Slide 0 - Instructions :
To move slides, use the arrow keys or swipe on your mobile device
To see the speaker notes, press "s"
To go to full screen, press "f"
To print as PDF, go to this URL: ?print-pdf , then print. To get a PDF with speaker notes, add ?print-pdf&showNotes=true to the URL.
ALMA · ACES 3 mm
JWST NIRCam · GC Treasury (F480M / F212N)
Top: ALMA, the ACES 3 mm continuum of the whole CMZ. Bottom:
the JWST GC Treasury (GO 10678) NIRCam mosaic, F480M / F212N -- the same patch of sky at the
same scale; the black is sky the survey has not covered (yet). It zooms in on Sgr B2
until ACES is at its own resolution, then pans along the dust ridge past the Brick, and
loops.
2MASS via IPAC SPHEREx via NASA/JPL-Caltech
We can see through that obscuring dust in the infrared, revealing the youngest stars.
The Galaxy is filled with gas and dust.
Planck + HI4PI via lambda.gsfc.nasa.gov
The Galactic Center
Tons of open questions, many about to be answered
New surveys: ALMA, JWST, and soon Roman
The GC is the lab for star & planet formation spanning cosmic time
The Galactic Center: Star Formation, Gas Flows, and open questions
Postdocs: Nazar Budaiev (2026-) ,
Theo Richardson (2025-2026) ,
Miriam Garcia Santa Maria (2024-2025), Allison Towner (2020-2023)
PhD: Desmond Jeff (2025) , Theo Richardson (2025) ,
Alyssa Bulatek (2026) ,
Nazar Budaiev (2026) ,
Savannah Gramze (2027) ,
Taehwa Yoo (2028)
Undergrad/postbac:
Derod Deal ,
Aden Dawson ,
Ethan Bhula ,
Mario Antonio Daley ,
Prashant Sikhdar ,
Avery Lacon ,
Laya Damaraju
Supported by NSF 2008101, 2206511, CAREER 2142300, STSCI 1905, 2221, 3523, 5365, 6151, 10662, 10678, Roman 19008, 19012, Astropy
So with that open, my talk is on Galactic Center, star formation, gas flows, and open questions. I've been working with a large group of my students. You can see their names and my postdocs' names here. Of course, you aren't going to read all this. So if you would like to see any of the materials in this talk later on, you can go to my website. The whole talk is already online there. This is the URL, but you just go to my webpage and click talks.
Star formation drives the evolution of the universe
The backdrop is that star formation drives the evolution of the universe.
We had a couple thousand years at the early times where it was just a mixed hot plasma.
After that, gravity started to take over, form galaxies, and within those galaxies convert primordial materials into stars and, through the stars, metals.
Almost everything we see over cosmic history is star formation, with a little bit of salt sprinkled in in the form of galactic nuclei.
After the big bang, there was a brief dark period.
After the dark ages, almost all of the light we see comes from star formation in galaxies.
Galaxies are made of stars, the observable universe is made of galaxies
Star Formation oversimplified
Ṁ
The star formation rate, i.e., how much gas turns to stars
L / M
The light per unit mass, i.e., how stars and stellar populations turn matter into light
So star formation is very important, and we tend to oversimplify it. I like to take the extragalactic view and oversimplify it into two terms.
M-dot, which is the star formation rate: how much gas turns into stars per unit time…
…and L over M, the light per unit mass: so how stars and stellar populations turn that matter into light.
High-mass stars produce photons & heavy elementslow-mass stars live practically forever Point color shows effective temperature, point size shows
As astronomers, we like to look at stars in these diagrams.
On the right is the Hertzsprung–Russell diagram in non-astronomer coordinates, so temperature rising to the right and luminosity on the Y-axis.
On the left side, we also have temperature going to the right, mass on the y-axis.
This left plot tags the lifetimes of stars of a given mass.
These diagrams convey that high-mass stars produce almost all of the photons.
One high-mass star will outshine a million stars like our Sun.
They also produce the heavy elements when they die.
On the other hand, low-mass stars live practically forever, serving as mass sinks.
The color scheme shows the temperature, and is roughly how you might see them by eye.
describe the plots!
Points are colored by an approximation of how they appear on the sky
100 msun stars produce about a million times the light of a 1 msun star
1 and lower Msun stars live as long as the universe
URL links to script to make figures
The stellar initial mass function (IMF)
Stars are sampled from this distribution
We keep that color scheme on this plot showing the distribution of stellar masses.
This is the stellar initial mass function.
In this cartoon, I show the Kroupa IMF, and we sample from it to produce a cluster of a thousand solar masses.
There are very few higher-mass sources, only one about 50 solar massses.
The vast majority by number, and also the majority by mass, are on the low end.
So this is our best model for how we decide how the stellar mass distribution is populated.
But we know it's not universal, even though we often assume that it is.
Log-differential probability vs log mass
Sample cluster of 1000 Msun
Points are a real sampling from the IMF, shows what the curve means in practice
URL links to script to make figures
The IMF is generally assumed to be universal in time and space
primarily because of the weakness of evidence of variation
Say "it may not be universal" -- whether it varies is Q7, asked later for the CMZ
I want to break it down a little further.
The low-mass stars, anything below about eight solar masses, are dominant by number and by mass.
But they produce almost none of the light when the other ones are around: about 1% of the luminosity.
On the opposite end, the blue shading, you have the 20 solar mass plus stars.
These are the massive OB stars.
They produce 95% of the light when they're around, but there's only 1 in 1,000 by number.
They're not negligible by either mass or fraction of supernovae, but neither are they dominant.
And then this middle group, 8 to 20 solar masses, don't produce that much light.
They don't account for that much by number.
They're kind of the same-ish bin in mass, but they're actually dominant for supernovae.
The supernovae are coming from this relatively narrow mass range, because by number these are a lot greater than the more massive ones.
Sample cluster of 10000 Msun
High-mass stars populate the stellar graveyard
Almost all of the light in star-forming galaxies is produced by high-mass stars
The stars form in and from gas (as traced by dust )
When we switch to an infrared view that shows where the dust emission is, from Herschel, you can see these are well correlated.
The ultraviolet light shows up in the same places that the dust does, although there is anti-correlation in detail, because the dust blocks the ultraviolet light.
Herschel FIR dust image of M31
Dust is closely correlated with blue starlight = young stars
Dust traces gas, stars form where there is gas
Most of what we know of star formation in detail comes from small local clouds
Most of what we know about star formation, most of the detailed knowledge we have, comes from looking at small local clouds, like the Taurus molecular cloud, pictured here in a Herschel image.
You see the dust in emission, which reveals the mass distribution of the stuff that will form stars.
When we switch back to the optical, we see a few young stellar objects poking out of the cloud.
These are the objects that have informed most of our models of star formation over the last several decades.
Herschel FIR dust image of Taurus mol cld
distance ~140 pc, local neighborhood
say in words: (and simulations)
click: the same field in visible light -- the bright Herschel filament is the dark
lane: the dust that glows in the far-IR is what blocks the starlight
Cartoon of isolated low-mass star formation
Missing: Neighbors, environment
The cartoon we get from this picture of isolated low-mass star formation is that you have a molecular cloud, and it's a little bit clumpy.
In it, there will be cores, regions of overdensity, that will collapse under their own gravity.
As they collapse, they will form stars in the center, and those stars will start to produce feedback.
At the same time, there will be a disk.
Disks are unavoidable.
And eventually, the feedback from that central star is going to further heat its core and blow away part of it in an outflow, and then eventually consume the rest.
We end up with a star system with a disk around it, and eventually a bare star with maybe some planets hanging around.
What's missing in this picture is any concept of neighbors or a broader environment. And that is the problem.
The cartoon is not universal
The cartoon picture isn't universal.
Most stars form in regions that are more crowded than today's solar neighborhood.
Most of them are forming near massive stars, certainly nearer than our simple picture would allow.
Forming stars care about the neighbors and the environment.
This picture is from the Westerlund 1 cluster, one of the most massive clusters in our galaxy (shout out to the EWOCS team members in the room).
It represents one extreme: only about 5 to 10% of stars in our galaxy today form in an environment as extreme as this.
But it's not a negligible population, and their conditions are quite different.
cosmic noon
Crowded ~ cosmic noon
Stars formed in smaller, denser, more turbulent volumes
The GC is special on the sky but cosmically commonplace
These crowded conditions are commonplace across the universe.
Take this very popular plot from Madau and Dickinson that shows the star formation density per unit time.
And that star formation peaked around 10 gigayears ago, around what we call cosmic noon.
In this period, stars were forming in smaller and denser and more turbulent volumes.
Galaxies were smaller, gas densities were generally higher, and they were more turbulent.
This points us to the Galactic Center: the Galactic Center is special on our sky,
but its environmental conditions are commonplace in a cosmic perspective focused on where typical stars formed.
The solar neighborhood is not crowded
young clusters: <15 Myr
· <60 Myr
1.25 kpc
Src I
Cep A HW2
The solar neighborhood is not particularly crowded.
This is a wonderful top-down view of the solar neighborhood, created using Gaia measurements of star distances plus extinction measurements to perform tomography, three-dimensional reconstruction, of the gas.
This was done involving many folks in this room.
In our local kiloparsec-ish, we have beautiful maps of where the gas is.
We also have, again largely from Gaia, exquisite measurements of young cluster positions.
We know where clusters of stars are formed.
We have a nearly complete census of the OB stars.
You should notice immediately that there are not that many.
You could count them on a couple hands and feet.
These are the massive stars at the top end.
They're the ones that are doing the interesting light production.
Not so many of them.
There are even fewer massive stars forming today, and they all happen to occupy the next mass bin.
The O stars in cyan already exist.
They're already bare stars.
The yellow circles are stars that are still accreting material.
They're still forming.
And within this local kiloparsec, we have fewer than ten.
These are not proto-O stars.
All of these look to be lower mass than the truly luminous objects that produce most of the important feedback.
Top-down: 3D dust and young clusters around the Sun
young clusters: <15 Myr
· <60 Myr
1.25 kpc
Orion
λ Ori
Per OB2
NGC 2264
γ Vel · ζ Pup
Cep OB2 · IC 1396
Cep OB3
Berkeley 59
Coll 419
NGC 7000
Cam OB1
CMa OB1
Vela
Cir OB1
Ara OB1
NGC 6383
Sgr OB1
Rosette
(1.5 kpc)
M8
(1.3 kpc)
The O stars of the previous view, labeled by region. Orion (theta1 Ori C, iota Ori) is the
closest; the Cepheus associations, NGC 7000 (the O3.5 "Bajamar star") and Berkeley 59
are toward l = 90-120; Ara OB1 (NGC 6193) and NGC 6383 toward the inner Galaxy. Dashed rings:
Rosette (NGC 2244, 1.5 kpc) and M8 (NGC 6530, 1.3 kpc), just outside the map. Unlabeled
rings are field / runaway O stars (zeta Pup is in the gamma Vel label; 10 Lac; 68 Cyg ...).
The local kpc calibrates star formation relations now
$\Sigma_{\rm SFR} = \dfrac{N_{\rm YSO}\,\langle m_*\rangle}{t_{\rm YSO}\,A}$
$\langle m_*\rangle = 0.5\,M_\odot$, $t_{\rm YSO} = 0.5$ Myr
Galactic Center
The most crowded region in the Galaxy is its center
We have at best a cartoony view of the Galactic Center
The same dust map, now our own top-down rendering of Edenhofer+ 2024. Click: pull out --
that whole map is the little circle; the Galactic Center is 8 kpc away. (Steps 0-1 of
#cmz-what-shape, which picks this up again later.)
Click: dive into the bar -- and here is what it does
(the bar-model cartoon).
The Galactic Center: open questions (a partial list)
Three major surveys change our perspective
ALMA · ACES
Steve Longmore · John Bally · Cara Battersby · Adam Ginsburg · Jonathan Henshaw · Paul Ho · Izaskun Jiménez‑Serra · Betsy Mills · Daniel Walker · Ashley Barnes · Mattia Sormani · Robin Tress · Jennifer Wallace · Laura Colzi · Maya Petkova · Danya Alboslani · Emad Alkhuja · Jairo Armijos‑Abendaño · Lucia Armillotta · N. Bijas · Nazar Budaiev · Rojita Buddhacharya · Alyssa Bulatek · Michael Burton · Laura Busch · Natalie Butterfield · Paola Caselli · Mélanie Chevance · Claire Cook · Samuel Crowe · Ana Karla Díaz‑Rodríguez · Simon Dicker · Katarzyna Dutkowska · Adam Fairley · Christoph Federrath · Rubén Fedriani · Zi‑Xuan Feng · Karl Fiteni · Gary Fuller · Pablo García · Philipp Girichidis · Simon Glover · Javier Goicoechea · Mark Gorski · Savannah Gramze · Qi‑Lao Gu · Perry Hatchfield · Christian Henkel · Rebecca Houghton · Pei‑Ying Hsieh · Yue Hu · Katharina Immer · Namitha Issac · Desmond Jeff · Janik Karoly · Jens Kauffmann · Ralf Klessen · Diederik Kruijssen · Mark Krumholz · Cornelia Lang · Alex Lazarian · Rachel Lee · Emily Levesque · Fu‑Heng Liang · Dani Lipman · Xunchuan Liu · Xing Lu · Qiu‑yi Luo · Alessandro Lupi · Sera Markoff · Sergio Martín · Brian Mason · Farideh Mazoochi · Laura McCafferty · Emily Moravec · Mark Morris · Francisco Nogueras‑Lara · Marie Nonhebel · Tomoharu Oka · John Orlowski‑Scherer · Jürgen Ott · Marco Padovani · Xing Pan · Dylan Paré · Thushara Pillai · Jaime Pineda · Marc Pound · Denise Riquelme‑Vásquez · Víctor Rivilla · Charles Romero · Galaxy Salo · Álvaro Sánchez‑Monge · Miriam Santa‑Maria · Anika Schmiedeke · Rainer Schödel · Matthias Schultheis · Howard Smith · Yoshiaki Sofue · Jack Sullivan · Enrico Di Teodoro · Leonardo Testi · Volker Tolls · Miguel Requena Torres · Grant Tremblay · Arianna Vasini · Gijs Vermariën · Alexey Vikhlinin · Serena Viti · Q. Daniel Wang · Gwenllian Williams · Fengwei Xu · Qizhou Zhang · Suinan Zhang
papers I–VIII out (2025–26)
JWST · GC Treasury
Rainer Schödel · Steve Longmore · Adam Ginsburg · Konstantina Anastasopoulou · Jay Anderson · John Bally · Ashley Barnes · Cara Battersby · Alessandra Mastrobuono Battisti · Giuseppe Bono · Aaron Bryant · Santi Cassisi · Mélanie Chevance · Anna Ciurlo · Angela Cotera · Samuel Crowe · Tuan Do · Rubén Fedriani · Anja Feldmeier‑Krause · Sebastiano von Fellenberg · Lucia Bravo Ferres · Dimitri Gadotti · Carme Gallart · Teresa Gallego‑Calvente · Eulalia Gallego‑Cano · Pablo García · Angela Gardini · Abhimat Gautam · Andrea Ghez · Stefan Gillessen · Miguel Cano González · Alessia Gualandris · Robert Gutermuth · Daryl Haggard · Matthew Hankins · Jonathan Henshaw · Matthew Hosek · Rebecca Houghton · Yue Hu · Tracy Huard · Ralf Klessen · Diederik Kruijssen · Ryan Lau · Alex Lazarian · Rebecca Levy · Mattia Libralato · Anan Lu · Jessica Lu · Xing Lu · Nora Luetzgendorf · John Magorrian · Shifra Mandel · Roeland van der Marel · Macarena García Marín · Sera Markoff · Alvaro Martinez · Maria Melamed · Betsy Mills · Kaya Mori · Mark Morris · Elena Murchikova · Paco Najarro · David Nataf · Nadine Neumayer · Shogo Nishiyama · Francisco Nogueras‑Lara · Dylan Paré · Florian Peissker · Maya Petkova · Thushara Pillai · Gabriele Ponti · Rodrigo Contreras Ramos · R. Michael Rich · Carlos Román‑Zúñiga · Michael Rugel · Nils Ryde · Nadeen Sabha · Joel Sánchez‑Bermúdez · Álvaro Sánchez‑Monge · Matthias Schultheis · Stephen Serjeant · Lijing Shao · Janet Simpson · Mattia Sormani · Christopher Sorrell · Jonathan Tan · Enrico Di Teodoro · Brian Thorsbro · Pablo Torne · Robin Tress · Elena Valenti · Sill Verberne · Pierre Vermot · Eduardo Vitral · Gunther Witzel · Siyao Xu · Farhad Yusef‑Zadeh · Michal Zajacek · Qizhou Zhang · Manuela Zoccali
half observed (68 of 139 pointings, Sep 2026); the rest Aug–Sep 2027
Roman surveys
et al.
coming
The Galactic Bulge Time Domain Survey & Galactic Plane Survey: proper motions,
variability, and microlensing planets.
The reason I'm telling you today about these open questions, and this handful of questions we've made some advances on, is that we have several new surveys that are completely changing our perspective.
The first is the ALMA ACES program.
This one has been basically published in the last year, and it surveys the entire Central Molecular Zone, so the central few hundred parsecs of the galaxy, at wavelengths of 3 millimeters.
Next we have the JWST Galactic Center Treasury program.
It was approved earlier this year, and the first observations were taken a month ago.
It is half observed.
The second half will be next year, in the next observing window.
Coming up very soon, starting in the next few months, are the Roman surveys: the Galactic Bulge Time Domain Survey, GBTDS, and the Galactic Plane Survey.
A shout-out that several people in the room are involved in all of these.
These are large collaborative programs.
Both the ACES and GC Treasury programs have of order a hundred co-Is on them, and it is an open collaboration.
You're welcome to join.
All you have to do is fill out the sign-up form.
For Roman, everybody is involved.
It's an even more open program, with community-level decisions on where to point and what to observe.
The ACES team · 113 people
Steve Longmore · John Bally · Cara Battersby · Adam Ginsburg · Jonathan Henshaw · Paul Ho · Izaskun Jiménez‑Serra · Betsy Mills · Daniel Walker · Ashley Barnes · Mattia Sormani · Robin Tress · Jennifer Wallace · Laura Colzi · Maya Petkova · Danya Alboslani · Emad Alkhuja · Jairo Armijos‑Abendaño · Lucia Armillotta · N. Bijas · Nazar Budaiev · Rojita Buddhacharya · Alyssa Bulatek · Michael Burton · Laura Busch · Natalie Butterfield · Paola Caselli · Mélanie Chevance · Claire Cook · Samuel Crowe · Ana Karla Díaz‑Rodríguez · Simon Dicker · Katarzyna Dutkowska · Adam Fairley · Christoph Federrath · Rubén Fedriani · Zi‑Xuan Feng · Karl Fiteni · Gary Fuller · Pablo García · Philipp Girichidis · Simon Glover · Javier Goicoechea · Mark Gorski · Savannah Gramze · Qi‑Lao Gu · Perry Hatchfield · Christian Henkel · Rebecca Houghton · Pei‑Ying Hsieh · Yue Hu · Katharina Immer · Namitha Issac · Desmond Jeff · Janik Karoly · Jens Kauffmann · Ralf Klessen · Diederik Kruijssen · Mark Krumholz · Cornelia Lang · Alex Lazarian · Rachel Lee · Emily Levesque · Fu‑Heng Liang · Dani Lipman · Xunchuan Liu · Xing Lu · Qiu‑yi Luo · Alessandro Lupi · Sera Markoff · Sergio Martín · Brian Mason · Farideh Mazoochi · Laura McCafferty · Emily Moravec · Mark Morris · Francisco Nogueras‑Lara · Marie Nonhebel · Tomoharu Oka · John Orlowski‑Scherer · Jürgen Ott · Marco Padovani · Xing Pan · Dylan Paré · Thushara Pillai · Jaime Pineda · Marc Pound · Denise Riquelme‑Vásquez · Víctor Rivilla · Charles Romero · Galaxy Salo · Álvaro Sánchez‑Monge · Miriam Santa‑Maria · Anika Schmiedeke · Rainer Schödel · Matthias Schultheis · Howard Smith · Yoshiaki Sofue · Jack Sullivan · Enrico Di Teodoro · Leonardo Testi · Volker Tolls · Miguel Requena Torres · Grant Tremblay · Arianna Vasini · Gijs Vermariën · Alexey Vikhlinin · Serena Viti · Q. Daniel Wang · Gwenllian Williams · Fengwei Xu · Qizhou Zhang · Suinan Zhang
ACES: the ALMA CMZ Exploration Survey. PI Steve Longmore; co-PIs John Bally, Cara Battersby, Adam
Ginsburg, Jonathan Henshaw, Paul Ho, Izaskun Jiménez-Serra, Betsy Mills; work-package leads Dan Walker, Ashley Barnes,
Mattia Sormani, Robin Tress, Jen Wallace, Laura Colzi, Maya Petkova. 113 people across papers I-VI.
The JWST GC Treasury team · 100 people
Rainer Schödel · Steve Longmore · Adam Ginsburg · Konstantina Anastasopoulou · Jay Anderson · John Bally · Ashley Barnes · Cara Battersby · Alessandra Mastrobuono Battisti · Giuseppe Bono · Aaron Bryant · Santi Cassisi · Mélanie Chevance · Anna Ciurlo · Angela Cotera · Samuel Crowe · Tuan Do · Rubén Fedriani · Anja Feldmeier‑Krause · Sebastiano von Fellenberg · Lucia Bravo Ferres · Dimitri Gadotti · Carme Gallart · Teresa Gallego‑Calvente · Eulalia Gallego‑Cano · Pablo García · Angela Gardini · Abhimat Gautam · Andrea Ghez · Stefan Gillessen · Miguel Cano González · Alessia Gualandris · Robert Gutermuth · Daryl Haggard · Matthew Hankins · Jonathan Henshaw · Matthew Hosek · Rebecca Houghton · Yue Hu · Tracy Huard · Ralf Klessen · Diederik Kruijssen · Ryan Lau · Alex Lazarian · Rebecca Levy · Mattia Libralato · Anan Lu · Jessica Lu · Xing Lu · Nora Luetzgendorf · John Magorrian · Shifra Mandel · Roeland van der Marel · Macarena García Marín · Sera Markoff · Alvaro Martinez · Maria Melamed · Betsy Mills · Kaya Mori · Mark Morris · Elena Murchikova · Paco Najarro · David Nataf · Nadine Neumayer · Shogo Nishiyama · Francisco Nogueras‑Lara · Dylan Paré · Florian Peissker · Maya Petkova · Thushara Pillai · Gabriele Ponti · Rodrigo Contreras Ramos · R. Michael Rich · Carlos Román‑Zúñiga · Michael Rugel · Nils Ryde · Nadeen Sabha · Joel Sánchez‑Bermúdez · Álvaro Sánchez‑Monge · Matthias Schultheis · Stephen Serjeant · Lijing Shao · Janet Simpson · Mattia Sormani · Christopher Sorrell · Jonathan Tan · Enrico Di Teodoro · Brian Thorsbro · Pablo Torne · Robin Tress · Elena Valenti · Sill Verberne · Pierre Vermot · Eduardo Vitral · Gunther Witzel · Siyao Xu · Farhad Yusef‑Zadeh · Michal Zajacek · Qizhou Zhang · Manuela Zoccali
Names: investigators of
JWST program 10678
·
gold : PI & co-PIs · faces from public pages, where one was found
JWST Galactic Center Treasury (program 10678): PI Rainer Schödel, co-PIs Steve Longmore and Adam
Ginsburg, and 97 co-investigators. Matt Hosek leads the cluster / tidal-stream work.
The Galactic Center: open questions (a partial list)
Q1 What planets form there? 2016 ← Q2 How does gas get into the CMZ? 1991 Q3 (How) does gas reach Sgr A*? 1996 Q4 What shape is the CMZ? 1995 Q5 Why does the CMZ form so few stars? 2013 Q6 Where do the CMZ’s young stars go? Do clusters dissolve? 2002 Q7 Is the IMF different here? 1999 Q8 Is the chemistry different? What is the metallicity? 2009
Almost every planet we know is within 1 kpc of the Sun
NASA Exoplanet Archive, Sept 2026
Roman will find tens of thousands of planets ~8 kpc away
(these numbers assume 0 planets in the GC pointing)
Planets within 1 kpc: 82% of those known today → ~10% after Roman
This is now the same histogram, but log-scaling the y-axis. And I'm highlighting where the Galactic Center is.
With Roman microlensing, we're going to start shifting the balance closer toward the Galactic Center distance.
With Roman transiting planets included, the predicted yields are that the average planet is going to be at distances greater than one kiloparsec within a couple of years.
Most will be from this Roman sample.
Caveat: the yields assume that there are zero planets in the actual Galactic Center pointing.
These are mostly from the Galactic bulge, which is a little bit off the Galactic Center.
I think that zero number is going to go up.
2 of the CMZ, F212N and F480M,
~5,300 per arcmin2 detected and crowding-limited. The planet yields above are
for the lower-extinction bulge fields, not the CMZ itself.
?
What happens to
Proplyds: disks photoevaporated
Tails in the ArchesHosek+ in prep
EWOCS : JWST in Westerlund 1 & 2
Monsch et al.,Lovell , Galarza
Wd1
Wd2
This census is going to let us study how star and planetary systems form in extreme environments like those at cosmic noon.
We have our local view of objects like Barnard 68, the prototypical prestellar core, and HL Tau, the prototype disk, in the solar neighborhood.
.
If we take those disks and put them in environments that are common at cosmic noon…
…we expect to see see more feedback similar to what we see in proplyds in Orion;
dynamical effects like accretion-driven truncation,
and photoevaporative tails like those we now see in the Arches.
And I want to give a shout-out: these environments are also similar to what we see in Westerlund 1 and 2, as part of the EWOCS program.
Kristina Monsch, Josh Lovell, Rafael Galarza, at least, in the audience here are involved in those programs, and specifically studying these sorts of photoevaporative effects.
We will be able to study the whole process end to end in the Galactic Center.
Builds: the plot alone. Click: the z = 1-3 band, B68 and HL Tau. Click: the question. Click: proplyds,
truncation, the Arches tails and EWOCS.
Most of the stars in the universe formed at cosmic noon, z ~ 1-3
Those conditions -- density, UV field, turbulence -- are what the CMZ reproduces locally
So: what happens to the disks, and to the planets, under those conditions?
Credit the truncation argument correctly: the mechanism is Wijnen+ 2016, 2017
(Pols, Pelupessy, Portegies Zwart — Nijmegen/Leiden): face-on accretion of ISM material
carrying no azimuthal angular momentum shrinks the disk; ram-pressure stripping removes the rest.
It shows up in the observations via Otter+ 2021 (Otter, Ginsburg, Ballering, Bally+) —
the OMC1 disks are embedded, hidden from the Trapezium's UV, and still small, so photoevaporation
cannot be what truncated them. Moeckel & Throop 2009 (Bondi–Hoyle–Lyttleton
accretion onto a protoplanetary disk) made essentially the same argument a decade earlier, as did Bally.
The picture on the slide is Bally+ 2020's; the physics is not his.
Bottom right, tails in the Arches (Hosek+ in prep).
EWOCS shout-out (Extended Westerlund 1 and 2 Open Clusters Survey; JWST GO-1905 in Wd1, GO-3523 in Wd2,
PI Guarcello): Kristina Monsch et al., including Josh Lovell and Juan Martínez-Galarza .
Left to right: Wd1, Wd2, Kristina's tail figure. The winds around Wd1's massive stars were Monsch, Lovell+'s
AAS 245 result; the MIRI shell around the M supergiant W237 is in Guarcello+ 2025 (EWOCS-III).
If asked about disks there: in Wd2, 73 of 353 substellar candidates show an infrared excess, i.e. candidate
disks (Almendros-Abad+ 2026, EWOCS-IX).
Disks exist in the CMZ
Sgr C ☉ ☉
We can study planet formation in this environment.
We have detected disks in the Galactic Center.
We've measured some that are clearly rotating, rotationally supported objects that are feeding into central stars.
The caveat is that right now we are limited to these very gas-rich and massive disks, simply because of resolution limitations, we're not going to be doing disk substructures in the Galactic Center.
We _are_ going to be doing disk fractions and demographics.
We have of order hundreds of disks detected.
For planets in cosmically common conditions, the future is bright, but right now we don't know very much.
Q1 What planets form there? 2016 future! Q2 How does gas get into the CMZ? 1991 ← Q3 (How) does gas reach Sgr A*? 1996 Q4 What shape is the CMZ? 1995 Q5 Why does the CMZ form so few stars? 2013 Q6 Where do the CMZ’s young stars go? Do clusters dissolve? 2002 Q7 Is the IMF different here? 1999 Q8 Is the chemistry different? What is the metallicity? 2009
The bar feeds the CMZ
I gave you the picture that the solar neighborhood is very well mapped in 3D.
The Galactic Center is not.
We have this very cartoony view of the Galactic Center.
We're going to now zoom into that cartoon and explain where it comes from.
The bar feeds the CMZ
simulation: D. Lipman
We have a picture of our Galactic Center that comes largely from looking at other galaxies with similar structures
Material falls in along dust lanes.
One third of that gas is going to land on the Central Molecular Zone.
The rest is going to overshoot, and when it overshoots, it hits stuff on the far side.
In this movie, watch in particular where gas overshoots the center and hits the far side gas.
These collisions both preventing that gas from falling straight in, and result in high-velocity collisions.
Colliding gas stagnates flows and feeds the CMZ
Gas comes in molecular but mostly without forming stars
G5 (l ≈ 5.4°)
l ≈ 3.3° (top) and 2.9° (bottom)
We have clear evidence that overshooting gas crashes into infalling gas.
This is work by my student, Savannah Gramze.
On the left side, published in 2023, on the right, work she is going to submit early next year.
These are sites along that inflowing bar where there are high-velocity collisions.
Each image is a position-velocity diagram: velocity on the y-axis, position on the x-axis.
There are two clouds separated by of order 100 kilometers per second, with only a little bit of material in the interaction zone between them.
These sharp-edged features show that clouds are actually smashing into each other.
Despite the hight delta-v, they're not producing high-energy emission at all.
They're also not producing star formation.
Why not is one of the open mysteries.
Why are there high-velocity collisions that produce very little energy, very little star formation?
Future: gas into the CMZ
We also want to know what's happening as we go further in, and that is one of our future prospects.
Right now, members of the ACES team like Dani Lipman and Robin Tress are working on simulations that are going to accurately bring material all the way inside.
And in the ACES program, at high resolution in the inner few hundred parsecs, we're starting to see hints of similar velocity-bridge collision features.
But it's a lot more uncertain, because the observational picture is more confused.
Q1 What planets form there? 2016 future! Q2 How does gas get into the CMZ? 1991 progress Q3 (How) does gas reach Sgr A*? 1996 ← Q4 What shape is the CMZ? 1995 Q5 Why does the CMZ form so few stars? 2013 Q6 Where do the CMZ’s young stars go? Do clusters dissolve? 2002 Q7 Is the IMF different here? 1999 Q8 Is the chemistry different? What is the metallicity? 2009
What gas reaches Sgr A*?
We have measured bounds on the amount of gas coming in to the CMZ and ending in Sgr A*.
We have only models in between.
The bar lanes supply a solar mass per year into the Central Molecular Zone.
Faraday rotation measurements from SMA and ALMA bound the final accretion flow.
Out at the Bondi radius, 10,000 AU, X-ray observations give limits.
Those set our boundary conditions.
Between those, we have six orders of magnitude where we don't know exactly what's going on.
Magnetic fields and star-formation driven feedback are likely the main drivers based on simulations.
The open question remains: how much gas reaches Sgr A*, or what feeds our black hole?
Bar lanes → CMZ: ~1 M☉ yr−1
SB19 Fig. 1: CO of the near-side (blue) and far-side (red) dust lanes; top l–b (gives the mass), bottom l–v
SB19 Fig. 3: inflow over the next 16 Myr from the gas now on the lanes: instantaneous (top), cumulative (bottom)
Hatchfield+ 2021 Fig. 3: simulation, face-on: lane gas that misses the CMZ (“overshooting”) and hits the far lane
Sormani & Barnes 2019: 2.7 (+1.5/−1.7) M☉ yr−1 — the circle
The two bar dust lanes, picked out of the Dame+ 2001 12 CO cube in (l, b, v):
near side (the “connecting arm”) and far side
Mass from CO × XCO . A geometric bar model (bar angle φ = 15–30°) turns each
lane's (l, v) into a position and an inward speed
Ṁ = mass ÷ time to reach the CMZ (R = 300 pc): near 1.2, far 1.5 → 2.7, averaged
over the next ~10 Myr. The instantaneous rate swings by a factor of a few (episodic)
Errors: XCO and φ
Hatchfield+ 2021: 0.8 ± 0.6 M☉ yr−1 — the square
Not all lane gas lands on the first pass: some overshoots the CMZ and hits the far lane. SB19 flagged
this as an unknown factor f ≤ 1
Tracer particles in an Arepo bar simulation: 30 ± 12% accretes on first approach; most
of the rest arrives within ~25 Myr
2.7 × 0.30 = 0.8: an observed mass flux times a simulated efficiency
On the ladder: both points at the CMZ edge
(~200 pc). The only step where the rate comes from tracing the inflowing gas itself.
Bar lanes: Sormani & Barnes 2019 measured the mass on the two dust lanes in CO and used a bar model to get the
inward speed: 2.7 Msun/yr. Hatchfield+ 2021 found in simulations that only ~30% lands on the first pass: 0.8.
CMZ ring → center: magnetic transport (simulation)
Tress+ 2024 Fig. 6: MHD run, central ±0.5 kpc, t = 147–293 Myr: the empty inside of the ring fills with gas
Fig. 23: inflow into R < 100 pc: MHD (solid) vs. no field (dashed)
Fig. B1: same, at cell mass 103 , 102 , 10 M☉ (bottom)
Tress+ 2024: the magneto-rotational instability (MRI) moves gas in from the ring
MHD of gas in a Milky Way bar potential (Arepo). No self-gravity, star formation or feedback
The bar fills the CMZ ring at ~1 M☉ yr−1 , with or without a field
Inside the ring only the MHD run has inflow: magnetic (Maxwell) stresses carry angular momentum out and
gas drifts in at 0.01–0.1 M☉ yr−1 . Without a field it is ~zero: all the gas
stays in the ring
Not converged: each 10× finer cell mass cuts the inflow ~5–10× (Fig. B1). The
authors call 0.01–0.1 an upper limit
What 3D data can test here
Predicted, and observable: the ordered field follows the gas flow; B ≈ 100 μG
(n/103 cm−3 )0.33 ; a slow radial drift of gas inside the ring
Ṁ(R) = 2πR Σ(R) vR (R) needs face-on Σ and vR . Seen
edge-on, the line-of-sight velocity mixes rotation and inflow; positions along the line of sight separate
them
On the ladder: the pale bar at ~100 pc with the
downward arrow.
Blue, magnetic: Tress+ 2024 MHD simulations. The MRI drives 0.01-0.1 Msun/yr from the ring inward; with no field,
nothing gets in. It drops with resolution, so it is an upper limit.
CMZ → CND → 1 pc: stellar feedback (simulation)
Feng+ 2026 Fig. 26: time-averaged inflow vs. radius (red: SN + radiation; blue: SN only; black: no feedback), with the observational estimates (dots) and Tress+ 2020 / 2024 (stars, squares)
Fig. 27: lower resolution (pale) → more inflow
Feng+ 2026: supernovae + radiation feedback push gas inward
Arepo: Milky Way bar potential, self-gravity, star formation, supernovae, on-the-fly radiative transfer;
cells ≤ 20 M☉ inside 100 pc
Time-averaged inflow falls steeply inward: 5×10−3 (100 pc) →
2×10−4 (10 pc) → 10−6 M☉ yr−1 (1 pc)
Two parts: a smooth, viscous-disk-like drift from feedback-driven turbulence, plus clouds falling in:
bursts up to 10−3 M☉ yr−1 for 3–5 Myr at 10 pc
No feedback → almost no inflow inside 100 pc
Higher resolution → less inflow (Fig. 27): upper limits
The data here are few and model-based: ~10
−2 M
☉ yr
−1 from
cloud kinematics fit with an accretion-disk model (
von Linden+ 1993 ; blue dot), and
10
−4 –10
−3 in the CND from disk models of colliding clouds and from CS
streamers (
Vollmer & Duschl 2001 ,
2002 ;
Hsieh+ 2017 ; green dot). Feng's curves lie below
both. Measuring Σ and v
R in 3D, ring to CND, is the test.
On the ladder: the open circles and
line (the 10 pc point is drawn at 10−4 ; the paper says ~2×10−4 );
the dashed arrow is the 3–5 Myr bursts.
Blue, feedback: Feng+ 2026. Supernovae and radiation stir the gas and it drifts in, 5e-3 at 100 pc down to 1e-6
at 1 pc, with bursts of 1e-3 at 10 pc. Also an upper limit. The few observational estimates here are older and
model-based; this is the zone a 3D map of the CMZ can measure.
The red diamond: Bondi capture, from Chandra
Wang+ 2013 Fig. 1a: Chandra 1–9 keV; dashed circle: their Bondi radius (4″)
Baganoff+ 2003 Fig. 12: spectrum of the diffuse gas within 10″: hot plasma + Fe line
Ressler+ 2018 Fig. 6: simulated winds of the ~30 Wolf–Rayet stars, 0.5 pc across
Ressler+ 2018 Fig. 12: inflow (grey) and outflow (green) rates vs. radius in that simulation
Measured (Baganoff+ 2003, Chandra): hot plasma, kT ≈ 1.3 keV and
n
e ≈ 26 cm
−3 , within ~1.5″ of Sgr A*.
Derived: Bondi (1952): gas inside R
B = 2GM/c
s 2 is captured.
c
s ≈ 550 km/s → R
B ≈ 0.07 pc (they adopt 0.06).
Ṁ
B ∝ M
2 ρ/c
s 3 ≈ 10
−6 M
☉ yr
−1 ,
for the 2003 mass, 2.6×10
6 M
☉ . At today's 4.3×10
6 : ~3×10
−6
(and R
B ≈ 0.1 pc);
Wang+ 2013 (3 Ms Chandra) quote ~10
−5 .
So: an estimate of what Sgr A*
could capture, not a measured inflow; Baganoff call
it a lower limit (cold gas not counted).
Whose gas: mostly shocked winds of the ~30 Wolf–Rayet stars in the central parsec
(
Cuadra+ 2008 ;
Ressler+ 2018 reproduce the Chandra emission with them) — not the molecular gas
coming down from the CMZ.
How much gets in: Wang+ 2013: a flat density profile means outflow nearly balances inflow;
≲1% of the captured gas reaches the hole. Ressler+ 2018 extrapolate to a few ×10
−8 at the horizon.
On the ladder: the diamond
(10−6 at 0.055 pc) is Baganoff's number at the 2003 mass, drawn filled but model-based;
today ~3×10−6 –10−5 at ~0.1 pc.
Red diamond: not EHT. Chandra sees hot gas around Sgr A* (Baganoff+ 2003); the Bondi formula says Sgr A* could
capture ~1e-6 Msun/yr of it (~3e-6 with today's mass, ~1e-5 in Wang+ 2013). The gas is mostly Wolf-Rayet wind
material, and less than ~1% of it gets in.
The red bar: Faraday rotation, not the EHT
Marrone+ 2007 Fig. 1: polarization angle vs. λ2 , SMA 227 + 343 GHz; slope = RM
Bower+ 2018 Fig. 10: |RM| over 2002–2016 (BIMA, SMA, ALMA)
Ressler+ 2020 Fig. 6: inflow (top), outflow (bottom) in 10−8 M☉ yr−1 vs. radius (rg ): WR winds → MHD → GRMHD
EHT 2022 Paper V Fig. 16 (top): horizon Ṁ of the GRMHD models vs. spin, normalized to the 230 GHz flux
Measured (Marrone+ 2007, SMA): the polarization angle of Sgr A* rotates ∝
λ
2 : RM = (−5.6 ± 0.7)×10
5 rad m
−2 ,
constant over 10 epochs.
Derived: RM ∝ ∫n
e B
∥ dl through the accretion flow,
from R
B in to r
in , where electrons turn relativistic. For n ∝ r
−β
and an equipartition field:
upper limit 5×10−8 –2×10−7 M☉ yr−1 for
rin = 30–100 RS , if the field is ordered and radial. A weaker field
(fraction ε) raises it by ε−2/3
lower limit 2×10−9 , even for a weak, tangled or toroidal field
Since: ALMA mean RM ≈ −5×10
5 , as for ~20 yr (it varies on hours–months) →
Ṁ ≈ 10
−8 (
Bower+ 2018 ). EHT “best-bet” models (MAD; pass every constraint
but variability): (5.2–9.5)×10
−9 (
EHT 2022 Paper V ). Wind-fed GRMHD: ≈ 10
−8 ,
magnetically arrested (
Ressler+ 2020 ).
On the ladder: the bar (2×10−9 –2×10−7 )
is Marrone's range, drawn at ~10 RS : the RM constrains the inner flow (≲100 RS ),
so “event horizon” is approximate. The EHT values fall inside it.
−6 –10−5 could be captured;
~10−8 gets in.
Red bar: not EHT either. It is Faraday rotation measured with the SMA (Marrone+ 2007) turned into an accretion
rate with a hot-flow model: between 2e-9 and 2e-7. Later ALMA RM (Bower+ 2018), the EHT models, and wind-fed
GRMHD all land near 1e-8. Say: at the Bondi radius, Chandra's hot gas (mostly stellar-wind material) implies
Sgr A* could capture ~1e-6 to 1e-5 Msun/yr; Faraday rotation and EHT-calibrated models put what reaches the
horizon near 1e-8 -- about a percent or less of the Bondi rate.
Evidence of flow inward?
High-velocity-width features in ACES are shocks — like the bar-lane collisions (Q2)
— Lipman+ 2026
Is this inflow?
Future: Roman tomography. Sub-pc MHD sims.
The simulations tell us what to look for, but right now they're resolution-limited.
With the ACES program, we've been mapping out high-velocity structures that indicate either collisions or stretching along the line of sight.
This ACES program, led by Dani Lipman with major contributions from Savannah, shows the very high-velocity features we found in the central hundred parsecs.
In short, there are possible inflow-driving shock signatures filling the whole CMZ.
From Cara Battersby's work, we saw hints that linear filaments could feed into the central black hole.
The problem, though: we can see this kinematically.
We know how the gas is moving, but we don't know where it is along the line of sight.
That's been the biggest outstanding problem in Galactic Center studies.
We see hints of inflowing material, but it could also be tens of parsecs away along the line of sight.
The future will be 3D tomography, which I'll tell you more about soon.
Q1 What planets form there? 2016 future! Q2 How does gas get into the CMZ? 1991 progress Q3 (How) does gas reach Sgr A*? 1996 future Q4 What shape is the CMZ? 1995 ← Q5 Why does the CMZ form so few stars? 2013 Q6 Where do the CMZ’s young stars go? Do clusters dissolve? 2002 Q7 Is the IMF different here? 1999 Q8 Is the chemistry different? What is the metallicity? 2009
Question 4: What shape is the CMZ? Open since 1995. Where we are: most clouds on the near side; nested x2 orbits, no single orbit fits. What comes next: JWST extinction + Roman proper-motion tomography (~10 pc along the line of sight).
We know the solar neighborhood in 3D, but not the GC
Tress+ 2026 (Rhea-RT): total gas surface density, face-on, 1 kpc across, over 125 Myr
We have beautiful edge-on, two-dimensional maps of the Galactic Center, but really poor three-dimensional maps.
And that has resulted in a lot of discussion and controversy.
One hypothesis is that the molecular gas is predominantly constrained to a ring.
This is known as the Molinari ring.
This was hinted at by Herschel data.
And it's the simplest view, simplest geometry.
It doesn't hold up in a lot of ways, but it's simple and captures several bulk features well.
Nested spiral arms have been around as an idea for 30-50 years and still are reasonable models.
The inner ring could instead be an inner disk, roughly uniformly filled in.
But the reality is almost certainly the most complex option, that it's a mix of everything.
The simulations hint at this.
They also tell us that, depending on when we are viewing the Galactic Center, any one of these features could be dominant.
How are we going to distinguish these different scenarios?
We have to measure the line-of-sight depth.
We have had access to position-velocity for a long time, but we still argue about position
Dani Lipman, Robin Tress, Mattia Sormani, et al — inner-Galaxy simulationl –v (right)
Observationally, we have access to position on the sky and the velocity on the sky.
We don't have access to the time domain, and we still don't have access to the third dimension.
This simulation color-codes where the gas is in three dimensions.
The top left shows the top-down view.
Red shows front; the blue is behind.
And you can see where each component lands on the sky.
We'd like to invert these PV diagrams to line-of-sight positions, but we can't simply do that.
The ambiguities are just too great.
Progress: the 3D CMZ (SMA's CMZOOM), 2025–26
~3/4 of known clouds are in front of Sgr A*, including the 20 and 50 km s−1 clouds
For most clouds, the best we can say is "in front"
Most gas is unconstrained (not in "clouds")
Nevertheless, we have made progress.
We can tell that about 3/4 of the identified clouds are on the front half.
In the 3D CMZ program, led by Cara Battersby building on SMA's CMZoom, we mapped 3D locations
This paper by Dani Lipman summarizes the progress: the millimeter-selected clouds can roughly be assigned LOS positions.
For most clouds, with the error bars we have on them, the most we can say is they're in front.
The material that isn't in identifiable clouds remains unconstrained.
Cloud Tomography with Stellar Proper Motions
Stellar orbital motion is measurable w/JWST + Roman
Clouds block and redden stars behind and within
The measurable stellar density with Roman is enough to achieve ~10 pc line-of-sight resolution
This idea comes from Arianna Vasini, who is a postdoc working with Mattia Sormani: as long as we can measure the stellar orbital motion, we can figure out the position of stars near the Galactic Center.
Focus on this plot in the bottom right for just a moment.
The proper motion of stars on typical orbits slowly increases from zero kiloparsecs, next to the Sun, up to eight kiloparsecs.
And then, right at that eight-kiloparsec point, there is a sharp turnaround, where, in the central hundred parsecs, the proper motion goes from left to right.
That's shown in the animated diagram.
All the stars we see are moving slowly on the sky.
Those in front are moving left, those behind are moving right.
If a cloud sits in the Galactic Center, it preferentially blocks the stars moving right, and not block the ones that are in front, moving left.
By counting the relative numbers of those, and ideally measuring the extinction toward ones in front of and behind a given cloud,
we're able to do the same kind of tomography that was used in local clouds to make that beautiful top-down view that I opened with.
From Arianna's modeling work, the prediction is that with Roman, we should have enough stars to get roughly 10-parsec line-of-sight resolution, 10-parsec three-dimensional resolution.
We can already start to do some proper motion work with just JWST, but Roman is really going to open it up.
Same live page as the next slide, just the face-on panel — no controls, nothing to click
Orange / teal is proper motion: the nuclear stellar disk rotates, so near-side and far-side stars move in opposite directions on the sky. ~5 mas/yr apart, against ~1 mas/yr scatter
The purple ellipse is a cloud; the stars behind it fade. Measure how much each star is dimmed and you locate the cloud along the line of sight
The cloud can be dragged live: move it behind the Galactic Center and fewer stars fade
Next slide is the same model with the tomography diagnostics attached
Cloud location inversion: ~10 pc resolution along the line of sight
~104 stars → ~10 pc accuracy.−2 at mK <19
gives ~10–20 pc
Q1 What planets form there? 2016 future! Q2 How does gas get into the CMZ? 1991 progress Q3 (How) does gas reach Sgr A*? 1996 future Q4 What shape is the CMZ? 1995 progress Q5 Why does the CMZ form so few stars? 2013 ← Q6 Where do the CMZ’s young stars go? Do clusters dissolve? 2002 Q7 Is the IMF different here? 1999 Q8 Is the chemistry different? What is the metallicity? 2009
Dense gas models predict more SF than we see
And the question of star formation is fundamentally that dense gas models predict more star formation than we see. So basically what you should be looking at is: this gray zone is the star formation we measure, and the hatched region is what you would predict from many different star formation relations. So there's the density threshold law from Charlie Lada, or this constant efficiency per free-fall time at a fixed density from Krumholz, Dekel, and McKee. So this is the story: the Galactic Center produces less star formation than we would expect.
The CMZ
$\sim5\times10^7$ M$_\odot$ of gas, SFR ~0.1 $M_\odot~\mathrm{yr}^{-1}$
$>\frac{1}{3}$ of CMZ SF is in bound clusters (3-8$\times$ local)
$\sim$50% of CMZ SF occurs in the Sgr B2 cloud.
We're going to zoom in ~300× to measure star formation by counting young stars.
The CMZ contains about 108 solar masses of gas.
Most of it is very dense compared to local.
It has a star formation rate of about 0.1 solar masses per year.
About a third of it is in bound clusters, which is a much higher value, about three to eight times that of local.
The Sagittarius B2 cloud is responsible for about half the star formation in the entire Central Molecular Zone.
I'm going to zoom in to that now by a factor of 300, to show you how we measure the star formation rate by counting these YSOs.
Zoom-in to the Sgr B2 cloud:
Most massive, actively star-forming cloud in the galaxy: 107 M⊙
ALMA 3mm image showing mix of dust and free-free emission
Where the ~50% comes from, if anyone asks:
Barnes+ 2017 (MNRAS, "Star formation rates and efficiencies in the Galactic Center"):
CMZ total SFR = 0.09 ± 0.02 M⊙ /yr averaged over IR-luminosity, YSO-counting
and free-free methods; their Sgr B2 entry is 0.005 M⊙ /yr from the IR, rising to
0.045 when the embedded stellar mass is taken from VLA+WMAP instead — i.e. half the CMZ total.
Hatchfield+ 2024 (CMZoom IV): "we find a total SFR (excluding the CND) of ~0.05 M⊙ /yr,
the majority of which is due to the estimate of the SFR of Sgr B2 (~0.036 M⊙ /yr)" —
that is ~70%, so 50% is the conservative statement.
Ginsburg+ 2018b ("Distributed star formation throughout the Galactic Center cloud Sgr B2")
is where the 0.036 M⊙ /yr itself comes from — it is the number Hatchfield adopts.
So the two independent totals (0.09 and 0.05) bracket the answer and both put Sgr B2 at half or more.
The caveat is that Sgr B2's SFR depends on how you count embedded stellar mass; the IR-only
route gives ~10x less.
Dense-gas mass measurements (down-slides 45.1–45.10, below the PPVII slide):
Total: 2–6 × 107 M⊙ within R ≈ 300 pc (Dahmen+ 1998,
Ferrière+ 2007, Henshaw+ 2023); Longmore+ 2013 4.1 × 107 ; Battersby+ 2025a 2.2 × 107
(3.9 without background subtraction). ~108 only from the standard XCO (which Dahmen+ say
overestimates H2 10×), from Launhardt+ 2002's ISM incl. He out to |l| ~ 4°, or from a 7° × 2°
box without background subtraction. So the slide's “~108 ” is high; “a few × 107 ”
is the defensible number.
Dense by threshold (AV > 8, or n > 104 ): ~90–100% of the gas, and SF is
~10× below Lada / Gao–Solomon (Longmore+ 2013, Kauffmann+ 2017, Henshaw+ 2023).
Dense by structure: N > 7 × 1022 32% (Mills & Battersby 2017); SMA-compact ≲10%
in most regions (CMZoom); gravitationally bound <1% (Lu+ 2019). Where compact, bound gas exists, SF looks
closer to normal (Lu+ 2019: ~1 Myr consumption time; Battersby+ 2020, 2025b).
So “10× too few stars” holds for solar-neighborhood definitions of dense; the shortfall is in
turning that gas into compact, bound structure.
The Central Molecular Zone of the Galaxy represents one extreme of star forming conditions in the Galaxy
Total CMZ molecular gas: 2–6 × 107 M☉
Measurement Tracer Region M (107 M☉ ) Dahmen+ 1998 C18 O + dust central ~600 pc 3 (+2/−1) Tsuboi+ 1999 CS 1–0, LTE 160′ × 40′ 3–8 Pierce-Price+ 2000 SCUBA 850 μm inner 400 pc 5.3 ± 1.0 Launhardt+ 2002 IRAS/COBE dust central ~1 kpc H: 6; ISM: ~10 Ferrière+ 2007 review −1.5° < l < 3.5° 2–6 Molinari+ 2011 Hi-GAL dust 100-pc ring only ~3 Longmore+ 2013 Hi-GAL dust −2.5° < l < 3.5° 4.1 Henshaw+ 2023 review R ≈ 300 pc 2–6 Battersby+ 2025a Herschel SED −1.3° < l < 1.8° 2.2 (3.9*) Launhardt ISM includes He and metals; * without fore/background subtraction
Battersby+ 2025a Fig. 4: mass per degree of l (inner 40°, 10°, 3°); dark blue: fore/background subtracted, green: not
Dense = nearly all of it, in recent usage: mean n ≳ 104 cm−3 , f(n > 104 ) ≈ 0.95 (Henshaw+ 2023)
Mass: ~108 only with the standard XCO (Dahmen+ 1998: it “overestimates the H2 column density by an order of magnitude” ), with He out to |l| ≈ 4° (Launhardt+ 2002), or with no background subtraction over 7° × 2° (Battersby+ 2025a: 7.8 × 107 )
SF link: these measure the reservoir. Multiplied by a dense-gas SF law, it predicts ~10× the observed SFR (next panes)
Total molecular gas in the CMZ: every modern optically thin or dust measurement gives 2-6e7 Msun within R ~ 300 pc.
Henshaw+ 2023 (PPVII): "The total mass of molecular gas within R ~ 300 pc is estimated to be 2-6e7 Msun (Dahmen 1998,
Ferriere 2007)". Battersby+ 2025a adopt 2.2e7 (3.9e7 without fore/background subtraction).
The “10× too few stars” depends on what “dense” means
“Dense” = Who Dense share of CMZ gas SF per unit of that gas AV > 8 mag (N > 7.5 × 1021 ) Longmore+ 2013 ~90% ≥10× below prediction n > 104 cm−3 Lu+ 2019; Henshaw+ 2023 95–100% ~10× below Lada+ 2010/2012 cloud mass at AV ≳ 100 mag Kauffmann+ 2017 (cloud by cloud) ~10× below Lada+ 2010 N > 7 × 1022 Mills & Battersby 2017 32% 80% of HCN comes from gas below it SMA-compact (~0.1 pc) in N ≥ 1023 Battersby+ 2020 ≲10% (most regions) “as expected” where compact gravitationally bound cores Lu+ 2019 <1% (Sgr C 9%, Sgr B2 3%) ~1 Myr depletion time in them whole CMZ vs. its clouds Battersby+ 2025b — CMZ 10× low; clouds closer
Threshold definitions (AV ~ 8, n ~ 104 ) count nearly all CMZ gas as dense, and the CMZ falls ~10× below dense-gas SF relations.
Structure definitions (compact, bound) find little dense gas; where it exists, star formation looks closer to normal (Lu+ 2019; Battersby+ 2020, 2025b).
Usage: in recent papers, “dense gas mass of the CMZ” often means the total (Henshaw+ 2023: 2–6 × 107 M☉ “at high density”; Battersby+ 2025a).
The deficit depends on the definition. Solar-neighborhood thresholds (Lada's A_V ~ 7-8, or n ~ 1e4) make ~90-100% of
CMZ gas dense, and the CMZ sits ~10x below Lada/Gao-Solomon (Longmore 13, Kauffmann 17, Henshaw 23, Battersby 25b
"CMZ overall"). Higher or structure-based definitions (N > 7e22: 32%; SMA-compact: <~10%; bound cores: <1%) find
little dense gas, and where it exists SF looks closer to normal: Lu+ 19 find SFR ~ 1e-6/yr x M_bound (a ~1 Myr
consumption time); Battersby+ 20 "Where these compact substructures do form, star formation seems to proceed as
expected"; Battersby+ 25b "the individual clouds in the CMZ show better agreement".
Longmore+ 2013: ~90% of CMZ gas is “dense”; the SFR is 10× low
Longmore+ 2013 Fig. 4b: Hi-GAL mass above a column-density threshold, |l| < 1°, |b| < 0.5°; vertical line: AV = 8
Dense = AV > 8 mag, N(H2 ) > 7.5 × 1021 cm−2 (the Lada+ 2012 threshold); also n ≳ 104 cm−3
Mass: 4.1 × 107 M☉ (Hi-GAL dust, −2.5° < l < 3.5°, |b| ≤ 0.5°); 3.7 × 107 of it above AV = 8. Hi-GAL masses may be low by 2–3× (vs. DustEM)
SF link: the origin of the “10× deficit”: “The measured SFRs are at least a factor 10 smaller than the predicted values” from column-density-threshold and volumetric relations (Krumholz–McKee: 20×)
Longmore+ 2013: "The total derived mass of the CMZ gas from the HiGAL data is 4.1e7 Msun". Table 2 by region
(total / A_V >= 8, 1e7 Msun): -2.5
Kauffmann+ 2017: star formation in CMZ dense gas is suppressed ~10×
Kauffmann+ 2017b Fig. 1: SFR vs. mass at AV > 7: Lada+ 2010 clouds and Gao & Solomon galaxies (yellow), CMZ clouds (red), CMZ averages from Longmore+ 2013 (green)
Dense = for the SF law, Lada+ 2010's AV ≳ 7 mag, evaluated with cloud masses at AV ≳ 100 mag. For structure, “all material detected by the interferometer” (SMA, ~0.1 pc), an approximate measure
Mass: individual clouds (GCMS Paper I); no CMZ total. Dense gas fractions “by factors of 2 to 10 below” Orion A
SF link: “star formation (SF) in dense gas of the CMZ is suppressed by a factor ~10” . Proposed causes: shallow density profiles (little gas at ~0.1 pc) and high SF density thresholds, ~107 –108 cm−3
Kauffmann+ 2017b (GCMS II): "We stress that here we can only present approximate measures of the dense gas fraction
... we adopt the working definition that all material detected by the interferometer resides in dense structures".
SF comparison: Lada 2010 reference rate 4.6e-8/yr x M(A_V > 7); "Here we use the total cloud masses from Table 1 in
Paper I, representing masses at A_V >~ 100 mag". (My reading, not theirs: M(A_V > 100) <= M(A_V > 7), so this is
if anything conservative.) Cloud masses in Paper I not re-checked.
Mills & Battersby 2017: most HCN comes from gas below the “dense” threshold
Mills & Battersby 2017 Fig. 1: HCN 1–0 (top) vs. Herschel N(H2 ) (bottom), central ~300 pc
Dense = N(H2 ) > 7 × 1022 cm−2 (Herschel column density), used for the cores of 13 clouds
Mass: no total. “only 32% of the total gas mass” lies above the threshold
SF link: 80% of the HCN 1–0 comes from gas below 7 × 1022 (68% of the mass), so HCN-based dense gas (Gao & Solomon) is mostly lower-column gas here: “the Gao-Solomon relation cannot be entirely a relationship built up from ... individual dense, high-mass clumps”
Mills & Battersby 2017: "only 32% of the total gas mass in the Galactic center as traced by the column density map lies
above the chosen column density threshold of N(H2) = 7e22". "80% of the HCN 1-0 emission comes from gas with column
density less than 7e22, accounting for 68% of the total mass". Figure rendered from the arXiv source (1701.04822).
Lu+ 2019: <1% of the cloud gas is in bound cores
Lu+ 2019 Fig. 6: SFR vs. cloud mass (a) and vs. gravitationally bound mass (b). Black dashed: Lada+ 2010; blue dash-dotted: 10× lower; cyan dashed: fit to the bound masses, 10−6 yr−1
Cloud Cloud mass (104 M☉ ) Bound mass (102 M☉ ) Bound fraction 20 km s−1 32.0 22.5 0.7% 50 km s−1 6.1 1.1 0.2% Brick 8.8 0.7 0.08% Sgr B1-off 13.7 5.4 0.4% Sgr C 2.4 21.0 9% Sgr B2 140 450 3.2%
Dense = two definitions side by side: n ≳ 104 cm−3 (Lada) gives 100% of these clouds; gas in gravitationally bound cores (SMA 1.3 mm) gives <1%
Mass: clouds of 2.4–32 × 104 M☉ (Sgr B2: 1.4 × 106 ); no CMZ total
SF link: vs. cloud mass, four of five clouds sit ~10× below Lada+ 2010; Sgr C is on the line. vs. bound mass, the five follow SFR ≈ 10−6 yr−1 × Mbound , “a gas consumption time of 1 Myr” in bound cores
Lu+ 2019: "The mean H2 densities of these clouds are >~ 1e4 cm^-3, therefore the dense gas fraction as defined in Lada et al.
is 100%" (but see Mills+ 2018: ~85% of the gas below 1e4). "at the cloud scale, except for Sgr B2 and Sgr C, the star
formation is ~10 times less efficient because less than 1% of gas is confined in gravitationally bound regions".
"The five clouds can be fit by a linear function, with a slope of 1e-6 /yr, which indicates a gas consumption time of
1 Myr in these gravitationally bound cores." Bound masses uncertain by ~3x; SFE in bound cores 10-90%.
Table: Lu Table 7 (cloud masses from Kauffmann+ 17a, scaled to 8.1 kpc).
CMZoom (Battersby+ 2020): little of the high-column gas is compact
Battersby+ 2020 Fig. 12: compact dense gas fraction (SMA / single-dish flux) vs. median N(H2 ); CMZ (blue) vs. non-CMZ clouds (red); ~2× uncertainty between the two sets
Dense = “compact dense gas”: SMA 1.3 mm structure (~0.1 pc) within all gas with N(H2 ) ≥ 1023 cm−2 . “a low CDGF does not mean a low fraction of the gas is at high density”
Mass: no CMZ total. “most regions in the CMZoom survey have CDGFs of less than 10%” , below Galactic-disk clouds despite higher column (exceptions: Sgr B2, Sgr C-dense, Dust Ridge cloud c)
SF link: the low SFE “is likely a direct result of the inability of this gas to fragment into compact substructures. Where these compact substructures do form, star formation seems to proceed as expected.”
CMZoom I: "a low CDGF does not mean a low fraction of the gas is at high density, but rather, that a small fraction of the
total gas is clumped up into compact substructures". Typical densities 1e4-1e7. No total mass; the intro cites
3e7 Msun for the CMZ. My own sum of their Table 4 (not a number in the paper): SMA 1.2e5 / Herschel 4.9e6 Msun = 2.4%
over 36 masks, 1.2% without Sgr B2 and the CND.
CMZoom catalog: how much star formation is possible?
Hatchfield+ 2020 Fig. 4: CMZoom compact-source mass per pc2 surveyed vs. longitude (top), over the Herschel N(H2 ) map
Dense = N(H2 ) ≥ 1023 cm−2 . The fraction of pixels hosting a compact source “appears to sharply increase around” 1–2 × 1023
Mass: no catalog total. Compact sources (“leaves”) up to ~7.6 × 104 M☉ , median ~86 M☉
SF link: if stars form only above 1023 : “maximum star formation potential of the CMZ is between 0.08–2.20 M☉ yr−1 , or, excluding Sgr B2, between 0.04–0.47” . CMZoom IV: incipient SFR ~0.08. No explicit dense-gas-law comparison
CMZoom II (Hatchfield+ 2020): dense = "all high column density gas (N(H2) >= 1e23)". Max SF potential assumes "stars are
not formed ... in gas with a column density lower than 1e23". CMZoom IV (Hatchfield+ 2024): "we estimate an incipient
SFR for the entire CMZ of ~0.08 Msun/yr", upper limit ~0.45; the volume-density-threshold test: "Neither of these
situations appears to be the case for the CMZoom catalog objects."
PPVII (Henshaw+ 2023): SFR ~10× below the dense-gas relation
Henshaw+ 2023: ΣSFR vs. Σgas (left) and SFR vs. dense-gas mass (right); the Milky Way CMZ in blue
Dense = n > 104 cm−3 : a fraction ≈ 0.95 of CMZ gas (from Longmore+ 2013b). Overdense substructure detectable with the SMA: ≲10% in most clouds
Mass: 2–6 × 107 M☉ within R ≈ 300 pc (Dahmen+ 1998; Ferrière+ 2007); ≈1.8 × 107 M☉ of high-column gas within |l| ≤ 1°, |b| ≤ 0.5°
SF link: the SFR “is an order of magnitude below that expected for the available reservoir of dense molecular gas” ; right panel: CMZ vs. Lada+ 2012
PPVII (Henshaw+ 2023): "The total mass of molecular gas within R ~ 300 pc is estimated to be 2-6e7 Msun"; "The CMZ contains
~2-6e7 Msun of gas at high density"; "bulk of the high column density material (total mass of ~1.8e7 Msun) is
contained within the inner |l| <= 1.0, |b| <= 0.5". Table: DGMF f(n > 1e4) = 0.95 (Longmore 2013b). "most clouds
have <~10% of their gas in overdense substructures detectable with the SMA".
Battersby+ 2025a,b: the CMZ is 10× low; its clouds, less so
Paper I Fig. 4: mass per degree of l; dark blue: background-subtracted, green: not
Paper II Fig. 8 (right): SFR vs. dense-gas mass; cyan circles: CMZ clouds; bold circle with error bars: CMZ overall
Box (Paper I Table 1) bg-subtracted (107 M☉ ) not subtracted inner 7° × 2° 2.8 7.8 Dahmen+ 1998 box 2.3 4.1 Henshaw+ 2023 box 2.0 3.1 inner 1.05° × 1.05° 0.6 1.0 adopted: −1.3° < l < 1.8°, |b| < 0.75°2.2 3.9
Dense = “total dense gas (N(H2 ) > 1023 cm−2 )” in the abstract, >1022 in Sec. 5.3; in practice the integrated background-subtracted dust mass, i.e. the total
Mass: 2 (+2/−1) × 107 M☉ ; 3.9 × 107 without fore/background subtraction; systematics ~2×
SF link: “While the CMZ overall ... falls an order of magnitude below the dense gas star formation relation, the individual clouds in the CMZ show better agreement.” Much of the dense gas “is simply not engaged in the star formation process”
Battersby+ 2025a: "Using the extent defined in this work of l = -1.3 to 1.8 deg, b = +-0.75 deg, we find a total dense gas
mass of the CMZ of 2.2e7 Msun (3.9e7 Msun without fore/background subtraction), which we adopt as our standard value."
Dust: OH94, kappa0 = 4.0 cm^2/g at 505 GHz, beta = 1.75, gas-to-dust 100, mu = 2.8 (He included). The paper states
both N > 1e23 (abstract) and N > 1e22 (Sec. 5.3) as "dense", and 8.2 vs 8.1 kpc. Paper II: structure "Entire CMZ"
1.3e7 Msun, SFR 0.03 Msun/yr; "much of the CMZ's dense gas ... is simply not engaged in the star formation process".
Sgr B2 with JWST/
NIRCam
MIRI
on MEERKAT: Ionized gas shows HII regions (OB stars)
Protoclusters
This is my former student's work, Nazar Budaiev.
This is the JWST NIRCam view, of the cloud.
If we switch to MIRI in the mid-infrared, we can see through more of the dust.
We see HII regions…
…and we see these three protoclusters.
These are analogs to the massive clusters I showed earlier.
mostly continuum-subtracted view of Sgr B2: blue is paA, red is f480m
JWST detects some, not all, of the free-free emission
NIRCam tricolor image: red is 4.8 μm excess (F480M – [F410M – F405N]), green is continuum-subtracted Brα ([F405N – F410M]), and blue is continuum-subtracted Paα ([F187N – F182M]).
MIRI sees deeper into the cloud
MIRI 25 micron shows the outflow:
first IR light from within Sgr B2 N;
$\dot{M} \sim 10^{-1} \mathrm{~M}_\odot \mathrm{~yr}^{-1}$, $M_{cl} \sim 3\times10^3 \mathrm{M}_\odot$
…the 25 micron light is escaping from this cluster, even though it has column densities of 1025 particles per square centimeter, or AV of over 1,000 in the center.
This is the most likely site for truly massive star formation, up to $M_{max}$, today.
The central region is accreting at a tenth of a solar mass per year, and there's already a cluster of about 3,000 solar masses, but this one is still growing.
The end state is likely to be 104 solar masses, but we're catching it in this very early state.
It's the only place in the galaxy we do that.
And what we see in those red things is the outflow from the cluster.
It's a whole cluster producing the outflow, not just one star.
In regions like this, the whole IMF is populated, but we are limited to inferring the total star formation from an IMF-based extrapolation.
Zooming in, this is the most massive cluster-forming region in the galaxy
It's accreting at 0.1 msun/yr - fast enough to assemble a massive open cluster in < 1 Myr
ALMA shows the ultra-dense gas, JWST shows the outflow cavities
Budaiev+ in prep
Massive clusters account for 30-40% of the star formation
A dark, dusty swath hides ongoing unclustered SF
◯ JWST 4.8μm excess sources (Budaiev+ 2026)
■ ALMA 3 mm continuum sources (Ginsburg+ 2018, Budaiev+ 2024, Daley+ in prep)
Those are the clusters.
Zooming out...
Clusters are forming here, and they account for about a third of the total star formation.
We switch now to the southern region.
We can see where more distributed star formation is happening.
The yellow dots are ALMA sources, embedded young stars, and the blue are JWST infrared-detected YSOs.
Zooming back out...
we switch to some ongoing, only moderately high-mass star formation
Click: JWST 4.8 μm excess sources (cyan) and ALMA 3 mm sources (yellow) in the dark swath
Budaiev+ in prep
Sgr B2 Deep South — extended star formation with JWST
Budaiev+ in prep
Sgr B2 Deep South — mostly seen by ALMA
NIRCam starless F480M (hot dust &/or CO+H2 ) Brα Paα
ALMA 1.3 mm continuum
contours: ALMA SiO 5–4 12–40 / 54–88 km s−1
Red is hot dust
It's associated with outflows, but doesn't simply track them as you might expect for H2 emission
Budaiev+ in prep
Sgr B2 Deep South — Bipolar flow; 5$\mu$m from the cavity?
NIRCam starless F480M (hot dust &/or CO+H2 ) Brα Paα
ALMA 1.3 mm continuum
contours: ALMA SiO 5–4 12–40 / 54–88 km s−1
Budaiev+ in prep
Sgr B2 Deep South — DS9, a hot core containing a massive star, along a thin filament
NIRCam starless F480M (hot dust &/or CO+H2 ) Brα Paα
ALMA 1.3 mm continuum
contours: ALMA SiO 5–4 12–40 / 54–88 km s−1
ALMA and JWST see different parts of the YSO
ALMA sees younger (gas-richer) sources
So ALMA is showing you the millimeter continuum.
It's where the dense gas and dust is.
It's on this narrow ridge.
On top, you see something that's not in the same direction.
In this case, the JWST 4.8 micron emission is coming from outflows.
ALMA detects the whole core in emission, the disk plus the core.
The four plus micron light seen with JWST NIRCam and MIRI bands is blocked out from the central star.
The envelope completely absorbs it.
But there are outflow cavities, and some of the light either escapes through the outflow cavity, or escapes into the cavity and is scattered into the line of sight through the optically thinner zones.
The short version is: ALMA and Webb are seeing different parts of young stellar objects.
And by selection, then, ALMA is seeing the younger and gas-richer sources, which Webb generally misses.
Zooming back out…
The puffy red stuff, only at 25μm, is still a mystery
JWST finds the extended, older star formation
ALMA finds the embedded, ongoing star formation
◯ JWST 4.8μm excess sources (Budaiev+ 2026)
■ ALMA 3 mm continuum sources (Ginsburg+ 2018, Budaiev+ 2024, Daley+ in prep)
━ JWST 4.8μm excess source density
There's some puffy 25 micron stuff.
This is one of the open mysteries.
We don't know what this is.
Zooming further out...
We show the extended, slightly older star formation population on the whole cloud.
We add the ALMA sources showing the embedded, ongoing star formation.
There is a long line of YSOs.
I'm going to switch the Webb part to contours.
You can see there's an overdensity near the Sagittarius B2 M cluster.
There is asymmetry in the star formation seen in JWST (HII regions, dust excess) and ALMA (embedded YSOs)
The transition is sharp
Star formation is ongoing
Sgr B2 asymmetry is compression/collision
?: Why does collision here end with star formation?
There's a very sharp transition from recent star formation, stars with infrared excess but minimal envelopes, to very active star formation on a very narrow ridge.
Star formation is ongoing in Sagittarius B2, but it has a sharp asymmetry that hints that there is compression and collision going on here.
This leaves open the question: why are colliding clouds ending with massive amounts of star formation, here, while in the Galactic bar those collisions did not produce star formation?
We don't have an answer to that.
We have a census of where ongoing star formation is in Sgr B2 from ALMA - it is at least as much as the rest of the CMZ
It shows a remarkable linear structure, which strongly hints at a dynamical driver, such as stream collision from inflow along the bar
The slide's "is compression/collision" is shorthand: say "hints at"
Gas & young stars can decouple
(a) A cluster forms, perhaps where the bar hits the CMZ
(b) Feedback opens a hole in cloud, doesn't destroy it
(c) Gas feels pressure, follows ~circular orbits; stars drift away
(d) stars may drift back in - feedback from non-birth cloud
What the inflow does once it reaches the CMZ, from Robin Tress's Rhea-RT simulations.
(a) Gas arrives along the shock lane and hits the ring, where a cloud is forming stars.
(b) Click: cloud collisions and the short (5-10 Myr) orbits decouple the stars from the gas,
and the cluster is exposed. (c) Click: gas can only follow nearly closed x2 orbits --
otherwise it collides with itself -- but stars are collisionless and follow open rosette
orbits, so they drift away from their birth gas on an orbital time. (d) Click: and they
run back into dense gas and are re-embedded. So in the CMZ, feedback does not get to
disperse a cloud from the inside the way it does near the Sun.
Sgr B2 is forming stars - is the CMZ?
Henshaw+ 2023; Longmore+ 2013 (Lada+ 2012, Krumholz+ 2012)
Cloud E/F: $10^5\mathrm{~M}_\odot$ gas, 1? YSO
Houghton, Esteve, Gutermuth, Longmore, Huard, Kauffmann subm.
The neighboring cloud, Cloud E/F, has 105 solar masses of gas, like a tenth or less that of Sagittarius B2, but still quite a lot.
In this whole region, there's only one YSO.
Webb has found it.
Webb was sensitive enough to go deep enough to see — it should have seen hundreds or thousands, and instead saw one.
So YSOs are missing.
We start putting data points on the star formation relation plot.
Cloud E/F is well below the extrapolation here.
Sagittarius B2 is actually on it, if you assume the lifetime of those Sagittarius B2 YSOs is a tenth of a million years.
That's where there is room for argument, because previously we assumed it was half a million years, and the rest of the curves make that assumption.
Star formation in the CMZ does not follow the local relation
So the story is that star formation in the Central Molecular Zone doesn't follow the local relation. It is underproducing stars given the amount of dense gas.
Nearby clouds all follow one relation, Sigma_SFR ~ Sigma_gas^2.0, with a uniform efficiency per free-fall time (eps_ff ~ 0.026)
Say the caveat out loud: Pokhrel+ fitted this only below 676 Msun/pc^2 (their N(H2) < 3e22 cut), and their own Fig. 3 shows eps_ff declining above ~1000. Every CMZ point here is 1–2 dex beyond the calibration — pure extrapolation, exactly as I said about Gutermuth in the 2018 paper
The bands are cloud-to-cloud scatter: 0.30 dex, so 2 sigma = 0.60 dex
Where each CMZ point comes from — one cloud-averaged point each, all at t_YSO = 0.1 Myr, and the two limits drawn as generously as the data allow, so the gap is not overstated:
Why 0.1 Myr: it is the Stage 0 (youngest protostar) lifetime, and the short age is the generous choice here — Sigma_SFR = Sigma_* / age, so a shorter age raises every CMZ rate toward the relation. The same age is used for all three points.
Sgr B2 (Ginsburg+ 2018), cloud-averaged: (log Sigma_gas, log Sigma_SFR) = (3.82, 3.19) — ~0.35 dex below the relation (0.23–0.48 across the paper's Total–Totalmax range), inside the 2 sigma band.
1.5e6 Msun over 15 x 15 pc (225 pc^2) → 6.7e3 Msun/pc^2; SFR 0.035–0.062 Msun/yr (Table 2), computed there with a 0.74 Myr age, rescaled by 0.74 / 0.1 to the same 0.1 Myr.
So at 0.1 Myr Sgr B2 as a whole is consistent with the local relation; the Brick and Cloud E/F are not.
Caveats to own: the YSO masses assume 95 Msun per source (21.8 Msun mean x 23% IMF sampling), and the sample misses Class II sources, so the stellar mass may be underestimated by up to 2x.
The Brick (Walker+ 2021), log Sigma_gas = 3.71, limit log Sigma_SFR < 1.72 — ~1.6 dex below the relation, and ~1.5 dex below Sgr B2 at nearly the same gas surface density.
Cloud-scale aperture, to match how Pokhrel measure and so the two limits are comparable: >1e5 Msun within R = 2–3 pc (Sec 1); R = 2.5 pc → 19.6 pc^2 → 5.1e3 Msun/pc^2.
Walker+ quote no SFR and no limit, so the ceiling is ours and deliberately generous: every Table 3 core at its 22 K mass, source 1 counted whole (64.2 Msun) not split, = 104 Msun of core gas , times a core-to-star efficiency of 1.0 — every gram already a star — / 0.1 Myr / 19.6 pc^2 → 53 Msun/pc^2/Myr.
Cross-check: one 5 Msun star (the measured core) extrapolated over a Kroupa IMF to 150 Msun is 117 Msun, within ~10% of the core sum.
These cores really are protostellar: 9 of 18 drive SiO outflows. Using the 19" ALMA field alone instead (7e3 Msun, 1.75 pc^2) moves it to (4.0e3, 594) — same cloud, aperture ten times smaller, and then only ~0.3 dex below.
Cloud E/F (Barnes+ 2019 + Walker+ 2018), log Sigma_gas = 3.78, limit log Sigma_SFR < 1.98 — ~1.5 dex below the relation, ~1.2 dex below Sgr B2.
Sigma_gas from Walker+ 2018 Table 1 , which tabulates the dust ridge on a consistent footing: cloud e = 11.2e4 Msun in R = 2.4 pc, cloud f = 7.3e4 Msun in R = 2.0 pc; Barnes treat them as one cloud, so 1.85e5 Msun over 30.7 pc^2 = 6.0e3 Msun/pc^2. The Immer+ 2012 column of the same table gives 2.6e3, so the horizontal position is good to a factor of ~2.
The ceiling is set by what is observed : no HII regions, no protostellar source catalog, and the only star formation signature in either cloud is one H2O + Class II CH3OH maser pair in the south. A Class II methanol maser marks a high-mass protostar, so allow one star of ≥ 10 Msun and the full Kroupa IMF that comes with it (to 150 Msun): 295 Msun. / 0.1 Myr / 30.7 pc^2 → 96 Msun/pc^2/Myr. (This is now above the Brick's ceiling, because of the full-IMF extrapolation.)
Not drawn any more (single points since 2026-10-19), in case it comes up: the range of column density each cloud actually covers: from the CMZoom threshold N(H2) = 1e23 cm^-2 (Walker+ 2018 Sec 2.1, "designed to target all regions within the CMZ that lie above a column density threshold of ~1e23 cm^-2") = 2.2e3 Msun/pc^2, up to the densest structure in each cloud — the Brick's maser core, 72 Msun in R = 0.04 pc (Rathborne+ 2015 via Walker+ 2021) = 1.4e4 = 6.4e23 cm^-2; and Cloud E/F's peak column, 3.7e24 cm^-2 (Barnes+ 2019 Sec 3.1) = 8.3e4.
Drawn as flat bars, at the cloud-averaged limit, they would be the conservative rendering: a contour-by-contour treatment tilts them upward to the right, since the area shrinks at higher contour — that is why Pokhrel's clouds are curves, not points.
Cross-check on which of the two is denser (Tang+ 2021, MNRAS 505, 2392, Fig. 10):
their AzTEC+Herschel dust-SED fit gives one uniform N(H2) map of the whole CMZ at 7", so the Brick and Cloud E/F are measured the same way .
Reading that map in matched R = 2.4 pc apertures: mean lg N(H2) = 22.90 for the Brick, 23.03 for cloud e and 23.01 for cloud f. Cloud E/F is the denser of the two, by ~0.13 dex (~40%).
That is the same ordering the Walker+ 2018 masses give (5.1e3 vs 6.0e3 Msun/pc^2, 0.07 dex), and by a slightly larger margin — so the separation plotted here is the conservative one. Both clouds saturate Tang's 10^23.4 color scale over a few per cent of their area, so these are means, not peaks.
Two things I got wrong first, in case they come up:
(i) I first took Sigma_gas from Barnes' intro prose, "~1e5 Msun, radii of ~1 pc". Multiplied together those give 3.2e4 Msun/pc^2, five to thirteen times too high, and put E/F to the right of most of Sgr B2. They are order-of-magnitude statements; Barnes never tabulate a cloud-scale mass or radius. Walker+ 2018 Table 1 is the number to use.
(ii) I first set the E/F ceiling from Barnes' "the core region could be up to ~600 Msun ... capable of forming one or several high-mass stars (~25 per cent efficiency)". That 150 Msun is star formation potential , not stars that exist — those cores are starless — and treating it as already formed put the limit inside the Sgr B2 cloud.
Rebuild: python3 colloquium_assets/make_sf_relation_figure.py (it overwrites sf_relation_cmz*.png, which other decks share). Edit CMZ_CLOUDS at the top of that file and rerun; it prints this provenance and the dex-below-relation for every point. Local tracks are digitised from Pokhrel+ Fig. 2a (10 of 12 clouds; Orion-A and Mon R2 are black/gray and cannot be separated from the axes).
A threshold separates Sgr B2 from The Brick
Gray histogram: the column-density distribution of Sgr B2; blue: The Brick, a CMZ cloud with very little star formation but still very high density by local standards (its few protostars sit at the high end)
Dark curve (right axis): the cumulative fraction of Sgr B2's protostellar point sources lying on a background column below N(H2 ). Almost none sit below the dotted line, ~2×1023 cm−2 (~1 g cm−2 )
Only a small fraction of The Brick's area reaches the columns at which Sgr B2 forms its protostars
The difference is consistent with a threshold that The Brick has not crossed and Sgr B2 has
What is different about the CMZ? From the solar neighborhood and cosmic noon?
What is different about the CMZ from the solar neighborhood ?
What is different about the CMZ from cosmic noon ? (blue: the same)
More turbulent , higher density , has inflow , shear , metallicity
We know stars are forming in the CMZ, but we've measured only a small portion
So far, though, those measurements only cover a handful of clouds. It's perhaps possible that we've missed the main site of star formation. That's not very likely at this point, but we want to be able to do the measurements more precisely over the whole Galactic Center.
The JWST Galactic Center Treasury Survey (GO 10678) — NIRCam: half observed
Stellar populations: star formation history
Extinction maps: where is the gas?
YSO candidates
■ observed Sep 2026 (68) ■ planned Aug–Sep 2027 (64) ■ skipped, to be rescheduled (7) status 2026-10-15
This leads us to one of the goals of the Galactic Center Treasury program.
This is a big program to measure stellar populations, make extinction maps, and identify young stellar object candidates.
It is about half done.
The green regions are the parts we got.
The blue regions are planned for next year.
The red ones we have to redo next year, because they failed to acquire guide stars.
We did this in coordinated parallel mode.
Progress:
68 of 139 pointings observed, Sep 11–20, 2026
F212N + F480M
the rest: Aug–Sep 2027
Early highlights
~4 million unique sources so far
astrometric precision $\sigma<1$ mas
proper motions to < 0.3 mas yr−1 w/earlier JWST epochs
□ observed — NIRCam F212N/F480M mosaic of 2026-10-15
JWST GC Treasury · NIRCam F212N + F480M
stars subtracted
NIRCam + ACES 3 mm continuum (red)
MIRI F770W, full frame incl. the 4QPM fields (Budaiev+ in prep)
Status report: we've gotten half the data, and we've pretty much reduced it all now.
We have of order 4 million unique sources so far; we'll be going deeper.
From repeatability experiments, we see the precision is less than a milliarcsecond per star for the top half of the stars.
That starts running out around 18th, 19th magnitude.
Where we have two JWST epochs, we're getting about 0.3 milliarcsecond per year PM accuracy.
So this is great.
It worked beautifully.
I'll show the image data for a few slides now.
This is the image.
It's a composite of the 212 and 480.
If we delete all the stars, you can see the gas being highlighted.
We can mix in the ACES data and see where the gas structures are relative to the stars.
And MIRI: we get a nice big swath.
It's incomplete, but not so bad.
I highlight that we've included the 4QPM fields.
These are the ones that are used for coronagraphy and spectroscopy.
Mostly they're not recommended for science, but my former student current postdoc, Nazar, has been working on recalibrating those and using them.
He found they actually have quite usable data, especially when the emission is as bright as it is in the Galactic Center.
The uncertainties are higher, but not by so much.
So we're very excited that we might actually have better spatial coverage than we had hoped for.
…and the MIRI parallel, ~6′ to the east
F770W: PAH emission, warm dust
□ observed — MIRI F770W, full frame incl. the 4QPM fields mosaic of 2026-10-09
(2026-10-15) The mosaic and the magnified strip are now the FULL-FRAME F770W mosaic, including the
4QPM coronagraph fields (HiPS gc_f770w_fullframe_hips, built 2026-10-09; layers from
colloquium_assets/make_miri_fullframe_layers.py), not the imager-only parallel mosaic -- so the
image runs past the green outlines, which are the imager apertures of the observed pointings.
Q1 What planets form there? 2016 future! Q2 How does gas get into the CMZ? 1991 progress Q3 (How) does gas reach Sgr A*? 1996 future Q4 What shape is the CMZ? 1995 progress Q5 Why does the CMZ form so few stars? 2013 progress Q6 Where do the CMZ’s young stars go? Do clusters dissolve? 2002 ← Q7 Is the IMF different here? 1999 ← Q8 Is the chemistry different? What is the metallicity? 2009
Where do the CMZ’s clusters go?
Arches · F212N / F323N / F480M / F770W
Quintuplet · F212N / F323N / F480M
Future: Arches and Quintuplet tidal streams from JWST GC Treasury + Roman
There are two massive clusters, the Arches and the Quintuplet.
They're analogous to Westerlund 1 and 2.
These are expected to dissolve after only a few tens of millions of years, and basically fade out and be impossible to detect in a few million years.
Now we have two epochs for each of these.
We'll be able to do proper motion measurements, and hopefully identify the tidal streams, showing how they're dissolving in this much different gravitational field.
These are also places where we so far have the best measurements of the IMF being different.
Q1 What planets form there? 2016 future! Q2 How does gas get into the CMZ? 1991 progress Q3 (How) does gas reach Sgr A*? 1996 future Q4 What shape is the CMZ? 1995 progress Q5 Why does the CMZ form so few stars? 2013 progress Q6 Where do the CMZ’s young stars go? Do clusters dissolve? 2002 future Q7 Is the IMF different here? 1999 ← Q8 Is the chemistry different? What is the metallicity? 2009
Question 7: Is the IMF different here? Open since 1999. Where we are: Arches top-heavy, α = 1.80 (HST proper motions). What comes next: Hosek: cleaner membership, deeper into the IMF, more clusters.
Is the IMF different here?
HST: ≥ 1.8 M☉
JWST: ≥ 0.4 M☉
(GO 2045 & 10678)
This is from Matt Hosek's work, showing the range of IMF slopes measured in different clusters.
Focus on the high-mass end: the normal Salpeter IMF is this black dashed line on top, and most regions in the galaxy fall up there.
But in the Arches cluster specifically, the slope is shallower.
The same is true in a handful of other massive clusters.
In the top right, I show that normal IMF as we were showing before.
And then the bottom right, we show what a top-heavy one would look like.
And notably, the critical number here is two.
For steeper than two, most of the mass is in low-mass stars.
For shallower than two, most of the mass is in high-mass stars.
It is possible that the Arches and Quintuplet and these other similar clusters are actually mass-dominated by high-mass stars.
In the work that made this original measurement, HST data were limited to 1.8 solar masses on the low end.
We expect to be able to go to 0.4 solar masses with Webb and constrain the peak of the IMF.
That's future work that JWST has enabled.
Q1 What planets form there? 2016 future! Q2 How does gas get into the CMZ? 1991 progress Q3 (How) does gas reach Sgr A*? 1996 future Q4 What shape is the CMZ? 1995 progress Q5 Why does the CMZ form so few stars? 2013 progress Q6 Where do the CMZ’s young stars go? Do clusters dissolve? 2002 future Q7 Is the IMF different here? 1999 future Q8 Is the chemistry different? What is the metallicity? 2009 ←
Ice measurements suggest higher metallicity
cloud c
the Brick
Ice measurements we've made suggest the metallicity is higher in the Galactic Center.
I'm going to focus on these two clouds that we observed starting in 2023.
This is cloud C, part of the dust ridge.
JWST's colors show ice
JWST's colors show ice
dust
ice
Smith+ 2025: 27 data points
Z
The highlighted cloud is in the foreground of the CMZ, probably somewhere along the bar.
On the x-axis of this plot, we have a normal near-infrared color that is sensitive to dust extinction.
On the y-axis we have 405 minus 466, and the 466 is blocked out by CO ice.
Measurements of stars are plotted.
Any stars that are extincted and reddened by dust should go up and to the right.
Red is up and to the right.
You can see the data do not do that.
The stars that are at higher extinction, as probed by the x-axis, are at bluer colors, as probed by the y-axis.
That is caused by ice.
We use this bluening (ballooning) effect to infer the total amount of ice.
There's more ice in these clouds than we see locally.
The local clouds, from the Chamaeleon region very nearby the Sun, are black data points
These come from spectra; you should trust them much more than you trust our measurements, which come from photometry.
The black dots all follow sort of the classic local relation, that the CO abundance is about 10−4 that of molecular hydrogen.
The 3 kiloparsec arm cloud, the one that's in this foreground here, is significantly higher, at something like 2.5 × 10−4 .
And the two clouds in the Galactic Center (the background is the Brick, and the blue is clouds C and D) seem to be higher still, maybe as much as 5 × 10−4 .
It's at the same column density, so it's not just that we're looking deeper into clouds.
Why is there more CO in the GC and spiral arm clouds?
It is likely that there is more CO there, and that implies higher metallicity by the same amount.
Too much CO ice for solar metallicity
The standard (local) carbon budget
With the measured GC CO-ice columns, the fraction of carbon in CO ice becomes uncomfortably large…
… or even exceeds the total budget
If
ZGC = 2.5 Z☉ , the CO-ice fraction goes back to normal (~10–20%)
(but some lines of sight still have ~double that CO ice…)
From the ice talk. Left: the CO-ice column against extinction for cloud c & D and the 3 kpc arm filament,
with the Smith+ 2025 points. Right: the standard carbon budget, C/H ~ 3e-4, with CO gas at 1e-4 per H2. Click: put the
measured Galactic Center CO-ice column (2.5e-4 per H2) into that budget -- CO ice is an uncomfortably large share. Click:
at the higher measured columns (5e-4) it exceeds the budget. Click: with Galactic Center carbon, C/H ~ 7.5e-4 -- a metallicity
of 2.5 times solar -- the CO-ice fraction is back to normal, ~10-20%, though some lines of sight still have about double that.
The Dust ridge: Cloud A (brick), C, D
All have different ice-to-dust ratios
D
C
A
Gramze+ (subm 2026)
The mystery deepens a little bit from that, though: when we look at different clouds in the Galactic Center, they have different ratios.
I'm showing three clouds: cloud C, cloud D, and cloud A, which has been more colorfully named the Brick.
Cloud D
Ice is everywhere
But it's different between clouds
3 kpc filament
Brick Head
Gramze+ (subm 2026)
Slope changes → metallicity differences.
Both CMZ ↔ bar & within CMZ Chemistry: CO/Water ratio?− / Ammonium salt abundance?
TBD: Ashby+ NIRSPEC 6927
CMZ-like
foreground-like
CMZ-like
foreground-like
Gramze+ (subm 2026)
We interpret those slope changes as metallicity differences.
The plots show slope on the X-axis; the y-axis is essentially a nuisance parameter.
Each of these colored points is a different part of the cloud.
The surprise is that we're seeing them both between the CMZ and the bar, and within the CMZ.
The variations within the CMZ are over only a 100-parsec region.
It is also possible, though, that the difference is chemistry, not metallicity.
Maybe the CO to water ratio, or perhaps the OCN− and ammonium salt abundance, could produce the same effect.
Excitingly, we actually are going to have the answers, because what I showed you is all from photometry, but we also have spectroscopy forthcoming.
Future: ice abundances
JWST GO 6927 · Cycle 4 · NIRSpec (obs Sep 2026)
PI Matt Ashby (SAO) · Cargile , Ginsburg, Hora , Lisse, Melnick , Paladini, Tolls , Werner, Young, Zucker , Kim
Matt Ashby's NIRSpec program targets over a hundred stars in the Brick.
These observations were executed in September of this year, so we've only had a first look.
This is an example of one particular star, and the ice aficionados in the audience will notice water, CO2 , CO, OCN− , and then some other things that we can talk about offline if you'd like.
I highlight the people in the audience who are part of this program in yellow.
Metallicity matters for star formation, chemistry for planets
Gas smash
CMZ SF - yes, still ??
CMZ is icy
3D coming!
question now future Q1 What planets form there? none known Roman microlensing, ALMA disks Q2 How does gas get into the CMZ? inflow measured ACES + matched sims Q3 What gas reaches Sgr A*? hints of inflow Roman + sub-pc sims Q4 What shape is the CMZ? mostly in front Roman tomography Q5 Why so few stars? 1–3 dex low Treasury YSO census Q6 Do clusters dissolve? no tails yet Treasury proper motions Q7 Is the IMF different? Arches top-heavy deeper, more clusters Q8 What is the metallicity? ices: higher? NIRSpec (GO 6927)
Answered: Why is the dense gas warm? (
Ginsburg+ 2016 )
Partly: the extinction law (
Bravo Ferres+ 2025 )
Still open: OB-star census · radio filaments · Sgr A*’s past · HVCCs · nuclear disk · cosmic rays · isolated massive stars · episodic SF
I'll wrap up with this super short summary.
For each of the questions, we have some sense of where we are now, and a strong sense of where we're going in the future.
There's a lot immediately on the horizon.
The GC is the laboratory where cosmologically common conditions are observed at planetary-system forming scales.
I leave you with the final message: the Galactic Center is this laboratory where these cosmologically common conditions are observable at a planet-forming scale.
This movie is going to pan through the new James Webb data, where it's showing in three colors: James Webb 212 and 480, and in red, the ACES 3 millimeter.
While I'm visiting, ask me about
Savannah Gramze+ 2023, in prep
