money buys an astounding
T.M. Eubanks et al. Page 9 of 24
Swarming Proxima Centauri
Figure 10: Worksheet for Swarm-to-Earth Communication, Link Budget Method.
Figure 11: Long Term Trends in Clock Frequency Stability.
The red icons are for various types of Cesium clocks and
frequency standards used in official timekeeping, blue icons
are real or predicted optical atomic clocks, and the black icons
predicted Thorium nuclear clocks. The solid horizontal line is
the gravitational red shift (-1.095 ×10−19) caused by a 1 mm
altitude difference on the surface of the Earth. As it is very
difficult to model or measure absolute vertical motions on the
Earth to better than ∼1 mm, when clocks improve beyond
this limit it is likely that the best time standards will have to
migrate into space.
twelve places of precision for measuring time.
Therefore, our innovation is to use advances in optical
clocks, mode-locked optical lasers, and network protocols
to enable a swarm of widely separated small spacecraft or
small flotillas of such to behave as a single distributed entity.
Optical frequency and reliable picosecond timing, synchro-
nized between Earth and Proxima b, is what underpins
the capability for useful data return despite the seemingly
low source power, very large space loss and low signal-to-
noise ratio. While quantum metrology would limit error in
interferometry to 1/N (N is number of photons received) vs.
classical proportionality (1/(𝑁)), the state of the art is not
quite there yet.
Figure 12: Space-rated radiation-hardened chip-scale
atomic clock. Credit: Microsemi Product Catalog:
https://microsemi.com/product-directory/embedded-clocks-
frequency-references/5207-space-csac.
Nevertheless, state-of-the-art chip-scale atomic clocks
(CSAC), existence now and likely to be space-rated in the
future, can enable the simple synchronization of optical
pulses amongst numerous individual emitters – which we
dub “operational coherence” – if not the proper “phase
coherence” required for true “interferometry in reverse”.
With a very limited energy budget, synchronization squeezes
the same signal photons into a smaller transmission window,
temporarily increasing the brightness of the signal relative
to unavoidable background noise. With sufficiently narrow
reception bins on Earth, this means data rates much higher
than from a single probe with the same mass. Such a swarm-
based mesh network would also provide information about
the target from a multiplicity of look-angles and distances,
T.M. Eubanks et al. Page 10 of 24
Swarming Proxima Centauri
information that would not be available in any other way.
Another intriguing very-long-term possibility is an atomic-
scale clock built around a single-atom oscillator. Hannah
and Brown (2007) Although we assess that such devices
would not be ready soon enough for BTS, improving a
swarm’s synchronization to beyond picosecond levels would
provide true phase coherence, enabling an entire fleet to
act as a single transmitter with vast aperture, which in
turn would focus many orders of magnitude more photons
to receivers in the Sol system than individual gram-scale
spacecraft could possibly manage on their own, even if they
were individually phase-coherent. In the more distant future,
beyond the time frame of this BTS mission, a truly phase-
coherent swarm would both require and enable autonomous
position-navigation-timing (PNT) down to 100-m precision
for receiving optical timing pulses from the Earth.
2.1.5. Signal Processing at Earth
We assume a large array of 796 “light buckets” (Figure
13) on Earth, summing to 1 km2(the size of a city in aggre-
gate), and ∼1019 photons in each signal chirp. These pulses
are assumed to occur once a minute and to be operationally-
coherent or synchronized from the whole fleet, so that a few
hundred 539-nm signal photons can be expected to arrive
at the receiver on Earth during particular narrow reception
windows in time and wavelength. This transmission mode
requires good clocks and good PNT at the source, and also
time synchronization with the Earth.
Hence the importance of good clocks to the entire
scheme. Therefore the final piece of our approach is to apply
clever time- and frequency-bandpass filtering back on Earth
to maximize the chance of seeing these photons, which
should stand out against Proxima Centauri’s weak UV flux.
Even without the cleverness, the amount of computation that
can be dedicated to signal processing at home is essentially
unlimited. While electronic receiver design is not covered
in this paper, with the essentially unlimited post-processing
that we anticipate to be available by 2050, a photon signal-
to-noise ratio (SNR) of 1:1, or 0 dB, is sufficient to receive
a signal with a reasonable bit error rate. For the actual
embodiment in a real Probe, we assume that additional
error-detection-correction schemes, such as Reed-Solomon
or Turbo codes, will be used. We are indeed fortunate.
2.2. Science Goals and Means During Boost,
Cruise, Flyby and Post-Encounter
2.2.1. Boost Phase
The operational, engineering, and scientific intelligence
gathering can commence immediately upon launch, since
the probes will have been boosted to the highest velocities
of the entire mission (they only slow down after launch, at
greater or lesser rates), and the density of collision hazards
(mostly dust in Sol system) is by far the highest it will ever
be.
Figure 13: A conceptual receiver implemented as a large
inflatable sphere, similar to widely used inflatable antenna
domes; the upper half is transparent, the lower half is silvered
to form a half-sphere mirror. At the top is a secondary mirror
which sends the light down into a cone-shaped accumulator
which gathers it into the receiver in the base. The optical
signals would be received and converted to electrical signals
– most probably with APDs at each station and combined
electrically at a central processing facility. Each bucket has a
10-nm wide band-pass filter, centered on the Doppler-shifted
received laser frequency. This could be made narrower, but
since the probes will be maneuvering and slowing in order
to meet up and form the swarm, and there will be some
deceleration on the whole swarm due to drag induced by the
ISM, there will be some uncertainty in the exact wavelength
of the received signal.
2.2.2. En Route
We can glean much technical information en route valu-
able to many fields of science both practical and theoretical.
Currently, our direct in situ observations of the “nearby”
ISM are limited to Voyager spacecraft at 100 AU, which have
not fully departed our heliopause. Indirect estimates rely on
absorption lines in direction of nearby stars. This means we
only have data along lines of sight. If there is no star in
some direction, there is no measurement in that direction!
The nearest cloud is the Local Interstellar Cloud (LIC), but
it is not even clear if our solar system is in that cloud or the
G Cloud, which contains the Alpha Centauri system Linsky,
Redfield, Ryder and Moebius (2022). We might be on the
boundary of G cloud, or a few thousand AU inside it, or away
from it (see Figures 14 &15).
1. Pathfinding in the Sol-to-Centauri ISM. Recordingthe
attrition rate of these pioneering pathfinding probes
due to collisions with ISM bigger than single atoms or
other hazards will both inform this mission’s operation
and also provide the first high-fidelity map of the
entire tube of ISM all the way between Sol and
Proxima, thus providing much basic astronomical
and stellar cartographic intelligence to the scientific
community and the world at large long before arrival.
T.M. Eubanks et al. Page 11 of 24
Swarming Proxima Centauri
Figure 14: Local clouds in the Interstellar Medium (ISM) from
the North Galactic Pole, based on inversion of Doppler shifted
absorption lines Linsky et al. (2019).
Figure 15: Another view of local clouds in the ISM, in the
Galactic X-Y plane as seen from the North Galactic Pole Linsky
et al. (2022). Our Solar System might be on the boundary of
“G”, or a few thousand AU inside or outside that boundary (a
distance the size of a period here on those plots). Resolving
this ambiguity is an important goal of initial deep precursor
missions, but is also crucial for planning of the BTS mission
itself, as drag from these clouds will be an important tool for
swarm coherence.
2. Chronometric geodesy. Accurate time synchroniza-
tion between state-of-the-art clocks on Earth (whose
size is unconstrained by spaceflight) tiny clocks aboard
the fleet will provide important clues about gravity
waves of such long wavelength they cannot be ob-
served on or near Earth. Scaling the known perfor-
mance of state-of-the-art clocks in existence on Earth
now would support determination of a spacecraft’s
position at Proxima to an accuracy of 100 meters, an
astounding thought. Perhaps a 10−21 frequency sta-
bility based on transitions inside the atomic nucleus,
which are even faster than optical transitions in the
electron cloud, will be attained by the time of the
fleet’s launch, which is the best attainable on Earth.
For better than 10−21, even the best clocks will have
to move into space.
2.2.3. Encounter
1. Approach Geomtery. We want to approach the planet
on its day side, obviously, in order to reveal macro-
scopic surface features such as bodies of water, ice,
continents. This approach also provides the best light-
ing geometry for finding moons, if any. Therefore
swarm must pass the star first (but not too close!),
which sets a ti Refer to Figure 16.
2. Gibbous Aspect. We want a gibbous aspect so we can
see the terminator, which is where life would be if it
exists anywhere on this world that probably is tidally
locked (we should know for sure by the time this
mission flies). The way to obtain a gibbous aspect
is to arrange the launch time 20 years beforehand
so as to arrive when Proxima b is at a moderate
phase angle from Proxima Centauri, which provides
a reasonably large fraction of sunlit hemisphere and
extent of terminator, as shown, while not getting too
close to the star. (See Figure 16). We have noted in
this work that the inclination of Proxima Centauri’s
invariable plane with respect to that of Sol’s, as well
as the other orbital parameters specific to Proxima
b such as the inclination of its ecliptic, its epoch,
eccentricity, periastron, are not known today to the
necessary accuracy if they are even known at all.
Resolving this basic lack of understanding about the
Proximan system is an operational challenge that must
be overcome decades before the fleet launches.
3. Impact Spectroscopy. Although the distribution pat-
tern of the probes in the fleet would be approxi-
mately hex-close-packing for maximal communica-
tion efficiency, over the course of 20 years of flight,
the mesh could gradually draw together and densify
itself around a point expected to intersect the planet.
The distance of one probe to another at the fleet’s
center could be as little as a few thousand kilometers,
in which case, one or more 1-gram probes may be
expected to impact the upper atmosphere, if one exists,
or the surface, it it does not. However, lack of an atmo-
sphere does not prove lack of life—Proxima b may be
an ice world like Europa. Each impact which would re-
lease ∼90 terajoules or ∼22 kilotonnes. (Being equiv-
alent to the yield of an old-fashioned “atomic” bomb,
that would be one helluva flashbulb.) Since a neutral
hydrogen in the ISM releases 20 MeV when it hits
the probe, the heavier elements in the probe can be
expected to release ∼1 GeV per atom, which is far
T.M. Eubanks et al. Page 12 of 24
Swarming Proxima Centauri
Figure 16: Geometry of swarm’s encounter with Proxima b. The Beta-plane is the plane orthogonal to the velocity vector of the
probe ”at infinity” as it approaches the planet; in this example the star is above (before) the Beta-plane. To ensure that the
elements of the swarm pass near the target, the probe-swarm is a disk oriented perpendicular to the velocity vector and extended
enough to cover the expected transverse uncertainty in the probe-Proxima b ephemeris.
more than merely ionizing, this much energy will
disrupt a nucleus. We note that Earth experiences such
blasts in the upper atmosphere at least annually with
no ill effects. The flash would be visible to nearly the
entire fleet, and yield important spectroscopic data
about the composition of Proxima’s atmosphere. It
would also serve as an unambiguous timestamp. If
Proxima b is determined by some other method to host
life then this scenario must be avoided.
4. Transmission Spectroscopy. Upon arrival we should
able to do transmission spectroscopy between pairs
of spacecraft at Proxima b, because some members
of the swarm will pass behind Proxima b as seen
from the Earth, due to inherent transverse motions of
Earth and Sol system with respect to Proxima b and
Proxima Centauri. Immediately after flyby of Proxima
b by most of the swarm, we want to look back from
the night side at the sunset of Proxima Centauri on
Proxima b’s limb through the thin ring of atmosphere,
if it exists. (At the same time, we might be able to
spot cities if they exist and are artificially lighted.)
Furthermore, with sufficiently good timekeeping and
absolute position measurement the drive laser could
be utilized at that time to backlight Proxima b in
monochromatic light which would be very useful for
spectroscopy / spectrometry. This would be a big deal
scientifically, on the order of Webb’s main missions
(a “Flagship” mission), or roughly equivalent to what
the Starshade is. (Being able to see each other fore
and aft is also why the swarm must be lens-shaped,
not be a completely flat disk, in order to have some
longitudinal dimension along the line of flight.)
2.2.4. Post-Encounter
. We assume that the probes require about 1 year to send
back the data collected during the flyby of the 𝛼Centauri
system. Assuming a “light bucket” on or near the Ear th that is
in aggregate the size of a city 1 km2, and order 100 joules per
pulse once a minute from the fleet (containing ∼1019 432-
nm photons), with good position-navigation-timing (PNT)
at both ends, then with clever time and spectral frequency
bandpass filtering, and sufficiently narrow reception win-
dows, a few hundred photons can be expected to arrive to
a 1 km2receiver on Earth with the pulse system described
in Figures 10 and 9. These photons, red-shifted to 539-nm
on arrival, should stand out against Proxima Centauri’s very
weak near-UV flux (see Figure 17). Again, even absent the
cleverness, the amount of computation that can be dedicated
to signal processing back on Earth is essentially unlimited,
which is fortunate.
T.M. Eubanks et al. Page 13 of 24
Swarming Proxima Centauri
Figure 17: Photon fluxes, in terms of photons m−2 ×s−1
nm−1 to and from Proxima Centauri 4.24 light years from each
star, with superimposed wavelengths of laser coms: probe-to-
probe, and swarm-to-Earth, as sent and received. Flux from
Sol at 1 𝜇m is ∼500 ×the quiescent Proxima flux. Note that
Proxima background emission drops off steeply below 1000
nm wavelength and would be very small at 432-nm source
transmission or even the red-shifted 539-nm received signal.
For realistic smallsat apertures and spectral resolution, within
a 1-ns-duration time gate, the laser photon flux from the target
would be very much greater that the flux from the star itself
(excepting stellar flares increasing brightness by ∼68, but are
almost entirely emitted on specific spectral lines).
2.3. Predicting Proxima b, or Necessary
Astrometry
. Since the goal is to fly this swarm past Proxima b, we
need to predict the position of Proxima b to order 10,000 km
at least 8.6 years before flyby (the time it takes for informa-
tion between the swarm and Earth to complete one round-
trip at the speed of light). Since four years is roughly 108
seconds, predicting the position of Proxima b to 108m (our
swarm diameter) means determining the planet’s velocity to
order 1 m/s, and its angular position to ∼0.1 microradians.
A good optical read from a terrestrial telescope yields a
position accuracy of order 4 x 109meters so obtaining a
velocity error of order 1 m/s implies a time baseline of order
100 years, which is too long for BTS. So, either one must go
to space with a really big telescope, or use other techniques.
Although we already have Proxima b’s period (11.68 days),
we need to determine its line of nodes, eccentricity, inclina-
tion and epoch, and also its perturbations by the other planets
in the system. At the time of flyby, the most recent Earth
update will be at least 8.5 years old. The Proxima b orbit
state will need to be propagated over at least that interval to
predict its position, and that prediction needs to be accuracy
to the order of the swarm diameter. That implies an orbital
velocity error ≲0.3 m/s at the time of the last update (or less
than one part in 10−5, and knowledge of the semi-major axis
at that time to ∼8.5 km or better.
In addition, the star’s ephemeris requires a position accu-
racy (both transverse and radial) of 10,000 km, requiring
knowledge of both the parallax and angular position (as seen
from Earth) of ∼40 microarcseconds at arrival, requiring
determination of the star’s proper motion and parallax to
one part in 105at launch. We think that this accuracy
could be provided with a small spacecraft in Earth orbit
tracking Proxima microlensing events (by matching the lens-
source relative velocity and thus significantly lengthening
the duration of the lensing events) and thus obtaining fine
details of the motion of Proxima b. But that’s a story for
another paper.
2.4. Implications of Our Approach for BTS
Our approach inevitably has profound implications for
some aspects of the overall BTS architecture, particularly
related to launch operations, however we have come to
believe that the objective cannot be achieved in any other
way if the principal constraints of “gram-scale spacecraft”
and “data return within a human lifetime” are retained.
Our operational modeling shows that even firing just
one probe per day from Earth for a year, it is possible to
assemble a functional swarm of 100s–1000s of probes at
Proxima b, noting that the Starshot baseline is 9 minutes
per shot. Larger swarms (104to 105) would provide greater
capability due to more eyes, more power and brightness,
and more computation at encounter, as well as better odds
of survival after 20 years transit through the ISM, would
be possible with a faster launch cadence. In addition, a
larger swarm would both support a significantly higher bit
rate (data return), and also provide the ability for some
probes to observe their brethren near and even behind the
planetAccoracy of time perspective impossible to gain with
just one probe, or from Earth).
3. Probe Concepts
3.1. Material
We propose to significantly increase the thickness of the
sail by utilizing extremely low-density materials to increase
the collision cross section of the sail vis-à-vis the oncoming
hydrogen flux. Candidate materials could be aerographene
(Shah, Patel, Patel, Bist and Sircar,2022) (density: 0.16 kg
m−3) and aerographite (Mecklenburg, Schuchardt, Mishra,
Kaps, Adelung, Lotnyk, Kienle and Schulte,2012) (density:
0.18 kg m−3) (Behera,2021). To illustrate the superb
performance of this material, a 1-mm thick aerographene
structure would have an area-to-mass ratio of order of 10−4
kg m−2. The ratio for a 1-micrometer thick structure would
be 10−7 kg m−2. Due to this exceptionally low sectional low
density, the performance of aerographene for the purpose
of deceleration is about 104better than for Mylar. Both
materials are completely opaque with an absorptivity of 1.
Also, both materials have been synthesized (Mecklenburg
et al.,2012) in the laboratory. No principal roadblocks seem
to exist towards mass production. Both materials share many
T.M. Eubanks et al. Page 14 of 24
Swarming Proxima Centauri
Figure 18: Microphotograph of graphite tetrapod (an are-
ographene material.
similarities but for simplicity, we will focus on aerographite
in this work.
Current structures based on this material are of order 10s–
100s of 𝜇m (Meija, Signetti, Schuchardt, Meurisch, Smazna,
Mecklenburg, Schulte, Erts, Lupan, Fiedler et al.,2017).
Further reducing the thickness to order 10s–100s of nanome-
ters should, in principle, be feasible, as the wall thickness of
the tetrapods in Figure 18 is on the order of 10 nm. The
use of smaller tetrapods as sacrificial material on which the
graphite is deposited should be possible. (Meija et al.,2017).
The tetrapod size and shape were based on the t-ZnO and t-
AG sacrificial material used for depositing the graphite on
those shapes. There do not seem to be principal obstacles
to shrinking the tetrapod size to sub-µm to generate a µm-
scale porous structure. While the synthesis seems feasible,
the main question is whether the mechanical properties of
such a thin, porous structure can satisfy the requirements for
an interstellar mission.
3.2. Layout and Features
The 10,000-gravity launch condition is so extreme that
concentrations of mass (“mascons”) must be absolutely
avoided. Therefore the thickness of the probe must be
as uniform as possible. The layout that results from this
constraint is reminiscent of a red corpuscle rather than the
simple flat disk which seems to be the default in the com-
munity. The central disk consists of a large-diameter (4m)
phase-coherent array of fresnel-like flat optics consisting
of 247 smaller 25-cm annular apertures arranged in hex-
close-packing (hcp, like a honeycomb). Each smaller 25-
cm element consists of an annular aperture above an optical
well containing a central sensor. Because they are part of a
common monolithic structure, each of those 247 elements
knows exactly where it is in relation to the others, to within
a few nanometers, which is what makes phase coherence
by the big optical element possible. The diffraction limit
for the beam is based on the diameter of the whole array.
The phased array is for both imaging the target and emitting
coherent photons for communicating back to Earth. Because
it contains 247 separate elements, the array will possess a
great deal of redundancy. Therefore the array will tolerate
a high rate of attrition due the microscopic holes that will
inevitably be formed by impacts with neutral hydrogen or
helium nuclei at those times when the probe must cruise
face-on to the ISM. Although photonic power would be lost
with each failure of an annular element, down to a certain
level, information would not be lost, like missing dots in an
LED traffic light.
Assuming that a third of the 1-gram payload mass is
partitioned to this one item, the entire probe must still be
extremely thin, no greater than 28 microns if it were a solid
plate (which it is not) but several hundred microns if a
sandwich of aerographene between meta-material films.
Another third of the probe’s mass is partitioned to the nuclear
battery.
Everything else comprises the remaining third.
Other than the main optical disk, the functional heart of
a probe (batteries, ultracapacitors, 2-cm-aperture lasers for
inter-probe communication, computation and other house-
keeping electronics) is concentrated the 160-degree long
trailing edge of a 2-cm high thickened raised rim perpen-
dicular to the disk.
Since the probes will fly edge-on without rotation on the roll
axis in order to minimize the radiation dose, the leading edge
of the probe will consist of a 2-cm thick layer of aerogel
faced with a thin sheet diamond to serve as an sacrificial
barrier to absorb impacts from the ISM and dissipate the
energy via evaporation. No electronics are located any closer
than 2 cm deep to the outer surface of this barrier, nor
do any instrument ports penetrate the surface. The barrier
extends around 200 degrees of the probe’s circumference.
No functionality is contained within this buffer except the
single layer of electronics buried 2-cm deep under the face
of the leading edge which helps assure overall connectivity.
3.3. Construction and Manufacturing
The 1-gram constraint is so extreme that it precludes tra-
ditional frameworks built up of subassemblies or conventionally-
packaged chips with pins on circuit boards or radiation
shielding. Even existing surface-mount technology such as
in Raspberry-Pi devices would not be good enough to avoid
mascons. Therefore, we assess that probes must be fabricated
with a combination of methods:
1. additive manufacturing such as stereolithography at
macro scale and the multi-pass processes used at
microscale now to manufacture flash memory (which
can contain >200 distinct layers);
2. direct ion implantation;
T.M. Eubanks et al. Page 15 of 24
Swarming Proxima Centauri
3. wafer-scale integration - state-of-the-art today is a
single chip 30 cm across containing trillions of tran-
sistors. The “30-cm” limit is defined by the maximum
size of silicon “log” that the semiconductor industry
is willing to produce today, which is an order of
magnitude smaller than the probe.
However, these methods would have to occur at
1. near atomic-scale, perhaps with atomic force micro-
scopes (AFM), and
2. at near-atomic precision, and
3. at very high speeds or massively parallel operations
due to the astronomical number of atoms contained
even with a 1-gram spacecraft.
4. Furthermore, the overall extent of the workpiece (up
to 4.1 meters diameter per baseline parameters) is 9
orders of magnitude larger than ∼1-nanometer size of
individual features that would be directly laid down.
5. An additional complication would be working with
radioactive materials which must be directly placed
without packaging or shielding due to the mass con-
straint. The facility that can accomplish all this would
only vaguely resemble a modern chip “fab”.
3.4. Power
In the course of this work, it became apparent that the
communications link budget must be grounded in a rational
power budget, which is why this team invested the time
to work out a defensible power system. Given the extreme
design constraint on the spacecraft, allowing just one gram
for everything except the dielectric layer during launch,
every speck of mass must “earn its keep” every day over
the duration of the entire voyage, not just during the brief
encounter. We have concluded that as high as mass fraction
as possible for stored energy must be brought along. For
comparison, the Voyager probes of two generations ago
has 15 percent of their total mass partitioned to the 238Pu
power system; we believe a 30 percent set-aside for nuclear
energy storage / power generation would be prudent for the
BTS probe. However—a subtle but key point—it would be
pointless to bring any stored mass-energy that would last
any longer than the transmission phase.
3.4.1. Photovoltaic Options
Even though triple-junction silicon cells have achieved
photoelectric conversion efficiencies ≥40 percent in the lab-
oratory, based on the criterion above, we immediately elimi-
nated photovoltaic from consideration as its effective duty
cycle would be less than 0.0001 percent (10s of minutes
out of a 10-million-minute-long mission), which conflicts
with the principle “earn your keep every day”. Furthermore,
the luminosity of Proxima Centauri, which peaks in the
infrared 17, is but ∼0.1% of Sol’s, i.e. ∼1 W m−2 at 1
AU, or ∼200W for less than a minute during flyby. At the
time of this writing, it is not prudent to count on generating
photovoltaic power with as-yet unknown infrared-optimized
semiconductors of unknown efficacy.
3.4.2. Other Electromagnetic Options
If they worked at all, “e-sails” and “mag-sails” would
have a 100-percent duty cycle during the entire voyage ,
but were nevertheless eliminated from consideration as their
technological readiness is far less than the isotopic power
option described below.
3.4.3. Radio-Isotope Thermoelectric Generators
The loss of mass in the fission of one uranium-235
nucleus yields ∼200 million electron-volts (MeV), i.e., ∼1
MeV per nucleon, which is 7 orders of magnitude greater
than the energy of chemical transitions taking place in an
electron shell. This intensity is due to the transition being
mediated by the strong force. Lesser transitions such as
radioactive decay (alpha, beta, gamma emissions) governed
by the electroweak force at 10s to 100s of kiloelectron-
volts (keV) per nucleon are 2-3 orders of magnitude smaller,
yet due to their nuclear origin are still thousands of times
greater than achievable with any known chemical method or
in-flight electromagnetic or photonic method. Furthermore,
time constants typical of middle-range radioactive decay
is well matched to the necessary duration (decades) of
interstellar flight even at relativistic speeds.
238Pu Highly miniaturized 150-mg plutonium-238 radioiso-
tope thermoelectric generators (RTG) were discussed in
early days of BTS. However, RTGs suffer from two fun-
damental physical constraints relevant to this BTS mission.
Thermal process requires a heat source and a cold sink. As
an object gets smaller, its surface-to-volume ratio increases
in inverse proportion to the length reduction, according to
the Square-Cube Principle, which means proportionately
more radiating area per unit mass to dissipate heat. This
is great for the cold sink. The more profound challenge
is the fundamental limits to miniaturizing anything whose
function is based maintaining continuously high temperature
in the heat source. According to the Stefan-Boltzmann Law,
radiation goes as T to the fourth power. Very small hot
objects cool off very quickly, hence why metal sparks and
meteors behave the way we observe them to, and whereas
much of the Earth is still liquid 4 billion years after its
accretion. This is a strong curve to be on the wrong side
of.
The second flaw is the relatively poor efficiency of thermionic
conversion, in the range of a few percent. Furthermore,
there is a logistical challenge, in that the demand for 238Pu
for space missions already far exceeds the supply. Even
though 238Pu is not fissionable hence no nuclear weapons
potential, which should exclude it from the class of “special
nuclear material” (SNM), it is a strong alpha emitter, which
T.M. Eubanks et al. Page 16 of 24
Swarming Proxima Centauri
Figure 19: NanoTritium𝑇𝑀 betavoltaic nuclear battery fueled
by 100 curies of tritium, which generates 100 microwatts (0.1
milliwatts). Picture courtesy of City Labs, Inc., 12491 SW
134th Court, Suite 23, Miami, FL 33186.
triggers extreme regulatory precautions all the same. Even
though a fleet of 1000 probes would require but a few grams
of 238Pu, distribution in any amount is tightly controlled.
Finally, according to the committee at the National Academy
of Sciences (NAS), which studied the matter of supply, no
one has even proposed studying the miniaturization of 238Pu
RTGs. Given the generally glacial pace of progress that is
peculiar to the nuclear field, laboratory-scale experiments
that commenced today to miniaturize RTGs would not yield
results for many decades.
90Sr RTGs fueled with strontium-90 (in the form of a
sintered ceramic, strontium titanate, SrTiO3) were widely
deployed in the Soviet Union for remote off-grid technical
applications. Strontium-90 is far more available (roughly 1
kilotonne in the world today), far cheaper (pennies per curie
versus hundreds of dollars per curie), and far less hazardous
to work with than plutonium-238, yet still has enough
radioactivity to get quite hot depending on concentration and
freshness.
3.4.4. Nuclear batteries
1. Past and Existing Betavoltaic Cells.
147𝑃 𝑚 In the 1970s, the “Betacel”, a prototype be-
tavoltaic battery based on promethium-147 was suc-
cessfully designed and built to provide years of power
for surgically implanted pacemakers. However, promethium-
146, a sister isotope that emits a hard gamma ray, was
inextricably mixed with the desired isotope. The two
were impossible to mass-separate given their mere
1-amu mass difference, therefore most of the weight
and volume of the device was for necessary shielding.
Furthermore, the idea of a radioactive implant faced
some resistance in the medical market. The invention
of the lithium battery at just about the same time
doomed the Betacel.
Tritium decays to helium-3 in a single step by emitting
a single 𝛽−ray (which is an electron, as opposed to
an 𝛼particle which is a much more massive ionized
helium nucleus). The sole commercial-off-the-shelf
(COTS) betavoltaic power source available today is
the NanoTritium𝑇 𝑀 battery, fueled with 100 curies
(10 milligram) of tritium, generating an initial 100
microwatts (0.1 milliwatt) per unit at the moment of
manufacture.
This COTS technology would be immediately appli-
cable to less demanding precursor missions within
the Solar system. Moreover, it appears that “Beta-
voltaics are like Jupiter!” meaning that commercially-
available products appear to maintain a constant form
factor regardless of a three-order-of-magnitude range
in power output. This in turn suggests that an nearly
arbitrary number of curies (from 225 millicuries in
early tritium cells to 100 curies now) could be stuffed
into a package of constant physical size depending on
mission/market.
To generate ∼10 milliwatts per interstellar probe
for decades as described below would require 100
times as much tritium as the battery described below,
i.e., 10,000 curies, or 1 gram of tritium. While no
vendors responded directly with price quotations or
budget numbers for tritium, in a recent work, it was
estimated by one of us that the inflation-adjusted fully-
burdened cost of tritium as produced by the (now-
decommissioned) Cold-War-era reactor complex at
Savannah River would amount to USD ∼200,000-
300,000 per gram in today’s money.Kennedy III (2018)
Today, tritium for U.S. nuclear weapons complex is
generated by lithium breeder rods at the civilian Ten-
nessee Valley Authority’s Watts Bar Nuclear Power
Plant. Tritium for all other purposes, including indus-
trial, is sourced from a fleet of ∼20 CANDU reactors,
mostly in Canada (hence the name) by periodically
“de-tritiating” their heavy water (D2O) which is used
for both a moderator and coolant. The worldwide
inventory of tritium is ≤20 kg (which is ∼5 orders
of magnitude less than the worldwide inventory of
90Sr below). The cost of tritium in 2003 was USD
∼30,000 per gram; worldwide sales of tritium at that
time amount to 100 g.
However, there is no room in the extremely-limited
interstellar mass budget for radiation shielding or
the packaging and pins of traditional chips. Rather,
the 𝛽−emitting layer and the receiving / electricity
conversion layers would have to be laid down by direct
implantation, and the battery flown without shielding.
Since tritium, an isotope of hydrogen, is normally
a gas, the 𝛽−-emitting layer of tritium used in the
NanoPower𝑇𝑀 above would have to be converted into
some solid form to be suitable for a long-duration
interstellar mission. While the proprietary details of
the Nanopower𝑇𝑀 cell pictured below are not publicly
available, it should be possible to prepare and work
with a mass-efficient solid compound of lithium tritide
(LiT), and plate that material directly to a semicon-
ductor such as crystalline silicon. Tritium’s short half-
life, 12.3 years, means that power from tritium-fueled
cell would decline by about two-thirds on the way to
Proxima.
T.M. Eubanks et al. Page 17 of 24
Swarming Proxima Centauri
Despite its extremely low isotopic mass, tritium’s
weak 𝛽−ray (0.0186 MeV), makes for a poor power-
to-mass ratio, thus providing a milliwatt at best, an
order of magnitude less power than the 90Sr candidate
analyzed below.
Despite to its many useful industrial applications,
tritium has a critical role in thermonuclear weapons,
which means its possession and distribution are tightly
controlled, albeit not to the severity of fissionable
materials (235U, 233 U and 239Pu). Fortunately, there
are viable alternatives subject to less regulation.
2. Feasible / Near-term Betavoltaics. A method exists
now (proven in the laboratory with commercial and
medical devices) for directly and continuously gen-
erating electricity at current/voltage/power (∼1-10
milliwatts per probe) suitable for deep space missions,
by sandwiching commercially-available crystalline
silicon (c-Si) ∼100-micron-thick photovoltaic layers
around a beta emitting layer, with a beta-ray-to-
electric conversion efficiency (∼20 percent) almost an
order of magnitude better than the thermionic conver-
sion (∼1-5 percent) in RTGs. This would sufficient
to support communication at interstellar range. Many
other photovoltaic materials exist, but have not been
put to the test for this application in a laboratory.
(Those experiments ought to commence.) More exotic
much tougher semiconductor photovoltaic materials
such as synthetic diamond are worth investigating in
this regard.
Several other isotopes were assessed as source ma-
terial for betavoltaic power consistent with the ex-
treme mass constraints and mission duration. We
specifically note that these candidates have not been
dismissed from consideration for application in in-
terstellar “chipsats”. They could be suitable for other
missions of different duration to targets either nearer
or farther away than Proxima b. Refining conceptual
designs, for example, exactly “length scale matching”
different types of specially-doped semiconductor to
particular thicknesses of each of the three “runners
up” isotopic fuel candidates (3H, 32Si, 63 Ni) is work
that we propose should be taken up in Phase II.
In order of isotopic mass, the betavoltaic candidates
beyond tritium that we analyzed and compared to
tritium are: 32Si, 63 Ni, 90 Sr.
32Si decays via phosphorus-32 to sulfur-32.
32
14𝑆 𝑖 →32
15 𝑃+𝛽−+𝜈 + (0.221 MeV) 𝐸𝛽,𝑃 (5)
32
15𝑃→32
16 𝑆+𝛽−+𝜈 + (1.709 MeV) 𝐸𝛽,𝑆 (6)
Of all the isotopic fuels we studied, ∼2500 distinct
radionuclides of the Chart of the Nuclides by Knolls
Atomic Power Laboratory, one in particular stands
far above the rest: strontium-90. Unlike tritium or
plutonium, strontium-90 is considered a nuisance,
a waste product. There is no shortage of supply,
as strontium-90 is one of the most frequent daugh-
ter products of uranium-235 fission. About 6 per-
cent of fission events produce strontium-90, there-
fore there are approximately 1,000 tons of the stuff
sitting in the spent fuel pools of the world’s nuclear
fleet. This conjecture was proven by an unofficial
quote from the National Isotope Development center
(NIDC) at www.isotopes.gov, in which the unit price
of strontium-90 came to a negligible USD 0.15 (15 US
cents) per full curie (plus a fixed handling fee of USD
57,000).NIDC (2022). This is ten orders of magnitude
less than the unit price for silicon-32, and five orders
less than the unit price for nickel-63. Strontium-90
has a 28.8-year half-life, which makes it an excellent
match for a 20-year mission to our closest neighbor.
Like 32Si above, it is at the top of a two-decay
chain; in this case, via yttrium-90 to zirconium-90.
It has a powerful combined 𝛽−ray (0.546 + 2.28
MeV), so, despite its moderate isotopic mass, it has
by far the best power-to-mass ratio. Sandwiching
semiconductor layers of sufficient thickness (“length
scale matching”) around the beta-emitting material
converts the primary beta-rays directly into charge, as
they interact with p- and n-layers. Strontium-90 could
be prepared as a thin pure metal foil, or as strontium
silicate that could be painted onto and should readily
adhere to the silicon semiconductor layer. This has
been demonstrated in the lab using a medical electron
gun and commercial-grade crystalline silicon.Dixon
et al. (2016).
A betavoltaic cell fueled on this isotope has a spe-
cific mass-to-power density of ∼32 milligrams per
milliwatt, including the ultracapacitor storage (whose
mass is but a small fraction of the energy system).
If 30 percent of the total 1-gram mass budget were
allocated to energy storage at nuclear densities, then
this source could provide about 10 milliwatts of con-
tinuous power on the day it is made, and 6 milliwatts
of power when flying by Proxima Centauri. We note
that triple-junction silicon cells have achieved pho-
toelectric conversion efficiencies ≥40 percent in the
laboratory, which would more than double the power
output. Betavoltaic cells fueled by strontium-90 could
be available within a decade with a modest technolog-
ical development effort, of order ∼107dollars.
Seven other 𝛽−emitters and one positron (𝛽+) emitter
were found amongst the ∼2400 nuclides on the “Chart
of the Nuclides” published by Knolls Atomic Power
Laboratory (Parrington, Knox, Breneman, Baum and
Feiner (1996)). We also note that while double-𝛽−
decay does exist, it is such a rare mode of emission
T.M. Eubanks et al. Page 18 of 24
Swarming Proxima Centauri
Figure 20: Decay chainsr four potential isotopic fuels for interstellar communications, from left to right: 3H, 32Si, 63Ni, 90 Sr.
Figure 21: Figure from Dixon et al. (2016). Length scale matching of a 90Sr+crystalline silicon betavoltaic device. (a) Normalized
energy spectrum of 𝛽rays emitted by two-step serial decay of 90 Sr then 90 Y. (b) Theoretical efficiency of energy collection in
semi-infinite materials. (c) Normalized energy spectrum of secondary electrons generated by 𝛽-spectrum of 90Sr+90 Y in comparison
to normalized 𝛽-spectrum itself. (d) 𝛽-particle energy deposition in Si and Sr. The “kink” near 1.5 mm is due to the higher energy
𝛽-particles (2.2 MeV) from the 90 Y decay. 50 percent of the beta-particle energy is intercepted by 300 microns of material,
comparable to thickness of commercial photovoltaic cells.
that it occurs on a timescale many orders of magnitude
longer than the age of the universe.
26Al: the sole positron emitter examined, has two
isomers, one with a half-life of 6.3 seconds, much
too short for the BTS mission, emits a high-energy
positron (𝛽+, 3.21 MeV); the other emits a high-
energy 𝛽+at 1.17 MeV, but has a half-life of 71,000
years, far too long for a good power-to-mass ratio.
39Ar: has a strong 𝛽−ray (0.565 MeV) and light
isotopic mass hence a good power-to-mass ratio, but
a light noble gas is difficult to stabilize in solid form
without excessive mass penalty; the half-life of 265
years is a much too long for the BTS mission.
85Kr: has two isomers, one with a half-life of 448
hours, far too short, which emits a high-energy 𝛽−
(0.840 MeV); the other one also emits a high-energy
𝛽−(0.687 MeV) and has a half-life of 10.76 years,
which is a good match for the mission. The isotopic
mass-to-power ratio is acceptable, but a medium noble
gas difficult to stabilize in solid form without exces-
sive mass penalty.
126Sn: i.e., tin-126, has a modest 𝛽−energy (0.25
MeV) and relatively high isotopic mass, thus a poor
power-to-mass ratio. Furthermore, it comes with a
moderately strong 𝛾ray (0.085 MeV) that makes it
difficult to work with, and its half-life of 250,000 years
puts it far out of consideration for the BTS mission.
T.M. Eubanks et al. Page 19 of 24
Swarming Proxima Centauri
Figure 22: Figure from Dixon et al. (2016). Betavoltaic device and setup of experiment at Georgia Tech in 2016. (a) Overall
dimensions of 1-W device proposed by Dixon et al.Dixon et al. (2016). (b) Cross section. (c) Experimental setup and apparatus
to characterize efficacy of betavoltaic cell. (d) Equivalent circuit of setup in (c).
Figure 23: Worksheet for Establishing Power Budget by Strontium-90 Betavoltaic Cell.
129I: has a modest 𝛽−energy (0.15 MeV), relatively
high isotopic mass, thus a poor power-to-mass ratio.
Furthermore, its half-life of 15.7 million years puts it
far out of consideration for the BTS mission.
137Cs: like strontium-90 is one of the most significant
daughter products of uranium-235 fission, and has a
similar half-life, 30.07 years, which makes it an excel-
lent match for the BTS mission. It has a moderately
strong 𝛽−energy (0.514 MeV), but a relatively high
isotopic mass, thus a poor power-to-mass ratio; and a
hard 𝛾ray (0.662 MeV) that would make it difficult to
work with.
166Ho i.e., holmium-166 has two isomers, one with a
half-life of 1.12 days hours, far too short, which emits
a high-energy 𝛽(1.855 MeV); the other one also emits
a weak 𝛽ray (0.065 MeV) and has a half-life of 1200
years when decaying to 166Er, which is far too long for
the mission.
210Pb, i.e., lead-210 has two low-energy 𝛽−emis-
sions: one 0.017 MeV, the other 0.061 MeV. It has
a 22.6-year half-life that makes it an excellent match
for the BTS mission, but a high isotopic mass which,
combined with the weak betas, gives it a poor power-
to-mass ratio.
T.M. Eubanks et al. Page 20 of 24
Swarming Proxima Centauri
3.4.5. Harvesting the ISM and Other Exotic
Long-Term Options
1. Alphavoltaic cells are less preferred than betavoltaic,
due to the orders-of-magnitude greater material dam-
age caused by 𝛼particles, and a lesser state of readi-
ness and market availability. Furthermore, the emis-
sion of alpha particles is one of two phenomena that
causes something to trigger extreme regulatory pre-
cautions. (The other phenomenon is being fissionable,
i.e.,“special nuclear material” (SNM).) But a probe
traveling at relativistic speeds through the ISM must
already be designed to withstand a very high radiation
dose over its lifetime due to impacts by 20-MeV
hydrogen and 80-MeV helium atoms. Furthermore,
the 90strontium betavoltaic cells have already been
physically simulated in the laboratory using a medical
6-MeV electron gun (“Clinac”) fired at commercial
solar cells made of crystalline silicon.Dixon et al.
(2016) This is only a factor of 3 less than the 20-MeV
hydrogen particles that is the principal constituent of
radiation flux induced by traveling through the ISM at
0.2c. This particular experimental apparatus suggests
that may even be possible to utilize the ISM itself as
the electrical power source by making the protective
leading edge out of a tough semiconductor, thus effec-
tively behaving as a particle-voltaic cell, harvesting
the initial investment of launch energy during the
entire flight.
2. Another very long term possibility for stably storing
energy at the ultimate limit of mass-energy could be
antimatter inside fullerenes, colloquially known as
“buckyballs” (the smallest of which is composed of
60 carbons in a spherical lattice, or “C60”). Suppose
10 milliwatts of power is needed for 20 years, which is
∼6×108seconds, hence ∼6 x 106joules or 7 x 10−11
kg of annihilation, which would take in principle 35
nanograms of antimatter, amounting to order 2 x 1016
antiprotons. One method of storing single antiprotons
(not atoms) one at a time in a fullerene trap has
been proposed.Rejcek, Browder, Fry, Koymen, Weiss,
Lockheed, Missiles, Control and Usa (2003). If this is
not impossible, then that would require 120 protons
and neutrons of 12C for every antiproton, or ∼12
micrograms of C60. This simple thought experiment
does suggest it might be possible to power the BTS
“Starchips” with very little actual mass.
4. Precursor Missions
Earlier work Parkin (2018b) produced a point design for
a 0.2 c mission carrying 1 g of payload. Recent work by
the same author widens the design space to missions having
0.1 mg to 100 kt payload and 0.0001-0.99 c cruise velocity
Parkin (2022).
The first six innovations in our approach could be
proved with planned precursor missions, at velocities up to
0.01c. The team has done and continues to do extensive
study of such missions Hein, Eubanks and Kennedy III
(2019); Hibberd, Hein and Eubanks (2020); Hein, Eubanks,
Lingam, Hibberd, Fries, Schneider, Kervella, Kennedy III,
Perakis and Dachwald (2022); Hibberd and Hein (2020);
Eubanks, Schneider, Hein, Hibberd and Kennedy III (2021).
While precursor missions could go to any body in the solar
system, we further propose a targeted flyby of the ISO
1I/’Oumuamua. Much of the scientific discussion of this
enigmatic object concerns its anomalous acceleration, which
is associated with its currently very large ephemeris error.
The simple detection of 1I at the one-pixel level in a Starshot
precursor chipsat would provide a substantial scientific
return and thus enable future missions. A second precursor
mission could be tightly targeted and could realistically
return images of 1I to Earth by 2035-2040. Other fast-
flyby precursor missions to celestial bodies such as trans-
Neptunian objects (TNOs) beyond the Edgeworth-Kuiper
Belt in the outer solar system could return valuable science
far more rapidly than conventional deep-space missions
presently taking near lifetimes.
The existence of Interstellar Objects (ISOs) visiting the Solar
System has been predicted for many years (e.g. Sekanina
(1976); Stern (1990)). We live in an interesting epoch
were two of them have been found (1I/’Oumuamua) and
(2I/Borisov) and signs of other visitors have been proposed
(Siraj and Loeb,2019;Froncisz, Brown and Weryk,2020).
In addition, there are prospects to detect one or several of
them per year with the Vera Rubin Observatory (or LSST)
starting in 2022 (Trilling, Robinson, Roegge, Chandler,
Smith, Loeffler, Trujillo, Navarro-Meza and Glaspie,2017).
Ground-based telescopes will not be able to give reliable
answers to questions about the origin, chemical composition,
mineral structure, size, shape of these interstellar visitors -
that will require in situ exploration.
5. Conclusions
1. Interstellar communications are achievable with gram-
scale spacecraft using the spacecraft swarm tech-
niques introduced here if an adequate energy source,
clocks and a suitable communications protocol exist.
Our proposed communications system will filter out
Proxima’s background photon flux using a combina-
tion of frequency and time bandpasses, and also by
using a short wavelength at the source (432 nm, is
red-shifted to 539 nm at Earth) where the Proxima
flux is very weak.
2. A swarm of 100s to 1000s of spacecraft of order
∼100,000 km diameter can be established en route to
Proxima b! This can be accomplished with a combina-
tion of a gross “time-on-target” (ToT) technique, con-
sisting of modulating the initial velocity of each probe
by the launch laser such that the tail catches up with
the head, relative to each other, and a finer “velocity-
on-target” (VoT) technique based on controlled drag
imparted by the interstellar medium (ISM) by altering
the attitude of individual probes with respect to the
ISM, thus keeping swarm together in relative and
T.M. Eubanks et al. Page 21 of 24
Swarming Proxima Centauri
absolute position once formed. This will take time, but
time is something we have a lot of.
3. Attitude adjustment is also necessary to minimize the
severe erosion (∼100 microns per day) by particle
impacts and extremely high radiation dose (gigarads)
induced by traveling through the ISM at 0.2c. To
minimize frontal area, hence dose and erosion, we
anticipate flying edge-on most of the way, without
rotation, hence distinct leading and trailing edges,
even though the probe is mostly symmetric.
4. An operationally-coherent (if not phase-coherent) i.e.
synchronized swarm supported by forecast-state-of-
the-art space-rated clock metrology onboard (10−13)
and supported by our forecast-by-2050 state-of-the-
art clock metrology at Earth (10−19 or better), and
existing time- and frequency-bandpass filtering, can
feasibly support the transmission of sufficient signal
photons for reception at Earth of ≳100K bytes of
science data. Space-rated, chip-scale atomic clocks
(CSAC) are already commercially available at very
low cost ($2K) today, although not yet in the miniscule
mass that would fit it a probe’s budget.
5. At least one candidate construction metamaterial,
aerographene, exists today, that has sufficiently low
density to satisfy the payload mass constraint, yet
sufficient mechanical strength and other qualities to
serve as the basis for the main body of a probe
body has a variable density, tailored for the particular
mechanical requirements.
6. A true optical phased array of elements (247) can fill
the full diameter (4 m) of an individual probe, which
would support both imaging of the target and return of
the science data to Earth at far higher signal-to-noise
ratio than currently thought.
7. Intra-swarm (probe-to-probe) communication can be
established by an independent orthogonal system
of lightweight efficient infrared (12,000 nm) laser-
beacons of modest aperture (20 mm).
8. At least one near-term candidate method of storing
primary energy onboard in sufficiently compact
form, i.e., nuclear energy density.
In addition, a method exists now for organically
and continuously generating electricity therefrom,
at current/voltage/power (∼3-10 milliwatts per probe)
sufficient for interstellar coms. Betavoltaic batteries
fueled by the isotope strontium-90 (90Sr), sandwiched
inside commercially-available crystalline silicon (c-
Si) photovoltaic material, could be available in a
decade (by ∼2032) with modest technology and man-
ufacturing development of order ∼107dollars. 90Sr
is a waste product of uranium fission. It has the
additional logistical advantages of not being regulated
as a “special nuclear material”, and its supply is
practically unlimited (∼1,000 tonnes), seven orders
of magnitude greater than is needed to achieve the
Breakthrough Starshot mission. Already, a betavoltaic
cell fueled with tritium (3H) is commercially available
today, albeit at an energy density or power output four
orders of magnitude less than would be necessary for
a 1-gram interstellar probe.
A method exists now of storing electricity as high
mass efficiency, well as discharging it in the necessary
time (nanoseconds), namely ultracapacitors.
The entire power system falls within the available
mass budget, ∼330 milligrams out of 1 gram payload
limit, or 30 percent partition, which is of the same
order as the 15 percent allocation in the Voyager
probes of two generations ago.
9. The probes we propose cannot be manufactured by
conventional methods based on subtraction and as-
sembly of subassemblies, but rather a monolithic pro-
cess involving a combination of additive manufactur-
ing and wafer-scale integration which however must
be scaled up by order of magnitude in extent, from 30
cm at present to 4 m.
10. The ∼100-GW drive laser can feasibly be repurposed
for a number of important suporting tasks after its
principal use for launching the fleet:
serving as a beacon for individual probes to point at
when communication back toward Earth
for direct communication from Earth to the fleet for
the entire journey in order to provide astronomical and
astrometric updates about the target, and synchroniza-
tion for position-navigation-timing (PNT), albeit at an
increasing temporal latency (4.3 years at Proxima b)
to continuously calibrate and optimize the main com-
munication channel between the fleet and the Earth
over the course of the entire cruise phase, and after
during the data return phase.
as an “interstellar flashlight” finding small bodies at
encounter, as well as provide an additional marginal
but possibly useful controlled source of illumination
for photography of dark objects.
as a adjunct piece of scientific equipment namely a
monochromatic light source for spectroscopy by the
fleet during flyby.
11. Sine qua non: In order for any of this to happen, we
have identified a fundamental operational unknown
that must be solved well in advance: accurately de-
termining the orbital position of Proxima b at least
8.6 Earth years (∼300 revolutions) ahead of launch.
We find that this can be overcome at reasonable cost
through the targeted use of gravitational microlens-
ing observed by small telescopes in low Earth orbit
(LEO).
6. Acknowledgments
The Breakthrough Starshot Foundation funded this work,
which was also supported and executed by the Institute
for Interstellar Studies (US), the Initiative for Interstellar
Studies (UK), and Space Initiatives Inc.
T.M. Eubanks et al. Page 22 of 24
Swarming Proxima Centauri
References
Angel, R., 2006. Feasibility of cooling the earth with a cloud of small
spacecraft near the inner lagrange point (l1). Proceedings of the National
Academy of Sciences 103, 17184–17189.
Behera, A., 2021. Advanced Materials: An Introduction to Modern Mate-
rials Science. Springer Nature.
Blase, W.P., 2022. Paul’s animation of probe rotations in x and y axes. URL:
https://www.youtube.com/watch?v=iD3S5Qj5sR4.
CubesatMarket, 2022. Specs for exa mt101 magnetorquer. URL: https:
//www.cubesat.market/mt01-compact- magnetorquer.
Dixon, J., Rajan, A., Bohlemann, S., Coso, D., Upadhyaya, A., Rohatgi,
A., Chu, S., Majumdar, A., Yee, S., 2016. Evaluation of a Silicon 90 Sr
Betavoltaic Power Source. Sci. Rep. 6, 6. doi:10.1038/srep38182.
Eubanks, T.M., Schneider, J., Hein, A.M., Hibberd, A., Kennedy III, R.,
2021. Exobodies in Our Back Yard: Science from Missions to Nearby
Interstellar Objects, in: Bulletin of the American Astronomical Society,
p. 292. doi:10.3847/25c2cfeb.2ddcf231,arXiv:2007.12480.
Froncisz, M., Brown, P., Weryk, R.J., 2020. Possible interstellar meteoroids
detected by the Canadian Meteor Orbit Radar. Planet. Space Sci. 190,
104980. doi:10.1016/j.pss.2020.104980,arXiv:2005.10896.
Hannah, E.C., Brown, M.A., 2007. Conceptual Design of a Micron-Scale
Atomic Clock. arXiv e-prints , arXiv:0707.4624arXiv:0707.4624.
Hein, A.M., Eubanks, T.M., Kennedy III, R.G., 2019. Near Term Inter-
stellar Missions: Finding and Reaching Interstellar Objects, in: 70th
International Astronautical Federation Congress.
Hein, A.M., Eubanks, T.M., Lingam, M., Hibberd, A., Fries, D., Schneider,
J., Kervella, P., Kennedy III, R., Perakis, N., Dachwald, B., 2022. Inter-
stellar Now! Missions to Explore Nearby Interstellar Objects. Advances
in Space Research 69, 402–414. doi:10.1016/j.asr.2021.06.052.
Hein, A.M., Long, K.F., Fries, D., Perakis, N., Genovese, A., Zeidler,
S., Langer, M., Osborne, R., Swinney, R., Davies, J., Cress, B., Cas-
son, M., Mann, A., Armstrong, R., 2017. The Andromeda Study:
A Femto-Spacecraft Mission to Alpha Centauri. ArXiv e-prints ,
arXiv:1708.03556.
Heller, R., Anglada-Escudé, G., Hippke, M., Kervella, P., 2020. Low-
cost precursor of an interstellar mission. Astron. Astrophys. 641, A45.
doi:10.1051/0004-6361/202038687,arXiv:2007.12814.
Hibberd, A., 2022. Adam’s space research. URL: https://www.youtube.com/
watch?v=jMgfVMNxNQs.
Hibberd, A., Hein, A.M., 2020. Project Lyra: Catching
1I/’Oumuamua – Using Laser Sailcraft in 2030. arXiv e-prints ,
arXiv:2006.03891arXiv:2006.03891.
Hibberd, A., Hein, A.M., Eubanks, T.M., 2020. Project Lyra: Catching
1I/’Oumuamua - Mission opportunities after 2024. Acta Astronautica
170, 136–144. doi:10.1016/j.actaastro.2020.01.018,arXiv:1902.04935.
Hoang, T., Lazarian, A., Burkhart, B., Loeb, A., 2017. The interaction of
relativistic spacecrafts with the interstellar medium. The Astrophysical
Journal 837, 5.
Kennedy III, R., 2021. Few Weapons Are as Deadly as a
Good Clock: Military Implications of 1:10-to-the-19th PNT.
Center for Global Security Research, Lawrence Livermore
National Laboratory. chapter 36. p. 501. URL: https:
//www.worldcat.org/title/1294398426?oclcNum=1294398426,
arXiv:https://cgsr.llnl.gov/content/assets/docs/StratLatUnONLINE.pdf.
Kennedy III, R.G., 2018. The Interstellar Fusion Fuel Resource Base of Our
Solar System. Journal of the British Inter planetary Society 71, 298–305.
Linsky, J., Redfield, S., Ryder, D., Moebius, E., 2022. Inhomogeneity in the
Local ISM and Its Relation to the Heliosphere. Space Science Reviews
218, 16. doi:10.1007/s11214- 022- 00884-5,arXiv:2203.13280.
Linsky, J.L., Redfield, S., Tilipman, D., 2019. The Interface between
the Outer Heliosphere and the Inner Local ISM: Morphology of the
Local Interstellar Cloud, Its Hydrogen Hole, Strömgren Shells, and
60Fe Accretion. Astrophys. J. 886, 41. doi:10.3847/1538-4357/ab498a,
arXiv:1910.01243.
Marshall, W.K., Burk, B.D., 1986. Received optical power calculations for
optical communications link performance analysis, in: The Telecommu-
nications and Data Acquisition Report, pp. 32–40.
Mecklenburg, M., Schuchardt, A., Mishra, Y.K., Kaps, S., Adelung, R., Lot-
nyk, A., Kienle, L., Schulte, K., 2012. Aerographite: ultra lightweight,
flexible nanowall, carbon microtube material with outstanding mechan-
ical performance. Advanced Materials 24, 3486–3490.
Meija, R., Signetti, S., Schuchardt, A., Meurisch, K., Smazna, D., Meck-
lenburg, M., Schulte, K., Erts, D., Lupan, O., Fiedler, B., et al., 2017.
Nanomechanics of individual aerographite tetrapods. Nature communi-
cations 8, 1–9.
NIDC, 2022. unofficial price quote via email from national isotope
development center, oar ridge, tn. URL: https://www.isotopes.gov.
Parkin, K.L.G., 2018a. The Breakthrough Starshot system model.
Acta Astronaut. 152, 370–384. doi:10.1016/j.actaastro.2018.08.035,
arXiv:1805.01306.
Parkin, K.L.G., 2018b. The Breakthrough Starshot system model.
Acta Astronautica 152, 370–384. doi:10.1016/j.actaastro.2018.08.035,
arXiv:1805.01306.
Parkin, K.L.G., 2022. Cost-Optimal System Performance
Maps for Laser-Accelerated Sailcraft. arXiv e-prints ,
arXiv:2205.13138arXiv:2205.13138.
Parrington, J., Knox, H., Breneman, S., Baum, E., Feiner, F., 1996. Knolls
atomic power laboratory chart of the nuclides.
Rejcek, J., Browder, M., Fry, J., Koymen, A., Weiss, A., Lockheed, A.,
Missiles, M., Control, F., Usa, D., 2003. Alternative pathways to
antimatter containment. Radiation Physics and Chemistry 68, 655–661.
doi:10.1016/S0969-806X(03)00195- 6.
Sekanina, Z., 1976. A Probability of Encounter with Interstellar Comets
and the Likelihood of their Existence. Icarus 27, 123–133. doi:10.1016/
0019-1035(76)90189- 5.
Shah, T.K., Patel, M.D., Patel, S.B., Bist, N., Sircar, A., 2022. Synthesis of
aerographene for oil spill remediation. Materials Today: Proceedings .
Siraj, A., Loeb, A., 2019. Identifying Interstellar Objects Trapped in
the Solar System through Their Orbital Parameters. Ap. J. 872, L10.
doi:10.3847/2041-8213/ab042a,arXiv:1811.09632.
Stern, S.A., 1990. On the Number Density of Interstellar Comets as a
Constraint on the Formation Rate of Planetary Systems. Pub. Astron.
Soc. Pacific 102, 793. doi:10.1086/132704.
Trilling, D.E., Robinson, T., Roegge, A., Chandler, C.O., Smith, N., Loef-
fler, M., Trujillo, C., Navarro-Meza, S., Glaspie, L.M., 2017. Implica-
tions for Planetary System Formation from Interstellar Object 1I/2017
U1 (’Oumuamua). Ap. J. Lett. 850, L38. doi:10.3847/2041-8213/aa9989,
arXiv:1711.01344.
Wang, X., Guo, L., Zhang, L., Liu, Y., 2014. Link design of Moon-to-
Earth optical communication based on telescope array receivers. Optics
Communications 310, 12–18. doi:10.1016/j.optcom.2013.07.079.
Youinou, G., Lin, C., 2022. Preliminary conceptual design of fast neutron
spectrum nuclear thermal rocket cores using monolithic uranium
nitride fuel. Progress in Nuclear Energy 149, 104237. URL: https:
//www.sciencedirect.com/science/article/pii/S0149197022001123,
doi:https://doi.org/10.1016/j.pnucene.2022.104237.
Author biography without author photo. Author biography. Author bi-
ography. Author biography. Author biography. Author biography. Author
biography. Author biography. Author biography. Author biography. Author
biography. Author biography. Author biography. Author biography. Author
biography. Author biography. Author biography. Author biography. Author
biography. Author biography. Author biography. Author biography. Author
biography. Author biography. Author biography. Author biography. Author
biography. Author biography.
Author biography with author photo. Author biog-
raphy. Author biography. Author biography. Au-
thor biography. Author biography. Author biogra-
phy. Author biography. Author biography. Author
biography. Author biography. Author biography.
Author biography. Author biography. Author bi-
ography. Author biography. Author biography. Au-
thor biography. Author biography. Author biogra-
phy. Author biography. Author biography. Author
T.M. Eubanks et al. Page 23 of 24
Swarming Proxima Centauri
biography. Author biography. Author biography.
Author biography. Author biography. Author biog-
raphy.
Author biography with author photo. Author biog-
raphy. Author biography. Author biography. Au-
thor biography. Author biography. Author biogra-
phy. Author biography. Author biography. Author
biography. Author biography. Author biography.
Author biography.
T.M. Eubanks et al. Page 24 of 24
Swarming Proxima Centauri
Table 1: Mission Traceability Matrix
Mission Goals Problem Solution Implications
Surviving to Proxima b (1) Dust impact ends mission with 1 spacecraft Fly many spacecraft Swarm benefits described here
Surviving to Proxima b (2) 20-MeV proton radiation at 0.2 c Fly “edge on.” Extra shielding in the rim
Assembling the Swarm Swarm is launched sequentially Use ISM drag to cohere swarm Ability to tack spacecraft as needed
Close Observation of Proxima b Ephemeris errors of 105km 100 - 1000 Spacecraft swarm Assembling the Swarm
Complete observation of Proxima b Single spacecraft has only one vantage point Swarm has multiple vantage points Enables transmission spectroscopy
Communicate with Earth during flyby This will limit observations with that probe Swarm elements can be sacrificed Could have probes impact Proxima b
Interswarm Communication 1000 - 10,000 km separation Laser coms in spacecraft rim Rim IR laser system
Regular communication with Earth Need continuous power for decades Betavoltaic power with 90Sr 330 mg power system
Gigabytes of data return per year Low photon rate with 1 spacecraft “Photon Coherent” swarm laser coms Swarm clock synchronization
T.M. Eubanks et al. Page 25 of 24
Swarming Proxima Centauri
Swarm or Probe Parameter Value [units]
Transverse swarm diameter at flyby [km] 100,000
Number of surviving probes at flyby, after 4-LY cruise 300
Average probe spacing within swarm [km] 5700
Individual probe diameter [mm] 4000
Probe rim height [mm] 20
Main disk thickness [mm] 10
Mass budget: disk [mg] 330
Mass budget: betavoltaic battery and ultracapacitor pulsed storage [mg] 330
Mass budget: rim, inter-probe coms, computation and everything else [mg] 340
Overall input electrical power per probe, at flyby [mW] 6
Input electrical power to Swarm-Earth or inter-probe lasers, at flyby [mW] 4
Optical output power per probe, at flyby [mW] 0.4
Swarm-Earth communications wavelength source / as received red-shifted [nm] 432 / 539
Number probe-Earth communications apertures 247
Inter-swarm (rim) communications wavelength [nm] 12,000
Number of rim transceivers per probe 5
Table 2
Basic parameters of the proposed probe swarm. The aerographene metamaterial that forms the main probe body has a variable
density, tailored for the particular mechanical requirements. A denser layer will support the drive-beam dielectric reflector while
the middle layer will be very sparse. There would be thicker areas around the sensor/communications telescope array on the front
face and around the betavoltaic, capacitor, and electronics layers to support them.
Thickness of Spacecraft thickness Mass of Total spacecraft Velocity
aerographene layer with aerographene sheet aerographene layer mass Ratio
100 nm 200 nm 2.4 ×10−7 gm 3.6 gm 10−6
1𝜇m 1.1 𝜇m 0.0024 gm 3.6 gm 10−5
100 𝜇m 100.1 𝜇m 0.24 gm 3.84 gm 10−3
1 mm 1 mm 2.4 gm 6 gm 10−2
10 mm 10 mm 24 gm 27.6 gm 10−1
Table 3
Aerographene layers added on top of the spacecraft in flight direction with the same surface area as the spacecraft and their
impact on spacecraft mass and generated interstellar hydrogen beam drag (v = 0.2c; M = 3.6 grams; d = 4.1 m.) The fifth
column is the ratio of velocity lost from interaction with the ISM to initial launch velocity.
T.M. Eubanks et al. Page 26 of 24
Swarming Proxima Centauri
Table 4
Mass Partition of the Probe
Main optical disk (%) 33
Betavoltaic cells and Ultracapacitors (%) 33
Rim and Everything else (%) 34
T.M. Eubanks et al. Page 27 of 24
