Chapter 5, Technical Studies

All of the adaptive optics (AO) systems proposed in this document for GSMT will require very significant technical advances beyond the current state-of-the-art. Table 1 is a brief review of the AO component performance requirements derived in the preceding sections. These requirements are tabulated by AO system (direct cassegrain AO, multi-conjugate adaptive optics (MCAO), very-high-order AO, and prime focus AO) and system component or development task (system design, modeling, and control; lab and field tests; wavefront sensing, wavefront correction, and laser systems; and opto-mechanical design). The following paragraphs expand upon some of these requirements, review current status and recent progress, and suggest several near-term studies and risk-reduction efforts that are most essential to future progress. Additional information can be found in Sections 4.6.2 and 4.7.6.2.

The last year has seen good progress in developing efficient wavefront control algorithms for future MCAO and very-high-order AO systems on ELTs, and in implementing detailed simulations to evaluate MCAO performance on 8-m class telescopes. Detailed simulations of GSMT and other ELTs should be feasible in the near future, provided that existing codes are parallelized to run on supercomputers and upgraded to incorporate recent advances in control algorithms. These simulations will enable a new level of design and optimization studies for ELT AO systems, including:

• Final confirmation of first order AO system design parameters, including the order of correction, control bandwidths, deformable mirror (DM) conjugate ranges, and guide star constellation (including options for natural guide star (NGS) MCAO and Rayleigh guide star MCAO);
• Detailed point spread function (PSF) characterization and estimates of NGS magnitude limits to support the development of the science case;
• Analysis of wavefront sensor (WFS) performance with highly elongated laser guide stars (LGSs), as will be necessary to determine laser power requirements and assess a variety of suggestions that have been made to compensate for this effect;
• Evaluation of AO performance with a segmented primary mirror, and the hierarchical interaction between the AO and active optics (aO) control systems; and
• More rigorous error budgeting of many implementation error sources, such as noncommon path wavefront errors, DM-to-WFS misregistration, DM hysteresis.

Moreover, detailed performance predictions, including implementation error sources and aO errors, will soon be more accurately anchored against real-world AO performance on smaller telescopes, beginning first with NGS AO and LGS AO, and then proceeding to very-high-order AO and MCAO as 4-8-m class systems come online. Progress on advanced wavefront control algorithms is also nearing the point where work may begin on the wavefront reconstructor electronics that will eventually be needed to implement these techniques in hardware.

The wavefront sensing requirements for GSMT range from foreseeable extensions of existing technology to dramatically new concepts. The order of wavefront sensing envisioned in this report for MCAO and direct Cassegrain AO is about 64 subapertures across the telescope aperture, which could be implemented using a conventional Shack-Hartmann WFS with a high-speed, low-noise CCD array of 256² or 512² pixels (the larger number of pixels might be necessary for wavefront sensing with highly elongated LGSs). This type of device appears to be a feasible extension of existing 128² wavefront sensing CCDs, although it is unlikely to be developed without technology development funding. The situation is probably different for very-high-order AO, because CCD arrays with 1024² pixels would be required to implement Shack-Hartmann sensors with up to about 256² subapertures. Alternative wavefront sensing approaches requiring fewer pixels per subaperture should be considered, including the shearing interferometer and the Smartt, or point diffraction, interferometer. Low order versions of these sensors that can be tested in closed-loop AO systems should be developed, first in the laboratory and eventually in the field.

Additional areas where research into novel wavefront sensing concepts might eventually be justified include (1) high-order IR wavefront sensors for Mid-IR NGS AO in regions of the sky with no visible stars, and (2) LGS wavefront sensors optimized for use with highly elongated LGSs. More work on the science case and detailed AO simulations is necessary to establish an actual requirement for these concepts, however.

The situation is somewhat less advanced with respect to wavefront correctors. The DMs required for MCAO on GSMT are essentially factors of two extrapolations (in linear dimension) from the existing mirrors of order 35² with inter-actuator spacings of 7 to 9 mm. Such mirrors could probably be built today, although the cost could easily be much higher than the rate of $1,600 per actuator currently typical for smaller mirrors. However, there are at least four other areas where significant innovations in wavefront correction technology will or may be necessary:

• Adaptive secondary mirrors with 1000-2400 actuators and a large (~10 µm) dynamic range are required for all of the AO system concepts considered in this report.
• Very-high-order systems will require microelectrical mechanical systems (MEMS) DMs with good figure quality, relatively high bandwidths, moderate stroke, and somewhere between 128² and 256² actuators. The current state-of-the-art is 12² actuators.
• Cryogenic operation for both conventional and MEMS mirrors may be desirable for observations in K-band, but this is not yet confirmed by the science case.
• Finally, we believe that the bandwidth requirements for the 300-500 mm diameter tip-tilt mirror postulated for the present MCAO opto-mechanical design may be extremely stressing for a mirror of this size, but more work on the aO disturbance characteristics will be necessary to confirm this.

Ongoing R & D activities in the first two areas should be continued and expanded, and telescope design studies should work towards estimating the tip-tilt disturbance spectrum reasonably soon.

Sodium guide star laser technology remains an essential area for R & D funding, both for 8-m class LGS AO and MCAO systems and for ELTs. The clearest way to overcome the LGS elongation problem on ELTs would be to increase the LGS signal level by perhaps a factor of 3 to 5. Increased laser power levels would also be useful for improving AO performance at shorter wavelengths. Other possible approaches include (1) novel laser pulse formats that would allow a short pulse to be "tracked" through the sodium layer, (2) NGS MCAO systems, and (3) Rayleigh guide star MCAO systems. Laser and WFS design work for these alternate approaches can be postponed, pending analysis and simulation results confirming that they are in fact viable system concepts.

Finally, all of the AO system opto-mechanical concepts presented in this report must be subjected to further analysis and trade studies, as is planned for the follow-on stages of the GSMT design effort.

Table 1 Technology requirements for GSMT AO systems vs. the current state-of-the-art.
	Direct Cassegrain AO	MCAO	Very-high-order AO	Prime Focus AO	Work to Date
Design Trade Studies	Error budget balancing	--Guide star constellations and DM conjugates --Order of correction, LGS power, and control bandwidth	Error budget balancing	--PSF FWHM and uniformity vs. guide star constellation and FOV	This report
Modeling/Simulation	--Parallelized wave optics simulations to model (1) 3-D LGS, (2) diffraction effects in atmosphere, optics, and WFS, and (3) implementation error sources --PSF stability and uniformity studies				Gemini South and CfAO MCAO simulation codes
Novel Control Algorithms	--Reduce computational complexity from O(n log n) to O(n)	--Reduce computational complexity from O(n3/2) to O(n) --Develop efficient algorithms for closed-loop case	--Reduce computational complexity from O(n log n) to O(n) --Adaptive and very precise minimization of calibration errors		--Gemini minimal variance estimators --Sparse matrix work at Montana, Gemini, and CfAO
Lab and Field Tests	IR wavefront sensing on smaller telescopes	MCAO demonstrations on smaller telescopes	--Very-high-order AO tests on smaller telescopes --AO astrometry / photometry tests on smaller telescopes		--SOR and AEOS high-order AO systems --Gemini South, ESO, and Palomar MCAO plans
Control Electronics	Develop architectures to implement efficient control algorithms with minimum latency				SOR FPGA reconstructor designs
Shack-Hartmann WFS	Develop CCD arrays and lenslets for order 642 sensors		Develop CCD arrays and lenslets for order 1282 or 2562 sensors??		SOR and AEOS order 322 sensors
Novel WFS Concepts	High-order, cryrogenic, IR sensors	--Sensors optimized for elongated LGS --Sensors to track short-pulse LGS	Direct phase measurement sensors (Smartt interferometer)
Adaptive Secondaries	Work towards ~2-m secondary mirrors with ~ 2400 degrees of freedom and large dynamic range.				MMT adaptive secondary
Deformable Mirrors	Work towards Cryogenic DMs of order 642	Work towards Big "conventional" DMs of order 642 with 5-10 mm actuator spacing	Work towards MEMS DMs of order 1282 to 2562 with 150-300 µm actuator spacing (cryogenic?)		--SOR and AEOS order 352 mirrors --Order 122 MEMS DMs with good figure quality
Tip/Tilt Mirrors		High-performance 300-500-mm diameter tip-tilt mirrors
Lasers		--50-100 W class lasers --Fiber-optics beam transport at high powers --50 KHz pulse formats with 1-2 µsec pulses			--Keck and Lick laser systems --Gemini/CfAO/SOR laser risk reduction program --Gemini North laser procurement
Opto-mechanical designs/components		--New optical layouts with fewer surfaces and/or smaller pupils --LGS WFS designs with minimized aberrations over 90-200 km LGS range	High quality Lyot stops and occulting disks		Gemini South LGS WFS design

---

Table of Contents | Back to Section 5.5, Characterization of Wind Loading | On to Section 5.7, Design-to-Cost

March 2002