meeting 2025 05 07 n410 - JacobPilawa/TriaxSchwarzschild_wiki_6 GitHub Wiki

Context

  • We've started to work on some of the new galaxies, starting with N410, so here are some preliminary diagnostics for this galaxy.

High Level Information

  • Stats:

    • D = 75.5, Jensen+21 SBF
    • Brightest group member, 29 group members
    • sigma_e = 247 km/s
    • Mk = −25.90, log10(Mstar) = 11.86
    • eps = 0.25, lambda = 0.03 (Slow rotator)
    • log10(Mhalo) = 14.4
    • MASSIVE XIV: Note that MASSIVE XIV says this galaxy is aligned, whereas MASSIVE X says it is misaligned
      • PA GMOS = 311 +/- 22
      • PA Mitchell = 161.0 ± 18.5
      • PA Phot = 34.9 (MASSIEV XIV: 35.8 ± 0.9)
      • Psi GMOS = 4.8 ± 9.3 (MASSIVE XIV)
      • Psi Mitchell = 53.9 (MASSIVE X)
      • Kinematic Twist, with note about low S/N
    • Photometry:
      • has disky isophotes between ∼10–30 arcsec. Outside of this range the isophotes go back to B4≈0. The ellipticity also peaks at ∼30 arcsec, beyond which it is approximately constant

  • GMOS:

    • All files seem to be on BOX.
    • Has both symmetric and non-symmetric binning schemes.
    • Other details:
      • Files on BOX say:
        • +X = 148 deg. E of N. in orbit modeling files
        • +X = +X = -32 deg. E of N. in paperversion figures of kinematic moments
      • PSF listed as: avg. FWHM = 0.39" (Gaussian)
      • 128 bins (symmetric)
      • 131 bins (non-symmetric)

  • Mitchell

    • All Mitchell files for all galaxies are present! There are folded and unfolded versions for everything. I'm currently sticking with our usual unfolded data.
    • 64 bins (unfolded)

Some fiducial fit comparisons

  • I've run four different cases that I've been comparing. For the time being I've called them Case A, B and C.

  • Case A: treating this as the fidcucial case

    • z = 0.017359 (from NED)
    • 8420-8770 Angstrom fitting region
    • Mask from BOX
    • 15 Bart Stars
    • nMc = 100
    • nMoments = 8
    • Bias = 0.2
    • Starting guess = [0, 200] km/s
    • (adeg, mdeg) = (0, 3)
    • Edges masked (9 pixels from the left and right boundary)

  • Case B

    • Same settings as Case A except I've switched bias from 0.2 to 0.0

  • Case C

    • Same settings as Case A except I've switched the additive polynomial off (adeg=-1).

  • Case D

    • Same settings as Case A except I've used the full CAT library instead of the Barth stars alone.

  • Takeaways:

    • Case A and Case B are in excellent agremeent! Case A and Case C however show quite a bit of tension, but in general the Case A spectra are better fit than the Case C spectra, so I tend to believe the Case A/Case B spectra bit more.
    • Opening the fits up to the full CAT library significantly change the kinematics, but the spectra don't look all that great; they're quite "wiggly", and I suspect pPXF is selecting a strange group of stars as the "best fitting." I don't think it's worth too much effort here since the results are so all over the place, and we're quite confident that the Barth stars are appropriate for N410 (in terms of stellar populations).
    • For the plots below, I'll focus mostly on Case A/B/C -- I didn't run the full nMC=100 for the full CAT library case since it takes quite a bit longer to use the full library. I'll include a comparison of the preliminary moments, however.

Diagnostics

  • First, here are the spectral fits for Cases A/B/C:
Case A
Plot
Case B
Plot
Case C
Plot
Case D
Plot
  • Here's a comparison of the RMS for each of the 167 spectra:
Case A Case B Case C Case D
  • And here are some plots comparing the moments:
Case A vs. Case B Case A vs. Case C Case B vs. Case C Case A vs. Case D
Prelim
MC N/A
  • And here are some kinematic diagnostics for the three cases:
Case A Case B Case C
(UNSYM) Radial Moments
(UNSYM) GMOS Maps
(SYM) Radial Moments
(SYM) GMOS Maps

Case A

  • I've also plotted combined GMOS and Mitchell maps for Case A, which I've been using as my "fiducial" case so far:
    • The plot from Irina are from the unfolded "paper version" of the Mitchell kinematics and the
Full Extent Zoomed
Unsymmetrized
Symmetrized
  • It's useful to compare these to the data that's on BOX, but there are a few caveats:
    • The GMOS data (I think) uses the symmetric binning scheme, whereas I have been using the asymmetric binning scheme for the plots above.
    • These appear to be non-point symmetric, so these should be compared to the raw data.
GMOS BOX Data Mitchell BOX Data

Some symmetric-binning scheme testing

  • Irina/Melanie had been using the symmetric binning scheme for GMOS it seems, and I wanted to try to run pPXF over these spectra/binning scheme to reproduce their results. I can't seem to find the kindata.dat file for their final kinematics, but I've processed their spectra to the best of my ability and the kinematics seem to be in decent agreement.

MGE Preliminary Fitting

  • I don't have a lot done for this yet, but am working on the MGE now that we have some kinematic diagnostics to work with. Just putting these up here for now but will post more diagnostics soon:
    • There's clearly some residual structure here, potentially from the region that I'm using (currently using the full image), and potentially due to the lack of a mask (fitting everything in the image right now).
    • This also places no restrictions on the MGE parameters, so I suspect the deprojected profiles for this will have a "bump" that I'll investigate ASAP.
Plot
  • The parameters for this first fit are:
  I [Lsun/pc^2]        Sig [arcsec]         q            PA 
        16.9822      0.0370118        0.99611              0
        19.0906       0.363766              1              0
        20.5022       0.694563        0.99222              0
        12.8873        1.26092       0.929582              0
        5.39231        2.37984       0.828874              0
        2.76644        3.97886       0.910116              0
       0.713066        7.51758       0.416106              0
        1.35849        8.94864       0.818632              0
       0.337729        18.9346       0.532158              0
       0.319275        24.6935       0.766123              0
      0.0859009        47.7424       0.763851              0
      0.0222079           87.5       0.567236              0

A few side notes

Case D Strangeness
  • I wanted to check out why Case D was so strange by checking the stellar templates that are favored.
    • It seems like opening up to the full CAT library results in a very different template distribution (in fact, none of the Barth Table 2 stars are actually selected).
Case A Total Templates Case D Top 15 Templates

  • Taking a look at a few of the Case D preferences:

    • Scan156 = M3III Red Giant
    • Scan556 = K4III
    • Scan696 = Spectroscopic Binary
    • Scan249 = A3Vn
    • Scan200 = Spectrocopic Binary
    • Can check others, but I think there's quite a mix of random stars in here that we would not want to include.

  • All of this is to say that I think including the full CAT library is overfitting our data. We end up getting strange template preferences, even if the RMS of the fits is lower. I think this because there's simply way more flexibility to fit finer features in the spectra.

⚠️ **GitHub.com Fallback** ⚠️