A 30-m GSMT will have nine times the light gathering power of the Keck telescopes. But as we look at fainter and more distant objects, they become more crowded on the sky and are increasingly affected by background illumination in the form of scattered or emitted light.
In order to fully realize the promise of the larger aperture, it is important to enhance the spatial resolution as well. With modern telescope fabrication, we have reached the limit to the possibility of improving resolution with more perfect optics. For further improvements, we have to correct the effects of the atmosphere in one of two ways: put our telescope in space or use Adaptive Optics (AO). The former option is not the clear winner. To require the telescope to be in space would exert downward pressure on the size and lead to lower resolution due to diffraction.
As the aperture grows, the diffraction limit on the image size decreases, making the potential gains with AO even greater. Higher spatial resolution produces greater image contrast, which serves two somewhat different goals: (1) distinguishing discrete objects in confused fields; and (2) distinguishing faint compact objects against a background of diffuse emission from the sky or from neighboring objects.
|Bkgnd||Diffraction-limited pixel||Seeing-limited pixel|
|Table 1 Faint limits for large telescopes, with and without AO.|
These advantages increase as the telescope grows in aperture. The first point is familiar from recent results with AO on ground-based telescopes of the 4-m and 8-m class. To illustrate the second point, Table 1 compares the faint limit of recent generations of telescopes as a function of wavelength. The table compares in a uniform way the faint limit for a point source against atmospheric background, with and without AO. It shows the stellar magnitude of a source detected with a signal-to-noise ratio of 10:1, in an hour of integration through a broad-band filter R = 5. The Strehl ratio is a modest 0.1. The table uses a "pixel" with a diameter matched to the 50% encircled energy diameter, using the diffraction-limited PSF (point spread function) in the AO case and the Kolmogorov seeing-limited PSF in the other. Note that we are throwing away 90% of the light in the AO case! The result is that, with large apertures, even at low Strehls the advantage of discriminating against background can outweigh the significant throughput loss of ignoring the light lost to the uncorrected halo. A factor is also included (0.8) for the throughput loss of the AO elements.
The results for the V-band are grayed out because we consider achieving even 0.1 Strehl out of reach for current technology at the shorter wavelengths. Note that in terms of detectivity, there is little advantage to AO at this Strehl ratio with a 4-m telescope. However, with a 30-m telescope, about three magnitudes of improvement are possible.
Unfortunately, all of the enemies of good image quality gain strength as the telescope aperture grows: turbulence, wind buffeting, gravity, and temperature gradients.
In addition, we have raised the bar on the requirements of the corrective system by having a higher resolution diffraction limit. Although for a given Strehl ratio the required RMS (root mean square) wavefront residual is the same, the requirements of virtually all other aspects of the AO system are more demanding.
With extremely large telescopes (ELTs), Adaptive Optics are called upon not only to correct the atmosphere, but to allow the ELT to function at all. Simple extrapolations from the experience with current generation telescopes indicate that correction of wind buffeting is essential for good operation. Fortunately, the corrections required are only of the same order as those required to correct turbulence.
The concept of active control of telescope disturbances and adaptive control of atmospheric aberrations is well confirmed by experience with the 8-m class of telescopes. To achieve improved performance with larger apertures, we need to move toward a more integrated approach.
We are exploring concepts and intentionally pushing the limits of technology. The point design is a generator of further studies of the sort envisioned in the AO roadmap. At the same time, we are trying to envision a design that could be implemented on the timescale of a decade. We try to avoid postulating technologies for which there is no plausible path to success.
The New Initiatives Office (NIO) is building from the strength and experience of the Gemini multi-conjugate adaptive optics (MCAO) system development program. We also are drawing on and are keen to support a number of AO developments in the wider community:
- Keck/CELT work on active control of segmented mirrors
- University of Arizona-Arcetri work on adaptive secondaries
- Work on laser guide stars by USAF and others
- Commercial developments in solid-state lasers
- Development of concepts enabling diffraction-limited coronagraphy
- Deformable mirror development efforts
- Computer hardware and control algorithms for very-high-order AO systems
Four modes of AO correction are discussed in this book. They vary in complexity and performance and target different science goals:
Prime Focus AO (Seeing Enhancement) - This is accomplished by partial correction of the lower altitude contributions to seeing, resulting in an improved value of r. This enables wide-field, visible light applications that will permit, for example, spectroscopic surveys of galaxies. See Appendix 4.6.A for a conceptual description of this mode and initial performance estimates. The MOMFOS section presents an implementation at prime focus.
Direct Cassegrain AO - A traditional single guide star AO, this will be particularly useful in the mid-IR (infrared) because high Strehl ratios are achieved fairly easily at the longer wavelengths involved. This will permit, for example, very high resolution mid-IR spectroscopy, which is the kind of application that is unlikely to be available on foreseeable space platforms. See Section 4.2 for a discussion of error term in the implementation of traditional AO.
Very-High-Order AO - This will push to reach Strehl ratios in the range from 0.8 to 0.9, suppressing the halo of stellar objects to the point where enhanced high dynamic range imaging becomes possible. The most tantalizing application is the direct imaging of planets around nearby bright stars. Section 220.127.116.11 presents a coronagraphic implementation of this mode.
Multi-conjugate adaptive optics (MCAO) - This will extend beyond the limits of the isoplanatic patch the capabilities of imaging at the diffraction limit, thereby permitting crowded field photometry and Integral Field Unit (IFU) spectroscopy. The implementation of MCAO on the Point Design is presented in Section 4.6.2.
We propose to implement the last three modes at Cassegrain focus. A deformable secondary plays a key role in each. These three modes represent different levels of complexity in implementation, but bear a close family relationship to each other. Figure 2 presents a block diagram that illustrates the relationship. In a.) Direct Cassegrain AO, the deformable secondary is the unique adaptive element, accomplishing tip-tilt and higher order correction. In b.) high-order AO mode, an additional deformable element of very-high-order is incorporated into a coronagraphic instrument, together with a fast tip-tilt mirror. The deformable secondary plays the vital role of handling relatively large throw, moderate order corrections at a relatively low frequency. In c.) MCAO, laser guide stars (for higher order sensing) and multiple natural guide stars (for tip-tilt sensing) are used to derive full three-dimensional information about the atmospheric turbulence. An additional tip-tilt mirror and three deformable mirrors, conjugate to different altitudes in the atmosphere (0, 5, and 10 km), can now be used to undo the effects of turbulence over different angles of arrival through the atmosphere. The deformable secondary continues to reduce the demands on the adaptive elements downstream. The result is adaptive correction over a wide field.
In the diagram, sensor data paths are indicated in red. Tip-tilt corrections are in green, and higher order corrections are in blue. The dashed lines indicate the path of very low frequency corrections that are passed (offloaded) to low bandwidth "active" elements in the control of the primary and main axes. The distinction between active and adaptive is faint. All paths are closed loop; those we continue to call "active" are of a lower frequency.
We are carrying out point designs for each AO mode that:
- Test concepts
- Highlight difficulties
- Indicate areas of required technology development
- Illustrate how the AO system gets integrated with the rest of the telescope
- Serve as proof of existence
- Allow one to start iterating with science requirements
Table 2 briefly summarizes the design parameters and performance characteristics of the point designs for the three AO modes illustrated in Figure 2. A point design is preliminary to a trade study (see Section 4.1).
Simulations can predict performance of these kinds of AO systems. The GSMT, which involves higher order corrections than did previous systems, will require substantial enhancement of the numerical tools for modeling. In addition, the GSMT will integrate AO with active control systems in unprecedented ways that will have to be modeled. For the point designs, a key goal will be the development of tools and capabilities that will enable:
- Simulations of telescope and AO system optical performance, including laser systems, wavefront sensors, and deformable mirrors (See Appendix 4.6.D for a status report on this topic)
- Simulations of the behavior of layered control systems
We mention selected examples of outstanding problems, current achievements, and future studies that serve to encapsulate the present state of our work on AO for a GSMT: Two major challenges that will have to be met in order to achieve the kind of performance discussed above are attaining high Strehl in the high-order AO system and dealing with laser beacon elongation in MCAO.
Two recent achievements include the development of modeling techniques and efficient control algorithms that enable us to simulate wavefront reconstruction in detail for AO systems, with up to 10,000 actuators for MCAO and 100,000 actuators for high-order AO. See Appendices 4.6.B and 4.6.C for a detailed description of these methods. We have completed the system layout of a plausible coronagraphic instrument for the near-IR.
In two future studies, we will underwrite the development of a robust, higher power laser system for sodium beacons with fiber optics beam handling, and investigate novel pulse formats to avoid the degradation of the WFS (wavefront sensor) by guide star elongation. The point design systems will serve as test cases for these tools and capabilities.