The Science of Small Things with a Big Telescope

By Dilyar Barat
[email protected]

There is never a dull moment when you do science with the ESO VLT. In 2018, thanks to ESO and ASTRO3D, I was fortunate while an ANU graduate student to visit the Paranal Observatory, and carry out observations of 69 dwarf galaxies using the FLAMES instrument. FLAMES is a multi-object, intermediate and high resolution spectrograph mounted on UT2 (Kueyen), that can access targets over a field of view 25 arcmin in diameter, allowing the observation of up to 130 targets at a time, or (in our case) to do integral field spectroscopy.
 
It was a special moment in the telescope control room when my colleagues and I, after a few sleepless nights, finally embraced the astronomical ritual, and put together an acronym for our survey (special thanks to Boris Häußler). The Study of H𝛼 from Dwarf Emissions (SH𝛼DE) is a high spectral resolution (R=13500) H𝛼 integral field survey of dwarf galaxies with stellar masses 10^6 < 𝑀_* < 10^9 𝑀_⊙.

Studies of the kinematic scaling relations, such as the Tully-Fisher relation, Faber-Jackson relations, etc, have often focused on massive galaxies. The lowest masses typically only go down to ~10^9 𝑀⊙, yet galaxies below this threshold are the most common. These galaxies are typically left out because of their defining physical properties: dwarf galaxies are less luminous than regular galaxies, so observations require longer integration times and/or larger collecting areas; they are smaller, so studying their structure requires better spatial resolution; and finally, dwarf galaxies have lower intrinsic velocity dispersions, so an unbiased measurement of their kinematics requires either high spectral resolution or very high signal to noise.

The SH𝛼DE survey was designed to fill this gap, and to deliver a sample of 69 dwarf galaxies with high spectral and spatial resolution H𝛼 observations. The survey was designed with four goals in mind: (i) to test the linearity of galaxy scaling relations over a range in mass and with sufficient spectral resolution to not be observationally limited; (ii) to measure and explain the fraction of dwarf galaxies with asymmetric kinematics; (iii) to study the dynamical effect of star formation feedback in the low-mass regime; and (iv) to study angular momentum accretion.

Figure 1. Example data from the SH𝛼DE survey, spanning 0.003 ≤ 𝑧 ≤ 0.055 and 10^7.4 ≤ 𝑀_* ≤ 10^10.6 M_⊙. From left to right: SDSS 𝑖-band image; stacked continuum image; H𝛼 observed flux map; and H𝛼 velocity and velocity dispersion maps. SDSS 𝑖-band images also show the FLAMES-ARGUS instrument target footprint on the galaxy with a grid representing the SH𝛼DE spaxels; the small circles in the 𝑖-band images have diameter equal to the FWHM. The flux and kinematic maps only show spaxels with SNR > 5; each map also has an associated histogram showing the distribution of the observed quantity. The red ellipse in each SH𝛼DE data map represents the sampling region inside 1𝑅𝑒.
In our first paper of the series, we investigate the log 𝑀_* – log 𝑆_0.5 mass-kinematics scaling relation, which is a combination of the Tully-Fisher and Faber-Jackson relations, and has previously shown potential for combining galaxies of all morphologies in a single scaling relation. The log 𝑀_* – log 𝑆_0.5 relation is linear over three orders of magnitude in mass, but, within the limits imposed by the mass range and spectral resolution of current, state-of-the-art integral field spectroscopy surveys, it appears to become steeper below 𝑀_* = 10^9 𝑀_⊙. It is unclear if the change in slope is due to increasing gas fractions (as for the Tully-Fisher relation), insufficient spectral resolution, non-equilibrium dynamics, or increasing dark matter fraction in dwarf galaxies. Part of the uncertainty is due to the fact that current IFS surveys are designed to probe galaxies with 𝑀_* > 10^9 𝑀_⊙, so that we lack accurate data precisely where the relation becomes most interesting. It is clear that obtaining new data with better spectral resolution will extend the baseline in 𝑀_* and better constrain the log 𝑀_* − log 𝑆_0.5 relation. Figure 1 above illustrates the excellent data quality we managed to obtain with FLAMES.

With SH𝛼DE, we find that there indeed is a bend in the scaling relation! Figure 2 below shows that there exists a lower limit at 𝑆_0.5 ≈ 22.4 km/s, which corresponds to a stellar mass limit of 𝑀_* ≈ 10^8.6 𝑀_⊙. Above this limit, the scaling relation has a slope of 2.58 ± 0.02 and an intercept of 5.16 ± 0.05. This lower limit originates from an apparent lower limit in the observed H𝛼 velocity dispersion at ∼20 km/s. This demonstrates a physical origin for the low-mass bend in the H𝛼 version of the log 𝑀_* – log 𝑆_0.5 scaling relation, rather than a purely observational effect!

Figure 2. log 𝑀_* - log 𝑆_0.5 scaling relation from combining the SAMI and SH𝛼DE samples. Round points are SAMI measurements; triangular points are SH𝛼DE measurements; all points are colour-coded by gas surface density. The black solid line is the best fit and the shaded region represents the 99% confidence interval in the vertical axis given the uncertainties in the fitted slope and intercept. Magenta horizontal and vertical lines indicate the limit of the linear fit; below this mass threshold, points are modelled as normally distributed around the vertical line.
More science with SH𝛼DE is yet to come. Having access to ESO facilities such as the VLT has proven to be tremendously valuable; to date and to our knowledge there has not been a study focused on dwarf galaxies with such high spectral resolution in such quantity. Currently our team is working hard to make the data available for everyone in the community, stay tuned!

Contributors

Michael Murphy is the Australian representative on the ESO Science Technical Committee. Contact: [email protected]

Sarah Sweet is the Australian representative on the ESO Users Committee. Contact: [email protected]

Stuart Ryder is a Program Manager with AAL. Contact: [email protected]

Guest posts are also welcome – please submit these to [email protected]