Z-Spec: A MM-Wave Spectrometer for Measuring Redshifts of Submillimeter Galaxies J.J. Gromke 1, J.J. Bock 1,2, C.M. Bradford 1, M. Dragovan 2, L. Duband 4, L. Earle 3, J. Glenn 3, B.J. Naylor 1, H. Nguyen 1,2, H. Matsuhara 5, J. Zmuidzinas 1 1 California Institute of Technology, Pasadena, CA, 2 Jet Propulsion Laboratory, Pasadena, CA, 3 CASA, University of Colorado, Boulder, CO, 4 CEA, France, 5 Institute of Space and Astronautical Science, Sagamihara, Japan Realization We are developing Z-Spec, a direct-detection spectrometer (195 GHz < f < 310 GHz). The concept combines a classical diffraction grating with a two-dimensional waveguide in a hybrid configuration (Figure 3). A transmitting feed horn (left) illuminates the faceted grating surface (bottom). The reflected, converging beams for three frequencies are shown. Receiving feed horns couple the dispersed radiation onto an array of bolometers (upper left). Waveguide Spectrometer Advantages This grating configuration offers significant advantages compared to classical spectrometers. · Compact size. For a millimeter-wave system, the grating is extremely compact, with dimensions of only 56 cm by 42 cm by 1 cm For comparison, a classical diffraction grating spectrometer (i.e., not waveguide coupled) would require a size of approximately 90 cm by 75 cm by 75 cm. · Sensitivity. Dispersing the radiation over a large detector array enables a broad instantaneous band and minimizes the detector background loading. · Light-tightness. The grating is completely sealed. · No moving parts. This is highly desirable for space applications. · Modularity. The compact grating could be used as a back-end of a more complex spectrometer for higher resolution. Space-Borne Application The millimeter-wave spectrometer will serve as a demonstration of the waveguide-coupled diffraction grating technology with efficient coupling to background-limited detectors. The properties of the waveguide-coupled grating make it an excellent candidate for a spectrometer on-board a cold (4-5 K) infrared telescope in space, such as SPICA, a proposed Japanese telescope with a 3.5 meter primary. An FIR spectrometer onboard SPICA would have spectroscopic sensitivity three orders of magnitude beyond planned far-infrared missions. Such capability would, for example, allow spectroscopic identification of all the galaxies in the 250  m confusion-limited surveys of the Herschel Space Observatory. Z-spec: Revealing the History of the Dusty Universe An extragalactic, far-infrared background radiation with integrated power nearly equal to the integrated optical & UV starlight in the universe was discovered by NASA's Cosmic Background Explorer. Ongoing and imminent submillimeter surveys with SCUBA, MAMBO, and BOLOCAM are resolving this background into individual galaxies (Figure 1). These surveys promise to provide a unique probe of star formation, nuclear activity, and structure formation in the universe. Waveguide-Coupled Diffraction Grating: A new Spectrometer Architecture Figure 3: Schematic of a waveguide grating spectrometer Detectors We require a large-format array (N ~ 150) of millimeter-wave direct detectors with high sensitivity (NEP ~ (4 – 6) W/  Hz) to achieve background-limited performance. Silicon nitride bolometers fabricated at the Micro Device Laboratory at JPL have achieved a sensitivity of 1.5 x W/  Hz at an operating temperature of 100 mK. Cryogenic system Cryogenic Features: · 100 mK stage: ADR · 300 mK stage: 3 He/ 4 He sorption fridge with 30 l 3 He and 60 l 4 He (STP) · Liquid He volume: 29 l · Liquid N 2 volume: 20 l · toroidal liquid N 2 & He tanks enable the JFET-modules on the 77 K stage close to be mounted close to the bolometers · Expected hold time: 48 hours Grating Prototype We have built a millimeter-wave room-temperature spectrometer as a prototype. (Figure 4). Since the number of bolometers and the size of the grating increase linearly with the resolution, we chose a moderate spectral resolution corresponding to 155 detector elements. The theoretical spectral resolution ranges from R = 350 at 310 GHz to R = 200 at 195 GHz. This design compromises resolution, total bandwidth and size. Figure 1: The Hubble Deep Field (Ref. 1) observed with the HSO (left) and at 0.87 mm with SCUBA. (right) (Ref. 2). Figure 2 demonstrates the feasibility of detecting high-redshift sources with a background-limited 1 mm spectrometer. The curves are drawn for each CO transition over the range of redshifts for which it is observed in the 1mm atmospheric window. The CO line luminosities are taken as fractions of the total FIR source luminosity (Blain et al. 2000). The fractions vary from 3x10 -8 to 3x The CII luminosity fraction is taken as 1x Sensitivity at the IRAM 30 m telescope will be 4-5x better. The redshift distribution of the submillimeter galaxies is not well known. Some galaxies are so dusty that they do not have observable optical counterparts, while others have multiple possible counterparts within the large submm beams. Far-IR fine structure ([CII], [OI], [OIII]) and millimeter-wave rotational (CO) spectral lines can be used to measure the redshifts of these galaxies. The narrow submillimeter windows do not permit routine observations of redshifted far-IR lines from the ground. At 1 mm, the atmosphere becomes reliably transparent over a broad band ( GHz). For 0.9 < z < 4 two or more CO transitions are always observable in the window, providing an unambiguous redshift probe. With this prototype we have demonstrated the functionality of the waveguide grating. The setup uses a tunable power source for frequencies up to 240 GHz. The signal is coupled in to and out of the spectrometer through custom feed horns. The horns are presently uncoated and holding only modest tolerances and are the dominant loss in the system. We anticipate substantial throughput Requirements To make efficient, reliable use of this long-wavelength redshift probe, a spectrometer would require: · resolution (R ~ 800) matched to extragalactic line widths · background-limited sensitivity with high throughput. · bandwidth as broad as possible · compact size (classical broadband spectrometers at 1 mm are prohibitively large) Figure 4: A picture of the prototype with the top plate removed. A 12-inch ruler is shown for scale. References: 1. R. Williams and the HDF Team (ST ScI) and NASA 2. Hughes, D., et al. 1998, Nature 394, Blain, A.W., et al. 2000, MNRAS, 313, 559 Figure 2: The feasibility of detecting high-redshift sources with a background-limited 1 mm spectrometer. improvements from better fabrication of the horns and in reduced waveguide losses upon cooling. The intrinsic spectral resolution of the system is higher than shown, because the plotted profiles include the convolution with a finite-sized output feed. A measurement with a small output feed at 230 GHz shows a resolving power of around 320, very close to the calculated value.