UV(oorwerp) from Space

We have some new results to show off, Hanny’s Voorwerp observed using a space telescope. No, not that space telescope, that’s still coming up (shortly, we hope).

Soon after the initial results showed what a fascinating object Hanny’s
Voorwerp was proving to be, it was entered in the observing schedule for NASA’S GALEX satellite (GALaxy Evolution EXplorer). Alex Szalay, who belongs to both the GALEX and Galaxy Zoo science teams, played a key role in making this happen). Alex has interesting career parallels with Brian May, but that’s another story.

GALEX was designed to make the first sensitive ultraviolet survey of most of the whole sky (skipping only areas where there are such bright stars that they would damage the detector array), with a major goal of tracing the recent evolution of galaxies. The ultraviolet spectral range gives a lot of leverage for this, since this is where the most massive and recently-formed stars are brightest (unless they are hidden by dust). Its 50-cm telescope couples to microchannel arrays to show a 1.2-degree patch of sky at once. Not only does it obtain images, in two bands at once (near- and far-UV), but can obtain the spectra of everything in the field of view by inserting a grating-prism combination (grism). They have a very nice sky browser – GALEXView – which will show you the UV-color images for any piece of sky GALEX has observed, and give you object lists and file retrieval options. GALEX observes in orbital night, for about 45 minutes of each orbit, because the tiny residual atmosphere even at its altitude creates a UV glow (which became best known as “Shuttle glow”) when encountering fast-moving objects, and this glow impairs sensitivity just like light pollution on the ground.

With its wide field, GALEX is very good at capturing large objects,and particularly large diffuse objects (such as the Voorwerp). It observed IC 2497 and its surroundings for 1.7 hours in the imaging filters, and for nearly five hours for the spectra. The spectra look odd compared to most we use in astronomy, where we analyze only the light coming through a narrow slit or a tiny optical fiber. Here, the entire field of the telescope is spread out, letting spectra overlap where they may, so big galaxies have big fuzzy spectra where one wavelength from one side overlaps a different wavelength from the other side, and both are overlapped by the light of the “blank” sky at yet a different wavelength. This technique – slitless spectroscopy – is hard to use from the ground because of light pollution and airglow. But in the UV, the background sky brightness in space is so low that the main limitation is confusion as spectra from different objects land atop one another. For these observations, all the data were taken with the spacecraft rolled so that the spectra of Hanny’s Voorwerp would cleanly fit between those of neighboring objects (such as the galaxy IC 2497). Even so, it just barely fit.

Here are some views of the data. First, a color UV image (where blue means brighter in the far-UV, such as hotter stars). Compare that to the SDSS image of the same area. The GALEX image isn’t as sharp (what you get from covering such a wide area at once with a single detector), but you can easily see how different the UV sky looks from the visible-light appearance. For example, the spiral galaxy to the southeast has really blue arms, and its central bulge almost vanishes in the ultraviolet. Stars come and go between the images depending on temperature. More to the point of our story, Hanny’s Voorwerp is brighter than IC 2497 at these wavelengths. We knew that much already – early on, the Swift satellite observed it in both UV and X-rays, giving us a somewhat sharper view at one UV wavelength.

IC 2497 field in UV with GALEX

IC 2497 field from SDSS matching GALEX image

Where does all that UV radiation come from? That’s the point of observing the spectrum – it will tell whether that comes from ionized gas, light from the AGN in IC 2497 scattered from dust grains mixed with the gas, or maybe even stars that used to belong to a galaxy which has been shredded by a gravitational encounter with IC 2497. The UV spectra aren’t much to look at compared to the optical data, but they can be informative. Here are the far- and near-UV spectra as they look when the GALEX data are all added together:

GALEX far-UV spectrum of Voorwerp

GALEX near-UV spectrum of Voorwerp

Each object gives a spectrum making a horizontal streak. The spectrum of IC 2497 shows up just above the Voorwerp’s, and the blue galaxy to its southwest spreads just to its south in the spectra. We could see right away that a couple of strong emission lines show up in the far-UV data, smeared out by the large size of the Voorwerp. To get some numbers, we folded in the wavelength corresponding to each location and the sensitivity of the GALEX system at each wavelength, to get this
composite spectrum from both data sets:

GALEX UV spectrum plot of Hanny's Voorwerp

We’re missing some of the wavelengths at the ends of each piece, where the sensitivity of the system is so low that statistical “noise” is amplified and even minor contamination from surrounding objects makes a huge difference in the result. But we do see some of the emission lines accompanying gas ionized by quasars and their kin, which are familiar to Zooites who look at high-redshift quasars. They are marked
with their expected locations at the redshift of Hanny’s Voorwerp. C IV and He II are unmistakeable, each blurred by the size of the emitting cloud. An optical line of He II
at 4686 A was one of the things that put us on the trail when we got the first optical spectra, and in the UV there is a stronger one. This stronger line, emitted at 1640 A, plays the same role for He⁺⁺ that Hα does for ionized hydrogen, at 1/4 the wavelength because the binding energy of a single lone electron depends on the square of the charge in the nucleus (2 instead of 1, comparing H and He).

It’s also interesting that the spectrum doesn’t go to zero between the strong emission peaks. Sources of this continuum could include the recombining hydrogen (except at the very short UV end), dust grains scattering the light of the AGN toward us, or young stars in the Voorwerp for example, if its gas originated in a galaxy disrupted by a tidal encounter). I don’t think these data let us distinguish yet how much of each there might be, but its brightness in the UV bodes well for seeing lots of interesting detail in a Hubble ultraviolet image. More shortly on the Hubble data!