SIGGMA: A SURVEY OF IONIZED GAS IN THE GALAXY, MADE WITH THE ARECIBO TELESCOPE

Recombination lines are emitted when electrons in ionized gas recombine with atomic nuclei in an excited state and cascade down in energy level, n. The most probable transitions are for Δn = 1 changes in energy level and are called "α" lines. Transitions with Δn = 2 are known as "β" lines and so on. For hydrogen, α lines with n ≳ 40 are in the radio regime (λ ≳ 3 mm) and are termed radio recombination lines (RRLs; see Gordon & Sorochenko 2002 for a full account of the generation of RRLs).

There are three main sources of RRL emission in our Galaxy: H ii regions, diffuse ionized gas, and photodissociation regions (PDRs). H ii regions are zones of plasma surrounding massive young stars. Astrophysical RRLs were first detected toward the Omega H ii region in 1964. These observations were reported by Dravskikh et al. (1966) and Sorochenko & Borodzich (1966) and shortly afterward were followed by high signal-to-noise (S/N) RRL detections toward Orion and M17 (Höglund & Mezger 1965). There have been many subsequent RRL surveys of H ii regions (e.g., Reifenstein 1970; Wilson et al. 1970; Wilson 1980; Lockman 1989; Anderson et al. 2011). H ii regions are the most intense sources of recombination line emission, although at low frequencies diffuse ionized gas becomes a relatively bright source of RRL emission (see Alves et al. 2010; Lee et al. 2012).

In the Galaxy, the diffuse ionized gas consists of a low-density component (<1 cm⁻³), referred to as the warm (T_e ∼ 3000–8000 K) ionized medium. This component has a scale height of ∼1000 pc and is usually studied using optical recombination lines and pulsar dispersion measures (Taylor & Manchester 1977; Reynolds 1990). Observations of low-frequency (<a few GHz) RRLs have established the presence of another diffuse ionized component with a density in the range 1 to 10 cm⁻³ close to the Galactic plane and with a scale height of ∼100 pc (Gottesman & Gordon 1970). In the literature, this component is referred to as Galactic Ridge RRL emission by Davies et al. (1972), the extended low-density medium (ELDM) by Mezger (1978), evolved H ii regions by Shaver (1976), H ii envelopes by Lockman (1976) and Anantharamaiah (1986; e.g., Roshi & Anantharamaiah 2000, 2001; Baddi 2012), the extended low-density warm ionized medium by Petuchowski & Bennett (1993) and Heiles (1994), and the warm ionized medium by Heiles et al. (1996). Following Mezger (1978), in the present paper, this diffuse ionized component is referred to as the ELDM. Recently, extensive higher angular resolution (148) RRL observations of the ELDM were performed by Alves et al. (2012). They found that the distribution of the ELDM is strongly correlated with the location of Galactic H ii regions, confirming the observations by Lockman (1976), Hart & Pedlar (1976), and Anantharamaiah (1986). The origin of the ELDM is, however, still unclear, in large part because previous studies used low-resolution observations. One possibility is that it originates from the ionization of low-density regions surrounding giant H ii regions, from which photons "leak" (see Anderson et al. 2011 for a discussion on the W43 region), although this scenario does not fit all observations (Roshi et al. 2012).

In addition to hydrogen lines, carbon RRLs have been detected in several directions in the Galaxy. They are generally observed from interfaces between neutral and fully ionized regions, referred to as PDRs. Photons with wavelengths longer than the Lyman limit can escape the H ii region and ionize carbon and other atoms with lower ionization potentials than hydrogen. Carbon RRLs were first detected by Palmer et al. (1967) and have been the focus of many subsequent studies (e.g., Pankonin et al. 1977; Roshi & Kantharia 2011; Wenger et al. 2013).

The Survey of Ionized Gas in the Galaxy, made with Arecibo (SIGGMA) will fully sample the entire Galactic plane observable with the 305 m William E. Gordon Telescope at the Arecibo Observatory (30° ⩽ l ⩽ 75° and −2° ⩽ b ⩽ 2° in the inner Galaxy; 175° ⩽ l ⩽ 207° and −2° ⩽ b ⩽ 1° in the outer Galaxy) and will be the most sensitive large-scale RRL survey ever made. The survey data will permit a wide range of science, including studies of: (1) H ii regions, planetary nebulae, and novae; (2) the Galactic temperature; (3) the large scale structure of the Milky Way; (4) carbon recombination line emitting regions; and, possibly, (5) the ELDM.

RRLs can distinguish between thermal and non-thermal sources. In the section of the inner Galactic plane observable with the Arecibo telescope, there are thousands of continuum sources that were revealed by the L-band NRAO VLA Sky Survey (NVSS; Condon et al. 1998). To date, only a few hundred H ii regions have been identified in this zone (Lockman 1989; Bania et al. 2010). With a much greater sensitivity than existing surveys, SIGGMA will detect hydrogen RRLs in sources with peak line intensities ≳1.5 mJy (3 σ threshold).

H ii regions are ideal tracers of spiral arm structure in galaxies. SIGGMA will offer a large-area sample of Galactic H ii regions which, together with the ALFA HI survey of the Galactic plane (Peek et al. 2011), will permit a comprehensive study of Galactic structure and kinematics within the Galactic longitude range 30°–75°. This survey will also be useful for checking the current velocity field models of the Galaxy since a rotation curve for the 4.5–8 kpc galactocentric distance range can be derived from the data.

Churchwell & Wamsley (1975) found, for the first time, that the average electron temperatures (T_e) of H ii regions gradually increase with galactocentric radius R in the Galaxy. The same trend of T_e was also obtained by Shaver et al. (1983), who directly related this result to the metallicity gradient with R. Using high angular resolution RRL observations toward ultra-compact H ii (UCHII) regions, Afflerbach et al. (1996) derived the slope of the temperature gradient with R to be 320 K kpc⁻¹. Afflerbach et al. (1997) directly determined the metal abundance in these UCHII regions using IR fine structure line observations. These authors showed, for the first time, that the inferred temperature gradient from the measured metallicity gradient is consistent with that obtained using RRL observations. However, the slope of the electron temperature gradient obtained from data toward UCHII regions is shallower than that determined from data toward classical H ii regions (Shaver et al. 1983; Afflerbach et al. 1996). The difference in slope obtained from the two sets of observations needs to be resolved. Moreover, there is still a considerable scatter both in the temperature and metal abundance gradients. This scatter is partly due to local metal abundance anomalies produced by supernovae and winds from evolved stars, but is also due to measurement errors along with uncertainties associated with the estimation of the distance to the objects. The high sensitivity of SIGGMA will help to improve the estimation of the electron temperature gradient from the data toward H ii regions.

We will also be able to measure RRLs from heavier elements such as carbon. Maps of carbon RRL emission in a variety of sources can be used to study PDRs and to test PDR models. SIGGMA can make a great contribution to the study of PDRs because the Arecibo telescope has the sensitivity to map carbon RRLs in a substantial number of sources having a wide range of metallicities.

SIGGMA can help to improve our understanding of the origin and ionization of the ELDM since, owing to its high spatial resolution and high sensitivity, it will map the ELDM up to latitudes of ∼2°, thus allowing us to associate the ELDM with individual H ii regions.

In summary, we expect the data products and results from SIGGMA to be comparable to those obtained from other surveys such as the Two Micron All Sky Survey (Skrutskie et al. 2006), the Infrared Space Observatory (Kessler et al. 1996), the Midcourse Space Experiment (Egan et al. 2003), the NVSS (Condon et al. 1998), the Galactic Legacy Infrared Mid-Plane Survey Extraordinaire (Benjamin et al. 2003; Churchwell et al. 2009), and the International Galactic Plane Survey,¹¹ as well as the G-ALFA Continuum Transit Survey (GALFACTS; Taylor & Salter 2010) and the ALFA Galactic HI Surveys.

SIGGMA uses the Arecibo L-band Feed Array (ALFA)¹² receiver on the Arecibo telescope. The ALFA receiver has seven independent beams, each recording two orthogonal linear polarizations. The beams are arranged in a hexagonal pattern such that there is one central beam (Beam 0) surrounded by six outer beams (Beams 1–6). Due to the optics of the telescope, the outer beams are projected onto an ellipse in the sky. This ellipse is centered on Beam 0 and has semi-axes of 64 in zenith angle by 55 in azimuth. The orientation of the long axis of the ellipse changes with the parallactic angle. Although the array is derotated during observations, the positions of the outer beams change on the sky due to the elliptical projection. This change of outer beams with respect to Beam 0 is typically less than the FWHM of the beam width.

The average FWHM of the beams varies between 33 and 34 across our frequency range.¹³ The footprint of all the beams on the sky, out to the ellipse enclosing the FWHM of the outer beams, is (153–186) × (136–165), covering an area of 163–241 arcmin². The area falling within the FWHM of the beams is 53–78 arcmin², or around a third of the total footprint. The gain of Beams 1–6 is ∼8.5 K Jy⁻¹, while Beam 0 has a gain of ∼11 K Jy⁻¹. All seven beams have measured system temperatures of 27–28 K at the low zenith angles (≲ 15°) where most of our observations will occur, producing a system equivalent flux density of ∼2.5 Jy for Beam 0 and ∼3.2 Jy for Beams 1–6. The receiver has a bandwidth of 300 MHz, covering 1225–1525 MHz.

Spectra are recorded once per second using the Mock Spectrometer, a Fourier-transform device that has two groups, each of 14 boards. This setup enables the Mock Spectrometer to process data from the seven ALFA beams, each of which is divided into two intermediate frequency (IF) sub-bands. The first group is used in a high time-resolution, low spectral-resolution mode to obtain data for the commensal "P-ALFA" pulsar survey (Cordes et al. 2006), while the second group is used to acquire data simultaneously for the RRL survey. Each IF sub-band covers 172 MHz, with the first IF centered at 1450 MHz and the second IF centered at 1300 MHz. Together, these sub-bands cover the entire 300 MHz ALFA bandpass, with the filter roll-off being either in the overlap region between the two sub-bands or outside the band of the receiver.

For the RRL survey, data were accumulated for 1 s before being Fourier transformed to form spectra with 8192 channels per IF sub-band. This procedure gives a spectral resolution of 21 kHz, equivalent to 4.2–5.1 km s⁻¹ for the Hnα recombination lines within the ALFA bandpass. The required integration time is then built up from multiple 1 s spectra.

The survey uses a "leapfrog"-style observing technique (Spitzak & Schneider 1998) with a grid of points that repeat along lines of constant declination (see Figure 1). Each observation integrates on each position for 270 s (180 s in the outer Galaxy). Combined with the slew time, this procedure gives a spacing of nearly five minutes between observations. By choosing pointings from the grid with the same declinations and separated by five minutes in right ascension, we track the same azimuth and zenith angle in consecutive points. This technique allows consecutive points to be used as ON–OFF pairs in order to form a bandpass-corrected (ON–OFF)/OFF spectrum (see Section 3). The five minute separation in time between ON and OFF points provides the best baselines because ripples caused by internal reflections in the antenna, ground pick up, and atmospheric effects mostly cancel out in the corrected spectrum.

Zoom In Zoom Out Reset image size

The survey tiling pattern is made up of clusters of three pointings that are offset from each other by one beam FWHM beamwidth, labeled as the north (N), central (C), and south (S) pointings in Figure 1. This pattern is a modification of the original P-ALFA pulsar survey tiling pattern (Cordes et al. 2006), with the axis of the pattern (and of the ALFA receiver) rotated by 19 deg from celestial north so that adjacent clusters of pointings fall on lines of constant declination. By leapfrogging through this tiling pattern, the survey covers approximately 10 deg² of sky in 100 hr of telescope time. In the outer Galaxy, where SIGGMA is commensal with the ALFA Zone of Avoidance survey, integration times per pointing are shorter but the pattern is observed three times with a small offset (∼19) in order to produce a Nyquist-sampled map of 10 deg² of sky to a similar depth in 200 hr of telescope time.