Future Projects

If you are interested in any of these projects, please feel free to contact me.

The paragraph descriptions are ideas that could be done as summer projects or expanded into 1+ year projects. I have several thesis-scale projects too, such as "reduce and analyze the ALMA-IMF large program", but I haven't gotten around to writing those up. Many of these have been started, but not all!

The first group of projects are pieces of four programs I'm running now: the JWST Treasury Survey of the Galactic Center, a Roman program to map the inner Galaxy in 3D, the Panta Rei ALMA Large Program and its long-baseline follow-ups, and a five-band ALMA program on the salts in the Orion Source I disk. The data for these are either in hand or scheduled, and each has a team behind it.

Software Development Projects

I always have a wide range of software development projects available. If you're interested in working on astro-related tools and techniques, I will have something for you depending on your skill set and interests.

Several project variants are always available:
  • Develop new modules for astroquery or solve issues (like this one and this one)
  • Test out and add features to spectral-cube
  • Develop examples showing how to make particular figures or do specific analyses with any astropy or radio-astro-tools functionality
  • Build interactive, zoomable Aladin Lite tours of our data sets (see our ACES tour for an example), and help write the tooling to turn a FITS cube into a HiPS cube. This is part of the outreach component of my current NSF grant.

Projects within currently-running programs

The projects in this section are small pieces of four programs I'm running now. Each one is sized so that a student can finish it in a summer, but each sits inside a larger effort with a team and a data set that already exists or is arriving on a known schedule.

JWST Treasury Survey of the Galactic Center (GO 10678)

We are mapping the inner Central Molecular Zone with JWST/NIRCam in F212N (2.12 μm) and F480M (4.8 μm), observed in parallel. Galactic center stars have nearly identical intrinsic F212N−F480M colors (<0.05 mag), so the observed color gives the extinction to each star individually. Most of the projects below start from that measurement. Observations run through 2026-2027.

Extinction mapping of a single Galactic center cloud

Extinction toward the Galactic center is large (AV ~ 30-100 mag) and varies by up to 0.5 mag over a few arcseconds at 2 μm, so it has to be measured star-by-star rather than assumed.

This project is to take one cloud, use the F212N−F480M color of each star around and behind it to make an extinction map, and compare that map to the dust emission column density from Herschel and from the ACES continuum data. The two need not agree: emission depends on the dust temperature and on the assumed opacity, and absorption only counts dust in front of the stars we can detect. Sorting out where and why they differ for one cloud is a self-contained project.

How complete are our protostar counts?

The star formation rate measured in the CMZ is 1-2 orders of magnitude lower than dense gas star formation relations predict. Counting young stellar objects is the most direct test of this, but the counts are only useful if we know the completeness and the contamination rate.

This project is to take spectral energy distributions of nearby Class 0 and Class I protostars, redden them to Galactic center extinctions, inject them into public JWST images of CMZ clouds, and measure what fraction are recovered as a function of luminosity and local crowding. The result is a completeness curve that converts source counts into a star formation rate, plus an estimate of how many reddened giants get mistaken for protostars.

Photometry of saturated stars

The Galactic center contains both the faintest stars we want to count and some of the brightest stars in the Galaxy, and NIRCam saturates on the latter. JWST reads its detectors non-destructively up the ramp, though, so the first few groups of a saturated exposure still have usable signal.

This project is to build and validate a saturation-recovery photometry pipeline: measure bright stars from the earliest groups, then calibrate against unsaturated measurements of the same stars in shorter exposures. The globular clusters M92, M4, and NGC 6397, observed with the same instrument setup, are uncrowded testbeds with well-known stellar populations. This is mostly a software project. The result is a catalog with several magnitudes more dynamic range than the standard pipeline produces.

3D structure of the inner Galaxy with Roman (program 19008)

Kinematic distances don't work toward the Galactic center: the bar drives non-circular motions, and all circular orbits pile up at vLSR = 0 km/s toward ℓ = 0°. Existing 3D dust maps stop at 1-4 kpc. We are using the Roman Galactic Plane Survey and the Galactic Bulge Time Domain Survey to measure where the inner Galaxy's gas and dust are, using three independent distance indicators: proper motions compared to stellar orbit models, red clump extinction jumps, and RR Lyrae. Roman data are a few months off, and each of these methods can be prototyped right now on VVV, GALACTICNUCLEUS, HST, and JWST data.

Finding RR Lyrae in the inner Galaxy

RR Lyrae are standard candles good to ~5% through the period-luminosity-metallicity relation, and there should be ~500 per square degree along the inner Galaxy sightlines. Roman will produce light curves for ~107 candidate stars, and we need a pipeline that can pick the RR Lyrae out of them in a regime where crowding and variable extinction are both problems.

This project is to build that pipeline and test it on existing data (VVV light curves plus injected synthetic RR Lyrae), measuring the completeness and false positive rate as a function of crowding and amplitude. The student will learn period-finding algorithms (Lomb-Scargle and its variants) and light curve template fitting.

We may do this in collaboration with Q. Daniel Wang and collaborators, who have a Roman program to use nuclear bulge RR Lyrae to constrain its assembly history (Roman program 19024).

Cloud distances from red clump extinction jumps

Stars in front of a molecular cloud are lightly reddened and stars behind it are heavily reddened. The transition shows up as a jump in a color-magnitude diagram and gives the cloud's distance. The red clump works well for this because it is intrinsically narrow in luminosity for the old populations that dominate the inner Galaxy.

This project is to pick one cloud along the Galactic center line of sight - there are many with unknown distances between 300 pc and 5 kpc - measure its extinction jump with existing near-IR photometry, and cross-match the extinction structure against dense gas emission (H13CO+, and CO surveys like SEDIGISM and CHIMPS) to assign it a velocity and a spatial extent. One cloud is a few-month project; the same method applies to the hundreds of clouds Roman will reach.

Panta Rei: gas flow from clouds to cores (ALMA Large Program 2025.1.00383.L)

Panta rei means "everything flows". This program is mapping 259 star-forming clumps at 2.5-5 kpc in 3 mm lines: N2H+(1-0) and H13CO+(1-0) for optically thin kinematics, HCO+(1-0) and HNC(1-0) for infall, and SiO(2-1) for outflows. The clumps span the full evolutionary sequence (luminosity-to-mass ratio from 0.1 to 100) and all of them have larger-scale cloud data (SEDIGISM, OGHReS) and smaller-scale core data (ALMAGAL) already. Data started arriving in October 2025.

Project selection here will be done in coordination with the Panta Rei PIs, so talk to me before settling on one of these. With 259 clumps, almost any per-clump measurement makes a reasonable student project: do a few clumps carefully, and the method scales to the rest of the sample.

Velocity fields from N2H+ hyperfine fitting

N2H+(1-0) is split into hyperfine components with different optical depths, so a single fit gives the centroid velocity, the line width, and the optical depth (and therefore the column density) at each pixel. Over a full mosaic that means fitting hundreds of thousands of spectra.

This project is to fit a few Panta Rei clumps with pyspeckit and compare against the other codes the team uses (mwydyn, BTS): where do they agree, where do they disagree, and what does the disagreement say about how much of the structure is real? From the fitted cubes, build velocity dispersion profiles as a function of radius and look for the drop that marks where the clump has decoupled from its parent cloud. This project will involve some pyspeckit debugging and development.

Infall signatures in HCO+

An optically thick line toward a contracting cloud shows a blue-asymmetric, self-absorbed profile, and HCO+(1-0) has been used this way for decades. Simulations show that collapsing clumps can also produce red-asymmetric profiles depending on the excitation, though, so the standard interpretation needs testing on a large sample.

This project is to build an automated line asymmetry classifier, run it on a subset of Panta Rei clumps, and see whether the fraction of clumps with infall signatures changes with evolutionary stage. The classifications can then be checked against the resolved velocity fields from the optically thin lines, which the analytic infall models never had. This one pairs well with the N2H+ project above.

SiO outflows across the clump evolutionary sequence

SiO is released into the gas phase by shocks, so it traces protostellar outflows selectively and is much less confused than CO. We detect ~300 SiO outflows across the 15 ALMA-IMF regions; Panta Rei covers 259 clumps in SiO(2-1), from starless to feedback-dominated.

This project is to catalog outflows in a subset of clumps, by hand first (as we did in ALMA-IMF) and then with an automated structure-finding routine trained on the hand-made catalog. Once the outflows are identified, measure their masses, momenta, and energy injection rates, and ask whether outflows ever inject enough momentum to reverse the inflow. See also the CO/SiO outflow fraction project below.

Three ways to measure a clump's temperature

Gas temperature sets the Jeans mass and therefore controls fragmentation, so it matters for interpreting core masses. Panta Rei's spectral setup gives three thermometers: the HNC(1-0)/HCN(1-0) ratio, the optically thinner HN13C/H13CN ratio, and CH3CCH(6-5) K-ladder fitting. They are sensitive to different density layers, so they won't agree exactly.

This project is to make temperature maps with all three methods for a set of clumps and work out where and why they differ. The result is a calibration of three widely used tracers, plus a set of temperature maps that the rest of the program needs. The student will learn about excitation and line ratio diagnostics and will work with spectral-cube.

From cores to massive stars: long-baseline ALMA (DIHCA, MAGMAR, HIPPOS)

Panta Rei covers tens of parsecs down to ~0.1 pc, and ALMAGAL reaches ~1000 AU. Getting from 1000 AU cores to ~100 AU disks takes ALMA's longest baselines. DIHCA, MAGMAR, and HIPPOS have observed more than 50 sites of ongoing high-mass star formation at <0.2" (in some cases 0.04") resolution, and 52 of their targets overlap with the larger-scale programs. These are the scales at which we can see whether an inflow ends up in one star or many.

Fragmentation at 100 AU: how many stars is a core making?

At 100 AU resolution, ALMA continuum is very sensitive to dust mass: a 5σ detection at 0.5 mJy is ~0.3 M☉ at 5 kpc. We detect hundreds of individual young stellar objects per high-mass star-forming region. The number of fragments at each scale - cloud, clump, core, disk - is the quantity that separates hierarchical from monolithic collapse in the analytic frameworks now available.

This project is to measure the fragmentation in a handful of regions: build continuum catalogs at the highest available resolution, count sources as a function of scale, and place the results in the fragmentation rate parameter space where different star formation theories make different predictions. There is already a hint that more massive cores fragment more, which is not what a simple thermal Jeans argument predicts.

Dynamical masses of massive protostars

The most direct way to measure a stellar mass is from the kinematics of orbiting material: find a line that traces the disk, extract a position-velocity diagram along the disk major axis, and fit the rotation curve. We have done this for ~30 sources, and the long-baseline surveys contain many more candidates.

This project would work through a set of candidate disks systematically: identify which lines trace the disk rather than the envelope or the outflow, fit the PV structure, and propagate the uncertainties from inclination, distance, and the disk's own mass. These masses calibrate the much cheaper millimeter-flux-based mass estimates we have to use for the rest of the sample. Salt lines (see below) are among the better disk tracers for these sources.

Streamers: late-stage accretion onto forming stars

Streamers are asymmetric flows of gas falling onto the outskirts of protoplanetary disks long after the star was supposed to have finished accreting. They matter for planet formation: cosmochemistry indicates that the early solar system accreted at least two distinct generations of material, and streamers are one way to deliver thermally processed material from elsewhere in the parent cloud.

This project is to search the long-baseline data for streamer candidates around high-mass protostars, check whether their kinematics are consistent with infalling trajectories in the star's potential, and estimate how much mass they deliver. With SiO and hot core tracers we can also ask whether the accreted material has been processed by other stars' outflows, which is not possible in the low-mass studies.

What excites the salts in the Orion Source I disk?

NaCl and KCl are the only molecules we know of that trace only the disks around high-mass young stellar objects; they are absent from the surrounding hot core and outflow because Na and K stay locked in dust grains except where it is hot enough (>500 K) to put them back into the gas. That makes them a useful tool, but we do not understand their excitation: in Orion Source I the vibrational temperature (1200-3300 K) is far above the rotational temperature (120-150 K), and nobody knows why. We have a new five-band ALMA program (2026.1.00197.S; Bands 3, 4, 6, 7, and 8; 88-420 GHz; 0.02-0.06" resolution) to measure rotational level populations across a range of vibrational states and settle it, along with the companion programs 2025.1.00236.S and 2025.1.00274.S.

Rotation-vibration diagrams for NaCl and KCl

The new observations cover more than a dozen NaCl and KCl transitions per band, spanning v=0 through v=5 and J from below 11 up to ~40. Each one needs its flux measured, which includes deblending from the other lines in these spectral scans and accounting for the bright dust continuum.

This project is to measure those fluxes and build the population diagrams. A turnover at low J favors mid-infrared pumping; a knee at high J favors UV/optical pumping; suppression of the high-frequency lines instead points to dust hiding the deeper disk layers.

Modeling salt excitation: infrared pumping, UV pumping, or dust?

Each of the three explanations for the salt excitation makes a different, calculable prediction for the level populations, but producing those predictions requires a radiative transfer calculation with a realistic radiation field: a stellar photosphere, a warm disk, and, for the mid-IR model, a strong emission feature near 29 μm for NaCl and 37 μm for KCl.

This project is to build that model and produce the predicted rotation-vibration diagrams under each hypothesis, so that the observations have something quantitative to be compared against. The same machinery applies to the other refractory species covered by the observations - AlO, FeO, and AlCl - which appear in the outflow rather than the disk. The student will learn statistical equilibrium and radiative transfer.

Hot Core Line Cataloging


We have identified a lot of hot cores in Sgr B2, W51, ALMA-IMF, and throughout the Galaxy. Each of these cores contains many molecular species. We haven't cataloged them all yet! This project would involve using LTE modeling tools: either XCLASS, pyspeckit, or MADCUBA. For a few-month project, a student can just focus on measuring all the lines in 1-2 cores.

Use ML to infer column densities from molecular lines in the CMZ


We have maps of column density (i.e., the number of H2 molecules per cm2) at only poor (10-30") resolution in the Central Molecular Zone. The ACES project produced maps of molecular line intensity at about 2" resolution. Following Gratier+ 2021, can we train a machine-learning model on the low-resolution data (e.g., Jones+ 2012) to derive column density maps at high resolution?

Line Radiative Transfer Modeling: What effect does turbulence have on observed lines?

Much of what we know of the interstellar medium and of star formation derives from properties we measure using spectral line intensities and line ratios. The standard tools for interpreting lines and line ratios are local thermodynamic equilibrium (LTE) models and large velocity gradient (LVG) models. In both cases, the radiative excitation and transfer is simplified to a one-zone models that reduce the entire line of sight to a single assumed state. This assumption is wrong (e.g., Leroy+ 2017).

To solve this, we will apply radiative transfer models using both LTE and LVG to individual cells in a simulation, but will then do ray-tracing radiative transfer through different viewing angles to see what the effects of varying optical depth and excitation are. This will be done using RADMC-3D, and the initial simulations used will be the CATS database. We will post-process these simulations to produce synthetic line emission maps, and we will analyze them as if the whole simulation box were a single zone. We will derive a mapping from the simulation physical conditions (linewidth, moments of the density, column density, and velocity fields, etc) to line ratios and determine whether they follow a trend that will allow inversion of line ratio to inferred physical properties. This work will also produce, as a side effect, synthetic image for other types of analysis.

Archival Research: Catalog of "core" catalogs

Several surveys of pre- and proto-stellar cores have been created, including dozens of local clouds using single-dish instruments and rapidly growing numbers of Galactic Disk clouds with ALMA. This project will be to assemble a table of these catalogs and determine whether they can be drawn from the same parent distribution, or not. If they cannot, that may indicate that the star formation process differs in these environments. If they can, the data hint at a universal star formation process. The key difficulty in this project is to determine how the catalogs are made and how much of the inevitable differences come from observational effects rather than physical effects.

Core Catalog Comparison

Literature catalogs of cores adopt different methods, with the most popular being getsf, dendrograms, and variants of derivative-based peak finders like gext2d and cutex. These cannot be fairly compared. Re-running these algorithms on public data will establish a library to enable fair comparison.

What fraction of molecular outflows have SiO?

We detect ~300 SiO outflows in ALMA-IMF. We haven't really cataloged CO outflows. I'd like to answer the question in the title, but we could ask the inverse too - what are the CO properties of SiO-emitting outflows?

Salt mining in Chile's high desert

We found salt (NaCl and KCl) in the disk around Orion's Source I, and a follow-up survey found nine salt-bearing disk candidates. Are there more salt detections waiting to be extracted from existing data in the ALMA archive? Each detection is likely to result in a dynamical mass measurement of the central source. Together, these detections might tell us more about the physical conditions that allow gas-phase salts to exist around protostars.

Train Machine Learning algorithms to find stars in HII regions

Stellar photometry on complex backgrounds - like HII regions and PDRs - is difficult. Commonly-used algorithms (e.g., those in photutils) fail because they are unable to distinguish substructure in the background from stars. However, it's easy to tell by eye which objects are stars and which are not, at least up to a point. That suggests that machine learning algorithms may be well-suited to solving this problem.

This project would involve creating training sets and testing algorithms against those training sets. The training sets will come from a few samples of "pure background" images: radio continuum images of HII regions, simulations of HII regions and turbulent clouds, and far-infrared images of parts of the Galactic plane. We will add simulated stars to the image using realistic PSFs (e.g., for JWST - STPSF, formerly WebbPSF) to train the algorithm. After proving the tools on simulated data, we will apply them to real JWST data to obtain more complete catalogs of stars in HII regions.

Measure the protostellar population in the W51 star forming region: Part II

High-mass star-cluster-forming regions are the most active areas of star formation in the Galaxy and may have been the dominant path for star formation in the early universe. The W51 protocluster region is one of the closest and most massive in our Galaxy. While the high-mass stars have been identified and measured with ALMA and VLA data, there are abundant lower-mass (but still possibly massive) stars in the surroundings that have only recently been discovered with very high resolution data. These objects need to be cataloged, characterized, and described. This detailed characterization will be used as the 'calibration' for our analysis of a large data set in the ALMA-IMF program. We will also look for binaries and attempt to characterize the binary distribution function at our ~300 AU resolution.
Previously, as part of a student project, a cataloging tool was developed and applied to one set of ALMA images. For this project, we need to take these existing catalogs and cross-match them to measure a few key quantities, including the source multiplicity as a function of resolution and source temperature.
This is a follow-up of another project completed by Connor McClellan (U. Florida). Summer Project Repository and associated software package

Ongoing Projects (that were once advertised on this page)


Which lines trace what processes in the Galactic Center?

This project was led by Alyssa Bulatek (PhD 2026).
There are a lot of molecular species that trace the entire "Central Molecular Zone" that are not seen in molecular clouds in the solar neighborhood. Some of these, like HNCO and SiO, are typically considered "shock tracers" when observed locally, but they are observed to be ubiquitous in the CMZ, so they're not useful for tracing high-velocity (protostellar jet, for example) material. I have a large ALMA program (not an ALMA large program...) to perform a full spectral line survey of one line of sight through the CMZ that includes diffuse gas, dense gas, molecular cores, and protostellar outflows, and our goal is to determine which molecular species, if any, uniquely trace one of these physical features. The data were taken in Cycle 7 (Oct. 2019 - Sep. 2020) in Bands 3, 4, and 6; Alyssa's dasar discovery came out of this survey, and her thesis analyzed the full line survey.

Measure the protostellar population in Galactic Center cloud Sgr B2 DS

This project was led by Desmond Jeff (PhD 2025); see Jeff+ 2024 (ten new hot cores in Sgr B2 DS) and Budaiev+ 2024 (YSO catalog).
How is star formation in the Galactic center different from the Galactic disk? We have a pretty good idea that it is (see, e.g., Longmore+ 2013 and references thereto), but why? Sgr B2 is the only richly star-forming cloud in the Galactic center, and I have been leading star-counting measurements of its ongoing star formation. In Ginsburg+ 2018, I discovered a large population of protostars that I asserted were all high-mass. About half of them are distributed along an elongated, possibly filamentary string in the "Deep South" subregion of the Sgr B2 molecular cloud. I obtained high-sensitivity and high-resolution data at 1 mm to measure these sources more accurately and search for their purported low-mass counterparts. The data are extremely rich, so there are several projects here:
  1. Catalog the continuum sources at 0.5" and 0.1" resolution. Compare the catalog to the existing 3 mm continuum 0.5" resolution data. Determine what the sources are and whether the IMF extrapolations of Ginsburg+ 2018 were correct. Characterize the source multiplicity on the observed 4000 and 1000 AU scales.
  2. Catalog and measure the outflows in Sgr B2 DS. Quick look imaging revealed at least a dozen SiO J=5-4 bipolar outflows. We need to compare these outflows to Galactic disk and other Galactic center outflows. We will also search for other molecular tracers of outflow within the same data set. We will infer source luminosities and perhaps accretion rates from these measurements.

Measure the Kinematic Structure of Sgr B2, the most massive cloud in our galaxy

UF Undergraduate Madeline Hall worked on this project. It was taken up by a large group as part of the ACES large program (data released February 2026), led by Jonny Henshaw, including Anika Schmiedeke, Adam Fairley, and Dylan Pare. Molecular clouds are the regions in which stars form, and they are turbulent. This project aims to measure the kinematic structure (the velocity of difference gas blobs) of the most massive cloud in our Galaxy. The student will use the SCOUSE software to fit spectral profiles to millions of spectra semi-automatically. Because SCOUSE is new and experimental software, this project will involve some software debugging and development. The end result should be an essential measurement of the velocity field and turbulent statistics in a very massive cloud, which will serve as the key measurement in a paper testing theoretical models of star formation. This project will be done in collaboration with Jonny Henshaw, who developed SCOUSE.

Ammonia masers in W51

UF Undergraduate Derod Deal worked on this project.
Ammonia masers are a very rare astronomical phenomenon. They have been detected toward only a handful of sources, including W51 IRS2, in which there are dozens of different masing transitions. I have obtained new high spatial resolution (0.04-0.1 arcsecond, 200-500 AU) data with the VLA with the goal of determining exactly where these masers come from and which source(s) drive them. This project will involve comparison between the VLA data and similar high-resolution data sets from ALMA.

Radio Continuum Imaging & Measurement for the ALMA-IMF project

In collaboration with Roberto Galvan-Madrid
The ALMA-IMF program is a large project to survey to image high-mass star forming regions and count the number of forming new stars of each mass. This measurement of the proto-initial-mass-function (proto-IMF, sometimes core mass function, CMF) is one of the two key ingredients of star formation theory, and if we find environmental variation, it will have profound impact on application of these models in the context of both simulations and extragalactic observations. This sub-project is to measure the "free-free contamination" in our millimeter continuum observations to help get accurate masses toward the highest-mass stars. It will involve imaging radio continuum data from the JVLA, and maybe poking at those same data sets to look for some interesting lines. The first paper measuring the free-free contamination with H41α (Galván-Madrid+ 2024) has been published.

Completed Projects

NRAO REU projects for Summer 2018

Implement a CASA region parser in astropy

Completed by GSoC student Sushobhana Patra. CRTF Documentation
CASA is the main data reduction package for ALMA and VLA data. Regions on the sky such as ellipses, circles, and boxes can be specified in a few different formats. CASA has a region format that is currently supported only by CASA itself, but it will be helpful for researchers to be able to translate these regions to other formats. Astropy is a general-use library for astronomy written in python, and it includes a package for handling sky regions. The student will write a CASA region parser for the astropy regions project. They will gain familiarity both with CASA and astropy. Some experience with python or similar scripted languages is required. This is primarily a software development project and therefore may be of interest to a broader range of students.

Measure the protostellar population in the W51 star forming region

Completed by Connor McClellan (U. Florida). Summer Project Repository and associated software package
High-mass star-cluster-forming regions are the most active areas of star formation in the Galaxy and may have been the dominant path for star formation in the early universe. The W51 protocluster region is one of the closest and most massive in our Galaxy. While the high-mass stars have been identified and measured with ALMA and VLA data, there are abundant lower-mass (but still possibly massive) stars in the surroundings that have only recently been discovered with very high resolution data. These objects need to be cataloged, characterized, and described. The student will examine VLA and ALMA images with resolution ~0.05" and look at source spectra to determine the protostars' luminosities and, ideally, masses. The student will learn to use astropy tools for source finding, Gaussian fitting, and spectroscopic analysis.

Protostars in Orion: Spectral Energy Distributions and proper motions

Completed by Justin Otter (Haverford College). Summer research repository
Published in ApJ: Small Protoplanetary Disks in the Orion Nebula Cluster and OMC1 with ALMA
The Orion nebula is adjacent to the closest high-mass star-forming region in the Galaxy. New ALMA data have revealed several new sources in this region that are likely to be protostars. We would like to measure their positions and luminosities and determine what sorts of protostars they are. Learning about these stars and their motions will tell us what the gravitational potential of the Orion cluster looks like and will be important for determining how high-mass stars have formed there. The student will examine VLA and ALMA images with resolution ~0.05" and look at source spectra to determine the protostars' luminosities and, ideally, masses. The student will learn to use astropy tools for source finding, Gaussian fitting, and spectroscopic analysis. The student will also have opportunities to look at the VLA archive and attempt to reduce archival data if they have the time and ambition.

NRAO REU projects for Summer 2017




Other ideas:

Sgr B2 43 GHz methanol masers: We have VLA A-array (high-resolution) data with dozens of detected methanol masers. We don't know what these masers trace or mean. The data need to be reduced more completely.

Orion disk chemistry: There are ~70 disks in the Orion Source I data set for which we have full spectra, but we haven't analyzed them yet. Are there any lines to be found?


Determine how much mass is ejected in a high-mass outflow

Completed by Terry Melo (Agnes Scott College). Presented at AAS

Goal: Measure the CO/H ratio across a symmetric high-mass outflow. Provide a fundamental calibration for outflow mass-loading rates.

One of the fundamental problems affecting observations of molecular gas is that we don't know precisely how much of each molecule there is relative to the total amount of mass (or the total amount of hydrogen, since it's mostly hydrogen). This outflow, known as the "Lacy Jet" for its discoverer, is unique in that it is (almost) perfectly symmetric, but one half is molecular and the other half is ionized. Because we can measure the hydrogen mass directly in an ionized medium, we should be able to measure the CO abundance directly in this outflow. The project should involve a relatively straightforward measurement followed by an examination of the possible confusing effects that might impact the measurement.

The student will learn about molecular abundances and protostellar outflows. They will be trained to use interpolation and reprojection algorithms to match datasets on different grids. The project is based on data from ALMA, the VLA, and the IRTF.




Geometry of Molecular Clouds in the CMZ

Part of Natalie Butterfield's PhD thesis (U. Iowa)

Goal: Determine the geometry of the clouds using the method demonstrated in Ginsburg+ 2015 (fig 9) and figure out how far the clouds are from the central black hole.

While improved models of the CMZ's geometry have recently been developed (1,2), we still know little about where exactly the clouds are relative to the central black hole and one another. This geometric information is crucial for understanding the physical properties and evolutionary sequence of clouds in the CMZ.

The basic method is to compare H2CO absorption with 13CO emission in position-velocity cubes over regions where we know that HII regions are present in the CMZ along the line of sight. By comparing the emission and absorption, we can determine which clouds are in the foreground or background of the HII regions and therefore where along the "CMZ Ring" they lay.

This project will use ATCA, GBT, and VLA data sets that are already reduced. The analysis will be primarily visual, using ds9 and glue to compare the images. The student will learn these visualization tools and the details of the CMZ's layout.