FIRST LIGHT WITH SPIFI ON THE JCMT

We report the results of our first light with the South Pole Imaging Fabry-Perot Interferometer (SPIFI) on the JCMT. SPIFI is a direct detection imaging spectrometer designed for use on the 15 m JCMT and on the 1.7m AST/RO at the South Pole in the 350, and 450 m m submillimeter windows. SPIFI's detector is a 5 ´ 5 square array of silicon bolometers held at 60 mK in an adiabatic demagnetization refrigerator (ADR). The bolometers are fed by an array of Winston cones, yielding a beamsize of 7", and field of view is 35" ´ 35" on the JCMT. Three cryogenic Fabry-Perot interferometers (FPIs) in series deliver resolving powers, R = lamda/delta(lamda)~ 500 to 10,000 (delta(v) ~ 600 to 30 km s^-1) over the entire field of view. SPIFI mounts at the f/16 right Nasmyth platform of the JCMT. The full field of view is delivered through the elevation arm by a polyethylene lens relay system.

First light with SPIFI was on the JCMT in April, 1999. We focussed on mapping the CO(7-6) rotational line in the Galactic Center and external galaxies. Despite rather mediocre weather, we mapped the entire circumnuclear ring (200 spectra) at 7" spatial, and 67 km s^-1 velocity resolution (Figure 3). We also mapped the inner regions of M82 and NGC 253 in their CO(7-6) line emission (Figure 4) -- to our knowledge, the first detections of the CO(7-6) line from external galaxies. The scientific implications of these results are discussed below. SPIFI performed up to our expectations. At R = lamda/delta(lamda) = 4500 (delta(v) = 67 km s^-1), the front end system noise equivalent power (NEP) was ~ 9´ 10^-16 W Hz^-1/2 - within a factor of two of the background limit given the (measured) system throughput (~25%), and detector quantum efficiencies (~50%). This NEP corresponds to a receiver temperature < 100 K(DSB), so that SPIFI is very competitive on a per pixel basis with the best heterodyne receivers for spectroscopy of broadline (delta(v) > 30 km s^-1) Galactic and extragalactic sources. For mapping projects, SPIFI's 25 pixel imaging array provides a factor of 25 gain in mapping speed over a single pixel system.

1. SCIENCE WITH SPIFI
1. Background. In its current configuration, SPIFI can access any line in the 350 m m telluric window. In the near future, we hope to upgrade SPIFI so that we can access the 450 window on the JCMT and the 200 m m window at the South Pole. For the Mauna Kea windows, the primary lines are the ³P₂ -³P₁ [CI] 370 m m fine structure line and the CO (J=7- 6) and (J=6- 5) rotational transitions. JCMT goals include mapping the Galactic Center, nearby external galaxies, and ultraluminous IRAS galaxies in these lines, and (hopefully someday!) detecting redshifted fine-structure line emission at cosmologically significant distances (e.g. [CII] 158 m m @ z > 1.2, [OIII] 88 m m @ z > 3, and [OI] 63 m m @ z > 4.5).

The mid-J rotational lines of CO probe the warm dense gas associated with photodissociation regions (PDRs) and molecular shocks. They are important probes, as warm, dense molecular gas is common in Galactic star forming regions, the Galactic Center, and external galaxies (cf. Harris et al.1991). For PDR gas, the CO(6 -5) and (7 -6)/CO(1 -0) line intensity ratios are sensitive indicators of the strength of the far-UV radiation field, and the gas density, n (Kaufman et al. 1999). The mid-J lines indicate that much (>35%) of the total molecular gas mass is both warm (T > 50 K), and dense (n(H₂) > 10⁴cm^-3) in both Galactic star forming regions and starburst nuclei. Since the warm gas is an important component, its study is critical to understanding the interplay between star formation and the natal molecular clouds on galactic scales.

Neutral carbon is abundant in the ISM, amounting to 10% of CO in the Milky Way (MW) and up to 50% in starburst nuclei (cf. Keene et al. 1985). The [CI] lines are easily excited, and are therefore important coolants of PDRs, and perhaps cloud interiors as well. In dense clouds, the line ratio is temperature sensitive, while for more diffuse regions, there is a density dependence as well. Both [CI] lines are normally optically thin, thus tracing mass. The [CI] 370.415 um and CO(7- 6) (371.651 um) lines are only 1000 km s^-1 apart, so they both can be included in a single spectral scan. SPIFI can therefore map in the two lines at once, saving time, and resulting in "perfect" spatial registration and excellent flux calibration between the maps. We demonstrated this mode of operation during our September 1999 observing run at the JCMT (see Figure 1). Note that the line separation plus typical extragalactic linewidths (~ 300 km s^-1) and sufficient baseline (± 100 km s^-1) for a good spectrum, corresponds to >4 GHz at 370 um, so these experiments are very difficult using heterodyne receivers with typical (1 GHz wide) backends.

 

2. First Light Results. During our first observing run with SPIFI, we were somewhat conservative with the instrument. For the first half of the run, we focused on observations in the CO(7 - 6) line since the instrument is easier to set up at this wavelength. We chose a velocity resolution of 70 km s^-1 so that we could resolve the major velocity features in the Galactic Center, as well as the lines in bright galaxies. This was the lowest velocity resolution we could obtain for this run, so that it was not possible to scan far enough to include both the CO(7- 6) (372 mm) and the [CI] 371 mm lines. After mapping the entire Galactic Center Circumnuclear Ring (CNR) in CO(7- 6), we made the switch to the [CI] line with four nights remaining in the run, but, unfortunately, the skies were Band 3 or worse for these nights, so that we were not able to take [CI] spectra. During our second run in September 1999, we set up at a lower velocity resolution, since we were focused on extragalactic science. Unfortunately, the weather was quite poor during this run so that we obtained very little data. However, we obtained a small map of M82 in CO(7- 6), weakly detected the [CI] line in M82, and obtained a few spectra of Orion that contain both the CO(7 -6) and [CI] lines demonstrating our large bandwidth.

Galactic Center Circumnuclear Ring (CNR). The CNR is an inclined ring of cloudlets at r > 1.5 pc that orbits the dynamical center of the Galaxy, Sgr A^*. We imaged the CNR at 5" resolution in the 31.5 and 37.7 um continuum with our imaging Fabry-Perot, KWIC on the Kuiper Airborne Observatory (Latvakoski et al. 1999, Figure 1). Prominent in these images are the complete inner edge of the CNR, and each of the mini-spiral features (the northern arm, eastern arm/bar, and western arc). The observed far-IR emission arises from warm dust in photodissociation regions (PDRs). These images trace the deposition of far-UV flux (energetics), trace the morphology, and give hints of clumpy structures. However, spectral lines are required for kinematics, and are much better tracers of clumps. We therefore plan to map the entire CNR in the [CI] and CO(7- 6) lines with SPIFI on the JCMT. These JCMT maps will have spatial resolution comparable to our far-IR images, so that we expect many of the same features as in the KWIC images, but velocity resolved.

We began this project on our first (April 1999) run with a partially filled (12/25 pixel) array, focussing on the CO(7- 6) line. Despite rather mediocre weather, we mapped the inner 1' by 2' regions (> 200 spectra!, Figure 2 and 3) in the CO(7- 6) line at R ~ 4500 (delta(v) ~ 67 km s^-1). The spectra cover a velocity range of V_LSR -200 to + 400 km s^-1. This is the first large-scale mapping of the CNR in the CO(7- 6) line. The circumnuclear ring and streamer velocities are quite evident in the spectra and in map. The high-J CO line traces the excitation of the ring, delineating clumps, shocks from clump-clump collisions, and shocks formed where streamers entering the central cavity collide with the ring. These data were obtained during 2 shifts of Band 2 (~ 5 to 7% transmission towards the source) weather, and one shift of marginally Band 1 weather (~ 11% transmission towards the source).

We are planning to continue this project in our May 2000 JCMT run. We plan to lower the velocity resolution, so that we can map the entire ring in both the CO (7- 6) and [CI] 370 um lines simultaneously. The [CI] line will deliver a complete rotation curve for the ring unfettered by foreground absorptions, and give for the first time good measurements of the overall ring dimensions and mass, and when combined with the ³P₁ - ³P₀ 609 um [CI] maps in the literature, yield cloud temperature, and mass. The CO(7- 6) line is sensitive to density enhancements, the local UV radiation fields, and shocks. We expect enhanced CO(7- 6)/[CI] line ratios from the shocks of clump-clump collisions in the ring, or where the gas from the northern arm crosses the CNR. Morphological and kinematic information from both lines may link the streamers that cross the ring to the infalling gas. The imaging array yields near perfect registration and relative calibration between pixels and spectral lines in a map, greatly facilitating the analysis

GALAXIES: NGC 253 NGC 253 is the best example of a nearby (D ~ 2.5 Mpc) spiral galaxy with a starburst nucleus. It is very dusty and highly inclined (i ~78° ). The inner 500 pc contain a massive molecular bar which hosts much of the starburst activity. The far-infrared luminosity of the starburst region is ~ 1.6 ´ 10¹⁰ Lo so that the average far-UV interstellar radiation field, G_o, is very high, heating and disrupting the ambient molecular clouds

During one shift in April, we had fairly good transmission (10% or better towards the source) for about half an hour on NGC 253 before the sun began to influence the pointing. However, in this time, we clearly detected CO(7- 6) line emission in all 12 of the pixels in the array (Figure 4). The main beam brightness temperature was typically ~ 3 K.

Comparing our CO(7- 6) map with low-J maps from the literature characterizes the physical conditions of the molecular ISM. The CO(7- 6) line, when compared with lower J CO transitions in an LVG model constrain both the gas temperature and pressure. From our measured value at the nucleus we derive: T(gas) ~ 120 K, n_gas ~ 4 ´ 10⁴ cm^-3. By itself, the CO(7- 6)/CO(1- 0) line intensity ratio, is a sensitive indicator of G_o and n (Kaufmann et al. 1999). Comparing our CO(7- 6) data to the OVRO CO(1- 0) line (similar sized beam), we estimate G_o ~ 3 ´ 10⁴ times the local interstellar radiation field. These physical conditions are similar to those obtained from the far-IR fine-structure lines (Carrol et al. 1994), but the area filling factor of the CO(7- 6) emitting molecular gas is much lower (~ 0.10) than that of the PDR emitting gas (~ 1.4 -- more than one PDR along the line of sight) indicating that the molecular cores of the clouds in NGC 253 are small relative to their PDR envelopes.

 

2. A FEW NOTES ON THE INSTRUMENT

Array. For first light, we ran with 12 working pixels in the array. During our September 1999 run, we had all 25 pixels installed, and 22 were working reasonably well. We have made some repairs, and are hopeful that the full array will be operational for our May run.

Velocity Resolution. When warm, SPIFI can be set up at any velocity resolution between 300 km s^-1, and 30 km s^-1. For any given cooldown of the spectrometer, the range of velocity resolutions available is limited by our translation stage travel, so that the complete selection of velocity resolutions is not available. For a typical set up, we should be able to change velocity resolution from delta(v) ~ 150 km s^-1 to 40 km s^-1.

Scan Length. Our maximum scan length is determined by the stretch of the scanning PZT, and is typically 15 resolution elements. For example, if we operated at a velocity resolution appropriate for galaxies (100 km s^-1), SPIFI can obtain a spectrum that covers about 1500 km s^-1. For narrow line sources, (D v = 30 km s^-1), the maximum spectrum would cover 450 km s^-1. To obtain a spectrum with a proper baseline, the minimum scan length is about 5 resolution elements.

Sensitivity. Our best estimates for SPIFI's current sensitivity are tabulated below.

Resolving Power	Velocity Resolution	D TMB
1000	300 km s-1	11 mK
1500	200 km s-1	16 mK
2000	150 km s-1	22 mK
5000	60 km s-1	50 mK
10000	30 km s-1	100 mK

These noise estimates assume a 5 spectral resolution element scan, (so that for dellta(v) = 200 km s^-1, the resulting spectrum will be 1000 km s^-1 wide), and assume a very good night, 40% zenith transmission at 370 um. These numbers are for the zenith, so they must be adjusted for airmass. Our measurements in April indicated that the total coupling of the telescope to a point source at 370 um is 20%, consistent with h _fss = 0.65 and h _MB = 0.3.

 

3. ACKNOWLEDGEMENTS

Many people contibuted to the success of our first run. We are extremely grateful to the management and staff of the JCMT for giving us the opportunity to bring a complicated new instrument onto the superb JCMT telescope, and for their help with logistics, emergency repairs to components of the instrument and mount, software and interfacing, and shipping and packing. We thank the Director, Ian Robson for his continued support and enthusiasm for the SPIFI project. Perhaps most important is the wonderful support we received from Wayne Holland, and Richard Prestage at the telescope during the first nights of observing.

The SPIFI instrument and the science called out above is the result of much hard work, and writing a few good proposals. Personnel involved include Gordon Stacey, Matt Bradford, Thomas Nikola, Jim Jackson, Alberto Bolatto, Mark Swain, Jackie Davidson, Maureen Savage, Sarah Unger, Peter Ade, and Frank Israel. We are eternally grateful for Peter Ade's filters which made our lives a whole lot easier.

More information on SPIFI can be found at: URL address: http://.astrosun.tn.cornell.edu/research/projects/spifi.html, and in Stacey et al. 1998, Swain et al. 1998, and Bradford et al. 2000.