Adaptive Optics at the Keck Observatory: Goals and Near-term Priorities AO Working Group Report; 13 March 2003 1. Overview The Keck Adaptive Optics Working Group (AOWG) recommends an aggressive path forward to maintain the leadership of the Keck Observatory at the forefront of astronomical adaptive optics. Our recommendations have three thrusts: (1) improved performance of the present AO systems, (2) completion of the laser guide star system and associated instrumentation, and (3) next generation, advanced AO systems. Our primary recommendation remains the development of a next-generation, high- performance, general-use AO system, which would deliver stable, high Strehl ratio, infrared images in moderate field-of-view areas throughout the sky. This system, named Keck Precision Adaptive Optics (KPAO), would allow Keck to take the lead in general- use high-precision adaptive optics where many of the scientific advances in high- resolution imaging are expected. The development of a precision AO system at Keck is predicated on the assumption that the current AO system will be operating at full ability, including the deployment of a laser guide star and the integral field spectrograph OSIRIS. In several cases, significant work still needs to be carried out to achieve this beginning point for future systems. Most critically, before any serious work on new AO systems begin, the under-performance of the current AO system needs to be understood and fixed. Work is currently underway which could answer these questions. This work needs to be given extremely high priority and support. Near-term opportunities also exist for performing important clearly needed upgrades to the current system. One such upgrade presently identified is to replace the first- generation wavefront control system with a new state-of-the-art system including new lower noise cameras, new tip/tilt sensors, and new real-time processors allowing critically needed real-time diagnostics and telemetry. Implementation of these types of important upgrades will be critical to keeping the current AO system on the cutting edge while work begins on the next generation of systems; in addition with careful planning many of these improvements should be directly applicable to the development of the next generation systems. We also continue to strongly advocate the support of the external development of a powerful adaptive optics systems designed solely for high contrast observations of the brightest stars. While limited in its uses, such a system would allow the observatory to command an undisputable lead in the extremely important field of direct detection of extra-solar planets. Owing to the relatively small investment for such a high scientific payback, the AOWG strongly recommends that the observatory pursue this opportunity. For FY04, we have worked with Peter Wizinowich, David LeMignant and the Keck AO team to define a set of priorities that will maximize science capabilities in the near-term and that will work towards the long term goal of an advanced system. These priorities are: 1. AO operations support. 2. K2 LGS AO facility available for science operation. 3. Performance characterization and optimization for the existing AO systems. (including an upgrade to PCS) 4. OSIRIS implementation with K2 AO. 5. Interferometer-required AO modifications (Interferometer-funded). 6. New wavefront controller implementation (externally-funded). 7. KPAO studies. 8. XAOPI interface design support and design review (externally-funded). 9. NSF-funded laser progress monitoring. The next two sections discuss the next generation adaptive optics systems: KPAO (Keck Precision Adaptive Optics), and XAOPI (Extreme AO) along with the most important preliminary tasks that need to be taken. 2. Keck Precision Adaptive Optics (KPAO) The AOWG strategic plan of November 2002 recommended the initiation of the development stages of a new high-precision adaptive optics system, recently dubbed the Keck Precision Adaptive Optics (KPAO). Such a system is envisioned to be one which would deliver extremely high infrared Strehl ratios, high stability, near-complete sky coverage, a good knowledge of the delivered PSF, and the ability to observe at high resolution into the visible (albeit with lower Strehl ratios). The AOWG believes that the largest opportunity in a future AO system at Keck will be to overcome the single most important set of limitations in all current AO systems: imperfect Strehl, PSF variability, and a lack of concrete PSF knowledge (note that we are implicitly assuming that the current limitation on sky coverage will be solved in the short term with laser guide stars thus we do not list that here). Quantitative measurements of many types currently suffer in the face of these problems: large wings to the PSF make precise spatially resolved spectroscopy unreliable and searches for faint companions more difficult, PSF variability and lack of PSF knowledge make deconvolution of sources suspect, and imperfect Strehl ratios make precision photometry impossible. The current Keck AO observers, and those who will become AO users with the advent of the laser guide star system, have learned (or will learn) to work around these limitations and find the scientific problems that are best addressed within these constraints, but the full power of high resolution observations at the diffraction limit of the Keck telescopes will not be met until these limitations can be largely overcome. 2.1 Summary Scientific Justification The development of a high precision adaptive optics system would enable currently impossible observations in all fields of astronomy. One of our strongest recommendations is that in the FY04 timescale a project scientist for KPAO should be selected to work with the AO team to ensure that the maximum scientific capabilities are achieved. We anticipate that the development of a thorough science case will be one of the project scientist’s first roles, but here we summarize a few important scientific examples in the areas of solar system, galactic, and extragalactic astronomy. 2.1.1 Solar system astronomy: global change on Pluto Much of the AO work in the solar system has been done on the large satellites and outer planets which are in the 1-4 arcsecond size range and thus easily resolved with current AO. Results have been generally spectacular, including the discovery of clouds and the highest-resolution ever maps of Titan, volcanic outbursts on Io, and atmospheric structure on Neptune. Another important class of solar system objects, the smaller icy planets and satellites, are just out of range of current AO systems. With sizes of a couple of tenths of an arcsecond or smaller, these objects are only marginally resolved and with low and variable Strehl no reliable results on surface structures, shapes, and compositional gradients can be obtained. As an example, one of the current outstanding questions on the climate of Pluto – when and if the nitrogen atmosphere is going to freeze out – would best be answered by even a crude compositional map of the surface of Pluto determining the amount, state, and location of nitrogen, methane, and water ices on the surface. Pluto subtends only 0.1 arcseconds, so current AO systems, while able to show that Pluto is resolved, are incapable of reliably mapping spatial resolved spectra across the surface. With high-resolution stable PSFs (coupled with planetary rotation), however, the problem becomes tractable; HST UV observations have allowed the construction of a crude map which shows that Pluto has more contrast than almost any other body in the solar system, but we have no current idea what the extremely bright and extremely dark areas on the surface are composed of and how these change with Pluto’s seasons. Similar types of research on other solar system bodies will be made possible by such a system. 2.1.2 Galactic astronomy: low mass stars in binaries and clusters Moderate Strehl AO systems are particularly good at resolving pairs of point sources, so much of the galactic AO work has focused on companion searches. The current limitation to much of this work is the contrast achievable by the current AO system, which roughly scales as Strehl/(1-Strehl). The development of the XAOPI system will address the need for an extremely high contrast instrument to be used on the brightest targets, but many interesting targets will be too faint for this NGS system even in its faint target mode. For example, studies of large numbers of stars in clusters of various ages and studies of companions to low-mass stars will be difficult with the modest Strehls expected from the current AO system with LGS. XAOPI is expected to allow ultra-high contrast observations of extremely bright targets, but only the creation of an all sky high Strehl system will allow high contrast observations of any target in the sky. One well known limitation of moderate and variable Strehl is the inability to perform precise photometry. Infrared surveys of clusters are finding increasing numbers of free floating planetary-mass objects, and observations at the limits of adaptive optics promise to be able to answer questions of whether or not these objects extend to Jupiter masses and below. Unfortunately, the observations require the types of precise photometry which are impossible with moderate Strehl systems. A next generation stable high Strehl system would be the first to allow the detection and photometric measurement of these faint objects. Precision adaptive optics on faint stellar populations would be possible at last. 2.1.3 Extragalactic astronomy: star formation in quasar host galaxies The deployment of the laser guide star should usher in a new era of extragalactic AO studies, but these studies will face the same or worse Strehl and stability limitations already faced by NGS observers. The development of a new generation adaptive optics system would lead to the ability to perform the first precision adaptive optics work on extragalactic sources. As an example, current AO work on the host galaxies of quasars has already shown the presence of young stellar populations and suggested that mergers may be playing a crucial role. One serious limitation of this work is precisely and believably separating the quasar from the host galaxy. With moderate and variably Strehl the problem is largely intractable except in a qualitative way. High stable Strehl ratios achieved by a next generation Keck AO system would allow the first true images of quasar host galaxies, with which one could look for smaller scale features such as dust lanes and star forming knots to which previous studies were not sensitive. By resolving such structures we can also separate the older, more distributed, stellar population. This is crucial since even a relatively small number of new stars will wipe out the signature of the presence of a large older population in the optical and UV, where HST work has been done. 2.2 Initial steps At a minimum, precision adaptive optics will require reduction of the final delivered wavefront error to a suitably small value to increase the Strehl ratio. For example, a goal of a minimum Strehl ratio of 90% for K-band observations would require a final wavefront error of 100 nm or less. Understanding how to achieve wavefront errors this low -- or indeed if it is even a possibility -- will require a significantly increased knowledge of the error budget of the telescope, the structure of turbulence in the atmosphere above the observatory, and the performance of the current AO system. Acquisition of this knowledge is a necessary first step before any serious design considerations should begin. 2.2.1 Telescope Error Budget Precision adaptive optics requires a precise telescope. The individual factors affecting the wavefront delivered by the telescope are currently not well characterized. The predicted aberrations used for the design of the current AO system are in many cases no more than educated guesses; some of these guesses are now suspected to be quite far from reality. A high priority therefore needs to be placed on developing the tools to measure telescope aberration sources such as warping harness residuals, segment phasing and stacking errors, ACS noise, and segment vibration. We strongly recommend the immediate start to the PCS upgrade proposed by Chanan as one part of this effort. Although the new PCS system will not provide a new capability, it will set the stage for future upgrades and ensures that we do not lose a key capability. Once the sources of aberration are known, it is entirely possible that a significant amount of the spending for the development of a precision AO instrument will actually be in developing methods of determining and correcting telescope aberrations. 2.2.2 Atmospheric characterization One limitation to the wavefront error possible for an advanced laser guide star based AO system will be the inherent focal anisoplanatism (FA) imposed by the atmosphere. The magnitude of this effect depends directly on the height of typical turbulent layers in the atmosphere and can only be made smaller with multiple laser guide stars. The number of lasers required will depend directly on the magnitude of the FA at Mauna Kea, which is currently unknown. Current estimates suggest that a best value might be ~170 nm, implying that as many as 9 lasers would be required to bring the FA down to levels necessary for a 100 nm system. It is critical to know the true value and range of values, however, as the required number of lasers goes as approximately the square of the FA and it is possible that this number could be incorrect by a factor of 2. The AOWG therefore urges that a high priority be placed on working with the ongoing campaigns to characterize the Mauna Kea atmosphere and measure typical C n2 profiles. Ideally, data would be collected over several years before final decisions would be made on critical parts of the design like the total numbers of lasers. 2.2.3 Current AO system characterization Before design of any next generation AO system commences, it is critical to fully understand the reasons for the underperformance of the current system. As discussed at the beginning of this document, continued characterization and optimization of the current AO system remains a top priority of the AOWG. 3. Extreme Adaptive Optics Planet Imager (XAOPI) For the first time in history, technology has permitted astronomers to image the environments of nearby stars on scales comparable to the size of our solar system. Exploitation of these capabilities has led to the discovery of new classes of objects including dusty circumstellar debris disks and brown dwarfs. These discoveries have galvanized intense public interest in science and technology and have led to profound new insights into the formation and evolution of planetary systems. It is now within our grasp to develop an AO system with the potential to make the first direct detection of the light from planets in orbit around solar-type stars. XAOPI, the eXtreme Adaptive Optics Planet Imager, will be a 3000-actuator AO system optimized for ultra-high-contrast imaging around bright (mR<7-10) stars Figure 1: Simulated XAOPI 15 minute integration on a solar-type star (centered behind the occulting spot) showing a 8 Jupiter-mass planet. The characteristic square null, 1.6” on a side, in scattered light from the primary star is caused by XAOPI’s spatially-filtered wavefront sensor. Knowledge of brown dwarfs and extra-solar planets is growing rapidly. Since 1995, the first sub-stellar objects have been discovered using low order AO coronagraph (Nakajima et al. 1995), precision Doppler measurements (Mayor & Queloz 1995) and field surveys (Kirkpatrick et al. 1997; Strauss et al. 1999). Doppler searches have succeeded spectacularly; the technique is mature with more than seventy known extra-solar planets. The 2MASS, DENIS and Sloan field surveys have been fruitful for finding faint brown dwarfs; many examples are now known. The latter two techniques are handicapped by strong biases: Doppler searches will continue to find planets in close orbits with short periods (<10 Yr), while field surveys only find massive, isolated objects. None of the planetary companions detected so far resembles the giants of the solar system. This fact is, however, consistent with the notion that our solar system is a typical planetary system. Selection effects can easily account for the lack of detections of Jupiter-like planets - exoplanets detected to date cannot resemble the planets of our solar system because the Doppler technique used to detect them is not sufficiently sensitive and has not been in use long enough to detect a Jupiter-mass planet in a 5 AU orbit. If the sun were a target star no definitive Doppler signature would yet have been detected. The detected exoplanets may well be the observable 5% tail of the main concentration of massive planets of which Jupiter is typical. Moreover, Doppler searches are indirect: they can neither image nor study the physics, chemistry or thermodynamics of the objects they discover. XAOPI is a proposed AO system which will be the next step toward the detailed study of extra-solar planets, brown dwarfs and circumstellar disks. In this high- contrast regime, sensitivity scales roughly with Strehl ratio S and telescope diameter D as SD2/(1-S), giving an enormous advantage to the S=0.95 XAOPI system over conventional AO or even KPAO. In addition, high-contrast imaging is often limited by small residual static phase errors. XAOPI will be designed from the start to minimize these static errors and to use a specially-designed wavefront sensor to precisely control phase errors, even those caused by discontinuities in the primary mirror, out to its spatial frequency cutoff. Simulations indicate that XAOPI will achieve contrast levels of 10 7-108 and be able to detect a significant population of extrasolar planets through their near-IR emission. Once a planet has been detected it can be characterized through IFU spectroscopy, allowing measurements of temperature and composition and inferences about formation history. Such an instrument will allow Keck to dominate the field of direct detection and characterization of extrasolar planets as it has so far dominated radial-velocity detection. XAOPI will also be capable of near-IR imaging of circumstellar dust around a large population of stars, imaging of the environments of evolved stars, and ultra-high-resolution imaging at high Strehl ratios of bright solar-system targets. Figure 2: Monte Carlo simulation (Graham et al 2003) of extrasolar planet detection. Stars indicate actual planets discovered by radial-velocity techniques; small dots indicate planets discovered by a simulated astrometric survey using the Keck outriggers; and filled circles indicate planets that would be found by a XAOPI survey of field stars.