Download 2013. CCAT. All Rights Reserved.

X-Spec: a Multi-Object Wideband
Spectrograph for CCAT:
Concept and Technology Development
C.M. Bradford for the X-Spec study team
We are developing X-Spec, a multi-object wide-band direct-detection spectrometer for CCAT. X-Spec is designed for rapid full-band (195-510 GHz), moderate resolution (R~700)
spectral surveys of galaxies, measuring the bright atomic fine-structure and molecular rotational transitions which cool the interstellar gas. Detected lines will provide redshifts for
and interstellar gas conditions in tens of thousands of galaxies ranging from the early universe (z > 6) to the present day, a survey scale which is impossible with ALMA. X-Spec is
based on an on-chip spectrometer in which the spectrometer and kinetic inductance detectors are defined lithographically in a single few-cm2-sized package which offers the
potential for a massive integral-field spectrometer limited only by available detector format. For the initial X-Spec to be deployed at or near CCAT first light, we are designing
toward 50,000 total detectors; they will be arrayed in 25 spectrometer units, each individually steered to a target galaxy with a (warm) articulated quasioptical feed.
www.CCATObservatory.org
Motivation
X-Spec at a Glance
The top scientific priority identified for CCAT is to chart the cosmic history of
the star formation and the buildup of heavy elements. Of particular interest
is the early period, from approximately 500 million to 3 billion years after the
Big Bang (corresponding to redshift ~10 to redshift ~2), in which the first
generation of stars began enriching the medium, the Universe was reionized,,
and the star formation rate increased to a what appears to have been a broad
peak or plateau. This period of galaxy infancy witnessed the greatest
logarithmic growth in the density of stars, black holes, and heavy elements,
and set the stage for the relatively modest evolution for the remaining 10
billion years of the Universe’s history.
On-Chip Spectrometer
• The first multi-object spectrometer for the mm / submm.
• Covers the full 195-520 GHz range at R~700.
• Based on an on-chip lithographically-patterned spectrometer.
• Uses RF-multiplexed kinetic inductance detectors (KIDs) based on
titanium nitride (TiN), like the CCAT SWCam and LWCam. Shares
readouts. 250 mK operating temperature.
• ~2000 detectors per dual-polarization module.
• First-light X-Spec will have at least 25 spectrometer elements.
• Source density is 1/100 to 1/10 per beam, so each spectrometer is fed
with an articulated quasioptical feed. Will be faster than ALMA for fullband surveys of galaxies with unknown z.
• Will eventually be a large integral-field imaging spectrometer with
~1000 spectrometer ‘pixels’, ~30x faster than ALMA for spectral surveys.
Sensitivity
Hopkins and Beacom, 2006
Chary+Pope
350/870 μm flux ratio
SPT
4900
2300
Joaquin Vieira
350 μm
dropouts
1000
350/870 micron ‘color’ as a function of redshift for submm / mm selected galaxies. The large
range in dust properties translates into a large redshift uncertainty.
© 2013. CCAT. All Rights Reserved.
CCAT Engineering Design Phase is partially supported by
funding by the National Science Foundation’s Division of
Astronomical Sciences
J. Zmuidzinas
The
system
uses
noibium microstrip on
silicon nitride. Each KID
consists of and inductor
that absorbs radiation,
and a capacitor, both
patterned in TiN. We
have built and begun
testing a Band 1 80channel
prototype,
shown
here.
(E.
Shirokoff et al., chip
design.)
7 mm
mm-wave feedline
(niobium, traveling
horizontally)
Star formation rate history compiled by Hopkins and Beacom, 2006. The epoch prior to z~3
relies on rest-frame UV measurements, some of which have substantial dust opacity
corrections.
The bulk of our information about this early epoch originates in rest-frame
optical/ultraviolet (observed-frame optical/near-IR). In many galaxies, the UV
is a tiny fraction of the total bolometric energy, so substantial (and uncertain)
dust corrections are required. CCAT cameras will identify broadly high-redshift
dusty galaxies using flux ratios, but flux ratios do not provide an accurate
redshift (unlike in the UV / optical in which there is a break). Charting a
detailed history requires spectroscopic redshifts, and ALMA is not optimal for
survey large numbers of objects. CCAT X-Spec is designed for spectral surveys.
As redshifts are measured, atomic line fluxes (e.g. C+) also provide constraints
on the temperature, density and UV field strength in the interstellar gas,
revealing the conditions of star formation.
12700
We are developing an on-chip
millimeter-wave
spectrometer
technology for CCAT/X-Spec and other
imaging spectroscopy applications
such as intensity mapping.
The
approach is to create a filterbank in
superconducting transmission line:
each channel of the filterbank is a
half-wave resonator coupled to a KID.
mm-wave half-wave resonator
(U-shape, niobium)
mm-wave absorber =
meandered KID inductor
(titanium nitride)
First light X-Spec will cover both Band 1 and Band 2 for each of 25 independently-steered feeds.
The upper black curve shows the worst case sensitivity (single pol, 2×photon-noise, line extracted
from a total bandwidth of R=250), while the lower curve is the goal (dual pol, sqrt(2) × photon
noise, effective noise bandwidth of R=700). For both, the instrument efficiency is 25%, forward
efficiency is 92% (8% warm spillover), taper efficiency is 85%, and the Ruze efficiency factor
assumes a total surface RMS of 12 microns. The sky is assumed to have 700 μm of water and be
viewed at a 40 deg zenith angle. Model galaxy spectra are overplotted in color.
Steering System Approach
X-Spec
will
couple
each
spectrometer to a galaxy position
with a 2-axis quasioptical steering
system located just exterior to the
cryostat vacuum jacket.
The
concept shown here is based on a
pair of commercially-available
Aerotech rotary stages.
Lupe Banales et al.
KID resonator
capacitors
(titanium nitride,
interdigitated)
KID
coupling
capacitors
Initial optical tests show basic functionality and that materials have sufficiently
low loss. Next step is smaller KID inductor for increased response.
5 cm
Goldsmith &
Seiffert, 2009
S. Hailey-Dunsheath, T. Reck
Full 1-deg (3-m, curved) CCAT focal plane
~30 cm
accommodates ~96 units
Canadian Consortium