Guanying Zhu

 Nanjing University
 REU Program -- Summer 2014
 University of Wisconsin - Madison, Madison, WI 53706


  Studying Galaxy Evolution with Unique State-of-the-Art 3-D Spectrographs


Advisors: Dr. Eric Hooper & Dr. Marsha Wolf


  It has been widely accepted that galaxy merger can be a likely mode of galaxy evolution. Theoretically(Figure 1), when two nearly equal mass gas-rich galaxies coalesce, a large amount of gas flows toward the center, which can initiate a nuclear starburst and feed the black hole, so at this point the galaxy is in an optically obscured AGN phase. Then, due to feedback from supernovae of stars formed during the burst and/or the black hole accretion, the obscuring materials can "blow out", which makes the accreting black hole appear as an observable optical quasar. After that, as leftover gas is consumed and dispersed, star formation and quasar activity decline. Since the remaining gas is soon used up, future star formation and black hole accretion will be quenched on the one hand. On the other, galactic disc formation is almost impossible without a gas supply. Eventually, the remaining galaxy fades to a red elliptical.

Fig.1 galaxy evolution theory flow chart

  However, there are still something uncertain about this process. For example, how AGN feedback quenches the star formation? Is it AGN feedback or feedback from supernovae of stars that quenches the star formation? What is the time scale of feedback and star formation quenching? In order to put some constraints on this process, we chose a particular phase in galaxy evolution, which interests many astronomers, post-starburst phase. It is a phase between starburst and final "red and dead" phase, which means almost every galaxy that evolves from an active phase to quiescence must pass through it. I think it is just like a remnant of an intense activity, so we may find some clues about the whole evolution, such as major merger parameters, gas consumption, AGN formation and duty cycles through the careful study of this particular phase.

  During this summer, we hope to put some observational constraints on this evolution theory based on the study of post-starburst galaxies with weak variable radio signals. More specifically, in terms of stars, we can derive stellar population ages by fitting optical spectra, which will tell the timing of the starbursts in the galaxies; in terms of AGN, we can estimate the age of the source with the radio synchrotron spectra. Working with Dr. Eric Hooper and Dr. Marsha Wolf, I mainly focused on the optical part. The object I studied was G515(SDSS J152426.50+080908.0), and we basically believe that G515(Figure 2) is an archetypal post-starburst galaxy. Its tidal tail and extended crescent show that it is very likely that G515 has formed from galaxy mergers. No OII or H-alpha emission shows that there is no on-going star formation, but we have detected a variable point source radio signal which may come from a weak AGN hidden in the galaxy. In addition, its spectrum in SDSS(Figure 3) is also consistent with that of a post-starburst galaxy -- superposition of an early type galaxy's spectrum and deep Balmer lines from A stars. All of these make G515 become an ideal object to study galaxy evolution.


Fig.2 G515

Fig.3 G515 spectrum in SDSS

  The data I analyzed were taken using WIYN 3.5m telescope at the Kitt Peak National Observatory in May 2014. The instrument we used to obtain spectra is called as Integral Field Spectrograph (IFS). In general, an IFS consists of two parts: an Integral Field Unit (IFU), and a spectrograph. We put a bundle of optical fibers on the focal plane, and those fibers can transfer the light from object to the slit. Then the slit will direct the light to the spectrograph, so finally we can obtain the spectra of the whole object at the same time(Figure 4).

Fig.4 IFS

Credit: M. Westmoquette, adapted from Allington-Smith et al. 1998

  Here comes a question, how to design those fibers? It depends on what kind of spectra we want, and I think it is obvious that we hope the spectra have a high signal-to-noise ratio and a high spatial resolution. However, it is hard to balance, because a high spatial resolution means we have to use small fibers, while small fibers cannot collect many photons, which means we may have a low signal-to-noise ratio. Fortunately, if we already know the basic structure of the objects we plan to observe, we can design an IFU to match. At WIYN telescope, scientists have developed new IFUs called as HexPak and GradPak(Figure 5), which are the first IFUs to provide formatted fiber integral field spectroscopy with simultaneous sampling of varying angular scales. Those two IFUs are in a single cable with a dual-head design, so you can switch them without changing the spectrograph feed: the two heads feed a variable-width double-slit. The layout and diameters of the fibers within each array are scientifically-driven for observations of galaxies: HexPak is designed to observe face-on spiral or spheroidal galaxies while GradPak is optimized for edge-on studies of galaxy disks, that is, the fibers sizes increase as surface brightness decreases. In that way, we can gain a higher resolution but avoid losing signal-to-noise ratio too much.

Fig.5 HexPak and GradPak

  It is really exciting that we could use the new IFU, although for data reduction, it is more challenging! G515 is a face-on elliptical galaxy, so we used HexPak. To better understand how powerful new technology is, please see a comparison between Hexpak and an old IFU--SparsePak in Figure 6.

Fig.6 comparison of HexPak(left) and SparsePak(right), both of them are located in G515

Back to Top