Hubble, meet Galaxy Zoo. Galaxy Zoo, meet Hubble.
Regular blog readers will know that we were all hugely pleased to find out that our proposal to observe Hanny’s Voorwerp with Hubble was approved. This was especially welcome because we expected a very high oversubscription rate for next year – new and repaired instruments meant that there was pent-up demand for some kinds of observations which have not been possible for several years. Nearly 1000 proposals were submitted to the Space Telescope Science Institute (STScI). which managed a complex review process involving about 200 astronomers from all over the world (noting that Hubble is a cooperative project of NASA and the European Space Agency). Specialized panels of reviewers looked at various subfields of astronomy, comparing the likely scientific fruitfulness of a wide range of projects.This last week saw the deadline for the next step in preparing for next year’s Hubble observations – what’s known as Phase II. This uses software distributed by STScI to plan each operation in detail – every exposure, filter change, and minute telescope motion. The astronomer can find out whether reordering certain operations uses precious telescope time more efficiently, and whether the results can be improved by restricting the observations to certain orientations of the telescope or times of year. The software will also overlay requested fields of view on sky surveys such as Sloan images), a welcome reality check that you’ve told it to look in the right place. This stage also gives us a chance to see whether anything we’ve learned since the proposal was submitted in early March gave us reason to change any of our originally proposed measurements.
By the way – “we” for this project has to mean broadly the whole Galaxy Zoo community. For the Hubble project, the other names included were Chris, Kevin, and Daniel from the Zoo core team; Nicola Bennert from the University of California; and Hanny herself. You can see the “phase II” scheduling results here.We proposed three kinds of observations, totalling 7 orbits. (The usual currency on allocating Hubble time is the viewing period during one orbit. Of each 94-minute orbit of the Earth, at least 54 minutes is available to look at a target depending on its celestial coordinates – which means that Hubble spends much of the rest pointed at the Earth with the shutter closed or doing internal instrumental calibrations). Each of our observations has specific goals in testing our ideas about Hanny’s Voorwerp (although we always hope for surprises, and the Universe seldom fails to oblige).In this case, we want to use three instruments to collect specific kinds of data. First up, we want to get a really close look at the nucleus of the spiral galaxy IC 2497, which is the most likely source of all the excitement (that is, where we suspect a quasar blazed until yesterday on the cosmic calendar). We’ll do this using the Space Telescope Imaging Spectrograph (STIS), which observes the spectrum along a slit about 50 arcseconds long. While we’ll be looking in the optical part of the spectrum, Hubble’s resolution allows it to isolate a small area around the core free of contamination by starlight from the large surrounding area (as happens when observing from the ground). In this way, we get a more accurate picture of what the nucleus is doing “now” (well, actually 700 million years ago when we include light travel time, but the verb tenses get complicated). We will look from any gas which sees the kind of high-energy radiation shown in the spectrum of the Voorwerp. The same data will also tell us how rapidly the stars are moving around in the galaxy core, by measuring how much the spectral lines are broadened by the stars’ various Doppler shifts, and thus whether it has an unusually massive central black hole (which we might expect if it was recently growing in a quasar episode). The galaxy is bit distant to get a solid measurement of this, but since we’re taking the data for other reasons anyway, it makes sense to include the possibility. Starting with these science goals, we have choices about the instrumental setup -what choices of diffraction gratings and spectral ranges, what slit opening, how do we point the telescope to center precisely at the core? The point of the spectra is to isolate smaller regions than we can from the ground; the spectrograph incorporates slits as narrow as 0.1 arcsecond, but that coms at a penalty in the amount of light that gets to the detector and requires a longer pointing process. Based on several colleagues’ experience, we chose the 0.2 arcsecond width, which will still isolate a central area about 25 times smaller than typical ground-based data. We still have to let the instrument tell the telescope exactly where to point – even with all the care that went into the Sloan survey’s positional accuracy, the SDSS data may not tell where the galaxy core is to 0.1 arcsecond. his is a common issue, one dealt with by a standard command to have the instrument take a short-exposure acquisition image and offset the telescope so that the spectrograph slit crosses its brightest point. We even have choices on how to define that brightest point, and could opt for one more likely to get it right if the core is slightly asymmetric or has adjacent dust absorption. Once the instrument is pointed at the galaxy core, we can start work. Since the spectrograph uses a slit, we get information along its full 50-arcsecond length. Unfortunately, we can’t align the slit in any direction we want – it’s fixed in the telescope and the telescope’s orientation at a particular time and direction is fixed by needing to keep the solar arrays face-on to the Sun. Looking toward Leo Minor, there are almost no opportunities to include the Voorwerp or the odd blob just off the nucleus of IC 2497. This visualization shows a typical orientation and the length of the slit. Anyway – we will use single-orbit observations with two grating settings, one for the blue and one for the red. Each orbit will give us 45-50 minutes’ worth of observations. Together these cover about the same slice of the spectrum as the Sloan data. The emission lines are especially important, since they can tell so much about what kind of radiation lights up the gas.Next, our plan turns to the Voorwerp itself. The Advanced Camera for Surveys (ACS) includes special filters known as ramps. These have coatings which vary the transmitted wavelength with position, so we can get a narrow band at any desired wavelength to map the ionized gas in exquisite detail. We proposed to observe the [O III] line and H-alpha, whose ratio gives the strength of the ionizing radiation. We might expect to see annular structure in this ratio if there were changes in the illuminating source (although the distribution of gas in depth will factor in to what we actually see). The fields of view where we get the desired wavelengths are small and oddly shaped due to the way the filters are mounted, so we had to carefully select the pointing position and restrict the telescope orientation. It turned out to be possible to always include the Voorwerp and the inner part of IC 2497, as shown here: Finally, we address stars and dust in the Voorwerp. With our current data, although it is likely that the Voorwerp was once a dwarf galaxy, we can’t actually detect any stars in it. One way to do this is looking in the deep red or infrared, where light scattered form dust grains is weakest, to see whether we see any star clusters. To do this, we plan exposures with different detectors in the new Wide-Field Camera 3 (WFC3). Since the detectors are vulnerable to cosmic-ray impacts, which show up as artificial bright spots, it has always been standard practice to take multiple images to reject them. As software has improved, the default procedure now is to take multiple images with small motions in between (dithers), which can be combined to eliminate changes in hot pixels as well as cosmic rays. The same procedure can sometimes be used to improve resolution somewhat if the dither motions encompass fractional pixels, For these images, we will get data at 2-3 dither positions. Our final image also uses WFC3, now taking advantage of its improvement in UV sensitivity over its predecessors. We’ll get a wide-band ultraviolet image. Based on the brightness we seeing the UV image from Swift, most of the light in this part of the spectrum is likely to be reflected (or more pedantically, scattered) from dust particles, so we should see the dust structure almost by itself. The software also shows us when observations might be scheduled, combining the telescope constraints on sun angle, power, and pointing. Here are our windows for each observation: This means that we should expect data lat in the spring of 2009. There was a time that all this had to be done by hand with a pile of manuals about 10 cm thick. This is better… The alert reader will notice that of the three instruments we are scheduled to use on Hubble, one is still in a clean room on the ground, while the other two are in orbit but non-functioning and awaiting electronic repair. The next scheduled year of Hubble observations , known as Cycle 17, will start only after the final servicing mission to be carried out by the crew of the space
shuttle Atlantis . Designated STS-125, this mission is now slated to occur in October 2008. (More information can be found here, here, and here. (This will be Servicing Mission 4, although this is the fifth visit made by Shuttle crews to work on Hubble. A gyro failure led to breaking mission 3 into 3A and 3B, which was the final completed mission of Columbia). Like previous missions, it will be followed by a period of Orbital Verification (OV) in which the telescope andinstruments are recommissioned for full scientific function. At its end, the OV period interleaves with the first cycle 17 regular observations.And this isn’t all we have planned in trying to understand Hanny’s Voorwerp. We have a proposal pending to look with NASA’s Chandra X-ray observatory, which would be able to spot weak activity in the nucleus of IC 2497 even if hidden by vast amounts of intervening gas. Their review panels met only a couple of weeks ago, so it may be some time yet before we find out its fate. Meanwhile, the service-observing program at La Palma has gotten spectroscopic mapping at the 4.2m William Herschel Telescope, in a rather complicated data format that we’re just starting to work with.