Stellar streams in the DESI Survey

drawing

I am a member of the DESI Milky Way survey. DESI is on track to obtain mid-resolution optical spectra for >7 million stars in the Galaxy with measured radial velocities at a precision of 1-10 km/s and a metallicity precision of 0.1 dex, allowing us to probe the Milky Way halo in detail. I’m particularly interested in finding previously unidentified remnants of disrupting(ed) dwarf galaxies and globular clusters. In the end, we would like to compare their properties to expectations from cosmological simulations by generating mock observations that take into account the survey selection. To do this, I am currently using 3 sets of dark matter-only simulations: Symphony (Nadler et al. 2023), MW-est (Buch et al. 2024) and EDEN (Wang et al. 2024). In the future, we would also like to extend these mocks to make predictions for the upcoming Rubin Observatory and other upcoming surveys.

Stay tuned for our results.

Gaps in Globular Cluster Stellar Streams as Probes for Dark Matter in M31 and Other Galaxies

Cosmological simulations that include dark matter predict the existence of low-mass subhalos. These subhalos can perturb thin stellar streams formed by tidal disruption of globular clusters, creating a gap-like feature like those detected in Gaia data (e.g. GD-1). In Aganze et al. 2023, we predict that the Nancy Grace Roman Telescope should be able to detect these gaps in galaxies within 3 Mpc with resolved stars. This research is done in collaboration primarily with Dr. Sarah Pearson and Dr. Tjitske Starkenburg.

drawing A simulation of a mock image of a metal-poor GC stream in an external galaxy for 1-hour exposure with Roman with resolved stars. The gap from a dark matter subhalo of $\approx 5 \times 10^6$ solar masses is show at the center.

Galactic Archeology with Brown Dwarfs at Large Distances

Finding Brown Dwarfs and Low-mass Stars in Deep HST Parallel Fields with Machine Learning

In Aganze et al. 2022a, we used traditional index-selection methods, and machine learning techniques (random forests and neural networks) to isolate a sample of 164 M7-T9 ultracool dwarfs up to distances of ~2~kpc in the Galaxy. Similar surveys with JWST (such as the PASSAGES survey) will uncover a comparable number of objects in the galactic thick disk and halo.

drawing

Top Left: Cutout from a processed image of an L dwarf in the WISP survey with the F140W filter (1.4 µm). Top Right: a 2D G141 (1.1-1.7 µm) spectrum of the same pointing. Bottom: 1D- Spectrum fitted to best-match spectral templates of dwarfs and metal-poor subdwarfs.

Links:

Scaleheights and Ages Brown Dwarfs and Low-Mass Stars in Deep HST Parallel Fields

In Aganze et al. 2022b, we used the sample from Aganze et al. 2022a to constrain the scaleheight and ages using a simulation that takes into account the stellar mass function, multiplicity, UCD evolutionary models, galactic structure and survey selection function. We found that late-M dwarfs are generally older than L and T dwarfs as expected from cooling models, with median ages ranging from 3.6-2.4 Gyr.

drawing

Figure from the paper showing the simulated age (blue) vs inferred age (green) of UCD sub-populations. Dots show the median age of a population.

Local M Dwarfs

Metal Poor GJ 660.1AB System

I’m also interested in the study of nearby UCDs. GJ 660.1AB is a binary system (M1 and M7 secondary) previously overlooked; in Aganze et al. 2016, we showed that the peculiarities in the spectrum of the secondary can be explained by its low metallicity.

drawing

Moderate resolution SpeX SXD spectrum of the primary showing prominent atomic and molecular features in its atmosphere

Undergraduate Mentoring

I have mentored UCSD undergraduate students on summer projects that involve using supervised learning techniques (random forests and neural networks) to classify brown dwarfs and to identify rare populations (binaries, subdwarfs).

Links: