Performance Simulations for Kola: Achieving High-Resolution, Visible-Light AO Correction Over a 1 Arcminute Field

TL;DR

This work investigates achieving high-resolution visible-light imaging over a wide field on Keck I via a visible-light MCAO concept called KOLA. It employs the MAOS end-to-end simulator with GPU acceleration to explore design trade-offs in LGS configurations and adaptive secondary mirror actuator density, validating the approach against the current Keck SCAO system. The nominal design—a 10 LGS, 3 TTNGS configuration with 3600 ASM actuators—achieves about 15 mas angular resolution and a Strehl ratio of roughly 34% at 652 nm on-axis, across a 60'' field and 350–2500 nm wavelengths. These results demonstrate the potential of diffraction-limited visible-light AO over wide fields on Keck I and provide a roadmap for further optimization of LGS asterisms, DM conjugation, and performance under varying seeing conditions.

Abstract

We present performance simulations for a proposed visible-light, multi-conjugate adaptive optics system for the 10-meter W. M. Keck I telescope that aims to deliver near diffraction-limited angular resolution at optical wavelengths. Our proposed architecture, the Keck Optical Laser Guide Star Adaptive Optics System (KOLA), combines multiple laser guide stars (LGS) and deformable mirrors to enable wide-field correction across a 60 arcsecond field of view. Simulations were conducted using the open-source Multi-Threaded Adaptive Optics Simulator (MAOS), which we validated against on-sky data for the current Keck I adaptive optics system. We evaluated KOLA performance across a range of design parameters and report key point spread function metrics, including Strehl ratio, full width at half maximum, and encircled energy radius. Example science-driven requirements include resolving black hole spheres of influence, probing crowded stellar fields, and imaging protoplanetary disks. Trade studies on actuator count and laser guide star configuration help inform future design decisions. We present a nominal KOLA design (10 LGS, 3 tip-tilt natural guide stars (TTNGS), and 3600 actuators on the adaptive secondary mirror). Performance simulations show a 15 mas angular resolution with a Strehl ratio of 34% at 652 nm on-axis. More work is needed to explore alternative LGS/TTNGS asterisms, optimize conjugation heights for high-altitude deformable mirrors, and test performance under poorer seeing conditions.

Figures (9)

• Figure 1: Structure of the MAOS simulation package. A master configuration file defines global simulation settings and sequentially calls the configuration files for the DMs, atmosphere, and wavefront sensors. This hierarchical structure enables coordinated control of all simulation components while allowing global parameter changes without modifying each configuration file individually. Simulation outputs (e.g., PSFs, telemetry, and wavefront error data) are written to a designated results directory.
• Figure 2: Simulated astrometric microlensing event comparing performance with the current Keck AO system (middle) and with KOLA (right). KOLA's higher angular resolution enables the two lensed stellar images to be separately resolved during the event, allowing direct measurement of the lensing geometry. This capability will enable systematic detection and weighing of stellar-mass black holes in the Milky Way, one of KOLA's key science drivers.
• Figure 3: Angular resolution required to resolve key features in protoplanetary disks as a function of distance. At $150\,\text{pc}$ (Taurus cluster), current Keck AO can detect large-scale disk strucutre, while KOLA will reach $\sim 3\,\text{AU}$ resolution, sufficient to probe the terrestrial planet-forming region. At $400-450\,\text{pc}$ (Orion cluster), KOLA will still achieve $\sim10\,\text{AU}$ resolution, allowing direct imaging of gaps, structure, and snowlines sculpted by giant planets. This level of detail is inaccessible to current ground-based AO or space telescopes.
• Figure 4: Simulated imaging ($810\,\text{nm}$) of a young stellar cluster at $4\,\text{kpc}$ using MAOS PSFs. Left: performance with the current Keck AO system. Right: performance with KOLA. The KOLA simulation resolves significantly more faint stars and disentangles crowded sources that remain blended in current AO imaging. The 30" field of view is shown on a logarithmic scale, with stellar magnitudes assigned using the SPISEA package. This demonstrates KOLA's capability to identify faint populations in dense environments across the Galaxy. Due to the finite size of the PSFs produced by MAOS, clipping occurs for the KOLA image. However, it is still possible to see the higher angular resolution capabilities.
• Figure 5: A nominal schematic of the KOLA visible-light multi-conjugate adaptive optics system. An adaptive secondary mirror (conjugated to $-100\,\text{m}$) provides atmospheric wavefront correction, complemented by two high-altitude DMs conjugated to $6\,\text{km}$ and $10\,\text{km}$, respectively. The inset figure shows the nominal on-sky asterism of laser guide stars and tip-tilt stars: eight LGS placed at a $30"$ radius for higher-order correction, and three TTNGS at a $30"$ radius. Performance trade studies in this paper explore varying the number of ASM actuators and the count of laser guide stars.