Integral Field Spectroscopy of Collisional Ring Galaxies I: Stellar Populations Analysis

Abstract

Collisional ring galaxies are produced by the collision of a disk galaxy with a compact galaxy plunging through the disk, forming a ring-shaped expanding density wave, triggering star formation at its wake. The wave expansion is expected to produce negative stellar age gradients in radial profiles of post-collision stellar populations. Integral field spectroscopy combined with stellar population synthesis allows us to spatially resolve the stellar populations, to separate the post-collision and pre-collision components, and to produce the radial profiles. We analyse three candidate galaxies: Arp~143, NGC~2793, and VII~Zw~466. Observations were performed with the Calar Alto 3.5~m~telescope using the PMAS/PPak spectrophotometer. NGC 2793 presents a positive stellar age gradient, dismissing the collision hypothesis. For Arp~143 and VII~Zw~466, we found negative stellar age gradients for the youngest stellar populations, up to the ring radii, consistent with the collision hypothesis. We estimated that the collisions occurred $\sim$300~Myr and $\sim$100~Myr (expansion velocities of 33~$\pm$~10 km s$^{-1}$ and 108~$\pm$~26 km s$^{-1}$), respectively, before the density waves reached the observed ring radii. A spatially resolved analysis of the specific star formation histories (sSFH), reveals an expected star formation enhancement following the collision. The sSFH also allowed to identify the most probable intruder galaxy for VII~Zw~466. We report new redshifts for its group members. Finally, radial profiles of light contributions from pre-collisional and post-collisional stars show that the density wave dragged old pre-collisional stars along, as predicted by simulations.

Figures (12)

• Figure 1: Representative spectral fits for three spaxels in Arp 143. The locations of these spaxels, labelled as A, B, and C, are indicated in Figure \ref{['fig:age_map']}, top-left panel. The flux of each spectrum is normalized at $\lambda = 5100$ Å. The observed spectrum is plotted with a solid-black line and the synthetic spectrum with the best-fit solution is plotted with a red-dashed line. The individual SSP spectra contributing to the solution are plotted as solid coloured lines. Each of these lines represents an SSP spectrum of a given age, integrated over the six metallicities, and scaled by its corresponding population vector component, $x_j$. The colour of each line corresponds to the age of that stellar population, as indicated by the colour bar at the bottom. In each panel, the central inset legend lists the ages of the three SSPs that most significantly contribute to the synthetic spectrum (the Top 3 components). The solid gray line represents the sum of these three SSPs. The shaded regions indicate the spectral intervals that were masked out during the STARLIGHT fit to exclude emission lines. The goodness-of-fit parameters, $\chi^2/$ndf and the RMS, are shown in the top-left corner, along with the light-weighted and mass-weighted mean stellar ages for each spaxel, computed using Equations \ref{['eq:xmean_age']} and \ref{['eq:mmean_age']}
• Figure 2: Left panels: RGB images of the observed galaxies produced by extracting the BVR Johnson synthetic bands from the IFS data cubes. Central panels: Light-Weighted Mean Age maps ($\langle \log{t_\star} \rangle_L$). Right panels: Mass-Weighted Mean Age maps ($\langle \log{t_\star} \rangle_M$). Contours in all maps represent H$\alpha$ surface brightness at 0.1 $L_\odot$ kpc$^{-2}$ obtained from spaxels with S/N > 3 for H$\alpha$. Ellipses were taken from the romano2008 fit over the H$\alpha$ band, with the crosses marking the centres of those ellipses. The relative coordinates of these maps are in relation to the IFS image centre (see coordinates in Table \ref{['tab:obs_gal']}). North is up and East is to the left. For Arp 143 (top-left panel), the red labels A, B, and C correspond to the position of the spaxels presented in Figure \ref{['fig:spec_grid']}.
• Figure 3: Box plots of the distribution of light-weighted ($\langle \log{t_\star} \rangle_L$) and mass-weighted ($\langle \log{t_\star} \rangle_M$) ages for the three galaxies of our sample. As for any classical box plot, the box contains data points between the first and third quartiles, and the whiskers represent the distance to the farthest data points within 1.5 of the interquartile range (IQR). Bold lines mark the median, while the notches represent the confidence interval around that median. The S/N > 10 criterion was applied beforehand. For VII Zw 466, only spaxels on the ring galaxy were included.
• Figure 4: De-projected radius vs mean age (light-weighted) radial profiles for each galaxy. The radii of the spaxels were de-projected using the H$\alpha$ ellipse parameters of romano2008. The black contours represent the 90% probability of finding a point. The green dashed lines represent the median age of the spaxel distributions. The red dashed lines represents the least square fit to the points. Blue vertical lines mark the semi-major axes distance (ring position). Histograms of the radial and age distribution of the spaxels are shown along box borders. Only spaxels within 1.2 times the semi-major axis distances were considered.
• Figure 5: Box plot profiles for the light-weighted age distribution for the three galaxies. The inset maps highlight the spaxels used to produce the profiles. The top panel includes the general profile (i.e., the four quadrants of the disk together) while each of the other panels correspond to each of the quadrants. The galactocentric radii are measured respect to the centre of the romano2008 H$\alpha$ ellipses and the distances are de-projected from the plane of the sky. The bins have a width of $3.0$ arcsec (i.e. twice FWHM of the seeing). Box plots are drawn only if the number of spaxels in the bin is higher than five. The maximum bin reaches a distance of 1.2 times the major semi-axis radius, represented with the vertical dotted line. Boxplots are defined as in Figure \ref{['fig:age_boxplot']}.