Radio Galaxy Zoo Talk

ARG0003ojf - can anyone explain the weird spectrum?

  • JeanTate by JeanTate

    In SDSS, the object which is at the same location as the peak of the radio contours (±a pixel or two), SDSS J101927.37+022521.3, is a photometric STAR:

    enter image description here

    It has a spectrum, which looks really weird to me; the redshift is given as 0.943. Here's the squished PNG version of the SDSS spectrum:

    enter image description here

    The NED entry gives the classification as BAL QSO (broad absorption line quasar), and its redshift as 1.3643 (and a reference which classes it as "FeLoBAL").

    What is its redshift? What are the unmarked lines (and other prominent features) in the spectrum?

    Posted

  • ivywong by ivywong scientist, admin

    Hi Jean,
    I'm not an optical spectra expert but I found a very nice introduction written by Duncan Farrah here in one of his 2010 papers: http://arxiv.org/abs/1005.3540

    Does this help?
    Ivy

    Posted

  • JeanTate by JeanTate in response to ivywong's comment.

    Thanks Ivy.

    This paper is indeed good, but for mid-IR spectra, not optical (restframe ~200-500nm in this case).

    It seems Becker+ 1997 was (one of the?) first to identify FeLoBAL QSOs as a distinct class of LoBAL QSOs; apparently the relevant region of their spectra is awash with broad absorption lines due to Al II, Al III, Mg II, and Fe II (and more), and these lines are very distinctive.

    However, the most extraordinary (to me) features in the SDSS spectrum of SDSS J101927.37+022521.3 do not seem to be explained in any of the papers I've read so far. 😦

    What are they?

    The wild swings between ~620-720nm (observer frame); by comparison, the FeLoBAL species blue-ward of the Mg II 2799 line are very tame (assuming z=0.943).

    Posted

  • ivywong by ivywong scientist, admin

    I think that you found a source with strong Fe lines. See the following paper about quasar spectra in SDSS:

    http://iopscience.iop.org/1538-3881/122/2/549/fulltext/

    Posted

  • JeanTate by JeanTate in response to ivywong's comment.

    Thanks! 😃

    An even more interesting paper also has 'our' Amanda Bauer as a co-author, Hall+2002 "Unusual Broad Absorption Line Quasars from the Sloan Digital Sky Survey". It looks specifically at a number of QSOs, among them SDSS J112526.12+002901.3. They find its redshift to be 0.8654, but the SDSS spectroscopic pipeline gives 1.0162 (albeit with the warnings SMALL_DELTA_CHI2 and NEGATIVE_EMISSION). There are some nice similarities with SDSS J101927.37+022521.3's spectrum:

    enter image description here

    So, my initial suspicion that z=0.9434 is wrong, for 'my' QSO were strengthened (notwithstanding the 'No Warnings').

    A couple of hours or so* later, I became pretty convinced that the redshift is likely close to 1.364, though I had great difficulty fitting the various lines and troughs between ~620-720nm (observer frame).

    By now, readers with more sense than me will be laughing their heads off ... why didn't I simply check out NED?!? Had I done that, I would have had those couple of hours to do other things ... because in Trump+ 2006 the redshift of 'my' QSO is given as ... wait for it ... 1.364! 😮

    But no, it wasn't a waste of time; I was able to independently estimate that the redshift is ~1.364, and also to strengthen my intuition when it comes to 'wrong' SDSS redshifts. 😄

    Anyone interested in how I, armed with nothing more than the SDSS webpages and a few cells in a speadsheet, concluded that z=1.364?

    *actually surely more like three than two

    Posted

  • ivywong by ivywong scientist, admin

    Great work! That's good detective work. Yes, automated algorithms are known to fail on occasions so it's good to have independent confirmation.

    There is a free piece of analysis tool called IRAF that many astronomers use. Within it, if you install the "rvsao" package, you'll be able input any spectrum and it'll try to fit it to known templates of lines. Unfortunately, this piece of software is not for the faint-hearted as it was not made to be user-friendly. But if you're curious, you can find it at: http://iraf.noao.edu/

    On the other hand, a spreadsheet works just as well so it's up to you..... 😉

    Posted

  • JeanTate by JeanTate in response to ivywong's comment.

    Thanks.

    Actually, much more fun is teaching yourself Python, downloading and installing an endless series of files, learning a dozen or so packages/modules, trying to understand FITS and how SDSS spectra FITS files work, then writing the code, making a bazillion mistakes, ... and finally producing this*:

    enter image description here

    x-axis is restframe wavelength, in Angstroms; y-axis flux, in units of 10-17 erg/sec/cm2/Angstroms. The black plot is the DR7 spectrum of SDSS J101927.37+022521.3, transformed assuming z=1.364; the blue plot the DR7 spectrum of SDSS J112828.31+011337.9, a different FeLoBAL in Hall+ 2002, with z=0.831, and shifted up by 5.0 (so you can compare the two plots more easily).

    Here's a piece zoomed in:

    enter image description here

    The [OII] doublet (blended) 3727, 3729 is perhaps the most obvious line; [NeIII] 3869 is noisy, but unmistakable. Perhaps even OIII 3445.

    *yeah, I have some work to do mastering the finer points of matplotlib 😉

    Posted

  • JeanTate by JeanTate in response to JeanTate's comment.

    I have some work to do mastering the finer points of matplotlib

    FeLoBAL QSOs are really wild! 😮

    Here's a v2 of my 'comparison, restframe spectrum', taking an object from Hall+ 2002 ('0338+0056') and the redshift given there (1.627):

    enter image description here

    The legend is off to the right, the image too big to display (and v2 Talk has no stable way - that I know of - of resizing); blue is 0338+0056 (DR7 ObjId 588015510366978288), offset vertically (by 5.0 flux units) so you can compare more clearly, and black is ARG0003ojf. Here's the region around Mg II:

    enter image description here

    Still some fine-tuning to do, but it's pretty obvious that analyzing the spectra of FeLoBAL QSOs is not for suitable for ordinary zooites (and small children).

    Posted

  • ivywong by ivywong scientist, admin

    I think FeLoBALs are an active area of research but you probably know more that I do about them at this point. Very nice mastering of Python! Definitely a very worthwhile language to learn.

    Posted