The past few years have seen rapid advances in cosmology and exoplanet science. Ground- and space-based observations are continuing to advance our understanding of the birth, expansion, and growth of the universe. The combination of the Kepler mission with ground-based follow-up has made great strides in determining the frequency of Earth-like planets around nearby stars. In the sections below, the committee places the scientific rationale for Wide-Field Infrared Survey Telescope (WFIRST) in this context and assesses the WFIRST/Astrophysics Focused Telescope Assets (AFTA) implementation in terms of its ability to achieve the scientific goals set for WFIRST by 2010 decadal survey New Worlds, New Horizons in Astronomy and Astrophysics1 (NWNH).
As described in NWNH, the two decades from 1990 to 2010 saw a remarkable maturation of our understanding of the birth, expansion, and growth of the universe. A standard model was developed, beginning with an epoch of rapid acceleration driven by an unknown mechanism (inflation) when the seeds of contemporary large-scale structure were sown. This event was followed by a long interval of gravitational deceleration, eventually dominated by an unknown substance (dark
1 National Research Council, New Worlds, New Horizons in Astronomy and Astrophysics, The National Academies Press, Washington, D.C., 2010.
matter), which then transitioned to the current epoch—a second accelerating phase driven by another mechanism (dark energy).
The standard model has survived many challenges. Its defining parameters and consequences have either been detected for the first time or measured with much greater precision. These include the size of the observable universe, and its age, its near-zero spatial curvature, the relative contributions of its various components, gravitational lens distortion of galaxy images, correlated polarization and lens distortion of the microwave background radiation, and the initial fluctuation spectrum (which is almost scale-free, adiabatic, and consistent with a quasi-Gaussian random field). The sound waves that had been observed from the young universe in the light patterns of the cosmic microwave background (CMB) have been matched to features in the distribution of positions and speeds of galaxies (known as baryon acoustic oscillations, BAOs) seen in the older (more nearby) universe. Observations of highly elliptical cluster potentials and the bullet cluster have shown that dark matter is effectively collisionless. The best estimate for the maximum mass of a neutrino comes from cosmological studies rather than particle physics. Most significantly, dark energy has been shown to have properties similar to Einstein’s cosmological constant with an accuracy of about 10 percent. All of this has been accomplished using extremely careful observations obtained with the Wilkinson Microwave Anisotropy Probe, Planck, the Hubble Space Telescope (HST), the Chandra X-ray Observatory, and the Fermi Gamma-ray Space Telescope, as well as many ground-based telescopes, including the Sloan Digital Sky Survey, the South Pole Telescope, and the Atacama Cosmology Telescope.
Despite this remarkably rapid progress, there are now hints of tensions among different data sets that could be indicating that researchers are on the threshold of new discoveries. The way forward is to make more accurate measurements with yet more capable telescopes and to pursue several approaches in parallel in order to eliminate the systematic biases that otherwise bedevil observational cosmology. Meeting this objective is one of the three goals of WFIRST.
The primary goal of WFIRST for cosmology is to provide significant advances in understanding the nature of dark energy by providing precise astronomical measurements of the expansion history and growth of structure in the universe (see Figure 1.1) using multiple probes. WFIRST will also provide critical tests of the underlying theoretical framework, which is Einstein’s theory of gravity. If the growth and expansion measurements point to different properties of dark energy, then perhaps the theory of gravity, that is, Einstein’s theory of general relativity, breaks down on cosmological scales.
The WFIRST cosmological probes are as follows: (1) standard candles, discovering Type Ia supernovae and measuring precise distances; (2) standard rulers, measuring features in the distribution of galaxies imprinted from the early universe (BAOs); and (3) growth, measuring redshift space distortions in the distribution
FIGURE 1.1 Constraints on the dark energy equation-of-state parameter (w) and its rate of change as the universe expands (dw/da). The green ellipse represents the current constraints and is centered on the values w = −1 and dw/da = 0 specified by Einstein’s cosmological constant (Λ). The black ellipse (for an arbitrary choice of w and dw/da) shows the forecasted constraints for the final baseline WFIRST/AFTA survey data when combined with the other independent data that will be available at that time. The red ellipse shows the constraints possible if the precision of the WFIRST/AFTA measurements is increased by a factor of two compared to the baseline. Legends indicate different possible regimes: scalar field models that are “freezing towards” or “thawing from” w = −1, and models with w < −1 in which increasing acceleration leads to a future “big rip.” SOURCE: WFIRST-AFTA Science Definition Team and WFIRST Project, Wide-Field InfraRed Survey Telescope-Astrophysics Focused Telescope Assets: Final Report by the Science Definition Team (SDT) and WFIRST Project, May 23, 2013, http://wfirst.gsfc.nasa.gov/science/sdt_public/WFIRST-AFTA_SDT_Final_Report_Rev1_130523.pdf, p. 28.
of galaxies and measuring the bending of light caused by the largest structures in the universe.
WFIRST/AFTA provides advantages relative to the capabilities described in NWNH for all three cosmological probes. The gains are realized by a larger collecting area and higher angular resolution, as well as the ability to obtain redshifts of the distant lensed galaxies with the Integral Field Unit (IFU). The larger aperture and ability to make precise measurements over a wide redshift range will enable WFIRST/AFTA to place powerful constraints on theories of cosmic acceleration.
A number of new approaches from the ground are also pursuing the expansion history of the universe from the era dominated by dark energy to redshifts beyond 2. Proposed measurements by the Dark Energy Spectroscopic Instrument (DESI), the Hobby-Eberly Telescope Dark Energy Experiment (HETDEX), the Canadian Hydrogen Intensity Mapping Experiment (CHIME), and the Square Kilometer Array Phase 1 together will enable a more nuanced understanding of how the universe has expanded and structure has formed. The unique spectral coverage of WFIRST and low systematics inherent to space ensures that WFIRST observations will remain scientifically compelling and will complement the rapidly advancing ground-based observations.
The distance probe that first indicated the current acceleration of the expansion of the universe is the luminosity distance to distant Type Ia supernovae. WFIRST/AFTA will discover supernovae and measure their light curves and spectra. WFIRST/AFTA will obtain 2,700 spectra with a low-spectral-resolution IFU that has been added to the mission and was not planned for WFIRST/Interim Design Reference Mission (IDRM). This addition improves the ability to separate the light of the supernova spectrum from that of the host galaxy. Spectral features in the supernovae can be used to reduce the scatter of the redshift-magnitude diagram to 0.12 mag, implying distance accuracies per supernova of 6 percent. The IDRM would also detect supernovae, but the larger aperture of WFIRST/AFTA leads to greater sensitivity and the discovery of more supernovae overall, reaching more distant supernovae. The WFIRST/AFTA supernova program should provide distance accuracies better than 1 percent, in 15 redshift (z) bins (dz = 0.1) from z = 0.2 to z = 1.7. The Large Synoptic Survey Telescope (LSST), which will provide a major advance over programs like the Dark Energy Survey (DES), is projected to find 2 million Type Ia supernovae over 10 years. But LSST will only find supernovae with redshifts less than about 1, measure photometric redshifts, and measure rest-frame optical light for supernovae with a redshift 0.5. A supplementary ground-based spectroscopy program will be needed. The European Space Agency’s (ESA’s) Euclid mission has no plans to study supernovae.
The BAO measurements of WFIRST/AFTA will be performed using the grism mounted in the filter wheel. A 2,000-square-degree field will be surveyed to a depth that detects 11,600 galaxies per square degree. The IDRM would also measure the BAO signal using a parallel field of view covering a larger field, but with lower sensitivity leading to a lower density of sources. WFIRST/AFTA is slightly better than the IDRM for BAO overall, and it is much better for the high redshift range. The redshift accuracy of the major BAO high-latitude surveys planned for WFIRST/AFTA and WFIRST/IDRM missions is virtually identical; however, the AFTA BAO survey will achieve that accuracy for fainter galaxies (H-alpha line fluxes of 1.0 × 10−16 erg sec−1 cm−2 versus 1.5 × 10−16 erg sec−1 cm−2 for IDRM) over a smaller field of view (2,000 sq deg versus 2,700 sq deg for IDRM) and will achieve a H-α emitting galaxy density at z ~ 1.9, about 3 times higher than that anticipated for the IDRM. Euclid will also measure BAO in a very large 15,000 sq deg field, but with a shallower survey detecting only 1,700 galaxies per square degree. The primary difference between WFIRST/AFTA and Euclid is that WFIRST/AFTA is much better coverage at high redshifts. The accuracy of WFIRST/AFTA BAO data is expected to be 0.4 percent in angular diameter distance (DA) and 0.7 percent in the expansion (H(z)) for the z = 1-2 range. In principle, ground-based 21-cm observations, such as those planned for CHIME, could reach similar precision over this redshift range, although this technique remains to be proven.
Density peaks in the matter distribution will be surrounded by infalling matter, which drives the growth of large-scale structures. This inward flow reduces the redshift space difference corresponding to a given physical separation along the line of sight. Thus, the transverse axes and line-of-sight axis of the three-dimensional correlation function will be different by an amount that depends on the growth rate of the structures. This effect is called the redshift space distortion, because it makes the contours of the apparent three-dimensional redshift space correlation function oblate, even though the real-space correlation is symmetric, depending only on the spatial separation. Thus, a large-scale redshift survey of galaxies provides the growth rate over cosmic time, while a large-scale weak lensing survey provides the density contrast versus redshift. Redshift space distortions complement the weak lensing survey for structure growth measurements. Accuracy of 1 percent in measuring this effect is projected over the important redshift interval covered by the WFIRST/AFTA grism survey. Comparing the history of structure growth to the expansion history of the universe allows a test of general relativity over the largest possible scales.
The planned WFIRST/AFTA High Latitude Survey has higher sensitivity than WFIRST/IDRM over a smaller area (2,000 sq deg versus 2,700 sq deg for IDRM). The higher sensitivity leads to measurement of higher galaxy densities on the sky (effective source densities of 54, 61, and 44 galaxies arcmin2 in the J, H, and F184 filters, respectively), which in turn leads to a factor of 1.5 improvement in the figure of merit for cosmic shear dark energy measurements. Furthermore, the higher source density enables cross-checks among the three shear auto-correlation functions and the three cross-filter correlation functions, allowing detailed tests of the weak lensing analysis. Euclid will conduct a larger survey (15,000 sq deg) with lower sensitivity. It will result in more shape measurements and, therefore, higher statistical precision. WFIRST/AFTA is complementary in that it will have much tighter control and cross-checks of systematics, making it more likely than WFIRST/IDRM or Euclid to achieve statistics-limited results. In addition, WFIRST/AFTA will make shear measurements out to higher redshifts than Euclid.
The high source density provided by WFIRST/AFTA should lead to the detection of 40,000 galaxy clusters with mass greater than 1014 Msun over 2,000 sq deg. WFIRST/AFTA will also provide mean mass profiles calibrated by weak lensing for this sample, as well as an absolute mass calibration of clusters discovered by other methods, such as current and planned X-ray and CMB-SZ measurements, greatly increasing the precision of the cosmological results. Armed with precise absolute mass measurements, the measured cluster evolution extending to redshift 2 will provide a powerful independent cross-check of the redshift-distance relation and growth function.
WFIRST/AFTA, as a large-aperture near-infrared (NIR) space mission, will provide a unique combination of sensitivity and wavelength coverage that complements Euclid and LSST. Both Euclid and LSST will do very-wide-field surveys, while WFIRST/AFTA will do a much deeper survey in a smaller but still quite large field. Thus, WFIRST/AFTA will provide an accurate measurement of the universe at redshifts from 1 to 2, an important extension to the lower-redshift Euclid and LSST data sets. For supernovae, Euclid is not planning any measurements, but the combination of WFIRST/AFTA and LSST will provide a densely sampled data set out to z = 2. For BAO, LSST does not have spectroscopic capabilities, but the combination of Euclid and WFIRST/AFTA will provide densely sampled spectroscopic data from z = 0.3-2. The low redshift range requires the large survey area of Euclid to cover a significant volume, while the smaller solid angle but increased depth of
WFIRST/AFTA will provide a significant volume at high redshift. For weak lensing, both Euclid and LSST will be densely sampled out to z = 1, but WFIRST/AFTA will see a much higher density of z = 2 lensed galaxies, which are needed to measure the dark matter density contrast in gravitational lenses at z = 1-1.5.
Finding 1-1: WFIRST/AFTA observations will provide a very strong complement to the Euclid and LSST data sets.
Finding 1-2: For each of the cosmological probes described in NWNH, WFIRST/AFTA exceeds the goals set out in NWNH. These are the goals that led to the specifications of the WFIRST/IDRM (with 2.0 μm cut-off).
NWNH emphasized the breadth of discoveries and rapid progress in the study of extrasolar planets and identified a microlensing survey with WFIRST as an important element in a program to characterize the complete diversity of planetary system orbital elements. Microlensing from space is complementary to other methods that have been widely applied on the ground and in space. Quoting from the report, NWNH2 recommended the following:
[A] program to explore the diversity and properties of planetary systems around other stars, and to prepare for the long-term goal of discovering and investigating nearby, habitable planets (p. 191).
The ground-based radial velocity and transit surveys are most sensitive to large planets with small orbits, as is the Kepler satellite, although it should be capable of detecting Earth-size planets out to almost Earth-like orbits. Together these techniques will determine the probability of planets with certain orbital characteristics around different types of stars. To complete the planetary census, it will be necessary to use techniques that are sensitive to Earth-mass planets on large orbits. One such technique is called gravitational microlensing, whereby the presence of planets is inferred through the tiny deflections that they impose on passing light rays from background stars. A survey for such events is one of the two main tasks of the proposed WFIRST satellite (pp. 192-193).
Ground-based observations continue to demonstrate the power of microlensing with discoveries of individual systems, but a space-based survey is required for statistical precision. Kepler has also shown that 2 to 4 Earth-radius planets are common and that many of these planets have low density. This result is counter to the expectation of a strong dichotomy between rocky terrestrial planets and gas giants, an expectation based on the situation in our own solar system. The composition of
2 NRC, New Worlds, New Horizons in Astronomy and Astrophysics, 2010.
these low-density super-Earth planets has not been determined. Microlensing with WFIRST/AFTA will answer the question as to whether 2 to 4 Earth-radius planets are also common in the outer parts (r > 1 AU) of solar systems (see Figure 1.2). Some models of planet formation predict that there will be a large population of planets that have been ejected into interstellar space, which could only be studied using microlensing.
Finding 1-3: The importance of exploring the diversity of planetary systems in the parameter space probed by the WFIRST microlensing survey is as vital as it was in NWNH. No other current mission or technique can address this issue. Both WFIRST/IDRM and WFIRST/AFTA can carry out the envisioned survey.
The larger number of pixels and better spatial sampling with WFIRST/AFTA should provide some improvements for drift scan astrometry over the 20-40 microarcseconds achieved by Wide-Field Camera 3 (WFC3) on HST. This capability would address some of the astrometry-based exoplanet science envisioned by the NWNH panels and will also help to break microlensing degeneracies, either by measuring the lens-source angular separation well after the microlensing event or by measuring the time-varying centroid shift during the lensing event due to the multiple images produced by the lensing.
Finding 1-4: The WFIRST/AFTA telescope’s large number of pixels and better point spread function (PSF) sampling will allow astrometry derived from drift scanning to break degeneracies inherent in interpreting the microlensing data.
The recommended program for WFIRST in NWNH consists of three components. In addition to its dark energy and microlensing planet-finding programs, NWNH recommended a program that would address a broad range of general astrophysics. This science was to be accomplished through the parallel use of the data taken in the dark energy and microlensing surveys, through dedicated large surveys optimized for general astrophysics, and through a Guest Observer program analogous to those on NASA’s Great Observatories. This strategy recognized the enormous discovery power of a wide-field NIR telescope in space.
A major increase in the capability to undertake NIR observations in space occurred when the WFC3 was installed on HST in May 2009. Since then, the WFC3 infrared channel has accounted for nearly 25 percent of the observing cycles on HST and a correspondingly large fraction of HST’s scientific output. These observations comprise both large dedicated surveys and a wide range of Guest Observer
FIGURE 1.2 The colored, shaded regions show approximate regions of sensitivity for Kepler (red) and WFIRST/AFTA (blue) (note the regions do not delineate 100 percent completeness for all stars in either technique). The solar system planets are also shown, as well as the Moon, Ganymede, and Titan. Kepler is sensitive to the abundant, hot and warm terrestrial planets with separations less than about 1.5 AU. On the other hand, WFIRST/AFTA is sensitive to Earth-mass planets with separations greater than 1 AU, as well as planets down to roughly twice the mass of the Moon at slightly larger separations. WFIRST/AFTA is also sensitive to unbound planets with masses as low as Mars. The small red points show candidate planets from Kepler, whereas the small blue points show simulated detections by WFIRST/AFTA. Based on assumptions of mass and semi-major axis distribution functions informed by dynamical simulations (see Appendix D of the WFIRST/AFTA SDT report), the number of such discoveries will be large with roughly 2,800 bound planet discoveries. Thus, WFIRST/AFTA and Kepler complement each other, and together they cover a large planet discovery space in mass and orbital separation, providing the understanding of exoplanet demographics necessary to constrain the formation and evolution of planetary systems. Furthermore, the large area of WFIRST/AFTA discovery space, combined with the large number of detections, essentially guarantees a number of unexpected and surprising discoveries. SOURCE: Caption and image from WFIRST-AFTA Science Definition Team and WFIRST Project, Wide-Field InfraRed Survey Telescope-Astrophysics Focused Telescope Assets: Final Report by the Science Definition Team (SDT) and WFIRST Project, May 23, 2013, http://wfirst.gsfc.nasa.gov/science/sdt_public/WFIRST-AFTA_SDT_Final_Report_Rev1_130523.pdf, p. 44.
programs. Given that WFIRST/IDRM would represent a further increase of about a factor of 50 in NIR survey speed relative to WFC3, there is little doubt about its impact in general astrophysics. Much of this would take place through mining the data archive created by the cosmology and exoplanet studies. However, the telescope is ideally suited for targeted use in quite different investigations.
In the limited space available, the committee cannot describe all of the major science areas that WFIRST would strongly advance. NWNH specifically called out surveys of the Milky Way and nearby galaxies. The astrometric capabilities of WFIRST/AFTA would complement those of the ESA Gaia mission and could touch on science ranging from galactic structure to Kuiper Belt objects. Here, the committee focuses on one of the three areas that NWNH recommended as a focused science objective, namely, “Cosmic Dawn: Searching for the First Stars, Galaxies, and Black Holes.” This area has progressed very rapidly since NWNH, driven largely by deep-NIR surveys of the early universe using the WFC3 on HST, supported by observations with large ground-based telescopes and multi-waveband telescopes in space. This work not only has pushed out the frontier to redshifts beyond 7, but also has established the basic properties of the population of galaxies in place when the universe was only about a gigayear old (near the end of the epoch of reionization). Getting to this point has required a huge investment in telescopic resources. The next major step would be to use the enormous increase in the field of view provided by WFIRST to discover and characterize the rare population of objects at still higher redshifts.
The committee notes that the development of the ESA-led Euclid space mission and the welcome progress in the National Science Foundation (NSF)/Department of Energy (DOE) LSST (the highest ranked, large, ground-based facility in NWNH) make the general astrophysics case of WFIRST even more compelling. The LSST will conduct wide-field surveys entirely in the visible wavebands, and Euclid devotes nearly 90 percent of its field of view to imaging in the visible band. WFIRST’s unmatched capabilities in the NIR nicely complement these facilities (see Figure 1.3). As envisaged in NWNH, there would be a significant overlap in the operations phases of WFIRST and the James Webb Space Telescope (JWST) with the potential for strong synergy between the survey speed of WFIRST and the capabilities of JWST for detailed follow-up. Analogous synergies exist between WFIRST and both the next generation of ground-based optical and infrared Giant Segmented Mirror Telescopes and the Atacama Large Millimeter Array (ALMA). These synergies extend across a broad range of science, but are particularly strong in the Cosmic Dawn program described above.
WFIRST/AFTA has a primary mirror with about a factor of 2.5 times more collecting area, more than twice as many NIR pixels, and a spatial resolution about 60 percent better than IDRM. This not only increases the survey depth, but it also substantially improves the ability to characterize the structures of detected objects.
FIGURE 1.3 The depth of the WFIRST/AFTA high latitude survey (2,000 square degrees, labeled WFIRST) is compared to the depths of the Large Synoptic Survey Telescope (LSST) and Euclid imaging surveys. The complementarity between the near-infrared survey by WFIRST/AFTA and the optical surveys by LSST and Euclid is clear. The numbers below each color bar indicate the size of the point spread function in units of 0.01 arcsec. SOURCE: WFIRST-AFTA Science Definition Team and WFIRST Project, Wide-Field InfraRed Survey Telescope-Astrophysics Focused Telescope Assets: Final Report by the Science Definition Team (SDT) and WFIRST Project, May 23, 2013, http://wfirst.gsfc.nasa.gov/science/sdt_public/WFIRST-AFTA_SDT_Final_Report_Rev1_130523.pdf, p. 22.
These benefits are generic for a general astrophysics program. The very steep faint end slope of the galaxy luminosity function at these high redshifts means that AFTA’s additional depth of one magnitude relative to the IDRM will lead to a roughly order-of-magnitude increase in the number of very-high-redshift (z > 8) galaxies that can be detected and characterized.
These gains are less clear in the F184 band owing to the warmer operating temperature planned for the AFTA primary mirror (270 K). If operated even warmer, then science longward of the H-band would be seriously compromised. More specifically, thermal emission from warm optics at the long end of the grism passband will degrade the grism sensitivity at all wavelengths, so the red edge of the passband will have to move shortward if the telescope runs warmer, which loses the highest-z
end of survey science. At an operating temperature of 270 K, the AFTA telescope still outperforms the IDRM by a factor 1.66 in the longest wavelength filter.
NWNH did not specify a detailed time allocation among the primary science goals, but in the first 5 years of the mission, somewhat more than 2 years were thought to be required for the cosmic-acceleration program, with somewhat over a year devoted to the microlensing survey. The planned WFIRST/AFTA observing program is presented in Table 1-2 of the Science Definition Team’s report.3 Of the 5-year mission (assuming no coronagraph), 6 × 72 days would be devoted to microlensing studies, 1.3 years to high-latitude survey imaging, 0.6 years to high-latitude spectroscopy, and 0.5 years to the supernova survey, including spectroscopic follow-up. The remaining 1.4 years of the 5-year prime mission would be devoted to Guest Observer programs.
Finding 1-5: The observing program envisioned for WFIRST/AFTA is consistent with the science program described for WFIRST in NWNH.
If the coronagraph is added, then the prime mission duration will be increased to 6 years to accommodate its use.
In developing its recommendations, NWNH recognized both the central importance of and the extraordinary pace of discovery in exoplanet science. In addition to its recommendation of a WFIRST microlensing program, it therefore proposed that NASA take the necessary steps to prepare for a future planet imaging mission. Specifically, NWNH recommended the following:
NASA and NSF should support an aggressive program of ground-based high-precision radial velocity surveys of nearby stars to identify potential candidates. In the first part of the decade NASA should support competed technology development to advance multiple possible technologies for a next-decade planet imager, and should accelerate measurements of exozodiacal light levels that will determine the size and complexity of such missions. The committee recommends an initial NASA funding level of $4 million per year so as to achieve a clear set of design requirements and technology gateways to be passed. If, by mid-decade, a [decadal survey advisory
3 WFIRST-AFTA Science Definition Team and WFIRST Project, Wide-Field InfraRed Survey Telescope Astrophysics Focused Telescope Assets: Final Report by the Science Definition Team (SDT) and WFIRST Project, May 23, 2013, http://wfirst.gsfc.nasa.gov/science/sdt_public/WFIRST-AFTA_SDT_Final_Report_Rev1_130523.pdf.
committee] review determines that sufficient information has become or is becoming available on key issues such as planet frequency and exozodiacal dust distribution, a technology down-select should be made and the level of support increased to enable a mission capable of studying nearby Earth-like planets to be mature for consideration by the 2020 decadal survey, with a view to a start early in the 2020 decade. The committee estimates that an additional $100 million will be required for the mission-specific development (p. 20).
Since the writing of NWNH, there have been scientific and programmatic changes affecting this recommendation. On the programmatic side, cost increases in JWST, coupled with reductions to NASA’s budget, have delayed the implementation of WFIRST, and so the start of a planet-imaging mission early in 2020 is highly unlikely. Nevertheless, the high scientific priority of such a mission is unaffected, even if the timeframe is likely to shift.
On the scientific side, significant progress has been made in determining the frequency of rocky planets. Results from Kepler show that small planets are very common, at least at small orbital radii, retiring much of this risk. As discussed earlier in this report, the WFIRST microlensing survey will also extend our knowledge to larger orbital distances.
The exozodiacal light levels are, however, still very poorly constrained. Bright dust distributions around nearby stars have been imaged in reflected light using coronagraphs on HST and from ground-based telescopes. The luminosity of the dust ring shown in Figure 1.4 is more than 10,000 times that of the solar system’s zodiacal dust. Detection of exozodiacal dust distributions as optically thin as the solar system’s are not yet possible with any technique around any type of star. Investments have been made in ground-based exozodiacal dust measurements using the Keck Interferometer Nuller and the Large Binocular Telescope Interferometer (LBTI), and these promise future improvements. However, the Keck Nuller and LBTI are unresolved detections and, therefore, do not reveal whether there is azimuthal structure in the disk, which is important to determine for planet detection. In addition, the LBTI measurements are in the thermal infrared, and there is uncertainty in extrapolating to the visible. NWNH, therefore, recognized that space-based exozodiacal light measurements may also be important and suggested that they might be possible on an Explorer-scale mission.
Finding 1-6: Budget constraints will slip the start of an Earth-like-planet imaging mission beyond the horizon envisioned by NWNH; however, developing the technologies for such a mission and addressing the key uncertainties, such as the levels of exozodiacal light and identifying targets, remains a high priority.
The goal of the WFIRST/AFTA coronagraph is to achieve high contrast with the stated primary objective of demonstrating some of the key technologies re-
FIGURE 1.4 HR 4796A dust ring imaged by Gemini Planet Imager. Ground-based adaptive optics is limited to operation at longer wavelengths (typically 1.6 micron) than WFIRST/AFTA and cannot achieve the same level of precision due to the fast fluctuation timescale of the atmosphere. The left panel shows total intensity (not point spread function [PSF] subtracted). The right panel shows polarized intensity, with PSF self-subtraction by simultaneous differential polarimetry. SOURCE: Courtesy of Gemini Observatory/AURA/Marshall Perrin.
quired for coronagraphs being considered for future exoplanet imaging missions. The details of the WFIRST/AFTA coronagraph design are immature at this stage. Two options, a primary and a backup, were selected through an open proposal call. Both are being studied and will be taken to technology readiness level (TRL) 5. The primary option for the optical design is an Occulting Mask Coronagraph (OMC) that has two options for masks. The “safe” mask option uses a Shaped Pupil (SP). A higher-performance but riskier Hybrid Lyot Coronagraph (HLC) mask is also being studied. A back-up optical design using a Phase Induced Amplitude Apodization (PIAA) coronagraph is also under consideration and would provide higher throughput. Both optical designs (OMC and PIAA) involve a coronagraphic system utilizing wavefront and jitter control independent of the telescope itself. This systems approach to a coronagraphic instrument is more sophisticated than the approach employed on HST or for those planned for JWST, which accept the wavefront and tracking provided by the telescope. The WFIRST/AFTA coronagraph will also demonstrate the use of a deformable mirror to remove telescope aberrations and a tip-tilt system to remove telescope jitter.
The performance of the complete system in space ultimately depends on the
final design choice, the degree to which the spacecraft jitter can be removed by the tip-tilt system, and the level at which the PSF can be subtracted. Estimates for the contrast at 200 milliarcseconds from a star range from somewhat better than 10−10 for the HLC, with the more aggressive jitter assumptions, to 5 × 10−9 for the SP, which has a sensitivity essentially independent of the jitter (see Figure 1.5).
The coronagraph designs being considered will all demonstrate techniques that are more advanced than what have been used to date in space and that are very likely to be applicable to any future planet-imaging mission that employs a coronagraph. These include the high-order wavefront control, the low-order wavefront sensing and correction of telescope jitter and thermal drift, and the high-throughput spectrograph. The committee was not charged with evaluating whether advancing these techniques would be achieved better and more cost effectively in a laboratory development or by the coronagraph on WFIRST/AFTA. Independent of this question, the planned coronagraph options will certainly advance broadly applicable techniques.
Finding 1-7: The WFIRST/AFTA coronagraph satisfies some aspects of the broader exoplanet technology program recommended by NWNH by developing and demonstrating advanced coronagraph starlight suppression techniques in space.
If the contrast levels projected for the HLC are achieved, then the coronagraph will measure exozodiacal dust emission at a level equal to that of the solar system’s (the luminosity of zodiacal dust in the solar system is 10−7 of the solar luminosity, defining 1 Zodi). It could be used to observe several dozen nearby stars, constraining the statistical distribution of dust brightnesses and the range of the dust spatial structure. The data would also yield constraints on dust composition and grain sizes. This would be a major advance over what is currently possible or projected for future ground-based measurements. The Keck Nuller and LBTI are unresolved detections and, therefore, do not reveal whether there is azimuthal structure in the disk, which is important to determine for planet detection. The AFTA coronagraph would make images in scattered light. If the WFIRST/AFTA coronagraph reaches the levels projected in Figure 1.5 (i.e., the HLC is chosen and the jitter and point spread function), then the improvement over LBTI would be significant.
Finding 1-8: Whether the WFIRST/AFTA coronagraph satisfies the NWNH goal to establish exozodiacal light levels at a precision required to plan an Earth-like-exoplanet imaging mission is uncertain due to the immaturity of the coronagraph design and uncertainty in the ultimate performance.
In addition to addressing some aspects of the NWNH medium-scale exoplanet technology development program, the AFTA coronagraph has the potential to advance general exoplanet science objectives. How significant this advance is, again,
FIGURE 1.5 Shown are contrast curves (detection limits) for the possible coronagraph configurations selected for WFIRST/AFTA: Shaped Pupil (SP), Hybrid Lyot Coronagraph (HLC), and Phased Induced Amplitude Apodization (PIAA). The solid lines indicate the systematic detection floor for the SP (black), HLC (blue), and PIAA (red), assuming 0.2 mas jitter after internal coronagraph correction and post-processing with point spread function (PSF) subtraction to an accuracy of 1/30 of the local PSF intensity. Points indicate the expected contrast for known radial velocity (RV) exoplanets that can be detected with a 5-sigma signal-to-noise ratio (SNR) with 10 percent bandwidth at 550 nm in less than 10 days of integration. The color and shape of the points indicate that a given planet can be detected with each of the coronagraph modes: SP (black), HLC (blue), PIAA (red). The median integration time required to reach SNR = 5 for the detected planets is 0.5 days for HLC and 0.2 days for PIAA. The dotted black line shows the intensity of an Edgeworth-Kuiper Belt (EKB) dust dusk with 100 times the brightness of Earth’s as viewed from a distance of 10 pc. The short dashed and long dashed lines show, respectively, the contrast per resolution element of an exozodiacal dust disk with intensity 10 times the solar system viewed from 10 pc and 2 times the solar system viewed from 5 pc. SOURCE: Courtesy of NASA/JPL/Caltech.
depends on the very uncertain coronagraph performance. In the best case, with 6 months of dedicated observations, it could be capable of detecting ~20 known radial velocity planets, including obtaining color information and R ~ 70 spectra of the brightest candidates. In the worst case, 6 months of observation could yield 4 to 5 detections. These estimates assume a total line-of-sight background brightness equivalent to 3 Zodis. The imagery would complement the system architecture information gleaned from the microlensing studies. The Kepler results show that large (in radius) super-Earth planets are common in the inner parts of solar systems. Because these lower-mass planets have not been detected in present radial velocity surveys, if these planets are also common in the outer parts of solar systems, then the number of planets detected with the AFTA coronagraph could be much higher. These planets will include some significantly farther from their parent stars than the exoplanets with transit spectra, and so the surfaces and atmospheres of these planets will not have been as strongly modified by stellar irradiation and will extend the observed exoplanet parameter space. At the current time, the sensitivity levels and observing program for the coronagraph are very uncertain, so it is not clear how much of this added science could be achieved in the 1 year allocated to coronagraph observations.