Research Overview

Steven Gibson, Western Kentucky University

The stars that light up the visible universe are each condensed from clouds of hydrogen gas and other trace elements within their host galaxies. This process is ongoing but mysterious: how can ambient interstellar matter -- more tenuous than any vacuum we can create in an Earthly laboratory -- become concentrated enough for its own gravity to pull it together into stars? To answer that question, we need to observe this transition, which probably begins in cold diffuse clouds (CDCs) whose free atoms are moving just slowly enough to start joining together as molecules. CDCs are largely invisible to the human eye but can be detected with radio and infrared telescopes. However, they are less conspicuous than other types of interstellar clouds; their cold atoms and still-forming molecules radiate so weakly that they are sometimes called "dark gas". To identify CDCs and determine their properties and origins, we must find ways to bring them out of the shadows.


Figure 1:
Interstellar matter observed in one section of the Galaxy, including total mass from dust infrared heat radiation, neutral atomic hydrogen from radio spectroscopy, molecules from shorter-wavelength radio waves, and "dark gas" implied by the difference between the dust and the other two.

At Western Kentucky University, we have assembled a large body of radio and infrared survey data covering much of our home Galaxy, applying novel analysis methods to uncover populations of cold interstellar clouds and to assess their physical stages of development. These surveys focus on areas near the plane of the Galactic disk, where the majority of suitable interstellar material is found. CDCs should contain mixtures of neutral atoms and simple molecules that are normally observed with radio spectroscopy of neutral atomic hydrogen (HI) and carbon monoxide (CO) emission lines at wavelengths of 21 centimeters and 2.6 millimeters, respectively. In addition, about 1% of the cloud mass is in small, solid "dust" particles that absorb starlight, heat up, and glow at infrared wavelengths. With some care, the dust emission can be used to track the total amount of material present in the clouds, including any dark gas not evident in HI and CO emission maps (Figure 1).


Figure 2:
Face-on view of a model of our Galaxy's spiral arms, where most star formation occurs. Left: Major arms and lines of sight from Earth corresponding to Galactic longitudes. We are 25,000 light years from the center. Right: Detailed radio surveys of neutral atomic hydrogen gas.

We use several HI surveys covering most of the Galactic plane region: the Canadian, Very Large Array, and Southern Galactic Plane Surveys, the Australia Telescope Galactic Center survey, and the Galactic Arecibo L-band Feed Array survey (CGPS, VGPS, SGPS, ATGC, and GALFA; Figure 2). CO spectral data were taken from the original and extended Outer Galaxy Surveys, the Galactic Ring Survey, the University of Massachusetts-Stony Brook survey, and the Harvard-Smithsonian Center for Astrophysics composite survey (OGS, EOGS, GRS, UMSB, and CfA). Dust emission maps were obtained from the Infrared Astronomical Satellite (IRAS) and Planck all-sky surveys, with the latter also providing supplementary CO maps.

The cold and dark atomic gas in many CDCs should appear as HI self-absorption (HISA) against warmer, brighter HI emission behind them (Figure 3). HISA features look similar to denser molecular clouds, and some have matching CO emission; others may have dark molecular gas in which CO is still forming. Searching through the HI surveys noted above, we find HISA wherever backgrounds are bright enough to absorb against, but with enhancements where gas is predicted to pile up in spiral arms prior to new star formation (Figure 4). We also find that HISA positions are more likely to have CO in the inner parts of the Galaxy than in areas outside the Sun's orbit, but when chance cloud alignments are accounted for, most HISA lacks matching CO, regardless of location (Figure 5). This suggests we are capturing different stages of cloud evolution, but how can we tell if dark molecular gas is present or not?


Figure 3:
Neutral atomic hydrogen (HI) line emission (blue) with carbon monoxide (CO) molecular emission (magenta contours) in a map (left) and in spectra (right) from the marked positions. Drops in HI brightness are HI self-absorption (HISA) from cold HI in front of warmer HI emission (bottom).

Figure 4:
HISA found across the Galactic disk using the surveys in Figure 1 and extracted with custom software (GALFA data pending). HISA detections are plotted vs. longitude and Doppler velocity, which relates to position in the Galaxy; contours are bright HI emission (green) and CO (magenta).

Figure 5:
Fraction of HISA clouds containing measurable CO, plotted vs. longitude and for velocity ranges corresponding to different parts of the Galaxy. Colors are for different HI surveys. Most HISA clouds have no apparent CO, but many may still have significant dark molecular gas.

To decide such issues, we developed a new method to map HISA cloud properties via combined analysis of the gas and dust data, finding temperatures, densities, and atomic/molecular mixes in the ranges expected for clouds making this transition, with significant amounts of dark atomic and dark molecular gas both present (Figure 6). In parallel, we devised a new way to identify CDCs without bright HI backgrounds from their HI emission line shapes and analyze them with CO and dust data in the same manner as HISA features to map their properties. These combined approaches present a picture of CDCs as indeed the probable stepping stones between ambient interstellar material and star-forming molecular clouds.


Figure 6:
Maps of gas temperature (left) and density (right) vs. location in a number of HISA clouds stretching over a 1500-light-year section of the Perseus spiral arm, determined from combined analysis of the HISA features, CO images, and dust emission.

The initial analysis of the cloud dust content was somewhat simplistic, but we have subsequently pursued more detailed investigations fitting the dust infrared emission with models over a broad range of wavelengths for improved constraints on the cloud properties, which shows great promise (Figure 7).


Figure 7:
Infrared dust emission of the clouds in Figure 3. Left: images at wavelengths of 8, 24, 60, 140, 250, and 850 microns from Spitzer, IRAS, AKARI, Herschel, and Planck prior to background emission subtraction; ON and OFF photometry boxes are marked. Right: Spectral energy distribution (SED) at the cloud center from all available survey bands, including WISE and MSX. Vertical bars indicate photometric uncertainties; horizontal bars mark the wavelength range of each band. The curves show a trial fit of the Li & Draine (2001) emission model and grain component contributions. The lower plot gives fit ratios as residuals.

Products of this work, including maps and analysis tools, will be made publicly available in support of future Galactic survey efforts. This research program has also helped a number of students prepare for scientific and technical careers by giving them real research experience while enrolled at a comprehensive, teaching-intensive, and primarily undergraduate institution in an economically disadvantaged region of the United States. WKU students presented over two dozen talks and posters on different aspects of the work at local, regional, and national conferences, winning awards at each level. A total of 10 students have been actively involved in this research, with 5 pursuing subsequent graduate studies.

Complementing these research efforts with professional facilities and surveys around the world, we have pursued parallel development of local radio astronomy facilities where students can gain real hands-on experience that will help them prepare for scientific and technical careers. This began with Radio JOVE but has branched out into other areas, including a 10-foot dish telescope under current development.


Steven Gibson