Postmanetal.

:ATLASTScienceDrivers

Science drivers and requirements for an Advanced Technology Large

Aperture Space Telescope (ATLAST): Implications for technology
development and synergies with other future facilities
Marc Postmana, Tom Browna, Kenneth Sembacha, Mauro Giavaliscob, Wesley Traubc, Karl
Stapelfeldtc, Daniela Calzettib, William Oegerled, R. Michael Riche, H. Phillip Stahlf, Jason
Tumlinsona, Matt Mountaina, Rmi Soummera, Tupper Hyded
a

Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD USA 21218;
b
University of Massachusetts, Dept. of Astronomy, Amherst, MA USA 01003;
c
Jet Propulsion Laboratory/California Institute of Technology, Pasadena, CA USA 91109;
d
Goddard Space Flight Center, Greenbelt, MD USA 20771;
e
University of California, Division of Astronomy, Los Angeles, CA USA 90095;
f
Marshall Space Flight Center, MS SD70 SOMTC, Huntsville, AL USA 35812-0262
ABSTRACT
The Advanced Technology Large-Aperture Space Telescope (ATLAST) is a concept for an 8-meter to 16-meter UVOIR
space observatory for launch in the 2025-2030 era. ATLAST will allow astronomers to answer fundamental questions at
the forefront of modern astronphysics, including "Is there life elsewhere in the Galaxy?" We present a range of science
drivers that define the main performance requirements for ATLAST (8 to 16 milliarcsec angular resolution, diffraction
limited imaging at 0.5 m wavelength, minimum collecting area of 45 square meters, high sensitivity to light
wavelengths from 0.1 m to 2.4 m, high stability in wavefront sensing and control). We will also discuss the synergy
between ATLAST and other anticipated future facilities (e.g., TMT, EELT, ALMA) and the priorities for technology
development that will enable the construction for a cost that is comparable to current generation observatory-class space
missions.
Keywords: Advanced Technology Large-Aperture Space Telescope (ATLAST); ultraviolet/optical space telescopes;
astrophysics; astrobiology; technology development.

1. Introduction
The most compelling astrophysical questions to be addressed in the 2020 era will, like those today, be pursued using data
obtained from both space-based and ground-based telescopes. The impressive capabilities anticipated for ground-based
observatories in the upcoming decade (e.g., 20 to 40-meter class optical telescopes, ALMA, and potentially the square
kilometer array (SKA) near the end of the decade) will redefine the existing synergy between ground and space
telescopes. Advances, over the next decade, in Multi-Conjugate Adaptive Optics (MCAO) and Ground-Layer Adaptive
Optics (GLAO) for large aperture ground-based telescopes1,2,3,4 may enable intermediate to high Strehl ratio (~40-80%)
performance over fields of view of perhaps up to 2 arcminutes across for wavelengths longwards of ~1 m. Advances in
Extreme AO5,6,7 may enable high Strehl ratio performance down to wavelengths as short as 0.55 m but only over very
small fields of view of 1 or 2 arcseconds. Space-based telescopes, however, are the optimal facilities for observations
that require any combination of very high-angular resolution and precise wavefront control over fields of view larger
than ~2 arcminutes or at all wavelengths shorter than 1 m, very high sensitivity (nanoJansky levels), very stable PSF
performance across the field of view, high photometric precision (< 0.0001 mag) and accuracy in crowded fields, and
very high stability of all these performance parameters over tens to hundreds of hours of exposure time.

Email: postman@stsci.edu; phone (410) 338-4340

Postmanetal.:ATLASTScienceDrivers
Several independent scientific drivers that we will highlight here require primary aperture diameters of at least 8 meters
and, for some applications, as large as 16 meters. The next large UVOIR space telescope must be diffraction-limited at
least down to 0.5 m and must have good sensitivity down to 0.11 m. A space telescope with such capability will be
revolutionary because it enables fundamental breakthroughs in astrophysics both on its own and in combination with
other telescopes with different capabilities. We have studied several designs for this next generation space telescope, a
suite of concepts we collectively refer to as the Advanced Technology Large-Aperture Space Telescope (ATLAST)
8,9,10,11
. ATLAST has the performance required to detect the potentially rare occurrence of biosignatures in the spectra of
terrestrial exoplanets, to reveal the underlying physics that drives star formation, and to trace the complex interactions
between dark matter, galaxies, and the intergalactic medium.
The challenge the astronomy community faces, given the compelling science enabled by ATLAST, is how to construct it
for a cost that is comparable to current flagship-class missions. Modern ground-based telescopes (post-1992) follow an
aperture size cost relationship that is significantly shallower than older designs, owing to the advance of technology.
Similar trends are seen for space-based telescopes12,13. The construction costs for the James Webb Space Telescope
(JWST) are similar to those of the Hubble Space Telescope (HST) and yet JWST has an aperture that is 2.7 times larger
and a mass that is a factor of 1.7 smaller. Key technologies that can continue to flatten the slope of the space telescope
aperture-cost relationship include lightweight mirror materials and fabrication methods14,15,16, closed-loop wavefront
control of active optics, disturbance isolation systems, modular design, high efficiency UV detectors, and ultra-low noise
optical/IR detectors. If significant investments are made in these technologies, it is plausible that 8-meter and larger
UVOIR space telescopes would be affordable by NASA in the 2020 era. An important element to achieving maturity of
the technology needed for a large UVOIR (non-cryogenic) space telescope is partnership with other communities that
have similar technology drivers (e.g., national reconnaisance, defense, and remote sensing of Earth).

2. Astrophysical Science Drivers

Conceptual breakthroughs in understanding astrophysical phenomena are made when our ability to probe structures on
the relevant angular scales is enabled by our astronomical observatories. HST tremendously advanced our understanding
of galaxy evolution because its angular resolution (~65 mas) at its diffraction limit (0.63 m) is, for all redshifts, less
than or equal to one half of the angular scale of the kiloparsec-size features within galactic systems that reveal key
morphological information. ATLAST will be poised to make equally large scientific breakthroughs if it can achieve
spatial resolutions that are more than 4 times better than HST and JWST. Achieving ultra-high dynamic range (highcontrast) imaging and/or long-term imaging with high spatial and temporal PSF stability at angular resolutions of ~8 to
15 mas in the 0.11 m to 2.4 m wavelength range are required for some extremely compelling science that will not be
readily achieved by any other facility. For example, various combinations of high spatial resolution, high sensitivity, and
high contrast are needed if we wish to definitively detect the potentially rare occurrence of biosignatures in the
atmospheres of terrestrialmass exoplanets, or to study stellar populations across the full range of galaxy environments,
or to trace the kinematics of gas and dark matter on galactic scales to directly map the growth of structure over time. We
discuss these cases in brief detail below. These are just a few of the many exciting investigations requiring a large
UVOIR space telescope (and which will enable an equal or greater number of as yet unimagined discoveries).
2.1 Terrestrial exoplanet characterization: Are We Alone?
We are at the brink of answering two paradigm-changing questions: Do Earth-sized planets exist in the Habitable Zones
of their host stars? Do any of them harbor life? The tools for answering the first question already exist (e.g., Kepler,
CoRoT); those that can address the second can be developed within the next 10-20 years17. ATLAST is our best option
for an extrasolar life-finding facility and is consistent with the long-range strategy for space-based initiatives
recommended by the AAAC Exoplanet Task Force18. Four significant drivers dictate the need for a large space-based
telescope if one wishes to conduct a successful search for biosignatures on exoplanets. First, and foremost, Earth-mass
planets are faint an Earth twin at 10 pc, seen at maximum elongation around a G-dwarf solar star, will have V ~ 29.8
AB mag. Detecting a biosignature, such as the presence of molecular oxygen in the exoplanets atmosphere, will require
the ability to obtain direct low-resolution spectroscopy of such extremely faint sources. Second, the average projected
angular radius of the Habitable Zone (HZ) around nearby F,G,K stars is less than 100 milli-arcseconds (mas). One thus

Postmanetal.:ATLASTScienceDrivers
needs an imaging system capable of angular resolutions of ~10 to 25 mas to adequately sample the HZ and isolate the
exoplanet point source in the presence of an exo-zodiacal background. Third, direct detection of an Earth-sized planet in
the HZ requires high contrast imaging, typically requiring starlight suppression factors of 109 to 1010. Several
techniques 19 are, in principle, capable of delivering such high contrast levels but all require levels of wavefront stability
not possible with ground-based telescopes. Guyon20 and Mountain et al.21 show that there is a limit to the achievable
dynamic range in high contrast imaging from the ground. This is because any AO system is fundamentally limited by the
capacity to analyze the wavefront due to the finite number of photons available for wavefront sensing. Even with the 30
to 40 m class telescopes expected to come on line later this decade, this limit corresponds to a contrast ratio of ~10-8 (see
Figure 4 in Mountain et al.21). A space-based platform is thus required to achieve the wavefront stability that is needed
for such high contrast imaging of terrestrial mass planets in the HZ. Lastly, biosignature-bearing planets may well be
rare, requiring one to search tens or even several hundred stars to find even a handful with compelling signs of life.
Therefore if we have to survey a number of stars, and hence limit the exposure time on each candidate, the number of
stars for which one can obtain an exoplanets spectrum at a given SNR scales approximately as D 3, where D is the
telescope aperture diameter. This is demonstrated in Figure 1 where we have averaged over different simulations done
using various starlight suppression options (internal coronagraphs of various kinds as well as an external occulter). To
estimate the number of potentially habitable worlds detected, one
must multiply the numbers in Figure 1 by the fraction of the stars
that have an exoplanet with detectable biosignatures in their HZ
(EARTH). The value of EARTH is currently not constrained but it is
not likely to be close to unity. One must conclude that to maximize
the chance for a successful search for life in the solar neighborhood
requires a space telescope with an aperture size of at least 8 meters.
Figure 2 shows two simulated ATLAST spectra for an Earth-twin at
10 pc, one at R=100 and one at R=500, taken with sufficient
exposure to reach SNR=10 at 0.75 m in the continuum. A 3-zodi
background was used (local plus exosolar). For these calculations we
use a fully validated model of the Earths spectrum22,23 in
combination with the observed visible reflection spectrum of the
present Earth. We assume that the exoplanet is at maximum
Figure 1: The number of spectral type F,G,K stars elongation. The R=100 exposure times are 46 ksec and 8 ksec,
as a function of telescope aperture where an R=70 respectively, for an 8-m and 16-m space telescope. The
SNR=10 spectrum could be obtained of an Earth- corresponding exposure times for the R=500 spectrum are 500 ksec
twin in the HZ in less than 500 ksec. Light grey
and 56 ksec, respectively, for the 8-m and 16-m telescopes. The
shows total number of stars that could be observed
2/3
at least once. Dark grey shows number of stars that reflected flux from an Earth-like rocky planet increases as M ,
could be visited 3 times in 5 years without where M is the exoplanet mass. Hence, the exposure times for a 5
exceeding 20% of available telescope time. It is Earth-mass exoplanet would be ~3 times shorter. At both
assumed every star has an Earth twin.
resolutions, the O2 features at 0.68 m and 0.76 m are detected, as
are the H2O features at 0.72, 0.82, 0.94, 1.13, 1.41, and 1.88 m.
Rayleigh scattering is detected as an increase in reflectivity bluewards of 0.55 m. The higher spectral resolutions
enabled by large-aperture space telescopes enable the detection of molecular oxygen in exoplanets with lower
abundances than those on Earth and provide constraints on the kinematics and thermal structure of the atmosphere that
are not accessible at lower resolution.
For a 16-m class space telescope, time-resolved spectroscopy over intervals of a few hours may reveal surface
composition variations, if the planet is not cloud dominated, as the exoplanet rotates. However, even broadband
photometry can be used to detect short-term variations in albedo that can determine the rotation period and constrain the
amount of cloud cover. ATLAST will allow us to glean substantial information about an exo-Earth from temporal
variations in its features. Such variations inform us about the nature of the dominant surface features, changes in climate,
changes in cloud cover, and potentially, seasonal variations in surface vegetation. Ford et al.24 generated model light
curves for the Earth over 6 consecutive days using data from real satellite observations. Photometric variations of 20
30% on timescales of 6 hours were typical in the B,V,R,I passbands. To track such variations with a SNR of 20 on an

Postmanetal.:ATLASTScienceDrivers
Earth-like planet (with a similar rotation period as Earth) at a distance of 20 pc would require a space telescope with an
aperture of at least 8 meters. A 4-meter space telescope would be able to perform such observations only for planetary
systems within 10 pc. As the number of terrestrial planetary systems scales as the cube of the distance, the ability to
reach to 20 pc provides nearly an order of magnitude more targets.
The instrumentation required to perform the above observations on an Earth-twin at up to 20 pc distance requires a
starlight suppression system that allows detection of exoplanets that are ~25 magnitudes fainter than their host star at an
inner working angle of ~40 mas. This is the baseline mission starlight suppression performance requirement. There are
several options for the suppression system an internal coronagraph or external occulter. Achieving the above level of
suppression with a segmented telescope will require development of a nulling coronagraph and/or an external occulter
(starshade)1. While the design of the
external occulter is independent of
the telescopes optical design, the
viable options for an internal
coronagraph are dependent upon the
telescopes optical design: a Lyot or
masked-based coronagraph can be
used with telescope that employs an
off-axis secondary mirror (SM) and
a monolithic primary mirror (PM)
and, possibly, with one that has an
on-axis SM and monolithic PM if
the SM is supported by a single
linear structure. However, for an onaxis SM with standard spider
supports or for a segmented PM, the
Figure 2: Simulated ATLAST spectra of an Earth-twin at 10 pc, shown at R=100 only internal coronagraph concept
(top) and R=500 (bottom). The SNR=10 at 0.75 micron in both cases. Key molecular
that would, in principle, do the job is
oxygen and water features are shown. Increased reflectcance at the blue end is due
a visible nulling coronagraph
to Rayleigh scattering.
(VNC). As part of the early
technology development plan for ATLAST, these options would be investigated and a downselect made prior to entry
into phase A.
An 8-m ATLAST (with an internal coronagraph) will be able to observe ~100 different star systems 3 times each in a 5year interval and not exceed 20% of the total observing time available to the community. A 16-m version (with an
internal coronagraph) could visit up to ~250 stars three times each in a five-year period. ATLAST used in conjunction
with a single external occulter can observe ~85 stars 3 times each in a 5-year period, limited by the transit times of the
occulter.
The impact on the ATLAST science objectives if a starlight suppression system is not able to permit detection of sources
at ~10-10 contrast ratio at the indicated inner working angle (IWA) would be the inability to directly detect and
characterize Earth-mass planets around solar-type stars (indirect characterization via transit spectroscopy would be
unaffected). Characterization of super-Earths (up to 10x Earths mass) or Earth-like planets around M-dwarfs would still
be possible, however, if contrasts of 10-9 are reachable. We adopt 10-9 as the minimum mission criterion for starlight
suppression performance. The ATLAST science flowdown for terrestrial exoplanet characterization is shown in Table 1.

1AthoroughdiscussionofstarlightsuppressionforATLASTisavailableintheappendicesofourpublicNASAstudyreportat
http://www.stsci.edu/institute/atlast.

Postmanetal.:ATLASTScienceDrivers
Table 1. ATLAST Science Flowdown Requirements for Biosignature Detection on Exoplanets
Science
Science Requirements
Measurements Needed
Design and Implementation
Question
Detect at least 10 Earthlike Planets in HZ with
95% confidence if
EARTH = 0.15
Is there life
elsewhere in
the Galaxy?

High contrast (Mag>25 mag)

SNR=10 broadband (R=5)
imaging with IWA ~ 40 mas for
~100 target stars.

At least 8 m PM aperture to achieve

sample size (~100 stars) and the desired
SNR spectra in <500 ksec for most
distant (~25 pc) exoplanets in the sample.
-10

Stable 10 starlight suppression on

detector at IWA.

Detect the presence of

habitability and biosignatures in the spectra
of Earth-like HZ planets

High contrast (Mag>25 mag)

SNR=10 low-resolution (R=70100) spectroscopy with an IWA
~ 40 mas. Exposure times <500
ksec.

0.1 nm WFE at 0.6 m over 2 hr;

~1.3 to 1.6 mas pointing stability.
IFU with sensitivity from 0.3 to 2.4 m,
with broadband imaging and two
spectroscopic modes (R=70, R=500).

2.2 Stellar population characterization

ATLAST will, for the first time, enable the reconstruction of complete star formation histories (spanning 10 Gyr) for
hundreds of galaxies beyond the Local Group, opening the full range of star formation environments to exploration. A
comprehensive and predictive theory of galaxy formation and evolution
requires that we accurately determine how and when galaxies assemble
their stellar populations, and how this assembly varies with environment.
By definition the dwarf galaxies we see today are not the same as the
dwarf galaxies and proto-galaxies that were disrupted during assembly.
Our only insight into those disrupted building blocks comes from sifting
through the resolved field populations of the surviving giant galaxies to
reconstruct the star formation history, chemical evolution, and kinematics
of their various structures25. Resolved stellar populations are cosmic
clocks. Their most direct and accurate age diagnostic comes from
observations that can resolve the individual, older stars that comprise the
main sequence turnoff. But the main sequence turnoff rapidly becomes
too faint to detect with any existing telescope for any galaxy beyond the
Local Group. This greatly limits our ability to infer much about the
details of galactic assembly because the galaxies in the Local Group are
not representative of the galaxy population at large. ATLAST will allow
us to reach well beyond the Local Group as shown in Figure 3. HST and
Figure 3: Nearby galaxy distribution, colorJWST cannot reach any large galaxies besides our Milky Way and M31 coded by morphological type (blue = spiral,
because they lack the required angular resolution. An 8-meter space orange = elliptical, olive: dwarf galaxy). Large
telescope can reach 10 Gyr old stars in 140 galaxies including 12 giant circles show how far away a solar luminosity
spirals and the nearest giant elliptical. A 16-meter space telescope star can be detected at SNR=5 in V and I band
extends our reach to the Coma Sculptor Cloud, netting a total of 370 in a 100 hour exposure.
galaxies including 45 giant spirals and 6 ellipticals. Deriving ages and other galactic properties from color-magnitude
data requires photometry for thousands of stars spanning 4 orders of magnitude in luminosity. Such observations require
a wide-field imager on ATLAST with half-Nyquist sampling over a field of view of at least 4 arcminutes. The ATLAST
science flowdown for star formation history reconstruction is shown in Table 2.
ATLAST will work in concert with 30m-class ground-based telescopes (e.g., TMT), expanding our reach to other wellpopulated galaxy groups, with ATLAST obtaining photometry of V~35 magnitude G dwarf stars and TMT obtaining
kinematics of much brighter giants out to the Coma Sculptor Cloud. The dwarf stars in the Coma Sculptor Cloud are

Postmanetal.:ATLASTScienceDrivers
effectively inaccessible to TMT, requiring gigaseconds of integration even for an isolated star. Ground-based telescopes
and ATLAST will also complement each other in establishing the universality of the local initial stellar mass function
(IMF) a fundamental observable that must be predicted by any viable comprehensive theory of star formation. Large
ground-based telescopes with NIR AO-enabled imaging will establish the universality of the low-mass (<2 MSUN) end of
the IMF and ATLAST will definitively determine the same for the high-mass end of the IMF, which can only be directly
measured in UV / optical.
Table 2. ATLAST Science Flowdown Requirements for Star Formation History Reconstruction
Science
Science Requirements
Measurements Needed
Design and Implementation
Question

What are the

star
formation
histories of
galaxies?

Determine the ages and

metallicities of stellar
populations over a broad
range of galactic
environments. Accuracy
requirements: age bins
~1 Gyr over lifetime of
galaxy; metallicity bins
of ~0.2 dex over full
range of abundances.

Color-magnitude diagrams using

broadband imaging (SNR=5) for
individual solar analog stars
(absolute V mag ~ 5) in spiral,
lenticular, and elliptical galaxies.
Imaging must be done in at least
two passbands, with bluer band
below 0.6 m wavelength.

At least 8 m PM aperture to observe

nearest giant elliptical galaxy and get
SNR in 2 passbands in a total exposure
<400 ksec.
WFE: Diffraction limited at 0.5 m;
~1.3 to 1.6 mas pointing stability;
Symmetric PSF highly desirable
VIS/NIR wide-field (4 - 8 arcmin FOV)
imager at TMA focus for simultaneous
photometry of >10,000 stars

2.3 Galaxy Halo and Gas Physics Revealed in Unprecedented Detail

The ultraviolet (UV) region of the electromagnetic spectrum is highly sensitive to many fundamental astrophysical
processes and, hence, measurements at rest-UV wavelengths provide robust, and often unique, diagnostics of the roles of
these processes in a establishing a variety of astronomical environments and in controlling the evolution of a variety of
objects. Our knowledge of star formation and evolution, of the growth of structure in the universe, of the physics of jet
phenomena on many scales, of the nature of
AGN, of aurora on and atmospheric composition
of the gas giant planets, and of the physics of
protoplanetary disks has either been gleaned or
greatly expanded through UV observations.
There is great scientific power in combining high
spatial
resolution
with
sensitive
UV
spectrographic capabilities. One very important
application is in the area of galaxy formation. We
know that galaxies form and evolve but we know
little about how this happens. The physical
processes involve complex interactions between
the baryonic matter in the galaxies, the energy
exchanged during the birth and death of stars, the
gas outside the galaxies in the intergalactic
medium (IGM), other neighboring galaxies, and
the dark matter that dominates and shapes the
underlying gravitational potential. Revealing the
Figure 4: Simulations of the gas density (as a function of ionization
state) and temperature distribution for 4 different size galaxies physics behind galaxy formation and evolution
requires making a broad array of observations
(Oppenheimer et al. 2009 26). ATLAST will map these distributions.
from the current epoch to the epoch of the first
stars. By enabling deep and extensive probes of the IGM in the UV and optical regime, ATLAST will provide many of
the key pieces needed to solve this puzzle, particularly in the redshift range z < 3 when the cosmic star formation rate
peaks and then fades, and galaxies develop their current morphologies.

Postmanetal.:ATLASTScienceDrivers

Understanding how gas in the IGM gets into galaxies and how galaxies respond to inflow lies at the heart of
understanding galactic evolution. The mode of accretion depends on the depth of the potential well (galaxy type) and the
location at which the intergalactic gas is shocked as it encounters that potential27,28. Depending on the mass of the galaxy
halo, the infalling gas may be shocked and heated or accrete in cold mode along narrow filaments. Gas can also be
removed from galaxies via tidal and ram pressure stripping, or during the accretion of gas-rich dwarfs onto giant
galaxies. Metal-enriched gas introduced into the IGM by these processes will be dynamically cool. All of these accretion
and gas removal theories have observational consequences that can be tested if the properties of gas (e.g., temperature,
density, velocity dispersion, metallicity) in and around galaxies can be characterized through absorption and emission
line spectroscopy. Figure 4 shows the variation in these properties that models predict as a function of galaxy halo mass.
Access to UV wavelengths is required to observe the (slightly redshifted) diagnostic lines (e.g., OVI, SiIII, Ly, NV,
SiIV, CIV) needed to characterize the warm IGM at low redshift or OIII, OV, NIV, NeVIII for characterization of the
intermediate redshift (0.3 < z < 2) IGM. The observational challenge is to acquire datasets of sufficient spatial sampling
and with enough diagnostic power (i.e., spectral resolution) to identify and characterize gas in galactic halos. The key
requirement is to observe a sufficient number of background sources around each galaxy as well as a large range in
galaxy parameter space (e.g., mass, morphology, star formation rate, redshift). Covering a broad range in all parameters
is currently beyond the capability of HST. ATLAST will have sufficient UV absorption line sensitivity to be able to
survey up to ~100 quasars per deg2 (corresponds to a flux limit of ~1 x 10-17 erg cm-2 sec-1 -1) over an area that subtends
~10 Mpc on a side (0.25 deg2 at z = 0.4) and obtain SNR=10-20 high-resolution (R=20,000) spectra of each QSO in the
region. At this limit, ~20% of randomly selected fields on the sky would have a sufficient number of background sources
to enable detailed mapping of the spatial distribution of IGM structure around specific galaxies. ATLASTs large
aperture coupled with efficient (~40% QE) UV detectors and an efficient spectrograph are needed to enable an
individual QSO observation to be completed in ~25 ksec (which permits a 0.25 deg2 area to be surveyed in ~1 week).
Table 3. ATLAST Science Flowdown Requirements for IGM Characterization
Science
Science Requirements
Measurements Needed
Design and Implementation
Question
How do
galaxies and Map, at high spatial
High efficiency UV detectors with at
the IGM
sampling, the properties
least an 8 m PM aperture to survey wide
SNR=20 high resolution
interact?
and kinematics of the
areas in less than 2 weeks of time.
(R=20,000) UV spectroscopy of
How does
intergalactic medium,
this
quasars down to FUV mag = 24. Hi-res UV/blue spectrograph at Cass
over contiguous regions
interaction
of the sky, on scales up
focus with sensitivity from 0.11 0.4 m.
effect galaxy to ~10 Mpc.
evolution?
The dramatically increased absorption line sensitivity at UV and optical wavelengths of ATLAST is crucial for reaching
the required background source densities. At the required sampling given above, one can select sight lines next to
thousands of examples of any common galaxy, group, or cluster. ATLAST could then be used to produce a highresolution map of the gas and metals surrounding these structures, which could be used to compare directly against
simulation predictions29,30. With ATLAST one could also use multiple quasars and distant galaxies as background
continuum sources to dissect the gas distribution in fields known to have galaxies and gas at the same redshift31.
ATLASTs large aperture will enable contiguous regions of ~10 Mpc on a side (like in Figure 4) to be surveyed in about
2 weeks of exposure time (with an 8-m ATLAST with enhanced UV detectors) or in ~3 days (with a 16-m ATLAST
with enhanced UV detectors)32. ATLAST could also be used systematically to target individual nearby galactic coronae
and groups of galaxies, for which it would be possible to observe the production sites of heavy elements (star-forming
regions, SNe, emission nebulae), follow the processes by which the elements are transported within galaxies, and trace
the expulsion of metals into the IGM. The ATLAST science flowdown for IGM characterization is shown in Table 3.
Large ground based telescopes will probe IGM kinematics and structure at higher redshifts where the key spectral
features get redshifted into optical or near-IR. However, the increasingly ubiquitous Ly-alpha absorption at higher z is

Postmanetal.:ATLASTScienceDrivers
challenging. Hence, the combination of a space-based survey covering z < 2 regime and ground-based covering z>2
makes for an ideal combination.
Recent observations with the Cosmic Origins Spectrograph (COS) on Hubble have shown that it is now possible to
detect million-degree intergalactic gas in the ultraviolet lines of O VI and H I Ly-alpha33. Studies of intergalactic gas at
these temperatures have been the province of x-ray observatories, but may now be broadened significantly to include
dedicated studies with future UV-optical telescopes in space. The higher resolution (10-100x) and higher sensitivity
(100-1000x) achievable with UV-optical observatories such as ATLAST will at last permit the observational study of
this important component of the hot IGM in the context of galaxy environments for large samples (100s-1000s)
of intergalactic sight lines. Hence, joint observations with ATLAST and an x-ray observatory will provide very
powerful constraints on the hot gas distribution.
2.4 Dark Matter Dynamics
Dwarf spheroidal galaxies (dSph), the faintest galaxies known, are extraordinary sites to explore the properties of nonbaryonic dark matter (DM). There are several reasons for this. First, their mass is dominated by DM they are observed
to have mass-to-light ratios 10 to 100 times higher than the typical L* galaxy, such as M31 or the Milky Way34,35,36.
Second, they are relatively abundant nearby to date ~40 dSph galaxies have been found in the Local Group and more
will be discovered. Third, and perhaps most striking, is the discovery that all nineteen dSph satellites of the Milky Way,
covering more than four orders of magnitude in luminosity, inhabit dark matter halos with the same mass (~107 MSUN)
within their central 300 pc37. The ability of DM to cluster in phase space is limited by intrinsic properties such as mass
and kinetic temperature. Cold dark matter particles have negligible velocity dispersion and very large central phasespace density, resulting in cuspy density profiles. Warm dark matter halos, in contrast, have smaller central phase-space
densities, so that density profiles saturate to form constant central cores. Owing to their small masses, dSphs have the
highest average phase space densities of any galaxy type, and this implies that for a given DM model, phase-space
limited cores will occupy a larger fraction of the virial radii. Hence, the mean density profile of dSph galaxies is a
fundamental constraint on the nature of dark matter.
Current observations are unable to measure the density profile slopes within dSph galaxies because of a strong
degeneracy between the inner slope of the DM density profile and the velocity anisotropy of the stellar orbits. Radial
velocities alone cannot break this degeneracy even if the present samples of radial velocities are increased to several
thousand stars37. Combining proper motions with the radial velocities is the only robust means of breaking the
anisotropy-inner slope degeneracy. The required measurements include proper motions for ~100 stars per galaxy with
accuracies better than 10 km/sec (< 40 as/yr at 60 kpc) and ~1000 line-of-sight velocities. In the case of the brightest of
these dSph galaxies, such as Fornax and Sculptor, sufficient velocities and proper motions can be obtained using stellar
giants. ATLAST, however, can perform the astrometric measurements. To accomplish this, ATLAST will measure
transverse stellar velocities to an accuracy of 5 km/sec. At a distance of 50 kpc, this corresponds to an angular
displacement of 0.1 mas over 5 years (about 40 times better than what HST can currently measure38). This is
approximately one two-hundredths of a pixel, or equivalently, one two-hundredth of the FWHM of the point spread
function (PSF). For reference, the Advanced Camera for Surveys (ACS) on HST has centroiding errors of about onehundredth of the pixel and PSF (the pixel size and PSF have nearly identical widths). However, HST suffers from large
thermal stresses on orbit, which cause significant changes in the lengths, positions and alignment of the supporting
structures. At the Sun-Earth second Lagrange point (SEL2), ATLAST is far more thermally stable, and the sensors and
actuators put in place to maintain the structure to the precision necessary for exoplanet science allows ATLAST to
achieve one-sigma astrometric errors of 0.005 pixels. In addition to the instrumental capability for astrometry to 0.1 mas,
background objects are needed to provide a stable astrometric reference frame. Quasars alone are likely to be too sparse
to provide this frame, so they will be supplemented with background galaxies. While an individual galaxy is far less
valuable as an astrometric source than a quasar of a similar magnitude, if the imager used for this investigation has a
FOV wide enough to contain thousands of galaxies in a single exposure, then their internal structures, which will be
resolved down to 15 mas, will provide the required reference frame. These observations require a wide-field imager on
ATLAST with a field of view of ~5 arcminutes. This experiment is not signal-to-noise limited, as solar mass stars at 50
kpc will have signal-to-noise ratios exceeding 100 in ~1 ksec exposures. Thus, even in some of the faintest known dwarf

Postmanetal.:ATLASTScienceDrivers
spheriodals, with mass to light ratios exceeding 10,000, the transverse velocity measurements for the 200+ stars per
dwarf required by this experiment can be obtained. The dark matter kinematics science flowdown is shown in Table 4.
Ground-based 8-m class telescopes could measure the spectra, and SIM could measure the proper motions for the nearest
dwarfs. For the less massive dwarfs, where the dark matter dominance is the greatest, main sequence stars will have to
be used to obtain the numbers of velocity and proper motion measurements needed. This will require larger (30-m class)
ground-based telescopes for the velocities. The 30-m class telescopes may also be able to obtain the necessary proper
motions but it will be extremely challenging: it will require precisely stitching many fields together, most of which are
unlikely to contain background quasars of sufficient brightness to be useful as astrometric references. However, the
necessary astrometric precision would be readily achieved with ATLAST, given its comparatively wide field of view
and stability. ATLAST will, thus, provide some of the best constraints on the nature of dark matter.
Table 4. ATLAST Science Flowdown Requirements for Dark Matter Kinematics
Science
Science Requirements
Measurements Needed
Design and Implementation
Question
Proper motions of ~200 stars
Aperture diameter: 8 m to achieve the
angular resolution to enable stellar
per galaxy with accuracies ~20
centroids to be determined to 0.1 mas.
as/yr at 50 kpc (yielding
Focal plane metrology must be
required transverse velocity
What are the Determine the mean
accuracy of 5 km/sec). Augment maintained to enable 0.005 pixel centroid
accuracy over a 5-year baseline.
kinematic
mass density profile of
with stellar radial velocities from
properties of high M/L dwarf
ground.
WFE: Diffraction limited at 0.5 m;
Dark Matter Spheroidal Galaxies
~1.3 to 1.6 mas pointing stability;
Need wide-FOV to ensure
symmetric PSF is essential.
sufficient number of background
VIS/NIR wide-field (5 - 8 arcmin FOV)
astrometric references (e.g.,
imager at TMA focus.
galaxies and QSO).

3. Technology Development
All ATLAST concepts require many of the same key technologies because they have several fundamental design
features in common. All ATLAST concepts are designed to operate in a halo orbit at the Sun-Earth L2 point. The optical
designs are diffraction limited at 0.5 m (36 nm RMS WFE) and the optical telescope assembly (OTA) operates near
room temperature (280 K 290 K). All OTAs employ two simultaneously usable foci: a three-mirror anastigmat (TMA)
channel for multiple, wide field of view instruments and a Cassegrain channel for high-throughput UV instruments and
instruments for imaging and spectroscopy of exoplanets. All designs have an RMS WFE of < 5 nm at < 2 arcseconds
radial offset from Cassegrain optical axis. See the ATLAST concept study8,9 and also Stahl et al., Oegerle et al., and
Hopkins et al. (these proceedings) for additional details.
3.1 Telescope Technology
The diffraction-limited imaging at 0.5 m that is needed for much of ATLAST science requires HST-quality mirror
surface errors (5-10 nm rms) to meet the overall system wavefront error of 36 nm rms. For the monolithic 8-m mirror
ATLAST concept, solid meniscus monolithic glass, as demonstrated on ground-based telescopes, requires no new
technology, but will require engineering to ensure survival of launch vibrations/acoustics and the proper gravity sag
unloading. For the segmented mirror ATLAST designs, hexagonal mirrors measuring 1.3 m and 2.4 m are baselined for
the 9.2 m and 16 m apertures, respectively. Some potential options for the mirror composition include the Advanced
Mirror System Demonstrator (AMSD) ULE glass segments developed for JWST, Actuated Hybrid Mirrors (AHM), and
corrugated glass technology. ATLAST-9.2m has baselined the AMSD-like mirror segments of 1.3m size and 25 kg/m2.
At this mass and size, mirror segments are sufficiently stiff such that demonstration of better than 10 nm rms wavefront
error and vibration and acoustic testing will be straightforward. An ATLAST technology development program would
downselect between AMSD-like and AHM mirror segment technologies. Assuming both technologies meet launch loads

Postmanetal.:ATLASTScienceDrivers
and figure error, the downslect will be based on studies of thermal stability, cost, and manufacturing schedule. The TRL
6 milestone is a mirror segment that meets environmental and wavefront error budget requirements.
Wavefront sensing and control (WFS&C) also needs to be advanced from JWST to meet visible image quality with an
allocation of 10-15 nm rms wavefront error for WFS&C residual. The segments positions must be measured to a few
nanometers at a rate faster than their support structures time constants. Periodic image-based WFS employs phase
retrieval to measure the wavefront error. For ATLAST-8m, this is done using WFS modules on each side of the three
foci, updated every few weeks using stars of opportunity. For the segmented versions, this is done using several WFS
modules that provide updates every few minutes (without interrupting science observations) using the guide stars
simultaneously for wavefront sensing and guiding. Technology development requires a demonstration of WFS
algorithms to the required performance within the processing capabilities of a flight computer with realistic guide star
scenes. Actuators supporting the segments require a modest technology development to reduce the resolution to 2 nm
and increase bandwidth. Since WFS&C is inherently a system-wide performance of the observatory, a demonstration of
the telescopes operation would be planned with a subscale (~6 m class), partially populated (three-segment) testbed
meeting error budgets under full thermal, vacuum, vibration, and jitter environments.
3.2 Detector Technology
Gigapixel detector arrays for visible imaging and ~500 Megapixel arrays for NIR imaging are required for studies of
resolved stellar populations, galaxy evolution, and structure formation. Such arrays can be built with existing
technology. However, development would result in better science performance (lower noise), lower risk (less complex
electronics), lower cost, and lower power consumption. The wide field cameras in all ATLAST designs are envisioned to
have ~1 gigapixel per channel. Exposure times invested in exoplanet and other faint object spectroscopy could be
reduced by up to five times using photon counting detectors. Technology development of photon-counting CCDs is
based on the low-light level CCDs built by E2V and the similar technology of Texas Instruments. The current TRL is 4;
improvements in anti-reflection coatings, voltage swing, and charge induced clock noise would be demonstrated before
the completion of a TRL 6 qualification program. In the longer term, CMOS based detectors are likely to be used.
UV detectors have ample room for improvements in efficiency and format size. UV spectroscopy, in particular, requires
coatings, optics and detectors that are highly efficient. Current flight UV imaging detectors use CsI and Cs 2Te
photocathodes with 10% - 30% quantum efficiencies (QE). Photocathodes with QE of 30% - 80% using cesiated pdoped GaN have been produced in the lab, but have not yet been integrated into large format detector systems suitable
for flight. Materials such as p-doped AlGaN and MgZnO should be developed for higher QE over a wider band, and
better AR-coatings may be matched with p-channel radiation-hardened photon-counting CCDs. For the far UV, higher
QEs also are required, and III/V semiconductor materials are effective. All must be coupled with large-format intensifier
and readout systems. EBCCDs and microchannel plate methods (ceramic and glass) compete for the highest QE and
largest formats, and both should be pursued.
3.3 Starlight Suppression Technology
An internal coronagraph or external occulter that provides 10-10 starlight suppression (output-to-input beam flux ratio) is
required to characterize Earth-sized exoplanets. The best starlight suppression technology is not yet obvious.
Fortunately, there are multiple options. The technology development plan addresses the viability of 1) a visible nulling
coronagraph (VNC) that any telescope could use (development costs might be shared with large ground-based
telescopes); 2) a Lyot-type coronagraph (for use with monolithic telescopes); and 3) an external starshade (a separate
spacecraft creating a star shadow at the telescope). High contrast Lyot coronagraphs39 are at TRL ~6 whereas the VNC
and starshade are currently at or below TRL 4.
There are a number of coronagraphic configurations, each having tradeoffs between achievable contrast, IWA,
throughput, aberration sensitivity, and ease-of-fabrication. Most techniques do not work on conventional (4-arm spider)
obscured or segmented systems, but some non-conventional spider configurations (linear support) appear to allow
performance competitive with off-axis systems. The VNC approach is a viable solution for internal suppression with a

10

Postmanetal.:ATLASTScienceDrivers
segmented telescope. VNC development is required to demonstrate 10-10 suppression ratios over at least a 23% passband.
This requires the development of a spatial filter array (1027 fiber bundle), deformable mirror (MEMS 1027 segment),
and an achromatic phase shifter.
Starshade theoretical performance has been validated by at least 4 independent algorithms and, in the lab, by two
beamline testbeds40. Detailed CAD models exist for the New Worlds Observer41 50-meter starshade, which use high
TRL components (membranes, hinges, latches, booms). For ATLAST, a larger starshade is required: 60 m to 80 m. The
key challenges are primarily deployment reliability and shape control. The ATLAST starshade separation of ~165,000
km, while large, does not present any challenging formation flying or orbital dynamics issues but put additional
requirements on the starshade propulsion system. The required starlight suppression performance with a segmented
telescope is achievable with an external starshade42. The starshade technology developments are addressed through
increasingly larger subscale models with TRL 6 being demonstrated through beamline tests, a half-scale quarter-section
deployment, and a full-scale single petal deployment, performance, and environmental testing.
3.4 Rocket and Propulsion Technologies
All ATLAST concepts we investigated rely on the availability of a heavy lift launch vehicle with a high-volume fairing.
The least massive of our ATLAST concepts (segmented 9.2m design) requires a payload-to-SEL2 capacity of ~16 metric
tons. We assumed a launch of ATLAST-9.2m from KSC on an enhanced version of the Delta IV-Heavy launch vehicle.
The enhancements, which are outlined in the Delta IV users guide, consist of fairing with a 6.5 m outer diameter, a
more powerful main engine (the RS68A which is already in testing) and the addition of 6 solid boosters already in use on
the smaller version of the Delta IV. Informal estimates indicate that this enhanced version could carry 18 metric tons of
payload mass to a C3 of about -0.69 km2/sec2, which is required for reaching SEL2. The most massive ATLAST concept
(monolithic 8m) requires a payload-to-SEL2 capacity of ~55 metric tons and an inner fairing diameter of 8.8 meters.
These requirements are consistent with the capabilities of the proposed Ares V launch vehicle (configuration LV
51.00.48). Cylindrical fairing segments of ~10 meters in height with an additional 10 12 meters of clearance over the
central payload region in a conical or ogive segment are also required for our current ATLAST designs.
NASAs Evolutionary Xenon Thruster (NEXT) ion engine will achieve TRL 6 or greater through NASAs in-space
propulsion programs. Such engines will be essential for efficiently enabling alignment of an external starshade with its
companion observatory.

4. Concluding Remarks
The most compelling science cases for a future large-aperture UVOIR space telescope require major increases in angular
resolution, sensitivity, and wavefront error stability over existing or planned facilities in the 0.11 m to 2.4 m
wavelength range. ATLAST will definitively establish the frequency of detectable life on terrestrial-like exoplanets.
ATLAST will revolutionize our understanding of star formation, galaxy evolution, and the nature of dark matter by its
ability to provide unique and essential data to complement that from a new generation of very large ground-based optical
and radio observatories anticipated to come on line in the 2010 2020 timeframe. Incremental advances in telescope
aperture are not compelling, so radical advance is needed. Fortunately, the technology needs of several communities are
aligning to make large space-based optical systems affordable. An affordable ATLAST observatory concept can satisfy
the challenging science-driven performance requirements discussed here in time for a 2025-2030 launch if progress is
made in maturing the key technologies needed.

REFERENCES
[1] Tokovinin, A. Seeing Improvement with Ground-Layer Adaptive Optics, PASP 116, 941951 (2004).
[2] Berkefeld, T., Glindemann, A., and Hippler, S., Multi-Conjugate Adaptive Optics with Two Deformable Mirrors Requirements and Performance, Exp. Astron. Vol. 11, 121 (2001).

11

Postmanetal.:ATLASTScienceDrivers
[3] Andersen, D., et al., Performance Modeling of a Wide-Field Ground-Layer Adaptive Optics System, PASP, 118,
1574-1590 (2006).
[4] Ammons, S.M., et al., First results from the UCSC Laboratory for Adaptive Optics multi-conjugate and multi-object
adaptive optics testbed, Proc. SPIE, 6272, (2006).
[5] Nicolle, M., Fusco, T., and Michau, V., Optics Letters, Vol. 29, Issue 23, pp. 2743-2745 (2004).
[6] Serabyn, G., Mawet, D., and Burruss, R., American Astronomical Society Meeting Abstracts, 216, #311.01 (2010).
[7] Guyon, O., et al., PASP, 122, 71, (2010).
[8] Postman, M., et al., Advanced Technology Large-Aperture Space Telescope (ATLAST): A Technology Roadmap for
the Next Decade, arXiv:0904.0941, (2009). Report also on line at http://www.stsci.edu/institute/atlast
[9] Postman et al., Science with an 8-meter to 16-meter optical/UV space telescope, Proc. SPIE, 7010, (2008).
[10] Oegerle et al., ATLAST-9.2: A Deployable Large Aperture UVOIR Space Telescope, Bulletin of the American
Astronomical Society, 41, 569 (2010).
[11] Stahl, H. P. et al., Design for an 8-meter monolithic UV/OIR space telescope, Proc. SPIE, 7436, (2009).
[12] Stahl, H. P., Survey of Cost Models for Space Telescopes, Optical Engineering, Vol.49, No.05, (2010).
[13] Stahl, H. P. and Hinrichs, T., Multi-Variable Cost Model for Space Telescopes, Proc. SPIE, 7731, (2010).
[14] Hickey, G. S., Lih, S.-S., Barbee, T. W., Jr., 2002, Proc. SPIE, 4849, 63, (2002).
[15] Strafford, D. N., DeSmitt, S. M., Kupinski, P. T., and Sebring, T. A., Proc. SPIE, 6273, 0R, (2006).
[16] Werner, M. et al., Active Space Telescope Systems A New Paradigm, Astro2010 White Paper, (2010).
[17] Kasting, J., Traub, W. et al., arXiv:0911.2936, (2009).
[18] Lunine, J. I., et al., arXiv:0808.2754, (2008).
[19] Levine, M., et al., The Astronomy and Astrophysics Decadal Survey, 2010, 37, (2009).
[20] Guyon, O., Ap.J., 629, 592, (2005).
[21] Mountain et al., arXiv:0909.4503 (2009).
[22] Des Marais, D. J., et al., Astrobiology, 2, 153, (2002).
[23] Woolf, N. J., Smith, P. S., Traub, W. A., and Jucks, K. W., Ap.J., 574, 430, (2002).
[24] Ford, E. B., Seager, S., and Turner, E. L., Scientific Frontiers in Research on Extrasolar Planets, 294, 639, (2003).
[25] Brown, T. et al., The History of Star Formation in Galaxies, Astro2010 Science White Paper, (2009).
[26] Oppenheimer, B. D., Dav, R., and Finlator, K., MNRAS, 396, 729, (2009).
[27] Birnboim, Y., and Deckel, A., MNRAS 345, 349, (2003).
[28] Keres, D., et al., MNRAS, 363, 2, (2003).
[29] Sembach, K. et al., The Cosmic Web, Astro2010 Science White Paper, (2010).
[30] Bertone, S., Schaye, J., and Dolag, K., Space Science Reviews, 134, 295, (2008).
[31] Giavalisco, M. et al., The Quest for a Physical Understanding of Galaxies Across Cosmic Time, Astro2010
Science White Paper, (2009).
[32] Sembach, K. et al., Technology Investments to Meet the Needs of Astronomy at Ultraviolet Wavelengths in the 21st
Century, Astro2010 Technology Development White Paper, (2009).
[33] Savage, B. D., et al., Bulletin of the American Astronomical Society, 41, 502, (2010).
[34] Martin, N. F., Ibata, R. A., Chapman, S. C., Irwin, M., and Lewis, G. F., MNRAS, 380, 281, (2007).
[35] Simon, J. D., and Geha, M., ApJ, 670, 313, (2007).
[36] Strigari, L. E. et al., Nature, 454, 1096, (2008).
[37] Strigari, L. E., Bullock, J. S., and Kaplinghat, M., Ap.J., 657, L1, (2007).
[38] Soto, M., Kuijken, K., and Rich, R. M., IAU Symposium, 245, 347, (2010).
[39] Trauger, J. and Traub, W., Nature, 446, 771, (2007).
[40] Leviton, D. B., et al., Proc. SPIE, 6687, (2007).
[41] Cash, W., et al., Proc. SPIE, 7436, (2009).
[42] Soummer, R. et al., A Starshade for JWST: Science Goals and Optimization, Proc. SPIE, 7440, 8, (2009).

12