Keck Cosmic Web Imager

We are designing the Keck Cosmic Web Imager (KCWI) as a new facility instrument for the Keck II telescope at the W. M. Keck Observatory (WMKO). KCWI is based on the Cosmic Web Imager (CWI), an instrument that has recently had first light at the Hale Telescope. KCWI is a wide-field integral-field spectrograph (IFS) optimized for precision sky limited spectroscopy of low surface brightness phenomena. KCWI will feature high throughput, and flexibility in field of view (FOV), spatial sampling, bandpass, and spectral resolution. KCWI will provide full wavelength coverage (0.35 to 1.05 μm) using optimized blue and red channels. KCWI will provide a unique and complementary capability at WMKO (optical band integral field spectroscopy) that is directly connected to one of the Observatoryís strategic goals (faint object, high precision spectroscopy), at a modest cost and on a competitive time scale, made possible by its simple concept and the prior demonstration of CWI.


The Keck Cosmic Web Imager (KCWI) is a wide field, seeing limited, integral field spectrograph (IFS) optimized for precision sky limited spectroscopy of low surface brightness phenomena. KCWI will feature high throughput and flexibility in field of view (FOV), spatial sampling, bandpass, and spectral resolution. KCWI will provide full wavelength coverage (0.35 to 1.05 μm) using optimized blue and red channels. The preliminary design phase of KCWI has recently been initiated through collaboration between CIT, UCSC, and WMKO.

KCWI is based on the Cosmic Web Imager (CWI) recently commissioned at the Palomar Observatory[1]. CWI is a single channel instrument with a fixed spatial sampling scale and a single adjustable VPH grating and an articulated camera with 2k x 4k mid-band optimized CCD. KCWI will build on the heritage of CWI with a two channel design offering three selectable spatial sampling scales and a selection of VPH gratings. KCWI will be located at the right Nasmyth focus of the Keck II telescope. A plan view of an initial concept for the opto-mechanical layout of KCWI is shown in Figure 4.


KCWI is designed to provide visible band, integral field spectroscopy (IFS) at WMKO with moderate to high spectral resolution. KCWI will benefit from the excellent seeing, low sky background, and large aperture, providing a world-leading capability. KCWI will provide a number of substantial benefits, including flexible image resolution and binning, flexible spectral resolution, excellent sky subtraction, and the highest possible throughput optimized for both the blue and red ends of the visible wavelength range.

KCWI will address a broad range of science applications, a number of which are highlighted in this section.

(1) Young Stars and Jets. KCWI will probe the kinematics and ionization physics of young stellar jets (such as Herbig-Haro objects) with high resolution that can resolve the <50 km/s line widths. With its excellent sensitivity KCWI will be suited to low surface brightness targets and can search for faint shocks and relic signatures of prior outflow history.

(2) Evolved Star Nebulae offer a wide variety of physical phenomena which can be highly informative about poorly-modeled mass-loss processes and the physics of outflows. Pulsar wind nebulae are excellent candidates for 2D kinematic, ionization, and density/pressure mapping. These nebula are now believed to be strong ultra-high-energy gamma-ray sources. Asymptotic Giant Branch stars in the thermal pulse phase show copious mass loss episodes. But many unusual nebulae are extremely faint in the optical[2],[3] and show complex morphology, requiring 2D spectroscopic imaging with high sensitivity.

(3) Light Echoes in the Milky Way and nearby galaxies are historical evidence for supernovae and possibly other energetic explosions. Extended, low surface brightness structures mapped with KCWI could provide spectroscopic evidence confirming their SNe or GRB origin and object class.

(4) Star Clusters. There has been tantalizing evidence for intermediate mass black holes at the center of several globular clusters[4].KCWI would provide high quality 2D kinematics with 15 km/s resolution along with abundance mapping in old and young clusters.

(5) Low Surface Brightness Galaxies and Star Formation in Extreme Regimes. An exciting area of research is the study of low surface brightness, low mass density galaxies and associated star formation. Topics include extending star formation laws to the lowest densities and galaxy masses, testing the universality of the initial mass function in the low-density regime, studying the stellar populations and searching for direct evidence of feedback. Low mass galaxies are often dark matter dominated, and provide a unique opportunity to probe the distribution of dark matter in galaxy cores and test dark matter models. A long-standing problem for these models that can be investigated is the lack of the "cusps" predicted in galaxy cores if the dark matter is cold and not self-interacting.

(6) Galactic Halos, Thick Disks, Streams, Intracluster Light. Recently, deep photometry has produced spectacular results showing that galaxies very often host thick disks, large diffuse halos, complex tidal streams and other extended features. These features are believed to be relics of prior minor and major mergers and interactions with satellites. The more extended structures have cosmological coherence timescales and KCWI offers the possibility of attempting kinematic, stellar population, and even abundance measurements of these highly extended low surface brightness structures. Spectroscopic analysis combined with modeling could provide strong constraints on merger histories and their impact on galaxy evolution. Intracluster light has been detected in many nearby clusters, and is believed to be produced by stars stripped from their galaxies as the groups and cluster assemble, as well as by stars formed during interactions in tidal streams. KCWI would provide kinematic maps and line-index diagnostics that could yield stellar age and metallicity information crucial to unraveling these stellar relics.

(7) AGN/QSO/Galaxy Co-evolution. One of the most interesting and controversial issues in galaxy evolution theory is the role of AGN feedback on determining the star formation history of galaxies and the black hole mass/stellar velocity dispersion relationship. One approach to this is to study nearby galaxies with AGN and extended emission to understand the physical properties, nature of ionization, mechanical energy flux, and ultimately the power deposited in the ISM to determine the effectiveness of feedback for self-regulation.

(8) Circum-QSO Medium. Prior to and shortly after QSOs are born, the Circum-QSO medium (CQM) may be a fair representation of the typical QSO environment. In particular, the CQM, illuminated by the QSO could provide a very useful picture of the CGM around galaxies at 2 < z < 6 through high signal to noise ratio (SNR) emission maps. KCWIís simple high throughput optical design should allow observations with a high dynamic range. Coupled with an occulting spot, it should be possible to explore the QSO host galaxy, host environment, and intervening absorber galaxy properties.

(9) Low and High Redshift Galaxy Structure

Galaxy stellar kinematics and populations: Bulges. KCWI would provide exquisite maps of stellar kinematics and populations traced by the usual absorption line indices in the high surface brightness cores of typical bulge systems, and thus furnish high order kinematic moments and spatial mapping which could separate multiple structural components of bulges assembled over cosmological time. KCWI could extend kinematic and stellar population measurements to the unexplored low surface brightness regions in the outskirts of galaxies, probing relic structures with long dynamical times and memories of their assembly history. In low surface brightness dwarf galaxies, kinematic maps will provide constraints on dark matter distribution and properties, allowing an attempt to address the cusp problem.

 Galaxy stellar kinematics and populations: Star-forming galaxies. In star forming galaxies it may be possible to map age vs. position in galaxies and reconstruct the local and global star formation history, and relate these to causal factors such as gas density, dynamical instabilities, ISM turbulence, and global star formation laws. KCWI extends these observations to low surface brightness irregulars, providing maps of complex star forming regions that could test for the impact of feedback, ionization, and turbulent regulation on the history of star formation in these low mass systems.

High z galaxy environments / assembly. High redshift galaxies are likely to lie in complex, multi-component systems with fainter companions and extended stellar and gaseous features. KCWI will detect these fainter companions and measure relative velocities and other properties (e.g., star formation rate). 2D coverage permits summing light from extended components and producing the highest resolution spectra with the highest possible SNR.

 Strong lens systems are complex, 2D systems that could be analyzed in detail using KCWI spectral images, providing information about the lensing galaxy (redshift, stellar velocity dispersion, mass, stellar populations) and lensed galaxy (stellar populations, star formation rate, age, metallicity, etc.).

Galactic feedback at low and high redshift. Feedback from massive star formation is one of the critical missing components in a complete understanding of galaxy formation and evolution. At high redshift, KCWI will probe outflows and their radiative impacts (if any) on the surrounding CGM/IGM (see below). At low redshift, KCWI can be used to map galactic outflows on large scales (more distant galaxies) and in nearby galaxies on the scale of star formation regions. It should be possible to detect very low surface brightness gaseous halos, galactic fountains, and superwind outflows even if the primary energy loss is not detectable.

(10) Circum‑Galactic Medium. One of the forefront topics today is the connection between galaxies and the gas in their dark matter halos (we term this the Circum-Galactic Medium or CGM) as well as with the IGM gas in the cosmic web. KCWI is optimal for detecting low surface brightness emission from redshifted Lyα as well as resonance lines of OVI1033, CIV1550≈ over the redshift range 2 < z(Lyα) < 6. For example, there have been tentative detections of Lyα halos around Lyman Break Galaxies, which may trace gas in dark matter halos which is either fueling star formation and galaxy formation, or gas that has been energized by galactic superwinds. A simulated observation is shown in Figure 1. Separating the different emission components, isolating the excitation mechanism for resonance line emission (scattering, recombination, shocks), and removing foreground objects and companions associated with the target galaxy requires 2D spectral imaging with excellent sky subtraction, the highest possible sensitivity, good imaging resolution, and moderate (R ~3000) spectral resolution to separate components at different velocities and to resolve velocity structure produced by bulk motions and radiative transfer in the resonance line.

(11) Lyα Emitters and Blobs have been detected using narrow-band imaging but are still poorly understood. 2D spectral mapping will provide kinematic information, constraints on associated metal line emission (OVI, CIV, etc.), and constraints on continuum sources that may be powering the blobs.

(12) Emission from the IGM/Cosmic Web

Metagalactic ionizing background can be measured in principal at high redshift using Lyα fluorescence, and at low redshift using Hα or Hβ emission. Reaching the very low surface brightness required, and separating out other mechanisms for producing the emission is best accomplished with2D spectral mapping designed to probe emission strengths < 0.1 to 1% of sky.

Emission from the Cosmic Web may be detectable at the limits of KCWI sensitivity. Figure 2 shows a prediction of Lyα emission at z ~6 from filaments of the IGM[, with the KCWI coarse FOV superimposed as a mosaic of 15 pointings. Emission at the green-yellow level is definitely detectable using statistical filament stacking techniques and may be even detected via direct imaging.